4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.1.1 and 3.1.2 [xx XXX xxxx]
29 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
31 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
32 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
34 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
35 numeric text form. For gigantic sub-identifiers, this would take a very
36 long time, the time complexity being O(n^2) where n is the size of that
37 sub-identifier. ([CVE-2023-2650])
39 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
40 IDENTIFIER to canonical numeric text form if the size of that OBJECT
41 IDENTIFIER is 586 bytes or less, and fail otherwise.
43 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
44 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
45 most 128 sub-identifiers, and that the maximum value that each sub-
46 identifier may have is 2^32-1 (4294967295 decimal).
48 For each byte of every sub-identifier, only the 7 lower bits are part of
49 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
50 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
55 * Multiple algorithm implementation fixes for ARM BE platforms.
59 * Added a -pedantic option to fipsinstall that adjusts the various
60 settings to ensure strict FIPS compliance rather than backwards
65 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
66 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
67 trigger a crash of an application using AES-XTS decryption if the memory
68 just after the buffer being decrypted is not mapped.
69 Thanks to Anton Romanov (Amazon) for discovering the issue.
74 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
75 The previous fix for this timing side channel turned out to cause
76 a severe 2-3x performance regression in the typical use case
77 compared to 3.0.7. The new fix uses existing constant time
78 code paths, and restores the previous performance level while
79 fully eliminating all existing timing side channels.
80 The fix was developed by Bernd Edlinger with testing support
85 * Add FIPS provider configuration option to disallow the use of
86 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
87 The option '-no_drbg_truncated_digests' can optionally be
88 supplied to 'openssl fipsinstall'.
92 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
93 that it does not enable policy checking. Thanks to David Benjamin for
94 discovering this issue.
99 * Fixed an issue where invalid certificate policies in leaf certificates are
100 silently ignored by OpenSSL and other certificate policy checks are skipped
101 for that certificate. A malicious CA could use this to deliberately assert
102 invalid certificate policies in order to circumvent policy checking on the
103 certificate altogether.
108 * Limited the number of nodes created in a policy tree to mitigate
109 against CVE-2023-0464. The default limit is set to 1000 nodes, which
110 should be sufficient for most installations. If required, the limit
111 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
112 time define to a desired maximum number of nodes or zero to allow
118 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
120 * Add FIPS provider configuration option to enforce the
121 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
122 The option '-ems-check' can optionally be supplied to
123 'openssl fipsinstall'.
127 * The FIPS provider includes a few non-approved algorithms for
128 backward compatibility purposes and the "fips=yes" property query
129 must be used for all algorithm fetches to ensure FIPS compliance.
131 The algorithms that are included but not approved are Triple DES ECB,
132 Triple DES CBC and EdDSA.
136 * Added support for KMAC in KBKDF.
140 * RNDR and RNDRRS support in provider functions to provide
141 random number generation for Arm CPUs (aarch64).
145 * s_client and s_server apps now explicitly say when the TLS version
146 does not include the renegotiation mechanism. This avoids confusion
147 between that scenario versus when the TLS version includes secure
148 renegotiation but the peer lacks support for it.
152 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
154 *Tomasz Kantecki, Andrey Matyukov*
156 * The various OBJ_* functions have been made thread safe.
160 * Parallel dual-prime 1536/2048-bit modular exponentiation for
161 AVX512_IFMA capable processors.
163 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
165 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
166 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
167 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
168 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
169 `OPENSSL_NO_DEPRECATED_3_1`.
171 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
172 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
173 definitions for these functions regardless of whether
174 `OPENSSL_NO_DEPRECATED_3_1` is defined.
176 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
177 functions regardless of whether they are using them. It is recommended that
178 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
182 * When generating safe-prime DH parameters set the recommended private key
183 length equivalent to minimum key lengths as in RFC 7919.
187 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
188 maximum size that is smaller or equal to the digest length to comply with
189 FIPS 186-4 section 5. This is implemented by a new option
190 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
191 `rsa_pss_saltlen` parameter, which is now the default. Signature
192 verification is not affected by this change and continues to work as before.
199 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
200 listed here are only a brief description.
201 The migration guide contains more detailed information related to new features,
202 breaking changes, and mappings for the large list of deprecated functions.
204 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
206 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
208 * Fixed NULL dereference during PKCS7 data verification.
210 A NULL pointer can be dereferenced when signatures are being
211 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
212 algorithm used for the signature is known to the OpenSSL library but
213 the implementation of the hash algorithm is not available the digest
214 initialization will fail. There is a missing check for the return
215 value from the initialization function which later leads to invalid
216 usage of the digest API most likely leading to a crash.
219 PKCS7 data is processed by the SMIME library calls and also by the
220 time stamp (TS) library calls. The TLS implementation in OpenSSL does
221 not call these functions however third party applications would be
222 affected if they call these functions to verify signatures on untrusted
227 * Fixed X.400 address type confusion in X.509 GeneralName.
229 There is a type confusion vulnerability relating to X.400 address processing
230 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
231 but the public structure definition for GENERAL_NAME incorrectly specified
232 the type of the x400Address field as ASN1_TYPE. This field is subsequently
233 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
236 When CRL checking is enabled (i.e. the application sets the
237 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
238 pass arbitrary pointers to a memcmp call, enabling them to read memory
239 contents or enact a denial of service.
244 * Fixed NULL dereference validating DSA public key.
246 An invalid pointer dereference on read can be triggered when an
247 application tries to check a malformed DSA public key by the
248 EVP_PKEY_public_check() function. This will most likely lead
249 to an application crash. This function can be called on public
250 keys supplied from untrusted sources which could allow an attacker
251 to cause a denial of service attack.
253 The TLS implementation in OpenSSL does not call this function
254 but applications might call the function if there are additional
255 security requirements imposed by standards such as FIPS 140-3.
258 *Shane Lontis, Tomáš Mráz*
260 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
262 An invalid pointer dereference on read can be triggered when an
263 application tries to load malformed PKCS7 data with the
264 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
266 The result of the dereference is an application crash which could
267 lead to a denial of service attack. The TLS implementation in OpenSSL
268 does not call this function however third party applications might
269 call these functions on untrusted data.
274 * Fixed Use-after-free following BIO_new_NDEF.
276 The public API function BIO_new_NDEF is a helper function used for
277 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
278 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
279 be called directly by end user applications.
281 The function receives a BIO from the caller, prepends a new BIO_f_asn1
282 filter BIO onto the front of it to form a BIO chain, and then returns
283 the new head of the BIO chain to the caller. Under certain conditions,
284 for example if a CMS recipient public key is invalid, the new filter BIO
285 is freed and the function returns a NULL result indicating a failure.
286 However, in this case, the BIO chain is not properly cleaned up and the
287 BIO passed by the caller still retains internal pointers to the previously
288 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
289 then a use-after-free will occur. This will most likely result in a crash.
292 *Viktor Dukhovni, Matt Caswell*
294 * Fixed Double free after calling PEM_read_bio_ex.
296 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
297 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
298 data. If the function succeeds then the "name_out", "header" and "data"
299 arguments are populated with pointers to buffers containing the relevant
300 decoded data. The caller is responsible for freeing those buffers. It is
301 possible to construct a PEM file that results in 0 bytes of payload data.
302 In this case PEM_read_bio_ex() will return a failure code but will populate
303 the header argument with a pointer to a buffer that has already been freed.
304 If the caller also frees this buffer then a double free will occur. This
305 will most likely lead to a crash.
307 The functions PEM_read_bio() and PEM_read() are simple wrappers around
308 PEM_read_bio_ex() and therefore these functions are also directly affected.
310 These functions are also called indirectly by a number of other OpenSSL
311 functions including PEM_X509_INFO_read_bio_ex() and
312 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
313 internal uses of these functions are not vulnerable because the caller does
314 not free the header argument if PEM_read_bio_ex() returns a failure code.
317 *Kurt Roeckx, Matt Caswell*
319 * Fixed Timing Oracle in RSA Decryption.
321 A timing based side channel exists in the OpenSSL RSA Decryption
322 implementation which could be sufficient to recover a plaintext across
323 a network in a Bleichenbacher style attack. To achieve a successful
324 decryption an attacker would have to be able to send a very large number
325 of trial messages for decryption. The vulnerability affects all RSA padding
326 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
329 *Dmitry Belyavsky, Hubert Kario*
331 * Fixed X.509 Name Constraints Read Buffer Overflow.
333 A read buffer overrun can be triggered in X.509 certificate verification,
334 specifically in name constraint checking. The read buffer overrun might
335 result in a crash which could lead to a denial of service attack.
336 In a TLS client, this can be triggered by connecting to a malicious
337 server. In a TLS server, this can be triggered if the server requests
338 client authentication and a malicious client connects.
343 * Fixed X.509 Policy Constraints Double Locking security issue.
345 If an X.509 certificate contains a malformed policy constraint and
346 policy processing is enabled, then a write lock will be taken twice
347 recursively. On some operating systems (most widely: Windows) this
348 results in a denial of service when the affected process hangs. Policy
349 processing being enabled on a publicly facing server is not considered
350 to be a common setup.
355 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
356 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
357 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
358 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
359 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
360 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
361 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
362 for legacy EC and SM2 keys is also changed similarly to honor the
363 equivalent conversion format flag as specified in the underlying
364 `EC_KEY` object being exported to a provider, when this function is
365 called through `EVP_PKEY_export()`.
369 * Fixed a type confusion vulnerability relating to X.400 address processing
370 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
371 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
372 vulnerability may allow an attacker who can provide a certificate chain and
373 CRL (neither of which need have a valid signature) to pass arbitrary pointers
374 to a `memcmp` call, creating a possible read primitive, subject to some
375 constraints. Refer to the advisory for more information. Thanks to David
376 Benjamin for discovering this issue. ([CVE-2023-0286])
378 This issue has been fixed by changing the public header file definition of
379 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
380 possible for any existing application to successfully use the existing
381 definition; however, if any application references the `x400Address` field
382 (e.g. in dead code), note that the type of this field has changed. There is
387 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
389 * Fixed two buffer overflows in punycode decoding functions.
391 A buffer overrun can be triggered in X.509 certificate verification,
392 specifically in name constraint checking. Note that this occurs after
393 certificate chain signature verification and requires either a CA to
394 have signed the malicious certificate or for the application to continue
395 certificate verification despite failure to construct a path to a trusted
398 In a TLS client, this can be triggered by connecting to a malicious
399 server. In a TLS server, this can be triggered if the server requests
400 client authentication and a malicious client connects.
402 An attacker can craft a malicious email address to overflow
403 an arbitrary number of bytes containing the `.` character (decimal 46)
404 on the stack. This buffer overflow could result in a crash (causing a
408 An attacker can craft a malicious email address to overflow four
409 attacker-controlled bytes on the stack. This buffer overflow could
410 result in a crash (causing a denial of service) or potentially remote code
411 execution depending on stack layout for any given platform/compiler.
416 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
417 parameters in OpenSSL code.
418 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
419 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
420 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
421 Using these invalid names may cause algorithms to use slower methods
422 that ignore the CRT parameters.
426 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
431 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
432 data to be signed before signing the certificate.
436 * Added RIPEMD160 to the default provider.
440 * Ensured that the key share group sent or accepted for the key exchange
441 is allowed for the protocol version.
445 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
447 * OpenSSL supports creating a custom cipher via the legacy
448 EVP_CIPHER_meth_new() function and associated function calls. This function
449 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
450 to use the new provider mechanism in order to implement custom ciphers.
452 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
453 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
454 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
455 and decryption initialisation functions). Instead of using the custom cipher
456 directly it incorrectly tries to fetch an equivalent cipher from the
457 available providers. An equivalent cipher is found based on the NID passed to
458 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
459 given cipher. However it is possible for an application to incorrectly pass
460 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
461 is used in this way the OpenSSL encryption/decryption initialisation function
462 will match the NULL cipher as being equivalent and will fetch this from the
463 available providers. This will succeed if the default provider has been
464 loaded (or if a third party provider has been loaded that offers this
465 cipher). Using the NULL cipher means that the plaintext is emitted as the
468 Applications are only affected by this issue if they call
469 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
470 encryption/decryption initialisation function. Applications that only use
471 SSL/TLS are not impacted by this issue.
476 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
481 * Fixed the linux-mips64 Configure target which was missing the
482 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
487 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
492 * Correctly handle a retransmitted ClientHello in DTLS
496 * Fixed detection of ktls support in cross-compile environment on Linux
500 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
505 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
506 report correct results in some cases
510 * Fix UWP builds by defining VirtualLock
514 * For known safe primes use the minimum key length according to RFC 7919.
515 Longer private key sizes unnecessarily raise the cycles needed to compute the
516 shared secret without any increase of the real security. This fixes a
517 regression from 1.1.1 where these shorter keys were generated for the known
522 * Added the loongarch64 target
526 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
527 only passed to the FIPS provider and not to the default or legacy provider.
531 * Fixed reported performance degradation on aarch64. Restored the
532 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
533 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
534 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
535 The new algorithm is still used for 32 bit targets.
539 * Added a missing header for memcmp that caused compilation failure on some
544 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
546 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
547 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
548 This issue makes the RSA implementation with 2048 bit private keys
549 incorrect on such machines and memory corruption will happen during
550 the computation. As a consequence of the memory corruption an attacker
551 may be able to trigger a remote code execution on the machine performing
554 SSL/TLS servers or other servers using 2048 bit RSA private keys running
555 on machines supporting AVX512IFMA instructions of the X86_64 architecture
556 are affected by this issue.
561 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
562 implementation would not encrypt the entirety of the data under some
563 circumstances. This could reveal sixteen bytes of data that was
564 preexisting in the memory that wasn't written. In the special case of
565 "in place" encryption, sixteen bytes of the plaintext would be revealed.
567 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
568 they are both unaffected.
571 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
573 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
575 * In addition to the c_rehash shell command injection identified in
576 CVE-2022-1292, further bugs where the c_rehash script does not
577 properly sanitise shell metacharacters to prevent command injection have been
580 When the CVE-2022-1292 was fixed it was not discovered that there
581 are other places in the script where the file names of certificates
582 being hashed were possibly passed to a command executed through the shell.
584 This script is distributed by some operating systems in a manner where
585 it is automatically executed. On such operating systems, an attacker
586 could execute arbitrary commands with the privileges of the script.
588 Use of the c_rehash script is considered obsolete and should be replaced
589 by the OpenSSL rehash command line tool.
592 *Daniel Fiala, Tomáš Mráz*
594 * Case insensitive string comparison no longer uses locales. It has instead
595 been directly implemented.
599 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
601 * Case insensitive string comparison is reimplemented via new locale-agnostic
602 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
603 comparison. The previous implementation had problems when the Turkish locale
608 * Fixed a bug in the c_rehash script which was not properly sanitising shell
609 metacharacters to prevent command injection. This script is distributed by
610 some operating systems in a manner where it is automatically executed. On
611 such operating systems, an attacker could execute arbitrary commands with the
612 privileges of the script.
614 Use of the c_rehash script is considered obsolete and should be replaced
615 by the OpenSSL rehash command line tool.
620 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
621 certificate on an OCSP response. The bug caused the function in the case
622 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
623 response (meaning a successful verification) even in the case where the
624 response signing certificate fails to verify.
626 It is anticipated that most users of `OCSP_basic_verify` will not use the
627 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
628 a negative value (indicating a fatal error) in the case of a certificate
629 verification failure. The normal expected return value in this case would be
632 This issue also impacts the command line OpenSSL "ocsp" application. When
633 verifying an ocsp response with the "-no_cert_checks" option the command line
634 application will report that the verification is successful even though it
635 has in fact failed. In this case the incorrect successful response will also
636 be accompanied by error messages showing the failure and contradicting the
637 apparently successful result.
642 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
643 AAD data as the MAC key. This made the MAC key trivially predictable.
645 An attacker could exploit this issue by performing a man-in-the-middle attack
646 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
647 that the modified data would still pass the MAC integrity check.
649 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
650 endpoint will always be rejected by the recipient and the connection will
651 fail at that point. Many application protocols require data to be sent from
652 the client to the server first. Therefore, in such a case, only an OpenSSL
653 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
655 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
656 sent in both directions. In this case both clients and servers could be
657 affected, regardless of the application protocol.
659 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
660 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
661 the handshake when using this ciphersuite.
663 The confidentiality of data is not impacted by this issue, i.e. an attacker
664 cannot decrypt data that has been encrypted using this ciphersuite - they can
667 In order for this attack to work both endpoints must legitimately negotiate
668 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
669 OpenSSL 3.0, and is not available within the default provider or the default
670 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
671 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
672 following must have occurred:
674 1) OpenSSL must have been compiled with the (non-default) compile time option
675 enable-weak-ssl-ciphers
677 2) OpenSSL must have had the legacy provider explicitly loaded (either
678 through application code or via configuration)
680 3) The ciphersuite must have been explicitly added to the ciphersuite list
682 4) The libssl security level must have been set to 0 (default is 1)
684 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
686 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
687 others that both endpoints have in common
692 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
693 occupied by the removed hash table entries.
695 This function is used when decoding certificates or keys. If a long lived
696 process periodically decodes certificates or keys its memory usage will
697 expand without bounds and the process might be terminated by the operating
698 system causing a denial of service. Also traversing the empty hash table
699 entries will take increasingly more time.
701 Typically such long lived processes might be TLS clients or TLS servers
702 configured to accept client certificate authentication.
705 *Hugo Landau, Aliaksei Levin*
707 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
708 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
709 statistics are no longer supported. For compatibility, these statistics are
710 still listed in the output but are now always reported as zero.
714 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
716 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
717 for non-prime moduli.
719 Internally this function is used when parsing certificates that contain
720 elliptic curve public keys in compressed form or explicit elliptic curve
721 parameters with a base point encoded in compressed form.
723 It is possible to trigger the infinite loop by crafting a certificate that
724 has invalid explicit curve parameters.
726 Since certificate parsing happens prior to verification of the certificate
727 signature, any process that parses an externally supplied certificate may thus
728 be subject to a denial of service attack. The infinite loop can also be
729 reached when parsing crafted private keys as they can contain explicit
730 elliptic curve parameters.
732 Thus vulnerable situations include:
734 - TLS clients consuming server certificates
735 - TLS servers consuming client certificates
736 - Hosting providers taking certificates or private keys from customers
737 - Certificate authorities parsing certification requests from subscribers
738 - Anything else which parses ASN.1 elliptic curve parameters
740 Also any other applications that use the BN_mod_sqrt() where the attacker
741 can control the parameter values are vulnerable to this DoS issue.
746 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
747 to the list of ciphersuites providing Perfect Forward Secrecy as
748 required by SECLEVEL >= 3.
750 *Dmitry Belyavskiy, Nicola Tuveri*
752 * Made the AES constant time code for no-asm configurations
753 optional due to the resulting 95% performance degradation.
754 The AES constant time code can be enabled, for no assembly
755 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
759 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
764 * The negative return value handling of the certificate verification callback
765 was reverted. The replacement is to set the verification retry state with
766 the SSL_set_retry_verify() function.
770 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
772 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
773 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
774 verify a certificate supplied by a server. That function may return a
775 negative return value to indicate an internal error (for example out of
776 memory). Such a negative return value is mishandled by OpenSSL and will cause
777 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
778 success and a subsequent call to SSL_get_error() to return the value
779 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
780 returned by OpenSSL if the application has previously called
781 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
782 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
783 totally unexpected and applications may not behave correctly as a result. The
784 exact behaviour will depend on the application but it could result in
785 crashes, infinite loops or other similar incorrect responses.
787 This issue is made more serious in combination with a separate bug in OpenSSL
788 3.0 that will cause X509_verify_cert() to indicate an internal error when
789 processing a certificate chain. This will occur where a certificate does not
790 include the Subject Alternative Name extension but where a Certificate
791 Authority has enforced name constraints. This issue can occur even with valid
797 * Corrected a few file name and file reference bugs in the build,
798 installation and setup scripts, which lead to installation verification
799 failures. Slightly enhanced the installation verification script.
803 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
808 * Fixed PVK encoder to properly query for the passphrase.
812 * Multiple fixes in the OSSL_HTTP API functions.
816 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
817 OSSL_PARAM_INTEGER data type and return error on negative numbers
818 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
819 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
823 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
827 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
831 * Multiple threading fixes.
835 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
839 * Allow fetching an operation from the provider that owns an unexportable key
840 as a fallback if that is still allowed by the property query.
844 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
846 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
851 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
852 S390X capability vector to zero. This simplifies testing of different code
853 paths on S390X architecture.
857 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
858 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
859 SP 800-38D". The communication will fail at this point.
863 * The EC_GROUP_clear_free() function is deprecated as there is nothing
864 confidential in EC_GROUP data.
868 * The byte order mark (BOM) character is ignored if encountered at the
869 beginning of a PEM-formatted file.
873 * Added CMS support for the Russian GOST algorithms.
877 * Due to move of the implementation of cryptographic operations
878 to the providers, validation of various operation parameters can
879 be postponed until the actual operation is executed where previously
880 it happened immediately when an operation parameter was set.
882 For example when setting an unsupported curve with
883 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
884 fail but later keygen operations with the EVP_PKEY_CTX will fail.
886 *OpenSSL team members and many third party contributors*
888 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
889 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
890 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
891 instead to retrieve these algorithms from a provider.
895 * On build targets where the multilib postfix is set in the build
896 configuration the libdir directory was changing based on whether
897 the lib directory with the multilib postfix exists on the system
898 or not. This unpredictable behavior was removed and eventual
899 multilib postfix is now always added to the default libdir. Use
900 `--libdir=lib` to override the libdir if adding the postfix is
905 * The triple DES key wrap functionality now conforms to RFC 3217 but is
906 no longer interoperable with OpenSSL 1.1.1.
910 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
911 function codes, this function can only cause problems for calling
916 * Add a configurable flag to output date formats as ISO 8601. Does not
917 change the default date format.
921 * Version of MSVC earlier than 1300 could get link warnings, which could
922 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
923 Support for this flag has been removed.
927 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
928 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
929 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
930 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
931 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
935 * The signatures of the functions to get and set options on SSL and
936 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
937 Some source code changes may be required.
941 * The public definitions of conf_method_st and conf_st have been
942 deprecated. They will be made opaque in a future release.
944 *Rich Salz and Tomáš Mráz*
946 * Client-initiated renegotiation is disabled by default. To allow it, use
947 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
948 flag, or the "ClientRenegotiation" config parameter as appropriate.
952 * Add "abspath" and "includedir" pragma's to config files, to prevent,
953 or modify relative pathname inclusion.
957 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
958 validated. Please consult the README-FIPS and
959 README-PROVIDERS files, as well as the migration guide.
961 *OpenSSL team members and many third party contributors*
963 * For the key types DH and DHX the allowed settable parameters are now different.
967 * The openssl commands that read keys, certificates, and CRLs now
968 automatically detect the PEM or DER format of the input files.
970 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
972 * Added enhanced PKCS#12 APIs which accept a library context.
976 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
980 * Added support for Kernel TLS (KTLS).
982 *Boris Pismenny, John Baldwin and Andrew Gallatin*
984 * Support for RFC 5746 secure renegotiation is now required by default for
985 SSL or TLS connections to succeed.
989 * The signature of the `copy` functional parameter of the
990 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
991 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
992 the signature of the `pub_decode` functional parameter of the
993 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
994 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
998 * The error return values from some control calls (ctrl) have changed.
1002 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1006 * Many functions in the EVP_ namespace that are getters of values from
1007 implementations or contexts were renamed to include get or get0 in their
1008 names. Old names are provided as macro aliases for compatibility and
1013 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1014 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1015 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1020 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1023 * The output from the command line applications may have minor
1028 * The output from numerous "printing" may have minor changes.
1032 * Windows thread synchronization uses read/write primitives (SRWLock) when
1033 supported by the OS, otherwise CriticalSection continues to be used.
1037 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1038 work on read only BIO source/sinks that do not support these functions.
1039 This allows piping or redirection of a file BIO using stdin to be buffered
1040 into memory. This is used internally in OSSL_DECODER_from_bio().
1044 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1045 this function would return one of the values OSSL_STORE_INFO_NAME,
1046 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1047 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1048 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1049 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1050 using this function should be amended to handle the changed return value.
1054 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1055 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1056 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1057 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1058 contains more than one certificate identifier: This means that all
1059 certificates referenced there MUST be part of the validation chain.
1063 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1064 RC5, DESX and DES have been moved to the legacy provider.
1068 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1069 RIPEMD-160 have been moved to the legacy provider.
1073 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1078 * The deprecated functions EVP_PKEY_get0_RSA(),
1079 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1080 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1081 well as the similarly named "get1" functions behave differently in
1086 * A number of functions handling low-level keys or engines were deprecated
1087 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1088 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1089 EVP_PKEY_get0_siphash().
1093 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1094 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1095 will need to load the legacy crypto provider. This includes these PBE
1096 algorithms which use this KDF:
1097 - NID_pbeWithMD2AndDES_CBC
1098 - NID_pbeWithMD5AndDES_CBC
1099 - NID_pbeWithSHA1AndRC2_CBC
1100 - NID_pbeWithMD2AndRC2_CBC
1101 - NID_pbeWithMD5AndRC2_CBC
1102 - NID_pbeWithSHA1AndDES_CBC
1106 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1107 BIO_debug_callback() functions.
1111 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1112 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1116 * The RAND_METHOD APIs have been deprecated.
1120 * The SRP APIs have been deprecated.
1124 * Add a compile time option to prevent the caching of provider fetched
1125 algorithms. This is enabled by including the no-cached-fetch option
1126 at configuration time.
1130 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1131 count of PKCS12_DEFAULT_ITER.
1133 *Tomáš Mráz and Sahana Prasad*
1135 * The openssl speed command does not use low-level API calls anymore.
1139 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1142 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1144 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1148 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1149 providers may supply their own group implementations (using either the "key
1150 exchange" or the "key encapsulation" methods) which will automatically be
1151 detected and used by libssl.
1153 *Matt Caswell, Nicola Tuveri*
1155 * The undocumented function X509_certificate_type() has been deprecated;
1159 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1163 * Removed RSA padding mode for SSLv23 (which was only used for
1164 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1165 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1170 * Deprecated the obsolete X9.31 RSA key generation related functions.
1172 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1173 is not allowed to return a value > 1, this is no more taken as failure.
1175 *Viktor Dukhovni and David von Oheimb*
1177 * Deprecated the obsolete X9.31 RSA key generation related functions
1178 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1179 BN_X931_generate_prime_ex().
1183 * The default key generation method for the regular 2-prime RSA keys was
1184 changed to the FIPS 186-4 B.3.6 method.
1188 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1192 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1196 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1197 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1199 *Rich Salz, Richard Levitte, and David von Oheimb*
1201 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1205 * Deprecated `OCSP_parse_url()`.
1209 * Validation of SM2 keys has been separated from the validation of regular EC
1214 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1215 switches: a validation failure triggers an early exit, returning a failure
1216 exit status to the parent process.
1220 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1221 to ignore unknown ciphers.
1225 * The `-cipher-commands` and `-digest-commands` options
1226 of the command line utility `list` have been deprecated.
1227 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1231 * Added convenience functions for generating asymmetric key pairs:
1232 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1233 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1237 * All of the low-level EC_KEY functions have been deprecated.
1239 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1241 * Deprecated all the libcrypto and libssl error string loading
1246 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1247 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1252 * The `-crypt` option to the `passwd` command line tool has been removed.
1256 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1261 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1265 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1266 EVP_PKEY_get1_tls_encodedpoint().
1270 * The security callback, which can be customised by application code, supports
1271 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1272 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1276 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1277 interface. Their functionality remains unchanged.
1281 * Added new option for 'openssl list', '-providers', which will display the
1282 list of loaded providers, their names, version and status. It optionally
1283 displays their gettable parameters.
1287 * Removed EVP_PKEY_set_alias_type().
1291 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1292 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1296 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1297 parameters are still checked for type safety at compile time via helper
1302 * Remove the RAND_DRBG API
1304 *Paul Dale and Matthias St. Pierre*
1306 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1307 as well as actual hostnames.
1311 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1312 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1313 conversely, silently ignore DTLS protocol version bounds when configuring
1314 TLS-based contexts. The commands can be repeated to set bounds of both
1315 types. The same applies with the corresponding "min_protocol" and
1316 "max_protocol" command-line switches, in case some application uses both TLS
1319 SSL_CTX instances that are created for a fixed protocol version (e.g.
1320 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1321 attempts to apply bounds to these protocol versions would result in an
1322 error. Now only the "version-flexible" SSL_CTX instances are subject to
1323 limits in configuration files in command-line options.
1327 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1332 * Reworked the recorded ERR codes to make better space for system errors.
1333 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1334 given code is a system error (true) or an OpenSSL error (false).
1338 * Reworked the test perl framework to better allow parallel testing.
1340 *Nicola Tuveri and David von Oheimb*
1342 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1343 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1347 * 'Configure' has been changed to figure out the configuration target if
1348 none is given on the command line. Consequently, the 'config' script is
1349 now only a mere wrapper. All documentation is changed to only mention
1352 *Rich Salz and Richard Levitte*
1354 * Added a library context `OSSL_LIB_CTX` that applications as well as
1355 other libraries can use to form a separate context within which
1356 libcrypto operations are performed.
1360 * Added various `_ex` functions to the OpenSSL API that support using
1361 a non-default `OSSL_LIB_CTX`.
1365 * Handshake now fails if Extended Master Secret extension is dropped
1370 * Dropped interactive mode from the `openssl` program.
1374 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1376 *David von Oheimb and Shane Lontis*
1378 * Deprecated `EC_METHOD_get_field_type()`.
1382 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1383 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1384 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1388 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1392 * Add CAdES-BES signature verification support, mostly derived
1393 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1395 *Filipe Raimundo da Silva*
1397 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1401 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1402 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1406 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1410 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1411 EC_KEY_precompute_mult().
1415 * Deprecated EC_POINTs_mul().
1419 * Removed FIPS_mode() and FIPS_mode_set().
1423 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1427 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1428 EC_POINT_get_Jprojective_coordinates_GFp().
1432 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1433 arrays to be more easily constructed via a series of utility functions.
1434 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1435 the various push functions and finally convert to a passable OSSL_PARAM
1436 array using OSSL_PARAM_BLD_to_param().
1440 * The security strength of SHA1 and MD5 based signatures in TLS has been
1445 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1446 contain a provider side internal key.
1450 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1454 * Project text documents not yet having a proper file name extension
1455 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1456 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1460 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1461 have been converted to Markdown with the goal to produce documents
1462 which not only look pretty when viewed online in the browser, but
1463 remain well readable inside a plain text editor.
1465 To achieve this goal, a 'minimalistic' Markdown style has been applied
1466 which avoids formatting elements that interfere too much with the
1467 reading flow in the text file. For example, it
1469 * avoids [ATX headings][] and uses [setext headings][] instead
1470 (which works for `<h1>` and `<h2>` headings only).
1471 * avoids [inline links][] and uses [reference links][] instead.
1472 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1474 [ATX headings]: https://github.github.com/gfm/#atx-headings
1475 [setext headings]: https://github.github.com/gfm/#setext-headings
1476 [inline links]: https://github.github.com/gfm/#inline-link
1477 [reference links]: https://github.github.com/gfm/#reference-link
1478 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1479 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1481 *Matthias St. Pierre*
1483 * The test suite is changed to preserve results of each test recipe.
1484 A new directory test-runs/ with subdirectories named like the
1485 test recipes are created in the build tree for this purpose.
1489 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1490 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1491 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1493 *David von Oheimb, Martin Peylo*
1495 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1496 It supports arbitrary request and response content types, GET redirection,
1497 TLS, connections via HTTP(S) proxies, connections and exchange via
1498 user-defined BIOs (allowing implicit connections), persistent connections,
1499 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1500 The legacy OCSP-focused (and only partly documented) API
1501 is retained for backward compatibility, while most of it is deprecated.
1505 * Added `util/check-format.pl`, a tool for checking adherence to the
1506 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1507 The checks performed are incomplete and yield some false positives.
1508 Still the tool should be useful for detecting most typical glitches.
1512 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1513 If domain name resolution yields multiple IP addresses all of them are tried
1514 after `connect()` failures.
1518 * All of the low-level RSA functions have been deprecated.
1522 * X509 certificates signed using SHA1 are no longer allowed at security
1527 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1528 modified to use PKEY APIs. These commands are now in maintenance mode
1529 and no new features will be added to them.
1533 * The command line utility rsautl has been deprecated.
1537 * The command line utilities genrsa and rsa have been modified to use PKEY
1538 APIs. They now write PKCS#8 keys by default. These commands are now in
1539 maintenance mode and no new features will be added to them.
1543 * All of the low-level DH functions have been deprecated.
1545 *Paul Dale and Matt Caswell*
1547 * All of the low-level DSA functions have been deprecated.
1551 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1552 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1556 * Deprecated low-level ECDH and ECDSA functions.
1560 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1564 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1565 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1566 a new formulation to include all the things it can be used for,
1567 as well as words of caution.
1571 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1575 * All of the low-level HMAC functions have been deprecated.
1577 *Paul Dale and David von Oheimb*
1579 * Over two thousand fixes were made to the documentation, including:
1580 - Common options (such as -rand/-writerand, TLS version control, etc)
1581 were refactored and point to newly-enhanced descriptions in openssl.pod.
1582 - Added style conformance for all options (with help from Richard Levitte),
1583 documented all reported missing options, added a CI build to check
1584 that all options are documented and that no unimplemented options
1586 - Documented some internals, such as all use of environment variables.
1587 - Addressed all internal broken L<> references.
1591 * All of the low-level CMAC functions have been deprecated.
1595 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1596 functions have been deprecated.
1598 *Paul Dale and David von Oheimb*
1600 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1601 set of functions. The documentation mentioned negative values for some
1602 errors, but this was never the case, so the mention of negative values
1605 Code that followed the documentation and thereby check with something
1606 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1610 * All of the low-level cipher functions have been deprecated.
1612 *Matt Caswell and Paul Dale*
1614 * Removed include/openssl/opensslconf.h.in and replaced it with
1615 include/openssl/configuration.h.in, which differs in not including
1616 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1617 was added to include both.
1619 This allows internal hacks where one might need to modify the set
1620 of configured macros, for example this if deprecated symbols are
1621 still supposed to be available internally:
1623 #include <openssl/configuration.h>
1625 #undef OPENSSL_NO_DEPRECATED
1626 #define OPENSSL_SUPPRESS_DEPRECATED
1628 #include <openssl/macros.h>
1630 This should not be used by applications that use the exported
1631 symbols, as that will lead to linking errors.
1635 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1636 used in exponentiation with 512-bit moduli. No EC algorithms are
1637 affected. Analysis suggests that attacks against 2-prime RSA1024,
1638 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1639 difficult to perform and are not believed likely. Attacks against DH512
1640 are considered just feasible. However, for an attack the target would
1641 have to re-use the DH512 private key, which is not recommended anyway.
1642 Also applications directly using the low-level API BN_mod_exp may be
1643 affected if they use BN_FLG_CONSTTIME.
1648 * Most memory-debug features have been deprecated, and the functionality
1649 replaced with no-ops.
1653 * Added documentation for the STACK API.
1657 * Introduced a new method type and API, OSSL_ENCODER, to represent
1658 generic encoders. These do the same sort of job that PEM writers
1659 and d2i functions do, but with support for methods supplied by
1660 providers, and the possibility for providers to support other
1665 * Introduced a new method type and API, OSSL_DECODER, to represent
1666 generic decoders. These do the same sort of job that PEM readers
1667 and i2d functions do, but with support for methods supplied by
1668 providers, and the possibility for providers to support other
1673 * Added a .pragma directive to the syntax of configuration files, to
1674 allow varying behavior in a supported and predictable manner.
1675 Currently added pragma:
1679 This allows dollar signs to be a keyword character unless it's
1680 followed by a opening brace or parenthesis. This is useful for
1681 platforms where dollar signs are commonly used in names, such as
1682 volume names and system directory names on VMS.
1686 * Added functionality to create an EVP_PKEY from user data.
1690 * Change the interpretation of the '--api' configuration option to
1691 mean that this is a desired API compatibility level with no
1692 further meaning. The previous interpretation, that this would
1693 also mean to remove all deprecated symbols up to and including
1694 the given version, no requires that 'no-deprecated' is also used
1695 in the configuration.
1697 When building applications, the desired API compatibility level
1698 can be set with the OPENSSL_API_COMPAT macro like before. For
1699 API compatibility version below 3.0, the old style numerical
1700 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1701 For version 3.0 and on, the value is expected to be the decimal
1702 value calculated from the major and minor version like this:
1704 MAJOR * 10000 + MINOR * 100
1708 -DOPENSSL_API_COMPAT=30000 For 3.0
1709 -DOPENSSL_API_COMPAT=30200 For 3.2
1711 To hide declarations that are deprecated up to and including the
1712 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1713 given when building the application as well.
1717 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1718 access to certificate and CRL stores via URIs and OSSL_STORE
1721 This adds the following functions:
1723 - X509_LOOKUP_store()
1724 - X509_STORE_load_file()
1725 - X509_STORE_load_path()
1726 - X509_STORE_load_store()
1727 - SSL_add_store_cert_subjects_to_stack()
1728 - SSL_CTX_set_default_verify_store()
1729 - SSL_CTX_load_verify_file()
1730 - SSL_CTX_load_verify_dir()
1731 - SSL_CTX_load_verify_store()
1735 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1736 The presence of this system service is determined at run-time.
1740 * Added functionality to create an EVP_PKEY context based on data
1741 for methods from providers. This takes an algorithm name and a
1742 property query string and simply stores them, with the intent
1743 that any operation that uses this context will use those strings
1744 to fetch the needed methods implicitly, thereby making the port
1745 of application written for pre-3.0 OpenSSL easier.
1749 * The undocumented function NCONF_WIN32() has been deprecated; for
1750 conversion details see the HISTORY section of doc/man5/config.pod
1754 * Introduced the new functions EVP_DigestSignInit_ex() and
1755 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1756 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1757 pages for further details.
1761 * Over two thousand fixes were made to the documentation, including:
1762 adding missing command flags, better style conformance, documentation
1765 *Rich Salz, Richard Levitte*
1767 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1768 X25519, X448, Ed25519 and Ed448.
1772 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1777 * Deprecated the public definition of `ERR_STATE` as well as the function
1778 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1783 * Added ERR functionality to give callers access to the stored function
1784 names that have replaced the older function code based functions.
1786 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1787 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1788 ERR_peek_error_all() and ERR_peek_last_error_all().
1790 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1791 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1792 ERR_func_error_string().
1796 * Extended testing to be verbose for failing tests only. The make variables
1797 VERBOSE_FAILURE or VF can be used to enable this:
1799 $ make VF=1 test # Unix
1800 $ mms /macro=(VF=1) test ! OpenVMS
1801 $ nmake VF=1 test # Windows
1805 * Added the `-copy_extensions` option to the `x509` command for use with
1806 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1807 all extensions in the request are copied to the certificate or vice versa.
1809 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1811 * Added the `-copy_extensions` option to the `req` command for use with
1812 `-x509`. When given with the `copy` or `copyall` argument,
1813 all extensions in the certification request are copied to the certificate.
1817 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1818 they generate are by default RFC 5280 compliant in the following sense:
1819 There is a subjectKeyIdentifier extension with a hash value of the public key
1820 and for not self-signed certs there is an authorityKeyIdentifier extension
1821 with a keyIdentifier field or issuer information identifying the signing key.
1822 This is done unless some configuration overrides the new default behavior,
1823 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1827 * Added several checks to `X509_verify_cert()` according to requirements in
1828 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1829 (which may be done by using the CLI option `-x509_strict`):
1830 * The basicConstraints of CA certificates must be marked critical.
1831 * CA certificates must explicitly include the keyUsage extension.
1832 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1833 * The issuer name of any certificate must not be empty.
1834 * The subject name of CA certs, certs with keyUsage crlSign,
1835 and certs without subjectAlternativeName must not be empty.
1836 * If a subjectAlternativeName extension is given it must not be empty.
1837 * The signatureAlgorithm field and the cert signature must be consistent.
1838 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1839 must not be marked critical.
1840 * The authorityKeyIdentifier must be given for X.509v3 certs
1841 unless they are self-signed.
1842 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1846 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1847 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1851 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1852 used even when parsing explicit parameters, when loading a encoded key
1853 or calling `EC_GROUP_new_from_ecpkparameters()`/
1854 `EC_GROUP_new_from_ecparameters()`.
1855 This prevents bypass of security hardening and performance gains,
1856 especially for curves with specialized EC_METHODs.
1857 By default, if a key encoded with explicit parameters is loaded and later
1858 encoded, the output is still encoded with explicit parameters, even if
1859 internally a "named" EC_GROUP is used for computation.
1863 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1864 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1865 NULL. After this change, only the cofactor parameter can be NULL. It also
1866 does some minimal sanity checks on the passed order.
1871 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1872 An attack is simple, if the first CMS_recipientInfo is valid but the
1873 second CMS_recipientInfo is chosen ciphertext. If the second
1874 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1875 encryption key will be replaced by garbage, and the message cannot be
1876 decoded, but if the RSA decryption fails, the correct encryption key is
1877 used and the recipient will not notice the attack.
1878 As a work around for this potential attack the length of the decrypted
1879 key must be equal to the cipher default key length, in case the
1880 certificate is not given and all recipientInfo are tried out.
1881 The old behaviour can be re-enabled in the CMS code by setting the
1882 CMS_DEBUG_DECRYPT flag.
1886 * Early start up entropy quality from the DEVRANDOM seed source has been
1887 improved for older Linux systems. The RAND subsystem will wait for
1888 /dev/random to be producing output before seeding from /dev/urandom.
1889 The seeded state is stored for future library initialisations using
1890 a system global shared memory segment. The shared memory identifier
1891 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1892 the desired value. The default identifier is 114.
1896 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1897 when primes for RSA keys are computed.
1898 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1899 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1900 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1901 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1902 This avoids possible fingerprinting of newly generated RSA modules.
1906 * Correct the extended master secret constant on EBCDIC systems. Without this
1907 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1908 negotiate EMS will fail. Unfortunately this also means that TLS connections
1909 between EBCDIC systems with this fix, and EBCDIC systems without this
1910 fix will fail if they negotiate EMS.
1914 * Changed the library initialisation so that the config file is now loaded
1915 by default. This was already the case for libssl. It now occurs for both
1916 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1917 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1921 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1922 where the former acts as a replacement for `ERR_put_error()`, and the
1923 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1924 `ERR_raise_data()` adds more flexibility by taking a format string and
1925 an arbitrary number of arguments following it, to be processed with
1930 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1931 to check if a named provider is loaded and available. When called, it
1932 will also activate all fallback providers if such are still present.
1936 * Enforce a minimum DH modulus size of 512 bits.
1940 * Changed DH parameters to generate the order q subgroup instead of 2q.
1941 Previously generated DH parameters are still accepted by DH_check
1942 but DH_generate_key works around that by clearing bit 0 of the
1943 private key for those. This avoids leaking bit 0 of the private key.
1947 * Significantly reduce secure memory usage by the randomness pools.
1951 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1956 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1957 algorithms. An implementation of a key exchange algorithm can be obtained
1958 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1959 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1960 the older EVP_PKEY_derive_init() function. See the man pages for the new
1961 functions for further details.
1965 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1969 * Removed the function names from error messages and deprecated the
1974 * Removed NextStep support and the macro OPENSSL_UNISTD
1978 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1979 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1980 Also removed "export var as function" capability; we do not export
1981 variables, only functions.
1985 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1986 an error and 1 indicating success. In previous versions of OpenSSL this
1987 was a void type. If a key was set longer than the maximum possible this
1992 * Support SM2 signing and verification schemes with X509 certificate.
1996 * Use SHA256 as the default digest for TS query in the `ts` app.
2000 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2004 * Default cipher lists/suites are now available via a function, the
2005 #defines are deprecated.
2009 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2010 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2011 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2015 * Join the directories crypto/x509 and crypto/x509v3
2019 * Added command 'openssl kdf' that uses the EVP_KDF API.
2023 * Added command 'openssl mac' that uses the EVP_MAC API.
2027 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2028 as default directories. Also added the command 'openssl info'
2029 for scripting purposes.
2033 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2038 * Add prediction resistance to the DRBG reseeding process.
2042 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2043 mandated by IEEE Std 1619-2018.
2047 * Added newline escaping functionality to a filename when using openssl dgst.
2048 This output format is to replicate the output format found in the `*sum`
2049 checksum programs. This aims to preserve backward compatibility.
2051 *Matt Eaton, Richard Levitte, and Paul Dale*
2053 * Removed the heartbeat message in DTLS feature, as it has very
2054 little usage and doesn't seem to fulfill a valuable purpose.
2055 The configuration option is now deprecated.
2059 * Changed the output of 'openssl {digestname} < file' to display the
2060 digest name in its output.
2064 * Added a new generic trace API which provides support for enabling
2065 instrumentation through trace output.
2067 *Richard Levitte & Matthias St. Pierre*
2069 * Added build tests for C++. These are generated files that only do one
2070 thing, to include one public OpenSSL head file each. This tests that
2071 the public header files can be usefully included in a C++ application.
2073 This test isn't enabled by default. It can be enabled with the option
2074 'enable-buildtest-c++'.
2078 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2082 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2086 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2090 * Added KMAC to EVP_MAC.
2094 * Added property based algorithm implementation selection framework to
2099 * Added SCA hardening for modular field inversion in EC_GROUP through
2100 a new dedicated field_inv() pointer in EC_METHOD.
2101 This also addresses a leakage affecting conversions from projective
2102 to affine coordinates.
2104 *Billy Bob Brumley, Nicola Tuveri*
2106 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2107 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2108 those algorithms that were already supported through the EVP_PKEY API
2109 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2110 and scrypt are now wrappers that call EVP_KDF.
2114 * Build devcrypto engine as a dynamic engine.
2116 *Eneas U de Queiroz*
2118 * Add keyed BLAKE2 to EVP_MAC.
2122 * Fix a bug in the computation of the endpoint-pair shared secret used
2123 by DTLS over SCTP. This breaks interoperability with older versions
2124 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2125 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2126 interoperability with such broken implementations. However, enabling
2127 this switch breaks interoperability with correct implementations.
2129 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2130 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2134 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2138 * Changed the license to the Apache License v2.0.
2142 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2144 - Major releases (indicated by incrementing the MAJOR release number)
2145 may introduce incompatible API/ABI changes.
2146 - Minor releases (indicated by incrementing the MINOR release number)
2147 may introduce new features but retain API/ABI compatibility.
2148 - Patch releases (indicated by incrementing the PATCH number)
2149 are intended for bug fixes and other improvements of existing
2150 features only (like improving performance or adding documentation)
2151 and retain API/ABI compatibility.
2155 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2159 * Remove the 'dist' target and add a tarball building script. The
2160 'dist' target has fallen out of use, and it shouldn't be
2161 necessary to configure just to create a source distribution.
2165 * Recreate the OS390-Unix config target. It no longer relies on a
2166 special script like it did for OpenSSL pre-1.1.0.
2170 * Instead of having the source directories listed in Configure, add
2171 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2176 * Add GMAC to EVP_MAC.
2180 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2184 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2185 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2186 to facilitate the continued use of MACs through raw private keys in
2187 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2191 * Deprecate ECDH_KDF_X9_62().
2195 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2196 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2197 are retained for backwards compatibility.
2201 * AES-XTS mode now enforces that its two keys are different to mitigate
2202 the attacked described in "Efficient Instantiations of Tweakable
2203 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2204 Details of this attack can be obtained from:
2205 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2209 * Rename the object files, i.e. give them other names than in previous
2210 versions. Their names now include the name of the final product, as
2211 well as its type mnemonic (bin, lib, shlib).
2215 * Added new option for 'openssl list', '-objects', which will display the
2216 list of built in objects, i.e. OIDs with names.
2220 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2221 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2226 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2227 improves application performance by removing data copies and providing
2228 applications with zero-copy system calls such as sendfile and splice.
2232 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2236 * `PKCS12_parse` now maintains the order of the parsed certificates
2237 when outputting them via `*ca` (rather than reversing it).
2241 * Deprecated pthread fork support methods.
2245 * Added support for FFDHE key exchange in TLS 1.3.
2249 * Added a new concept for OpenSSL plugability: providers. This
2250 functionality is designed to replace the ENGINE API and ENGINE
2251 implementations, and to be much more dynamic, allowing provider
2252 authors to introduce new algorithms among other things, as long as
2253 there's an API that supports the algorithm type.
2255 With this concept comes a new core API for interaction between
2256 libcrypto and provider implementations. Public libcrypto functions
2257 that want to use providers do so through this core API.
2259 The main documentation for this core API is found in
2260 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2261 refer to other manuals describing the API specific for supported
2262 algorithm types (also called operations).
2269 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2271 * Avoid loading of a dynamic engine twice.
2275 * Prioritise DANE TLSA issuer certs over peer certs
2279 * Fixed random API for MacOS prior to 10.12
2281 These MacOS versions don't support the CommonCrypto APIs
2285 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2287 * Fixed an SM2 Decryption Buffer Overflow.
2289 In order to decrypt SM2 encrypted data an application is expected to
2290 call the API function EVP_PKEY_decrypt(). Typically an application will
2291 call this function twice. The first time, on entry, the "out" parameter
2292 can be NULL and, on exit, the "outlen" parameter is populated with the
2293 buffer size required to hold the decrypted plaintext. The application
2294 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2295 again, but this time passing a non-NULL value for the "out" parameter.
2297 A bug in the implementation of the SM2 decryption code means that the
2298 calculation of the buffer size required to hold the plaintext returned
2299 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2300 size required by the second call. This can lead to a buffer overflow
2301 when EVP_PKEY_decrypt() is called by the application a second time with
2302 a buffer that is too small.
2304 A malicious attacker who is able present SM2 content for decryption to
2305 an application could cause attacker chosen data to overflow the buffer
2306 by up to a maximum of 62 bytes altering the contents of other data held
2307 after the buffer, possibly changing application behaviour or causing
2308 the application to crash. The location of the buffer is application
2309 dependent but is typically heap allocated.
2314 * Fixed various read buffer overruns processing ASN.1 strings
2316 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2317 structure which contains a buffer holding the string data and a field
2318 holding the buffer length. This contrasts with normal C strings which
2319 are represented as a buffer for the string data which is terminated
2320 with a NUL (0) byte.
2322 Although not a strict requirement, ASN.1 strings that are parsed using
2323 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2324 well as any string whose value has been set with the ASN1_STRING_set()
2325 function will additionally NUL terminate the byte array in the
2326 ASN1_STRING structure.
2328 However, it is possible for applications to directly construct valid
2329 ASN1_STRING structures which do not NUL terminate the byte array by
2330 directly setting the "data" and "length" fields in the ASN1_STRING
2331 array. This can also happen by using the ASN1_STRING_set0() function.
2333 Numerous OpenSSL functions that print ASN.1 data have been found to
2334 assume that the ASN1_STRING byte array will be NUL terminated, even
2335 though this is not guaranteed for strings that have been directly
2336 constructed. Where an application requests an ASN.1 structure to be
2337 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2338 been directly constructed by the application without NUL terminating
2339 the "data" field, then a read buffer overrun can occur.
2341 The same thing can also occur during name constraints processing
2342 of certificates (for example if a certificate has been directly
2343 constructed by the application instead of loading it via the OpenSSL
2344 parsing functions, and the certificate contains non NUL terminated
2345 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2346 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2348 If a malicious actor can cause an application to directly construct an
2349 ASN1_STRING and then process it through one of the affected OpenSSL
2350 functions then this issue could be hit. This might result in a crash
2351 (causing a Denial of Service attack). It could also result in the
2352 disclosure of private memory contents (such as private keys, or
2353 sensitive plaintext).
2358 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2360 * Fixed a problem with verifying a certificate chain when using the
2361 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2362 the certificates present in a certificate chain. It is not set by default.
2364 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2365 the chain that have explicitly encoded elliptic curve parameters was added
2366 as an additional strict check.
2368 An error in the implementation of this check meant that the result of a
2369 previous check to confirm that certificates in the chain are valid CA
2370 certificates was overwritten. This effectively bypasses the check
2371 that non-CA certificates must not be able to issue other certificates.
2373 If a "purpose" has been configured then there is a subsequent opportunity
2374 for checks that the certificate is a valid CA. All of the named "purpose"
2375 values implemented in libcrypto perform this check. Therefore, where
2376 a purpose is set the certificate chain will still be rejected even when the
2377 strict flag has been used. A purpose is set by default in libssl client and
2378 server certificate verification routines, but it can be overridden or
2379 removed by an application.
2381 In order to be affected, an application must explicitly set the
2382 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2383 for the certificate verification or, in the case of TLS client or server
2384 applications, override the default purpose.
2389 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2390 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2391 renegotiation ClientHello omits the signature_algorithms extension (where it
2392 was present in the initial ClientHello), but includes a
2393 signature_algorithms_cert extension then a NULL pointer dereference will
2394 result, leading to a crash and a denial of service attack.
2396 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2397 (which is the default configuration). OpenSSL TLS clients are not impacted by
2401 *Peter Kästle and Samuel Sapalski*
2403 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2405 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2406 create a unique hash value based on the issuer and serial number data
2407 contained within an X509 certificate. However, it was failing to correctly
2408 handle any errors that may occur while parsing the issuer field (which might
2409 occur if the issuer field is maliciously constructed). This may subsequently
2410 result in a NULL pointer deref and a crash leading to a potential denial of
2416 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2417 padding mode to correctly check for rollback attacks. This is considered a
2418 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2423 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2424 functions. Previously they could overflow the output length argument in some
2425 cases where the input length is close to the maximum permissible length for
2426 an integer on the platform. In such cases the return value from the function
2427 call would be 1 (indicating success), but the output length value would be
2428 negative. This could cause applications to behave incorrectly or crash.
2433 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2434 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2435 could be exploited in a side channel attack to recover the password. Since
2436 the attack is local host only this is outside of the current OpenSSL
2437 threat model and therefore no CVE is assigned.
2439 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2444 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2446 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2447 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2448 If an attacker can control both items being compared then this could lead
2449 to a possible denial of service attack. OpenSSL itself uses the
2450 GENERAL_NAME_cmp function for two purposes:
2451 1) Comparing CRL distribution point names between an available CRL and a
2452 CRL distribution point embedded in an X509 certificate
2453 2) When verifying that a timestamp response token signer matches the
2454 timestamp authority name (exposed via the API functions
2455 TS_RESP_verify_response and TS_RESP_verify_token)
2460 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2462 * Certificates with explicit curve parameters are now disallowed in
2463 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2467 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2468 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2469 conversely, silently ignore DTLS protocol version bounds when configuring
2470 TLS-based contexts. The commands can be repeated to set bounds of both
2471 types. The same applies with the corresponding "min_protocol" and
2472 "max_protocol" command-line switches, in case some application uses both TLS
2475 SSL_CTX instances that are created for a fixed protocol version (e.g.
2476 TLSv1_server_method()) also silently ignore version bounds. Previously
2477 attempts to apply bounds to these protocol versions would result in an
2478 error. Now only the "version-flexible" SSL_CTX instances are subject to
2479 limits in configuration files in command-line options.
2483 * Handshake now fails if Extended Master Secret extension is dropped
2488 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2490 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2492 * Fixed segmentation fault in SSL_check_chain()
2493 Server or client applications that call the SSL_check_chain() function
2494 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2495 dereference as a result of incorrect handling of the
2496 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2497 or unrecognised signature algorithm is received from the peer. This could
2498 be exploited by a malicious peer in a Denial of Service attack.
2503 * Added AES consttime code for no-asm configurations
2504 an optional constant time support for AES was added
2505 when building openssl for no-asm.
2506 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2507 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2508 At this time this feature is by default disabled.
2509 It will be enabled by default in 3.0.
2513 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2515 * Revert the change of EOF detection while reading in libssl to avoid
2516 regressions in applications depending on the current way of reporting
2517 the EOF. As the existing method is not fully accurate the change to
2518 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2519 branch and will be present in the 3.0 release.
2523 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2524 when primes for RSA keys are computed.
2525 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2526 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2527 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2528 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2529 This avoids possible fingerprinting of newly generated RSA modules.
2533 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2535 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2536 while reading in libssl then we would report an error back to the
2537 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2538 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2539 therefore give a hint as to what went wrong.
2543 * Check that ed25519 and ed448 are allowed by the security level. Previously
2544 signature algorithms not using an MD were not being checked that they were
2545 allowed by the security level.
2549 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2550 was not quite right. The behaviour was not consistent between resumption
2551 and normal handshakes, and also not quite consistent with historical
2552 behaviour. The behaviour in various scenarios has been clarified and
2553 it has been updated to make it match historical behaviour as closely as
2558 * *[VMS only]* The header files that the VMS compilers include automatically,
2559 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2560 that the C++ compiler doesn't understand. This is a shortcoming in the
2561 compiler, but can be worked around with `__cplusplus` guards.
2563 C++ applications that use OpenSSL libraries must be compiled using the
2564 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2565 functions. Otherwise, only functions with symbols of less than 31
2566 characters can be used, as the linker will not be able to successfully
2567 resolve symbols with longer names.
2571 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2572 The presence of this system service is determined at run-time.
2576 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2581 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2583 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2584 number generator (RNG). This was intended to include protection in the
2585 event of a fork() system call in order to ensure that the parent and child
2586 processes did not share the same RNG state. However, this protection was not
2587 being used in the default case.
2589 A partial mitigation for this issue is that the output from a high
2590 precision timer is mixed into the RNG state so the likelihood of a parent
2591 and child process sharing state is significantly reduced.
2593 If an application already calls OPENSSL_init_crypto() explicitly using
2594 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2597 *Matthias St. Pierre*
2599 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2600 used even when parsing explicit parameters, when loading a encoded key
2601 or calling `EC_GROUP_new_from_ecpkparameters()`/
2602 `EC_GROUP_new_from_ecparameters()`.
2603 This prevents bypass of security hardening and performance gains,
2604 especially for curves with specialized EC_METHODs.
2605 By default, if a key encoded with explicit parameters is loaded and later
2606 encoded, the output is still encoded with explicit parameters, even if
2607 internally a "named" EC_GROUP is used for computation.
2611 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2612 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2613 NULL. After this change, only the cofactor parameter can be NULL. It also
2614 does some minimal sanity checks on the passed order.
2619 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2620 An attack is simple, if the first CMS_recipientInfo is valid but the
2621 second CMS_recipientInfo is chosen ciphertext. If the second
2622 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2623 encryption key will be replaced by garbage, and the message cannot be
2624 decoded, but if the RSA decryption fails, the correct encryption key is
2625 used and the recipient will not notice the attack.
2626 As a work around for this potential attack the length of the decrypted
2627 key must be equal to the cipher default key length, in case the
2628 certificate is not given and all recipientInfo are tried out.
2629 The old behaviour can be re-enabled in the CMS code by setting the
2630 CMS_DEBUG_DECRYPT flag.
2635 * Early start up entropy quality from the DEVRANDOM seed source has been
2636 improved for older Linux systems. The RAND subsystem will wait for
2637 /dev/random to be producing output before seeding from /dev/urandom.
2638 The seeded state is stored for future library initialisations using
2639 a system global shared memory segment. The shared memory identifier
2640 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2641 the desired value. The default identifier is 114.
2645 * Correct the extended master secret constant on EBCDIC systems. Without this
2646 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2647 negotiate EMS will fail. Unfortunately this also means that TLS connections
2648 between EBCDIC systems with this fix, and EBCDIC systems without this
2649 fix will fail if they negotiate EMS.
2653 * Use Windows installation paths in the mingw builds
2655 Mingw isn't a POSIX environment per se, which means that Windows
2656 paths should be used for installation.
2661 * Changed DH_check to accept parameters with order q and 2q subgroups.
2662 With order 2q subgroups the bit 0 of the private key is not secret
2663 but DH_generate_key works around that by clearing bit 0 of the
2664 private key for those. This avoids leaking bit 0 of the private key.
2668 * Significantly reduce secure memory usage by the randomness pools.
2672 * Revert the DEVRANDOM_WAIT feature for Linux systems
2674 The DEVRANDOM_WAIT feature added a select() call to wait for the
2675 /dev/random device to become readable before reading from the
2676 /dev/urandom device.
2678 It turned out that this change had negative side effects on
2679 performance which were not acceptable. After some discussion it
2680 was decided to revert this feature and leave it up to the OS
2681 resp. the platform maintainer to ensure a proper initialization
2682 during early boot time.
2684 *Matthias St. Pierre*
2686 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2688 * Add build tests for C++. These are generated files that only do one
2689 thing, to include one public OpenSSL head file each. This tests that
2690 the public header files can be usefully included in a C++ application.
2692 This test isn't enabled by default. It can be enabled with the option
2693 'enable-buildtest-c++'.
2697 * Enable SHA3 pre-hashing for ECDSA and DSA.
2701 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2702 This changes the size when using the `genpkey` command when no size is given.
2703 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2704 generation commands to use 2048 bits by default.
2708 * Reorganize the manual pages to consistently have RETURN VALUES,
2709 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2710 util/fix-doc-nits accordingly.
2712 *Paul Yang, Joshua Lock*
2714 * Add the missing accessor EVP_PKEY_get0_engine()
2718 * Have commands like `s_client` and `s_server` output the signature scheme
2719 along with other cipher suite parameters when debugging.
2723 * Make OPENSSL_config() error agnostic again.
2727 * Do the error handling in RSA decryption constant time.
2731 * Prevent over long nonces in ChaCha20-Poly1305.
2733 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2734 for every encryption operation. RFC 7539 specifies that the nonce value
2735 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2736 and front pads the nonce with 0 bytes if it is less than 12
2737 bytes. However it also incorrectly allows a nonce to be set of up to 16
2738 bytes. In this case only the last 12 bytes are significant and any
2739 additional leading bytes are ignored.
2741 It is a requirement of using this cipher that nonce values are
2742 unique. Messages encrypted using a reused nonce value are susceptible to
2743 serious confidentiality and integrity attacks. If an application changes
2744 the default nonce length to be longer than 12 bytes and then makes a
2745 change to the leading bytes of the nonce expecting the new value to be a
2746 new unique nonce then such an application could inadvertently encrypt
2747 messages with a reused nonce.
2749 Additionally the ignored bytes in a long nonce are not covered by the
2750 integrity guarantee of this cipher. Any application that relies on the
2751 integrity of these ignored leading bytes of a long nonce may be further
2752 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2753 is safe because no such use sets such a long nonce value. However user
2754 applications that use this cipher directly and set a non-default nonce
2755 length to be longer than 12 bytes may be vulnerable.
2757 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2763 * Add DEVRANDOM_WAIT feature for Linux systems
2765 On older Linux systems where the getrandom() system call is not available,
2766 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2767 Contrary to getrandom(), the /dev/urandom device will not block during
2768 early boot when the kernel CSPRNG has not been seeded yet.
2770 To mitigate this known weakness, use select() to wait for /dev/random to
2771 become readable before reading from /dev/urandom.
2773 * Ensure that SM2 only uses SM3 as digest algorithm
2777 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2779 * Change the info callback signals for the start and end of a post-handshake
2780 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2781 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2782 confused by this and assume that a TLSv1.2 renegotiation has started. This
2783 can break KeyUpdate handling. Instead we no longer signal the start and end
2784 of a post handshake message exchange (although the messages themselves are
2785 still signalled). This could break some applications that were expecting
2786 the old signals. However without this KeyUpdate is not usable for many
2791 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2793 * Timing vulnerability in DSA signature generation
2795 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2796 timing side channel attack. An attacker could use variations in the signing
2797 algorithm to recover the private key.
2799 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2804 * Timing vulnerability in ECDSA signature generation
2806 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2807 timing side channel attack. An attacker could use variations in the signing
2808 algorithm to recover the private key.
2810 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2815 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2816 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2817 of two gigabytes and the error handling improved.
2819 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2820 categorized as a normal bug, not a security issue, because the DRBG reseeds
2821 automatically and is fully functional even without additional randomness
2822 provided by the application.
2824 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2826 * Add a new ClientHello callback. Provides a callback interface that gives
2827 the application the ability to adjust the nascent SSL object at the
2828 earliest stage of ClientHello processing, immediately after extensions have
2829 been collected but before they have been processed. In particular, this
2830 callback can adjust the supported TLS versions in response to the contents
2835 * Add SM2 base algorithm support.
2839 * s390x assembly pack: add (improved) hardware-support for the following
2840 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2841 aes-cfb/cfb8, aes-ecb.
2845 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2846 parameter is no longer accepted, as it leads to a corrupt table. NULL
2847 pem_str is reserved for alias entries only.
2851 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2852 step for prime curves. The new implementation is based on formulae from
2853 differential addition-and-doubling in homogeneous projective coordinates
2854 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2855 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2856 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2857 to work in projective coordinates.
2859 *Billy Bob Brumley, Nicola Tuveri*
2861 * Change generating and checking of primes so that the error rate of not
2862 being prime depends on the intended use based on the size of the input.
2863 For larger primes this will result in more rounds of Miller-Rabin.
2864 The maximal error rate for primes with more than 1080 bits is lowered
2867 *Kurt Roeckx, Annie Yousar*
2869 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2873 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2874 moving between systems, and to avoid confusion when a Windows build is
2875 done with mingw vs with MSVC. For POSIX installs, there's still a
2876 symlink or copy named 'tsget' to avoid that confusion as well.
2880 * Revert blinding in ECDSA sign and instead make problematic addition
2881 length-invariant. Switch even to fixed-length Montgomery multiplication.
2885 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2886 step for binary curves. The new implementation is based on formulae from
2887 differential addition-and-doubling in mixed Lopez-Dahab projective
2888 coordinates, modified to independently blind the operands.
2890 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2892 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2893 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2894 EC_METHODs to implement their own specialized "ladder step", to take
2895 advantage of more favorable coordinate systems or more efficient
2896 differential addition-and-doubling algorithms.
2898 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2900 * Modified the random device based seed sources to keep the relevant
2901 file descriptors open rather than reopening them on each access.
2902 This allows such sources to operate in a chroot() jail without
2903 the associated device nodes being available. This behaviour can be
2904 controlled using RAND_keep_random_devices_open().
2908 * Numerous side-channel attack mitigations have been applied. This may have
2909 performance impacts for some algorithms for the benefit of improved
2910 security. Specific changes are noted in this change log by their respective
2915 * AIX shared library support overhaul. Switch to AIX "natural" way of
2916 handling shared libraries, which means collecting shared objects of
2917 different versions and bitnesses in one common archive. This allows to
2918 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2919 doesn't affect the way 3rd party applications are linked, only how
2920 multi-version installation is managed.
2924 * Make ec_group_do_inverse_ord() more robust and available to other
2925 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2926 mitigations are applied to the fallback BN_mod_inverse().
2927 When using this function rather than BN_mod_inverse() directly, new
2928 EC cryptosystem implementations are then safer-by-default.
2932 * Add coordinate blinding for EC_POINT and implement projective
2933 coordinate blinding for generic prime curves as a countermeasure to
2934 chosen point SCA attacks.
2936 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2938 * Add blinding to ECDSA and DSA signatures to protect against side channel
2939 attacks discovered by Keegan Ryan (NCC Group).
2943 * Enforce checking in the `pkeyutl` command to ensure that the input
2944 length does not exceed the maximum supported digest length when performing
2945 a sign, verify or verifyrecover operation.
2949 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2950 I/O in combination with something like select() or poll() will hang. This
2951 can be turned off again using SSL_CTX_clear_mode().
2952 Many applications do not properly handle non-application data records, and
2953 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2954 around the problems in those applications, but can also break some.
2955 It's recommended to read the manpages about SSL_read(), SSL_write(),
2956 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2957 SSL_CTX_set_read_ahead() again.
2961 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2962 now allow empty (zero character) pass phrases.
2966 * Apply blinding to binary field modular inversion and remove patent
2967 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2971 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2972 binary and prime elliptic curves.
2976 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2977 constant time fixed point multiplication.
2981 * Revise elliptic curve scalar multiplication with timing attack
2982 defenses: ec_wNAF_mul redirects to a constant time implementation
2983 when computing fixed point and variable point multiplication (which
2984 in OpenSSL are mostly used with secret scalars in keygen, sign,
2985 ECDH derive operations).
2986 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2989 * Updated CONTRIBUTING
2993 * Updated DRBG / RAND to request nonce and additional low entropy
2994 randomness from the system.
2996 *Matthias St. Pierre*
2998 * Updated 'openssl rehash' to use OpenSSL consistent default.
3002 * Moved the load of the ssl_conf module to libcrypto, which helps
3003 loading engines that libssl uses before libssl is initialised.
3007 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3011 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3013 *Ingo Schwarze, Rich Salz*
3015 * Added output of accepting IP address and port for 'openssl s_server'
3019 * Added a new API for TLSv1.3 ciphersuites:
3020 SSL_CTX_set_ciphersuites()
3021 SSL_set_ciphersuites()
3025 * Memory allocation failures consistently add an error to the error
3030 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3031 in libcrypto when run as setuid/setgid.
3035 * Load any config file by default when libssl is used.
3039 * Added new public header file <openssl/rand_drbg.h> and documentation
3040 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3042 *Matthias St. Pierre*
3044 * QNX support removed (cannot find contributors to get their approval
3045 for the license change).
3049 * TLSv1.3 replay protection for early data has been implemented. See the
3050 SSL_read_early_data() man page for further details.
3054 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3055 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3056 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3057 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3058 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3059 configuration has been separated out. See the ciphers man page or the
3060 SSL_CTX_set_ciphersuites() man page for more information.
3064 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3065 in responder mode now supports the new "-multi" option, which
3066 spawns the specified number of child processes to handle OCSP
3067 requests. The "-timeout" option now also limits the OCSP
3068 responder's patience to wait to receive the full client request
3069 on a newly accepted connection. Child processes are respawned
3070 as needed, and the CA index file is automatically reloaded
3071 when changed. This makes it possible to run the "ocsp" responder
3072 as a long-running service, making the OpenSSL CA somewhat more
3073 feature-complete. In this mode, most diagnostic messages logged
3074 after entering the event loop are logged via syslog(3) rather than
3079 * Added support for X448 and Ed448. Heavily based on original work by
3084 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3085 objects loaded. This adds the functions OSSL_STORE_expect() and
3086 OSSL_STORE_find() as well as needed tools to construct searches and
3087 get the search data out of them.
3091 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3092 version of OpenSSL should review their configuration settings to ensure
3093 that they are still appropriate for TLSv1.3. For further information see:
3094 <https://wiki.openssl.org/index.php/TLS1.3>
3098 * Grand redesign of the OpenSSL random generator
3100 The default RAND method now utilizes an AES-CTR DRBG according to
3101 NIST standard SP 800-90Ar1. The new random generator is essentially
3102 a port of the default random generator from the OpenSSL FIPS 2.0
3103 object module. It is a hybrid deterministic random bit generator
3104 using an AES-CTR bit stream and which seeds and reseeds itself
3105 automatically using trusted system entropy sources.
3107 Some of its new features are:
3108 - Support for multiple DRBG instances with seed chaining.
3109 - The default RAND method makes use of a DRBG.
3110 - There is a public and private DRBG instance.
3111 - The DRBG instances are fork-safe.
3112 - Keep all global DRBG instances on the secure heap if it is enabled.
3113 - The public and private DRBG instance are per thread for lock free
3116 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3118 * Changed Configure so it only says what it does and doesn't dump
3119 so much data. Instead, ./configdata.pm should be used as a script
3120 to display all sorts of configuration data.
3124 * Added processing of "make variables" to Configure.
3128 * Added SHA512/224 and SHA512/256 algorithm support.
3132 * The last traces of Netware support, first removed in 1.1.0, have
3137 * Get rid of Makefile.shared, and in the process, make the processing
3138 of certain files (rc.obj, or the .def/.map/.opt files produced from
3139 the ordinal files) more visible and hopefully easier to trace and
3140 debug (or make silent).
3144 * Make it possible to have environment variable assignments as
3145 arguments to config / Configure.
3149 * Add multi-prime RSA (RFC 8017) support.
3153 * Add SM3 implemented according to GB/T 32905-2016
3154 *Jack Lloyd <jack.lloyd@ribose.com>,*
3155 *Ronald Tse <ronald.tse@ribose.com>,*
3156 *Erick Borsboom <erick.borsboom@ribose.com>*
3158 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3159 as documented in RFC6066.
3160 Based on a patch from Tomasz Moń
3162 *Filipe Raimundo da Silva*
3164 * Add SM4 implemented according to GB/T 32907-2016.
3165 *Jack Lloyd <jack.lloyd@ribose.com>,*
3166 *Ronald Tse <ronald.tse@ribose.com>,*
3167 *Erick Borsboom <erick.borsboom@ribose.com>*
3169 * Reimplement -newreq-nodes and ERR_error_string_n; the
3170 original author does not agree with the license change.
3174 * Add ARIA AEAD TLS support.
3178 * Some macro definitions to support VS6 have been removed. Visual
3179 Studio 6 has not worked since 1.1.0
3183 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3184 without clearing the errors.
3188 * Add "atfork" functions. If building on a system that without
3189 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3190 requirements. The RAND facility now uses/requires this.
3198 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3199 not possible to disable entirely. However, it's still possible to
3200 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3203 To disable, configure with 'no-ui-console'. 'no-ui' is still
3204 possible to use as an alias. Check at compile time with the
3205 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3206 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3210 * Add a STORE module, which implements a uniform and URI based reader of
3211 stores that can contain keys, certificates, CRLs and numerous other
3212 objects. The main API is loosely based on a few stdio functions,
3213 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3214 OSSL_STORE_error and OSSL_STORE_close.
3215 The implementation uses backends called "loaders" to implement arbitrary
3216 URI schemes. There is one built in "loader" for the 'file' scheme.
3220 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3221 then adjusted to work on FreeBSD 8.4 as well.
3222 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3223 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3227 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3228 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3229 error code calls like this:
3231 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3233 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3234 that can be encoded in C. For the foreseeable future, this will only
3237 *Richard Levitte and Tim Hudson*
3239 * Removed BSD cryptodev engine.
3243 * Add a build target 'build_all_generated', to build all generated files
3244 and only that. This can be used to prepare everything that requires
3245 things like perl for a system that lacks perl and then move everything
3246 to that system and do the rest of the build there.
3250 * In the UI interface, make it possible to duplicate the user data. This
3251 can be used by engines that need to retain the data for a longer time
3252 than just the call where this user data is passed.
3256 * Ignore the '-named_curve auto' value for compatibility of applications
3259 *Tomáš Mráz <tmraz@fedoraproject.org>*
3261 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3262 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3263 alerts across multiple records (some of which could be empty). In practice
3264 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3265 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3266 support this at all. Supporting it adds significant complexity to the
3267 record layer, and its removal is unlikely to cause interoperability
3272 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3273 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3274 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3279 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3280 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3282 *Richard Levitte, Andy Polyakov*
3284 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3289 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3290 platform rather than 'mingw'.
3294 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3295 success if they are asked to add an object which already exists
3296 in the store. This change cascades to other functions which load
3297 certificates and CRLs.
3301 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3302 facilitate stack unwinding even from assembly subroutines.
3306 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3307 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3311 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3312 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3313 which is the minimum version we support.
3317 * Certificate time validation (X509_cmp_time) enforces stricter
3318 compliance with RFC 5280. Fractional seconds and timezone offsets
3319 are no longer allowed.
3323 * Add support for ARIA
3327 * s_client will now send the Server Name Indication (SNI) extension by
3328 default unless the new "-noservername" option is used. The server name is
3329 based on the host provided to the "-connect" option unless overridden by
3330 using "-servername".
3334 * Add support for SipHash
3338 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3339 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3340 prevent issues where no progress is being made and the peer continually
3341 sends unrecognised record types, using up resources processing them.
3345 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3346 using the algorithm defined in
3347 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3351 * Heartbeat support has been removed; the ABI is changed for now.
3353 *Richard Levitte, Rich Salz*
3355 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3359 * The RSA "null" method, which was partially supported to avoid patent
3360 issues, has been replaced to always returns NULL.
3367 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3369 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3370 used even when parsing explicit parameters, when loading a encoded key
3371 or calling `EC_GROUP_new_from_ecpkparameters()`/
3372 `EC_GROUP_new_from_ecparameters()`.
3373 This prevents bypass of security hardening and performance gains,
3374 especially for curves with specialized EC_METHODs.
3375 By default, if a key encoded with explicit parameters is loaded and later
3376 encoded, the output is still encoded with explicit parameters, even if
3377 internally a "named" EC_GROUP is used for computation.
3381 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3382 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3383 NULL. After this change, only the cofactor parameter can be NULL. It also
3384 does some minimal sanity checks on the passed order.
3389 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3390 An attack is simple, if the first CMS_recipientInfo is valid but the
3391 second CMS_recipientInfo is chosen ciphertext. If the second
3392 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3393 encryption key will be replaced by garbage, and the message cannot be
3394 decoded, but if the RSA decryption fails, the correct encryption key is
3395 used and the recipient will not notice the attack.
3396 As a work around for this potential attack the length of the decrypted
3397 key must be equal to the cipher default key length, in case the
3398 certificate is not given and all recipientInfo are tried out.
3399 The old behaviour can be re-enabled in the CMS code by setting the
3400 CMS_DEBUG_DECRYPT flag.
3405 * Use Windows installation paths in the mingw builds
3407 Mingw isn't a POSIX environment per se, which means that Windows
3408 paths should be used for installation.
3413 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3415 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3416 This changes the size when using the `genpkey` command when no size is given.
3417 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3418 generation commands to use 2048 bits by default.
3422 * Prevent over long nonces in ChaCha20-Poly1305.
3424 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3425 for every encryption operation. RFC 7539 specifies that the nonce value
3426 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3427 and front pads the nonce with 0 bytes if it is less than 12
3428 bytes. However it also incorrectly allows a nonce to be set of up to 16
3429 bytes. In this case only the last 12 bytes are significant and any
3430 additional leading bytes are ignored.
3432 It is a requirement of using this cipher that nonce values are
3433 unique. Messages encrypted using a reused nonce value are susceptible to
3434 serious confidentiality and integrity attacks. If an application changes
3435 the default nonce length to be longer than 12 bytes and then makes a
3436 change to the leading bytes of the nonce expecting the new value to be a
3437 new unique nonce then such an application could inadvertently encrypt
3438 messages with a reused nonce.
3440 Additionally the ignored bytes in a long nonce are not covered by the
3441 integrity guarantee of this cipher. Any application that relies on the
3442 integrity of these ignored leading bytes of a long nonce may be further
3443 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3444 is safe because no such use sets such a long nonce value. However user
3445 applications that use this cipher directly and set a non-default nonce
3446 length to be longer than 12 bytes may be vulnerable.
3448 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3454 * Added SCA hardening for modular field inversion in EC_GROUP through
3455 a new dedicated field_inv() pointer in EC_METHOD.
3456 This also addresses a leakage affecting conversions from projective
3457 to affine coordinates.
3459 *Billy Bob Brumley, Nicola Tuveri*
3461 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3462 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3466 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3470 * Remove the 'dist' target and add a tarball building script. The
3471 'dist' target has fallen out of use, and it shouldn't be
3472 necessary to configure just to create a source distribution.
3476 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3478 * Timing vulnerability in DSA signature generation
3480 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3481 timing side channel attack. An attacker could use variations in the signing
3482 algorithm to recover the private key.
3484 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3489 * Timing vulnerability in ECDSA signature generation
3491 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3492 timing side channel attack. An attacker could use variations in the signing
3493 algorithm to recover the private key.
3495 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3500 * Add coordinate blinding for EC_POINT and implement projective
3501 coordinate blinding for generic prime curves as a countermeasure to
3502 chosen point SCA attacks.
3504 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3506 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3508 * Client DoS due to large DH parameter
3510 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3511 malicious server can send a very large prime value to the client. This will
3512 cause the client to spend an unreasonably long period of time generating a
3513 key for this prime resulting in a hang until the client has finished. This
3514 could be exploited in a Denial Of Service attack.
3516 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3521 * Cache timing vulnerability in RSA Key Generation
3523 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3524 a cache timing side channel attack. An attacker with sufficient access to
3525 mount cache timing attacks during the RSA key generation process could
3526 recover the private key.
3528 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3529 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3534 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3535 parameter is no longer accepted, as it leads to a corrupt table. NULL
3536 pem_str is reserved for alias entries only.
3540 * Revert blinding in ECDSA sign and instead make problematic addition
3541 length-invariant. Switch even to fixed-length Montgomery multiplication.
3545 * Change generating and checking of primes so that the error rate of not
3546 being prime depends on the intended use based on the size of the input.
3547 For larger primes this will result in more rounds of Miller-Rabin.
3548 The maximal error rate for primes with more than 1080 bits is lowered
3551 *Kurt Roeckx, Annie Yousar*
3553 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3557 * Add blinding to ECDSA and DSA signatures to protect against side channel
3558 attacks discovered by Keegan Ryan (NCC Group).
3562 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3563 now allow empty (zero character) pass phrases.
3567 * Certificate time validation (X509_cmp_time) enforces stricter
3568 compliance with RFC 5280. Fractional seconds and timezone offsets
3569 are no longer allowed.
3573 * Fixed a text canonicalisation bug in CMS
3575 Where a CMS detached signature is used with text content the text goes
3576 through a canonicalisation process first prior to signing or verifying a
3577 signature. This process strips trailing space at the end of lines, converts
3578 line terminators to CRLF and removes additional trailing line terminators
3579 at the end of a file. A bug in the canonicalisation process meant that
3580 some characters, such as form-feed, were incorrectly treated as whitespace
3581 and removed. This is contrary to the specification (RFC5485). This fix
3582 could mean that detached text data signed with an earlier version of
3583 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3584 signed with a fixed OpenSSL may fail to verify with an earlier version of
3585 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3586 and use the "-binary" flag (for the "cms" command line application) or set
3587 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3591 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3593 * Constructed ASN.1 types with a recursive definition could exceed the stack
3595 Constructed ASN.1 types with a recursive definition (such as can be found
3596 in PKCS7) could eventually exceed the stack given malicious input with
3597 excessive recursion. This could result in a Denial Of Service attack. There
3598 are no such structures used within SSL/TLS that come from untrusted sources
3599 so this is considered safe.
3601 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3607 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3609 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3610 effectively reduced to only comparing the least significant bit of each
3611 byte. This allows an attacker to forge messages that would be considered as
3612 authenticated in an amount of tries lower than that guaranteed by the
3613 security claims of the scheme. The module can only be compiled by the
3614 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3616 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3622 * Add a build target 'build_all_generated', to build all generated files
3623 and only that. This can be used to prepare everything that requires
3624 things like perl for a system that lacks perl and then move everything
3625 to that system and do the rest of the build there.
3629 * Backport SSL_OP_NO_RENGOTIATION
3631 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3632 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3633 changes this is no longer possible in 1.1.0. Therefore, the new
3634 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3635 1.1.0 to provide equivalent functionality.
3637 Note that if an application built against 1.1.0h headers (or above) is run
3638 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3639 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3643 * Removed the OS390-Unix config target. It relied on a script that doesn't
3648 * rsaz_1024_mul_avx2 overflow bug on x86_64
3650 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3651 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3652 Analysis suggests that attacks against RSA and DSA as a result of this
3653 defect would be very difficult to perform and are not believed likely.
3654 Attacks against DH1024 are considered just feasible, because most of the
3655 work necessary to deduce information about a private key may be performed
3656 offline. The amount of resources required for such an attack would be
3657 significant. However, for an attack on TLS to be meaningful, the server
3658 would have to share the DH1024 private key among multiple clients, which is
3659 no longer an option since CVE-2016-0701.
3661 This only affects processors that support the AVX2 but not ADX extensions
3662 like Intel Haswell (4th generation).
3664 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3665 was originally found via the OSS-Fuzz project.
3670 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3672 * bn_sqrx8x_internal carry bug on x86_64
3674 There is a carry propagating bug in the x86_64 Montgomery squaring
3675 procedure. No EC algorithms are affected. Analysis suggests that attacks
3676 against RSA and DSA as a result of this defect would be very difficult to
3677 perform and are not believed likely. Attacks against DH are considered just
3678 feasible (although very difficult) because most of the work necessary to
3679 deduce information about a private key may be performed offline. The amount
3680 of resources required for such an attack would be very significant and
3681 likely only accessible to a limited number of attackers. An attacker would
3682 additionally need online access to an unpatched system using the target
3683 private key in a scenario with persistent DH parameters and a private
3684 key that is shared between multiple clients.
3686 This only affects processors that support the BMI1, BMI2 and ADX extensions
3687 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3689 This issue was reported to OpenSSL by the OSS-Fuzz project.
3694 * Malformed X.509 IPAddressFamily could cause OOB read
3696 If an X.509 certificate has a malformed IPAddressFamily extension,
3697 OpenSSL could do a one-byte buffer overread. The most likely result
3698 would be an erroneous display of the certificate in text format.
3700 This issue was reported to OpenSSL by the OSS-Fuzz project.
3705 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3707 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3708 platform rather than 'mingw'.
3712 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3713 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3714 which is the minimum version we support.
3718 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3720 * Encrypt-Then-Mac renegotiation crash
3722 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3723 negotiated where it was not in the original handshake (or vice-versa) then
3724 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3725 and servers are affected.
3727 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3732 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3734 * Truncated packet could crash via OOB read
3736 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3737 cipher is being used, then a truncated packet can cause that host to
3738 perform an out-of-bounds read, usually resulting in a crash.
3740 This issue was reported to OpenSSL by Robert Święcki of Google.
3745 * Bad (EC)DHE parameters cause a client crash
3747 If a malicious server supplies bad parameters for a DHE or ECDHE key
3748 exchange then this can result in the client attempting to dereference a
3749 NULL pointer leading to a client crash. This could be exploited in a Denial
3752 This issue was reported to OpenSSL by Guido Vranken.
3757 * BN_mod_exp may produce incorrect results on x86_64
3759 There is a carry propagating bug in the x86_64 Montgomery squaring
3760 procedure. No EC algorithms are affected. Analysis suggests that attacks
3761 against RSA and DSA as a result of this defect would be very difficult to
3762 perform and are not believed likely. Attacks against DH are considered just
3763 feasible (although very difficult) because most of the work necessary to
3764 deduce information about a private key may be performed offline. The amount
3765 of resources required for such an attack would be very significant and
3766 likely only accessible to a limited number of attackers. An attacker would
3767 additionally need online access to an unpatched system using the target
3768 private key in a scenario with persistent DH parameters and a private
3769 key that is shared between multiple clients. For example this can occur by
3770 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3771 similar to CVE-2015-3193 but must be treated as a separate problem.
3773 This issue was reported to OpenSSL by the OSS-Fuzz project.
3778 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3780 * ChaCha20/Poly1305 heap-buffer-overflow
3782 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3783 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3784 crash. This issue is not considered to be exploitable beyond a DoS.
3786 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3791 * CMS Null dereference
3793 Applications parsing invalid CMS structures can crash with a NULL pointer
3794 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3795 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3796 structure callback if an attempt is made to free certain invalid encodings.
3797 Only CHOICE structures using a callback which do not handle NULL value are
3800 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3805 * Montgomery multiplication may produce incorrect results
3807 There is a carry propagating bug in the Broadwell-specific Montgomery
3808 multiplication procedure that handles input lengths divisible by, but
3809 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3810 and DH private keys are impossible. This is because the subroutine in
3811 question is not used in operations with the private key itself and an input
3812 of the attacker's direct choice. Otherwise the bug can manifest itself as
3813 transient authentication and key negotiation failures or reproducible
3814 erroneous outcome of public-key operations with specially crafted input.
3815 Among EC algorithms only Brainpool P-512 curves are affected and one
3816 presumably can attack ECDH key negotiation. Impact was not analyzed in
3817 detail, because pre-requisites for attack are considered unlikely. Namely
3818 multiple clients have to choose the curve in question and the server has to
3819 share the private key among them, neither of which is default behaviour.
3820 Even then only clients that chose the curve will be affected.
3822 This issue was publicly reported as transient failures and was not
3823 initially recognized as a security issue. Thanks to Richard Morgan for
3824 providing reproducible case.
3829 * Removed automatic addition of RPATH in shared libraries and executables,
3830 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3834 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3836 * Fix Use After Free for large message sizes
3838 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3839 message larger than approx 16k is received then the underlying buffer to
3840 store the incoming message is reallocated and moved. Unfortunately a
3841 dangling pointer to the old location is left which results in an attempt to
3842 write to the previously freed location. This is likely to result in a
3843 crash, however it could potentially lead to execution of arbitrary code.
3845 This issue only affects OpenSSL 1.1.0a.
3847 This issue was reported to OpenSSL by Robert Święcki.
3852 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3854 * OCSP Status Request extension unbounded memory growth
3856 A malicious client can send an excessively large OCSP Status Request
3857 extension. If that client continually requests renegotiation, sending a
3858 large OCSP Status Request extension each time, then there will be unbounded
3859 memory growth on the server. This will eventually lead to a Denial Of
3860 Service attack through memory exhaustion. Servers with a default
3861 configuration are vulnerable even if they do not support OCSP. Builds using
3862 the "no-ocsp" build time option are not affected.
3864 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3869 * SSL_peek() hang on empty record
3871 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3872 sends an empty record. This could be exploited by a malicious peer in a
3873 Denial Of Service attack.
3875 This issue was reported to OpenSSL by Alex Gaynor.
3880 * Excessive allocation of memory in tls_get_message_header() and
3881 dtls1_preprocess_fragment()
3883 A (D)TLS message includes 3 bytes for its length in the header for the
3884 message. This would allow for messages up to 16Mb in length. Messages of
3885 this length are excessive and OpenSSL includes a check to ensure that a
3886 peer is sending reasonably sized messages in order to avoid too much memory
3887 being consumed to service a connection. A flaw in the logic of version
3888 1.1.0 means that memory for the message is allocated too early, prior to
3889 the excessive message length check. Due to way memory is allocated in
3890 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3891 to service a connection. This could lead to a Denial of Service through
3892 memory exhaustion. However, the excessive message length check still takes
3893 place, and this would cause the connection to immediately fail. Assuming
3894 that the application calls SSL_free() on the failed connection in a timely
3895 manner then the 21Mb of allocated memory will then be immediately freed
3896 again. Therefore, the excessive memory allocation will be transitory in
3897 nature. This then means that there is only a security impact if:
3899 1) The application does not call SSL_free() in a timely manner in the event
3900 that the connection fails
3902 2) The application is working in a constrained environment where there is
3903 very little free memory
3905 3) The attacker initiates multiple connection attempts such that there are
3906 multiple connections in a state where memory has been allocated for the
3907 connection; SSL_free() has not yet been called; and there is insufficient
3908 memory to service the multiple requests.
3910 Except in the instance of (1) above any Denial Of Service is likely to be
3911 transitory because as soon as the connection fails the memory is
3912 subsequently freed again in the SSL_free() call. However there is an
3913 increased risk during this period of application crashes due to the lack of
3914 memory - which would then mean a more serious Denial of Service.
3916 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3917 (CVE-2016-6307 and CVE-2016-6308)
3921 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3922 had to be removed. Primary reason is that vendor assembler can't
3923 assemble our modules with -KPIC flag. As result it, assembly
3924 support, was not even available as option. But its lack means
3925 lack of side-channel resistant code, which is incompatible with
3926 security by todays standards. Fortunately gcc is readily available
3927 prepackaged option, which we firmly point at...
3931 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3933 * Windows command-line tool supports UTF-8 opt-in option for arguments
3934 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3935 (to any value) allows Windows user to access PKCS#12 file generated
3936 with Windows CryptoAPI and protected with non-ASCII password, as well
3937 as files generated under UTF-8 locale on Linux also protected with
3942 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3943 have been disabled by default and removed from DEFAULT, just like RC4.
3944 See the RC4 item below to re-enable both.
3948 * The method for finding the storage location for the Windows RAND seed file
3949 has changed. First we check %RANDFILE%. If that is not set then we check
3950 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3951 all else fails we fall back to C:\.
3955 * The EVP_EncryptUpdate() function has had its return type changed from void
3956 to int. A return of 0 indicates and error while a return of 1 indicates
3961 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3962 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3963 off the constant time implementation for RSA, DSA and DH have been made
3964 no-ops and deprecated.
3968 * Windows RAND implementation was simplified to only get entropy by
3969 calling CryptGenRandom(). Various other RAND-related tickets
3972 *Joseph Wylie Yandle, Rich Salz*
3974 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3975 and `OPENSSL_LH_`, respectively. The old names are available
3976 with API compatibility. They new names are now completely documented.
3980 * Unify TYPE_up_ref(obj) methods signature.
3981 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3982 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3983 int (instead of void) like all others TYPE_up_ref() methods.
3984 So now these methods also check the return value of CRYPTO_atomic_add(),
3985 and the validity of object reference counter.
3987 *fdasilvayy@gmail.com*
3989 * With Windows Visual Studio builds, the .pdb files are installed
3990 alongside the installed libraries and executables. For a static
3991 library installation, ossl_static.pdb is the associate compiler
3992 generated .pdb file to be used when linking programs.
3996 * Remove openssl.spec. Packaging files belong with the packagers.
4000 * Automatic Darwin/OSX configuration has had a refresh, it will now
4001 recognise x86_64 architectures automatically. You can still decide
4002 to build for a different bitness with the environment variable
4003 KERNEL_BITS (can be 32 or 64), for example:
4005 KERNEL_BITS=32 ./config
4009 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4010 256 bit AES and HMAC with SHA256.
4014 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4018 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4022 * To enable users to have their own config files and build file templates,
4023 Configure looks in the directory indicated by the environment variable
4024 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4025 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4026 name and is used as is.
4030 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4031 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4032 X509_CERT_FILE_CTX was removed.
4036 * "shared" builds are now the default. To create only static libraries use
4037 the "no-shared" Configure option.
4041 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4042 All of these option have not worked for some while and are fundamental
4047 * Make various cleanup routines no-ops and mark them as deprecated. Most
4048 global cleanup functions are no longer required because they are handled
4049 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4050 Explicitly de-initing can cause problems (e.g. where a library that uses
4051 OpenSSL de-inits, but an application is still using it). The affected
4052 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4053 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4054 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4055 COMP_zlib_cleanup().
4059 * --strict-warnings no longer enables runtime debugging options
4060 such as REF_DEBUG. Instead, debug options are automatically
4061 enabled with '--debug' builds.
4063 *Andy Polyakov, Emilia Käsper*
4065 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4066 have been moved out of the public header files. New functions for managing
4067 these have been added.
4071 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4072 objects have been moved out of the public header files. New
4073 functions for managing these have been added.
4077 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4078 have been moved out of the public header files. New functions for managing
4079 these have been added.
4083 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4084 moved out of the public header files. New functions for managing these
4089 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4093 * Removed the mk1mf build scripts.
4097 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4098 it is always safe to #include a header now.
4102 * Removed the aged BC-32 config and all its supporting scripts
4106 * Removed support for Ultrix, Netware, and OS/2.
4110 * Add support for HKDF.
4112 *Alessandro Ghedini*
4114 * Add support for blake2b and blake2s
4118 * Added support for "pipelining". Ciphers that have the
4119 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4120 encryptions/decryptions simultaneously. There are currently no built-in
4121 ciphers with this property but the expectation is that engines will be able
4122 to offer it to significantly improve throughput. Support has been extended
4123 into libssl so that multiple records for a single connection can be
4124 processed in one go (for >=TLS 1.1).
4128 * Added the AFALG engine. This is an async capable engine which is able to
4129 offload work to the Linux kernel. In this initial version it only supports
4130 AES128-CBC. The kernel must be version 4.1.0 or greater.
4134 * OpenSSL now uses a new threading API. It is no longer necessary to
4135 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4136 are two supported threading models: pthreads and windows threads. It is
4137 also possible to configure OpenSSL at compile time for "no-threads". The
4138 old threading API should no longer be used. The functions have been
4139 replaced with "no-op" compatibility macros.
4141 *Alessandro Ghedini, Matt Caswell*
4143 * Modify behavior of ALPN to invoke callback after SNI/servername
4144 callback, such that updates to the SSL_CTX affect ALPN.
4148 * Add SSL_CIPHER queries for authentication and key-exchange.
4152 * Changes to the DEFAULT cipherlist:
4153 - Prefer (EC)DHE handshakes over plain RSA.
4154 - Prefer AEAD ciphers over legacy ciphers.
4155 - Prefer ECDSA over RSA when both certificates are available.
4156 - Prefer TLSv1.2 ciphers/PRF.
4157 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4162 * Change the ECC default curve list to be this, in order: x25519,
4163 secp256r1, secp521r1, secp384r1.
4167 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4168 disabled by default. They can be re-enabled using the
4169 enable-weak-ssl-ciphers option to Configure.
4173 * If the server has ALPN configured, but supports no protocols that the
4174 client advertises, send a fatal "no_application_protocol" alert.
4175 This behaviour is SHALL in RFC 7301, though it isn't universally
4176 implemented by other servers.
4180 * Add X25519 support.
4181 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4182 for public and private key encoding using the format documented in
4183 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4184 key generation and key derivation.
4186 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4191 * Deprecate SRP_VBASE_get_by_user.
4192 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4193 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4194 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4195 seed, even if the seed is configured.
4197 Users should use SRP_VBASE_get1_by_user instead. Note that in
4198 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4199 also that even though configuring the SRP seed attempts to hide
4200 invalid usernames by continuing the handshake with fake
4201 credentials, this behaviour is not constant time and no strong
4202 guarantees are made that the handshake is indistinguishable from
4203 that of a valid user.
4207 * Configuration change; it's now possible to build dynamic engines
4208 without having to build shared libraries and vice versa. This
4209 only applies to the engines in `engines/`, those in `crypto/engine/`
4210 will always be built into libcrypto (i.e. "static").
4212 Building dynamic engines is enabled by default; to disable, use
4213 the configuration option "disable-dynamic-engine".
4215 The only requirements for building dynamic engines are the
4216 presence of the DSO module and building with position independent
4217 code, so they will also automatically be disabled if configuring
4218 with "disable-dso" or "disable-pic".
4220 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4221 are also taken away from openssl/opensslconf.h, as they are
4226 * Configuration change; if there is a known flag to compile
4227 position independent code, it will always be applied on the
4228 libcrypto and libssl object files, and never on the application
4229 object files. This means other libraries that use routines from
4230 libcrypto / libssl can be made into shared libraries regardless
4231 of how OpenSSL was configured.
4233 If this isn't desirable, the configuration options "disable-pic"
4234 or "no-pic" can be used to disable the use of PIC. This will
4235 also disable building shared libraries and dynamic engines.
4239 * Removed JPAKE code. It was experimental and has no wide use.
4243 * The INSTALL_PREFIX Makefile variable has been renamed to
4244 DESTDIR. That makes for less confusion on what this variable
4245 is for. Also, the configuration option --install_prefix is
4250 * Heartbeat for TLS has been removed and is disabled by default
4251 for DTLS; configure with enable-heartbeats. Code that uses the
4252 old #define's might need to be updated.
4254 *Emilia Käsper, Rich Salz*
4256 * Rename REF_CHECK to REF_DEBUG.
4260 * New "unified" build system
4262 The "unified" build system is aimed to be a common system for all
4263 platforms we support. With it comes new support for VMS.
4265 This system builds supports building in a different directory tree
4266 than the source tree. It produces one Makefile (for unix family
4267 or lookalikes), or one descrip.mms (for VMS).
4269 The source of information to make the Makefile / descrip.mms is
4270 small files called 'build.info', holding the necessary
4271 information for each directory with source to compile, and a
4272 template in Configurations, like unix-Makefile.tmpl or
4275 With this change, the library names were also renamed on Windows
4276 and on VMS. They now have names that are closer to the standard
4277 on Unix, and include the major version number, and in certain
4278 cases, the architecture they are built for. See "Notes on shared
4279 libraries" in INSTALL.
4281 We rely heavily on the perl module Text::Template.
4285 * Added support for auto-initialisation and de-initialisation of the library.
4286 OpenSSL no longer requires explicit init or deinit routines to be called,
4287 except in certain circumstances. See the OPENSSL_init_crypto() and
4288 OPENSSL_init_ssl() man pages for further information.
4292 * The arguments to the DTLSv1_listen function have changed. Specifically the
4293 "peer" argument is now expected to be a BIO_ADDR object.
4295 * Rewrite of BIO networking library. The BIO library lacked consistent
4296 support of IPv6, and adding it required some more extensive
4297 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4298 which hold all types of addresses and chains of address information.
4299 It also introduces a new API, with functions like BIO_socket,
4300 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4301 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4302 have been adapted accordingly.
4306 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4311 * CRIME protection: disable compression by default, even if OpenSSL is
4312 compiled with zlib enabled. Applications can still enable compression
4313 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4314 using the SSL_CONF library to configure compression.
4318 * The signature of the session callback configured with
4319 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4320 was explicitly marked as `const unsigned char*` instead of
4325 * Always DPURIFY. Remove the use of uninitialized memory in the
4326 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4330 * Removed many obsolete configuration items, including
4331 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4332 MD2_CHAR, MD2_INT, MD2_LONG
4334 IDEA_SHORT, IDEA_LONG
4335 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4337 *Rich Salz, with advice from Andy Polyakov*
4339 * Many BN internals have been moved to an internal header file.
4341 *Rich Salz with help from Andy Polyakov*
4343 * Configuration and writing out the results from it has changed.
4344 Files such as Makefile include/openssl/opensslconf.h and are now
4345 produced through general templates, such as Makefile.in and
4346 crypto/opensslconf.h.in and some help from the perl module
4349 Also, the center of configuration information is no longer
4350 Makefile. Instead, Configure produces a perl module in
4351 configdata.pm which holds most of the config data (in the hash
4352 table %config), the target data that comes from the target
4353 configuration in one of the `Configurations/*.conf` files (in
4358 * To clarify their intended purposes, the Configure options
4359 --prefix and --openssldir change their semantics, and become more
4360 straightforward and less interdependent.
4362 --prefix shall be used exclusively to give the location INSTALLTOP
4363 where programs, scripts, libraries, include files and manuals are
4364 going to be installed. The default is now /usr/local.
4366 --openssldir shall be used exclusively to give the default
4367 location OPENSSLDIR where certificates, private keys, CRLs are
4368 managed. This is also where the default openssl.cnf gets
4370 If the directory given with this option is a relative path, the
4371 values of both the --prefix value and the --openssldir value will
4372 be combined to become OPENSSLDIR.
4373 The default for --openssldir is INSTALLTOP/ssl.
4375 Anyone who uses --openssldir to specify where OpenSSL is to be
4376 installed MUST change to use --prefix instead.
4380 * The GOST engine was out of date and therefore it has been removed. An up
4381 to date GOST engine is now being maintained in an external repository.
4382 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4383 support for GOST ciphersuites (these are only activated if a GOST engine
4388 * EGD is no longer supported by default; use enable-egd when
4391 *Ben Kaduk and Rich Salz*
4393 * The distribution now has Makefile.in files, which are used to
4394 create Makefile's when Configure is run. *Configure must be run
4395 before trying to build now.*
4399 * The return value for SSL_CIPHER_description() for error conditions
4404 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4406 Obtaining and performing DNSSEC validation of TLSA records is
4407 the application's responsibility. The application provides
4408 the TLSA records of its choice to OpenSSL, and these are then
4409 used to authenticate the peer.
4411 The TLSA records need not even come from DNS. They can, for
4412 example, be used to implement local end-entity certificate or
4413 trust-anchor "pinning", where the "pin" data takes the form
4414 of TLSA records, which can augment or replace verification
4415 based on the usual WebPKI public certification authorities.
4419 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4420 continues to support deprecated interfaces in default builds.
4421 However, applications are strongly advised to compile their
4422 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4423 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4424 or the 1.1.0 releases.
4426 In environments in which all applications have been ported to
4427 not use any deprecated interfaces OpenSSL's Configure script
4428 should be used with the --api=1.1.0 option to entirely remove
4429 support for the deprecated features from the library and
4430 unconditionally disable them in the installed headers.
4431 Essentially the same effect can be achieved with the "no-deprecated"
4432 argument to Configure, except that this will always restrict
4433 the build to just the latest API, rather than a fixed API
4436 As applications are ported to future revisions of the API,
4437 they should update their compile-time OPENSSL_API_COMPAT define
4438 accordingly, but in most cases should be able to continue to
4439 compile with later releases.
4441 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4442 0x10000000L and 0x00908000L, respectively. However those
4443 versions did not support the OPENSSL_API_COMPAT feature, and
4444 so applications are not typically tested for explicit support
4445 of just the undeprecated features of either release.
4449 * Add support for setting the minimum and maximum supported protocol.
4450 It can bet set via the SSL_set_min_proto_version() and
4451 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4452 MaxProtocol. It's recommended to use the new APIs to disable
4453 protocols instead of disabling individual protocols using
4454 SSL_set_options() or SSL_CONF's Protocol. This change also
4455 removes support for disabling TLS 1.2 in the OpenSSL TLS
4456 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4460 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4464 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4465 and integrates ECDSA and ECDH functionality into EC. Implementations can
4466 now redirect key generation and no longer need to convert to or from
4469 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4470 include the ec.h header file instead.
4474 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4475 ciphers who are no longer supported and drops support the ephemeral RSA key
4476 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4480 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4481 opaque. For HMAC_CTX, the following constructors and destructors
4484 HMAC_CTX *HMAC_CTX_new(void);
4485 void HMAC_CTX_free(HMAC_CTX *ctx);
4487 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4488 destroy such methods has been added. See EVP_MD_meth_new(3) and
4489 EVP_CIPHER_meth_new(3) for documentation.
4492 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4493 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4494 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4495 an already created structure.
4496 2) For consistency with the majority of our object creators and
4497 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4498 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4499 for deprecated builds.
4503 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4504 cryptographic operations to be performed asynchronously as long as an
4505 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4506 further details. Libssl has also had this capability integrated with the
4507 introduction of the new mode SSL_MODE_ASYNC and associated error
4508 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4509 pages. This work was developed in partnership with Intel Corp.
4513 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4514 always enabled now. If you want to disable the support you should
4515 exclude it using the list of supported ciphers. This also means that the
4516 "-no_ecdhe" option has been removed from s_server.
4520 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4521 SSL_{CTX_}set1_curves() which can set a list.
4525 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4526 curve you want to support using SSL_{CTX_}set1_curves().
4530 * State machine rewrite. The state machine code has been significantly
4531 refactored in order to remove much duplication of code and solve issues
4532 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4533 further details). This change does have some associated API changes.
4534 Notably the SSL_state() function has been removed and replaced by
4535 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4536 SSL_set_state() has been removed altogether. The previous handshake states
4537 defined in ssl.h and ssl3.h have also been removed.
4541 * All instances of the string "ssleay" in the public API were replaced
4542 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4543 Some error codes related to internal RSA_eay API's were renamed.
4547 * The demo files in crypto/threads were moved to demo/threads.
4551 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4554 *Matt Caswell, Rich Salz*
4556 * New ASN.1 embed macro.
4558 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4559 structure is not allocated: it is part of the parent. That is instead of
4567 This reduces memory fragmentation and make it impossible to accidentally
4568 set a mandatory field to NULL.
4570 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4571 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4572 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4577 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4581 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4582 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4583 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4584 DES and RC4 ciphersuites.
4588 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4589 This changes the decoding behaviour for some invalid messages,
4590 though the change is mostly in the more lenient direction, and
4591 legacy behaviour is preserved as much as possible.
4595 * Fix no-stdio build.
4596 *David Woodhouse <David.Woodhouse@intel.com> and also*
4597 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4599 * New testing framework
4600 The testing framework has been largely rewritten and is now using
4601 perl and the perl modules Test::Harness and an extended variant of
4602 Test::More called OpenSSL::Test to do its work. All test scripts in
4603 test/ have been rewritten into test recipes, and all direct calls to
4604 executables in test/Makefile have become individual recipes using the
4605 simplified testing OpenSSL::Test::Simple.
4607 For documentation on our testing modules, do:
4609 perldoc test/testlib/OpenSSL/Test/Simple.pm
4610 perldoc test/testlib/OpenSSL/Test.pm
4614 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4615 are used; the latter aborts on memory leaks (usually checked on exit).
4616 Some undocumented "set malloc, etc., hooks" functions were removed
4617 and others were changed. All are now documented.
4621 * In DSA_generate_parameters_ex, if the provided seed is too short,
4624 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4626 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4627 from RFC4279, RFC4785, RFC5487, RFC5489.
4629 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4630 original RSA_PSK patch.
4634 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4635 era flag was never set throughout the codebase (only read). Also removed
4636 SSL3_FLAGS_POP_BUFFER which was only used if
4637 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4641 * Changed the default name options in the "ca", "crl", "req" and "x509"
4642 to be "oneline" instead of "compat".
4646 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4647 not aware of clients that still exhibit this bug, and the workaround
4648 hasn't been working properly for a while.
4652 * The return type of BIO_number_read() and BIO_number_written() as well as
4653 the corresponding num_read and num_write members in the BIO structure has
4654 changed from unsigned long to uint64_t. On platforms where an unsigned
4655 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4660 * Given the pervasive nature of TLS extensions it is inadvisable to run
4661 OpenSSL without support for them. It also means that maintaining
4662 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4663 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4667 * Removed support for the two export grade static DH ciphersuites
4668 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4669 were newly added (along with a number of other static DH ciphersuites) to
4670 1.0.2. However the two export ones have *never* worked since they were
4671 introduced. It seems strange in any case to be adding new export
4672 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4676 * Version negotiation has been rewritten. In particular SSLv23_method(),
4677 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4678 and turned into macros which simply call the new preferred function names
4679 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4680 should use the new names instead. Also as part of this change the ssl23.h
4681 header file has been removed.
4685 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4686 code and the associated standard is no longer considered fit-for-purpose.
4690 * RT2547 was closed. When generating a private key, try to make the
4691 output file readable only by the owner. This behavior change might
4692 be noticeable when interacting with other software.
4694 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4699 * Added HTTP GET support to the ocsp command.
4703 * Changed default digest for the dgst and enc commands from MD5 to
4708 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4712 * Added support for TLS extended master secret from
4713 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4714 initial patch which was a great help during development.
4718 * All libssl internal structures have been removed from the public header
4719 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4720 now redundant). Users should not attempt to access internal structures
4721 directly. Instead they should use the provided API functions.
4725 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4726 Access to deprecated functions can be re-enabled by running config with
4727 "enable-deprecated". In addition applications wishing to use deprecated
4728 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4729 will, by default, disable some transitive includes that previously existed
4730 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4734 * Added support for OCB mode. OpenSSL has been granted a patent license
4735 compatible with the OpenSSL license for use of OCB. Details are available
4736 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4737 for OCB can be removed by calling config with no-ocb.
4741 * SSLv2 support has been removed. It still supports receiving an SSLv2
4742 compatible client hello.
4746 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4747 done while fixing the error code for the key-too-small case.
4749 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4751 * CA.sh has been removed; use CA.pl instead.
4755 * Removed old DES API.
4759 * Remove various unsupported platforms:
4765 Sinix/ReliantUNIX RM400
4770 16-bit platforms such as WIN16
4774 * Clean up OPENSSL_NO_xxx #define's
4775 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4776 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4777 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4778 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4779 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4780 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4781 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4782 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4783 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4784 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4788 * Cleaned up dead code
4789 Remove all but one '#ifdef undef' which is to be looked at.
4793 * Clean up calling of xxx_free routines.
4794 Just like free(), fix most of the xxx_free routines to accept
4795 NULL. Remove the non-null checks from callers. Save much code.
4799 * Add secure heap for storage of private keys (when possible).
4800 Add BIO_s_secmem(), CBIGNUM, etc.
4801 Contributed by Akamai Technologies under our Corporate CLA.
4805 * Experimental support for a new, fast, unbiased prime candidate generator,
4806 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4808 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4810 * New output format NSS in the sess_id command line tool. This allows
4811 exporting the session id and the master key in NSS keylog format.
4813 *Martin Kaiser <martin@kaiser.cx>*
4815 * Harmonize version and its documentation. -f flag is used to display
4818 *mancha <mancha1@zoho.com>*
4820 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4821 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4823 *mancha <mancha1@zoho.com>*
4825 * Fix some double frees. These are not thought to be exploitable.
4827 *mancha <mancha1@zoho.com>*
4829 * A missing bounds check in the handling of the TLS heartbeat extension
4830 can be used to reveal up to 64k of memory to a connected client or
4833 Thanks for Neel Mehta of Google Security for discovering this bug and to
4834 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4835 preparing the fix ([CVE-2014-0160])
4837 *Adam Langley, Bodo Moeller*
4839 * Fix for the attack described in the paper "Recovering OpenSSL
4840 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4841 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4842 <http://eprint.iacr.org/2014/140>
4844 Thanks to Yuval Yarom and Naomi Benger for discovering this
4845 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4847 *Yuval Yarom and Naomi Benger*
4849 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4850 this fixes a limitation in previous versions of OpenSSL.
4854 * Experimental encrypt-then-mac support.
4856 Experimental support for encrypt then mac from
4857 draft-gutmann-tls-encrypt-then-mac-02.txt
4859 To enable it set the appropriate extension number (0x42 for the test
4860 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4862 For non-compliant peers (i.e. just about everything) this should have no
4865 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4869 * Add EVP support for key wrapping algorithms, to avoid problems with
4870 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4871 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4872 algorithms and include tests cases.
4876 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4881 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4882 MGF1 digest and OAEP label.
4886 * Make openssl verify return errors.
4888 *Chris Palmer <palmer@google.com> and Ben Laurie*
4890 * New function ASN1_TIME_diff to calculate the difference between two
4891 ASN1_TIME structures or one structure and the current time.
4895 * Update fips_test_suite to support multiple command line options. New
4896 test to induce all self test errors in sequence and check expected
4901 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4902 sign or verify all in one operation.
4906 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4907 test programs and fips_test_suite. Includes functionality to parse
4908 the minimal script output of fipsalgest.pl directly.
4912 * Add authorisation parameter to FIPS_module_mode_set().
4916 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4920 * Use separate DRBG fields for internal and external flags. New function
4921 FIPS_drbg_health_check() to perform on demand health checking. Add
4922 generation tests to fips_test_suite with reduced health check interval to
4923 demonstrate periodic health checking. Add "nodh" option to
4924 fips_test_suite to skip very slow DH test.
4928 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4933 * More extensive health check for DRBG checking many more failure modes.
4934 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4935 combination: call this in fips_test_suite.
4939 * Add support for canonical generation of DSA parameter 'g'. See
4942 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4943 POST to handle HMAC cases.
4947 * Add functions FIPS_module_version() and FIPS_module_version_text()
4948 to return numerical and string versions of the FIPS module number.
4952 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4953 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4954 outside the validated module in the FIPS capable OpenSSL.
4958 * Minor change to DRBG entropy callback semantics. In some cases
4959 there is no multiple of the block length between min_len and
4960 max_len. Allow the callback to return more than max_len bytes
4961 of entropy but discard any extra: it is the callback's responsibility
4962 to ensure that the extra data discarded does not impact the
4963 requested amount of entropy.
4967 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4968 information in FIPS186-3, SP800-57 and SP800-131A.
4972 * CCM support via EVP. Interface is very similar to GCM case except we
4973 must supply all data in one chunk (i.e. no update, final) and the
4974 message length must be supplied if AAD is used. Add algorithm test
4979 * Initial version of POST overhaul. Add POST callback to allow the status
4980 of POST to be monitored and/or failures induced. Modify fips_test_suite
4981 to use callback. Always run all selftests even if one fails.
4985 * XTS support including algorithm test driver in the fips_gcmtest program.
4986 Note: this does increase the maximum key length from 32 to 64 bytes but
4987 there should be no binary compatibility issues as existing applications
4988 will never use XTS mode.
4992 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4993 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4994 performs algorithm blocking for unapproved PRNG types. Also do not
4995 set PRNG type in FIPS_mode_set(): leave this to the application.
4996 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4997 the standard OpenSSL PRNG: set additional data to a date time vector.
5001 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5002 This shouldn't present any incompatibility problems because applications
5003 shouldn't be using these directly and any that are will need to rethink
5004 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5008 * Extensive self tests and health checking required by SP800-90 DRBG.
5009 Remove strength parameter from FIPS_drbg_instantiate and always
5010 instantiate at maximum supported strength.
5014 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5018 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5022 * New function DH_compute_key_padded() to compute a DH key and pad with
5023 leading zeroes if needed: this complies with SP800-56A et al.
5027 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5028 anything, incomplete, subject to change and largely untested at present.
5032 * Modify fipscanisteronly build option to only build the necessary object
5033 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5037 * Add experimental option FIPSSYMS to give all symbols in
5038 fipscanister.o and FIPS or fips prefix. This will avoid
5039 conflicts with future versions of OpenSSL. Add perl script
5040 util/fipsas.pl to preprocess assembly language source files
5041 and rename any affected symbols.
5045 * Add selftest checks and algorithm block of non-fips algorithms in
5046 FIPS mode. Remove DES2 from selftests.
5050 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5051 return internal method without any ENGINE dependencies. Add new
5052 tiny fips sign and verify functions.
5056 * New build option no-ec2m to disable characteristic 2 code.
5060 * New build option "fipscanisteronly". This only builds fipscanister.o
5061 and (currently) associated fips utilities. Uses the file Makefile.fips
5062 instead of Makefile.org as the prototype.
5066 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5067 Update fips_gcmtest to use IV generator.
5071 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5072 setting output buffer to NULL. The `*Final` function must be
5073 called although it will not retrieve any additional data. The tag
5074 can be set or retrieved with a ctrl. The IV length is by default 12
5075 bytes (96 bits) but can be set to an alternative value. If the IV
5076 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5081 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5082 underlying do_cipher function handles all cipher semantics itself
5083 including padding and finalisation. This is useful if (for example)
5084 an ENGINE cipher handles block padding itself. The behaviour of
5085 do_cipher is subtly changed if this flag is set: the return value
5086 is the number of characters written to the output buffer (zero is
5087 no longer an error code) or a negative error code. Also if the
5088 input buffer is NULL and length 0 finalisation should be performed.
5092 * If a candidate issuer certificate is already part of the constructed
5093 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5097 * Improve forward-security support: add functions
5099 void SSL_CTX_set_not_resumable_session_callback(
5100 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5101 void SSL_set_not_resumable_session_callback(
5102 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5104 for use by SSL/TLS servers; the callback function will be called whenever a
5105 new session is created, and gets to decide whether the session may be
5106 cached to make it resumable (return 0) or not (return 1). (As by the
5107 SSL/TLS protocol specifications, the session_id sent by the server will be
5108 empty to indicate that the session is not resumable; also, the server will
5109 not generate RFC 4507 (RFC 5077) session tickets.)
5111 A simple reasonable callback implementation is to return is_forward_secure.
5112 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5113 by the SSL/TLS server library, indicating whether it can provide forward
5116 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5118 * New -verify_name option in command line utilities to set verification
5123 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5124 Add CMAC pkey methods.
5128 * Experimental renegotiation in s_server -www mode. If the client
5129 browses /reneg connection is renegotiated. If /renegcert it is
5130 renegotiated requesting a certificate.
5134 * Add an "external" session cache for debugging purposes to s_server. This
5135 should help trace issues which normally are only apparent in deployed
5136 multi-process servers.
5140 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5141 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5142 BIO_set_cipher() and some obscure PEM functions were changed so they
5143 can now return an error. The RAND changes required a change to the
5144 RAND_METHOD structure.
5148 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5149 a gcc attribute to warn if the result of a function is ignored. This
5150 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5151 whose return value is often ignored.
5155 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5156 These allow SCTs (signed certificate timestamps) to be requested and
5157 validated when establishing a connection.
5159 *Rob Percival <robpercival@google.com>*
5164 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5166 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5167 used even when parsing explicit parameters, when loading a encoded key
5168 or calling `EC_GROUP_new_from_ecpkparameters()`/
5169 `EC_GROUP_new_from_ecparameters()`.
5170 This prevents bypass of security hardening and performance gains,
5171 especially for curves with specialized EC_METHODs.
5172 By default, if a key encoded with explicit parameters is loaded and later
5173 encoded, the output is still encoded with explicit parameters, even if
5174 internally a "named" EC_GROUP is used for computation.
5178 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5179 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5180 NULL. After this change, only the cofactor parameter can be NULL. It also
5181 does some minimal sanity checks on the passed order.
5186 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5187 An attack is simple, if the first CMS_recipientInfo is valid but the
5188 second CMS_recipientInfo is chosen ciphertext. If the second
5189 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5190 encryption key will be replaced by garbage, and the message cannot be
5191 decoded, but if the RSA decryption fails, the correct encryption key is
5192 used and the recipient will not notice the attack.
5193 As a work around for this potential attack the length of the decrypted
5194 key must be equal to the cipher default key length, in case the
5195 certificate is not given and all recipientInfo are tried out.
5196 The old behaviour can be re-enabled in the CMS code by setting the
5197 CMS_DEBUG_DECRYPT flag.
5202 * Document issue with installation paths in diverse Windows builds
5204 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5205 binaries and run-time config file.
5210 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5212 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5213 This changes the size when using the `genpkey` command when no size is given.
5214 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5215 generation commands to use 2048 bits by default.
5219 * Add FIPS support for Android Arm 64-bit
5221 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5222 Module in Version 2.0.10. For some reason, the corresponding target
5223 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5224 built with FIPS support on Android Arm 64-bit. This omission has been
5227 *Matthias St. Pierre*
5229 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5231 * 0-byte record padding oracle
5233 If an application encounters a fatal protocol error and then calls
5234 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5235 then OpenSSL can respond differently to the calling application if a 0 byte
5236 record is received with invalid padding compared to if a 0 byte record is
5237 received with an invalid MAC. If the application then behaves differently
5238 based on that in a way that is detectable to the remote peer, then this
5239 amounts to a padding oracle that could be used to decrypt data.
5241 In order for this to be exploitable "non-stitched" ciphersuites must be in
5242 use. Stitched ciphersuites are optimised implementations of certain
5243 commonly used ciphersuites. Also the application must call SSL_shutdown()
5244 twice even if a protocol error has occurred (applications should not do
5245 this but some do anyway).
5247 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5248 Aviram, with additional investigation by Steven Collison and Andrew
5249 Hourselt. It was reported to OpenSSL on 10th December 2018.
5254 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5258 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5260 * Microarchitecture timing vulnerability in ECC scalar multiplication
5262 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5263 shown to be vulnerable to a microarchitecture timing side channel attack.
5264 An attacker with sufficient access to mount local timing attacks during
5265 ECDSA signature generation could recover the private key.
5267 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5268 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5274 * Timing vulnerability in DSA signature generation
5276 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5277 timing side channel attack. An attacker could use variations in the signing
5278 algorithm to recover the private key.
5280 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5285 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5286 Module, accidentally introduced while backporting security fixes from the
5287 development branch and hindering the use of ECC in FIPS mode.
5291 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5293 * Client DoS due to large DH parameter
5295 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5296 malicious server can send a very large prime value to the client. This will
5297 cause the client to spend an unreasonably long period of time generating a
5298 key for this prime resulting in a hang until the client has finished. This
5299 could be exploited in a Denial Of Service attack.
5301 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5306 * Cache timing vulnerability in RSA Key Generation
5308 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5309 a cache timing side channel attack. An attacker with sufficient access to
5310 mount cache timing attacks during the RSA key generation process could
5311 recover the private key.
5313 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5314 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5319 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5320 parameter is no longer accepted, as it leads to a corrupt table. NULL
5321 pem_str is reserved for alias entries only.
5325 * Revert blinding in ECDSA sign and instead make problematic addition
5326 length-invariant. Switch even to fixed-length Montgomery multiplication.
5330 * Change generating and checking of primes so that the error rate of not
5331 being prime depends on the intended use based on the size of the input.
5332 For larger primes this will result in more rounds of Miller-Rabin.
5333 The maximal error rate for primes with more than 1080 bits is lowered
5336 *Kurt Roeckx, Annie Yousar*
5338 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5342 * Add blinding to ECDSA and DSA signatures to protect against side channel
5343 attacks discovered by Keegan Ryan (NCC Group).
5347 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5348 now allow empty (zero character) pass phrases.
5352 * Certificate time validation (X509_cmp_time) enforces stricter
5353 compliance with RFC 5280. Fractional seconds and timezone offsets
5354 are no longer allowed.
5358 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5360 * Constructed ASN.1 types with a recursive definition could exceed the stack
5362 Constructed ASN.1 types with a recursive definition (such as can be found
5363 in PKCS7) could eventually exceed the stack given malicious input with
5364 excessive recursion. This could result in a Denial Of Service attack. There
5365 are no such structures used within SSL/TLS that come from untrusted sources
5366 so this is considered safe.
5368 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5374 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5376 * Read/write after SSL object in error state
5378 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5379 mechanism. The intent was that if a fatal error occurred during a handshake
5380 then OpenSSL would move into the error state and would immediately fail if
5381 you attempted to continue the handshake. This works as designed for the
5382 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5383 SSL_connect()), however due to a bug it does not work correctly if
5384 SSL_read() or SSL_write() is called directly. In that scenario, if the
5385 handshake fails then a fatal error will be returned in the initial function
5386 call. If SSL_read()/SSL_write() is subsequently called by the application
5387 for the same SSL object then it will succeed and the data is passed without
5388 being decrypted/encrypted directly from the SSL/TLS record layer.
5390 In order to exploit this issue an application bug would have to be present
5391 that resulted in a call to SSL_read()/SSL_write() being issued after having
5392 already received a fatal error.
5394 This issue was reported to OpenSSL by David Benjamin (Google).
5399 * rsaz_1024_mul_avx2 overflow bug on x86_64
5401 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5402 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5403 Analysis suggests that attacks against RSA and DSA as a result of this
5404 defect would be very difficult to perform and are not believed likely.
5405 Attacks against DH1024 are considered just feasible, because most of the
5406 work necessary to deduce information about a private key may be performed
5407 offline. The amount of resources required for such an attack would be
5408 significant. However, for an attack on TLS to be meaningful, the server
5409 would have to share the DH1024 private key among multiple clients, which is
5410 no longer an option since CVE-2016-0701.
5412 This only affects processors that support the AVX2 but not ADX extensions
5413 like Intel Haswell (4th generation).
5415 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5416 was originally found via the OSS-Fuzz project.
5421 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5423 * bn_sqrx8x_internal carry bug on x86_64
5425 There is a carry propagating bug in the x86_64 Montgomery squaring
5426 procedure. No EC algorithms are affected. Analysis suggests that attacks
5427 against RSA and DSA as a result of this defect would be very difficult to
5428 perform and are not believed likely. Attacks against DH are considered just
5429 feasible (although very difficult) because most of the work necessary to
5430 deduce information about a private key may be performed offline. The amount
5431 of resources required for such an attack would be very significant and
5432 likely only accessible to a limited number of attackers. An attacker would
5433 additionally need online access to an unpatched system using the target
5434 private key in a scenario with persistent DH parameters and a private
5435 key that is shared between multiple clients.
5437 This only affects processors that support the BMI1, BMI2 and ADX extensions
5438 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5440 This issue was reported to OpenSSL by the OSS-Fuzz project.
5445 * Malformed X.509 IPAddressFamily could cause OOB read
5447 If an X.509 certificate has a malformed IPAddressFamily extension,
5448 OpenSSL could do a one-byte buffer overread. The most likely result
5449 would be an erroneous display of the certificate in text format.
5451 This issue was reported to OpenSSL by the OSS-Fuzz project.
5455 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5457 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5458 platform rather than 'mingw'.
5462 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5464 * Truncated packet could crash via OOB read
5466 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5467 cipher is being used, then a truncated packet can cause that host to
5468 perform an out-of-bounds read, usually resulting in a crash.
5470 This issue was reported to OpenSSL by Robert Święcki of Google.
5475 * BN_mod_exp may produce incorrect results on x86_64
5477 There is a carry propagating bug in the x86_64 Montgomery squaring
5478 procedure. No EC algorithms are affected. Analysis suggests that attacks
5479 against RSA and DSA as a result of this defect would be very difficult to
5480 perform and are not believed likely. Attacks against DH are considered just
5481 feasible (although very difficult) because most of the work necessary to
5482 deduce information about a private key may be performed offline. The amount
5483 of resources required for such an attack would be very significant and
5484 likely only accessible to a limited number of attackers. An attacker would
5485 additionally need online access to an unpatched system using the target
5486 private key in a scenario with persistent DH parameters and a private
5487 key that is shared between multiple clients. For example this can occur by
5488 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5489 similar to CVE-2015-3193 but must be treated as a separate problem.
5491 This issue was reported to OpenSSL by the OSS-Fuzz project.
5496 * Montgomery multiplication may produce incorrect results
5498 There is a carry propagating bug in the Broadwell-specific Montgomery
5499 multiplication procedure that handles input lengths divisible by, but
5500 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5501 and DH private keys are impossible. This is because the subroutine in
5502 question is not used in operations with the private key itself and an input
5503 of the attacker's direct choice. Otherwise the bug can manifest itself as
5504 transient authentication and key negotiation failures or reproducible
5505 erroneous outcome of public-key operations with specially crafted input.
5506 Among EC algorithms only Brainpool P-512 curves are affected and one
5507 presumably can attack ECDH key negotiation. Impact was not analyzed in
5508 detail, because pre-requisites for attack are considered unlikely. Namely
5509 multiple clients have to choose the curve in question and the server has to
5510 share the private key among them, neither of which is default behaviour.
5511 Even then only clients that chose the curve will be affected.
5513 This issue was publicly reported as transient failures and was not
5514 initially recognized as a security issue. Thanks to Richard Morgan for
5515 providing reproducible case.
5520 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5521 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5522 prevent issues where no progress is being made and the peer continually
5523 sends unrecognised record types, using up resources processing them.
5527 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5529 * Missing CRL sanity check
5531 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5532 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5533 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5535 This issue only affects the OpenSSL 1.0.2i
5540 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5542 * OCSP Status Request extension unbounded memory growth
5544 A malicious client can send an excessively large OCSP Status Request
5545 extension. If that client continually requests renegotiation, sending a
5546 large OCSP Status Request extension each time, then there will be unbounded
5547 memory growth on the server. This will eventually lead to a Denial Of
5548 Service attack through memory exhaustion. Servers with a default
5549 configuration are vulnerable even if they do not support OCSP. Builds using
5550 the "no-ocsp" build time option are not affected.
5552 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5557 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5560 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5566 * OOB write in MDC2_Update()
5568 An overflow can occur in MDC2_Update() either if called directly or
5569 through the EVP_DigestUpdate() function using MDC2. If an attacker
5570 is able to supply very large amounts of input data after a previous
5571 call to EVP_EncryptUpdate() with a partial block then a length check
5572 can overflow resulting in a heap corruption.
5574 The amount of data needed is comparable to SIZE_MAX which is impractical
5577 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5582 * Malformed SHA512 ticket DoS
5584 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5585 DoS attack where a malformed ticket will result in an OOB read which will
5588 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5589 a custom server callback and ticket lookup mechanism.
5591 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5596 * OOB write in BN_bn2dec()
5598 The function BN_bn2dec() does not check the return value of BN_div_word().
5599 This can cause an OOB write if an application uses this function with an
5600 overly large BIGNUM. This could be a problem if an overly large certificate
5601 or CRL is printed out from an untrusted source. TLS is not affected because
5602 record limits will reject an oversized certificate before it is parsed.
5604 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5609 * OOB read in TS_OBJ_print_bio()
5611 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5612 the total length the OID text representation would use and not the amount
5613 of data written. This will result in OOB reads when large OIDs are
5616 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5621 * Pointer arithmetic undefined behaviour
5623 Avoid some undefined pointer arithmetic
5625 A common idiom in the codebase is to check limits in the following manner:
5628 Where "p" points to some malloc'd data of SIZE bytes and
5631 "len" here could be from some externally supplied data (e.g. from a TLS
5634 The rules of C pointer arithmetic are such that "p + len" is only well
5635 defined where len <= SIZE. Therefore the above idiom is actually
5636 undefined behaviour.
5638 For example this could cause problems if some malloc implementation
5639 provides an address for "p" such that "p + len" actually overflows for
5640 values of len that are too big and therefore p + len < limit.
5642 This issue was reported to OpenSSL by Guido Vranken
5647 * Constant time flag not preserved in DSA signing
5649 Operations in the DSA signing algorithm should run in constant time in
5650 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5651 implementation means that a non-constant time codepath is followed for
5652 certain operations. This has been demonstrated through a cache-timing
5653 attack to be sufficient for an attacker to recover the private DSA key.
5655 This issue was reported by César Pereida (Aalto University), Billy Brumley
5656 (Tampere University of Technology), and Yuval Yarom (The University of
5657 Adelaide and NICTA).
5662 * DTLS buffered message DoS
5664 In a DTLS connection where handshake messages are delivered out-of-order
5665 those messages that OpenSSL is not yet ready to process will be buffered
5666 for later use. Under certain circumstances, a flaw in the logic means that
5667 those messages do not get removed from the buffer even though the handshake
5668 has been completed. An attacker could force up to approx. 15 messages to
5669 remain in the buffer when they are no longer required. These messages will
5670 be cleared when the DTLS connection is closed. The default maximum size for
5671 a message is 100k. Therefore, the attacker could force an additional 1500k
5672 to be consumed per connection. By opening many simultaneous connections an
5673 attacker could cause a DoS attack through memory exhaustion.
5675 This issue was reported to OpenSSL by Quan Luo.
5680 * DTLS replay protection DoS
5682 A flaw in the DTLS replay attack protection mechanism means that records
5683 that arrive for future epochs update the replay protection "window" before
5684 the MAC for the record has been validated. This could be exploited by an
5685 attacker by sending a record for the next epoch (which does not have to
5686 decrypt or have a valid MAC), with a very large sequence number. This means
5687 that all subsequent legitimate packets are dropped causing a denial of
5688 service for a specific DTLS connection.
5690 This issue was reported to OpenSSL by the OCAP audit team.
5695 * Certificate message OOB reads
5697 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5698 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5699 theoretical DoS risk but this has not been observed in practice on common
5702 The messages affected are client certificate, client certificate request
5703 and server certificate. As a result the attack can only be performed
5704 against a client or a server which enables client authentication.
5706 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5711 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5713 * Prevent padding oracle in AES-NI CBC MAC check
5715 A MITM attacker can use a padding oracle attack to decrypt traffic
5716 when the connection uses an AES CBC cipher and the server support
5719 This issue was introduced as part of the fix for Lucky 13 padding
5720 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5721 constant time by making sure that always the same bytes are read and
5722 compared against either the MAC or padding bytes. But it no longer
5723 checked that there was enough data to have both the MAC and padding
5726 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5730 * Fix EVP_EncodeUpdate overflow
5732 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5733 Base64 encoding of binary data. If an attacker is able to supply very large
5734 amounts of input data then a length check can overflow resulting in a heap
5737 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5738 the `PEM_write_bio*` family of functions. These are mainly used within the
5739 OpenSSL command line applications, so any application which processes data
5740 from an untrusted source and outputs it as a PEM file should be considered
5741 vulnerable to this issue. User applications that call these APIs directly
5742 with large amounts of untrusted data may also be vulnerable.
5744 This issue was reported by Guido Vranken.
5749 * Fix EVP_EncryptUpdate overflow
5751 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5752 is able to supply very large amounts of input data after a previous call to
5753 EVP_EncryptUpdate() with a partial block then a length check can overflow
5754 resulting in a heap corruption. Following an analysis of all OpenSSL
5755 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5756 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5757 the first called function after an EVP_EncryptInit(), and therefore that
5758 specific call must be safe. The second form is where the length passed to
5759 EVP_EncryptUpdate() can be seen from the code to be some small value and
5760 therefore there is no possibility of an overflow. Since all instances are
5761 one of these two forms, it is believed that there can be no overflows in
5762 internal code due to this problem. It should be noted that
5763 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5764 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5765 of these calls have also been analysed too and it is believed there are no
5766 instances in internal usage where an overflow could occur.
5768 This issue was reported by Guido Vranken.
5773 * Prevent ASN.1 BIO excessive memory allocation
5775 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5776 a short invalid encoding can cause allocation of large amounts of memory
5777 potentially consuming excessive resources or exhausting memory.
5779 Any application parsing untrusted data through d2i BIO functions is
5780 affected. The memory based functions such as d2i_X509() are *not* affected.
5781 Since the memory based functions are used by the TLS library, TLS
5782 applications are not affected.
5784 This issue was reported by Brian Carpenter.
5791 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5792 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5793 in arbitrary stack data being returned in the buffer.
5795 This issue was reported by Guido Vranken.
5800 * Modify behavior of ALPN to invoke callback after SNI/servername
5801 callback, such that updates to the SSL_CTX affect ALPN.
5805 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5810 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5811 methods are enabled and ssl2 is disabled the methods return NULL.
5815 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5817 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5818 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5819 provide any "EXPORT" or "LOW" strength ciphers.
5823 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5824 is by default disabled at build-time. Builds that are not configured with
5825 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5826 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5827 will need to explicitly call either of:
5829 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5831 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5833 as appropriate. Even if either of those is used, or the application
5834 explicitly uses the version-specific SSLv2_method() or its client and
5835 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5836 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5837 ciphers, and SSLv2 56-bit DES are no longer available.
5842 * Fix a double-free in DSA code
5844 A double free bug was discovered when OpenSSL parses malformed DSA private
5845 keys and could lead to a DoS attack or memory corruption for applications
5846 that receive DSA private keys from untrusted sources. This scenario is
5849 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5855 * Disable SRP fake user seed to address a server memory leak.
5857 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5859 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5860 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5861 was changed to ignore the "fake user" SRP seed, even if the seed
5864 Users should use SRP_VBASE_get1_by_user instead. Note that in
5865 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5866 also that even though configuring the SRP seed attempts to hide
5867 invalid usernames by continuing the handshake with fake
5868 credentials, this behaviour is not constant time and no strong
5869 guarantees are made that the handshake is indistinguishable from
5870 that of a valid user.
5875 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5877 In the BN_hex2bn function the number of hex digits is calculated using an
5878 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5879 large values of `i` this can result in `bn_expand` not allocating any
5880 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5881 field as NULL leading to a subsequent NULL ptr deref. For very large values
5882 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5883 In this case memory is allocated to the internal BIGNUM data field, but it
5884 is insufficiently sized leading to heap corruption. A similar issue exists
5885 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5886 is ever called by user applications with very large untrusted hex/dec data.
5887 This is anticipated to be a rare occurrence.
5889 All OpenSSL internal usage of these functions use data that is not expected
5890 to be untrusted, e.g. config file data or application command line
5891 arguments. If user developed applications generate config file data based
5892 on untrusted data then it is possible that this could also lead to security
5893 consequences. This is also anticipated to be rare.
5895 This issue was reported to OpenSSL by Guido Vranken.
5900 * Fix memory issues in `BIO_*printf` functions
5902 The internal `fmtstr` function used in processing a "%s" format string in
5903 the `BIO_*printf` functions could overflow while calculating the length of a
5904 string and cause an OOB read when printing very long strings.
5906 Additionally the internal `doapr_outch` function can attempt to write to an
5907 OOB memory location (at an offset from the NULL pointer) in the event of a
5908 memory allocation failure. In 1.0.2 and below this could be caused where
5909 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5910 could be in processing a very long "%s" format string. Memory leaks can
5913 The first issue may mask the second issue dependent on compiler behaviour.
5914 These problems could enable attacks where large amounts of untrusted data
5915 is passed to the `BIO_*printf` functions. If applications use these functions
5916 in this way then they could be vulnerable. OpenSSL itself uses these
5917 functions when printing out human-readable dumps of ASN.1 data. Therefore
5918 applications that print this data could be vulnerable if the data is from
5919 untrusted sources. OpenSSL command line applications could also be
5920 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5921 as command line arguments.
5923 Libssl is not considered directly vulnerable. Additionally certificates etc
5924 received via remote connections via libssl are also unlikely to be able to
5925 trigger these issues because of message size limits enforced within libssl.
5927 This issue was reported to OpenSSL Guido Vranken.
5932 * Side channel attack on modular exponentiation
5934 A side-channel attack was found which makes use of cache-bank conflicts on
5935 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5936 of RSA keys. The ability to exploit this issue is limited as it relies on
5937 an attacker who has control of code in a thread running on the same
5938 hyper-threaded core as the victim thread which is performing decryptions.
5940 This issue was reported to OpenSSL by Yuval Yarom, The University of
5941 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5942 Nadia Heninger, University of Pennsylvania with more information at
5943 <http://cachebleed.info>.
5948 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5949 if no keysize is specified with default_bits. This fixes an
5950 omission in an earlier change that changed all RSA/DSA key generation
5951 commands to use 2048 bits by default.
5955 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5957 * DH small subgroups
5959 Historically OpenSSL only ever generated DH parameters based on "safe"
5960 primes. More recently (in version 1.0.2) support was provided for
5961 generating X9.42 style parameter files such as those required for RFC 5114
5962 support. The primes used in such files may not be "safe". Where an
5963 application is using DH configured with parameters based on primes that are
5964 not "safe" then an attacker could use this fact to find a peer's private
5965 DH exponent. This attack requires that the attacker complete multiple
5966 handshakes in which the peer uses the same private DH exponent. For example
5967 this could be used to discover a TLS server's private DH exponent if it's
5968 reusing the private DH exponent or it's using a static DH ciphersuite.
5970 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5971 TLS. It is not on by default. If the option is not set then the server
5972 reuses the same private DH exponent for the life of the server process and
5973 would be vulnerable to this attack. It is believed that many popular
5974 applications do set this option and would therefore not be at risk.
5976 The fix for this issue adds an additional check where a "q" parameter is
5977 available (as is the case in X9.42 based parameters). This detects the
5978 only known attack, and is the only possible defense for static DH
5979 ciphersuites. This could have some performance impact.
5981 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5982 default and cannot be disabled. This could have some performance impact.
5984 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5989 * SSLv2 doesn't block disabled ciphers
5991 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5992 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5993 been disabled, provided that the SSLv2 protocol was not also disabled via
5996 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5997 and Sebastian Schinzel.
6002 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6004 * BN_mod_exp may produce incorrect results on x86_64
6006 There is a carry propagating bug in the x86_64 Montgomery squaring
6007 procedure. No EC algorithms are affected. Analysis suggests that attacks
6008 against RSA and DSA as a result of this defect would be very difficult to
6009 perform and are not believed likely. Attacks against DH are considered just
6010 feasible (although very difficult) because most of the work necessary to
6011 deduce information about a private key may be performed offline. The amount
6012 of resources required for such an attack would be very significant and
6013 likely only accessible to a limited number of attackers. An attacker would
6014 additionally need online access to an unpatched system using the target
6015 private key in a scenario with persistent DH parameters and a private
6016 key that is shared between multiple clients. For example this can occur by
6017 default in OpenSSL DHE based SSL/TLS ciphersuites.
6019 This issue was reported to OpenSSL by Hanno Böck.
6024 * Certificate verify crash with missing PSS parameter
6026 The signature verification routines will crash with a NULL pointer
6027 dereference if presented with an ASN.1 signature using the RSA PSS
6028 algorithm and absent mask generation function parameter. Since these
6029 routines are used to verify certificate signature algorithms this can be
6030 used to crash any certificate verification operation and exploited in a
6031 DoS attack. Any application which performs certificate verification is
6032 vulnerable including OpenSSL clients and servers which enable client
6035 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6040 * X509_ATTRIBUTE memory leak
6042 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6043 memory. This structure is used by the PKCS#7 and CMS routines so any
6044 application which reads PKCS#7 or CMS data from untrusted sources is
6045 affected. SSL/TLS is not affected.
6047 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6053 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6054 This changes the decoding behaviour for some invalid messages,
6055 though the change is mostly in the more lenient direction, and
6056 legacy behaviour is preserved as much as possible.
6060 * In DSA_generate_parameters_ex, if the provided seed is too short,
6063 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6065 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6067 * Alternate chains certificate forgery
6069 During certificate verification, OpenSSL will attempt to find an
6070 alternative certificate chain if the first attempt to build such a chain
6071 fails. An error in the implementation of this logic can mean that an
6072 attacker could cause certain checks on untrusted certificates to be
6073 bypassed, such as the CA flag, enabling them to use a valid leaf
6074 certificate to act as a CA and "issue" an invalid certificate.
6076 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6081 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6083 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6084 incompatibility in the handling of HMAC. The previous ABI has now been
6089 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6091 * Malformed ECParameters causes infinite loop
6093 When processing an ECParameters structure OpenSSL enters an infinite loop
6094 if the curve specified is over a specially malformed binary polynomial
6097 This can be used to perform denial of service against any
6098 system which processes public keys, certificate requests or
6099 certificates. This includes TLS clients and TLS servers with
6100 client authentication enabled.
6102 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6107 * Exploitable out-of-bounds read in X509_cmp_time
6109 X509_cmp_time does not properly check the length of the ASN1_TIME
6110 string and can read a few bytes out of bounds. In addition,
6111 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6114 An attacker can use this to craft malformed certificates and CRLs of
6115 various sizes and potentially cause a segmentation fault, resulting in
6116 a DoS on applications that verify certificates or CRLs. TLS clients
6117 that verify CRLs are affected. TLS clients and servers with client
6118 authentication enabled may be affected if they use custom verification
6121 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6122 independently by Hanno Böck.
6127 * PKCS7 crash with missing EnvelopedContent
6129 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6130 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6131 with missing content and trigger a NULL pointer dereference on parsing.
6133 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6134 structures from untrusted sources are affected. OpenSSL clients and
6135 servers are not affected.
6137 This issue was reported to OpenSSL by Michal Zalewski (Google).
6142 * CMS verify infinite loop with unknown hash function
6144 When verifying a signedData message the CMS code can enter an infinite loop
6145 if presented with an unknown hash function OID. This can be used to perform
6146 denial of service against any system which verifies signedData messages using
6148 This issue was reported to OpenSSL by Johannes Bauer.
6153 * Race condition handling NewSessionTicket
6155 If a NewSessionTicket is received by a multi-threaded client when attempting to
6156 reuse a previous ticket then a race condition can occur potentially leading to
6157 a double free of the ticket data.
6162 * Only support 256-bit or stronger elliptic curves with the
6163 'ecdh_auto' setting (server) or by default (client). Of supported
6164 curves, prefer P-256 (both).
6168 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6170 * ClientHello sigalgs DoS fix
6172 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6173 invalid signature algorithms extension a NULL pointer dereference will
6174 occur. This can be exploited in a DoS attack against the server.
6176 This issue was was reported to OpenSSL by David Ramos of Stanford
6180 *Stephen Henson and Matt Caswell*
6182 * Multiblock corrupted pointer fix
6184 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6185 feature only applies on 64 bit x86 architecture platforms that support AES
6186 NI instructions. A defect in the implementation of "multiblock" can cause
6187 OpenSSL's internal write buffer to become incorrectly set to NULL when
6188 using non-blocking IO. Typically, when the user application is using a
6189 socket BIO for writing, this will only result in a failed connection.
6190 However if some other BIO is used then it is likely that a segmentation
6191 fault will be triggered, thus enabling a potential DoS attack.
6193 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6198 * Segmentation fault in DTLSv1_listen fix
6200 The DTLSv1_listen function is intended to be stateless and processes the
6201 initial ClientHello from many peers. It is common for user code to loop
6202 over the call to DTLSv1_listen until a valid ClientHello is received with
6203 an associated cookie. A defect in the implementation of DTLSv1_listen means
6204 that state is preserved in the SSL object from one invocation to the next
6205 that can lead to a segmentation fault. Errors processing the initial
6206 ClientHello can trigger this scenario. An example of such an error could be
6207 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6210 This issue was reported to OpenSSL by Per Allansson.
6215 * Segmentation fault in ASN1_TYPE_cmp fix
6217 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6218 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6219 certificate signature algorithm consistency this can be used to crash any
6220 certificate verification operation and exploited in a DoS attack. Any
6221 application which performs certificate verification is vulnerable including
6222 OpenSSL clients and servers which enable client authentication.
6227 * Segmentation fault for invalid PSS parameters fix
6229 The signature verification routines will crash with a NULL pointer
6230 dereference if presented with an ASN.1 signature using the RSA PSS
6231 algorithm and invalid parameters. Since these routines are used to verify
6232 certificate signature algorithms this can be used to crash any
6233 certificate verification operation and exploited in a DoS attack. Any
6234 application which performs certificate verification is vulnerable including
6235 OpenSSL clients and servers which enable client authentication.
6237 This issue was was reported to OpenSSL by Brian Carpenter.
6242 * ASN.1 structure reuse memory corruption fix
6244 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6245 memory corruption via an invalid write. Such reuse is and has been
6246 strongly discouraged and is believed to be rare.
6248 Applications that parse structures containing CHOICE or ANY DEFINED BY
6249 components may be affected. Certificate parsing (d2i_X509 and related
6250 functions) are however not affected. OpenSSL clients and servers are
6256 * PKCS7 NULL pointer dereferences fix
6258 The PKCS#7 parsing code does not handle missing outer ContentInfo
6259 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6260 missing content and trigger a NULL pointer dereference on parsing.
6262 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6263 otherwise parse PKCS#7 structures from untrusted sources are
6264 affected. OpenSSL clients and servers are not affected.
6266 This issue was reported to OpenSSL by Michal Zalewski (Google).
6271 * DoS via reachable assert in SSLv2 servers fix
6273 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6274 servers that both support SSLv2 and enable export cipher suites by sending
6275 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6277 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6278 (OpenSSL development team).
6283 * Empty CKE with client auth and DHE fix
6285 If client auth is used then a server can seg fault in the event of a DHE
6286 ciphersuite being selected and a zero length ClientKeyExchange message
6287 being sent by the client. This could be exploited in a DoS attack.
6292 * Handshake with unseeded PRNG fix
6294 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6295 with an unseeded PRNG. The conditions are:
6296 - The client is on a platform where the PRNG has not been seeded
6297 automatically, and the user has not seeded manually
6298 - A protocol specific client method version has been used (i.e. not
6299 SSL_client_methodv23)
6300 - A ciphersuite is used that does not require additional random data from
6301 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6303 If the handshake succeeds then the client random that has been used will
6304 have been generated from a PRNG with insufficient entropy and therefore the
6305 output may be predictable.
6307 For example using the following command with an unseeded openssl will
6308 succeed on an unpatched platform:
6310 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6315 * Use After Free following d2i_ECPrivatekey error fix
6317 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6318 could cause a use after free condition. This, in turn, could cause a double
6319 free in several private key parsing functions (such as d2i_PrivateKey
6320 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6321 for applications that receive EC private keys from untrusted
6322 sources. This scenario is considered rare.
6324 This issue was discovered by the BoringSSL project and fixed in their
6330 * X509_to_X509_REQ NULL pointer deref fix
6332 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6333 the certificate key is invalid. This function is rarely used in practice.
6335 This issue was discovered by Brian Carpenter.
6340 * Removed the export ciphers from the DEFAULT ciphers
6344 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6346 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6347 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6348 So far those who have to target multiple platforms would compromise
6349 and argue that binary targeting say ARMv5 would still execute on
6350 ARMv8. "Universal" build resolves this compromise by providing
6351 near-optimal performance even on newer platforms.
6355 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6356 (other platforms pending).
6358 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6360 * Add support for the SignedCertificateTimestampList certificate and
6361 OCSP response extensions from RFC6962.
6365 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6366 for corner cases. (Certain input points at infinity could lead to
6367 bogus results, with non-infinity inputs mapped to infinity too.)
6371 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6372 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6373 common cases are optimized and there still is room for further
6374 improvements. Vector Permutation AES for Altivec is also added.
6378 * Add support for little-endian ppc64 Linux target.
6380 *Marcelo Cerri (IBM)*
6382 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6383 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6384 are optimized and there still is room for further improvements.
6385 Both 32- and 64-bit modes are supported.
6387 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6389 * Improved ARMv7 NEON support.
6393 * Support for SPARC Architecture 2011 crypto extensions, first
6394 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6395 SHA256/512, MD5, GHASH and modular exponentiation.
6397 *Andy Polyakov, David Miller*
6399 * Accelerated modular exponentiation for Intel processors, a.k.a.
6402 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6404 * Support for new and upcoming Intel processors, including AVX2,
6405 BMI and SHA ISA extensions. This includes additional "stitched"
6406 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6409 This work was sponsored by Intel Corp.
6413 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6414 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6415 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6419 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6420 this fixes a limitation in previous versions of OpenSSL.
6424 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6425 MGF1 digest and OAEP label.
6429 * Add EVP support for key wrapping algorithms, to avoid problems with
6430 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6431 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6432 algorithms and include tests cases.
6436 * Add functions to allocate and set the fields of an ECDSA_METHOD
6439 *Douglas E. Engert, Steve Henson*
6441 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6442 difference in days and seconds between two tm or ASN1_TIME structures.
6446 * Add -rev test option to s_server to just reverse order of characters
6447 received by client and send back to server. Also prints an abbreviated
6448 summary of the connection parameters.
6452 * New option -brief for s_client and s_server to print out a brief summary
6453 of connection parameters.
6457 * Add callbacks for arbitrary TLS extensions.
6459 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6461 * New option -crl_download in several openssl utilities to download CRLs
6462 from CRLDP extension in certificates.
6466 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6470 * New function X509_CRL_diff to generate a delta CRL from the difference
6471 of two full CRLs. Add support to "crl" utility.
6475 * New functions to set lookup_crls function and to retrieve
6476 X509_STORE from X509_STORE_CTX.
6480 * Print out deprecated issuer and subject unique ID fields in
6485 * Extend OCSP I/O functions so they can be used for simple general purpose
6486 HTTP as well as OCSP. New wrapper function which can be used to download
6487 CRLs using the OCSP API.
6491 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6495 * `SSL_CONF*` functions. These provide a common framework for application
6496 configuration using configuration files or command lines.
6500 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6501 message callback and prints the results. Needs compile time option
6502 "enable-ssl-trace". New options to s_client and s_server to enable
6507 * New ctrl and macro to retrieve supported points extensions.
6508 Print out extension in s_server and s_client.
6512 * New functions to retrieve certificate signature and signature
6517 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6522 * New Suite B modes for TLS code. These use and enforce the requirements
6523 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6524 only use Suite B curves. The Suite B modes can be set by using the
6525 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6529 * New chain verification flags for Suite B levels of security. Check
6530 algorithms are acceptable when flags are set in X509_verify_cert.
6534 * Make tls1_check_chain return a set of flags indicating checks passed
6535 by a certificate chain. Add additional tests to handle client
6536 certificates: checks for matching certificate type and issuer name
6541 * If an attempt is made to use a signature algorithm not in the peer
6542 preference list abort the handshake. If client has no suitable
6543 signature algorithms in response to a certificate request do not
6544 use the certificate.
6548 * If server EC tmp key is not in client preference list abort handshake.
6552 * Add support for certificate stores in CERT structure. This makes it
6553 possible to have different stores per SSL structure or one store in
6554 the parent SSL_CTX. Include distinct stores for certificate chain
6555 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6556 to build and store a certificate chain in CERT structure: returning
6557 an error if the chain cannot be built: this will allow applications
6558 to test if a chain is correctly configured.
6560 Note: if the CERT based stores are not set then the parent SSL_CTX
6561 store is used to retain compatibility with existing behaviour.
6565 * New function ssl_set_client_disabled to set a ciphersuite disabled
6566 mask based on the current session, check mask when sending client
6567 hello and checking the requested ciphersuite.
6571 * New ctrls to retrieve and set certificate types in a certificate
6572 request message. Print out received values in s_client. If certificate
6573 types is not set with custom values set sensible values based on
6574 supported signature algorithms.
6578 * Support for distinct client and server supported signature algorithms.
6582 * Add certificate callback. If set this is called whenever a certificate
6583 is required by client or server. An application can decide which
6584 certificate chain to present based on arbitrary criteria: for example
6585 supported signature algorithms. Add very simple example to s_server.
6586 This fixes many of the problems and restrictions of the existing client
6587 certificate callback: for example you can now clear an existing
6588 certificate and specify the whole chain.
6592 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6593 the certificate can be used for (if anything). Set valid_flags field
6594 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6595 to have similar checks in it.
6597 Add new "cert_flags" field to CERT structure and include a "strict mode".
6598 This enforces some TLS certificate requirements (such as only permitting
6599 certificate signature algorithms contained in the supported algorithms
6600 extension) which some implementations ignore: this option should be used
6601 with caution as it could cause interoperability issues.
6605 * Update and tidy signature algorithm extension processing. Work out
6606 shared signature algorithms based on preferences and peer algorithms
6607 and print them out in s_client and s_server. Abort handshake if no
6608 shared signature algorithms.
6612 * Add new functions to allow customised supported signature algorithms
6613 for SSL and SSL_CTX structures. Add options to s_client and s_server
6618 * New function SSL_certs_clear() to delete all references to certificates
6619 from an SSL structure. Before this once a certificate had been added
6620 it couldn't be removed.
6624 * Integrate hostname, email address and IP address checking with certificate
6625 verification. New verify options supporting checking in openssl utility.
6629 * Fixes and wildcard matching support to hostname and email checking
6630 functions. Add manual page.
6632 *Florian Weimer (Red Hat Product Security Team)*
6634 * New functions to check a hostname email or IP address against a
6635 certificate. Add options x509 utility to print results of checks against
6640 * Fix OCSP checking.
6642 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6644 * Initial experimental support for explicitly trusted non-root CAs.
6645 OpenSSL still tries to build a complete chain to a root but if an
6646 intermediate CA has a trust setting included that is used. The first
6647 setting is used: whether to trust (e.g., -addtrust option to the x509
6652 * Add -trusted_first option which attempts to find certificates in the
6653 trusted store even if an untrusted chain is also supplied.
6657 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6658 platform support for Linux and Android.
6662 * Support for linux-x32, ILP32 environment in x86_64 framework.
6666 * Experimental multi-implementation support for FIPS capable OpenSSL.
6667 When in FIPS mode the approved implementations are used as normal,
6668 when not in FIPS mode the internal unapproved versions are used instead.
6669 This means that the FIPS capable OpenSSL isn't forced to use the
6670 (often lower performance) FIPS implementations outside FIPS mode.
6674 * Transparently support X9.42 DH parameters when calling
6675 PEM_read_bio_DHparameters. This means existing applications can handle
6676 the new parameter format automatically.
6680 * Initial experimental support for X9.42 DH parameter format: mainly
6681 to support use of 'q' parameter for RFC5114 parameters.
6685 * Add DH parameters from RFC5114 including test data to dhtest.
6689 * Support for automatic EC temporary key parameter selection. If enabled
6690 the most preferred EC parameters are automatically used instead of
6691 hardcoded fixed parameters. Now a server just has to call:
6692 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6693 support ECDH and use the most appropriate parameters.
6697 * Enhance and tidy EC curve and point format TLS extension code. Use
6698 static structures instead of allocation if default values are used.
6699 New ctrls to set curves we wish to support and to retrieve shared curves.
6700 Print out shared curves in s_server. New options to s_server and s_client
6701 to set list of supported curves.
6705 * New ctrls to retrieve supported signature algorithms and
6706 supported curve values as an array of NIDs. Extend openssl utility
6707 to print out received values.
6711 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6712 between NIDs and the more common NIST names such as "P-256". Enhance
6713 ecparam utility and ECC method to recognise the NIST names for curves.
6717 * Enhance SSL/TLS certificate chain handling to support different
6718 chains for each certificate instead of one chain in the parent SSL_CTX.
6722 * Support for fixed DH ciphersuite client authentication: where both
6723 server and client use DH certificates with common parameters.
6727 * Support for fixed DH ciphersuites: those requiring DH server
6732 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6734 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6735 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6736 X509_CINF_get_signature were reverted post internal team review.
6741 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6743 * OCSP Status Request extension unbounded memory growth
6745 A malicious client can send an excessively large OCSP Status Request
6746 extension. If that client continually requests renegotiation, sending a
6747 large OCSP Status Request extension each time, then there will be unbounded
6748 memory growth on the server. This will eventually lead to a Denial Of
6749 Service attack through memory exhaustion. Servers with a default
6750 configuration are vulnerable even if they do not support OCSP. Builds using
6751 the "no-ocsp" build time option are not affected.
6753 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6758 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6761 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6767 * OOB write in MDC2_Update()
6769 An overflow can occur in MDC2_Update() either if called directly or
6770 through the EVP_DigestUpdate() function using MDC2. If an attacker
6771 is able to supply very large amounts of input data after a previous
6772 call to EVP_EncryptUpdate() with a partial block then a length check
6773 can overflow resulting in a heap corruption.
6775 The amount of data needed is comparable to SIZE_MAX which is impractical
6778 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6783 * Malformed SHA512 ticket DoS
6785 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6786 DoS attack where a malformed ticket will result in an OOB read which will
6789 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6790 a custom server callback and ticket lookup mechanism.
6792 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6797 * OOB write in BN_bn2dec()
6799 The function BN_bn2dec() does not check the return value of BN_div_word().
6800 This can cause an OOB write if an application uses this function with an
6801 overly large BIGNUM. This could be a problem if an overly large certificate
6802 or CRL is printed out from an untrusted source. TLS is not affected because
6803 record limits will reject an oversized certificate before it is parsed.
6805 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6810 * OOB read in TS_OBJ_print_bio()
6812 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6813 the total length the OID text representation would use and not the amount
6814 of data written. This will result in OOB reads when large OIDs are
6817 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6822 * Pointer arithmetic undefined behaviour
6824 Avoid some undefined pointer arithmetic
6826 A common idiom in the codebase is to check limits in the following manner:
6829 Where "p" points to some malloc'd data of SIZE bytes and
6832 "len" here could be from some externally supplied data (e.g. from a TLS
6835 The rules of C pointer arithmetic are such that "p + len" is only well
6836 defined where len <= SIZE. Therefore, the above idiom is actually
6837 undefined behaviour.
6839 For example this could cause problems if some malloc implementation
6840 provides an address for "p" such that "p + len" actually overflows for
6841 values of len that are too big and therefore p + len < limit.
6843 This issue was reported to OpenSSL by Guido Vranken
6848 * Constant time flag not preserved in DSA signing
6850 Operations in the DSA signing algorithm should run in constant time in
6851 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6852 implementation means that a non-constant time codepath is followed for
6853 certain operations. This has been demonstrated through a cache-timing
6854 attack to be sufficient for an attacker to recover the private DSA key.
6856 This issue was reported by César Pereida (Aalto University), Billy Brumley
6857 (Tampere University of Technology), and Yuval Yarom (The University of
6858 Adelaide and NICTA).
6863 * DTLS buffered message DoS
6865 In a DTLS connection where handshake messages are delivered out-of-order
6866 those messages that OpenSSL is not yet ready to process will be buffered
6867 for later use. Under certain circumstances, a flaw in the logic means that
6868 those messages do not get removed from the buffer even though the handshake
6869 has been completed. An attacker could force up to approx. 15 messages to
6870 remain in the buffer when they are no longer required. These messages will
6871 be cleared when the DTLS connection is closed. The default maximum size for
6872 a message is 100k. Therefore, the attacker could force an additional 1500k
6873 to be consumed per connection. By opening many simultaneous connections an
6874 attacker could cause a DoS attack through memory exhaustion.
6876 This issue was reported to OpenSSL by Quan Luo.
6881 * DTLS replay protection DoS
6883 A flaw in the DTLS replay attack protection mechanism means that records
6884 that arrive for future epochs update the replay protection "window" before
6885 the MAC for the record has been validated. This could be exploited by an
6886 attacker by sending a record for the next epoch (which does not have to
6887 decrypt or have a valid MAC), with a very large sequence number. This means
6888 that all subsequent legitimate packets are dropped causing a denial of
6889 service for a specific DTLS connection.
6891 This issue was reported to OpenSSL by the OCAP audit team.
6896 * Certificate message OOB reads
6898 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6899 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6900 theoretical DoS risk but this has not been observed in practice on common
6903 The messages affected are client certificate, client certificate request
6904 and server certificate. As a result the attack can only be performed
6905 against a client or a server which enables client authentication.
6907 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6912 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6914 * Prevent padding oracle in AES-NI CBC MAC check
6916 A MITM attacker can use a padding oracle attack to decrypt traffic
6917 when the connection uses an AES CBC cipher and the server support
6920 This issue was introduced as part of the fix for Lucky 13 padding
6921 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6922 constant time by making sure that always the same bytes are read and
6923 compared against either the MAC or padding bytes. But it no longer
6924 checked that there was enough data to have both the MAC and padding
6927 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6932 * Fix EVP_EncodeUpdate overflow
6934 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6935 Base64 encoding of binary data. If an attacker is able to supply very large
6936 amounts of input data then a length check can overflow resulting in a heap
6939 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6940 the `PEM_write_bio*` family of functions. These are mainly used within the
6941 OpenSSL command line applications, so any application which processes data
6942 from an untrusted source and outputs it as a PEM file should be considered
6943 vulnerable to this issue. User applications that call these APIs directly
6944 with large amounts of untrusted data may also be vulnerable.
6946 This issue was reported by Guido Vranken.
6951 * Fix EVP_EncryptUpdate overflow
6953 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6954 is able to supply very large amounts of input data after a previous call to
6955 EVP_EncryptUpdate() with a partial block then a length check can overflow
6956 resulting in a heap corruption. Following an analysis of all OpenSSL
6957 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6958 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6959 the first called function after an EVP_EncryptInit(), and therefore that
6960 specific call must be safe. The second form is where the length passed to
6961 EVP_EncryptUpdate() can be seen from the code to be some small value and
6962 therefore there is no possibility of an overflow. Since all instances are
6963 one of these two forms, it is believed that there can be no overflows in
6964 internal code due to this problem. It should be noted that
6965 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6966 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6967 of these calls have also been analysed too and it is believed there are no
6968 instances in internal usage where an overflow could occur.
6970 This issue was reported by Guido Vranken.
6975 * Prevent ASN.1 BIO excessive memory allocation
6977 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6978 a short invalid encoding can casuse allocation of large amounts of memory
6979 potentially consuming excessive resources or exhausting memory.
6981 Any application parsing untrusted data through d2i BIO functions is
6982 affected. The memory based functions such as d2i_X509() are *not* affected.
6983 Since the memory based functions are used by the TLS library, TLS
6984 applications are not affected.
6986 This issue was reported by Brian Carpenter.
6993 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6994 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6995 in arbitrary stack data being returned in the buffer.
6997 This issue was reported by Guido Vranken.
7002 * Modify behavior of ALPN to invoke callback after SNI/servername
7003 callback, such that updates to the SSL_CTX affect ALPN.
7007 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7012 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7013 methods are enabled and ssl2 is disabled the methods return NULL.
7017 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7019 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7020 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7021 provide any "EXPORT" or "LOW" strength ciphers.
7025 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7026 is by default disabled at build-time. Builds that are not configured with
7027 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7028 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7029 will need to explicitly call either of:
7031 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7033 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7035 as appropriate. Even if either of those is used, or the application
7036 explicitly uses the version-specific SSLv2_method() or its client and
7037 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7038 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7039 ciphers, and SSLv2 56-bit DES are no longer available.
7044 * Fix a double-free in DSA code
7046 A double free bug was discovered when OpenSSL parses malformed DSA private
7047 keys and could lead to a DoS attack or memory corruption for applications
7048 that receive DSA private keys from untrusted sources. This scenario is
7051 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7057 * Disable SRP fake user seed to address a server memory leak.
7059 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7061 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7062 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7063 was changed to ignore the "fake user" SRP seed, even if the seed
7066 Users should use SRP_VBASE_get1_by_user instead. Note that in
7067 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7068 also that even though configuring the SRP seed attempts to hide
7069 invalid usernames by continuing the handshake with fake
7070 credentials, this behaviour is not constant time and no strong
7071 guarantees are made that the handshake is indistinguishable from
7072 that of a valid user.
7077 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7079 In the BN_hex2bn function the number of hex digits is calculated using an
7080 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7081 large values of `i` this can result in `bn_expand` not allocating any
7082 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7083 field as NULL leading to a subsequent NULL ptr deref. For very large values
7084 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7085 In this case memory is allocated to the internal BIGNUM data field, but it
7086 is insufficiently sized leading to heap corruption. A similar issue exists
7087 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7088 is ever called by user applications with very large untrusted hex/dec data.
7089 This is anticipated to be a rare occurrence.
7091 All OpenSSL internal usage of these functions use data that is not expected
7092 to be untrusted, e.g. config file data or application command line
7093 arguments. If user developed applications generate config file data based
7094 on untrusted data then it is possible that this could also lead to security
7095 consequences. This is also anticipated to be rare.
7097 This issue was reported to OpenSSL by Guido Vranken.
7102 * Fix memory issues in `BIO_*printf` functions
7104 The internal `fmtstr` function used in processing a "%s" format string in
7105 the `BIO_*printf` functions could overflow while calculating the length of a
7106 string and cause an OOB read when printing very long strings.
7108 Additionally the internal `doapr_outch` function can attempt to write to an
7109 OOB memory location (at an offset from the NULL pointer) in the event of a
7110 memory allocation failure. In 1.0.2 and below this could be caused where
7111 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7112 could be in processing a very long "%s" format string. Memory leaks can
7115 The first issue may mask the second issue dependent on compiler behaviour.
7116 These problems could enable attacks where large amounts of untrusted data
7117 is passed to the `BIO_*printf` functions. If applications use these functions
7118 in this way then they could be vulnerable. OpenSSL itself uses these
7119 functions when printing out human-readable dumps of ASN.1 data. Therefore
7120 applications that print this data could be vulnerable if the data is from
7121 untrusted sources. OpenSSL command line applications could also be
7122 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7123 as command line arguments.
7125 Libssl is not considered directly vulnerable. Additionally certificates etc
7126 received via remote connections via libssl are also unlikely to be able to
7127 trigger these issues because of message size limits enforced within libssl.
7129 This issue was reported to OpenSSL Guido Vranken.
7134 * Side channel attack on modular exponentiation
7136 A side-channel attack was found which makes use of cache-bank conflicts on
7137 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7138 of RSA keys. The ability to exploit this issue is limited as it relies on
7139 an attacker who has control of code in a thread running on the same
7140 hyper-threaded core as the victim thread which is performing decryptions.
7142 This issue was reported to OpenSSL by Yuval Yarom, The University of
7143 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7144 Nadia Heninger, University of Pennsylvania with more information at
7145 <http://cachebleed.info>.
7150 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7151 if no keysize is specified with default_bits. This fixes an
7152 omission in an earlier change that changed all RSA/DSA key generation
7153 commands to use 2048 bits by default.
7157 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7159 * Protection for DH small subgroup attacks
7161 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7162 switched on by default and cannot be disabled. This could have some
7167 * SSLv2 doesn't block disabled ciphers
7169 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7170 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7171 been disabled, provided that the SSLv2 protocol was not also disabled via
7174 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7175 and Sebastian Schinzel.
7180 * Reject DH handshakes with parameters shorter than 1024 bits.
7184 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7186 * Certificate verify crash with missing PSS parameter
7188 The signature verification routines will crash with a NULL pointer
7189 dereference if presented with an ASN.1 signature using the RSA PSS
7190 algorithm and absent mask generation function parameter. Since these
7191 routines are used to verify certificate signature algorithms this can be
7192 used to crash any certificate verification operation and exploited in a
7193 DoS attack. Any application which performs certificate verification is
7194 vulnerable including OpenSSL clients and servers which enable client
7197 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7202 * X509_ATTRIBUTE memory leak
7204 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7205 memory. This structure is used by the PKCS#7 and CMS routines so any
7206 application which reads PKCS#7 or CMS data from untrusted sources is
7207 affected. SSL/TLS is not affected.
7209 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7215 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7216 This changes the decoding behaviour for some invalid messages,
7217 though the change is mostly in the more lenient direction, and
7218 legacy behaviour is preserved as much as possible.
7222 * In DSA_generate_parameters_ex, if the provided seed is too short,
7223 use a random seed, as already documented.
7225 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7227 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7229 * Alternate chains certificate forgery
7231 During certificate verfification, OpenSSL will attempt to find an
7232 alternative certificate chain if the first attempt to build such a chain
7233 fails. An error in the implementation of this logic can mean that an
7234 attacker could cause certain checks on untrusted certificates to be
7235 bypassed, such as the CA flag, enabling them to use a valid leaf
7236 certificate to act as a CA and "issue" an invalid certificate.
7238 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7244 * Race condition handling PSK identify hint
7246 If PSK identity hints are received by a multi-threaded client then
7247 the values are wrongly updated in the parent SSL_CTX structure. This can
7248 result in a race condition potentially leading to a double free of the
7254 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7256 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7257 incompatibility in the handling of HMAC. The previous ABI has now been
7260 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7262 * Malformed ECParameters causes infinite loop
7264 When processing an ECParameters structure OpenSSL enters an infinite loop
7265 if the curve specified is over a specially malformed binary polynomial
7268 This can be used to perform denial of service against any
7269 system which processes public keys, certificate requests or
7270 certificates. This includes TLS clients and TLS servers with
7271 client authentication enabled.
7273 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7278 * Exploitable out-of-bounds read in X509_cmp_time
7280 X509_cmp_time does not properly check the length of the ASN1_TIME
7281 string and can read a few bytes out of bounds. In addition,
7282 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7285 An attacker can use this to craft malformed certificates and CRLs of
7286 various sizes and potentially cause a segmentation fault, resulting in
7287 a DoS on applications that verify certificates or CRLs. TLS clients
7288 that verify CRLs are affected. TLS clients and servers with client
7289 authentication enabled may be affected if they use custom verification
7292 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7293 independently by Hanno Böck.
7298 * PKCS7 crash with missing EnvelopedContent
7300 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7301 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7302 with missing content and trigger a NULL pointer dereference on parsing.
7304 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7305 structures from untrusted sources are affected. OpenSSL clients and
7306 servers are not affected.
7308 This issue was reported to OpenSSL by Michal Zalewski (Google).
7313 * CMS verify infinite loop with unknown hash function
7315 When verifying a signedData message the CMS code can enter an infinite loop
7316 if presented with an unknown hash function OID. This can be used to perform
7317 denial of service against any system which verifies signedData messages using
7319 This issue was reported to OpenSSL by Johannes Bauer.
7324 * Race condition handling NewSessionTicket
7326 If a NewSessionTicket is received by a multi-threaded client when attempting to
7327 reuse a previous ticket then a race condition can occur potentially leading to
7328 a double free of the ticket data.
7333 * Reject DH handshakes with parameters shorter than 768 bits.
7335 *Kurt Roeckx and Emilia Kasper*
7337 * dhparam: generate 2048-bit parameters by default.
7339 *Kurt Roeckx and Emilia Kasper*
7341 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7343 * Segmentation fault in ASN1_TYPE_cmp fix
7345 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7346 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7347 certificate signature algorithm consistency this can be used to crash any
7348 certificate verification operation and exploited in a DoS attack. Any
7349 application which performs certificate verification is vulnerable including
7350 OpenSSL clients and servers which enable client authentication.
7355 * ASN.1 structure reuse memory corruption fix
7357 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7358 memory corruption via an invalid write. Such reuse is and has been
7359 strongly discouraged and is believed to be rare.
7361 Applications that parse structures containing CHOICE or ANY DEFINED BY
7362 components may be affected. Certificate parsing (d2i_X509 and related
7363 functions) are however not affected. OpenSSL clients and servers are
7369 * PKCS7 NULL pointer dereferences fix
7371 The PKCS#7 parsing code does not handle missing outer ContentInfo
7372 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7373 missing content and trigger a NULL pointer dereference on parsing.
7375 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7376 otherwise parse PKCS#7 structures from untrusted sources are
7377 affected. OpenSSL clients and servers are not affected.
7379 This issue was reported to OpenSSL by Michal Zalewski (Google).
7384 * DoS via reachable assert in SSLv2 servers fix
7386 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7387 servers that both support SSLv2 and enable export cipher suites by sending
7388 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7390 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7391 (OpenSSL development team).
7396 * Use After Free following d2i_ECPrivatekey error fix
7398 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7399 could cause a use after free condition. This, in turn, could cause a double
7400 free in several private key parsing functions (such as d2i_PrivateKey
7401 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7402 for applications that receive EC private keys from untrusted
7403 sources. This scenario is considered rare.
7405 This issue was discovered by the BoringSSL project and fixed in their
7411 * X509_to_X509_REQ NULL pointer deref fix
7413 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7414 the certificate key is invalid. This function is rarely used in practice.
7416 This issue was discovered by Brian Carpenter.
7421 * Removed the export ciphers from the DEFAULT ciphers
7425 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7427 * Build fixes for the Windows and OpenVMS platforms
7429 *Matt Caswell and Richard Levitte*
7431 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7433 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7434 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7435 dereference. This could lead to a Denial Of Service attack. Thanks to
7436 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7441 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7442 dtls1_buffer_record function under certain conditions. In particular this
7443 could occur if an attacker sent repeated DTLS records with the same
7444 sequence number but for the next epoch. The memory leak could be exploited
7445 by an attacker in a Denial of Service attack through memory exhaustion.
7446 Thanks to Chris Mueller for reporting this issue.
7451 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7452 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7453 method would be set to NULL which could later result in a NULL pointer
7454 dereference. Thanks to Frank Schmirler for reporting this issue.
7459 * Abort handshake if server key exchange message is omitted for ephemeral
7462 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7463 reporting this issue.
7468 * Remove non-export ephemeral RSA code on client and server. This code
7469 violated the TLS standard by allowing the use of temporary RSA keys in
7470 non-export ciphersuites and could be used by a server to effectively
7471 downgrade the RSA key length used to a value smaller than the server
7472 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7473 INRIA or reporting this issue.
7478 * Fixed issue where DH client certificates are accepted without verification.
7479 An OpenSSL server will accept a DH certificate for client authentication
7480 without the certificate verify message. This effectively allows a client to
7481 authenticate without the use of a private key. This only affects servers
7482 which trust a client certificate authority which issues certificates
7483 containing DH keys: these are extremely rare and hardly ever encountered.
7484 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7490 * Ensure that the session ID context of an SSL is updated when its
7491 SSL_CTX is updated via SSL_set_SSL_CTX.
7493 The session ID context is typically set from the parent SSL_CTX,
7494 and can vary with the CTX.
7498 * Fix various certificate fingerprint issues.
7500 By using non-DER or invalid encodings outside the signed portion of a
7501 certificate the fingerprint can be changed without breaking the signature.
7502 Although no details of the signed portion of the certificate can be changed
7503 this can cause problems with some applications: e.g. those using the
7504 certificate fingerprint for blacklists.
7506 1. Reject signatures with non zero unused bits.
7508 If the BIT STRING containing the signature has non zero unused bits reject
7509 the signature. All current signature algorithms require zero unused bits.
7511 2. Check certificate algorithm consistency.
7513 Check the AlgorithmIdentifier inside TBS matches the one in the
7514 certificate signature. NB: this will result in signature failure
7515 errors for some broken certificates.
7517 Thanks to Konrad Kraszewski from Google for reporting this issue.
7519 3. Check DSA/ECDSA signatures use DER.
7521 Re-encode DSA/ECDSA signatures and compare with the original received
7522 signature. Return an error if there is a mismatch.
7524 This will reject various cases including garbage after signature
7525 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7526 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7527 (negative or with leading zeroes).
7529 Further analysis was conducted and fixes were developed by Stephen Henson
7530 of the OpenSSL core team.
7536 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7537 results on some platforms, including x86_64. This bug occurs at random
7538 with a very low probability, and is not known to be exploitable in any
7539 way, though its exact impact is difficult to determine. Thanks to Pieter
7540 Wuille (Blockstream) who reported this issue and also suggested an initial
7541 fix. Further analysis was conducted by the OpenSSL development team and
7542 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7543 the OpenSSL core team.
7548 * Do not resume sessions on the server if the negotiated protocol
7549 version does not match the session's version. Resuming with a different
7550 version, while not strictly forbidden by the RFC, is of questionable
7551 sanity and breaks all known clients.
7553 *David Benjamin, Emilia Käsper*
7555 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7556 early CCS messages during renegotiation. (Note that because
7557 renegotiation is encrypted, this early CCS was not exploitable.)
7561 * Tighten client-side session ticket handling during renegotiation:
7562 ensure that the client only accepts a session ticket if the server sends
7563 the extension anew in the ServerHello. Previously, a TLS client would
7564 reuse the old extension state and thus accept a session ticket if one was
7565 announced in the initial ServerHello.
7567 Similarly, ensure that the client requires a session ticket if one
7568 was advertised in the ServerHello. Previously, a TLS client would
7569 ignore a missing NewSessionTicket message.
7573 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7577 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7578 sends a carefully crafted handshake message, to cause OpenSSL to fail
7579 to free up to 64k of memory causing a memory leak. This could be
7580 exploited in a Denial Of Service attack. This issue affects OpenSSL
7581 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7582 whether SRTP is used or configured. Implementations of OpenSSL that
7583 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7585 The fix was developed by the OpenSSL team.
7590 * Session Ticket Memory Leak.
7592 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7593 integrity of that ticket is first verified. In the event of a session
7594 ticket integrity check failing, OpenSSL will fail to free memory
7595 causing a memory leak. By sending a large number of invalid session
7596 tickets an attacker could exploit this issue in a Denial Of Service
7602 * Build option no-ssl3 is incomplete.
7604 When OpenSSL is configured with "no-ssl3" as a build option, servers
7605 could accept and complete an SSL 3.0 handshake, and clients could be
7606 configured to send them.
7609 *Akamai and the OpenSSL team*
7611 * Add support for TLS_FALLBACK_SCSV.
7612 Client applications doing fallback retries should call
7613 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7616 *Adam Langley, Bodo Moeller*
7618 * Add additional DigestInfo checks.
7620 Re-encode DigestInto in DER and check against the original when
7621 verifying RSA signature: this will reject any improperly encoded
7622 DigestInfo structures.
7624 Note: this is a precautionary measure and no attacks are currently known.
7628 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7630 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7631 SRP code can be overrun an internal buffer. Add sanity check that
7632 g, A, B < N to SRP code.
7634 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7635 Group for discovering this issue.
7640 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7641 TLS 1.0 instead of higher protocol versions when the ClientHello message
7642 is badly fragmented. This allows a man-in-the-middle attacker to force a
7643 downgrade to TLS 1.0 even if both the server and the client support a
7644 higher protocol version, by modifying the client's TLS records.
7646 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7647 researching this issue.
7652 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7653 to a denial of service attack. A malicious server can crash the client
7654 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7655 ciphersuite and sending carefully crafted handshake messages.
7657 Thanks to Felix Gröbert (Google) for discovering and researching this
7663 * By sending carefully crafted DTLS packets an attacker could cause openssl
7664 to leak memory. This can be exploited through a Denial of Service attack.
7665 Thanks to Adam Langley for discovering and researching this issue.
7670 * An attacker can force openssl to consume large amounts of memory whilst
7671 processing DTLS handshake messages. This can be exploited through a
7672 Denial of Service attack.
7673 Thanks to Adam Langley for discovering and researching this issue.
7678 * An attacker can force an error condition which causes openssl to crash
7679 whilst processing DTLS packets due to memory being freed twice. This
7680 can be exploited through a Denial of Service attack.
7681 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7687 * If a multithreaded client connects to a malicious server using a resumed
7688 session and the server sends an ec point format extension it could write
7689 up to 255 bytes to freed memory.
7691 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7697 * A malicious server can crash an OpenSSL client with a null pointer
7698 dereference (read) by specifying an SRP ciphersuite even though it was not
7699 properly negotiated with the client. This can be exploited through a
7700 Denial of Service attack.
7702 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7703 discovering and researching this issue.
7708 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7709 X509_name_oneline, X509_name_print_ex et al. to leak some information
7710 from the stack. Applications may be affected if they echo pretty printing
7711 output to the attacker.
7713 Thanks to Ivan Fratric (Google) for discovering this issue.
7716 *Emilia Käsper, and Steve Henson*
7718 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7719 for corner cases. (Certain input points at infinity could lead to
7720 bogus results, with non-infinity inputs mapped to infinity too.)
7724 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7726 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7727 handshake can force the use of weak keying material in OpenSSL
7728 SSL/TLS clients and servers.
7730 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7731 researching this issue. ([CVE-2014-0224])
7733 *KIKUCHI Masashi, Steve Henson*
7735 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7736 OpenSSL DTLS client the code can be made to recurse eventually crashing
7739 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7742 *Imre Rad, Steve Henson*
7744 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7745 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7746 client or server. This is potentially exploitable to run arbitrary
7747 code on a vulnerable client or server.
7749 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7751 *Jüri Aedla, Steve Henson*
7753 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7754 are subject to a denial of service attack.
7756 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7757 this issue. ([CVE-2014-3470])
7759 *Felix Gröbert, Ivan Fratric, Steve Henson*
7761 * Harmonize version and its documentation. -f flag is used to display
7764 *mancha <mancha1@zoho.com>*
7766 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7767 in i2d_ECPrivateKey.
7769 *mancha <mancha1@zoho.com>*
7771 * Fix some double frees. These are not thought to be exploitable.
7773 *mancha <mancha1@zoho.com>*
7775 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7777 * A missing bounds check in the handling of the TLS heartbeat extension
7778 can be used to reveal up to 64k of memory to a connected client or
7781 Thanks for Neel Mehta of Google Security for discovering this bug and to
7782 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7783 preparing the fix ([CVE-2014-0160])
7785 *Adam Langley, Bodo Moeller*
7787 * Fix for the attack described in the paper "Recovering OpenSSL
7788 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7789 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7790 <http://eprint.iacr.org/2014/140>
7792 Thanks to Yuval Yarom and Naomi Benger for discovering this
7793 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7795 *Yuval Yarom and Naomi Benger*
7797 * TLS pad extension: draft-agl-tls-padding-03
7799 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7800 TLS client Hello record length value would otherwise be > 255 and
7801 less that 512 pad with a dummy extension containing zeroes so it
7802 is at least 512 bytes long.
7804 *Adam Langley, Steve Henson*
7806 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7808 * Fix for TLS record tampering bug. A carefully crafted invalid
7809 handshake could crash OpenSSL with a NULL pointer exception.
7810 Thanks to Anton Johansson for reporting this issues.
7813 * Keep original DTLS digest and encryption contexts in retransmission
7814 structures so we can use the previous session parameters if they need
7815 to be resent. ([CVE-2013-6450])
7819 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7820 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7821 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7822 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7823 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7824 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7826 *Rob Stradling, Adam Langley*
7828 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7830 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7831 supporting platforms or when small records were transferred.
7833 *Andy Polyakov, Steve Henson*
7835 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7837 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7839 This addresses the flaw in CBC record processing discovered by
7840 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7841 at: <http://www.isg.rhul.ac.uk/tls/>
7843 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7844 Security Group at Royal Holloway, University of London
7845 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7846 Emilia Käsper for the initial patch.
7849 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7851 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7852 ciphersuites which can be exploited in a denial of service attack.
7853 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7854 and detecting this bug and to Wolfgang Ettlinger
7855 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7860 * Return an error when checking OCSP signatures when key is NULL.
7861 This fixes a DoS attack. ([CVE-2013-0166])
7865 * Make openssl verify return errors.
7867 *Chris Palmer <palmer@google.com> and Ben Laurie*
7869 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7870 the right response is stapled. Also change SSL_get_certificate()
7871 so it returns the certificate actually sent.
7872 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7874 *Rob Stradling <rob.stradling@comodo.com>*
7876 * Fix possible deadlock when decoding public keys.
7880 * Don't use TLS 1.0 record version number in initial client hello
7885 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7887 * Sanity check record length before skipping explicit IV in TLS
7888 1.2, 1.1 and DTLS to fix DoS attack.
7890 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7891 fuzzing as a service testing platform.
7896 * Initialise tkeylen properly when encrypting CMS messages.
7897 Thanks to Solar Designer of Openwall for reporting this issue.
7901 * In FIPS mode don't try to use composite ciphers as they are not
7906 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7908 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7909 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7910 mean any application compiled against OpenSSL 1.0.0 headers setting
7911 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7912 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7913 0x10000000L Any application which was previously compiled against
7914 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7915 will need to be recompiled as a result. Letting be results in
7916 inability to disable specifically TLS 1.1 and in client context,
7917 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7921 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7922 disable just protocol X, but all protocols above X *if* there are
7923 protocols *below* X still enabled. In more practical terms it means
7924 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7925 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7926 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7931 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7933 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7934 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7935 in CRYPTO_realloc_clean.
7937 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7938 issue and to Adam Langley <agl@chromium.org> for fixing it.
7941 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7943 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7947 * Workarounds for some broken servers that "hang" if a client hello
7948 record length exceeds 255 bytes.
7950 1. Do not use record version number > TLS 1.0 in initial client
7951 hello: some (but not all) hanging servers will now work.
7952 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7953 the number of ciphers sent in the client hello. This should be
7954 set to an even number, such as 50, for example by passing:
7955 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7956 Most broken servers should now work.
7957 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7958 TLS 1.2 client support entirely.
7962 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7966 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7968 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7969 STRING form instead of a DigestInfo.
7973 * The format used for MDC2 RSA signatures is inconsistent between EVP
7974 and the RSA_sign/RSA_verify functions. This was made more apparent when
7975 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7976 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7977 the correct format in RSA_verify so both forms transparently work.
7981 * Some servers which support TLS 1.0 can choke if we initially indicate
7982 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7983 encrypted premaster secret. As a workaround use the maximum permitted
7984 client version in client hello, this should keep such servers happy
7985 and still work with previous versions of OpenSSL.
7989 * Add support for TLS/DTLS heartbeats.
7991 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7993 * Add support for SCTP.
7995 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7997 * Improved PRNG seeding for VOS.
7999 *Paul Green <Paul.Green@stratus.com>*
8001 * Extensive assembler packs updates, most notably:
8003 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8004 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8005 - x86_64: bit-sliced AES implementation;
8006 - ARM: NEON support, contemporary platforms optimizations;
8007 - s390x: z196 support;
8008 - `*`: GHASH and GF(2^m) multiplication implementations;
8012 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8013 (removal of unnecessary code)
8015 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8017 * Add TLS key material exporter from RFC 5705.
8021 * Add DTLS-SRTP negotiation from RFC 5764.
8025 * Add Next Protocol Negotiation,
8026 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8027 disabled with a no-npn flag to config or Configure. Code donated
8030 *Adam Langley <agl@google.com> and Ben Laurie*
8032 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8033 NIST-P256, NIST-P521, with constant-time single point multiplication on
8034 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8035 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8036 Code made available under Apache License version 2.0.
8038 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8039 line to include this in your build of OpenSSL, and run "make depend" (or
8040 "make update"). This enables the following EC_METHODs:
8042 EC_GFp_nistp224_method()
8043 EC_GFp_nistp256_method()
8044 EC_GFp_nistp521_method()
8046 EC_GROUP_new_by_curve_name() will automatically use these (while
8047 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8050 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8052 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8053 all platforms. Move ssize_t definition from e_os.h to the public
8054 header file e_os2.h as it now appears in public header file cms.h
8058 * New -sigopt option to the ca, req and x509 utilities. Additional
8059 signature parameters can be passed using this option and in
8064 * Add RSA PSS signing function. This will generate and set the
8065 appropriate AlgorithmIdentifiers for PSS based on those in the
8066 corresponding EVP_MD_CTX structure. No application support yet.
8070 * Support for companion algorithm specific ASN1 signing routines.
8071 New function ASN1_item_sign_ctx() signs a pre-initialised
8072 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8073 the appropriate parameters.
8077 * Add new algorithm specific ASN1 verification initialisation function
8078 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8079 handling will be the same no matter what EVP_PKEY_METHOD is used.
8080 Add a PSS handler to support verification of PSS signatures: checked
8081 against a number of sample certificates.
8085 * Add signature printing for PSS. Add PSS OIDs.
8087 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8089 * Add algorithm specific signature printing. An individual ASN1 method
8090 can now print out signatures instead of the standard hex dump.
8092 More complex signatures (e.g. PSS) can print out more meaningful
8093 information. Include DSA version that prints out the signature
8098 * Password based recipient info support for CMS library: implementing
8103 * Split password based encryption into PBES2 and PBKDF2 functions. This
8104 neatly separates the code into cipher and PBE sections and is required
8105 for some algorithms that split PBES2 into separate pieces (such as
8106 password based CMS).
8110 * Session-handling fixes:
8111 - Fix handling of connections that are resuming with a session ID,
8112 but also support Session Tickets.
8113 - Fix a bug that suppressed issuing of a new ticket if the client
8114 presented a ticket with an expired session.
8115 - Try to set the ticket lifetime hint to something reasonable.
8116 - Make tickets shorter by excluding irrelevant information.
8117 - On the client side, don't ignore renewed tickets.
8119 *Adam Langley, Bodo Moeller (Google)*
8121 * Fix PSK session representation.
8125 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8127 This work was sponsored by Intel.
8131 * Add GCM support to TLS library. Some custom code is needed to split
8132 the IV between the fixed (from PRF) and explicit (from TLS record)
8133 portions. This adds all GCM ciphersuites supported by RFC5288 and
8134 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8135 add a special AESGCM string for GCM only.
8139 * Expand range of ctrls for AES GCM. Permit setting invocation
8140 field on decrypt and retrieval of invocation field only on encrypt.
8144 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8145 As required by RFC5289 these ciphersuites cannot be used if for
8146 versions of TLS earlier than 1.2.
8150 * For FIPS capable OpenSSL interpret a NULL default public key method
8151 as unset and return the appropriate default but do *not* set the default.
8152 This means we can return the appropriate method in applications that
8153 switch between FIPS and non-FIPS modes.
8157 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8158 ENGINE is used then we cannot handle that in the FIPS module so we
8159 keep original code iff non-FIPS operations are allowed.
8163 * Add -attime option to openssl utilities.
8165 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8167 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8171 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8172 FIPS EC methods unconditionally for now.
8176 * New build option no-ec2m to disable characteristic 2 code.
8180 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8181 all cases can be covered as some introduce binary incompatibilities.
8185 * Redirect RSA operations to FIPS module including keygen,
8186 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8190 * Add similar low-level API blocking to ciphers.
8194 * low-level digest APIs are not approved in FIPS mode: any attempt
8195 to use these will cause a fatal error. Applications that *really* want
8196 to use them can use the `private_*` version instead.
8200 * Redirect cipher operations to FIPS module for FIPS builds.
8204 * Redirect digest operations to FIPS module for FIPS builds.
8208 * Update build system to add "fips" flag which will link in fipscanister.o
8209 for static and shared library builds embedding a signature if needed.
8213 * Output TLS supported curves in preference order instead of numerical
8214 order. This is currently hardcoded for the highest order curves first.
8215 This should be configurable so applications can judge speed vs strength.
8219 * Add TLS v1.2 server support for client authentication.
8223 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8228 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8229 FIPS modules versions.
8233 * Add TLS v1.2 client side support for client authentication. Keep cache
8234 of handshake records longer as we don't know the hash algorithm to use
8235 until after the certificate request message is received.
8239 * Initial TLS v1.2 client support. Add a default signature algorithms
8240 extension including all the algorithms we support. Parse new signature
8241 format in client key exchange. Relax some ECC signing restrictions for
8242 TLS v1.2 as indicated in RFC5246.
8246 * Add server support for TLS v1.2 signature algorithms extension. Switch
8247 to new signature format when needed using client digest preference.
8248 All server ciphersuites should now work correctly in TLS v1.2. No client
8249 support yet and no support for client certificates.
8253 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8254 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8255 ciphersuites. At present only RSA key exchange ciphersuites work with
8256 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8257 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8258 and version checking.
8262 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8263 with this defined it will not be affected by any changes to ssl internal
8264 structures. Add several utility functions to allow openssl application
8265 to work with OPENSSL_NO_SSL_INTERN defined.
8269 * A long standing patch to add support for SRP from EdelWeb (Peter
8270 Sylvester and Christophe Renou) was integrated.
8271 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8272 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8275 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8279 * Permit abbreviated handshakes when renegotiating using the function
8280 SSL_renegotiate_abbreviated().
8282 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8284 * Add call to ENGINE_register_all_complete() to
8285 ENGINE_load_builtin_engines(), so some implementations get used
8286 automatically instead of needing explicit application support.
8290 * Add support for TLS key exporter as described in RFC5705.
8292 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8294 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8295 a few changes are required:
8297 Add SSL_OP_NO_TLSv1_1 flag.
8298 Add TLSv1_1 methods.
8299 Update version checking logic to handle version 1.1.
8300 Add explicit IV handling (ported from DTLS code).
8301 Add command line options to s_client/s_server.
8308 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8310 * X509_ATTRIBUTE memory leak
8312 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8313 memory. This structure is used by the PKCS#7 and CMS routines so any
8314 application which reads PKCS#7 or CMS data from untrusted sources is
8315 affected. SSL/TLS is not affected.
8317 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8323 * Race condition handling PSK identify hint
8325 If PSK identity hints are received by a multi-threaded client then
8326 the values are wrongly updated in the parent SSL_CTX structure. This can
8327 result in a race condition potentially leading to a double free of the
8333 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8335 * Malformed ECParameters causes infinite loop
8337 When processing an ECParameters structure OpenSSL enters an infinite loop
8338 if the curve specified is over a specially malformed binary polynomial
8341 This can be used to perform denial of service against any
8342 system which processes public keys, certificate requests or
8343 certificates. This includes TLS clients and TLS servers with
8344 client authentication enabled.
8346 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8351 * Exploitable out-of-bounds read in X509_cmp_time
8353 X509_cmp_time does not properly check the length of the ASN1_TIME
8354 string and can read a few bytes out of bounds. In addition,
8355 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8358 An attacker can use this to craft malformed certificates and CRLs of
8359 various sizes and potentially cause a segmentation fault, resulting in
8360 a DoS on applications that verify certificates or CRLs. TLS clients
8361 that verify CRLs are affected. TLS clients and servers with client
8362 authentication enabled may be affected if they use custom verification
8365 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8366 independently by Hanno Böck.
8371 * PKCS7 crash with missing EnvelopedContent
8373 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8374 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8375 with missing content and trigger a NULL pointer dereference on parsing.
8377 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8378 structures from untrusted sources are affected. OpenSSL clients and
8379 servers are not affected.
8381 This issue was reported to OpenSSL by Michal Zalewski (Google).
8386 * CMS verify infinite loop with unknown hash function
8388 When verifying a signedData message the CMS code can enter an infinite loop
8389 if presented with an unknown hash function OID. This can be used to perform
8390 denial of service against any system which verifies signedData messages using
8392 This issue was reported to OpenSSL by Johannes Bauer.
8397 * Race condition handling NewSessionTicket
8399 If a NewSessionTicket is received by a multi-threaded client when attempting to
8400 reuse a previous ticket then a race condition can occur potentially leading to
8401 a double free of the ticket data.
8406 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8408 * Segmentation fault in ASN1_TYPE_cmp fix
8410 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8411 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8412 certificate signature algorithm consistency this can be used to crash any
8413 certificate verification operation and exploited in a DoS attack. Any
8414 application which performs certificate verification is vulnerable including
8415 OpenSSL clients and servers which enable client authentication.
8420 * ASN.1 structure reuse memory corruption fix
8422 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8423 memory corruption via an invalid write. Such reuse is and has been
8424 strongly discouraged and is believed to be rare.
8426 Applications that parse structures containing CHOICE or ANY DEFINED BY
8427 components may be affected. Certificate parsing (d2i_X509 and related
8428 functions) are however not affected. OpenSSL clients and servers are
8434 * PKCS7 NULL pointer dereferences fix
8436 The PKCS#7 parsing code does not handle missing outer ContentInfo
8437 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8438 missing content and trigger a NULL pointer dereference on parsing.
8440 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8441 otherwise parse PKCS#7 structures from untrusted sources are
8442 affected. OpenSSL clients and servers are not affected.
8444 This issue was reported to OpenSSL by Michal Zalewski (Google).
8449 * DoS via reachable assert in SSLv2 servers fix
8451 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8452 servers that both support SSLv2 and enable export cipher suites by sending
8453 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8455 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8456 (OpenSSL development team).
8461 * Use After Free following d2i_ECPrivatekey error fix
8463 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8464 could cause a use after free condition. This, in turn, could cause a double
8465 free in several private key parsing functions (such as d2i_PrivateKey
8466 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8467 for applications that receive EC private keys from untrusted
8468 sources. This scenario is considered rare.
8470 This issue was discovered by the BoringSSL project and fixed in their
8476 * X509_to_X509_REQ NULL pointer deref fix
8478 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8479 the certificate key is invalid. This function is rarely used in practice.
8481 This issue was discovered by Brian Carpenter.
8486 * Removed the export ciphers from the DEFAULT ciphers
8490 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8492 * Build fixes for the Windows and OpenVMS platforms
8494 *Matt Caswell and Richard Levitte*
8496 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8498 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8499 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8500 dereference. This could lead to a Denial Of Service attack. Thanks to
8501 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8506 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8507 dtls1_buffer_record function under certain conditions. In particular this
8508 could occur if an attacker sent repeated DTLS records with the same
8509 sequence number but for the next epoch. The memory leak could be exploited
8510 by an attacker in a Denial of Service attack through memory exhaustion.
8511 Thanks to Chris Mueller for reporting this issue.
8516 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8517 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8518 method would be set to NULL which could later result in a NULL pointer
8519 dereference. Thanks to Frank Schmirler for reporting this issue.
8524 * Abort handshake if server key exchange message is omitted for ephemeral
8527 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8528 reporting this issue.
8533 * Remove non-export ephemeral RSA code on client and server. This code
8534 violated the TLS standard by allowing the use of temporary RSA keys in
8535 non-export ciphersuites and could be used by a server to effectively
8536 downgrade the RSA key length used to a value smaller than the server
8537 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8538 INRIA or reporting this issue.
8543 * Fixed issue where DH client certificates are accepted without verification.
8544 An OpenSSL server will accept a DH certificate for client authentication
8545 without the certificate verify message. This effectively allows a client to
8546 authenticate without the use of a private key. This only affects servers
8547 which trust a client certificate authority which issues certificates
8548 containing DH keys: these are extremely rare and hardly ever encountered.
8549 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8555 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8556 results on some platforms, including x86_64. This bug occurs at random
8557 with a very low probability, and is not known to be exploitable in any
8558 way, though its exact impact is difficult to determine. Thanks to Pieter
8559 Wuille (Blockstream) who reported this issue and also suggested an initial
8560 fix. Further analysis was conducted by the OpenSSL development team and
8561 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8562 the OpenSSL core team.
8567 * Fix various certificate fingerprint issues.
8569 By using non-DER or invalid encodings outside the signed portion of a
8570 certificate the fingerprint can be changed without breaking the signature.
8571 Although no details of the signed portion of the certificate can be changed
8572 this can cause problems with some applications: e.g. those using the
8573 certificate fingerprint for blacklists.
8575 1. Reject signatures with non zero unused bits.
8577 If the BIT STRING containing the signature has non zero unused bits reject
8578 the signature. All current signature algorithms require zero unused bits.
8580 2. Check certificate algorithm consistency.
8582 Check the AlgorithmIdentifier inside TBS matches the one in the
8583 certificate signature. NB: this will result in signature failure
8584 errors for some broken certificates.
8586 Thanks to Konrad Kraszewski from Google for reporting this issue.
8588 3. Check DSA/ECDSA signatures use DER.
8590 Reencode DSA/ECDSA signatures and compare with the original received
8591 signature. Return an error if there is a mismatch.
8593 This will reject various cases including garbage after signature
8594 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8595 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8596 (negative or with leading zeroes).
8598 Further analysis was conducted and fixes were developed by Stephen Henson
8599 of the OpenSSL core team.
8605 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8607 * Session Ticket Memory Leak.
8609 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8610 integrity of that ticket is first verified. In the event of a session
8611 ticket integrity check failing, OpenSSL will fail to free memory
8612 causing a memory leak. By sending a large number of invalid session
8613 tickets an attacker could exploit this issue in a Denial Of Service
8619 * Build option no-ssl3 is incomplete.
8621 When OpenSSL is configured with "no-ssl3" as a build option, servers
8622 could accept and complete an SSL 3.0 handshake, and clients could be
8623 configured to send them.
8626 *Akamai and the OpenSSL team*
8628 * Add support for TLS_FALLBACK_SCSV.
8629 Client applications doing fallback retries should call
8630 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8633 *Adam Langley, Bodo Moeller*
8635 * Add additional DigestInfo checks.
8637 Reencode DigestInto in DER and check against the original when
8638 verifying RSA signature: this will reject any improperly encoded
8639 DigestInfo structures.
8641 Note: this is a precautionary measure and no attacks are currently known.
8645 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8647 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8648 to a denial of service attack. A malicious server can crash the client
8649 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8650 ciphersuite and sending carefully crafted handshake messages.
8652 Thanks to Felix Gröbert (Google) for discovering and researching this
8658 * By sending carefully crafted DTLS packets an attacker could cause openssl
8659 to leak memory. This can be exploited through a Denial of Service attack.
8660 Thanks to Adam Langley for discovering and researching this issue.
8665 * An attacker can force openssl to consume large amounts of memory whilst
8666 processing DTLS handshake messages. This can be exploited through a
8667 Denial of Service attack.
8668 Thanks to Adam Langley for discovering and researching this issue.
8673 * An attacker can force an error condition which causes openssl to crash
8674 whilst processing DTLS packets due to memory being freed twice. This
8675 can be exploited through a Denial of Service attack.
8676 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8682 * If a multithreaded client connects to a malicious server using a resumed
8683 session and the server sends an ec point format extension it could write
8684 up to 255 bytes to freed memory.
8686 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8692 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8693 X509_name_oneline, X509_name_print_ex et al. to leak some information
8694 from the stack. Applications may be affected if they echo pretty printing
8695 output to the attacker.
8697 Thanks to Ivan Fratric (Google) for discovering this issue.
8700 *Emilia Käsper, and Steve Henson*
8702 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8703 for corner cases. (Certain input points at infinity could lead to
8704 bogus results, with non-infinity inputs mapped to infinity too.)
8708 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8710 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8711 handshake can force the use of weak keying material in OpenSSL
8712 SSL/TLS clients and servers.
8714 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8715 researching this issue. ([CVE-2014-0224])
8717 *KIKUCHI Masashi, Steve Henson*
8719 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8720 OpenSSL DTLS client the code can be made to recurse eventually crashing
8723 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8726 *Imre Rad, Steve Henson*
8728 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8729 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8730 client or server. This is potentially exploitable to run arbitrary
8731 code on a vulnerable client or server.
8733 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8735 *Jüri Aedla, Steve Henson*
8737 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8738 are subject to a denial of service attack.
8740 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8741 this issue. ([CVE-2014-3470])
8743 *Felix Gröbert, Ivan Fratric, Steve Henson*
8745 * Harmonize version and its documentation. -f flag is used to display
8748 *mancha <mancha1@zoho.com>*
8750 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8751 in i2d_ECPrivateKey.
8753 *mancha <mancha1@zoho.com>*
8755 * Fix some double frees. These are not thought to be exploitable.
8757 *mancha <mancha1@zoho.com>*
8759 * Fix for the attack described in the paper "Recovering OpenSSL
8760 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8761 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8762 <http://eprint.iacr.org/2014/140>
8764 Thanks to Yuval Yarom and Naomi Benger for discovering this
8765 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8767 *Yuval Yarom and Naomi Benger*
8769 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8771 * Keep original DTLS digest and encryption contexts in retransmission
8772 structures so we can use the previous session parameters if they need
8773 to be resent. ([CVE-2013-6450])
8777 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8778 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8779 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8780 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8781 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8782 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8784 *Rob Stradling, Adam Langley*
8786 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8788 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8790 This addresses the flaw in CBC record processing discovered by
8791 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8792 at: <http://www.isg.rhul.ac.uk/tls/>
8794 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8795 Security Group at Royal Holloway, University of London
8796 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8797 Emilia Käsper for the initial patch.
8800 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8802 * Return an error when checking OCSP signatures when key is NULL.
8803 This fixes a DoS attack. ([CVE-2013-0166])
8807 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8808 the right response is stapled. Also change SSL_get_certificate()
8809 so it returns the certificate actually sent.
8810 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8811 (This is a backport)
8813 *Rob Stradling <rob.stradling@comodo.com>*
8815 * Fix possible deadlock when decoding public keys.
8819 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8821 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8824 * Sanity check record length before skipping explicit IV in DTLS
8827 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8828 fuzzing as a service testing platform.
8833 * Initialise tkeylen properly when encrypting CMS messages.
8834 Thanks to Solar Designer of Openwall for reporting this issue.
8838 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8840 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8841 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8842 in CRYPTO_realloc_clean.
8844 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8845 issue and to Adam Langley <agl@chromium.org> for fixing it.
8848 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8850 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8852 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8853 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8854 content decryption and always return the same error. Note: this attack
8855 needs on average 2^20 messages so it only affects automated senders. The
8856 old behaviour can be re-enabled in the CMS code by setting the
8857 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8858 an MMA defence is not necessary.
8859 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8860 this issue. ([CVE-2012-0884])
8864 * Fix CVE-2011-4619: make sure we really are receiving a
8865 client hello before rejecting multiple SGC restarts. Thanks to
8866 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8870 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8872 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8873 Thanks to Antonio Martin, Enterprise Secure Access Research and
8874 Development, Cisco Systems, Inc. for discovering this bug and
8875 preparing a fix. ([CVE-2012-0050])
8879 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8881 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8882 of the Vaudenay padding oracle attack on CBC mode encryption
8883 which enables an efficient plaintext recovery attack against
8884 the OpenSSL implementation of DTLS. Their attack exploits timing
8885 differences arising during decryption processing. A research
8886 paper describing this attack can be found at:
8887 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8888 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8889 Security Group at Royal Holloway, University of London
8890 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8891 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8892 for preparing the fix. ([CVE-2011-4108])
8894 *Robin Seggelmann, Michael Tuexen*
8896 * Clear bytes used for block padding of SSL 3.0 records.
8899 *Adam Langley (Google)*
8901 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8902 Kadianakis <desnacked@gmail.com> for discovering this issue and
8903 Adam Langley for preparing the fix. ([CVE-2011-4619])
8905 *Adam Langley (Google)*
8907 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8909 *Andrey Kulikov <amdeich@gmail.com>*
8911 * Prevent malformed RFC3779 data triggering an assertion failure.
8912 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8913 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8915 *Rob Austein <sra@hactrn.net>*
8917 * Improved PRNG seeding for VOS.
8919 *Paul Green <Paul.Green@stratus.com>*
8921 * Fix ssl_ciph.c set-up race.
8923 *Adam Langley (Google)*
8925 * Fix spurious failures in ecdsatest.c.
8927 *Emilia Käsper (Google)*
8929 * Fix the BIO_f_buffer() implementation (which was mixing different
8930 interpretations of the `..._len` fields).
8932 *Adam Langley (Google)*
8934 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8935 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8936 threads won't reuse the same blinding coefficients.
8938 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8939 lock to call BN_BLINDING_invert_ex, and avoids one use of
8940 BN_BLINDING_update for each BN_BLINDING structure (previously,
8941 the last update always remained unused).
8943 *Emilia Käsper (Google)*
8945 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8947 *Bob Buckholz (Google)*
8949 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8951 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8952 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8954 *Kaspar Brand <ossl@velox.ch>*
8956 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8957 for multi-threaded use of ECDH. ([CVE-2011-3210])
8959 *Adam Langley (Google)*
8961 * Fix x509_name_ex_d2i memory leak on bad inputs.
8965 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8966 signature public key algorithm by using OID xref utilities instead.
8967 Before this you could only use some ECC ciphersuites with SHA1 only.
8971 * Add protection against ECDSA timing attacks as mentioned in the paper
8972 by Billy Bob Brumley and Nicola Tuveri, see:
8973 <http://eprint.iacr.org/2011/232.pdf>
8975 *Billy Bob Brumley and Nicola Tuveri*
8977 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8979 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8981 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8983 * Fix bug in string printing code: if *any* escaping is enabled we must
8984 escape the escape character (backslash) or the resulting string is
8989 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8991 * Disable code workaround for ancient and obsolete Netscape browsers
8992 and servers: an attacker can use it in a ciphersuite downgrade attack.
8993 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8997 * Fixed J-PAKE implementation error, originally discovered by
8998 Sebastien Martini, further info and confirmation from Stefan
8999 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9003 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9005 * Fix extension code to avoid race conditions which can result in a buffer
9006 overrun vulnerability: resumed sessions must not be modified as they can
9007 be shared by multiple threads. CVE-2010-3864
9011 * Fix WIN32 build system to correctly link an ENGINE directory into
9016 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9018 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9021 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9023 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9025 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9026 context. The operation can be customised via the ctrl mechanism in
9027 case ENGINEs want to include additional functionality.
9031 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9035 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9036 output hashes compatible with older versions of OpenSSL.
9038 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9040 * Fix compression algorithm handling: if resuming a session use the
9041 compression algorithm of the resumed session instead of determining
9042 it from client hello again. Don't allow server to change algorithm.
9046 * Add load_crls() function to commands tidying load_certs() too. Add option
9047 to verify utility to allow additional CRLs to be included.
9051 * Update OCSP request code to permit adding custom headers to the request:
9052 some responders need this.
9056 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9059 *Julia Lawall <julia@diku.dk>*
9061 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9062 needlessly dereferenced structures, used obsolete functions and
9063 didn't handle all updated verify codes correctly.
9067 * Disable MD2 in the default configuration.
9071 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9072 indicate the initial BIO being pushed or popped. This makes it possible
9073 to determine whether the BIO is the one explicitly called or as a result
9074 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9075 it handles reference counts correctly and doesn't zero out the I/O bio
9076 when it is not being explicitly popped. WARNING: applications which
9077 included workarounds for the old buggy behaviour will need to be modified
9078 or they could free up already freed BIOs.
9082 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9083 renaming to all platforms (within the 0.9.8 branch, this was
9084 done conditionally on Netware platforms to avoid a name clash).
9086 *Guenter <lists@gknw.net>*
9088 * Add ECDHE and PSK support to DTLS.
9090 *Michael Tuexen <tuexen@fh-muenster.de>*
9092 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9097 * Add "missing" function EVP_MD_flags() (without this the only way to
9098 retrieve a digest flags is by accessing the structure directly. Update
9099 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9100 or cipher is registered as in the "from" argument. Print out all
9101 registered digests in the dgst usage message instead of manually
9102 attempting to work them out.
9106 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9107 this allows the use of compression and extensions. Change default cipher
9108 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9109 by default unless an application cipher string requests it.
9113 * Alter match criteria in PKCS12_parse(). It used to try to use local
9114 key ids to find matching certificates and keys but some PKCS#12 files
9115 don't follow the (somewhat unwritten) rules and this strategy fails.
9116 Now just gather all certificates together and the first private key
9117 then look for the first certificate that matches the key.
9121 * Support use of registered digest and cipher names for dgst and cipher
9122 commands instead of having to add each one as a special case. So now
9129 openssl dgst -sha256 foo
9131 and this works for ENGINE based algorithms too.
9135 * Update Gost ENGINE to support parameter files.
9137 *Victor B. Wagner <vitus@cryptocom.ru>*
9139 * Support GeneralizedTime in ca utility.
9141 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9143 * Enhance the hash format used for certificate directory links. The new
9144 form uses the canonical encoding (meaning equivalent names will work
9145 even if they aren't identical) and uses SHA1 instead of MD5. This form
9146 is incompatible with the older format and as a result c_rehash should
9147 be used to rebuild symbolic links.
9151 * Make PKCS#8 the default write format for private keys, replacing the
9152 traditional format. This form is standardised, more secure and doesn't
9153 include an implicit MD5 dependency.
9157 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9158 committed to OpenSSL should pass this lot as a minimum.
9162 * Add session ticket override functionality for use by EAP-FAST.
9164 *Jouni Malinen <j@w1.fi>*
9166 * Modify HMAC functions to return a value. Since these can be implemented
9167 in an ENGINE errors can occur.
9171 * Type-checked OBJ_bsearch_ex.
9175 * Type-checked OBJ_bsearch. Also some constification necessitated
9176 by type-checking. Still to come: TXT_DB, bsearch(?),
9177 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9182 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9183 seconds to a tm structure directly, instead of going through OS
9184 specific date routines. This avoids any issues with OS routines such
9185 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9186 and X509_time_adj_ex() to cover the extended range. The existing
9187 X509_time_adj() is still usable and will no longer have any date issues.
9191 * Delta CRL support. New use deltas option which will attempt to locate
9192 and search any appropriate delta CRLs available.
9194 This work was sponsored by Google.
9198 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9199 code and add additional score elements. Validate alternate CRL paths
9200 as part of the CRL checking and indicate a new error "CRL path validation
9201 error" in this case. Applications wanting additional details can use
9202 the verify callback and check the new "parent" field. If this is not
9203 NULL CRL path validation is taking place. Existing applications won't
9204 see this because it requires extended CRL support which is off by
9207 This work was sponsored by Google.
9211 * Support for freshest CRL extension.
9213 This work was sponsored by Google.
9217 * Initial indirect CRL support. Currently only supported in the CRLs
9218 passed directly and not via lookup. Process certificate issuer
9219 CRL entry extension and lookup CRL entries by bother issuer name
9220 and serial number. Check and process CRL issuer entry in IDP extension.
9222 This work was sponsored by Google.
9226 * Add support for distinct certificate and CRL paths. The CRL issuer
9227 certificate is validated separately in this case. Only enabled if
9228 an extended CRL support flag is set: this flag will enable additional
9229 CRL functionality in future.
9231 This work was sponsored by Google.
9235 * Add support for policy mappings extension.
9237 This work was sponsored by Google.
9241 * Fixes to pathlength constraint, self issued certificate handling,
9242 policy processing to align with RFC3280 and PKITS tests.
9244 This work was sponsored by Google.
9248 * Support for name constraints certificate extension. DN, email, DNS
9249 and URI types are currently supported.
9251 This work was sponsored by Google.
9255 * To cater for systems that provide a pointer-based thread ID rather
9256 than numeric, deprecate the current numeric thread ID mechanism and
9257 replace it with a structure and associated callback type. This
9258 mechanism allows a numeric "hash" to be extracted from a thread ID in
9259 either case, and on platforms where pointers are larger than 'long',
9260 mixing is done to help ensure the numeric 'hash' is usable even if it
9261 can't be guaranteed unique. The default mechanism is to use "&errno"
9262 as a pointer-based thread ID to distinguish between threads.
9264 Applications that want to provide their own thread IDs should now use
9265 CRYPTO_THREADID_set_callback() to register a callback that will call
9266 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9268 Note that ERR_remove_state() is now deprecated, because it is tied
9269 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9270 to free the current thread's error state should be replaced by
9271 ERR_remove_thread_state(NULL).
9273 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9274 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9275 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9276 application was previously providing a numeric thread callback that
9277 was inappropriate for distinguishing threads, then uniqueness might
9278 have been obtained with &errno that happened immediately in the
9279 intermediate development versions of OpenSSL; this is no longer the
9280 case, the numeric thread callback will now override the automatic use
9283 *Geoff Thorpe, with help from Bodo Moeller*
9285 * Initial support for different CRL issuing certificates. This covers a
9286 simple case where the self issued certificates in the chain exist and
9287 the real CRL issuer is higher in the existing chain.
9289 This work was sponsored by Google.
9293 * Removed effectively defunct crypto/store from the build.
9297 * Revamp of STACK to provide stronger type-checking. Still to come:
9298 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9299 ASN1_STRING, CONF_VALUE.
9303 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9304 RAM on SSL connections. This option can save about 34k per idle SSL.
9308 * Revamp of LHASH to provide stronger type-checking. Still to come:
9309 STACK, TXT_DB, bsearch, qsort.
9313 * Initial support for Cryptographic Message Syntax (aka CMS) based
9314 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9315 support for data, signedData, compressedData, digestedData and
9316 encryptedData, envelopedData types included. Scripts to check against
9317 RFC4134 examples draft and interop and consistency checks of many
9318 content types and variants.
9322 * Add options to enc utility to support use of zlib compression BIO.
9326 * Extend mk1mf to support importing of options and assembly language
9327 files from Configure script, currently only included in VC-WIN32.
9328 The assembly language rules can now optionally generate the source
9329 files from the associated perl scripts.
9333 * Implement remaining functionality needed to support GOST ciphersuites.
9334 Interop testing has been performed using CryptoPro implementations.
9336 *Victor B. Wagner <vitus@cryptocom.ru>*
9338 * s390x assembler pack.
9342 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9347 * Implement Opaque PRF Input TLS extension as specified in
9348 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9349 official specification yet and no extension type assignment by
9350 IANA exists, this extension (for now) will have to be explicitly
9351 enabled when building OpenSSL by providing the extension number
9352 to use. For example, specify an option
9354 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9356 to the "config" or "Configure" script to enable the extension,
9357 assuming extension number 0x9527 (which is a completely arbitrary
9358 and unofficial assignment based on the MD5 hash of the Internet
9359 Draft). Note that by doing so, you potentially lose
9360 interoperability with other TLS implementations since these might
9361 be using the same extension number for other purposes.
9363 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9364 opaque PRF input value to use in the handshake. This will create
9365 an internal copy of the length-'len' string at 'src', and will
9366 return non-zero for success.
9368 To get more control and flexibility, provide a callback function
9371 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9372 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9376 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9379 Callback function 'cb' will be called in handshakes, and is
9380 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9381 Argument 'arg' is for application purposes (the value as given to
9382 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9383 be provided to the callback function). The callback function
9384 has to return non-zero to report success: usually 1 to use opaque
9385 PRF input just if possible, or 2 to enforce use of the opaque PRF
9386 input. In the latter case, the library will abort the handshake
9387 if opaque PRF input is not successfully negotiated.
9389 Arguments 'peerinput' and 'len' given to the callback function
9390 will always be NULL and 0 in the case of a client. A server will
9391 see the client's opaque PRF input through these variables if
9392 available (NULL and 0 otherwise). Note that if the server
9393 provides an opaque PRF input, the length must be the same as the
9394 length of the client's opaque PRF input.
9396 Note that the callback function will only be called when creating
9397 a new session (session resumption can resume whatever was
9398 previously negotiated), and will not be called in SSL 2.0
9399 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9400 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9401 for applications that need to enforce opaque PRF input.
9405 * Update ssl code to support digests other than SHA1+MD5 for handshake
9408 *Victor B. Wagner <vitus@cryptocom.ru>*
9410 * Add RFC4507 support to OpenSSL. This includes the corrections in
9411 RFC4507bis. The encrypted ticket format is an encrypted encoded
9412 SSL_SESSION structure, that way new session features are automatically
9415 If a client application caches session in an SSL_SESSION structure
9416 support is transparent because tickets are now stored in the encoded
9419 The SSL_CTX structure automatically generates keys for ticket
9420 protection in servers so again support should be possible
9421 with no application modification.
9423 If a client or server wishes to disable RFC4507 support then the option
9424 SSL_OP_NO_TICKET can be set.
9426 Add a TLS extension debugging callback to allow the contents of any client
9427 or server extensions to be examined.
9429 This work was sponsored by Google.
9433 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9434 OpenSSL should now compile cleanly on gcc 4.2
9436 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9438 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9439 support including streaming MAC support: this is required for GOST
9440 ciphersuite support.
9442 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9444 * Add option -stream to use PKCS#7 streaming in smime utility. New
9445 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9446 to output in BER and PEM format.
9450 * Experimental support for use of HMAC via EVP_PKEY interface. This
9451 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9452 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9453 ENGINE support for HMAC keys which are unextractable. New -mac and
9454 -macopt options to dgst utility.
9458 * New option -sigopt to dgst utility. Update dgst to use
9459 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9460 alternative signing parameters such as X9.31 or PSS in the dgst
9465 * Change ssl_cipher_apply_rule(), the internal function that does
9466 the work each time a ciphersuite string requests enabling
9467 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9468 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9469 the order of disabled ciphersuites such that those ciphersuites
9470 that most recently went from enabled to disabled not only stay
9471 in order with respect to each other, but also have higher priority
9472 than other disabled ciphersuites the next time ciphersuites are
9475 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9476 the same ciphersuites as with "HIGH" alone, but in a specific
9477 order where the PSK ciphersuites come first (since they are the
9478 most recently disabled ciphersuites when "HIGH" is parsed).
9480 Also, change ssl_create_cipher_list() (using this new
9481 functionality) such that between otherwise identical
9482 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9487 * Change ssl_create_cipher_list() so that it automatically
9488 arranges the ciphersuites in reasonable order before starting
9489 to process the rule string. Thus, the definition for "DEFAULT"
9490 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9491 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9492 This makes it much easier to arrive at a reasonable default order
9493 in applications for which anonymous ciphers are OK (meaning
9494 that you can't actually use DEFAULT).
9496 *Bodo Moeller; suggested by Victor Duchovni*
9498 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9499 processing) into multiple integers instead of setting
9500 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9501 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9502 (These masks as well as the individual bit definitions are hidden
9503 away into the non-exported interface ssl/ssl_locl.h, so this
9504 change to the definition of the SSL_CIPHER structure shouldn't
9505 affect applications.) This give us more bits for each of these
9506 categories, so there is no longer a need to coagulate AES128 and
9507 AES256 into a single algorithm bit, and to coagulate Camellia128
9508 and Camellia256 into a single algorithm bit, which has led to all
9511 Thus, among other things, the kludge introduced in 0.9.7m and
9512 0.9.8e for masking out AES256 independently of AES128 or masking
9513 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9515 With the change, we also introduce new ciphersuite aliases that
9516 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9521 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9522 Use the leftmost N bytes of the signature input if the input is
9523 larger than the prime q (with N being the size in bytes of q).
9527 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9528 it yet and it is largely untested.
9532 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9536 * Initial incomplete changes to avoid need for function casts in OpenSSL
9537 some compilers (gcc 4.2 and later) reject their use. Safestack is
9538 reimplemented. Update ASN1 to avoid use of legacy functions.
9542 * Win32/64 targets are linked with Winsock2.
9546 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9547 to external functions. This can be used to increase CRL handling
9548 efficiency especially when CRLs are very large by (for example) storing
9549 the CRL revoked certificates in a database.
9553 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9554 new CRLs added to a directory can be used. New command line option
9555 -verify_return_error to s_client and s_server. This causes real errors
9556 to be returned by the verify callback instead of carrying on no matter
9557 what. This reflects the way a "real world" verify callback would behave.
9561 * GOST engine, supporting several GOST algorithms and public key formats.
9562 Kindly donated by Cryptocom.
9566 * Partial support for Issuing Distribution Point CRL extension. CRLs
9567 partitioned by DP are handled but no indirect CRL or reason partitioning
9568 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9569 selected via a scoring technique which handles IDP and AKID in CRLs.
9573 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9574 will ultimately be used for all verify operations: this will remove the
9575 X509_STORE dependency on certificate verification and allow alternative
9576 lookup methods. X509_STORE based implementations of these two callbacks.
9580 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9581 Modify get_crl() to find a valid (unexpired) CRL if possible.
9585 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9586 this would be called X509_CRL_cmp() but that name is already used by
9587 a function that just compares CRL issuer names. Cache several CRL
9588 extensions in X509_CRL structure and cache CRLDP in X509.
9592 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9593 this maps equivalent X509_NAME structures into a consistent structure.
9594 Name comparison can then be performed rapidly using memcmp().
9598 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9603 * Allow digests to supply their own micalg string for S/MIME type using
9604 the ctrl EVP_MD_CTRL_MICALG.
9608 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9609 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9610 ctrl. It can then customise the structure before and/or after signing
9615 * New function OBJ_add_sigid() to allow application defined signature OIDs
9616 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9617 to free up any added signature OIDs.
9621 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9622 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9623 digest and cipher tables. New options added to openssl utility:
9624 list-message-digest-algorithms and list-cipher-algorithms.
9628 * Change the array representation of binary polynomials: the list
9629 of degrees of non-zero coefficients is now terminated with -1.
9630 Previously it was terminated with 0, which was also part of the
9631 value; thus, the array representation was not applicable to
9632 polynomials where t^0 has coefficient zero. This change makes
9633 the array representation useful in a more general context.
9637 * Various modifications and fixes to SSL/TLS cipher string
9638 handling. For ECC, the code now distinguishes between fixed ECDH
9639 with RSA certificates on the one hand and with ECDSA certificates
9640 on the other hand, since these are separate ciphersuites. The
9641 unused code for Fortezza ciphersuites has been removed.
9643 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9644 (not "ECDHE"). For consistency with the code for DH
9645 certificates, use of ECDH certificates is now considered ECDH
9646 authentication, not RSA or ECDSA authentication (the latter is
9647 merely the CA's signing algorithm and not actively used in the
9650 The temporary ciphersuite alias "ECCdraft" is no longer
9651 available, and ECC ciphersuites are no longer excluded from "ALL"
9652 and "DEFAULT". The following aliases now exist for RFC 4492
9653 ciphersuites, most of these by analogy with the DH case:
9655 kECDHr - ECDH cert, signed with RSA
9656 kECDHe - ECDH cert, signed with ECDSA
9657 kECDH - ECDH cert (signed with either RSA or ECDSA)
9658 kEECDH - ephemeral ECDH
9659 ECDH - ECDH cert or ephemeral ECDH
9665 AECDH - anonymous ECDH
9666 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9670 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9671 Use correct micalg parameters depending on digest(s) in signed message.
9675 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9676 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9680 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9681 an engine to register a method. Add ENGINE lookups for methods and
9682 functional reference processing.
9686 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9687 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9692 * New -resign option to smime utility. This adds one or more signers
9693 to an existing PKCS#7 signedData structure. Also -md option to use an
9694 alternative message digest algorithm for signing.
9698 * Tidy up PKCS#7 routines and add new functions to make it easier to
9699 create PKCS7 structures containing multiple signers. Update smime
9700 application to support multiple signers.
9704 * New -macalg option to pkcs12 utility to allow setting of an alternative
9709 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9710 Reorganize PBE internals to lookup from a static table using NIDs,
9711 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9712 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9713 PRF which will be automatically used with PBES2.
9717 * Replace the algorithm specific calls to generate keys in "req" with the
9722 * Update PKCS#7 enveloped data routines to use new API. This is now
9723 supported by any public key method supporting the encrypt operation. A
9724 ctrl is added to allow the public key algorithm to examine or modify
9725 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9730 * Add a ctrl to asn1 method to allow a public key algorithm to express
9731 a default digest type to use. In most cases this will be SHA1 but some
9732 algorithms (such as GOST) need to specify an alternative digest. The
9733 return value indicates how strong the preference is 1 means optional and
9734 2 is mandatory (that is it is the only supported type). Modify
9735 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9736 use the default md. Update openssl utilities to use the default digest
9737 type for signing if it is not explicitly indicated.
9741 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9742 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9743 signing method from the key type. This effectively removes the link
9744 between digests and public key types.
9748 * Add an OID cross reference table and utility functions. Its purpose is to
9749 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9750 rsaEncryption. This will allow some of the algorithm specific hackery
9751 needed to use the correct OID to be removed.
9755 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9756 structures for PKCS7_sign(). They are now set up by the relevant public
9761 * Add provisional EC pkey method with support for ECDSA and ECDH.
9765 * Add support for key derivation (agreement) in the API, DH method and
9770 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9771 public and private key formats. As a side effect these add additional
9772 command line functionality not previously available: DSA signatures can be
9773 generated and verified using pkeyutl and DH key support and generation in
9780 *Oliver Tappe <zooey@hirschkaefer.de>*
9782 * New make target "install_html_docs" installs HTML renditions of the
9785 *Oliver Tappe <zooey@hirschkaefer.de>*
9787 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9788 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9789 support key and parameter generation and add initial key generation
9790 functionality for RSA.
9794 * Add functions for main EVP_PKEY_method operations. The undocumented
9795 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9796 `EVP_PKEY_{encrypt,decrypt}_old`.
9800 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9801 key API, doesn't do much yet.
9805 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9806 public key algorithms. New option to openssl utility:
9807 "list-public-key-algorithms" to print out info.
9811 * Implement the Supported Elliptic Curves Extension for
9812 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9816 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9817 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9821 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9822 utilities such as rsa, dsa, dsaparam etc except they process any key
9827 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9828 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9829 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9834 * Initial support for pluggable public key ASN1.
9835 De-spaghettify the public key ASN1 handling. Move public and private
9836 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9837 algorithm specific handling to a single module within the relevant
9838 algorithm directory. Add functions to allow (near) opaque processing
9839 of public and private key structures.
9843 * Implement the Supported Point Formats Extension for
9844 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9848 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9849 for the psk identity [hint] and the psk callback functions to the
9850 SSL_SESSION, SSL and SSL_CTX structure.
9853 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9857 SSL_CTX_use_psk_identity_hint
9858 SSL_get_psk_identity_hint
9859 SSL_get_psk_identity
9860 SSL_use_psk_identity_hint
9862 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9864 * Add RFC 3161 compliant time stamp request creation, response generation
9865 and response verification functionality.
9867 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9869 * Add initial support for TLS extensions, specifically for the server_name
9870 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9871 have new members for a hostname. The SSL data structure has an
9872 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9873 stored in that context to allow for session resumption, even after the
9874 SSL has been switched to a new SSL_CTX in reaction to a client's
9875 server_name extension.
9877 New functions (subject to change):
9879 SSL_get_servername()
9880 SSL_get_servername_type()
9883 New CTRL codes and macros (subject to change):
9885 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9886 - SSL_CTX_set_tlsext_servername_callback()
9887 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9888 - SSL_CTX_set_tlsext_servername_arg()
9889 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9891 openssl s_client has a new '-servername ...' option.
9893 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9894 '-key2 ...', '-servername_fatal' (subject to change). This allows
9895 testing the HostName extension for a specific single hostname ('-cert'
9896 and '-key' remain fallbacks for handshakes without HostName
9897 negotiation). If the unrecognized_name alert has to be sent, this by
9898 default is a warning; it becomes fatal with the '-servername_fatal'
9901 *Peter Sylvester, Remy Allais, Christophe Renou*
9903 * Whirlpool hash implementation is added.
9907 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9908 bn(64,32). Because of instruction set limitations it doesn't have
9909 any negative impact on performance. This was done mostly in order
9910 to make it possible to share assembler modules, such as bn_mul_mont
9911 implementations, between 32- and 64-bit builds without hassle.
9915 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9916 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9921 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9922 dedicated Montgomery multiplication procedure, is introduced.
9923 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9924 "64-bit" performance on certain 32-bit targets.
9928 * New option SSL_OP_NO_COMP to disable use of compression selectively
9929 in SSL structures. New SSL ctrl to set maximum send fragment size.
9930 Save memory by setting the I/O buffer sizes dynamically instead of
9931 using the maximum available value.
9935 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9936 in addition to the text details.
9940 * Very, very preliminary EXPERIMENTAL support for printing of general
9941 ASN1 structures. This currently produces rather ugly output and doesn't
9942 handle several customised structures at all.
9946 * Integrated support for PVK file format and some related formats such
9947 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9948 these in the 'rsa' and 'dsa' utilities.
9952 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9956 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9957 place for the (very old) "NETSCAPE" format certificates which are now
9958 handled using new ASN1 code equivalents.
9962 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9963 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9964 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9968 * Modify CRL distribution points extension code to print out previously
9969 unsupported fields. Enhance extension setting code to allow setting of
9974 * Add print and set support for Issuing Distribution Point CRL extension.
9978 * Change 'Configure' script to enable Camellia by default.
9985 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9987 * When rejecting SSL/TLS records due to an incorrect version number, never
9988 update s->server with a new major version number. As of
9989 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9990 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9991 the previous behavior could result in a read attempt at NULL when
9992 receiving specific incorrect SSL/TLS records once record payload
9993 protection is active. ([CVE-2010-0740])
9995 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9997 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9998 could be crashed if the relevant tables were not present (e.g. chrooted).
10000 *Tomas Hoger <thoger@redhat.com>*
10002 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10004 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10006 *Martin Olsson, Neel Mehta*
10008 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10009 accommodate for stack sorting, always a write lock!).
10013 * On some versions of WIN32 Heap32Next is very slow. This can cause
10014 excessive delays in the RAND_poll(): over a minute. As a workaround
10015 include a time check in the inner Heap32Next loop too.
10019 * The code that handled flushing of data in SSL/TLS originally used the
10020 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10021 the problem outlined in PR#1949. The fix suggested there however can
10022 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10023 of Apache). So instead simplify the code to flush unconditionally.
10024 This should be fine since flushing with no data to flush is a no op.
10028 * Handle TLS versions 2.0 and later properly and correctly use the
10029 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10030 off ancient servers have a habit of sticking around for a while...
10034 * Modify compression code so it frees up structures without using the
10035 ex_data callbacks. This works around a problem where some applications
10036 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10037 restarting) then use compression (e.g. SSL with compression) later.
10038 This results in significant per-connection memory leaks and
10039 has caused some security issues including CVE-2008-1678 and
10044 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10045 change when encrypting or decrypting.
10049 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10050 connect and renegotiate with servers which do not support RI.
10051 Until RI is more widely deployed this option is enabled by default.
10055 * Add "missing" ssl ctrls to clear options and mode.
10059 * If client attempts to renegotiate and doesn't support RI respond with
10060 a no_renegotiation alert as required by RFC5746. Some renegotiating
10061 TLS clients will continue a connection gracefully when they receive
10062 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10063 waiting for a server hello which it will never receive. Now we treat a
10064 received no_renegotiation alert as a fatal error. This is because
10065 applications requesting a renegotiation might well expect it to succeed
10066 and would have no code in place to handle the server denying it so the
10067 only safe thing to do is to terminate the connection.
10071 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10072 peer supports secure renegotiation and 0 otherwise. Print out peer
10073 renegotiation support in s_client/s_server.
10077 * Replace the highly broken and deprecated SPKAC certification method with
10078 the updated NID creation version. This should correctly handle UTF8.
10082 * Implement RFC5746. Re-enable renegotiation but require the extension
10083 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10084 turns out to be a bad idea. It has been replaced by
10085 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10086 SSL_CTX_set_options(). This is really not recommended unless you
10087 know what you are doing.
10089 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10091 * Fixes to stateless session resumption handling. Use initial_ctx when
10092 issuing and attempting to decrypt tickets in case it has changed during
10093 servername handling. Use a non-zero length session ID when attempting
10094 stateless session resumption: this makes it possible to determine if
10095 a resumption has occurred immediately after receiving server hello
10096 (several places in OpenSSL subtly assume this) instead of later in
10101 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10102 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10103 fixes for a few places where the return code is not checked
10106 *Julia Lawall <julia@diku.dk>*
10108 * Add --strict-warnings option to Configure script to include devteam
10109 warnings in other configurations.
10113 * Add support for --libdir option and LIBDIR variable in makefiles. This
10114 makes it possible to install openssl libraries in locations which
10115 have names other than "lib", for example "/usr/lib64" which some
10118 *Steve Henson, based on patch from Jeremy Utley*
10120 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10121 X690 8.9.12 and can produce some misleading textual output of OIDs.
10123 *Steve Henson, reported by Dan Kaminsky*
10125 * Delete MD2 from algorithm tables. This follows the recommendation in
10126 several standards that it is not used in new applications due to
10127 several cryptographic weaknesses. For binary compatibility reasons
10128 the MD2 API is still compiled in by default.
10132 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10137 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10138 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10141 *Guenter <lists@gknw.net>*
10143 * Fix the server certificate chain building code to use X509_verify_cert(),
10144 it used to have an ad-hoc builder which was unable to cope with anything
10145 other than a simple chain.
10147 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10149 * Don't check self signed certificate signatures in X509_verify_cert()
10150 by default (a flag can override this): it just wastes time without
10151 adding any security. As a useful side effect self signed root CAs
10152 with non-FIPS digests are now usable in FIPS mode.
10156 * In dtls1_process_out_of_seq_message() the check if the current message
10157 is already buffered was missing. For every new message was memory
10158 allocated, allowing an attacker to perform an denial of service attack
10159 with sending out of seq handshake messages until there is no memory
10160 left. Additionally every future message was buffered, even if the
10161 sequence number made no sense and would be part of another handshake.
10162 So only messages with sequence numbers less than 10 in advance will be
10163 buffered. ([CVE-2009-1378])
10165 *Robin Seggelmann, discovered by Daniel Mentz*
10167 * Records are buffered if they arrive with a future epoch to be
10168 processed after finishing the corresponding handshake. There is
10169 currently no limitation to this buffer allowing an attacker to perform
10170 a DOS attack with sending records with future epochs until there is no
10171 memory left. This patch adds the pqueue_size() function to determine
10172 the size of a buffer and limits the record buffer to 100 entries.
10175 *Robin Seggelmann, discovered by Daniel Mentz*
10177 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10178 parent structure is freed. ([CVE-2009-1379])
10182 * Handle non-blocking I/O properly in SSL_shutdown() call.
10184 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10186 * Add `2.5.4.*` OIDs
10188 *Ilya O. <vrghost@gmail.com>*
10190 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10192 * Disable renegotiation completely - this fixes a severe security
10193 problem ([CVE-2009-3555]) at the cost of breaking all
10194 renegotiation. Renegotiation can be re-enabled by setting
10195 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10196 run-time. This is really not recommended unless you know what
10201 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10203 * Don't set val to NULL when freeing up structures, it is freed up by
10204 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10205 zeroing past the valid field. ([CVE-2009-0789])
10207 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10209 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10210 checked correctly. This would allow some invalid signed attributes to
10211 appear to verify correctly. ([CVE-2009-0591])
10213 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10215 * Reject UniversalString and BMPString types with invalid lengths. This
10216 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10217 a legal length. ([CVE-2009-0590])
10221 * Set S/MIME signing as the default purpose rather than setting it
10222 unconditionally. This allows applications to override it at the store
10227 * Permit restricted recursion of ASN1 strings. This is needed in practice
10228 to handle some structures.
10232 * Improve efficiency of mem_gets: don't search whole buffer each time
10235 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10237 * New -hex option for openssl rand.
10241 * Print out UTF8String and NumericString when parsing ASN1.
10245 * Support NumericString type for name components.
10249 * Allow CC in the environment to override the automatically chosen
10250 compiler. Note that nothing is done to ensure flags work with the
10255 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10257 * Properly check EVP_VerifyFinal() and similar return values
10260 *Ben Laurie, Bodo Moeller, Google Security Team*
10262 * Enable TLS extensions by default.
10266 * Allow the CHIL engine to be loaded, whether the application is
10267 multithreaded or not. (This does not release the developer from the
10268 obligation to set up the dynamic locking callbacks.)
10270 *Sander Temme <sander@temme.net>*
10272 * Use correct exit code if there is an error in dgst command.
10274 *Steve Henson; problem pointed out by Roland Dirlewanger*
10276 * Tweak Configure so that you need to say "experimental-jpake" to enable
10277 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10281 * Add experimental JPAKE support, including demo authentication in
10282 s_client and s_server.
10286 * Set the comparison function in v3_addr_canonize().
10288 *Rob Austein <sra@hactrn.net>*
10290 * Add support for XMPP STARTTLS in s_client.
10292 *Philip Paeps <philip@freebsd.org>*
10294 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10295 to ensure that even with this option, only ciphersuites in the
10296 server's preference list will be accepted. (Note that the option
10297 applies only when resuming a session, so the earlier behavior was
10298 just about the algorithm choice for symmetric cryptography.)
10302 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10304 * Fix NULL pointer dereference if a DTLS server received
10305 ChangeCipherSpec as first record ([CVE-2009-1386]).
10309 * Fix a state transition in s3_srvr.c and d1_srvr.c
10310 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10312 *Nagendra Modadugu*
10314 * The fix in 0.9.8c that supposedly got rid of unsafe
10315 double-checked locking was incomplete for RSA blinding,
10316 addressing just one layer of what turns out to have been
10317 doubly unsafe triple-checked locking.
10319 So now fix this for real by retiring the MONT_HELPER macro
10320 in crypto/rsa/rsa_eay.c.
10322 *Bodo Moeller; problem pointed out by Marius Schilder*
10324 * Various precautionary measures:
10326 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10328 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10329 (NB: This would require knowledge of the secret session ticket key
10330 to exploit, in which case you'd be SOL either way.)
10332 - Change bn_nist.c so that it will properly handle input BIGNUMs
10333 outside the expected range.
10335 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10338 *Neel Mehta, Bodo Moeller*
10340 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10341 the load fails. Useful for distros.
10343 *Ben Laurie and the FreeBSD team*
10345 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10349 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10353 * Expand ENGINE to support engine supplied SSL client certificate functions.
10355 This work was sponsored by Logica.
10359 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10360 keystores. Support for SSL/TLS client authentication too.
10361 Not compiled unless enable-capieng specified to Configure.
10363 This work was sponsored by Logica.
10367 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10368 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10369 attribute creation routines such as certificate requests and PKCS#12
10374 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10376 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10377 handshake which could lead to a client crash as found using the
10378 Codenomicon TLS test suite ([CVE-2008-1672])
10380 *Steve Henson, Mark Cox*
10382 * Fix double free in TLS server name extensions which could lead to
10383 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10387 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10389 Clear the error queue to ensure that error entries left from
10390 older function calls do not interfere with the correct operation.
10392 *Lutz Jaenicke, Erik de Castro Lopo*
10394 * Remove root CA certificates of commercial CAs:
10396 The OpenSSL project does not recommend any specific CA and does not
10397 have any policy with respect to including or excluding any CA.
10398 Therefore, it does not make any sense to ship an arbitrary selection
10399 of root CA certificates with the OpenSSL software.
10403 * RSA OAEP patches to fix two separate invalid memory reads.
10404 The first one involves inputs when 'lzero' is greater than
10405 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10406 before the beginning of from). The second one involves inputs where
10407 the 'db' section contains nothing but zeroes (there is a one-byte
10408 invalid read after the end of 'db').
10410 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10412 * Partial backport from 0.9.9-dev:
10414 Introduce bn_mul_mont (dedicated Montgomery multiplication
10415 procedure) as a candidate for BIGNUM assembler implementation.
10416 While 0.9.9-dev uses assembler for various architectures, only
10417 x86_64 is available by default here in the 0.9.8 branch, and
10418 32-bit x86 is available through a compile-time setting.
10420 To try the 32-bit x86 assembler implementation, use Configure
10421 option "enable-montasm" (which exists only for this backport).
10423 As "enable-montasm" for 32-bit x86 disclaims code stability
10424 anyway, in this constellation we activate additional code
10425 backported from 0.9.9-dev for further performance improvements,
10426 namely BN_from_montgomery_word. (To enable this otherwise,
10427 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10429 *Andy Polyakov (backport partially by Bodo Moeller)*
10431 * Add TLS session ticket callback. This allows an application to set
10432 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10433 values. This is useful for key rollover for example where several key
10434 sets may exist with different names.
10438 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10439 This was broken until now in 0.9.8 releases, such that the only way
10440 a registered ENGINE could be used (assuming it initialises
10441 successfully on the host) was to explicitly set it as the default
10442 for the relevant algorithms. This is in contradiction with 0.9.7
10443 behaviour and the documentation. With this fix, when an ENGINE is
10444 registered into a given algorithm's table of implementations, the
10445 'uptodate' flag is reset so that auto-discovery will be used next
10446 time a new context for that algorithm attempts to select an
10449 *Ian Lister (tweaked by Geoff Thorpe)*
10451 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10452 implementation in the following ways:
10454 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10457 Lack of BER streaming support means one pass streaming processing is
10458 only supported if data is detached: setting the streaming flag is
10459 ignored for embedded content.
10461 CMS support is disabled by default and must be explicitly enabled
10462 with the enable-cms configuration option.
10466 * Update the GMP engine glue to do direct copies between BIGNUM and
10467 mpz_t when openssl and GMP use the same limb size. Otherwise the
10468 existing "conversion via a text string export" trick is still used.
10470 *Paul Sheer <paulsheer@gmail.com>*
10472 * Zlib compression BIO. This is a filter BIO which compressed and
10473 uncompresses any data passed through it.
10477 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10478 RFC3394 compatible AES key wrapping.
10482 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10483 sets string data without copying. X509_ALGOR_set0() and
10484 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10485 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10486 from an X509_ATTRIBUTE structure optionally checking it occurs only
10487 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10492 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10493 to get the expected BN_FLG_CONSTTIME behavior.
10495 *Bodo Moeller (Google)*
10499 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10500 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10501 - added some more tests to do_tests.pl
10502 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10503 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10504 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10505 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10506 - various changes to netware.pl to enable gcc-cross builds on Win32
10508 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10509 - various changes to fix missing prototype warnings
10510 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10511 - added AES, WHIRLPOOL and CPUID assembler code to build files
10512 - added missing AES assembler make rules to mk1mf.pl
10513 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10515 *Guenter Knauf <eflash@gmx.net>*
10517 * Implement certificate status request TLS extension defined in RFC3546.
10518 A client can set the appropriate parameters and receive the encoded
10519 OCSP response via a callback. A server can query the supplied parameters
10520 and set the encoded OCSP response in the callback. Add simplified examples
10521 to s_client and s_server.
10525 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10527 * Fix various bugs:
10528 + Binary incompatibility of ssl_ctx_st structure
10529 + DTLS interoperation with non-compliant servers
10530 + Don't call get_session_cb() without proposed session
10531 + Fix ia64 assembler code
10533 *Andy Polyakov, Steve Henson*
10535 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10537 * DTLS Handshake overhaul. There were longstanding issues with
10538 OpenSSL DTLS implementation, which were making it impossible for
10539 RFC 4347 compliant client to communicate with OpenSSL server.
10540 Unfortunately just fixing these incompatibilities would "cut off"
10541 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10542 server keeps tolerating non RFC compliant syntax. The opposite is
10543 not true, 0.9.8f client can not communicate with earlier server.
10544 This update even addresses CVE-2007-4995.
10548 * Changes to avoid need for function casts in OpenSSL: some compilers
10549 (gcc 4.2 and later) reject their use.
10550 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10553 * Add RFC4507 support to OpenSSL. This includes the corrections in
10554 RFC4507bis. The encrypted ticket format is an encrypted encoded
10555 SSL_SESSION structure, that way new session features are automatically
10558 If a client application caches session in an SSL_SESSION structure
10559 support is transparent because tickets are now stored in the encoded
10562 The SSL_CTX structure automatically generates keys for ticket
10563 protection in servers so again support should be possible
10564 with no application modification.
10566 If a client or server wishes to disable RFC4507 support then the option
10567 SSL_OP_NO_TICKET can be set.
10569 Add a TLS extension debugging callback to allow the contents of any client
10570 or server extensions to be examined.
10572 This work was sponsored by Google.
10576 * Add initial support for TLS extensions, specifically for the server_name
10577 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10578 have new members for a hostname. The SSL data structure has an
10579 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10580 stored in that context to allow for session resumption, even after the
10581 SSL has been switched to a new SSL_CTX in reaction to a client's
10582 server_name extension.
10584 New functions (subject to change):
10586 SSL_get_servername()
10587 SSL_get_servername_type()
10590 New CTRL codes and macros (subject to change):
10592 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10593 - SSL_CTX_set_tlsext_servername_callback()
10594 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10595 - SSL_CTX_set_tlsext_servername_arg()
10596 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10598 openssl s_client has a new '-servername ...' option.
10600 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10601 '-key2 ...', '-servername_fatal' (subject to change). This allows
10602 testing the HostName extension for a specific single hostname ('-cert'
10603 and '-key' remain fallbacks for handshakes without HostName
10604 negotiation). If the unrecognized_name alert has to be sent, this by
10605 default is a warning; it becomes fatal with the '-servername_fatal'
10608 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10610 * Add AES and SSE2 assembly language support to VC++ build.
10614 * Mitigate attack on final subtraction in Montgomery reduction.
10618 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10619 (which previously caused an internal error).
10623 * Squeeze another 10% out of IGE mode when in != out.
10627 * AES IGE mode speedup.
10629 *Dean Gaudet (Google)*
10631 * Add the Korean symmetric 128-bit cipher SEED (see
10632 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10633 add SEED ciphersuites from RFC 4162:
10635 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10636 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10637 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10638 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10640 To minimize changes between patchlevels in the OpenSSL 0.9.8
10641 series, SEED remains excluded from compilation unless OpenSSL
10642 is configured with 'enable-seed'.
10644 *KISA, Bodo Moeller*
10646 * Mitigate branch prediction attacks, which can be practical if a
10647 single processor is shared, allowing a spy process to extract
10648 information. For detailed background information, see
10649 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10650 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10651 and Necessary Software Countermeasures"). The core of the change
10652 are new versions BN_div_no_branch() and
10653 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10654 respectively, which are slower, but avoid the security-relevant
10655 conditional branches. These are automatically called by BN_div()
10656 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10657 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10658 remove a conditional branch.
10660 BN_FLG_CONSTTIME is the new name for the previous
10661 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10662 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10663 in the exponent causes BN_mod_exp_mont() to use the alternative
10664 implementation in BN_mod_exp_mont_consttime().) The old name
10665 remains as a deprecated alias.
10667 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10668 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10669 constant-time implementations for more than just exponentiation.
10670 Here too the old name is kept as a deprecated alias.
10672 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10673 the BN_BLINDING structure gets an independent copy of the
10674 modulus. This means that the previous `BIGNUM *m` argument to
10675 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10676 essentially becomes `const BIGNUM *m`, although we can't actually
10677 change this in the header file before 0.9.9. It allows
10678 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10679 enable BN_FLG_CONSTTIME.
10681 *Matthew D Wood (Intel Corp)*
10683 * In the SSL/TLS server implementation, be strict about session ID
10684 context matching (which matters if an application uses a single
10685 external cache for different purposes). Previously,
10686 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10687 set. This did ensure strict client verification, but meant that,
10688 with applications using a single external cache for quite
10689 different requirements, clients could circumvent ciphersuite
10690 restrictions for a given session ID context by starting a session
10691 in a different context.
10695 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10696 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10697 authentication-only ciphersuites.
10701 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10702 not complete and could lead to a possible single byte overflow
10703 ([CVE-2007-5135]) [Ben Laurie]
10705 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10707 * Since AES128 and AES256 (and similarly Camellia128 and
10708 Camellia256) share a single mask bit in the logic of
10709 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10710 kludge to work properly if AES128 is available and AES256 isn't
10711 (or if Camellia128 is available and Camellia256 isn't).
10715 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10716 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10717 When a point or a seed is encoded in a BIT STRING, we need to
10718 prevent the removal of trailing zero bits to get the proper DER
10719 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10720 of a NamedBitList, for which trailing 0 bits need to be removed.)
10724 * Have SSL/TLS server implementation tolerate "mismatched" record
10725 protocol version while receiving ClientHello even if the
10726 ClientHello is fragmented. (The server can't insist on the
10727 particular protocol version it has chosen before the ServerHello
10728 message has informed the client about his choice.)
10732 * Add RFC 3779 support.
10734 *Rob Austein for ARIN, Ben Laurie*
10736 * Load error codes if they are not already present instead of using a
10737 static variable. This allows them to be cleanly unloaded and reloaded.
10738 Improve header file function name parsing.
10742 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10743 or CAPABILITY handshake as required by RFCs.
10745 *Goetz Babin-Ebell*
10747 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10749 * Introduce limits to prevent malicious keys being able to
10750 cause a denial of service. ([CVE-2006-2940])
10752 *Steve Henson, Bodo Moeller*
10754 * Fix ASN.1 parsing of certain invalid structures that can result
10755 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10757 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10758 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10760 * Fix SSL client code which could crash if connecting to a
10761 malicious SSLv2 server. ([CVE-2006-4343])
10763 *Tavis Ormandy and Will Drewry, Google Security Team*
10765 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10766 match only those. Before that, "AES256-SHA" would be interpreted
10767 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10768 the same strength classification in 0.9.7h) as we currently only
10769 have a single AES bit in the ciphersuite description bitmap.
10770 That change, however, also applied to ciphersuite strings such as
10771 "RC4-MD5" that intentionally matched multiple ciphersuites --
10772 namely, SSL 2.0 ciphersuites in addition to the more common ones
10773 from SSL 3.0/TLS 1.0.
10775 So we change the selection algorithm again: Naming an explicit
10776 ciphersuite selects this one ciphersuite, and any other similar
10777 ciphersuite (same bitmap) from *other* protocol versions.
10778 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10779 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10781 Since SSL 2.0 does not have any ciphersuites for which the
10782 128/256 bit distinction would be relevant, this works for now.
10783 The proper fix will be to use different bits for AES128 and
10784 AES256, which would have avoided the problems from the beginning;
10785 however, bits are scarce, so we can only do this in a new release
10786 (not just a patchlevel) when we can change the SSL_CIPHER
10787 definition to split the single 'unsigned long mask' bitmap into
10788 multiple values to extend the available space.
10792 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10794 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10795 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10797 * Add AES IGE and biIGE modes.
10801 * Change the Unix randomness entropy gathering to use poll() when
10802 possible instead of select(), since the latter has some
10803 undesirable limitations.
10805 *Darryl Miles via Richard Levitte and Bodo Moeller*
10807 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10808 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10809 cannot be implicitly activated as part of, e.g., the "AES" alias.
10810 However, please upgrade to OpenSSL 0.9.9[-dev] for
10811 non-experimental use of the ECC ciphersuites to get TLS extension
10812 support, which is required for curve and point format negotiation
10813 to avoid potential handshake problems.
10817 * Disable rogue ciphersuites:
10819 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10820 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10821 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10823 The latter two were purportedly from
10824 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10827 Also deactivate the remaining ciphersuites from
10828 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10829 unofficial, and the ID has long expired.
10833 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10834 dual-core machines) and other potential thread-safety issues.
10838 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10839 versions), which is now available for royalty-free use
10840 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10841 Also, add Camellia TLS ciphersuites from RFC 4132.
10843 To minimize changes between patchlevels in the OpenSSL 0.9.8
10844 series, Camellia remains excluded from compilation unless OpenSSL
10845 is configured with 'enable-camellia'.
10849 * Disable the padding bug check when compression is in use. The padding
10850 bug check assumes the first packet is of even length, this is not
10851 necessarily true if compression is enabled and can result in false
10852 positives causing handshake failure. The actual bug test is ancient
10853 code so it is hoped that implementations will either have fixed it by
10854 now or any which still have the bug do not support compression.
10858 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10860 * When applying a cipher rule check to see if string match is an explicit
10861 cipher suite and only match that one cipher suite if it is.
10865 * Link in manifests for VC++ if needed.
10867 *Austin Ziegler <halostatue@gmail.com>*
10869 * Update support for ECC-based TLS ciphersuites according to
10870 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10871 TLS extensions, which are supported starting with the 0.9.9
10872 branch, not in the OpenSSL 0.9.8 branch).
10876 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10877 opaque EVP_CIPHER_CTX handling.
10881 * Fixes and enhancements to zlib compression code. We now only use
10882 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10883 to conform with the standards mentioned here:
10884 <http://www.zlib.net/DLL_FAQ.txt>
10885 Static zlib linking now works on Windows and the new --with-zlib-include
10886 --with-zlib-lib options to Configure can be used to supply the location
10887 of the headers and library. Gracefully handle case where zlib library
10892 * Several fixes and enhancements to the OID generation code. The old code
10893 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10894 handle numbers larger than ULONG_MAX, truncated printing and had a
10895 non standard OBJ_obj2txt() behaviour.
10899 * Add support for building of engines under engine/ as shared libraries
10900 under VC++ build system.
10904 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10905 Hopefully, we will not see any false combination of paths any more.
10909 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10911 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10912 (part of SSL_OP_ALL). This option used to disable the
10913 countermeasure against man-in-the-middle protocol-version
10914 rollback in the SSL 2.0 server implementation, which is a bad
10915 idea. ([CVE-2005-2969])
10917 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10918 for Information Security, National Institute of Advanced Industrial
10919 Science and Technology [AIST], Japan)*
10921 * Add two function to clear and return the verify parameter flags.
10925 * Keep cipherlists sorted in the source instead of sorting them at
10926 runtime, thus removing the need for a lock.
10930 * Avoid some small subgroup attacks in Diffie-Hellman.
10932 *Nick Mathewson and Ben Laurie*
10934 * Add functions for well-known primes.
10938 * Extended Windows CE support.
10940 *Satoshi Nakamura and Andy Polyakov*
10942 * Initialize SSL_METHOD structures at compile time instead of during
10943 runtime, thus removing the need for a lock.
10947 * Make PKCS7_decrypt() work even if no certificate is supplied by
10948 attempting to decrypt each encrypted key in turn. Add support to
10953 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10955 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10958 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10962 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10963 key into the same file any more.
10967 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10971 * Add -utf8 command line and config file option to 'ca'.
10973 *Stefan <stf@udoma.org*
10975 * Removed the macro des_crypt(), as it seems to conflict with some
10976 libraries. Use DES_crypt().
10980 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10981 involves renaming the source and generated shared-libs for
10982 both. The engines will accept the corrected or legacy ids
10983 ('ncipher' and '4758_cca' respectively) when binding. NB,
10984 this only applies when building 'shared'.
10986 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10988 * Add attribute functions to EVP_PKEY structure. Modify
10989 PKCS12_create() to recognize a CSP name attribute and
10990 use it. Make -CSP option work again in pkcs12 utility.
10994 * Add new functionality to the bn blinding code:
10995 - automatic re-creation of the BN_BLINDING parameters after
10996 a fixed number of uses (currently 32)
10997 - add new function for parameter creation
10998 - introduce flags to control the update behaviour of the
10999 BN_BLINDING parameters
11000 - hide BN_BLINDING structure
11001 Add a second BN_BLINDING slot to the RSA structure to improve
11002 performance when a single RSA object is shared among several
11007 * Add support for DTLS.
11009 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11011 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11012 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11016 * Remove buggy and incomplete DH cert support from
11017 ssl/ssl_rsa.c and ssl/s3_both.c
11021 * Use SHA-1 instead of MD5 as the default digest algorithm for
11022 the `apps/openssl` commands.
11026 * Compile clean with "-Wall -Wmissing-prototypes
11027 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11028 DEBUG_SAFESTACK must also be set.
11032 * Change ./Configure so that certain algorithms can be disabled by default.
11033 The new counterpiece to "no-xxx" is "enable-xxx".
11035 The patented RC5 and MDC2 algorithms will now be disabled unless
11036 "enable-rc5" and "enable-mdc2", respectively, are specified.
11038 (IDEA remains enabled despite being patented. This is because IDEA
11039 is frequently required for interoperability, and there is no license
11040 fee for non-commercial use. As before, "no-idea" can be used to
11041 avoid this algorithm.)
11045 * Add processing of proxy certificates (see RFC 3820). This work was
11046 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11047 EGEE (Enabling Grids for E-science in Europe).
11051 * RC4 performance overhaul on modern architectures/implementations, such
11052 as Intel P4, IA-64 and AMD64.
11056 * New utility extract-section.pl. This can be used specify an alternative
11057 section number in a pod file instead of having to treat each file as
11058 a separate case in Makefile. This can be done by adding two lines to the
11061 =for comment openssl_section:XXX
11063 The blank line is mandatory.
11067 * New arguments -certform, -keyform and -pass for s_client and s_server
11068 to allow alternative format key and certificate files and passphrase
11073 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11074 update associated structures and add various utility functions.
11076 Add new policy related verify parameters, include policy checking in
11077 standard verify code. Enhance 'smime' application with extra parameters
11078 to support policy checking and print out.
11082 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11083 Nehemiah processors. These extensions support AES encryption in hardware
11084 as well as RNG (though RNG support is currently disabled).
11086 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11088 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11092 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11094 *Andy Polyakov and a number of other people*
11096 * Improved PowerPC platform support. Most notably BIGNUM assembler
11097 implementation contributed by IBM.
11099 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11101 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11102 exponent rather than 'unsigned long'. There is a corresponding change to
11103 the new 'rsa_keygen' element of the RSA_METHOD structure.
11105 *Jelte Jansen, Geoff Thorpe*
11107 * Functionality for creating the initial serial number file is now
11108 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11110 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11111 number file to 1, which is bound to cause problems. To avoid
11112 the problems while respecting compatibility between different 0.9.7
11113 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11114 CA.pl for serial number initialization. With the new release 0.9.8,
11115 we can fix the problem directly in the 'ca' utility.)
11119 * Reduced header interdependencies by declaring more opaque objects in
11120 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11121 give fewer recursive includes, which could break lazy source code - so
11122 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11123 developers should define this symbol when building and using openssl to
11124 ensure they track the recommended behaviour, interfaces, [etc], but
11125 backwards-compatible behaviour prevails when this isn't defined.
11129 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11133 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11134 This will generate a random key of the appropriate length based on the
11135 cipher context. The EVP_CIPHER can provide its own random key generation
11136 routine to support keys of a specific form. This is used in the des and
11137 3des routines to generate a key of the correct parity. Update S/MIME
11138 code to use new functions and hence generate correct parity DES keys.
11139 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11140 valid (weak or incorrect parity).
11144 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11145 as looking them up. This is useful when the verified structure may contain
11146 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11147 present unless the new PKCS7_NO_CRL flag is asserted.
11151 * Extend ASN1 oid configuration module. It now additionally accepts the
11154 shortName = some long name, 1.2.3.4
11158 * Reimplemented the BN_CTX implementation. There is now no more static
11159 limitation on the number of variables it can handle nor the depth of the
11160 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11161 information can now expand as required, and rather than having a single
11162 static array of bignums, BN_CTX now uses a linked-list of such arrays
11163 allowing it to expand on demand whilst maintaining the usefulness of
11164 BN_CTX's "bundling".
11168 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11169 to allow all RSA operations to function using a single BN_CTX.
11173 * Preliminary support for certificate policy evaluation and checking. This
11174 is initially intended to pass the tests outlined in "Conformance Testing
11175 of Relying Party Client Certificate Path Processing Logic" v1.07.
11179 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11180 remained unused and not that useful. A variety of other little bignum
11181 tweaks and fixes have also been made continuing on from the audit (see
11186 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11187 associated ASN1, EVP and SSL functions and old ASN1 macros.
11191 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11192 and this should never fail. So the return value from the use of
11193 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11194 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11198 * BN_CTX_get() should return zero-valued bignums, providing the same
11199 initialised value as BN_new().
11201 *Geoff Thorpe, suggested by Ulf Möller*
11203 * Support for inhibitAnyPolicy certificate extension.
11207 * An audit of the BIGNUM code is underway, for which debugging code is
11208 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11209 is considered valid when processing BIGNUMs, and causes execution to
11210 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11211 further steps are taken to deliberately pollute unused data in BIGNUM
11212 structures to try and expose faulty code further on. For now, openssl will
11213 (in its default mode of operation) continue to tolerate the inconsistent
11214 forms that it has tolerated in the past, but authors and packagers should
11215 consider trying openssl and their own applications when compiled with
11216 these debugging symbols defined. It will help highlight potential bugs in
11217 their own code, and will improve the test coverage for OpenSSL itself. At
11218 some point, these tighter rules will become openssl's default to improve
11219 maintainability, though the assert()s and other overheads will remain only
11220 in debugging configurations. See bn.h for more details.
11222 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11224 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11225 that can only be obtained through BN_CTX_new() (which implicitly
11226 initialises it). The presence of this function only made it possible
11227 to overwrite an existing structure (and cause memory leaks).
11231 * Because of the callback-based approach for implementing LHASH as a
11232 template type, lh_insert() adds opaque objects to hash-tables and
11233 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11234 to clean up those corresponding objects before destroying the hash table
11235 (and losing the object pointers). So some over-zealous constifications in
11236 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11237 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11238 prototyped to have "const" restrictions on the object pointers they are
11239 given (and so aren't required to cast them away any more).
11243 * The tmdiff.h API was so ugly and minimal that our own timing utility
11244 (speed) prefers to use its own implementation. The two implementations
11245 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11246 its object type properly exposed (MS_TM) instead of casting to/from
11247 `char *`. This may still change yet if someone realises MS_TM and
11249 aren't necessarily the greatest nomenclatures - but this is what was used
11250 internally to the implementation so I've used that for now.
11254 * Ensure that deprecated functions do not get compiled when
11255 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11256 the self-tests were still using deprecated key-generation functions so
11257 these have been updated also.
11261 * Reorganise PKCS#7 code to separate the digest location functionality
11262 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11263 New function PKCS7_set_digest() to set the digest type for PKCS#7
11264 digestedData type. Add additional code to correctly generate the
11265 digestedData type and add support for this type in PKCS7 initialization
11270 * New function PKCS7_set0_type_other() this initializes a PKCS7
11271 structure of type "other".
11275 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11276 sure the loop does correctly stop and breaking ("division by zero")
11277 modulus operations are not performed. The (pre-generated) prime
11278 table crypto/bn/bn_prime.h was already correct, but it could not be
11279 re-generated on some platforms because of the "division by zero"
11280 situation in the script.
11282 *Ralf S. Engelschall*
11284 * Update support for ECC-based TLS ciphersuites according to
11285 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11286 SHA-1 now is only used for "small" curves (where the
11287 representation of a field element takes up to 24 bytes); for
11288 larger curves, the field element resulting from ECDH is directly
11289 used as premaster secret.
11291 *Douglas Stebila (Sun Microsystems Laboratories)*
11293 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11294 curve secp160r1 to the tests.
11296 *Douglas Stebila (Sun Microsystems Laboratories)*
11298 * Add the possibility to load symbols globally with DSO.
11300 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11302 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11303 control of the error stack.
11307 * Add support for STORE in ENGINE.
11311 * Add the STORE type. The intention is to provide a common interface
11312 to certificate and key stores, be they simple file-based stores, or
11313 HSM-type store, or LDAP stores, or...
11314 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11318 * Add a generic structure called OPENSSL_ITEM. This can be used to
11319 pass a list of arguments to any function as well as provide a way
11320 for a function to pass data back to the caller.
11324 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11325 works like BUF_strdup() but can be used to duplicate a portion of
11326 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11331 * Add the function sk_find_ex() which works like sk_find(), but will
11332 return an index to an element even if an exact match couldn't be
11333 found. The index is guaranteed to point at the element where the
11334 searched-for key would be inserted to preserve sorting order.
11338 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11339 takes an extra flags argument for optional functionality. Currently,
11340 the following flags are defined:
11342 OBJ_BSEARCH_VALUE_ON_NOMATCH
11343 This one gets OBJ_bsearch_ex() to return a pointer to the first
11344 element where the comparing function returns a negative or zero
11347 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11348 This one gets OBJ_bsearch_ex() to return a pointer to the first
11349 element where the comparing function returns zero. This is useful
11350 if there are more than one element where the comparing function
11355 * Make it possible to create self-signed certificates with 'openssl ca'
11356 in such a way that the self-signed certificate becomes part of the
11357 CA database and uses the same mechanisms for serial number generation
11358 as all other certificate signing. The new flag '-selfsign' enables
11359 this functionality. Adapt CA.sh and CA.pl.in.
11363 * Add functionality to check the public key of a certificate request
11364 against a given private. This is useful to check that a certificate
11365 request can be signed by that key (self-signing).
11369 * Make it possible to have multiple active certificates with the same
11370 subject in the CA index file. This is done only if the keyword
11371 'unique_subject' is set to 'no' in the main CA section (default
11372 if 'CA_default') of the configuration file. The value is saved
11373 with the database itself in a separate index attribute file,
11374 named like the index file with '.attr' appended to the name.
11378 * Generate multi-valued AVAs using '+' notation in config files for
11383 * Support for nameConstraints certificate extension.
11387 * Support for policyConstraints certificate extension.
11391 * Support for policyMappings certificate extension.
11395 * Make sure the default DSA_METHOD implementation only uses its
11396 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11397 and change its own handlers to be NULL so as to remove unnecessary
11398 indirection. This lets alternative implementations fallback to the
11399 default implementation more easily.
11403 * Support for directoryName in GeneralName related extensions
11408 * Make it possible to link applications using Makefile.shared.
11409 Make that possible even when linking against static libraries!
11413 * Support for single pass processing for S/MIME signing. This now
11414 means that S/MIME signing can be done from a pipe, in addition
11415 cleartext signing (multipart/signed type) is effectively streaming
11416 and the signed data does not need to be all held in memory.
11418 This is done with a new flag PKCS7_STREAM. When this flag is set
11419 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11420 is done after the data is output (and digests calculated) in
11421 SMIME_write_PKCS7().
11425 * Add full support for -rpath/-R, both in shared libraries and
11426 applications, at least on the platforms where it's known how
11431 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11432 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11433 will now compute a table of multiples of the generator that
11434 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11435 faster (notably in the case of a single point multiplication,
11436 scalar * generator).
11438 *Nils Larsch, Bodo Moeller*
11440 * IPv6 support for certificate extensions. The various extensions
11441 which use the IP:a.b.c.d can now take IPv6 addresses using the
11442 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11447 * Added an ENGINE that implements RSA by performing private key
11448 exponentiations with the GMP library. The conversions to and from
11449 GMP's mpz_t format aren't optimised nor are any montgomery forms
11450 cached, and on x86 it appears OpenSSL's own performance has caught up.
11451 However there are likely to be other architectures where GMP could
11452 provide a boost. This ENGINE is not built in by default, but it can be
11453 specified at Configure time and should be accompanied by the necessary
11454 linker additions, eg;
11455 ./config -DOPENSSL_USE_GMP -lgmp
11459 * "openssl engine" will not display ENGINE/DSO load failure errors when
11460 testing availability of engines with "-t" - the old behaviour is
11461 produced by increasing the feature's verbosity with "-tt".
11465 * ECDSA routines: under certain error conditions uninitialized BN objects
11466 could be freed. Solution: make sure initialization is performed early
11467 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11472 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11473 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11474 software implementations. For DSA and DH, parameter generation can
11475 also be overridden by providing the appropriate method callbacks.
11479 * Change the "progress" mechanism used in key-generation and
11480 primality testing to functions that take a new BN_GENCB pointer in
11481 place of callback/argument pairs. The new API functions have `_ex`
11482 postfixes and the older functions are reimplemented as wrappers for
11483 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11484 declarations of the old functions to help (graceful) attempts to
11485 migrate to the new functions. Also, the new key-generation API
11486 functions operate on a caller-supplied key-structure and return
11487 success/failure rather than returning a key or NULL - this is to
11488 help make "keygen" another member function of RSA_METHOD etc.
11490 Example for using the new callback interface:
11492 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11493 void *my_arg = ...;
11496 BN_GENCB_set(&my_cb, my_callback, my_arg);
11498 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11499 /* For the meaning of a, b in calls to my_callback(), see the
11500 * documentation of the function that calls the callback.
11501 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11502 * my_callback should return 1 if it wants BN_is_prime_ex()
11503 * to continue, or 0 to stop.
11508 * Change the ZLIB compression method to be stateful, and make it
11509 available to TLS with the number defined in
11510 draft-ietf-tls-compression-04.txt.
11514 * Add the ASN.1 structures and functions for CertificatePair, which
11515 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11517 CertificatePair ::= SEQUENCE {
11518 forward [0] Certificate OPTIONAL,
11519 reverse [1] Certificate OPTIONAL,
11520 -- at least one of the pair shall be present -- }
11522 Also implement the PEM functions to read and write certificate
11523 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11525 This needed to be defined, mostly for the sake of the LDAP
11526 attribute crossCertificatePair, but may prove useful elsewhere as
11531 * Make it possible to inhibit symlinking of shared libraries in
11532 Makefile.shared, for Cygwin's sake.
11536 * Extend the BIGNUM API by creating a function
11537 void BN_set_negative(BIGNUM *a, int neg);
11538 and a macro that behave like
11539 int BN_is_negative(const BIGNUM *a);
11541 to avoid the need to access 'a->neg' directly in applications.
11545 * Implement fast modular reduction for pseudo-Mersenne primes
11546 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11547 EC_GROUP_new_curve_GFp() will now automatically use this
11550 *Nils Larsch <nla@trustcenter.de>*
11552 * Add new lock type (CRYPTO_LOCK_BN).
11556 * Change the ENGINE framework to automatically load engines
11557 dynamically from specific directories unless they could be
11558 found to already be built in or loaded. Move all the
11559 current engines except for the cryptodev one to a new
11560 directory engines/.
11561 The engines in engines/ are built as shared libraries if
11562 the "shared" options was given to ./Configure or ./config.
11563 Otherwise, they are inserted in libcrypto.a.
11564 /usr/local/ssl/engines is the default directory for dynamic
11565 engines, but that can be overridden at configure time through
11566 the usual use of --prefix and/or --openssldir, and at run
11567 time with the environment variable OPENSSL_ENGINES.
11569 *Geoff Thorpe and Richard Levitte*
11571 * Add Makefile.shared, a helper makefile to build shared
11572 libraries. Adapt Makefile.org.
11576 * Add version info to Win32 DLLs.
11578 *Peter 'Luna' Runestig" <peter@runestig.com>*
11580 * Add new 'medium level' PKCS#12 API. Certificates and keys
11581 can be added using this API to created arbitrary PKCS#12
11582 files while avoiding the low-level API.
11584 New options to PKCS12_create(), key or cert can be NULL and
11585 will then be omitted from the output file. The encryption
11586 algorithm NIDs can be set to -1 for no encryption, the mac
11587 iteration count can be set to 0 to omit the mac.
11589 Enhance pkcs12 utility by making the -nokeys and -nocerts
11590 options work when creating a PKCS#12 file. New option -nomac
11591 to omit the mac, NONE can be set for an encryption algorithm.
11592 New code is modified to use the enhanced PKCS12_create()
11593 instead of the low-level API.
11597 * Extend ASN1 encoder to support indefinite length constructed
11598 encoding. This can output sequences tags and octet strings in
11599 this form. Modify pk7_asn1.c to support indefinite length
11600 encoding. This is experimental and needs additional code to
11601 be useful, such as an ASN1 bio and some enhanced streaming
11604 Extend template encode functionality so that tagging is passed
11605 down to the template encoder.
11609 * Let 'openssl req' fail if an argument to '-newkey' is not
11610 recognized instead of using RSA as a default.
11614 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11615 As these are not official, they are not included in "ALL";
11616 the "ECCdraft" ciphersuite group alias can be used to select them.
11618 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11620 * Add ECDH engine support.
11622 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11624 * Add ECDH in new directory crypto/ecdh/.
11626 *Douglas Stebila (Sun Microsystems Laboratories)*
11628 * Let BN_rand_range() abort with an error after 100 iterations
11629 without success (which indicates a broken PRNG).
11633 * Change BN_mod_sqrt() so that it verifies that the input value
11634 is really the square of the return value. (Previously,
11635 BN_mod_sqrt would show GIGO behaviour.)
11639 * Add named elliptic curves over binary fields from X9.62, SECG,
11640 and WAP/WTLS; add OIDs that were still missing.
11642 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11644 * Extend the EC library for elliptic curves over binary fields
11645 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11648 EC_GF2m_simple_method
11652 EC_GROUP_new_curve_GF2m
11653 EC_GROUP_set_curve_GF2m
11654 EC_GROUP_get_curve_GF2m
11655 EC_POINT_set_affine_coordinates_GF2m
11656 EC_POINT_get_affine_coordinates_GF2m
11657 EC_POINT_set_compressed_coordinates_GF2m
11659 Point compression for binary fields is disabled by default for
11660 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11663 As binary polynomials are represented as BIGNUMs, various members
11664 of the EC_GROUP and EC_POINT data structures can be shared
11665 between the implementations for prime fields and binary fields;
11666 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11667 are essentially identical to their `..._GFp` counterparts.
11668 (For simplicity, the `..._GFp` prefix has been dropped from
11669 various internal method names.)
11671 An internal 'field_div' method (similar to 'field_mul' and
11672 'field_sqr') has been added; this is used only for binary fields.
11674 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11676 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11677 through methods ('mul', 'precompute_mult').
11679 The generic implementations (now internally called 'ec_wNAF_mul'
11680 and 'ec_wNAF_precomputed_mult') remain the default if these
11681 methods are undefined.
11683 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11685 * New function EC_GROUP_get_degree, which is defined through
11686 EC_METHOD. For curves over prime fields, this returns the bit
11687 length of the modulus.
11689 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11691 * New functions EC_GROUP_dup, EC_POINT_dup.
11692 (These simply call ..._new and ..._copy).
11694 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11696 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11697 Polynomials are represented as BIGNUMs (where the sign bit is not
11698 used) in the following functions [macros]:
11701 BN_GF2m_sub [= BN_GF2m_add]
11702 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11703 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11704 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11706 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11707 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11708 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11709 BN_GF2m_cmp [= BN_ucmp]
11711 (Note that only the 'mod' functions are actually for fields GF(2^m).
11712 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11714 For some functions, an the irreducible polynomial defining a
11715 field can be given as an 'unsigned int[]' with strictly
11716 decreasing elements giving the indices of those bits that are set;
11717 i.e., p[] represents the polynomial
11718 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11720 p[0] > p[1] > ... > p[k] = 0.
11721 This applies to the following functions:
11724 BN_GF2m_mod_mul_arr
11725 BN_GF2m_mod_sqr_arr
11726 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11727 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11728 BN_GF2m_mod_exp_arr
11729 BN_GF2m_mod_sqrt_arr
11730 BN_GF2m_mod_solve_quad_arr
11734 Conversion can be performed by the following functions:
11739 bntest.c has additional tests for binary polynomial arithmetic.
11741 Two implementations for BN_GF2m_mod_div() are available.
11742 The default algorithm simply uses BN_GF2m_mod_inv() and
11743 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11744 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11745 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11747 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11749 * Add new error code 'ERR_R_DISABLED' that can be used when some
11750 functionality is disabled at compile-time.
11752 *Douglas Stebila <douglas.stebila@sun.com>*
11754 * Change default behaviour of 'openssl asn1parse' so that more
11755 information is visible when viewing, e.g., a certificate:
11757 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11758 mode the content of non-printable OCTET STRINGs is output in a
11759 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11760 avoid the appearance of a printable string.
11762 *Nils Larsch <nla@trustcenter.de>*
11764 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11766 EC_GROUP_set_asn1_flag()
11767 EC_GROUP_get_asn1_flag()
11768 EC_GROUP_set_point_conversion_form()
11769 EC_GROUP_get_point_conversion_form()
11770 These control ASN1 encoding details:
11771 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11772 has been set to OPENSSL_EC_NAMED_CURVE.
11773 - Points are encoded in uncompressed form by default; options for
11774 asn1_for are as for point2oct, namely
11775 POINT_CONVERSION_COMPRESSED
11776 POINT_CONVERSION_UNCOMPRESSED
11777 POINT_CONVERSION_HYBRID
11779 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11781 EC_GROUP_set_seed()
11782 EC_GROUP_get0_seed()
11783 EC_GROUP_get_seed_len()
11784 This is used only for ASN1 purposes (so far).
11786 *Nils Larsch <nla@trustcenter.de>*
11788 * Add 'field_type' member to EC_METHOD, which holds the NID
11789 of the appropriate field type OID. The new function
11790 EC_METHOD_get_field_type() returns this value.
11792 *Nils Larsch <nla@trustcenter.de>*
11795 EC_POINT_point2bn()
11796 EC_POINT_bn2point()
11797 EC_POINT_point2hex()
11798 EC_POINT_hex2point()
11799 providing useful interfaces to EC_POINT_point2oct() and
11800 EC_POINT_oct2point().
11802 *Nils Larsch <nla@trustcenter.de>*
11804 * Change internals of the EC library so that the functions
11805 EC_GROUP_set_generator()
11806 EC_GROUP_get_generator()
11807 EC_GROUP_get_order()
11808 EC_GROUP_get_cofactor()
11809 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11810 to methods, which would lead to unnecessary code duplication when
11811 adding different types of curves.
11813 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11815 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11816 arithmetic, and such that modified wNAFs are generated
11817 (which avoid length expansion in many cases).
11821 * Add a function EC_GROUP_check_discriminant() (defined via
11822 EC_METHOD) that verifies that the curve discriminant is non-zero.
11824 Add a function EC_GROUP_check() that makes some sanity tests
11825 on a EC_GROUP, its generator and order. This includes
11826 EC_GROUP_check_discriminant().
11828 *Nils Larsch <nla@trustcenter.de>*
11830 * Add ECDSA in new directory crypto/ecdsa/.
11832 Add applications 'openssl ecparam' and 'openssl ecdsa'
11833 (these are based on 'openssl dsaparam' and 'openssl dsa').
11835 ECDSA support is also included in various other files across the
11836 library. Most notably,
11837 - 'openssl req' now has a '-newkey ecdsa:file' option;
11838 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11839 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11840 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11841 them suitable for ECDSA where domain parameters must be
11842 extracted before the specific public key;
11843 - ECDSA engine support has been added.
11845 *Nils Larsch <nla@trustcenter.de>*
11847 * Include some named elliptic curves, and add OIDs from X9.62,
11848 SECG, and WAP/WTLS. Each curve can be obtained from the new
11850 EC_GROUP_new_by_curve_name(),
11851 and the list of available named curves can be obtained with
11852 EC_get_builtin_curves().
11853 Also add a 'curve_name' member to EC_GROUP objects, which can be
11855 EC_GROUP_set_curve_name()
11856 EC_GROUP_get_curve_name()
11858 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11860 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11861 was actually never needed) and in BN_mul(). The removal in BN_mul()
11862 required a small change in bn_mul_part_recursive() and the addition
11863 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11864 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11865 bn_sub_words() and bn_add_words() except they take arrays with
11870 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11872 * Cleanse PEM buffers before freeing them since they may contain
11875 *Benjamin Bennett <ben@psc.edu>*
11877 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11878 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11879 authentication-only ciphersuites.
11883 * Since AES128 and AES256 share a single mask bit in the logic of
11884 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11885 kludge to work properly if AES128 is available and AES256 isn't.
11889 * Expand security boundary to match 1.1.1 module.
11893 * Remove redundant features: hash file source, editing of test vectors
11894 modify fipsld to use external fips_premain.c signature.
11898 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11899 run algorithm test programs.
11903 * Make algorithm test programs more tolerant of whitespace.
11907 * Have SSL/TLS server implementation tolerate "mismatched" record
11908 protocol version while receiving ClientHello even if the
11909 ClientHello is fragmented. (The server can't insist on the
11910 particular protocol version it has chosen before the ServerHello
11911 message has informed the client about his choice.)
11915 * Load error codes if they are not already present instead of using a
11916 static variable. This allows them to be cleanly unloaded and reloaded.
11920 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11922 * Introduce limits to prevent malicious keys being able to
11923 cause a denial of service. ([CVE-2006-2940])
11925 *Steve Henson, Bodo Moeller*
11927 * Fix ASN.1 parsing of certain invalid structures that can result
11928 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11930 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11931 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11933 * Fix SSL client code which could crash if connecting to a
11934 malicious SSLv2 server. ([CVE-2006-4343])
11936 *Tavis Ormandy and Will Drewry, Google Security Team*
11938 * Change ciphersuite string processing so that an explicit
11939 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11940 will no longer include "AES128-SHA"), and any other similar
11941 ciphersuite (same bitmap) from *other* protocol versions (so that
11942 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11943 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11944 changes from 0.9.8b and 0.9.8d.
11948 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11950 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11951 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11953 * Change the Unix randomness entropy gathering to use poll() when
11954 possible instead of select(), since the latter has some
11955 undesirable limitations.
11957 *Darryl Miles via Richard Levitte and Bodo Moeller*
11959 * Disable rogue ciphersuites:
11961 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11962 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11963 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11965 The latter two were purportedly from
11966 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11969 Also deactivate the remaining ciphersuites from
11970 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11971 unofficial, and the ID has long expired.
11975 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11976 dual-core machines) and other potential thread-safety issues.
11980 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11982 * Adapt fipsld and the build system to link against the validated FIPS
11983 module in FIPS mode.
11987 * Fixes for VC++ 2005 build under Windows.
11991 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11992 from a Windows bash shell such as MSYS. It is autodetected from the
11993 "config" script when run from a VC++ environment. Modify standard VC++
11994 build to use fipscanister.o from the GNU make build.
11998 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12000 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12001 The value now differs depending on if you build for FIPS or not.
12002 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12003 safely run with a non-FIPSed libcrypto, as it may crash because of
12004 the difference induced by this change.
12008 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12010 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12011 (part of SSL_OP_ALL). This option used to disable the
12012 countermeasure against man-in-the-middle protocol-version
12013 rollback in the SSL 2.0 server implementation, which is a bad
12014 idea. ([CVE-2005-2969])
12016 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12017 for Information Security, National Institute of Advanced Industrial
12018 Science and Technology [AIST, Japan)]*
12020 * Minimal support for X9.31 signatures and PSS padding modes. This is
12021 mainly for FIPS compliance and not fully integrated at this stage.
12025 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12026 the exponentiation using a fixed-length exponent. (Otherwise,
12027 the information leaked through timing could expose the secret key
12028 after many signatures; cf. Bleichenbacher's attack on DSA with
12033 * Make a new fixed-window mod_exp implementation the default for
12034 RSA, DSA, and DH private-key operations so that the sequence of
12035 squares and multiplies and the memory access pattern are
12036 independent of the particular secret key. This will mitigate
12037 cache-timing and potential related attacks.
12039 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12040 and this is automatically used by BN_mod_exp_mont() if the new flag
12041 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12042 will use this BN flag for private exponents unless the flag
12043 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12044 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12046 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12048 * Change the client implementation for SSLv23_method() and
12049 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12050 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12051 (Previously, the SSL 2.0 backwards compatible Client Hello
12052 message format would be used even with SSL_OP_NO_SSLv2.)
12056 * Add support for smime-type MIME parameter in S/MIME messages which some
12061 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12062 a threadsafe manner. Modify rsa code to use new function and add calls
12063 to dsa and dh code (which had race conditions before).
12067 * Include the fixed error library code in the C error file definitions
12068 instead of fixing them up at runtime. This keeps the error code
12069 structures constant.
12073 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12075 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12078 * Fixes for newer kerberos headers. NB: the casts are needed because
12079 the 'length' field is signed on one version and unsigned on another
12080 with no (?) obvious way to tell the difference, without these VC++
12081 complains. Also the "definition" of FAR (blank) is no longer included
12082 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12083 some needed definitions.
12087 * Undo Cygwin change.
12091 * Added support for proxy certificates according to RFC 3820.
12092 Because they may be a security thread to unaware applications,
12093 they must be explicitly allowed in run-time. See
12094 docs/HOWTO/proxy_certificates.txt for further information.
12098 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12100 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12101 server and client random values. Previously
12102 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12103 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12105 This change has negligible security impact because:
12107 1. Server and client random values still have 24 bytes of pseudo random
12110 2. Server and client random values are sent in the clear in the initial
12113 3. The master secret is derived using the premaster secret (48 bytes in
12114 size for static RSA ciphersuites) as well as client server and random
12117 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12120 *Stephen Henson, reported by UK NISCC*
12122 * Use Windows randomness collection on Cygwin.
12126 * Fix hang in EGD/PRNGD query when communication socket is closed
12127 prematurely by EGD/PRNGD.
12129 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12131 * Prompt for pass phrases when appropriate for PKCS12 input format.
12135 * Back-port of selected performance improvements from development
12136 branch, as well as improved support for PowerPC platforms.
12140 * Add lots of checks for memory allocation failure, error codes to indicate
12141 failure and freeing up memory if a failure occurs.
12143 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12145 * Add new -passin argument to dgst.
12149 * Perform some character comparisons of different types in X509_NAME_cmp:
12150 this is needed for some certificates that re-encode DNs into UTF8Strings
12151 (in violation of RFC3280) and can't or won't issue name rollover
12156 * Make an explicit check during certificate validation to see that
12157 the CA setting in each certificate on the chain is correct. As a
12158 side effect always do the following basic checks on extensions,
12159 not just when there's an associated purpose to the check:
12161 - if there is an unhandled critical extension (unless the user
12162 has chosen to ignore this fault)
12163 - if the path length has been exceeded (if one is set at all)
12164 - that certain extensions fit the associated purpose (if one has
12169 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12171 * Avoid a race condition when CRLs are checked in a multi threaded
12172 environment. This would happen due to the reordering of the revoked
12173 entries during signature checking and serial number lookup. Now the
12174 encoding is cached and the serial number sort performed under a lock.
12175 Add new STACK function sk_is_sorted().
12179 * Add Delta CRL to the extension code.
12183 * Various fixes to s3_pkt.c so alerts are sent properly.
12185 *David Holmes <d.holmes@f5.com>*
12187 * Reduce the chances of duplicate issuer name and serial numbers (in
12188 violation of RFC3280) using the OpenSSL certificate creation utilities.
12189 This is done by creating a random 64 bit value for the initial serial
12190 number when a serial number file is created or when a self signed
12191 certificate is created using 'openssl req -x509'. The initial serial
12192 number file is created using 'openssl x509 -next_serial' in CA.pl
12193 rather than being initialized to 1.
12197 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12199 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12200 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12202 *Joe Orton, Steve Henson*
12204 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12207 *Joe Orton, Steve Henson*
12209 * Make it possible to have multiple active certificates with the same
12210 subject in the CA index file. This is done only if the keyword
12211 'unique_subject' is set to 'no' in the main CA section (default
12212 if 'CA_default') of the configuration file. The value is saved
12213 with the database itself in a separate index attribute file,
12214 named like the index file with '.attr' appended to the name.
12218 * X509 verify fixes. Disable broken certificate workarounds when
12219 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12220 keyUsage extension present. Don't accept CRLs with unhandled critical
12221 extensions: since verify currently doesn't process CRL extensions this
12222 rejects a CRL with *any* critical extensions. Add new verify error codes
12227 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12228 A clarification of RFC2560 will require the use of OCTET STRINGs and
12229 some implementations cannot handle the current raw format. Since OpenSSL
12230 copies and compares OCSP nonces as opaque blobs without any attempt at
12231 parsing them this should not create any compatibility issues.
12235 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12236 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12237 this HMAC (and other) operations are several times slower than OpenSSL
12242 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12244 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12246 * Use the correct content when signing type "other".
12250 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12252 * Fix various bugs revealed by running the NISCC test suite:
12254 Stop out of bounds reads in the ASN1 code when presented with
12255 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12257 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12259 If verify callback ignores invalid public key errors don't try to check
12260 certificate signature with the NULL public key.
12264 * New -ignore_err option in ocsp application to stop the server
12265 exiting on the first error in a request.
12269 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12270 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12275 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12276 extra data after the compression methods not only for TLS 1.0
12277 but also for SSL 3.0 (as required by the specification).
12279 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12281 * Change X509_certificate_type() to mark the key as exported/exportable
12282 when it's 512 *bits* long, not 512 bytes.
12286 * Change AES_cbc_encrypt() so it outputs exact multiple of
12287 blocks during encryption.
12291 * Various fixes to base64 BIO and non blocking I/O. On write
12292 flushes were not handled properly if the BIO retried. On read
12293 data was not being buffered properly and had various logic bugs.
12294 This also affects blocking I/O when the data being decoded is a
12299 * Various S/MIME bugfixes and compatibility changes:
12300 output correct application/pkcs7 MIME type if
12301 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12302 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12303 of files as .eml work). Correctly handle very long lines in MIME
12308 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12310 * Countermeasure against the Klima-Pokorny-Rosa extension of
12311 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12312 a protocol version number mismatch like a decryption error
12313 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12317 * Turn on RSA blinding by default in the default implementation
12318 to avoid a timing attack. Applications that don't want it can call
12319 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12320 They would be ill-advised to do so in most cases.
12322 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12324 * Change RSA blinding code so that it works when the PRNG is not
12325 seeded (in this case, the secret RSA exponent is abused as
12326 an unpredictable seed -- if it is not unpredictable, there
12327 is no point in blinding anyway). Make RSA blinding thread-safe
12328 by remembering the creator's thread ID in rsa->blinding and
12329 having all other threads use local one-time blinding factors
12330 (this requires more computation than sharing rsa->blinding, but
12331 avoids excessive locking; and if an RSA object is not shared
12332 between threads, blinding will still be very fast).
12336 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12337 ENGINE as defaults for all supported algorithms irrespective of
12338 the 'flags' parameter. 'flags' is now honoured, so applications
12339 should make sure they are passing it correctly.
12343 * Target "mingw" now allows native Windows code to be generated in
12344 the Cygwin environment as well as with the MinGW compiler.
12348 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12350 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12351 via timing by performing a MAC computation even if incorrect
12352 block cipher padding has been found. This is a countermeasure
12353 against active attacks where the attacker has to distinguish
12354 between bad padding and a MAC verification error. ([CVE-2003-0078])
12356 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12357 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12358 Martin Vuagnoux (EPFL, Ilion)*
12360 * Make the no-err option work as intended. The intention with no-err
12361 is not to have the whole error stack handling routines removed from
12362 libcrypto, it's only intended to remove all the function name and
12363 reason texts, thereby removing some of the footprint that may not
12364 be interesting if those errors aren't displayed anyway.
12366 NOTE: it's still possible for any application or module to have its
12367 own set of error texts inserted. The routines are there, just not
12368 used by default when no-err is given.
12372 * Add support for FreeBSD on IA64.
12374 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12376 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12377 Kerberos function mit_des_cbc_cksum(). Before this change,
12378 the value returned by DES_cbc_cksum() was like the one from
12379 mit_des_cbc_cksum(), except the bytes were swapped.
12381 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12383 * Allow an application to disable the automatic SSL chain building.
12384 Before this a rather primitive chain build was always performed in
12385 ssl3_output_cert_chain(): an application had no way to send the
12386 correct chain if the automatic operation produced an incorrect result.
12388 Now the chain builder is disabled if either:
12390 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12392 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12394 The reasoning behind this is that an application would not want the
12395 auto chain building to take place if extra chain certificates are
12396 present and it might also want a means of sending no additional
12397 certificates (for example the chain has two certificates and the
12402 * Add the possibility to build without the ENGINE framework.
12404 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12406 * Under Win32 gmtime() can return NULL: check return value in
12407 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12411 * DSA routines: under certain error conditions uninitialized BN objects
12412 could be freed. Solution: make sure initialization is performed early
12413 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12414 Nils Larsch <nla@trustcenter.de> via PR#459)
12418 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12419 checked on reconnect on the client side, therefore session resumption
12420 could still fail with a "ssl session id is different" error. This
12421 behaviour is masked when SSL_OP_ALL is used due to
12422 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12423 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12424 followup to PR #377.
12428 * IA-32 assembler support enhancements: unified ELF targets, support
12429 for SCO/Caldera platforms, fix for Cygwin shared build.
12433 * Add support for FreeBSD on sparc64. As a consequence, support for
12434 FreeBSD on non-x86 processors is separate from x86 processors on
12435 the config script, much like the NetBSD support.
12437 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12439 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12441 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12444 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12445 code (06) was taken as the first octet of the session ID and the last
12446 octet was ignored consequently. As a result SSLv2 client side session
12447 caching could not have worked due to the session ID mismatch between
12449 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12454 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12455 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12460 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12461 seems that in spite of existing for more than a year, many application
12462 author have done nothing to provide the necessary callbacks, which
12463 means that this particular engine will not work properly anywhere.
12464 This is a very unfortunate situation which forces us, in the name
12465 of usability, to give the hw_ncipher.c a static lock, which is part
12467 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12468 appear in 0.9.8 or later. We EXPECT application authors to have
12469 dealt properly with this when 0.9.8 is released (unless we actually
12470 make such changes in the libcrypto locking code that changes will
12471 have to be made anyway).
12475 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12476 octets have been read, EOF or an error occurs. Without this change
12477 some truncated ASN1 structures will not produce an error.
12481 * Disable Heimdal support, since it hasn't been fully implemented.
12482 Still give the possibility to force the use of Heimdal, but with
12483 warnings and a request that patches get sent to openssl-dev.
12487 * Add the VC-CE target, introduce the WINCE sysname, and add
12488 INSTALL.WCE and appropriate conditionals to make it build.
12490 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12492 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12493 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12494 edit numbers of the version.
12496 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12498 * Introduce safe string copy and catenation functions
12499 (BUF_strlcpy() and BUF_strlcat()).
12501 *Ben Laurie (CHATS) and Richard Levitte*
12503 * Avoid using fixed-size buffers for one-line DNs.
12505 *Ben Laurie (CHATS)*
12507 * Add BUF_MEM_grow_clean() to avoid information leakage when
12508 resizing buffers containing secrets, and use where appropriate.
12510 *Ben Laurie (CHATS)*
12512 * Avoid using fixed size buffers for configuration file location.
12514 *Ben Laurie (CHATS)*
12516 * Avoid filename truncation for various CA files.
12518 *Ben Laurie (CHATS)*
12520 * Use sizeof in preference to magic numbers.
12522 *Ben Laurie (CHATS)*
12524 * Avoid filename truncation in cert requests.
12526 *Ben Laurie (CHATS)*
12528 * Add assertions to check for (supposedly impossible) buffer
12531 *Ben Laurie (CHATS)*
12533 * Don't cache truncated DNS entries in the local cache (this could
12534 potentially lead to a spoofing attack).
12536 *Ben Laurie (CHATS)*
12538 * Fix various buffers to be large enough for hex/decimal
12539 representations in a platform independent manner.
12541 *Ben Laurie (CHATS)*
12543 * Add CRYPTO_realloc_clean() to avoid information leakage when
12544 resizing buffers containing secrets, and use where appropriate.
12546 *Ben Laurie (CHATS)*
12548 * Add BIO_indent() to avoid much slightly worrying code to do
12551 *Ben Laurie (CHATS)*
12553 * Convert sprintf()/BIO_puts() to BIO_printf().
12555 *Ben Laurie (CHATS)*
12557 * buffer_gets() could terminate with the buffer only half
12560 *Ben Laurie (CHATS)*
12562 * Add assertions to prevent user-supplied crypto functions from
12563 overflowing internal buffers by having large block sizes, etc.
12565 *Ben Laurie (CHATS)*
12567 * New OPENSSL_assert() macro (similar to assert(), but enabled
12570 *Ben Laurie (CHATS)*
12572 * Eliminate unused copy of key in RC4.
12574 *Ben Laurie (CHATS)*
12576 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12578 *Ben Laurie (CHATS)*
12580 * Fix off-by-one error in EGD path.
12582 *Ben Laurie (CHATS)*
12584 * If RANDFILE path is too long, ignore instead of truncating.
12586 *Ben Laurie (CHATS)*
12588 * Eliminate unused and incorrectly sized X.509 structure
12591 *Ben Laurie (CHATS)*
12593 * Eliminate unused and dangerous function knumber().
12595 *Ben Laurie (CHATS)*
12597 * Eliminate unused and dangerous structure, KSSL_ERR.
12599 *Ben Laurie (CHATS)*
12601 * Protect against overlong session ID context length in an encoded
12602 session object. Since these are local, this does not appear to be
12605 *Ben Laurie (CHATS)*
12607 * Change from security patch (see 0.9.6e below) that did not affect
12608 the 0.9.6 release series:
12610 Remote buffer overflow in SSL3 protocol - an attacker could
12611 supply an oversized master key in Kerberos-enabled versions.
12614 *Ben Laurie (CHATS)*
12616 * Change the SSL kerb5 codes to match RFC 2712.
12620 * Make -nameopt work fully for req and add -reqopt switch.
12622 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12624 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12626 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12628 * Make sure tests can be performed even if the corresponding algorithms
12629 have been removed entirely. This was also the last step to make
12630 OpenSSL compilable with DJGPP under all reasonable conditions.
12632 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12634 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12635 to allow version independent disabling of normally unselected ciphers,
12636 which may be activated as a side-effect of selecting a single cipher.
12638 (E.g., cipher list string "RSA" enables ciphersuites that are left
12639 out of "ALL" because they do not provide symmetric encryption.
12640 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12642 *Lutz Jaenicke, Bodo Moeller*
12644 * Add appropriate support for separate platform-dependent build
12645 directories. The recommended way to make a platform-dependent
12646 build directory is the following (tested on Linux), maybe with
12649 # Place yourself outside of the OpenSSL source tree. In
12650 # this example, the environment variable OPENSSL_SOURCE
12651 # is assumed to contain the absolute OpenSSL source directory.
12652 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12653 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12654 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12655 mkdir -p `dirname $F`
12656 ln -s $OPENSSL_SOURCE/$F $F
12659 To be absolutely sure not to disturb the source tree, a "make clean"
12660 is a good thing. If it isn't successful, don't worry about it,
12661 it probably means the source directory is very clean.
12665 * Make sure any ENGINE control commands make local copies of string
12666 pointers passed to them whenever necessary. Otherwise it is possible
12667 the caller may have overwritten (or deallocated) the original string
12668 data when a later ENGINE operation tries to use the stored values.
12670 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12672 * Improve diagnostics in file reading and command-line digests.
12674 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12676 * Add AES modes CFB and OFB to the object database. Correct an
12677 error in AES-CFB decryption.
12681 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12682 allows existing EVP_CIPHER_CTX structures to be reused after
12683 calling `EVP_*Final()`. This behaviour is used by encryption
12684 BIOs and some applications. This has the side effect that
12685 applications must explicitly clean up cipher contexts with
12686 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12690 * Check the values of dna and dnb in bn_mul_recursive before calling
12691 bn_mul_comba (a non zero value means the a or b arrays do not contain
12692 n2 elements) and fallback to bn_mul_normal if either is not zero.
12696 * Fix escaping of non-ASCII characters when using the -subj option
12697 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12701 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12702 form for "surname", serialNumber has no short form.
12703 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12704 therefore remove "mail" short name for "internet 7".
12705 The OID for unique identifiers in X509 certificates is
12706 x500UniqueIdentifier, not uniqueIdentifier.
12707 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12711 * Add an "init" command to the ENGINE config module and auto initialize
12712 ENGINEs. Without any "init" command the ENGINE will be initialized
12713 after all ctrl commands have been executed on it. If init=1 the
12714 ENGINE is initialized at that point (ctrls before that point are run
12715 on the uninitialized ENGINE and after on the initialized one). If
12716 init=0 then the ENGINE will not be initialized at all.
12720 * Fix the 'app_verify_callback' interface so that the user-defined
12721 argument is actually passed to the callback: In the
12722 SSL_CTX_set_cert_verify_callback() prototype, the callback
12723 declaration has been changed from
12726 int (*cb)(X509_STORE_CTX *,void *);
12727 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12728 i=s->ctx->app_verify_callback(&ctx)
12729 has been changed into
12730 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12732 To update applications using SSL_CTX_set_cert_verify_callback(),
12733 a dummy argument can be added to their callback functions.
12735 *D. K. Smetters <smetters@parc.xerox.com>*
12737 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12739 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12741 * Add and OPENSSL_LOAD_CONF define which will cause
12742 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12743 This allows older applications to transparently support certain
12744 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12745 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12746 load the config file and OPENSSL_add_all_algorithms_conf() which will
12747 always load it have also been added.
12751 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12752 Adjust NIDs and EVP layer.
12754 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12756 * Config modules support in openssl utility.
12758 Most commands now load modules from the config file,
12759 though in a few (such as version) this isn't done
12760 because it couldn't be used for anything.
12762 In the case of ca and req the config file used is
12763 the same as the utility itself: that is the -config
12764 command line option can be used to specify an
12769 * Move default behaviour from OPENSSL_config(). If appname is NULL
12770 use "openssl_conf" if filename is NULL use default openssl config file.
12774 * Add an argument to OPENSSL_config() to allow the use of an alternative
12775 config section name. Add a new flag to tolerate a missing config file
12776 and move code to CONF_modules_load_file().
12780 * Support for crypto accelerator cards from Accelerated Encryption
12781 Processing, www.aep.ie. (Use engine 'aep')
12782 The support was copied from 0.9.6c [engine] and adapted/corrected
12783 to work with the new engine framework.
12785 *AEP Inc. and Richard Levitte*
12787 * Support for SureWare crypto accelerator cards from Baltimore
12788 Technologies. (Use engine 'sureware')
12789 The support was copied from 0.9.6c [engine] and adapted
12790 to work with the new engine framework.
12794 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12795 make the newer ENGINE framework commands for the CHIL engine work.
12797 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12799 * Make it possible to produce shared libraries on ReliantUNIX.
12801 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12803 * Add the configuration target debug-linux-ppro.
12804 Make 'openssl rsa' use the general key loading routines
12805 implemented in `apps.c`, and make those routines able to
12806 handle the key format FORMAT_NETSCAPE and the variant
12809 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12811 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12813 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12815 * Add -keyform to rsautl, and document -engine.
12817 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12819 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12820 BIO_R_NO_SUCH_FILE error code rather than the generic
12821 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12825 * Add new functions
12826 ERR_peek_last_error
12827 ERR_peek_last_error_line
12828 ERR_peek_last_error_line_data.
12829 These are similar to
12831 ERR_peek_error_line
12832 ERR_peek_error_line_data,
12833 but report on the latest error recorded rather than the first one
12834 still in the error queue.
12836 *Ben Laurie, Bodo Moeller*
12838 * default_algorithms option in ENGINE config module. This allows things
12840 default_algorithms = ALL
12841 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12845 * Preliminary ENGINE config module.
12849 * New experimental application configuration code.
12853 * Change the AES code to follow the same name structure as all other
12854 symmetric ciphers, and behave the same way. Move everything to
12855 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12857 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12859 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12861 *Ben Laurie and Theo de Raadt*
12863 * Add option to output public keys in req command.
12865 *Massimiliano Pala madwolf@openca.org*
12867 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12868 (up to about 10% better than before for P-192 and P-224).
12872 * New functions/macros
12874 SSL_CTX_set_msg_callback(ctx, cb)
12875 SSL_CTX_set_msg_callback_arg(ctx, arg)
12876 SSL_set_msg_callback(ssl, cb)
12877 SSL_set_msg_callback_arg(ssl, arg)
12879 to request calling a callback function
12881 void cb(int write_p, int version, int content_type,
12882 const void *buf, size_t len, SSL *ssl, void *arg)
12884 whenever a protocol message has been completely received
12885 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12886 protocol version according to which the SSL library interprets
12887 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12888 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12889 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12890 specification (change_cipher_spec(20), alert(21), handshake(22)).
12891 'buf' and 'len' point to the actual message, 'ssl' to the
12892 SSL object, and 'arg' is the application-defined value set by
12893 SSL[_CTX]_set_msg_callback_arg().
12895 'openssl s_client' and 'openssl s_server' have new '-msg' options
12896 to enable a callback that displays all protocol messages.
12900 * Change the shared library support so shared libraries are built as
12901 soon as the corresponding static library is finished, and thereby get
12902 openssl and the test programs linked against the shared library.
12903 This still only happens when the keyword "shard" has been given to
12904 the configuration scripts.
12906 NOTE: shared library support is still an experimental thing, and
12907 backward binary compatibility is still not guaranteed.
12909 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12911 * Add support for Subject Information Access extension.
12913 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12915 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12916 additional bytes when new memory had to be allocated, not just
12917 when reusing an existing buffer.
12921 * New command line and configuration option 'utf8' for the req command.
12922 This allows field values to be specified as UTF8 strings.
12926 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12927 runs for the former and machine-readable output for the latter.
12931 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12932 of the e-mail address in the DN (i.e., it will go into a certificate
12933 extension only). The new configuration file option 'email_in_dn = no'
12934 has the same effect.
12936 *Massimiliano Pala madwolf@openca.org*
12938 * Change all functions with names starting with `des_` to be starting
12939 with `DES_` instead. Add wrappers that are compatible with libdes,
12940 but are named `_ossl_old_des_*`. Finally, add macros that map the
12941 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12942 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12943 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12946 Since we provide two compatibility mappings, the user needs to
12947 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12948 compatibility is desired. The default (i.e., when that macro
12949 isn't defined) is OpenSSL 0.9.6c compatibility.
12951 There are also macros that enable and disable the support of old
12952 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12953 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12954 are defined, the default will apply: to support the old des routines.
12956 In either case, one must include openssl/des.h to get the correct
12957 definitions. Do not try to just include openssl/des_old.h, that
12960 NOTE: This is a major break of an old API into a new one. Software
12961 authors are encouraged to switch to the `DES_` style functions. Some
12962 time in the future, des_old.h and the libdes compatibility functions
12963 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12964 default), and then completely removed.
12968 * Test for certificates which contain unsupported critical extensions.
12969 If such a certificate is found during a verify operation it is
12970 rejected by default: this behaviour can be overridden by either
12971 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12972 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12973 X509_supported_extension() has also been added which returns 1 if a
12974 particular extension is supported.
12978 * Modify the behaviour of EVP cipher functions in similar way to digests
12979 to retain compatibility with existing code.
12983 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12984 compatibility with existing code. In particular the 'ctx' parameter does
12985 not have to be to be initialized before the call to EVP_DigestInit() and
12986 it is tidied up after a call to EVP_DigestFinal(). New function
12987 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12988 EVP_MD_CTX_copy() changed to not require the destination to be
12989 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12990 requires the destination to be valid.
12992 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12993 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12997 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12998 so that complete 'Handshake' protocol structures are kept in memory
12999 instead of overwriting 'msg_type' and 'length' with 'body' data.
13003 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13005 *Massimo Santin via Richard Levitte*
13007 * Major restructuring to the underlying ENGINE code. This includes
13008 reduction of linker bloat, separation of pure "ENGINE" manipulation
13009 (initialisation, etc) from functionality dealing with implementations
13010 of specific crypto interfaces. This change also introduces integrated
13011 support for symmetric ciphers and digest implementations - so ENGINEs
13012 can now accelerate these by providing EVP_CIPHER and EVP_MD
13013 implementations of their own. This is detailed in
13014 [crypto/engine/README.md](crypto/engine/README.md)
13015 as it couldn't be adequately described here. However, there are a few
13016 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13017 were changed in the original introduction of ENGINE code have now
13018 reverted back - the hooking from this code to ENGINE is now a good
13019 deal more passive and at run-time, operations deal directly with
13020 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13021 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13022 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13023 they were not being used by the framework as there is no concept of a
13024 BIGNUM_METHOD and they could not be generalised to the new
13025 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13026 ENGINE_cpy() has been removed as it cannot be consistently defined in
13031 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13035 * Change mkdef.pl to sort symbols that get the same entry number,
13036 and make sure the automatically generated functions `ERR_load_*`
13037 become part of libeay.num as well.
13041 * New function SSL_renegotiate_pending(). This returns true once
13042 renegotiation has been requested (either SSL_renegotiate() call
13043 or HelloRequest/ClientHello received from the peer) and becomes
13044 false once a handshake has been completed.
13045 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13046 sends a HelloRequest, but does not ensure that a handshake takes
13047 place. SSL_renegotiate_pending() is useful for checking if the
13048 client has followed the request.)
13052 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13053 By default, clients may request session resumption even during
13054 renegotiation (if session ID contexts permit); with this option,
13055 session resumption is possible only in the first handshake.
13057 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13058 more bits available for options that should not be part of
13059 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13063 * Add some demos for certificate and certificate request creation.
13067 * Make maximum certificate chain size accepted from the peer application
13068 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13069 "Douglas E. Engert" <deengert@anl.gov>.
13073 * Add support for shared libraries for Unixware-7
13074 (Boyd Lynn Gerber <gerberb@zenez.com>).
13078 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13079 be done prior to destruction. Use this to unload error strings from
13080 ENGINEs that load their own error strings. NB: This adds two new API
13081 functions to "get" and "set" this destroy handler in an ENGINE.
13085 * Alter all existing ENGINE implementations (except "openssl" and
13086 "openbsd") to dynamically instantiate their own error strings. This
13087 makes them more flexible to be built both as statically-linked ENGINEs
13088 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13089 Also, add stub code to each that makes building them as self-contained
13090 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13094 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13095 implementations into applications that are completely implemented in
13096 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13097 commands that can be used to configure what shared-library to load and
13098 to control aspects of the way it is handled. Also, made an update to
13099 the [README-Engine.md](README-Engine.md) file
13100 that brings its information up-to-date and
13101 provides some information and instructions on the "dynamic" ENGINE
13102 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13106 * Make it possible to unload ranges of ERR strings with a new
13107 "ERR_unload_strings" function.
13111 * Add a copy() function to EVP_MD.
13115 * Make EVP_MD routines take a context pointer instead of just the
13116 md_data void pointer.
13120 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13121 that the digest can only process a single chunk of data
13122 (typically because it is provided by a piece of
13123 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13124 is only going to provide a single chunk of data, and hence the
13125 framework needn't accumulate the data for oneshot drivers.
13129 * As with "ERR", make it possible to replace the underlying "ex_data"
13130 functions. This change also alters the storage and management of global
13131 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13132 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13133 index counters. The API functions that use this state have been changed
13134 to take a "class_index" rather than pointers to the class's local STACK
13135 and counter, and there is now an API function to dynamically create new
13136 classes. This centralisation allows us to (a) plug a lot of the
13137 thread-safety problems that existed, and (b) makes it possible to clean
13138 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13139 such data would previously have always leaked in application code and
13140 workarounds were in place to make the memory debugging turn a blind eye
13141 to it. Application code that doesn't use this new function will still
13142 leak as before, but their memory debugging output will announce it now
13143 rather than letting it slide.
13145 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13146 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13147 has a return value to indicate success or failure.
13151 * Make it possible to replace the underlying "ERR" functions such that the
13152 global state (2 LHASH tables and 2 locks) is only used by the "default"
13153 implementation. This change also adds two functions to "get" and "set"
13154 the implementation prior to it being automatically set the first time
13155 any other ERR function takes place. Ie. an application can call "get",
13156 pass the return value to a module it has just loaded, and that module
13157 can call its own "set" function using that value. This means the
13158 module's "ERR" operations will use (and modify) the error state in the
13159 application and not in its own statically linked copy of OpenSSL code.
13163 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13164 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13165 the operation, and provides a more encapsulated way for external code
13166 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13167 to use these functions rather than manually incrementing the counts.
13169 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13173 * Add EVP test program.
13177 * Add symmetric cipher support to ENGINE. Expect the API to change!
13181 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13182 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13183 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13184 These allow a CRL to be built without having to access X509_CRL fields
13185 directly. Modify 'ca' application to use new functions.
13189 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13190 bug workarounds. Rollback attack detection is a security feature.
13191 The problem will only arise on OpenSSL servers when TLSv1 is not
13192 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13193 Software authors not wanting to support TLSv1 will have special reasons
13194 for their choice and can explicitly enable this option.
13196 *Bodo Moeller, Lutz Jaenicke*
13198 * Rationalise EVP so it can be extended: don't include a union of
13199 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13200 (similar to those existing for EVP_CIPHER_CTX).
13205 EVP_MD_CTX_init(&md); /* new function call */
13206 EVP_DigestInit(&md, EVP_sha1());
13207 EVP_DigestUpdate(&md, in, len);
13208 EVP_DigestFinal(&md, out, NULL);
13209 EVP_MD_CTX_cleanup(&md); /* new function call */
13213 * Make DES key schedule conform to the usual scheme, as well as
13214 correcting its structure. This means that calls to DES functions
13215 now have to pass a pointer to a des_key_schedule instead of a
13216 plain des_key_schedule (which was actually always a pointer
13219 des_key_schedule ks;
13221 des_set_key_checked(..., &ks);
13222 des_ncbc_encrypt(..., &ks, ...);
13224 (Note that a later change renames 'des_...' into 'DES_...'.)
13228 * Initial reduction of linker bloat: the use of some functions, such as
13229 PEM causes large amounts of unused functions to be linked in due to
13230 poor organisation. For example pem_all.c contains every PEM function
13231 which has a knock on effect of linking in large amounts of (unused)
13232 ASN1 code. Grouping together similar functions and splitting unrelated
13233 functions prevents this.
13237 * Cleanup of EVP macros.
13241 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13242 correct `_ecb suffix`.
13246 * Add initial OCSP responder support to ocsp application. The
13247 revocation information is handled using the text based index
13248 use by the ca application. The responder can either handle
13249 requests generated internally, supplied in files (for example
13250 via a CGI script) or using an internal minimal server.
13254 * Add configuration choices to get zlib compression for TLS.
13258 * Changes to Kerberos SSL for RFC 2712 compliance:
13259 1. Implemented real KerberosWrapper, instead of just using
13260 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13261 2. Implemented optional authenticator field of KerberosWrapper.
13263 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13264 and authenticator structs; see crypto/krb5/.
13266 Generalized Kerberos calls to support multiple Kerberos libraries.
13267 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13268 via Richard Levitte*
13270 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13271 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13272 values for each of the key sizes rather than having just
13273 parameters (and 'speed' generating keys each time).
13277 * Speed up EVP routines.
13280 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13281 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13282 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13283 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13285 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13286 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13287 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13290 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13292 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13296 * Added the OS2-EMX target.
13298 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13300 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13301 New functions to support `NCONF` routines in extension code.
13302 New function `CONF_set_nconf()`
13303 to allow functions which take an `NCONF` to also handle the old `LHASH`
13304 structure: this means that the old `CONF` compatible routines can be
13305 retained (in particular w.rt. extensions) without having to duplicate the
13306 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13310 * Enhance the general user interface with mechanisms for inner control
13311 and with possibilities to have yes/no kind of prompts.
13315 * Change all calls to low-level digest routines in the library and
13316 applications to use EVP. Add missing calls to HMAC_cleanup() and
13317 don't assume HMAC_CTX can be copied using memcpy().
13319 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13321 * Add the possibility to control engines through control names but with
13322 arbitrary arguments instead of just a string.
13323 Change the key loaders to take a UI_METHOD instead of a callback
13324 function pointer. NOTE: this breaks binary compatibility with earlier
13325 versions of OpenSSL [engine].
13326 Adapt the nCipher code for these new conditions and add a card insertion
13331 * Enhance the general user interface with mechanisms to better support
13332 dialog box interfaces, application-defined prompts, the possibility
13333 to use defaults (for example default passwords from somewhere else)
13334 and interrupts/cancellations.
13338 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13339 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13343 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13344 tidy up some unnecessarily weird code in 'sk_new()').
13346 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13348 * Change the key loading routines for ENGINEs to use the same kind
13349 callback (pem_password_cb) as all other routines that need this
13354 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13355 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13356 than this minimum value is recommended.
13360 * New random seeder for OpenVMS, using the system process statistics
13361 that are easily reachable.
13365 * Windows apparently can't transparently handle global
13366 variables defined in DLLs. Initialisations such as:
13368 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13370 won't compile. This is used by the any applications that need to
13371 declare their own ASN1 modules. This was fixed by adding the option
13372 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13373 needed for static libraries under Win32.
13377 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13378 setting of purpose and trust fields. New X509_STORE trust and
13379 purpose functions and tidy up setting in other SSL functions.
13383 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13384 structure. These are inherited by X509_STORE_CTX when it is
13385 initialised. This allows various defaults to be set in the
13386 X509_STORE structure (such as flags for CRL checking and custom
13387 purpose or trust settings) for functions which only use X509_STORE_CTX
13388 internally such as S/MIME.
13390 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13391 trust settings if they are not set in X509_STORE. This allows X509_STORE
13392 purposes and trust (in S/MIME for example) to override any set by default.
13394 Add command line options for CRL checking to smime, s_client and s_server
13399 * Initial CRL based revocation checking. If the CRL checking flag(s)
13400 are set then the CRL is looked up in the X509_STORE structure and
13401 its validity and signature checked, then if the certificate is found
13402 in the CRL the verify fails with a revoked error.
13404 Various new CRL related callbacks added to X509_STORE_CTX structure.
13406 Command line options added to 'verify' application to support this.
13408 This needs some additional work, such as being able to handle multiple
13409 CRLs with different times, extension based lookup (rather than just
13410 by subject name) and ultimately more complete V2 CRL extension
13415 * Add a general user interface API (crypto/ui/). This is designed
13416 to replace things like des_read_password and friends (backward
13417 compatibility functions using this new API are provided).
13418 The purpose is to remove prompting functions from the DES code
13419 section as well as provide for prompting through dialog boxes in
13420 a window system and the like.
13424 * Add "ex_data" support to ENGINE so implementations can add state at a
13425 per-structure level rather than having to store it globally.
13429 * Make it possible for ENGINE structures to be copied when retrieved by
13430 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13431 This causes the "original" ENGINE structure to act like a template,
13432 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13433 operational state can be localised to each ENGINE structure, despite the
13434 fact they all share the same "methods". New ENGINE structures returned in
13435 this case have no functional references and the return value is the single
13436 structural reference. This matches the single structural reference returned
13437 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13442 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13443 needs to match any other type at all we need to manually clear the
13448 * Changes to the "openssl engine" utility to include;
13449 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13450 about an ENGINE's available control commands.
13451 - executing control commands from command line arguments using the
13452 '-pre' and '-post' switches. '-post' is only used if '-t' is
13453 specified and the ENGINE is successfully initialised. The syntax for
13454 the individual commands are colon-separated, for example;
13455 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13459 * New dynamic control command support for ENGINEs. ENGINEs can now
13460 declare their own commands (numbers), names (strings), descriptions,
13461 and input types for run-time discovery by calling applications. A
13462 subset of these commands are implicitly classed as "executable"
13463 depending on their input type, and only these can be invoked through
13464 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13465 can be based on user input, config files, etc). The distinction is
13466 that "executable" commands cannot return anything other than a boolean
13467 result and can only support numeric or string input, whereas some
13468 discoverable commands may only be for direct use through
13469 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13470 pointers, or other custom uses. The "executable" commands are to
13471 support parameterisations of ENGINE behaviour that can be
13472 unambiguously defined by ENGINEs and used consistently across any
13473 OpenSSL-based application. Commands have been added to all the
13474 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13475 control over shared-library paths without source code alterations.
13479 * Changed all ENGINE implementations to dynamically allocate their
13480 ENGINEs rather than declaring them statically. Apart from this being
13481 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13482 this also allows the implementations to compile without using the
13483 internal engine_int.h header.
13487 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13488 'const' value. Any code that should be able to modify a RAND_METHOD
13489 should already have non-const pointers to it (ie. they should only
13490 modify their own ones).
13494 * Made a variety of little tweaks to the ENGINE code.
13495 - "atalla" and "ubsec" string definitions were moved from header files
13496 to C code. "nuron" string definitions were placed in variables
13497 rather than hard-coded - allowing parameterisation of these values
13498 later on via ctrl() commands.
13499 - Removed unused "#if 0"'d code.
13500 - Fixed engine list iteration code so it uses ENGINE_free() to release
13501 structural references.
13502 - Constified the RAND_METHOD element of ENGINE structures.
13503 - Constified various get/set functions as appropriate and added
13504 missing functions (including a catch-all ENGINE_cpy that duplicates
13505 all ENGINE values onto a new ENGINE except reference counts/state).
13506 - Removed NULL parameter checks in get/set functions. Setting a method
13507 or function to NULL is a way of cancelling out a previously set
13508 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13509 and doesn't justify the extra error symbols and code.
13510 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13511 flags from engine_int.h to engine.h.
13512 - Changed prototypes for ENGINE handler functions (init(), finish(),
13513 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13517 * Implement binary inversion algorithm for BN_mod_inverse in addition
13518 to the algorithm using long division. The binary algorithm can be
13519 used only if the modulus is odd. On 32-bit systems, it is faster
13520 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13521 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13522 up to 450 bits. In 64-bit environments, the binary algorithm
13523 appears to be advantageous for much longer moduli; here we use it
13524 for moduli up to 2048 bits.
13528 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13529 could not support the combine flag in choice fields.
13533 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13534 extensions from a certificate request to the certificate.
13538 * Allow multiple 'certopt' and 'nameopt' options to be separated
13539 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13540 file: this allows the display of the certificate about to be
13541 signed to be customised, to allow certain fields to be included
13542 or excluded and extension details. The old system didn't display
13543 multicharacter strings properly, omitted fields not in the policy
13544 and couldn't display additional details such as extensions.
13548 * Function EC_POINTs_mul for multiple scalar multiplication
13549 of an arbitrary number of elliptic curve points
13550 \sum scalars[i]*points[i],
13551 optionally including the generator defined for the EC_GROUP:
13552 scalar*generator + \sum scalars[i]*points[i].
13554 EC_POINT_mul is a simple wrapper function for the typical case
13555 that the point list has just one item (besides the optional
13560 * First EC_METHODs for curves over GF(p):
13562 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13563 operations and provides various method functions that can also
13564 operate with faster implementations of modular arithmetic.
13566 EC_GFp_mont_method() reuses most functions that are part of
13567 EC_GFp_simple_method, but uses Montgomery arithmetic.
13569 *Bodo Moeller; point addition and point doubling
13570 implementation directly derived from source code provided by
13571 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13573 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13574 crypto/ec/ec_lib.c):
13576 Curves are EC_GROUP objects (with an optional group generator)
13577 based on EC_METHODs that are built into the library.
13579 Points are EC_POINT objects based on EC_GROUP objects.
13581 Most of the framework would be able to handle curves over arbitrary
13582 finite fields, but as there are no obvious types for fields other
13583 than GF(p), some functions are limited to that for now.
13587 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13588 that the file contains a complete HTTP response.
13592 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13593 change the def and num file printf format specifier from "%-40sXXX"
13594 to "%-39s XXX". The latter will always guarantee a space after the
13595 field while the former will cause them to run together if the field
13596 is 40 of more characters long.
13600 * Constify the cipher and digest 'method' functions and structures
13601 and modify related functions to take constant EVP_MD and EVP_CIPHER
13606 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13607 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13611 * Modify `EVP_Digest*()` routines so they now return values. Although the
13612 internal software routines can never fail additional hardware versions
13617 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13619 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13620 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13623 ERR_R_NESTED_ASN1_ERROR
13625 ERR_R_MISSING_ASN1_EOS
13626 were 4 .. 9, conflicting with
13627 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13629 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13630 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13632 Add new error code 'ERR_R_INTERNAL_ERROR'.
13636 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13641 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13642 sets the subject name for a new request or supersedes the
13643 subject name in a given request. Formats that can be parsed are
13644 'CN=Some Name, OU=myOU, C=IT'
13646 'CN=Some Name/OU=myOU/C=IT'.
13648 Add options '-batch' and '-verbose' to 'openssl req'.
13650 *Massimiliano Pala <madwolf@hackmasters.net>*
13652 * Introduce the possibility to access global variables through
13653 functions on platform were that's the best way to handle exporting
13654 global variables in shared libraries. To enable this functionality,
13655 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13656 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13657 is normally done by Configure or something similar).
13659 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13660 in the source file (foo.c) like this:
13662 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13663 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13665 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13666 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13668 OPENSSL_DECLARE_GLOBAL(int,foo);
13669 #define foo OPENSSL_GLOBAL_REF(foo)
13670 OPENSSL_DECLARE_GLOBAL(double,bar);
13671 #define bar OPENSSL_GLOBAL_REF(bar)
13673 The #defines are very important, and therefore so is including the
13674 header file everywhere where the defined globals are used.
13676 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13677 of ASN.1 items, but that structure is a bit different.
13679 The largest change is in util/mkdef.pl which has been enhanced with
13680 better and easier to understand logic to choose which symbols should
13681 go into the Windows .def files as well as a number of fixes and code
13682 cleanup (among others, algorithm keywords are now sorted
13683 lexicographically to avoid constant rewrites).
13687 * In BN_div() keep a copy of the sign of 'num' before writing the
13688 result to 'rm' because if rm==num the value will be overwritten
13689 and produce the wrong result if 'num' is negative: this caused
13690 problems with BN_mod() and BN_nnmod().
13694 * Function OCSP_request_verify(). This checks the signature on an
13695 OCSP request and verifies the signer certificate. The signer
13696 certificate is just checked for a generic purpose and OCSP request
13701 * Add OCSP_check_validity() function to check the validity of OCSP
13702 responses. OCSP responses are prepared in real time and may only
13703 be a few seconds old. Simply checking that the current time lies
13704 between thisUpdate and nextUpdate max reject otherwise valid responses
13705 caused by either OCSP responder or client clock inaccuracy. Instead
13706 we allow thisUpdate and nextUpdate to fall within a certain period of
13707 the current time. The age of the response can also optionally be
13708 checked. Two new options -validity_period and -status_age added to
13713 * If signature or public key algorithm is unrecognized print out its
13714 OID rather that just UNKNOWN.
13718 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13719 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13720 ID to be generated from the issuer certificate alone which can then be
13721 passed to OCSP_id_issuer_cmp().
13725 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13726 ASN1 modules to export functions returning ASN1_ITEM pointers
13727 instead of the ASN1_ITEM structures themselves. This adds several
13728 new macros which allow the underlying ASN1 function/structure to
13729 be accessed transparently. As a result code should not use ASN1_ITEM
13730 references directly (such as &X509_it) but instead use the relevant
13731 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13732 use of the new ASN1 code on platforms where exporting structures
13733 is problematical (for example in shared libraries) but exporting
13734 functions returning pointers to structures is not.
13738 * Add support for overriding the generation of SSL/TLS session IDs.
13739 These callbacks can be registered either in an SSL_CTX or per SSL.
13740 The purpose of this is to allow applications to control, if they wish,
13741 the arbitrary values chosen for use as session IDs, particularly as it
13742 can be useful for session caching in multiple-server environments. A
13743 command-line switch for testing this (and any client code that wishes
13744 to use such a feature) has been added to "s_server".
13746 *Geoff Thorpe, Lutz Jaenicke*
13748 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13749 of the form `#if defined(...) || defined(...) || ...` and
13750 `#if !defined(...) && !defined(...) && ...`. This also avoids
13751 the growing number of special cases it was previously handling.
13755 * Make all configuration macros available for application by making
13756 sure they are available in opensslconf.h, by giving them names starting
13757 with `OPENSSL_` to avoid conflicts with other packages and by making
13758 sure e_os2.h will cover all platform-specific cases together with
13760 Additionally, it is now possible to define configuration/platform-
13761 specific names (called "system identities"). In the C code, these
13762 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13763 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13764 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13769 * New option -set_serial to 'req' and 'x509' this allows the serial
13770 number to use to be specified on the command line. Previously self
13771 signed certificates were hard coded with serial number 0 and the
13772 CA options of 'x509' had to use a serial number in a file which was
13777 * New options to 'ca' utility to support V2 CRL entry extensions.
13778 Currently CRL reason, invalidity date and hold instruction are
13779 supported. Add new CRL extensions to V3 code and some new objects.
13783 * New function EVP_CIPHER_CTX_set_padding() this is used to
13784 disable standard block padding (aka PKCS#5 padding) in the EVP
13785 API, which was previously mandatory. This means that the data is
13786 not padded in any way and so the total length much be a multiple
13787 of the block size, otherwise an error occurs.
13791 * Initial (incomplete) OCSP SSL support.
13795 * New function OCSP_parse_url(). This splits up a URL into its host,
13796 port and path components: primarily to parse OCSP URLs. New -url
13797 option to ocsp utility.
13801 * New nonce behavior. The return value of OCSP_check_nonce() now
13802 reflects the various checks performed. Applications can decide
13803 whether to tolerate certain situations such as an absent nonce
13804 in a response when one was present in a request: the ocsp application
13805 just prints out a warning. New function OCSP_add1_basic_nonce()
13806 this is to allow responders to include a nonce in a response even if
13807 the request is nonce-less.
13811 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13812 skipped when using openssl x509 multiple times on a single input file,
13813 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13817 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13818 set string type: to handle setting ASN1_TIME structures. Fix ca
13819 utility to correctly initialize revocation date of CRLs.
13823 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13824 the clients preferred ciphersuites and rather use its own preferences.
13825 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13826 Internet Explorer by ensuring unchanged hash method during stepup.
13827 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13831 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13832 to aes and add a new 'exist' option to print out symbols that don't
13837 * Additional options to ocsp utility to allow flags to be set and
13838 additional certificates supplied.
13842 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13843 OCSP client a number of certificate to only verify the response
13848 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13849 handle the new API. Currently only ECB, CBC modes supported. Add new
13852 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13853 Encryption Standard (AES) Ciphersuites for Transport Layer
13854 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13855 not enabled by default and were not part of the "ALL" ciphersuite
13856 alias because they were not yet official; they could be
13857 explicitly requested by specifying the "AESdraft" ciphersuite
13858 group alias. In the final release of OpenSSL 0.9.7, the group
13859 alias is called "AES" and is part of "ALL".)
13861 *Ben Laurie, Steve Henson, Bodo Moeller*
13863 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13864 request to response.
13868 * Functions for OCSP responders. OCSP_request_onereq_count(),
13869 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13870 extract information from a certificate request. OCSP_response_create()
13871 creates a response and optionally adds a basic response structure.
13872 OCSP_basic_add1_status() adds a complete single response to a basic
13873 response and returns the OCSP_SINGLERESP structure just added (to allow
13874 extensions to be included for example). OCSP_basic_add1_cert() adds a
13875 certificate to a basic response and OCSP_basic_sign() signs a basic
13876 response with various flags. New helper functions ASN1_TIME_check()
13877 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13878 (converts ASN1_TIME to GeneralizedTime).
13882 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13883 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13884 structure from a certificate. X509_pubkey_digest() digests the public_key
13885 contents: this is used in various key identifiers.
13889 * Make sk_sort() tolerate a NULL argument.
13891 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13893 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13894 passed by the function are trusted implicitly. If any of them signed the
13895 response then it is assumed to be valid and is not verified.
13899 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13900 to data. This was previously part of the PKCS7 ASN1 code. This
13901 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13902 *Steve Henson, reported by Kenneth R. Robinette
13903 <support@securenetterm.com>*
13905 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13906 routines: without these tracing memory leaks is very painful.
13907 Fix leaks in PKCS12 and PKCS7 routines.
13911 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13912 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13913 effectively meant GeneralizedTime would never be used. Now it
13914 is initialised to -1 but X509_time_adj() now has to check the value
13915 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13916 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13917 *Steve Henson, reported by Kenneth R. Robinette
13918 <support@securenetterm.com>*
13920 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13921 result in a zero length in the ASN1_INTEGER structure which was
13922 not consistent with the structure when d2i_ASN1_INTEGER() was used
13923 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13924 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13925 where it did not print out a minus for negative ASN1_INTEGER.
13929 * Add summary printout to ocsp utility. The various functions which
13930 convert status values to strings have been renamed to:
13931 OCSP_response_status_str(), OCSP_cert_status_str() and
13932 OCSP_crl_reason_str() and are no longer static. New options
13933 to verify nonce values and to disable verification. OCSP response
13934 printout format cleaned up.
13938 * Add additional OCSP certificate checks. These are those specified
13939 in RFC2560. This consists of two separate checks: the CA of the
13940 certificate being checked must either be the OCSP signer certificate
13941 or the issuer of the OCSP signer certificate. In the latter case the
13942 OCSP signer certificate must contain the OCSP signing extended key
13943 usage. This check is performed by attempting to match the OCSP
13944 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13945 in the OCSP_CERTID structures of the response.
13949 * Initial OCSP certificate verification added to OCSP_basic_verify()
13950 and related routines. This uses the standard OpenSSL certificate
13951 verify routines to perform initial checks (just CA validity) and
13952 to obtain the certificate chain. Then additional checks will be
13953 performed on the chain. Currently the root CA is checked to see
13954 if it is explicitly trusted for OCSP signing. This is used to set
13955 a root CA as a global signing root: that is any certificate that
13956 chains to that CA is an acceptable OCSP signing certificate.
13960 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13961 extensions from a separate configuration file.
13962 As when reading extensions from the main configuration file,
13963 the '-extensions ...' option may be used for specifying the
13966 *Massimiliano Pala <madwolf@comune.modena.it>*
13968 * New OCSP utility. Allows OCSP requests to be generated or
13969 read. The request can be sent to a responder and the output
13970 parsed, outputted or printed in text form. Not complete yet:
13971 still needs to check the OCSP response validity.
13975 * New subcommands for 'openssl ca':
13976 `openssl ca -status <serial>` prints the status of the cert with
13977 the given serial number (according to the index file).
13978 `openssl ca -updatedb` updates the expiry status of certificates
13981 *Massimiliano Pala <madwolf@comune.modena.it>*
13983 * New '-newreq-nodes' command option to CA.pl. This is like
13984 '-newreq', but calls 'openssl req' with the '-nodes' option
13985 so that the resulting key is not encrypted.
13987 *Damien Miller <djm@mindrot.org>*
13989 * New configuration for the GNU Hurd.
13991 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13993 * Initial code to implement OCSP basic response verify. This
13994 is currently incomplete. Currently just finds the signer's
13995 certificate and verifies the signature on the response.
13999 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14000 value of OPENSSLDIR. This is available via the new '-d' option
14001 to 'openssl version', and is also included in 'openssl version -a'.
14005 * Allowing defining memory allocation callbacks that will be given
14006 file name and line number information in additional arguments
14007 (a `const char*` and an int). The basic functionality remains, as
14008 well as the original possibility to just replace malloc(),
14009 realloc() and free() by functions that do not know about these
14010 additional arguments. To register and find out the current
14011 settings for extended allocation functions, the following
14012 functions are provided:
14014 CRYPTO_set_mem_ex_functions
14015 CRYPTO_set_locked_mem_ex_functions
14016 CRYPTO_get_mem_ex_functions
14017 CRYPTO_get_locked_mem_ex_functions
14019 These work the same way as CRYPTO_set_mem_functions and friends.
14020 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14021 extended allocation function is enabled.
14022 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14023 a conventional allocation function is enabled.
14025 *Richard Levitte, Bodo Moeller*
14027 * Finish off removing the remaining LHASH function pointer casts.
14028 There should no longer be any prototype-casting required when using
14029 the LHASH abstraction, and any casts that remain are "bugs". See
14030 the callback types and macros at the head of lhash.h for details
14031 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14035 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14036 If /dev/[u]random devices are not available or do not return enough
14037 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14039 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14040 /etc/entropy will be queried once each in this sequence, querying stops
14041 when enough entropy was collected without querying more sockets.
14045 * Change the Unix RAND_poll() variant to be able to poll several
14046 random devices, as specified by DEVRANDOM, until a sufficient amount
14047 of data has been collected. We spend at most 10 ms on each file
14048 (select timeout) and read in non-blocking mode. DEVRANDOM now
14049 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14050 (previously it was just the string "/dev/urandom"), so on typical
14051 platforms the 10 ms delay will never occur.
14052 Also separate out the Unix variant to its own file, rand_unix.c.
14053 For VMS, there's a currently-empty rand_vms.c.
14057 * Move OCSP client related routines to ocsp_cl.c. These
14058 provide utility functions which an application needing
14059 to issue a request to an OCSP responder and analyse the
14060 response will typically need: as opposed to those which an
14061 OCSP responder itself would need which will be added later.
14063 OCSP_request_sign() signs an OCSP request with an API similar
14064 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14065 response. OCSP_response_get1_basic() extracts basic response
14066 from response. OCSP_resp_find_status(): finds and extracts status
14067 information from an OCSP_CERTID structure (which will be created
14068 when the request structure is built). These are built from lower
14069 level functions which work on OCSP_SINGLERESP structures but
14070 won't normally be used unless the application wishes to examine
14071 extensions in the OCSP response for example.
14073 Replace nonce routines with a pair of functions.
14074 OCSP_request_add1_nonce() adds a nonce value and optionally
14075 generates a random value. OCSP_check_nonce() checks the
14076 validity of the nonce in an OCSP response.
14080 * Change function OCSP_request_add() to OCSP_request_add0_id().
14081 This doesn't copy the supplied OCSP_CERTID and avoids the
14082 need to free up the newly created id. Change return type
14083 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14084 This can then be used to add extensions to the request.
14085 Deleted OCSP_request_new(), since most of its functionality
14086 is now in OCSP_REQUEST_new() (and the case insensitive name
14087 clash) apart from the ability to set the request name which
14088 will be added elsewhere.
14092 * Update OCSP API. Remove obsolete extensions argument from
14093 various functions. Extensions are now handled using the new
14094 OCSP extension code. New simple OCSP HTTP function which
14095 can be used to send requests and parse the response.
14099 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14100 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14101 uses the special reorder version of SET OF to sort the attributes
14102 and reorder them to match the encoded order. This resolves a long
14103 standing problem: a verify on a PKCS7 structure just after signing
14104 it used to fail because the attribute order did not match the
14105 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14106 it uses the received order. This is necessary to tolerate some broken
14107 software that does not order SET OF. This is handled by encoding
14108 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14109 to produce the required SET OF.
14113 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14114 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14115 files to get correct declarations of the ASN.1 item variables.
14119 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14120 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14121 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14122 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14123 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14124 ASN1_ITEM and no wrapper functions.
14128 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14129 replace the old function pointer based I/O routines. Change most of
14130 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14134 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14135 lines, recognize more "algorithms" that can be deselected, and make
14136 it complain about algorithm deselection that isn't recognised.
14140 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14141 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14142 to use new functions. Add NO_ASN1_OLD which can be set to remove
14143 some old style ASN1 functions: this can be used to determine if old
14144 code will still work when these eventually go away.
14148 * New extension functions for OCSP structures, these follow the
14149 same conventions as certificates and CRLs.
14153 * New function X509V3_add1_i2d(). This automatically encodes and
14154 adds an extension. Its behaviour can be customised with various
14155 flags to append, replace or delete. Various wrappers added for
14156 certificates and CRLs.
14160 * Fix to avoid calling the underlying ASN1 print routine when
14161 an extension cannot be parsed. Correct a typo in the
14162 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14166 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14167 entries for variables.
14171 * Add functionality to `apps/openssl.c` for detecting locking
14172 problems: As the program is single-threaded, all we have
14173 to do is register a locking callback using an array for
14174 storing which locks are currently held by the program.
14178 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14179 SSL_get_ex_data_X509_STORE_idx(), which is used in
14180 ssl_verify_cert_chain() and thus can be called at any time
14181 during TLS/SSL handshakes so that thread-safety is essential.
14182 Unfortunately, the ex_data design is not at all suited
14183 for multi-threaded use, so it probably should be abolished.
14187 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14189 *Broadcom, tweaked and integrated by Geoff Thorpe*
14191 * Move common extension printing code to new function
14192 X509V3_print_extensions(). Reorganise OCSP print routines and
14193 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14197 * New function X509_signature_print() to remove duplication in some
14202 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14203 set (this was treated exactly the same as SET OF previously). This
14204 is used to reorder the STACK representing the structure to match the
14205 encoding. This will be used to get round a problem where a PKCS7
14206 structure which was signed could not be verified because the STACK
14207 order did not reflect the encoded order.
14211 * Reimplement the OCSP ASN1 module using the new code.
14215 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14216 for its ASN1 operations. The old style function pointers still exist
14217 for now but they will eventually go away.
14221 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14222 completely replaces the old ASN1 functionality with a table driven
14223 encoder and decoder which interprets an ASN1_ITEM structure describing
14224 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14225 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14226 has also been converted to the new form.
14230 * Change BN_mod_exp_recp so that negative moduli are tolerated
14231 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14232 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14233 for negative moduli.
14237 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14238 of not touching the result's sign bit.
14242 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14247 * Changed the LHASH code to use prototypes for callbacks, and created
14248 macros to declare and implement thin (optionally static) functions
14249 that provide type-safety and avoid function pointer casting for the
14250 type-specific callbacks.
14254 * Added Kerberos Cipher Suites to be used with TLS, as written in
14256 *Veers Staats <staatsvr@asc.hpc.mil>,
14257 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14259 * Reformat the FAQ so the different questions and answers can be divided
14260 in sections depending on the subject.
14264 * Have the zlib compression code load ZLIB.DLL dynamically under
14269 * New function BN_mod_sqrt for computing square roots modulo a prime
14270 (using the probabilistic Tonelli-Shanks algorithm unless
14271 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14272 be handled deterministically).
14274 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14276 * Make BN_mod_inverse faster by explicitly handling small quotients
14277 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14278 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14282 * New function BN_kronecker.
14286 * Fix BN_gcd so that it works on negative inputs; the result is
14287 positive unless both parameters are zero.
14288 Previously something reasonably close to an infinite loop was
14289 possible because numbers could be growing instead of shrinking
14290 in the implementation of Euclid's algorithm.
14294 * Fix BN_is_word() and BN_is_one() macros to take into account the
14295 sign of the number in question.
14297 Fix BN_is_word(a,w) to work correctly for w == 0.
14299 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14300 because its test if the absolute value of 'a' equals 'w'.
14301 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14302 it exists mostly for use in the implementations of BN_is_zero(),
14303 BN_is_one(), and BN_is_word().
14307 * New function BN_swap.
14311 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14312 the exponentiation functions are more likely to produce reasonable
14313 results on negative inputs.
14317 * Change BN_mod_mul so that the result is always non-negative.
14318 Previously, it could be negative if one of the factors was negative;
14319 I don't think anyone really wanted that behaviour.
14323 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14324 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14325 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14326 and add new functions:
14335 BN_mod_lshift1_quick
14337 BN_mod_lshift_quick
14339 These functions always generate non-negative results.
14341 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14342 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14344 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14345 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14346 be reduced modulo `m`.
14348 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14351 The following entry accidentally appeared in the CHANGES file
14352 distributed with OpenSSL 0.9.7. The modifications described in
14353 it do *not* apply to OpenSSL 0.9.7.
14355 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14356 was actually never needed) and in BN_mul(). The removal in BN_mul()
14357 required a small change in bn_mul_part_recursive() and the addition
14358 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14359 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14360 bn_sub_words() and bn_add_words() except they take arrays with
14366 * In 'openssl passwd', verify passwords read from the terminal
14367 unless the '-salt' option is used (which usually means that
14368 verification would just waste user's time since the resulting
14369 hash is going to be compared with some given password hash)
14370 or the new '-noverify' option is used.
14372 This is an incompatible change, but it does not affect
14373 non-interactive use of 'openssl passwd' (passwords on the command
14374 line, '-stdin' option, '-in ...' option) and thus should not
14375 cause any problems.
14379 * Remove all references to RSAref, since there's no more need for it.
14383 * Make DSO load along a path given through an environment variable
14384 (SHLIB_PATH) with shl_load().
14388 * Constify the ENGINE code as a result of BIGNUM constification.
14389 Also constify the RSA code and most things related to it. In a
14390 few places, most notable in the depth of the ASN.1 code, ugly
14391 casts back to non-const were required (to be solved at a later
14396 * Make it so the openssl application has all engines loaded by default.
14400 * Constify the BIGNUM routines a little more.
14404 * Add the following functions:
14406 ENGINE_load_cswift()
14408 ENGINE_load_atalla()
14409 ENGINE_load_nuron()
14410 ENGINE_load_builtin_engines()
14412 That way, an application can itself choose if external engines that
14413 are built-in in OpenSSL shall ever be used or not. The benefit is
14414 that applications won't have to be linked with libdl or other dso
14415 libraries unless it's really needed.
14417 Changed 'openssl engine' to load all engines on demand.
14418 Changed the engine header files to avoid the duplication of some
14419 declarations (they differed!).
14423 * 'openssl engine' can now list capabilities.
14427 * Better error reporting in 'openssl engine'.
14431 * Never call load_dh_param(NULL) in s_server.
14435 * Add engine application. It can currently list engines by name and
14436 identity, and test if they are actually available.
14440 * Improve RPM specification file by forcing symbolic linking and making
14441 sure the installed documentation is also owned by root.root.
14443 *Damien Miller <djm@mindrot.org>*
14445 * Give the OpenSSL applications more possibilities to make use of
14446 keys (public as well as private) handled by engines.
14450 * Add OCSP code that comes from CertCo.
14454 * Add VMS support for the Rijndael code.
14458 * Added untested support for Nuron crypto accelerator.
14462 * Add support for external cryptographic devices. This code was
14463 previously distributed separately as the "engine" branch.
14465 *Geoff Thorpe, Richard Levitte*
14467 * Rework the filename-translation in the DSO code. It is now possible to
14468 have far greater control over how a "name" is turned into a filename
14469 depending on the operating environment and any oddities about the
14470 different shared library filenames on each system.
14474 * Support threads on FreeBSD-elf in Configure.
14478 * Fix for SHA1 assembly problem with MASM: it produces
14479 warnings about corrupt line number information when assembling
14480 with debugging information. This is caused by the overlapping
14483 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14486 NCONF_get_number() has no error checking at all. As a replacement,
14487 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14488 promoted strongly. The old NCONF_get_number is kept around for
14489 binary backward compatibility.
14490 Make it possible for methods to load from something other than a BIO,
14491 by providing a function pointer that is given a name instead of a BIO.
14492 For example, this could be used to load configuration data from an
14497 * Fix for non blocking accept BIOs. Added new I/O special reason
14498 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14499 with non blocking I/O was not possible because no retry code was
14500 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14505 * Added the beginnings of Rijndael support.
14509 * Fix for bug in DirectoryString mask setting. Add support for
14510 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14511 to allow certificate printing to more controllable, additional
14512 'certopt' option to 'x509' to allow new printing options to be
14517 * Clean old EAY MD5 hack from e_os.h.
14521 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14523 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14524 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14526 *Joe Orton, Steve Henson*
14528 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14530 * Fix additional bug revealed by the NISCC test suite:
14532 Stop bug triggering large recursion when presented with
14533 certain ASN.1 tags ([CVE-2003-0851])
14537 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14539 * Fix various bugs revealed by running the NISCC test suite:
14541 Stop out of bounds reads in the ASN1 code when presented with
14542 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14544 If verify callback ignores invalid public key errors don't try to check
14545 certificate signature with the NULL public key.
14549 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14550 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14555 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14556 extra data after the compression methods not only for TLS 1.0
14557 but also for SSL 3.0 (as required by the specification).
14559 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14561 * Change X509_certificate_type() to mark the key as exported/exportable
14562 when it's 512 *bits* long, not 512 bytes.
14566 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14568 * Countermeasure against the Klima-Pokorny-Rosa extension of
14569 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14570 a protocol version number mismatch like a decryption error
14571 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14575 * Turn on RSA blinding by default in the default implementation
14576 to avoid a timing attack. Applications that don't want it can call
14577 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14578 They would be ill-advised to do so in most cases.
14580 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14582 * Change RSA blinding code so that it works when the PRNG is not
14583 seeded (in this case, the secret RSA exponent is abused as
14584 an unpredictable seed -- if it is not unpredictable, there
14585 is no point in blinding anyway). Make RSA blinding thread-safe
14586 by remembering the creator's thread ID in rsa->blinding and
14587 having all other threads use local one-time blinding factors
14588 (this requires more computation than sharing rsa->blinding, but
14589 avoids excessive locking; and if an RSA object is not shared
14590 between threads, blinding will still be very fast).
14594 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14596 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14597 via timing by performing a MAC computation even if incorrect
14598 block cipher padding has been found. This is a countermeasure
14599 against active attacks where the attacker has to distinguish
14600 between bad padding and a MAC verification error. ([CVE-2003-0078])
14602 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14603 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14604 Martin Vuagnoux (EPFL, Ilion)*
14606 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14608 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14609 memory from its contents. This is done with a counter that will
14610 place alternating values in each byte. This can be used to solve
14611 two issues: 1) the removal of calls to memset() by highly optimizing
14612 compilers, and 2) cleansing with other values than 0, since those can
14613 be read through on certain media, for example a swap space on disk.
14617 * Bugfix: client side session caching did not work with external caching,
14618 because the session->cipher setting was not restored when reloading
14619 from the external cache. This problem was masked, when
14620 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14621 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14625 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14626 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14628 *Zeev Lieber <zeev-l@yahoo.com>*
14630 * Undo an undocumented change introduced in 0.9.6e which caused
14631 repeated calls to OpenSSL_add_all_ciphers() and
14632 OpenSSL_add_all_digests() to be ignored, even after calling
14637 * Change the default configuration reader to deal with last line not
14638 being properly terminated.
14642 * Change X509_NAME_cmp() so it applies the special rules on handling
14643 DN values that are of type PrintableString, as well as RDNs of type
14644 emailAddress where the value has the type ia5String.
14646 *stefank@valicert.com via Richard Levitte*
14648 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14649 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14650 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14651 the bitwise-OR of the two for use by the majority of applications
14652 wanting this behaviour, and update the docs. The documented
14653 behaviour and actual behaviour were inconsistent and had been
14654 changing anyway, so this is more a bug-fix than a behavioural
14657 *Geoff Thorpe, diagnosed by Nadav Har'El*
14659 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14660 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14664 * Fix initialization code race conditions in
14665 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14666 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14667 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14668 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14669 ssl2_get_cipher_by_char(),
14670 ssl3_get_cipher_by_char().
14672 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14674 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14675 the cached sessions are flushed, as the remove_cb() might use ex_data
14676 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14677 (see [openssl.org #212]).
14679 *Geoff Thorpe, Lutz Jaenicke*
14681 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14682 length, instead of the encoding length to d2i_ASN1_OBJECT.
14686 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14688 * [In 0.9.6g-engine release:]
14689 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14691 *Lynn Gazis <lgazis@rainbow.com>*
14693 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14695 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14696 and get fix the header length calculation.
14697 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14698 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14700 * Use proper error handling instead of 'assertions' in buffer
14701 overflow checks added in 0.9.6e. This prevents DoS (the
14702 assertions could call abort()).
14704 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14706 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14708 * Add various sanity checks to asn1_get_length() to reject
14709 the ASN1 length bytes if they exceed sizeof(long), will appear
14710 negative or the content length exceeds the length of the
14713 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14715 * Fix cipher selection routines: ciphers without encryption had no flags
14716 for the cipher strength set and where therefore not handled correctly
14717 by the selection routines (PR #130).
14721 * Fix EVP_dsa_sha macro.
14726 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14727 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14728 that was added in OpenSSL 0.9.6d.
14730 As the countermeasure turned out to be incompatible with some
14731 broken SSL implementations, the new option is part of SSL_OP_ALL.
14732 SSL_OP_ALL is usually employed when compatibility with weird SSL
14733 implementations is desired (e.g. '-bugs' option to 's_client' and
14734 's_server'), so the new option is automatically set in many
14739 * Changes in security patch:
14741 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14742 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14743 Air Force Materiel Command, USAF, under agreement number
14746 * Add various sanity checks to asn1_get_length() to reject
14747 the ASN1 length bytes if they exceed sizeof(long), will appear
14748 negative or the content length exceeds the length of the
14749 supplied buffer. ([CVE-2002-0659])
14751 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14753 * Assertions for various potential buffer overflows, not known to
14754 happen in practice.
14756 *Ben Laurie (CHATS)*
14758 * Various temporary buffers to hold ASCII versions of integers were
14759 too small for 64 bit platforms. ([CVE-2002-0655])
14760 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14762 * Remote buffer overflow in SSL3 protocol - an attacker could
14763 supply an oversized session ID to a client. ([CVE-2002-0656])
14765 *Ben Laurie (CHATS)*
14767 * Remote buffer overflow in SSL2 protocol - an attacker could
14768 supply an oversized client master key. ([CVE-2002-0656])
14770 *Ben Laurie (CHATS)*
14772 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14774 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14775 encoded as NULL) with id-dsa-with-sha1.
14777 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14779 * Check various `X509_...()` return values in `apps/req.c`.
14781 *Nils Larsch <nla@trustcenter.de>*
14783 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14784 an end-of-file condition would erroneously be flagged, when the CRLF
14785 was just at the end of a processed block. The bug was discovered when
14786 processing data through a buffering memory BIO handing the data to a
14787 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14788 <ptsekov@syntrex.com> and Nedelcho Stanev.
14792 * Implement a countermeasure against a vulnerability recently found
14793 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14794 before application data chunks to avoid the use of known IVs
14795 with data potentially chosen by the attacker.
14799 * Fix length checks in ssl3_get_client_hello().
14803 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14804 to prevent ssl3_read_internal() from incorrectly assuming that
14805 ssl3_read_bytes() found application data while handshake
14806 processing was enabled when in fact s->s3->in_read_app_data was
14807 merely automatically cleared during the initial handshake.
14809 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14811 * Fix object definitions for Private and Enterprise: they were not
14812 recognized in their shortname (=lowercase) representation. Extend
14813 obj_dat.pl to issue an error when using undefined keywords instead
14814 of silently ignoring the problem (Svenning Sorensen
14815 <sss@sss.dnsalias.net>).
14819 * Fix DH_generate_parameters() so that it works for 'non-standard'
14820 generators, i.e. generators other than 2 and 5. (Previously, the
14821 code did not properly initialise the 'add' and 'rem' values to
14822 BN_generate_prime().)
14824 In the new general case, we do not insist that 'generator' is
14825 actually a primitive root: This requirement is rather pointless;
14826 a generator of the order-q subgroup is just as good, if not
14831 * Map new X509 verification errors to alerts. Discovered and submitted by
14832 Tom Wu <tom@arcot.com>.
14836 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14837 returning non-zero before the data has been completely received
14838 when using non-blocking I/O.
14840 *Bodo Moeller; problem pointed out by John Hughes*
14842 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14844 *Ben Laurie, Lutz Jaenicke*
14846 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14847 Yoram Zahavi <YoramZ@gilian.com>).
14851 * Add information about CygWin 1.3 and on, and preserve proper
14852 configuration for the versions before that.
14854 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14856 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14857 check whether we deal with a copy of a session and do not delete from
14858 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14859 <izhar@checkpoint.com>.
14863 * Do not store session data into the internal session cache, if it
14864 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14865 flag is set). Proposed by Aslam <aslam@funk.com>.
14869 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14874 * [In 0.9.6d-engine release:]
14875 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14877 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14879 * Add the configuration target linux-s390x.
14881 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14883 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14884 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14885 variable as an indication that a ClientHello message has been
14886 received. As the flag value will be lost between multiple
14887 invocations of ssl3_accept when using non-blocking I/O, the
14888 function may not be aware that a handshake has actually taken
14889 place, thus preventing a new session from being added to the
14892 To avoid this problem, we now set s->new_session to 2 instead of
14893 using a local variable.
14895 *Lutz Jaenicke, Bodo Moeller*
14897 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14898 if the SSL_R_LENGTH_MISMATCH error is detected.
14900 *Geoff Thorpe, Bodo Moeller*
14902 * New 'shared_ldflag' column in Configure platform table.
14906 * Fix EVP_CIPHER_mode macro.
14908 *"Dan S. Camper" <dan@bti.net>*
14910 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14911 type, we must throw them away by setting rr->length to 0.
14913 *D P Chang <dpc@qualys.com>*
14915 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14917 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14918 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14919 worked incorrectly for those cases where range = `10..._2` and
14920 `3*range` is two bits longer than range.)
14924 * Only add signing time to PKCS7 structures if it is not already
14929 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14930 OBJ_ld_ce should be OBJ_id_ce.
14931 Also some ip-pda OIDs in crypto/objects/objects.txt were
14932 incorrect (cf. RFC 3039).
14934 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14936 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14937 returns early because it has nothing to do.
14939 *Andy Schneider <andy.schneider@bjss.co.uk>*
14941 * [In 0.9.6c-engine release:]
14942 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14944 *Andy Schneider <andy.schneider@bjss.co.uk>*
14946 * [In 0.9.6c-engine release:]
14947 Add support for Cryptographic Appliance's keyserver technology.
14948 (Use engine 'keyclient')
14950 *Cryptographic Appliances and Geoff Thorpe*
14952 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14953 is called via tools/c89.sh because arguments have to be
14954 rearranged (all '-L' options must appear before the first object
14957 *Richard Shapiro <rshapiro@abinitio.com>*
14959 * [In 0.9.6c-engine release:]
14960 Add support for Broadcom crypto accelerator cards, backported
14963 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14965 * [In 0.9.6c-engine release:]
14966 Add support for SureWare crypto accelerator cards from
14967 Baltimore Technologies. (Use engine 'sureware')
14969 *Baltimore Technologies and Mark Cox*
14971 * [In 0.9.6c-engine release:]
14972 Add support for crypto accelerator cards from Accelerated
14973 Encryption Processing, www.aep.ie. (Use engine 'aep')
14975 *AEP Inc. and Mark Cox*
14977 * Add a configuration entry for gcc on UnixWare.
14979 *Gary Benson <gbenson@redhat.com>*
14981 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14982 messages are stored in a single piece (fixed-length part and
14983 variable-length part combined) and fix various bugs found on the way.
14987 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14988 instead. BIO_gethostbyname() does not know what timeouts are
14989 appropriate, so entries would stay in cache even when they have
14991 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14993 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14994 faced with a pathologically small ClientHello fragment that does
14995 not contain client_version: Instead of aborting with an error,
14996 simply choose the highest available protocol version (i.e.,
14997 TLS 1.0 unless it is disabled). In practice, ClientHello
14998 messages are never sent like this, but this change gives us
14999 strictly correct behaviour at least for TLS.
15003 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15004 never resets s->method to s->ctx->method when called from within
15005 one of the SSL handshake functions.
15007 *Bodo Moeller; problem pointed out by Niko Baric*
15009 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15010 (sent using the client's version number) if client_version is
15011 smaller than the protocol version in use. Also change
15012 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15013 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15014 the client will at least see that alert.
15018 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15023 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15024 client receives HelloRequest while in a handshake.
15026 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15028 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15029 should end in 'break', not 'goto end' which circumvents various
15030 cleanups done in state SSL_ST_OK. But session related stuff
15031 must be disabled for SSL_ST_OK in the case that we just sent a
15034 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15035 before just sending a HelloRequest.
15037 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15039 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15040 reveal whether illegal block cipher padding was found or a MAC
15041 verification error occurred. (Neither SSLerr() codes nor alerts
15042 are directly visible to potential attackers, but the information
15043 may leak via logfiles.)
15045 Similar changes are not required for the SSL 2.0 implementation
15046 because the number of padding bytes is sent in clear for SSL 2.0,
15047 and the extra bytes are just ignored. However ssl/s2_pkt.c
15048 failed to verify that the purported number of padding bytes is in
15053 * Add OpenUNIX-8 support including shared libraries
15054 (Boyd Lynn Gerber <gerberb@zenez.com>).
15058 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15059 'wristwatch attack' using huge encoding parameters (cf.
15060 James H. Manger's CRYPTO 2001 paper). Note that the
15061 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15062 encoding parameters and hence was not vulnerable.
15066 * BN_sqr() bug fix.
15068 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15070 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15071 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15072 followed by modular reduction.
15074 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15076 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15077 equivalent based on BN_pseudo_rand() instead of BN_rand().
15081 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15082 This function was broken, as the check for a new client hello message
15083 to handle SGC did not allow these large messages.
15084 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15088 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15092 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15093 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15097 * Rework the configuration and shared library support for Tru64 Unix.
15098 The configuration part makes use of modern compiler features and
15099 still retains old compiler behavior for those that run older versions
15100 of the OS. The shared library support part includes a variant that
15101 uses the RPATH feature, and is available through the special
15102 configuration target "alpha-cc-rpath", which will never be selected
15105 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15107 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15108 with the same message size as in ssl3_get_certificate_request().
15109 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15110 messages might inadvertently be reject as too long.
15112 *Petr Lampa <lampa@fee.vutbr.cz>*
15114 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15118 * Modified SSL library such that the verify_callback that has been set
15119 specifically for an SSL object with SSL_set_verify() is actually being
15120 used. Before the change, a verify_callback set with this function was
15121 ignored and the verify_callback() set in the SSL_CTX at the time of
15122 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15123 to allow the necessary settings.
15127 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15128 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15129 done automatically (in contradiction to the requirements of the C
15130 standard). This made problems when used from OpenSSH.
15134 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15135 dh->length and always used
15137 BN_rand_range(priv_key, dh->p).
15139 BN_rand_range() is not necessary for Diffie-Hellman, and this
15140 specific range makes Diffie-Hellman unnecessarily inefficient if
15141 dh->length (recommended exponent length) is much smaller than the
15142 length of dh->p. We could use BN_rand_range() if the order of
15143 the subgroup was stored in the DH structure, but we only have
15148 BN_rand(priv_key, l, ...)
15150 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15157 RSA_eay_public_encrypt
15158 RSA_eay_private_decrypt
15159 RSA_eay_private_encrypt (signing)
15160 RSA_eay_public_decrypt (signature verification)
15162 (default implementations for RSA_public_encrypt,
15163 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15164 always reject numbers >= n.
15168 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15169 to synchronize access to 'locking_thread'. This is necessary on
15170 systems where access to 'locking_thread' (an 'unsigned long'
15171 variable) is not atomic.
15175 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15176 *before* setting the 'crypto_lock_rand' flag. The previous code had
15177 a race condition if 0 is a valid thread ID.
15179 *Travis Vitek <vitek@roguewave.com>*
15181 * Add support for shared libraries under Irix.
15183 *Albert Chin-A-Young <china@thewrittenword.com>*
15185 * Add configuration option to build on Linux on both big-endian and
15186 little-endian MIPS.
15188 *Ralf Baechle <ralf@uni-koblenz.de>*
15190 * Add the possibility to create shared libraries on HP-UX.
15194 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15196 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15197 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15198 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15199 PRNG state recovery was possible based on the output of
15200 one PRNG request appropriately sized to gain knowledge on
15201 'md' followed by enough consecutive 1-byte PRNG requests
15202 to traverse all of 'state'.
15204 1. When updating 'md_local' (the current thread's copy of 'md')
15205 during PRNG output generation, hash all of the previous
15206 'md_local' value, not just the half used for PRNG output.
15208 2. Make the number of bytes from 'state' included into the hash
15209 independent from the number of PRNG bytes requested.
15211 The first measure alone would be sufficient to avoid
15212 Markku-Juhani's attack. (Actually it had never occurred
15213 to me that the half of 'md_local' used for chaining was the
15214 half from which PRNG output bytes were taken -- I had always
15215 assumed that the secret half would be used.) The second
15216 measure makes sure that additional data from 'state' is never
15217 mixed into 'md_local' in small portions; this heuristically
15218 further strengthens the PRNG.
15222 * Fix crypto/bn/asm/mips3.s.
15226 * When only the key is given to "enc", the IV is undefined. Print out
15227 an error message in this case.
15231 * Handle special case when X509_NAME is empty in X509 printing routines.
15235 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15236 positive and less than q.
15240 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15241 used: it isn't thread safe and the add_lock_callback should handle
15244 *Paul Rose <Paul.Rose@bridge.com>*
15246 * Verify that incoming data obeys the block size in
15247 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15253 *Ulf Möller, Bodo Möller*
15255 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15256 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15257 when fixing the server behaviour for backwards-compatible 'client
15258 hello' messages. (Note that the attack is impractical against
15259 SSL 3.0 and TLS 1.0 anyway because length and version checking
15260 means that the probability of guessing a valid ciphertext is
15261 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15264 Before 0.9.5, the countermeasure (hide the error by generating a
15265 random 'decryption result') did not work properly because
15266 ERR_clear_error() was missing, meaning that SSL_get_error() would
15267 detect the supposedly ignored error.
15269 Both problems are now fixed.
15273 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15274 (previously it was 1024).
15278 * Fix for compatibility mode trust settings: ignore trust settings
15279 unless some valid trust or reject settings are present.
15283 * Fix for blowfish EVP: its a variable length cipher.
15287 * Fix various bugs related to DSA S/MIME verification. Handle missing
15288 parameters in DSA public key structures and return an error in the
15289 DSA routines if parameters are absent.
15293 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15294 in the current directory if neither $RANDFILE nor $HOME was set.
15295 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15296 caused some confusion to Windows users who haven't defined $HOME.
15297 Thus RAND_file_name() is changed again: e_os.h can define a
15298 DEFAULT_HOME, which will be used if $HOME is not set.
15299 For Windows, we use "C:"; on other platforms, we still require
15300 environment variables.
15302 * Move 'if (!initialized) RAND_poll()' into regions protected by
15303 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15304 having multiple threads call RAND_poll() concurrently.
15308 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15309 combination of a flag and a thread ID variable.
15310 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15311 flag), *other* threads can enter ssleay_add_bytes without obeying
15312 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15313 that they do not hold after the first thread unsets add_do_not_lock).
15317 * Change bctest again: '-x' expressions are not available in all
15318 versions of 'test'.
15322 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15324 * Fix a couple of memory leaks in PKCS7_dataDecode()
15326 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15328 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15329 the default extension for executables, if any. Also, make the perl
15330 scripts that use symlink() to test if it really exists and use "cp"
15331 if it doesn't. All this made OpenSSL compilable and installable in
15336 * Fix for asn1_GetSequence() for indefinite length constructed data.
15337 If SEQUENCE is length is indefinite just set c->slen to the total
15338 amount of data available.
15340 *Steve Henson, reported by shige@FreeBSD.org*
15342 *This change does not apply to 0.9.7.*
15344 * Change bctest to avoid here-documents inside command substitution
15345 (workaround for FreeBSD /bin/sh bug).
15346 For compatibility with Ultrix, avoid shell functions (introduced
15347 in the bctest version that searches along $PATH).
15351 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15352 with des_encrypt() defined on some operating systems, like Solaris
15357 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15358 On the Importance of Eliminating Errors in Cryptographic
15359 Computations, J. Cryptology 14 (2001) 2, 101-119,
15360 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15364 * MIPS assembler BIGNUM division bug fix.
15368 * Disabled incorrect Alpha assembler code.
15372 * Fix PKCS#7 decode routines so they correctly update the length
15373 after reading an EOC for the EXPLICIT tag.
15377 *This change does not apply to 0.9.7.*
15379 * Fix bug in PKCS#12 key generation routines. This was triggered
15380 if a 3DES key was generated with a 0 initial byte. Include
15381 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15382 (but broken) behaviour.
15386 * Enhance bctest to search for a working bc along $PATH and print
15389 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15391 * Fix memory leaks in err.c: free err_data string if necessary;
15392 don't write to the wrong index in ERR_set_error_data.
15396 * Implement ssl23_peek (analogous to ssl23_read), which previously
15401 * Replace rdtsc with `_emit` statements for VC++ version 5.
15403 *Jeremy Cooper <jeremy@baymoo.org>*
15405 * Make it possible to reuse SSLv2 sessions.
15409 * In copy_email() check for >= 0 as a return value for
15410 X509_NAME_get_index_by_NID() since 0 is a valid index.
15412 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15414 * Avoid coredump with unsupported or invalid public keys by checking if
15415 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15416 PKCS7_verify() fails with non detached data.
15420 * Don't use getenv in library functions when run as setuid/setgid.
15421 New function OPENSSL_issetugid().
15425 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15426 due to incorrect handling of multi-threading:
15428 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15430 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15432 3. Count how many times MemCheck_off() has been called so that
15433 nested use can be treated correctly. This also avoids
15434 inband-signalling in the previous code (which relied on the
15435 assumption that thread ID 0 is impossible).
15439 * Add "-rand" option also to s_client and s_server.
15443 * Fix CPU detection on Irix 6.x.
15444 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15445 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15447 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15452 *This change does not apply to 0.9.7.*
15454 * Use the cached encoding of an X509_NAME structure rather than
15455 copying it. This is apparently the reason for the libsafe "errors"
15456 but the code is actually correct.
15460 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15461 Bleichenbacher's DSA attack.
15462 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15463 to be set and top=0 forces the highest bit to be set; top=-1 is new
15464 and leaves the highest bit random.
15466 *Ulf Moeller, Bodo Moeller*
15468 * In the `NCONF_...`-based implementations for `CONF_...` queries
15469 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15470 a temporary CONF structure with the data component set to NULL
15471 (which gives segmentation faults in lh_retrieve).
15472 Instead, use NULL for the CONF pointer in CONF_get_string and
15473 CONF_get_number (which may use environment variables) and directly
15474 return NULL from CONF_get_section.
15478 * Fix potential buffer overrun for EBCDIC.
15482 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15483 keyUsage if basicConstraints absent for a CA.
15487 * Make SMIME_write_PKCS7() write mail header values with a format that
15488 is more generally accepted (no spaces before the semicolon), since
15489 some programs can't parse those values properly otherwise. Also make
15490 sure BIO's that break lines after each write do not create invalid
15495 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15496 macros previously used would not encode an empty SEQUENCE OF
15497 and break the signature.
15501 *This change does not apply to 0.9.7.*
15503 * Zero the premaster secret after deriving the master secret in
15508 * Add some EVP_add_digest_alias registrations (as found in
15509 OpenSSL_add_all_digests()) to SSL_library_init()
15510 aka OpenSSL_add_ssl_algorithms(). This provides improved
15511 compatibility with peers using X.509 certificates
15512 with unconventional AlgorithmIdentifier OIDs.
15516 * Fix for Irix with NO_ASM.
15518 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15520 * ./config script fixes.
15522 *Ulf Moeller, Richard Levitte*
15524 * Fix 'openssl passwd -1'.
15528 * Change PKCS12_key_gen_asc() so it can cope with non null
15529 terminated strings whose length is passed in the passlen
15530 parameter, for example from PEM callbacks. This was done
15531 by adding an extra length parameter to asc2uni().
15533 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15535 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15536 call failed, free the DSA structure.
15540 * Fix to uni2asc() to cope with zero length Unicode strings.
15541 These are present in some PKCS#12 files.
15545 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15546 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15547 when writing a 32767 byte record.
15549 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15551 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15552 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15554 (RSA objects have a reference count access to which is protected
15555 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15556 so they are meant to be shared between threads.)
15557 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15558 "Reddie, Steven" <Steven.Reddie@ca.com>*
15560 * Fix a deadlock in CRYPTO_mem_leaks().
15564 * Use better test patterns in bntest.
15568 * rand_win.c fix for Borland C.
15572 * BN_rshift bugfix for n == 0.
15576 * Add a 'bctest' script that checks for some known 'bc' bugs
15577 so that 'make test' does not abort just because 'bc' is broken.
15581 * Store verify_result within SSL_SESSION also for client side to
15582 avoid potential security hole. (Re-used sessions on the client side
15583 always resulted in verify_result==X509_V_OK, not using the original
15584 result of the server certificate verification.)
15588 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15589 SSL3_RT_APPLICATION_DATA, return 0.
15590 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15595 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15596 releases, have been re-implemented by renaming the previous
15597 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15598 and ssl3_read_internal, respectively, and adding 'peek' parameters
15599 to them. The new ssl[23]_{read,peek} functions are calls to
15600 ssl[23]_read_internal with the 'peek' flag set appropriately.
15601 A 'peek' parameter has also been added to ssl3_read_bytes, which
15602 does the actual work for ssl3_read_internal.
15606 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15607 the method-specific "init()" handler. Also clean up ex_data after
15608 calling the method-specific "finish()" handler. Previously, this was
15609 happening the other way round.
15613 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15614 The previous value, 12, was not always sufficient for BN_mod_exp().
15618 * Make sure that shared libraries get the internal name engine with
15619 the full version number and not just 0. This should mark the
15620 shared libraries as not backward compatible. Of course, this should
15621 be changed again when we can guarantee backward binary compatibility.
15625 * Fix typo in get_cert_by_subject() in by_dir.c
15627 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15629 * Rework the system to generate shared libraries:
15631 - Make note of the expected extension for the shared libraries and
15632 if there is a need for symbolic links from for example libcrypto.so.0
15633 to libcrypto.so.0.9.7. There is extended info in Configure for
15636 - Make as few rebuilds of the shared libraries as possible.
15638 - Still avoid linking the OpenSSL programs with the shared libraries.
15640 - When installing, install the shared libraries separately from the
15645 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15647 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15648 and not in SSL_clear because the latter is also used by the
15649 accept/connect functions; previously, the settings made by
15650 SSL_set_read_ahead would be lost during the handshake.
15652 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15654 * Correct util/mkdef.pl to be selective about disabled algorithms.
15655 Previously, it would create entries for disabled algorithms no
15660 * Added several new manual pages for SSL_* function.
15664 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15666 * In ssl23_get_client_hello, generate an error message when faced
15667 with an initial SSL 3.0/TLS record that is too small to contain the
15668 first two bytes of the ClientHello message, i.e. client_version.
15669 (Note that this is a pathologic case that probably has never happened
15670 in real life.) The previous approach was to use the version number
15671 from the record header as a substitute; but our protocol choice
15672 should not depend on that one because it is not authenticated
15673 by the Finished messages.
15677 * More robust randomness gathering functions for Windows.
15679 *Jeffrey Altman <jaltman@columbia.edu>*
15681 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15682 not set then we don't setup the error code for issuer check errors
15683 to avoid possibly overwriting other errors which the callback does
15684 handle. If an application does set the flag then we assume it knows
15685 what it is doing and can handle the new informational codes
15690 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15691 a general "ANY" type, as such it should be able to decode anything
15692 including tagged types. However it didn't check the class so it would
15693 wrongly interpret tagged types in the same way as their universal
15694 counterpart and unknown types were just rejected. Changed so that the
15695 tagged and unknown types are handled in the same way as a SEQUENCE:
15696 that is the encoding is stored intact. There is also a new type
15697 "V_ASN1_OTHER" which is used when the class is not universal, in this
15698 case we have no idea what the actual type is so we just lump them all
15703 * On VMS, stdout may very well lead to a file that is written to
15704 in a record-oriented fashion. That means that every write() will
15705 write a separate record, which will be read separately by the
15706 programs trying to read from it. This can be very confusing.
15708 The solution is to put a BIO filter in the way that will buffer
15709 text until a linefeed is reached, and then write everything a
15710 line at a time, so every record written will be an actual line,
15711 not chunks of lines and not (usually doesn't happen, but I've
15712 seen it once) several lines in one record. BIO_f_linebuffer() is
15715 Currently, it's a VMS-only method, because that's where it has
15716 been tested well enough.
15720 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15721 it can return incorrect results.
15722 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15723 but it was in 0.9.6-beta[12].)
15727 * Disable the check for content being present when verifying detached
15728 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15729 include zero length content when signing messages.
15733 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15734 BIO_ctrl (for BIO pairs).
15738 * Add DSO method for VMS.
15742 * Bug fix: Montgomery multiplication could produce results with the
15747 * Add RPM specification openssl.spec and modify it to build three
15748 packages. The default package contains applications, application
15749 documentation and run-time libraries. The devel package contains
15750 include files, static libraries and function documentation. The
15751 doc package contains the contents of the doc directory. The original
15752 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15756 * Add a large number of documentation files for many SSL routines.
15758 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15760 * Add a configuration entry for Sony News 4.
15762 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15764 * Don't set the two most significant bits to one when generating a
15765 random number < q in the DSA library.
15769 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15770 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15771 the underlying transport is blocking) if a handshake took place.
15772 (The default behaviour is needed by applications such as s_client
15773 and s_server that use select() to determine when to use SSL_read;
15774 but for applications that know in advance when to expect data, it
15775 just makes things more complicated.)
15779 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15784 * Add a few more EBCDIC conditionals that make `req` and `x509`
15785 work better on such systems.
15787 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15789 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15790 Update PKCS12_parse() so it copies the friendlyName and the
15791 keyid to the certificates aux info.
15795 * Fix bug in PKCS7_verify() which caused an infinite loop
15796 if there was more than one signature.
15798 *Sven Uszpelkat <su@celocom.de>*
15800 * Major change in util/mkdef.pl to include extra information
15801 about each symbol, as well as presenting variables as well
15802 as functions. This change means that there's n more need
15803 to rebuild the .num files when some algorithms are excluded.
15807 * Allow the verify time to be set by an application,
15808 rather than always using the current time.
15812 * Phase 2 verify code reorganisation. The certificate
15813 verify code now looks up an issuer certificate by a
15814 number of criteria: subject name, authority key id
15815 and key usage. It also verifies self signed certificates
15816 by the same criteria. The main comparison function is
15817 X509_check_issued() which performs these checks.
15819 Lot of changes were necessary in order to support this
15820 without completely rewriting the lookup code.
15822 Authority and subject key identifier are now cached.
15824 The LHASH 'certs' is X509_STORE has now been replaced
15825 by a STACK_OF(X509_OBJECT). This is mainly because an
15826 LHASH can't store or retrieve multiple objects with
15827 the same hash value.
15829 As a result various functions (which were all internal
15830 use only) have changed to handle the new X509_STORE
15831 structure. This will break anything that messed round
15832 with X509_STORE internally.
15834 The functions X509_STORE_add_cert() now checks for an
15835 exact match, rather than just subject name.
15837 The X509_STORE API doesn't directly support the retrieval
15838 of multiple certificates matching a given criteria, however
15839 this can be worked round by performing a lookup first
15840 (which will fill the cache with candidate certificates)
15841 and then examining the cache for matches. This is probably
15842 the best we can do without throwing out X509_LOOKUP
15843 entirely (maybe later...).
15845 The X509_VERIFY_CTX structure has been enhanced considerably.
15847 All certificate lookup operations now go via a get_issuer()
15848 callback. Although this currently uses an X509_STORE it
15849 can be replaced by custom lookups. This is a simple way
15850 to bypass the X509_STORE hackery necessary to make this
15851 work and makes it possible to use more efficient techniques
15852 in future. A very simple version which uses a simple
15853 STACK for its trusted certificate store is also provided
15854 using X509_STORE_CTX_trusted_stack().
15856 The verify_cb() and verify() callbacks now have equivalents
15857 in the X509_STORE_CTX structure.
15859 X509_STORE_CTX also has a 'flags' field which can be used
15860 to customise the verify behaviour.
15864 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15865 excludes S/MIME capabilities.
15869 * When a certificate request is read in keep a copy of the
15870 original encoding of the signed data and use it when outputting
15871 again. Signatures then use the original encoding rather than
15872 a decoded, encoded version which may cause problems if the
15873 request is improperly encoded.
15877 * For consistency with other BIO_puts implementations, call
15878 buffer_write(b, ...) directly in buffer_puts instead of calling
15881 In BIO_puts, increment b->num_write as in BIO_write.
15883 *Peter.Sylvester@EdelWeb.fr*
15885 * Fix BN_mul_word for the case where the word is 0. (We have to use
15886 BN_zero, we may not return a BIGNUM with an array consisting of
15887 words set to zero.)
15891 * Avoid calling abort() from within the library when problems are
15892 detected, except if preprocessor symbols have been defined
15893 (such as REF_CHECK, BN_DEBUG etc.).
15897 * New openssl application 'rsautl'. This utility can be
15898 used for low-level RSA operations. DER public key
15899 BIO/fp routines also added.
15903 * New Configure entry and patches for compiling on QNX 4.
15905 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15907 * A demo state-machine implementation was sponsored by
15908 Nuron (<http://www.nuron.com/>) and is now available in
15909 demos/state_machine.
15913 * New options added to the 'dgst' utility for signature
15914 generation and verification.
15918 * Unrecognized PKCS#7 content types are now handled via a
15919 catch all ASN1_TYPE structure. This allows unsupported
15920 types to be stored as a "blob" and an application can
15921 encode and decode it manually.
15925 * Fix various signed/unsigned issues to make a_strex.c
15926 compile under VC++.
15928 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15930 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15931 length if passed a buffer. ASN1_INTEGER_to_BN failed
15932 if passed a NULL BN and its argument was negative.
15934 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15936 * Modification to PKCS#7 encoding routines to output definite
15937 length encoding. Since currently the whole structures are in
15938 memory there's not real point in using indefinite length
15939 constructed encoding. However if OpenSSL is compiled with
15940 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15944 * Added BIO_vprintf() and BIO_vsnprintf().
15948 * Added more prefixes to parse for in the strings written
15949 through a logging bio, to cover all the levels that are available
15950 through syslog. The prefixes are now:
15952 PANIC, EMERG, EMR => LOG_EMERG
15953 ALERT, ALR => LOG_ALERT
15954 CRIT, CRI => LOG_CRIT
15955 ERROR, ERR => LOG_ERR
15956 WARNING, WARN, WAR => LOG_WARNING
15957 NOTICE, NOTE, NOT => LOG_NOTICE
15958 INFO, INF => LOG_INFO
15959 DEBUG, DBG => LOG_DEBUG
15961 and as before, if none of those prefixes are present at the
15962 beginning of the string, LOG_ERR is chosen.
15964 On Win32, the `LOG_*` levels are mapped according to this:
15966 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15967 LOG_WARNING => EVENTLOG_WARNING_TYPE
15968 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15972 * Made it possible to reconfigure with just the configuration
15973 argument "reconf" or "reconfigure". The command line arguments
15974 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15975 and are retrieved from there when reconfiguring.
15981 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15983 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15987 * The obj_dat.pl script was messing up the sorting of object
15988 names. The reason was that it compared the quoted version
15989 of strings as a result "OCSP" > "OCSP Signing" because
15990 " > SPACE. Changed script to store unquoted versions of
15991 names and add quotes on output. It was also omitting some
15992 names from the lookup table if they were given a default
15993 value (that is if SN is missing it is given the same
15994 value as LN and vice versa), these are now added on the
15995 grounds that if an object has a name we should be able to
15996 look it up. Finally added warning output when duplicate
15997 short or long names are found.
16001 * Changes needed for Tandem NSK.
16003 *Scott Uroff <scott@xypro.com>*
16005 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16006 RSA_padding_check_SSLv23(), special padding was never detected
16007 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16008 version rollback attacks was not effective.
16010 In s23_clnt.c, don't use special rollback-attack detection padding
16011 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16012 client; similarly, in s23_srvr.c, don't do the rollback check if
16013 SSL 2.0 is the only protocol enabled in the server.
16017 * Make it possible to get hexdumps of unprintable data with 'openssl
16018 asn1parse'. By implication, the functions ASN1_parse_dump() and
16019 BIO_dump_indent() are added.
16023 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16024 these print out strings and name structures based on various
16025 flags including RFC2253 support and proper handling of
16026 multibyte characters. Added options to the 'x509' utility
16027 to allow the various flags to be set.
16031 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16032 Also change the functions X509_cmp_current_time() and
16033 X509_gmtime_adj() work with an ASN1_TIME structure,
16034 this will enable certificates using GeneralizedTime in validity
16035 dates to be checked.
16039 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16040 negative public key encodings) on by default,
16041 NO_NEG_PUBKEY_BUG can be set to disable it.
16045 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16046 content octets. An i2c_ASN1_OBJECT is unnecessary because
16047 the encoding can be trivially obtained from the structure.
16051 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16052 not read locks (`CRYPTO_r_[un]lock`).
16056 * A first attempt at creating official support for shared
16057 libraries through configuration. I've kept it so the
16058 default is static libraries only, and the OpenSSL programs
16059 are always statically linked for now, but there are
16060 preparations for dynamic linking in place.
16061 This has been tested on Linux and Tru64.
16065 * Randomness polling function for Win9x, as described in:
16066 Peter Gutmann, Software Generation of Practically Strong
16071 * Fix so PRNG is seeded in req if using an already existing
16076 * New options to smime application. -inform and -outform
16077 allow alternative formats for the S/MIME message including
16078 PEM and DER. The -content option allows the content to be
16079 specified separately. This should allow things like Netscape
16080 form signing output easier to verify.
16084 * Fix the ASN1 encoding of tags using the 'long form'.
16088 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16089 STRING types. These convert content octets to and from the
16090 underlying type. The actual tag and length octets are
16091 already assumed to have been read in and checked. These
16092 are needed because all other string types have virtually
16093 identical handling apart from the tag. By having versions
16094 of the ASN1 functions that just operate on content octets
16095 IMPLICIT tagging can be handled properly. It also allows
16096 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16097 and ASN1_INTEGER are identical apart from the tag.
16101 * Change the handling of OID objects as follows:
16103 - New object identifiers are inserted in objects.txt, following
16104 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16105 - objects.pl is used to process obj_mac.num and create a new
16107 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16110 This is currently kind of a hack, and the perl code in objects.pl
16111 isn't very elegant, but it works as I intended. The simplest way
16112 to check that it worked correctly is to look in obj_dat.h and
16113 check the array nid_objs and make sure the objects haven't moved
16114 around (this is important!). Additions are OK, as well as
16115 consistent name changes.
16119 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16123 * Addition of the command line parameter '-rand file' to 'openssl req'.
16124 The given file adds to whatever has already been seeded into the
16125 random pool through the RANDFILE configuration file option or
16126 environment variable, or the default random state file.
16130 * mkstack.pl now sorts each macro group into lexical order.
16131 Previously the output order depended on the order the files
16132 appeared in the directory, resulting in needless rewriting
16137 * Patches to make OpenSSL compile under Win32 again. Mostly
16138 work arounds for the VC++ problem that it treats func() as
16139 func(void). Also stripped out the parts of mkdef.pl that
16140 added extra typesafe functions: these no longer exist.
16144 * Reorganisation of the stack code. The macros are now all
16145 collected in safestack.h . Each macro is defined in terms of
16146 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16147 DEBUG_SAFESTACK is now handled in terms of function casts,
16148 this has the advantage of retaining type safety without the
16149 use of additional functions. If DEBUG_SAFESTACK is not defined
16150 then the non typesafe macros are used instead. Also modified the
16151 mkstack.pl script to handle the new form. Needs testing to see
16152 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16153 the default if no major problems. Similar behaviour for ASN1_SET_OF
16154 and PKCS12_STACK_OF.
16158 * When some versions of IIS use the 'NET' form of private key the
16159 key derivation algorithm is different. Normally MD5(password) is
16160 used as a 128 bit RC4 key. In the modified case
16161 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16162 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16163 as the old Netscape_RSA functions except they have an additional
16164 'sgckey' parameter which uses the modified algorithm. Also added
16165 an -sgckey command line option to the rsa utility. Thanks to
16166 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16167 algorithm to openssl-dev.
16171 * The evp_local.h macros were using 'c.##kname' which resulted in
16172 invalid expansion on some systems (SCO 5.0.5 for example).
16173 Corrected to 'c.kname'.
16175 *Phillip Porch <root@theporch.com>*
16177 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16178 a STACK of email addresses from a certificate or request, these look
16179 in the subject name and the subject alternative name extensions and
16180 omit any duplicate addresses.
16184 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16185 This makes DSA verification about 2 % faster.
16189 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16190 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16191 plus overhead for 1024 bit moduli).
16192 This makes exponentiations about 0.5 % faster for 1024 bit
16193 exponents (as measured by "openssl speed rsa2048").
16197 * Rename memory handling macros to avoid conflicts with other
16199 Malloc => OPENSSL_malloc
16200 Malloc_locked => OPENSSL_malloc_locked
16201 Realloc => OPENSSL_realloc
16202 Free => OPENSSL_free
16206 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16207 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16211 * CygWin32 support.
16213 *John Jarvie <jjarvie@newsguy.com>*
16215 * The type-safe stack code has been rejigged. It is now only compiled
16216 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16217 by default all type-specific stack functions are "#define"d back to
16218 standard stack functions. This results in more streamlined output
16219 but retains the type-safety checking possibilities of the original
16224 * The STACK code has been cleaned up, and certain type declarations
16225 that didn't make a lot of sense have been brought in line. This has
16226 also involved a cleanup of sorts in safestack.h to more correctly
16227 map type-safe stack functions onto their plain stack counterparts.
16228 This work has also resulted in a variety of "const"ifications of
16229 lots of the code, especially `_cmp` operations which should normally
16230 be prototyped with "const" parameters anyway.
16234 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16235 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16236 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16237 where all of 'md' is used each time the PRNG is used, but 'state'
16238 is used only indexed by a cyclic counter. As entropy may not be
16239 well distributed from the beginning, 'md' is important as a
16240 chaining variable. However, the output function chains only half
16241 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16242 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16243 in all of 'state' being rewritten, with the new values depending
16244 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16248 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16249 the handshake is continued after ssl_verify_cert_chain();
16250 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16251 can lead to 'unexplainable' connection aborts later.
16253 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16255 * Major EVP API cipher revision.
16256 Add hooks for extra EVP features. This allows various cipher
16257 parameters to be set in the EVP interface. Support added for variable
16258 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16259 setting of RC2 and RC5 parameters.
16261 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16264 Remove lots of duplicated code from the EVP library. For example *every*
16265 cipher init() function handles the 'iv' in the same way according to the
16266 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16267 for CFB and OFB modes they zero ctx->num.
16269 New functionality allows removal of S/MIME code RC2 hack.
16271 Most of the routines have the same form and so can be declared in terms
16274 By shifting this to the top level EVP_CipherInit() it can be removed from
16275 all individual ciphers. If the cipher wants to handle IVs or keys
16276 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16279 Change lots of functions like EVP_EncryptUpdate() to now return a
16280 value: although software versions of the algorithms cannot fail
16281 any installed hardware versions can.
16285 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16286 this option is set, tolerate broken clients that send the negotiated
16287 protocol version number instead of the requested protocol version
16292 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16293 i.e. non-zero for export ciphersuites, zero otherwise.
16294 Previous versions had this flag inverted, inconsistent with
16295 rsa_tmp_cb (..._TMP_RSA_CB).
16297 *Bodo Moeller; problem reported by Amit Chopra*
16299 * Add missing DSA library text string. Work around for some IIS
16300 key files with invalid SEQUENCE encoding.
16304 * Add a document (doc/standards.txt) that list all kinds of standards
16305 and so on that are implemented in OpenSSL.
16309 * Enhance c_rehash script. Old version would mishandle certificates
16310 with the same subject name hash and wouldn't handle CRLs at all.
16311 Added -fingerprint option to crl utility, to support new c_rehash
16316 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16320 * Fix for SSL server purpose checking. Server checking was
16321 rejecting certificates which had extended key usage present
16322 but no ssl client purpose.
16324 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16326 * Make PKCS#12 code work with no password. The PKCS#12 spec
16327 is a little unclear about how a blank password is handled.
16328 Since the password in encoded as a BMPString with terminating
16329 double NULL a zero length password would end up as just the
16330 double NULL. However no password at all is different and is
16331 handled differently in the PKCS#12 key generation code. NS
16332 treats a blank password as zero length. MSIE treats it as no
16333 password on export: but it will try both on import. We now do
16334 the same: PKCS12_parse() tries zero length and no password if
16335 the password is set to "" or NULL (NULL is now a valid password:
16336 it wasn't before) as does the pkcs12 application.
16340 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16341 perror when PEM_read_bio_X509_REQ fails, the error message must
16342 be obtained from the error queue.
16346 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16347 it in ERR_remove_state if appropriate, and change ERR_get_state
16348 accordingly to avoid race conditions (this is necessary because
16349 thread_hash is no longer constant once set).
16353 * Bugfix for linux-elf makefile.one.
16357 * RSA_get_default_method() will now cause a default
16358 RSA_METHOD to be chosen if one doesn't exist already.
16359 Previously this was only set during a call to RSA_new()
16360 or RSA_new_method(NULL) meaning it was possible for
16361 RSA_get_default_method() to return NULL.
16365 * Added native name translation to the existing DSO code
16366 that will convert (if the flag to do so is set) filenames
16367 that are sufficiently small and have no path information
16368 into a canonical native form. Eg. "blah" converted to
16369 "libblah.so" or "blah.dll" etc.
16373 * New function ERR_error_string_n(e, buf, len) which is like
16374 ERR_error_string(e, buf), but writes at most 'len' bytes
16375 including the 0 terminator. For ERR_error_string_n, 'buf'
16378 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16380 * CONF library reworked to become more general. A new CONF
16381 configuration file reader "class" is implemented as well as a
16382 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16383 old `CONF_*` functions are still there, but are reimplemented to
16384 work in terms of the new functions. Also, a set of functions
16385 to handle the internal storage of the configuration data is
16386 provided to make it easier to write new configuration file
16387 reader "classes" (I can definitely see something reading a
16388 configuration file in XML format, for example), called `_CONF_*`,
16389 or "the configuration storage API"...
16391 The new configuration file reading functions are:
16393 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16394 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16396 NCONF_default, NCONF_WIN32
16398 NCONF_dump_fp, NCONF_dump_bio
16400 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16401 NCONF_new creates a new CONF object. This works in the same way
16402 as other interfaces in OpenSSL, like the BIO interface.
16403 `NCONF_dump_*` dump the internal storage of the configuration file,
16404 which is useful for debugging. All other functions take the same
16405 arguments as the old `CONF_*` functions with the exception of the
16406 first that must be a `CONF *` instead of a `LHASH *`.
16408 To make it easier to use the new classes with the old `CONF_*` functions,
16409 the function CONF_set_default_method is provided.
16413 * Add '-tls1' option to 'openssl ciphers', which was already
16414 mentioned in the documentation but had not been implemented.
16415 (This option is not yet really useful because even the additional
16416 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16420 * Initial DSO code added into libcrypto for letting OpenSSL (and
16421 OpenSSL-based applications) load shared libraries and bind to
16422 them in a portable way.
16424 *Geoff Thorpe, with contributions from Richard Levitte*
16426 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16428 * Make sure _lrotl and _lrotr are only used with MSVC.
16430 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16431 (the default implementation of RAND_status).
16433 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16434 to '-clrext' (= clear extensions), as intended and documented.
16435 *Bodo Moeller; inconsistency pointed out by Michael Attili
16436 <attili@amaxo.com>*
16438 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16439 was larger than the MD block size.
16441 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16443 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16444 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16445 using the passed key: if the passed key was a private key the result
16446 of X509_print(), for example, would be to print out all the private key
16451 * des_quad_cksum() byte order bug fix.
16452 *Ulf Möller, using the problem description in krb4-0.9.7, where
16453 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16455 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16458 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16460 * For easily testing in shell scripts whether some command
16461 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16462 returns with exit code 0 iff no command of the given name is available.
16463 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16464 the output goes to stdout and nothing is printed to stderr.
16465 Additional arguments are always ignored.
16467 Since for each cipher there is a command of the same name,
16468 the 'no-cipher' compilation switches can be tested this way.
16470 ('openssl no-XXX' is not able to detect pseudo-commands such
16471 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16475 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16479 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16480 is set; it will be thrown away anyway because each handshake creates
16482 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16483 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16484 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16485 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16489 * New s_client option -ign_eof: EOF at stdin is ignored, and
16490 'Q' and 'R' lose their special meanings (quit/renegotiate).
16491 This is part of what -quiet does; unlike -quiet, -ign_eof
16492 does not suppress any output.
16496 * Add compatibility options to the purpose and trust code. The
16497 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16498 accepts a certificate or CA, this was the previous behaviour,
16499 with all the associated security issues.
16501 X509_TRUST_COMPAT is the old trust behaviour: only and
16502 automatically trust self signed roots in certificate store. A
16503 new trust setting X509_TRUST_DEFAULT is used to specify that
16504 a purpose has no associated trust setting and it should instead
16505 use the value in the default purpose.
16509 * Fix the PKCS#8 DSA private key code so it decodes keys again
16510 and fix a memory leak.
16514 * In util/mkerr.pl (which implements 'make errors'), preserve
16515 reason strings from the previous version of the .c file, as
16516 the default to have only downcase letters (and digits) in
16517 automatically generated reasons codes is not always appropriate.
16521 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16522 using strerror. Previously, ERR_reason_error_string() returned
16523 library names as reason strings for SYSerr; but SYSerr is a special
16524 case where small numbers are errno values, not library numbers.
16528 * Add '-dsaparam' option to 'openssl dhparam' application. This
16529 converts DSA parameters into DH parameters. (When creating parameters,
16530 DSA_generate_parameters is used.)
16534 * Include 'length' (recommended exponent length) in C code generated
16535 by 'openssl dhparam -C'.
16539 * The second argument to set_label in perlasm was already being used
16540 so couldn't be used as a "file scope" flag. Moved to third argument
16545 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16546 instead of RAND_bytes for encryption IVs and salts.
16550 * Include RAND_status() into RAND_METHOD instead of implementing
16551 it only for md_rand.c Otherwise replacing the PRNG by calling
16552 RAND_set_rand_method would be impossible.
16556 * Don't let DSA_generate_key() enter an infinite loop if the random
16557 number generation fails.
16561 * New 'rand' application for creating pseudo-random output.
16565 * Added configuration support for Linux/IA64
16567 *Rolf Haberrecker <rolf@suse.de>*
16569 * Assembler module support for Mingw32.
16573 * Shared library support for HPUX (in shlib/).
16575 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16577 * Shared library support for Solaris gcc.
16579 *Lutz Behnke <behnke@trustcenter.de>*
16581 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16583 * PKCS7_encrypt() was adding text MIME headers twice because they
16584 were added manually and by SMIME_crlf_copy().
16588 * In bntest.c don't call BN_rand with zero bits argument.
16590 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16592 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16593 case was implemented. This caused BN_div_recp() to fail occasionally.
16597 * Add an optional second argument to the set_label() in the perl
16598 assembly language builder. If this argument exists and is set
16599 to 1 it signals that the assembler should use a symbol whose
16600 scope is the entire file, not just the current function. This
16601 is needed with MASM which uses the format label:: for this scope.
16603 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16605 * Change the ASN1 types so they are typedefs by default. Before
16606 almost all types were #define'd to ASN1_STRING which was causing
16607 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16612 * Change names of new functions to the new get1/get0 naming
16613 convention: After 'get1', the caller owns a reference count
16614 and has to call `..._free`; 'get0' returns a pointer to some
16615 data structure without incrementing reference counters.
16616 (Some of the existing 'get' functions increment a reference
16617 counter, some don't.)
16618 Similarly, 'set1' and 'add1' functions increase reference
16619 counters or duplicate objects.
16623 * Allow for the possibility of temp RSA key generation failure:
16624 the code used to assume it always worked and crashed on failure.
16628 * Fix potential buffer overrun problem in BIO_printf().
16629 *Ulf Möller, using public domain code by Patrick Powell; problem
16630 pointed out by David Sacerdote <das33@cornell.edu>*
16632 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16633 RAND_egd() and RAND_status(). In the command line application,
16634 the EGD socket can be specified like a seed file using RANDFILE
16639 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16640 Some CAs (e.g. Verisign) distribute certificates in this form.
16644 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16645 list to exclude them. This means that no special compilation option
16646 is needed to use anonymous DH: it just needs to be included in the
16651 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16652 EVP_MD_type. The old functionality is available in a new macro called
16653 EVP_MD_md(). Change code that uses it and update docs.
16657 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16658 where the `void *` argument is replaced by a function pointer argument.
16659 Previously `void *` was abused to point to functions, which works on
16660 many platforms, but is not correct. As these functions are usually
16661 called by macros defined in OpenSSL header files, most source code
16662 should work without changes.
16666 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16667 sections with information on -D... compiler switches used for
16668 compiling the library so that applications can see them. To enable
16669 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16670 must be defined. E.g.,
16671 #define OPENSSL_ALGORITHM_DEFINES
16672 #include <openssl/opensslconf.h>
16673 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16675 *Richard Levitte, Ulf and Bodo Möller*
16677 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16682 * Change the 'other' type in certificate aux info to a STACK_OF
16683 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16684 the required ASN1 format: arbitrary types determined by an OID.
16688 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16689 argument to 'req'. This is not because the function is newer or
16690 better than others it just uses the work 'NEW' in the certificate
16691 request header lines. Some software needs this.
16695 * Reorganise password command line arguments: now passwords can be
16696 obtained from various sources. Delete the PEM_cb function and make
16697 it the default behaviour: i.e. if the callback is NULL and the
16698 usrdata argument is not NULL interpret it as a null terminated pass
16699 phrase. If usrdata and the callback are NULL then the pass phrase
16700 is prompted for as usual.
16704 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16705 the support is automatically enabled. The resulting binaries will
16706 autodetect the card and use it if present.
16708 *Ben Laurie and Compaq Inc.*
16710 * Work around for Netscape hang bug. This sends certificate request
16711 and server done in one record. Since this is perfectly legal in the
16712 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16713 the bugs/SSLv3 entry for more info.
16717 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16721 * Add -rand argument to smime and pkcs12 applications and read/write
16726 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16730 * Add command line password options to the remaining applications.
16734 * Bug fix for BN_div_recp() for numerators with an even number of
16739 * More tests in bntest.c, and changed test_bn output.
16743 * ./config recognizes MacOS X now.
16747 * Bug fix for BN_div() when the first words of num and divisor are
16748 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16752 * Add support for various broken PKCS#8 formats, and command line
16753 options to produce them.
16757 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16758 get temporary BIGNUMs from a BN_CTX.
16762 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16767 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16768 include a #define from the old name to the new. The original intent
16769 was that statically linked binaries could for example just call
16770 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16771 link with digests. This never worked because SSLeay_add_all_digests()
16772 and SSLeay_add_all_ciphers() were in the same source file so calling
16773 one would link with the other. They are now in separate source files.
16777 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16781 * Use a less unusual form of the Miller-Rabin primality test (it used
16782 a binary algorithm for exponentiation integrated into the Miller-Rabin
16783 loop, our standard modexp algorithms are faster).
16787 * Support for the EBCDIC character set completed.
16789 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16791 * Source code cleanups: use const where appropriate, eliminate casts,
16792 use `void *` instead of `char *` in lhash.
16796 * Bugfix: ssl3_send_server_key_exchange was not restartable
16797 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16798 this the server could overwrite ephemeral keys that the client
16803 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16804 using 50 iterations of the Rabin-Miller test.
16806 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16807 iterations of the Rabin-Miller test as required by the appendix
16808 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16809 As BN_is_prime_fasttest includes trial division, DSA parameter
16810 generation becomes much faster.
16812 This implies a change for the callback functions in DSA_is_prime
16813 and DSA_generate_parameters: The callback function is called once
16814 for each positive witness in the Rabin-Miller test, not just
16815 occasionally in the inner loop; and the parameters to the
16816 callback function now provide an iteration count for the outer
16817 loop rather than for the current invocation of the inner loop.
16818 DSA_generate_parameters additionally can call the callback
16819 function with an 'iteration count' of -1, meaning that a
16820 candidate has passed the trial division test (when q is generated
16821 from an application-provided seed, trial division is skipped).
16825 * New function BN_is_prime_fasttest that optionally does trial
16826 division before starting the Rabin-Miller test and has
16827 an additional BN_CTX * argument (whereas BN_is_prime always
16828 has to allocate at least one BN_CTX).
16829 'callback(1, -1, cb_arg)' is called when a number has passed the
16830 trial division stage.
16834 * Fix for bug in CRL encoding. The validity dates weren't being handled
16839 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16843 * New function BN_pseudo_rand().
16847 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16848 bignum version of BN_from_montgomery() with the working code from
16849 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16854 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16855 made it impossible to use the same SSL_SESSION data structure in
16856 SSL2 clients in multiple threads.
16860 * The return value of RAND_load_file() no longer counts bytes obtained
16861 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16862 to seed the PRNG (previously an explicit byte count was required).
16864 *Ulf Möller, Bodo Möller*
16866 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16867 used `char *` instead of `void *` and had casts all over the place.
16871 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16875 * Retain source code compatibility for BN_prime_checks macro:
16876 BN_is_prime(..., BN_prime_checks, ...) now uses
16877 BN_prime_checks_for_size to determine the appropriate number of
16878 Rabin-Miller iterations.
16882 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16883 DH_CHECK_P_NOT_SAFE_PRIME.
16884 (Check if this is true? OpenPGP calls them "strong".)
16888 * Merge the functionality of "dh" and "gendh" programs into a new program
16889 "dhparam". The old programs are retained for now but will handle DH keys
16890 (instead of parameters) in future.
16894 * Make the ciphers, s_server and s_client programs check the return values
16895 when a new cipher list is set.
16899 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16900 ciphers. Before when the 56bit ciphers were enabled the sorting was
16903 The syntax for the cipher sorting has been extended to support sorting by
16904 cipher-strength (using the strength_bits hard coded in the tables).
16905 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16907 Fix a bug in the cipher-command parser: when supplying a cipher command
16908 string with an "undefined" symbol (neither command nor alphanumeric
16909 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16910 an error is flagged.
16912 Due to the strength-sorting extension, the code of the
16913 ssl_create_cipher_list() function was completely rearranged. I hope that
16914 the readability was also increased :-)
16916 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16918 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16919 for the first serial number and places 2 in the serial number file. This
16920 avoids problems when the root CA is created with serial number zero and
16921 the first user certificate has the same issuer name and serial number
16926 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16927 the new code. Add documentation for this stuff.
16931 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16932 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16933 structures and behave in an analogous way to the X509v3 functions:
16934 they shouldn't be called directly but wrapper functions should be used
16937 So we also now have some wrapper functions that call the X509at functions
16938 when passed certificate requests. (TO DO: similar things can be done with
16939 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16940 things. Some of these need some d2i or i2d and print functionality
16941 because they handle more complex structures.)
16945 * Add missing #ifndefs that caused missing symbols when building libssl
16946 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16947 NO_RSA in `ssl/s2*.c`.
16949 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16951 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16952 has a return value which indicates the quality of the random data
16953 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16954 error queue. New function RAND_pseudo_bytes() generates output that is
16955 guaranteed to be unique but not unpredictable. RAND_add is like
16956 RAND_seed, but takes an extra argument for an entropy estimate
16957 (RAND_seed always assumes full entropy).
16961 * Do more iterations of Rabin-Miller probable prime test (specifically,
16962 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16963 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16964 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16965 false-positive rate of at most 2^-80 for random input.
16969 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16973 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16974 in the 0.9.5 release), this returns the chain
16975 from an X509_CTX structure with a dup of the stack and all
16976 the X509 reference counts upped: so the stack will exist
16977 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16980 Also make SSL_SESSION_print() print out the verify return
16985 * Add manpage for the pkcs12 command. Also change the default
16986 behaviour so MAC iteration counts are used unless the new
16987 -nomaciter option is used. This improves file security and
16988 only older versions of MSIE (4.0 for example) need it.
16992 * Honor the no-xxx Configure options when creating .DEF files.
16996 * Add PKCS#10 attributes to field table: challengePassword,
16997 unstructuredName and unstructuredAddress. These are taken from
16998 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16999 international characters are used.
17001 More changes to X509_ATTRIBUTE code: allow the setting of types
17002 based on strings. Remove the 'loc' parameter when adding
17003 attributes because these will be a SET OF encoding which is sorted
17008 * Initial changes to the 'req' utility to allow request generation
17009 automation. This will allow an application to just generate a template
17010 file containing all the field values and have req construct the
17013 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17014 used all over the place including certificate requests and PKCS#7
17015 structures. They are currently handled manually where necessary with
17016 some primitive wrappers for PKCS#7. The new functions behave in a
17017 manner analogous to the X509 extension functions: they allow
17018 attributes to be looked up by NID and added.
17020 Later something similar to the X509V3 code would be desirable to
17021 automatically handle the encoding, decoding and printing of the
17022 more complex types. The string types like challengePassword can
17023 be handled by the string table functions.
17025 Also modified the multi byte string table handling. Now there is
17026 a 'global mask' which masks out certain types. The table itself
17027 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17028 is useful when for example there is only one permissible type
17029 (as in countryName) and using the mask might result in no valid
17034 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17035 SSL_get_peer_finished to allow applications to obtain the latest
17036 Finished messages sent to the peer or expected from the peer,
17037 respectively. (SSL_get_peer_finished is usually the Finished message
17038 actually received from the peer, otherwise the protocol will be aborted.)
17040 As the Finished message are message digests of the complete handshake
17041 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17042 be used for external authentication procedures when the authentication
17043 provided by SSL/TLS is not desired or is not enough.
17047 * Enhanced support for Alpha Linux is added. Now ./config checks if
17048 the host supports BWX extension and if Compaq C is present on the
17049 $PATH. Just exploiting of the BWX extension results in 20-30%
17050 performance kick for some algorithms, e.g. DES and RC4 to mention
17051 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17056 * Add support for MS "fast SGC". This is arguably a violation of the
17057 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17058 weak crypto and after checking the certificate is SGC a second one
17059 with strong crypto. MS SGC stops the first handshake after receiving
17060 the server certificate message and sends a second client hello. Since
17061 a server will typically do all the time consuming operations before
17062 expecting any further messages from the client (server key exchange
17063 is the most expensive) there is little difference between the two.
17065 To get OpenSSL to support MS SGC we have to permit a second client
17066 hello message after we have sent server done. In addition we have to
17067 reset the MAC if we do get this second client hello.
17071 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17072 if a DER encoded private key is RSA or DSA traditional format. Changed
17073 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17074 format DER encoded private key. Newer code should use PKCS#8 format which
17075 has the key type encoded in the ASN1 structure. Added DER private key
17076 support to pkcs8 application.
17080 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17081 ciphersuites has been selected (as required by the SSL 3/TLS 1
17082 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17083 is set, we interpret this as a request to violate the specification
17084 (the worst that can happen is a handshake failure, and 'correct'
17085 behaviour would result in a handshake failure anyway).
17089 * In SSL_CTX_add_session, take into account that there might be multiple
17090 SSL_SESSION structures with the same session ID (e.g. when two threads
17091 concurrently obtain them from an external cache).
17092 The internal cache can handle only one SSL_SESSION with a given ID,
17093 so if there's a conflict, we now throw out the old one to achieve
17098 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17099 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17100 some routines that use cipher OIDs: some ciphers do not have OIDs
17101 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17106 * Simplify the trust setting structure and code. Now we just have
17107 two sequences of OIDs for trusted and rejected settings. These will
17108 typically have values the same as the extended key usage extension
17109 and any application specific purposes.
17111 The trust checking code now has a default behaviour: it will just
17112 check for an object with the same NID as the passed id. Functions can
17113 be provided to override either the default behaviour or the behaviour
17114 for a given id. SSL client, server and email already have functions
17115 in place for compatibility: they check the NID and also return "trusted"
17116 if the certificate is self signed.
17120 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17121 traditional format into an EVP_PKEY structure.
17125 * Add a password callback function PEM_cb() which either prompts for
17126 a password if usr_data is NULL or otherwise assumes it is a null
17127 terminated password. Allow passwords to be passed on command line
17128 environment or config files in a few more utilities.
17132 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17133 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17134 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17135 Update documentation.
17139 * Support for ASN1 "NULL" type. This could be handled before by using
17140 ASN1_TYPE but there wasn't any function that would try to read a NULL
17141 and produce an error if it couldn't. For compatibility we also have
17142 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17143 don't allocate anything because they don't need to.
17147 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17150 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17152 * Rebuild of the memory allocation routines used by OpenSSL code and
17153 possibly others as well. The purpose is to make an interface that
17154 provide hooks so anyone can build a separate set of allocation and
17155 deallocation routines to be used by OpenSSL, for example memory
17156 pool implementations, or something else, which was previously hard
17157 since Malloc(), Realloc() and Free() were defined as macros having
17158 the values malloc, realloc and free, respectively (except for Win32
17159 compilations). The same is provided for memory debugging code.
17160 OpenSSL already comes with functionality to find memory leaks, but
17161 this gives people a chance to debug other memory problems.
17163 With these changes, a new set of functions and macros have appeared:
17165 CRYPTO_set_mem_debug_functions() [F]
17166 CRYPTO_get_mem_debug_functions() [F]
17167 CRYPTO_dbg_set_options() [F]
17168 CRYPTO_dbg_get_options() [F]
17169 CRYPTO_malloc_debug_init() [M]
17171 The memory debug functions are NULL by default, unless the library
17172 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17173 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17174 gives the standard debugging functions that come with OpenSSL) or
17175 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17176 provided by the library user) must be used. When the standard
17177 debugging functions are used, CRYPTO_dbg_set_options can be used to
17178 request additional information:
17179 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17180 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17182 Also, things like CRYPTO_set_mem_functions will always give the
17183 expected result (the new set of functions is used for allocation
17184 and deallocation) at all times, regardless of platform and compiler
17187 To finish it up, some functions that were never use in any other
17188 way than through macros have a new API and new semantic:
17190 CRYPTO_dbg_malloc()
17191 CRYPTO_dbg_realloc()
17194 All macros of value have retained their old syntax.
17196 *Richard Levitte and Bodo Moeller*
17198 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17199 ordering of SMIMECapabilities wasn't in "strength order" and there
17200 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17205 * Some ASN1 types with illegal zero length encoding (INTEGER,
17206 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17208 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17210 * Merge in my S/MIME library for OpenSSL. This provides a simple
17211 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17212 functionality to handle multipart/signed properly) and a utility
17213 called 'smime' to call all this stuff. This is based on code I
17214 originally wrote for Celo who have kindly allowed it to be
17215 included in OpenSSL.
17219 * Add variants des_set_key_checked and des_set_key_unchecked of
17220 des_set_key (aka des_key_sched). Global variable des_check_key
17221 decides which of these is called by des_set_key; this way
17222 des_check_key behaves as it always did, but applications and
17223 the library itself, which was buggy for des_check_key == 1,
17224 have a cleaner way to pick the version they need.
17228 * New function PKCS12_newpass() which changes the password of a
17233 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17234 dynamic mix. In both cases the ids can be used as an index into the
17235 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17236 functions so they accept a list of the field values and the
17237 application doesn't need to directly manipulate the X509_TRUST
17242 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17247 * Modify the way the V3 extension code looks up extensions. This now
17248 works in a similar way to the object code: we have some "standard"
17249 extensions in a static table which is searched with OBJ_bsearch()
17250 and the application can add dynamic ones if needed. The file
17251 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17252 updated whenever a new extension is added to the core code and kept
17253 in ext_nid order. There is a simple program 'tabtest.c' which checks
17254 this. New extensions are not added too often so this file can readily
17255 be maintained manually.
17257 There are two big advantages in doing things this way. The extensions
17258 can be looked up immediately and no longer need to be "added" using
17259 X509V3_add_standard_extensions(): this function now does nothing.
17260 Side note: I get *lots* of email saying the extension code doesn't
17261 work because people forget to call this function.
17262 Also no dynamic allocation is done unless new extensions are added:
17263 so if we don't add custom extensions there is no need to call
17264 X509V3_EXT_cleanup().
17268 * Modify enc utility's salting as follows: make salting the default. Add a
17269 magic header, so unsalted files fail gracefully instead of just decrypting
17270 to garbage. This is because not salting is a big security hole, so people
17271 should be discouraged from doing it.
17275 * Fixes and enhancements to the 'x509' utility. It allowed a message
17276 digest to be passed on the command line but it only used this
17277 parameter when signing a certificate. Modified so all relevant
17278 operations are affected by the digest parameter including the
17279 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17280 DSA key was used because it didn't fix the digest.
17284 * Initial certificate chain verify code. Currently tests the untrusted
17285 certificates for consistency with the verify purpose (which is set
17286 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17288 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17289 this is because it will reject chains with invalid extensions whereas
17290 every previous version of OpenSSL and SSLeay made no checks at all.
17292 Trust code: checks the root CA for the relevant trust settings. Trust
17293 settings have an initial value consistent with the verify purpose: e.g.
17294 if the verify purpose is for SSL client use it expects the CA to be
17295 trusted for SSL client use. However the default value can be changed to
17296 permit custom trust settings: one example of this would be to only trust
17297 certificates from a specific "secure" set of CAs.
17299 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17300 which should be used for version portability: especially since the
17301 verify structure is likely to change more often now.
17303 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17304 to set them. If not set then assume SSL clients will verify SSL servers
17307 Two new options to the verify program: -untrusted allows a set of
17308 untrusted certificates to be passed in and -purpose which sets the
17309 intended purpose of the certificate. If a purpose is set then the
17310 new chain verify code is used to check extension consistency.
17314 * Support for the authority information access extension.
17318 * Modify RSA and DSA PEM read routines to transparently handle
17319 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17320 public keys in a format compatible with certificate
17321 SubjectPublicKeyInfo structures. Unfortunately there were already
17322 functions called *_PublicKey_* which used various odd formats so
17323 these are retained for compatibility: however the DSA variants were
17324 never in a public release so they have been deleted. Changed dsa/rsa
17325 utilities to handle the new format: note no releases ever handled public
17326 keys so we should be OK.
17328 The primary motivation for this change is to avoid the same fiasco
17329 that dogs private keys: there are several incompatible private key
17330 formats some of which are standard and some OpenSSL specific and
17331 require various evil hacks to allow partial transparent handling and
17332 even then it doesn't work with DER formats. Given the option anything
17333 other than PKCS#8 should be dumped: but the other formats have to
17334 stay in the name of compatibility.
17336 With public keys and the benefit of hindsight one standard format
17337 is used which works with EVP_PKEY, RSA or DSA structures: though
17338 it clearly returns an error if you try to read the wrong kind of key.
17340 Added a -pubkey option to the 'x509' utility to output the public key.
17341 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17342 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17343 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17344 that do the same as the `EVP_PKEY_assign_*()` except they up the
17345 reference count of the added key (they don't "swallow" the
17350 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17351 CRLs would fail if the file contained no certificates or no CRLs:
17352 added a new function to read in both types and return the number
17353 read: this means that if none are read it will be an error. The
17354 DER versions of the certificate and CRL reader would always fail
17355 because it isn't possible to mix certificates and CRLs in DER format
17356 without choking one or the other routine. Changed this to just read
17357 a certificate: this is the best we can do. Also modified the code
17358 in `apps/verify.c` to take notice of return codes: it was previously
17359 attempting to read in certificates from NULL pointers and ignoring
17360 any errors: this is one reason why the cert and CRL reader seemed
17361 to work. It doesn't check return codes from the default certificate
17362 routines: these may well fail if the certificates aren't installed.
17366 * Code to support otherName option in GeneralName.
17370 * First update to verify code. Change the verify utility
17371 so it warns if it is passed a self signed certificate:
17372 for consistency with the normal behaviour. X509_verify
17373 has been modified to it will now verify a self signed
17374 certificate if *exactly* the same certificate appears
17375 in the store: it was previously impossible to trust a
17376 single self signed certificate. This means that:
17377 openssl verify ss.pem
17378 now gives a warning about a self signed certificate but
17379 openssl verify -CAfile ss.pem ss.pem
17384 * For servers, store verify_result in SSL_SESSION data structure
17385 (and add it to external session representation).
17386 This is needed when client certificate verifications fails,
17387 but an application-provided verification callback (set by
17388 SSL_CTX_set_cert_verify_callback) allows accepting the session
17389 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17390 but returns 1): When the session is reused, we have to set
17391 ssl->verify_result to the appropriate error code to avoid
17394 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17396 * Fix a bug in the new PKCS#7 code: it didn't consider the
17397 case in PKCS7_dataInit() where the signed PKCS7 structure
17398 didn't contain any existing data because it was being created.
17400 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17402 * Add a salt to the key derivation routines in enc.c. This
17403 forms the first 8 bytes of the encrypted file. Also add a
17404 -S option to allow a salt to be input on the command line.
17408 * New function X509_cmp(). Oddly enough there wasn't a function
17409 to compare two certificates. We do this by working out the SHA1
17410 hash and comparing that. X509_cmp() will be needed by the trust
17415 * SSL_get1_session() is like SSL_get_session(), but increments
17416 the reference count in the SSL_SESSION returned.
17418 *Geoff Thorpe <geoff@eu.c2.net>*
17420 * Fix for 'req': it was adding a null to request attributes.
17421 Also change the X509_LOOKUP and X509_INFO code to handle
17422 certificate auxiliary information.
17426 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17431 * Add the possibility to add extra information to the memory leak
17432 detecting output, to form tracebacks, showing from where each
17433 allocation was originated: CRYPTO_push_info("constant string") adds
17434 the string plus current file name and line number to a per-thread
17435 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17436 is like calling CYRPTO_pop_info() until the stack is empty.
17437 Also updated memory leak detection code to be multi-thread-safe.
17441 * Add options -text and -noout to pkcs7 utility and delete the
17442 encryption options which never did anything. Update docs.
17446 * Add options to some of the utilities to allow the pass phrase
17447 to be included on either the command line (not recommended on
17448 OSes like Unix) or read from the environment. Update the
17449 manpages and fix a few bugs.
17453 * Add a few manpages for some of the openssl commands.
17457 * Fix the -revoke option in ca. It was freeing up memory twice,
17458 leaking and not finding already revoked certificates.
17462 * Extensive changes to support certificate auxiliary information.
17463 This involves the use of X509_CERT_AUX structure and X509_AUX
17464 functions. An X509_AUX function such as PEM_read_X509_AUX()
17465 can still read in a certificate file in the usual way but it
17466 will also read in any additional "auxiliary information". By
17467 doing things this way a fair degree of compatibility can be
17468 retained: existing certificates can have this information added
17469 using the new 'x509' options.
17471 Current auxiliary information includes an "alias" and some trust
17472 settings. The trust settings will ultimately be used in enhanced
17473 certificate chain verification routines: currently a certificate
17474 can only be trusted if it is self signed and then it is trusted
17479 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17480 The problem was that one of the replacement routines had not been working
17481 since SSLeay releases. For now the offending routine has been replaced
17482 with non-optimised assembler. Even so, this now gives around 95%
17483 performance improvement for 1024 bit RSA signs.
17487 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17488 handling. Most clients have the effective key size in bits equal to
17489 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17490 A few however don't do this and instead use the size of the decrypted key
17491 to determine the RC2 key length and the AlgorithmIdentifier to determine
17492 the effective key length. In this case the effective key length can still
17493 be 40 bits but the key length can be 168 bits for example. This is fixed
17494 by manually forcing an RC2 key into the EVP_PKEY structure because the
17495 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17496 the key length and effective key length are equal.
17500 * Add a bunch of functions that should simplify the creation of
17501 X509_NAME structures. Now you should be able to do:
17502 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17503 and have it automatically work out the correct field type and fill in
17504 the structures. The more adventurous can try:
17505 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17506 and it will (hopefully) work out the correct multibyte encoding.
17510 * Change the 'req' utility to use the new field handling and multibyte
17511 copy routines. Before the DN field creation was handled in an ad hoc
17512 way in req, ca, and x509 which was rather broken and didn't support
17513 BMPStrings or UTF8Strings. Since some software doesn't implement
17514 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17515 using the dirstring_type option. See the new comment in the default
17516 openssl.cnf for more info.
17520 * Make crypto/rand/md_rand.c more robust:
17521 - Assure unique random numbers after fork().
17522 - Make sure that concurrent threads access the global counter and
17523 md serializably so that we never lose entropy in them
17524 or use exactly the same state in multiple threads.
17525 Access to the large state is not always serializable because
17526 the additional locking could be a performance killer, and
17527 md should be large enough anyway.
17531 * New file `apps/app_rand.c` with commonly needed functionality
17532 for handling the random seed file.
17534 Use the random seed file in some applications that previously did not:
17536 dsaparam -genkey (which also ignored its '-rand' option),
17539 x509 (when signing).
17540 Except on systems with /dev/urandom, it is crucial to have a random
17541 seed file at least for key creation, DSA signing, and for DH exchanges;
17542 for RSA signatures we could do without one.
17544 gendh and gendsa (unlike genrsa) used to read only the first byte
17545 of each file listed in the '-rand' option. The function as previously
17546 found in genrsa is now in app_rand.c and is used by all programs
17547 that support '-rand'.
17551 * In RAND_write_file, use mode 0600 for creating files;
17552 don't just chmod when it may be too late.
17556 * Report an error from X509_STORE_load_locations
17557 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17561 * New function ASN1_mbstring_copy() this copies a string in either
17562 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17563 into an ASN1_STRING type. A mask of permissible types is passed
17564 and it chooses the "minimal" type to use or an error if not type
17569 * Add function equivalents to the various macros in asn1.h. The old
17570 macros are retained with an `M_` prefix. Code inside the library can
17571 use the `M_` macros. External code (including the openssl utility)
17572 should *NOT* in order to be "shared library friendly".
17576 * Add various functions that can check a certificate's extensions
17577 to see if it usable for various purposes such as SSL client,
17578 server or S/MIME and CAs of these types. This is currently
17579 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17580 verification. Also added a -purpose flag to x509 utility to
17581 print out all the purposes.
17585 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17590 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17591 for, obtain and decode and extension and obtain its critical flag.
17592 This allows all the necessary extension code to be handled in a
17593 single function call.
17597 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17598 platforms. See crypto/rc4/rc4_enc.c for further details.
17602 * New -noout option to asn1parse. This causes no output to be produced
17603 its main use is when combined with -strparse and -out to extract data
17604 from a file (which may not be in ASN.1 format).
17608 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17609 when producing the local key id.
17611 *Richard Levitte <levitte@stacken.kth.se>*
17613 * New option -dhparam in s_server. This allows a DH parameter file to be
17614 stated explicitly. If it is not stated then it tries the first server
17615 certificate file. The previous behaviour hard coded the filename
17620 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17621 a public key to be input or output. For example:
17622 openssl rsa -in key.pem -pubout -out pubkey.pem
17623 Also added necessary DSA public key functions to handle this.
17627 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17628 in the message. This was handled by allowing
17629 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17631 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17633 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17634 to the end of the strings whereas this didn't. This would cause problems
17635 if strings read with d2i_ASN1_bytes() were later modified.
17637 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17639 * Fix for base64 decode bug. When a base64 bio reads only one line of
17640 data and it contains EOF it will end up returning an error. This is
17641 caused by input 46 bytes long. The cause is due to the way base64
17642 BIOs find the start of base64 encoded data. They do this by trying a
17643 trial decode on each line until they find one that works. When they
17644 do a flag is set and it starts again knowing it can pass all the
17645 data directly through the decoder. Unfortunately it doesn't reset
17646 the context it uses. This means that if EOF is reached an attempt
17647 is made to pass two EOFs through the context and this causes the
17648 resulting error. This can also cause other problems as well. As is
17649 usual with these problems it takes *ages* to find and the fix is
17650 trivial: move one line.
17652 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17654 * Ugly workaround to get s_client and s_server working under Windows. The
17655 old code wouldn't work because it needed to select() on sockets and the
17656 tty (for keypresses and to see if data could be written). Win32 only
17657 supports select() on sockets so we select() with a 1s timeout on the
17658 sockets and then see if any characters are waiting to be read, if none
17659 are present then we retry, we also assume we can always write data to
17660 the tty. This isn't nice because the code then blocks until we've
17661 received a complete line of data and it is effectively polling the
17662 keyboard at 1s intervals: however it's quite a bit better than not
17663 working at all :-) A dedicated Windows application might handle this
17664 with an event loop for example.
17668 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17669 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17670 will be called when RSA_sign() and RSA_verify() are used. This is useful
17671 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17672 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17673 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17674 This necessitated the support of an extra signature type NID_md5_sha1
17675 for SSL signatures and modifications to the SSL library to use it instead
17676 of calling RSA_public_decrypt() and RSA_private_encrypt().
17680 * Add new -verify -CAfile and -CApath options to the crl program, these
17681 will lookup a CRL issuers certificate and verify the signature in a
17682 similar way to the verify program. Tidy up the crl program so it
17683 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17684 less strict. It will now permit CRL extensions even if it is not
17685 a V2 CRL: this will allow it to tolerate some broken CRLs.
17689 * Initialize all non-automatic variables each time one of the openssl
17690 sub-programs is started (this is necessary as they may be started
17691 multiple times from the "OpenSSL>" prompt).
17693 *Lennart Bang, Bodo Moeller*
17695 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17696 removing all other RSA functionality (this is what NO_RSA does). This
17697 is so (for example) those in the US can disable those operations covered
17698 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17703 * Non-copying interface to BIO pairs.
17704 (still largely untested)
17708 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17709 ASCII string. This was handled independently in various places before.
17713 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17714 UTF8 strings a character at a time.
17718 * Use client_version from client hello to select the protocol
17719 (s23_srvr.c) and for RSA client key exchange verification
17720 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17724 * Add various utility functions to handle SPKACs, these were previously
17725 handled by poking round in the structure internals. Added new function
17726 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17727 print, verify and generate SPKACs. Based on an original idea from
17728 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17732 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17736 * Allow the config file extension section to be overwritten on the
17737 command line. Based on an original idea from Massimiliano Pala
17738 <madwolf@comune.modena.it>. The new option is called -extensions
17739 and can be applied to ca, req and x509. Also -reqexts to override
17740 the request extensions in req and -crlexts to override the crl extensions
17745 * Add new feature to the SPKAC handling in ca. Now you can include
17746 the same field multiple times by preceding it by "XXXX." for example:
17749 this is the same syntax as used in the req config file.
17753 * Allow certificate extensions to be added to certificate requests. These
17754 are specified in a 'req_extensions' option of the req section of the
17755 config file. They can be printed out with the -text option to req but
17756 are otherwise ignored at present.
17760 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17761 data read consists of only the final block it would not decrypted because
17762 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17763 A misplaced 'break' also meant the decrypted final block might not be
17764 copied until the next read.
17768 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17769 a few extra parameters to the DH structure: these will be useful if
17770 for example we want the value of 'q' or implement X9.42 DH.
17774 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17775 provides hooks that allow the default DSA functions or functions on a
17776 "per key" basis to be replaced. This allows hardware acceleration and
17777 hardware key storage to be handled without major modification to the
17778 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17779 associated functions.
17783 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17784 as "read only": it can't be written to and the buffer it points to will
17785 not be freed. Reading from a read only BIO is much more efficient than
17786 a normal memory BIO. This was added because there are several times when
17787 an area of memory needs to be read from a BIO. The previous method was
17788 to create a memory BIO and write the data to it, this results in two
17789 copies of the data and an O(n^2) reading algorithm. There is a new
17790 function BIO_new_mem_buf() which creates a read only memory BIO from
17791 an area of memory. Also modified the PKCS#7 routines to use read only
17796 * Bugfix: ssl23_get_client_hello did not work properly when called in
17797 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17798 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17799 but a retry condition occurred while trying to read the rest.
17803 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17804 NID_pkcs7_encrypted by default: this was wrong since this should almost
17805 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17806 the encrypted data type: this is a more sensible place to put it and it
17807 allows the PKCS#12 code to be tidied up that duplicated this
17812 * Changed obj_dat.pl script so it takes its input and output files on
17813 the command line. This should avoid shell escape redirection problems
17818 * Initial support for certificate extension requests, these are included
17819 in things like Xenroll certificate requests. Included functions to allow
17820 extensions to be obtained and added.
17824 * -crlf option to s_client and s_server for sending newlines as
17825 CRLF (as required by many protocols).
17829 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17831 * Install libRSAglue.a when OpenSSL is built with RSAref.
17833 *Ralf S. Engelschall*
17835 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17837 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17839 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17844 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17845 DH parameters/keys (q is lost during that conversion, but the resulting
17846 DH parameters contain its length).
17848 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17849 much faster than DH_generate_parameters (which creates parameters
17850 where `p = 2*q + 1`), and also the smaller q makes DH computations
17851 much more efficient (160-bit exponentiation instead of 1024-bit
17852 exponentiation); so this provides a convenient way to support DHE
17853 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17854 utter importance to use
17855 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17857 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17858 when such DH parameters are used, because otherwise small subgroup
17859 attacks may become possible!
17863 * Avoid memory leak in i2d_DHparams.
17867 * Allow the -k option to be used more than once in the enc program:
17868 this allows the same encrypted message to be read by multiple recipients.
17872 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17873 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17874 it will always use the numerical form of the OID, even if it has a short
17879 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17880 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17881 otherwise bn_mod_exp was called. In the case of hardware keys for example
17882 no private key components need be present and it might store extra data
17883 in the RSA structure, which cannot be accessed from bn_mod_exp.
17884 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17885 private key operations.
17889 * Added support for SPARC Linux.
17893 * pem_password_cb function type incompatibly changed from
17894 typedef int pem_password_cb(char *buf, int size, int rwflag);
17896 ....(char *buf, int size, int rwflag, void *userdata);
17897 so that applications can pass data to their callbacks:
17898 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17899 additional void * argument, which is just handed through whenever
17900 the password callback is called.
17902 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17904 New function SSL_CTX_set_default_passwd_cb_userdata.
17906 Compatibility note: As many C implementations push function arguments
17907 onto the stack in reverse order, the new library version is likely to
17908 interoperate with programs that have been compiled with the old
17909 pem_password_cb definition (PEM_whatever takes some data that
17910 happens to be on the stack as its last argument, and the callback
17911 just ignores this garbage); but there is no guarantee whatsoever that
17914 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17915 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17916 problems not only on Windows, but also on some Unix platforms.
17917 To avoid problematic command lines, these definitions are now in an
17918 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17919 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17923 * MIPS III/IV assembler module is reimplemented.
17927 * More DES library cleanups: remove references to srand/rand and
17928 delete an unused file.
17932 * Add support for the free Netwide assembler (NASM) under Win32,
17933 since not many people have MASM (ml) and it can be hard to obtain.
17934 This is currently experimental but it seems to work OK and pass all
17935 the tests. Check out INSTALL.W32 for info.
17939 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17940 without temporary keys kept an extra copy of the server key,
17941 and connections with temporary keys did not free everything in case
17946 * New function RSA_check_key and new openssl rsa option -check
17947 for verifying the consistency of RSA keys.
17949 *Ulf Moeller, Bodo Moeller*
17951 * Various changes to make Win32 compile work:
17952 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17953 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17954 comparison" warnings.
17955 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17959 * Add a debugging option to PKCS#5 v2 key generation function: when
17960 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17961 derived keys are printed to stderr.
17965 * Copy the flags in ASN1_STRING_dup().
17967 *Roman E. Pavlov <pre@mo.msk.ru>*
17969 * The x509 application mishandled signing requests containing DSA
17970 keys when the signing key was also DSA and the parameters didn't match.
17972 It was supposed to omit the parameters when they matched the signing key:
17973 the verifying software was then supposed to automatically use the CA's
17974 parameters if they were absent from the end user certificate.
17976 Omitting parameters is no longer recommended. The test was also
17977 the wrong way round! This was probably due to unusual behaviour in
17978 EVP_cmp_parameters() which returns 1 if the parameters match.
17979 This meant that parameters were omitted when they *didn't* match and
17980 the certificate was useless. Certificates signed with 'ca' didn't have
17983 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17985 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17986 The interface is as follows:
17987 Applications can use
17988 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17989 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17990 "off" is now the default.
17991 The library internally uses
17992 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17993 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17994 to disable memory-checking temporarily.
17996 Some inconsistent states that previously were possible (and were
17997 even the default) are now avoided.
17999 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18000 with each memory chunk allocated; this is occasionally more helpful
18001 than just having a counter.
18003 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18005 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18010 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18011 which largely parallels "options", but is for changing API behaviour,
18012 whereas "options" are about protocol behaviour.
18013 Initial "mode" flags are:
18015 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18016 a single record has been written.
18017 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18018 retries use the same buffer location.
18019 (But all of the contents must be
18024 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18027 * Fix problems with no-hmac etc.
18029 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18031 * New functions RSA_get_default_method(), RSA_set_method() and
18032 RSA_get_method(). These allows replacement of RSA_METHODs without having
18033 to mess around with the internals of an RSA structure.
18037 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18038 Also really enable memory leak checks in openssl.c and in some
18041 *Chad C. Mulligan, Bodo Moeller*
18043 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18044 up the length of negative integers. This has now been simplified to just
18045 store the length when it is first determined and use it later, rather
18046 than trying to keep track of where data is copied and updating it to
18048 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18050 * Add a new function PKCS7_signatureVerify. This allows the verification
18051 of a PKCS#7 signature but with the signing certificate passed to the
18052 function itself. This contrasts with PKCS7_dataVerify which assumes the
18053 certificate is present in the PKCS#7 structure. This isn't always the
18054 case: certificates can be omitted from a PKCS#7 structure and be
18055 distributed by "out of band" means (such as a certificate database).
18059 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18060 function prototypes in pem.h, also change util/mkdef.pl to add the
18061 necessary function names.
18065 * mk1mf.pl (used by Windows builds) did not properly read the
18066 options set by Configure in the top level Makefile, and Configure
18067 was not even able to write more than one option correctly.
18068 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18072 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18073 file to be loaded from a BIO or FILE pointer. The BIO version will
18074 for example allow memory BIOs to contain config info.
18078 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18079 Whoever hopes to achieve shared-library compatibility across versions
18080 must use this, not the compile-time macro.
18081 (Exercise 0.9.4: Which is the minimum library version required by
18083 Note: All this applies only to multi-threaded programs, others don't
18088 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18089 through a BIO pair triggered the default case, i.e.
18090 SSLerr(...,SSL_R_UNKNOWN_STATE).
18094 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18095 can use the SSL library even if none of the specific BIOs is
18100 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18101 for the encoded length.
18103 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18105 * Add initial documentation of the X509V3 functions.
18109 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18110 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18111 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18112 secure PKCS#8 private key format with a high iteration count.
18116 * Fix determination of Perl interpreter: A perl or perl5
18117 *directory* in $PATH was also accepted as the interpreter.
18119 *Ralf S. Engelschall*
18121 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18122 wrong with it but it was very old and did things like calling
18123 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18124 unusual formatting.
18128 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18129 to use the new extension code.
18133 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18134 with macros. This should make it easier to change their form, add extra
18135 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18140 * Add to configuration table a new entry that can specify an alternative
18141 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18142 according to Mark Crispin <MRC@Panda.COM>.
18146 * DES CBC did not update the IV. Weird.
18150 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18151 Changing the behaviour of the former might break existing programs --
18152 where IV updating is needed, des_ncbc_encrypt can be used.
18155 * When bntest is run from "make test" it drives bc to check its
18156 calculations, as well as internally checking them. If an internal check
18157 fails, it needs to cause bc to give a non-zero result or make test carries
18158 on without noticing the failure. Fixed.
18162 * DES library cleanups.
18166 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18167 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18168 ciphers. NOTE: although the key derivation function has been verified
18169 against some published test vectors it has not been extensively tested
18170 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18175 * Instead of "mkdir -p", which is not fully portable, use new
18176 Perl script "util/mkdir-p.pl".
18180 * Rewrite the way password based encryption (PBE) is handled. It used to
18181 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18182 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18183 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18184 the 'parameter' field of the AlgorithmIdentifier is passed to the
18185 underlying key generation function so it must do its own ASN1 parsing.
18186 This has also changed the EVP_PBE_CipherInit() function which now has a
18187 'parameter' argument instead of literal salt and iteration count values
18188 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18192 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18193 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18194 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18195 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18196 value was just used as a "magic string" and not used directly its
18197 value doesn't matter.
18201 * Introduce some semblance of const correctness to BN. Shame C doesn't
18206 * "linux-sparc64" configuration (ultrapenguin).
18208 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18209 "linux-sparc" configuration.
18211 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18213 * config now generates no-xxx options for missing ciphers.
18217 * Support the EBCDIC character set (work in progress).
18218 File ebcdic.c not yet included because it has a different license.
18220 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18222 * Support BS2000/OSD-POSIX.
18224 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18226 * Make callbacks for key generation use `void *` instead of `char *`.
18230 * Make S/MIME samples compile (not yet tested).
18234 * Additional typesafe stacks.
18238 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18242 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18244 * New configuration variant "sco5-gcc".
18246 * Updated some demos.
18248 *Sean O Riordain, Wade Scholine*
18250 * Add missing BIO_free at exit of pkcs12 application.
18254 * Fix memory leak in conf.c.
18258 * Updates for Win32 to assembler version of MD5.
18262 * Set #! path to perl in `apps/der_chop` to where we found it
18263 instead of using a fixed path.
18267 * SHA library changes for irix64-mips4-cc.
18271 * Improvements for VMS support.
18275 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18277 * Bignum library bug fix. IRIX 6 passes "make test" now!
18278 This also avoids the problems with SC4.2 and unpatched SC5.
18280 *Andy Polyakov <appro@fy.chalmers.se>*
18282 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18283 These are required because of the typesafe stack would otherwise break
18284 existing code. If old code used a structure member which used to be STACK
18285 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18286 sk_num or sk_value it would produce an error because the num, data members
18287 are not present in STACK_OF. Now it just produces a warning. sk_set
18288 replaces the old method of assigning a value to sk_value
18289 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18290 that does this will no longer work (and should use sk_set instead) but
18291 this could be regarded as a "questionable" behaviour anyway.
18295 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18296 correctly handle encrypted S/MIME data.
18300 * Change type of various DES function arguments from des_cblock
18301 (which means, in function argument declarations, pointer to char)
18302 to des_cblock * (meaning pointer to array with 8 char elements),
18303 which allows the compiler to do more typechecking; it was like
18304 that back in SSLeay, but with lots of ugly casts.
18306 Introduce new type const_des_cblock.
18310 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18311 problems: find RecipientInfo structure that matches recipient certificate
18312 and initialise the ASN1 structures properly based on passed cipher.
18316 * Belatedly make the BN tests actually check the results.
18320 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18321 to and from BNs: it was completely broken. New compilation option
18322 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18323 key elements as negative integers.
18327 * Reorganize and speed up MD5.
18329 *Andy Polyakov <appro@fy.chalmers.se>*
18333 *Richard Levitte <richard@levitte.org>*
18335 * New option -out to asn1parse to allow the parsed structure to be
18336 output to a file. This is most useful when combined with the -strparse
18337 option to examine the output of things like OCTET STRINGS.
18341 * Make SSL library a little more fool-proof by not requiring any longer
18342 that `SSL_set_{accept,connect}_state` be called before
18343 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18344 in many applications because usually everything *appeared* to work as
18345 intended anyway -- now it really works as intended).
18349 * Move openssl.cnf out of lib/.
18353 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18354 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18355 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18357 *Ralf S. Engelschall*
18359 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18360 handle PKCS#7 enveloped data properly.
18362 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18364 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18365 copying pointers. The cert_st handling is changed by this in
18366 various ways (and thus what used to be known as ctx->default_cert
18367 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18368 any longer when s->cert does not give us what we need).
18369 ssl_cert_instantiate becomes obsolete by this change.
18370 As soon as we've got the new code right (possibly it already is?),
18371 we have solved a couple of bugs of the earlier code where s->cert
18372 was used as if it could not have been shared with other SSL structures.
18374 Note that using the SSL API in certain dirty ways now will result
18375 in different behaviour than observed with earlier library versions:
18376 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18377 does not influence s as it used to.
18379 In order to clean up things more thoroughly, inside SSL_SESSION
18380 we don't use CERT any longer, but a new structure SESS_CERT
18381 that holds per-session data (if available); currently, this is
18382 the peer's certificate chain and, for clients, the server's certificate
18383 and temporary key. CERT holds only those values that can have
18384 meaningful defaults in an SSL_CTX.
18388 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18389 from the internal representation. Various PKCS#7 fixes: remove some
18390 evil casts and set the enc_dig_alg field properly based on the signing
18395 * Allow PKCS#12 password to be set from the command line or the
18396 environment. Let 'ca' get its config file name from the environment
18397 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18402 * Allow certificate policies extension to use an IA5STRING for the
18403 organization field. This is contrary to the PKIX definition but
18404 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18409 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18410 without disallowing inline assembler and the like for non-pedantic builds.
18414 * Support Borland C++ builder.
18416 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18422 * SHA-1 cleanups and performance enhancements.
18424 *Andy Polyakov <appro@fy.chalmers.se>*
18426 * Sparc v8plus assembler for the bignum library.
18428 *Andy Polyakov <appro@fy.chalmers.se>*
18430 * Accept any -xxx and +xxx compiler options in Configure.
18434 * Update HPUX configuration.
18438 * Add missing `sk_<type>_unshift()` function to safestack.h
18440 *Ralf S. Engelschall*
18442 * New function SSL_CTX_use_certificate_chain_file that sets the
18443 "extra_cert"s in addition to the certificate. (This makes sense
18444 only for "PEM" format files, as chains as a whole are not
18449 * Support verify_depth from the SSL API.
18450 x509_vfy.c had what can be considered an off-by-one-error:
18451 Its depth (which was not part of the external interface)
18452 was actually counting the number of certificates in a chain;
18453 now it really counts the depth.
18457 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18458 instead of X509err, which often resulted in confusing error
18459 messages since the error codes are not globally unique
18460 (e.g. an alleged error in ssl3_accept when a certificate
18461 didn't match the private key).
18463 * New function SSL_CTX_set_session_id_context that allows to set a default
18464 value (so that you don't need SSL_set_session_id_context for each
18465 connection using the SSL_CTX).
18469 * OAEP decoding bug fix.
18473 * Support INSTALL_PREFIX for package builders, as proposed by
18478 * New Configure options "threads" and "no-threads". For systems
18479 where the proper compiler options are known (currently Solaris
18480 and Linux), "threads" is the default.
18484 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18488 * Install various scripts to $(OPENSSLDIR)/misc, not to
18489 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18490 such as /usr/local/bin.
18494 * "make linux-shared" to build shared libraries.
18496 *Niels Poppe <niels@netbox.org>*
18498 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18502 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18503 extension adding in x509 utility.
18507 * Remove NOPROTO sections and error code comments.
18511 * Partial rewrite of the DEF file generator to now parse the ANSI
18516 * New Configure options --prefix=DIR and --openssldir=DIR.
18520 * Complete rewrite of the error code script(s). It is all now handled
18521 by one script at the top level which handles error code gathering,
18522 header rewriting and C source file generation. It should be much better
18523 than the old method: it now uses a modified version of Ulf's parser to
18524 read the ANSI prototypes in all header files (thus the old K&R definitions
18525 aren't needed for error creation any more) and do a better job of
18526 translating function codes into names. The old 'ASN1 error code embedded
18527 in a comment' is no longer necessary and it doesn't use .err files which
18528 have now been deleted. Also the error code call doesn't have to appear all
18529 on one line (which resulted in some large lines...).
18533 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18537 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18538 0 (which usually indicates a closed connection), but continue reading.
18542 * Fix some race conditions.
18546 * Add support for CRL distribution points extension. Add Certificate
18547 Policies and CRL distribution points documentation.
18551 * Move the autogenerated header file parts to crypto/opensslconf.h.
18555 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18556 8 of keying material. Merlin has also confirmed interop with this fix
18557 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18559 *Merlin Hughes <merlin@baltimore.ie>*
18561 * Fix lots of warnings.
18563 *Richard Levitte <levitte@stacken.kth.se>*
18565 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18566 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18568 *Richard Levitte <levitte@stacken.kth.se>*
18570 * Fix problems with sizeof(long) == 8.
18572 *Andy Polyakov <appro@fy.chalmers.se>*
18574 * Change functions to ANSI C.
18578 * Fix typos in error codes.
18580 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18582 * Remove defunct assembler files from Configure.
18586 * SPARC v8 assembler BIGNUM implementation.
18588 *Andy Polyakov <appro@fy.chalmers.se>*
18590 * Support for Certificate Policies extension: both print and set.
18591 Various additions to support the r2i method this uses.
18595 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18596 return a const string when you are expecting an allocated buffer.
18600 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18601 types DirectoryString and DisplayText.
18605 * Add code to allow r2i extensions to access the configuration database,
18606 add an LHASH database driver and add several ctx helper functions.
18610 * Fix an evil bug in bn_expand2() which caused various BN functions to
18611 fail when they extended the size of a BIGNUM.
18615 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18616 support typesafe stack.
18620 * Fix typo in SSL_[gs]et_options().
18622 *Nils Frostberg <nils@medcom.se>*
18624 * Delete various functions and files that belonged to the (now obsolete)
18625 old X509V3 handling code.
18629 * New Configure option "rsaref".
18633 * Don't auto-generate pem.h.
18637 * Introduce type-safe ASN.1 SETs.
18641 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18643 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18645 * Introduce type-safe STACKs. This will almost certainly break lots of code
18646 that links with OpenSSL (well at least cause lots of warnings), but fear
18647 not: the conversion is trivial, and it eliminates loads of evil casts. A
18648 few STACKed things have been converted already. Feel free to convert more.
18649 In the fullness of time, I'll do away with the STACK type altogether.
18653 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18654 specified in `<certfile>` by updating the entry in the index.txt file.
18655 This way one no longer has to edit the index.txt file manually for
18656 revoking a certificate. The -revoke option does the gory details now.
18658 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18660 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18661 `-text` option at all and this way the `-noout -text` combination was
18662 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18664 *Ralf S. Engelschall*
18666 * Make sure a corresponding plain text error message exists for the
18667 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18668 verify callback function determined that a certificate was revoked.
18670 *Ralf S. Engelschall*
18672 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18673 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18674 all available ciphers including rc5, which was forgotten until now.
18675 In order to let the testing shell script know which algorithms
18676 are available, a new (up to now undocumented) command
18677 `openssl list-cipher-commands` is used.
18681 * Bugfix: s_client occasionally would sleep in select() when
18682 it should have checked SSL_pending() first.
18686 * New functions DSA_do_sign and DSA_do_verify to provide access to
18687 the raw DSA values prior to ASN.1 encoding.
18691 * Tweaks to Configure
18693 *Niels Poppe <niels@netbox.org>*
18695 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18700 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18704 * New config option to avoid instructions that are illegal on the 80386.
18705 The default code is faster, but requires at least a 486.
18709 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18710 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18711 same as SSL2_VERSION anyway.
18715 * New "-showcerts" option for s_client.
18719 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18720 application. Various cleanups and fixes.
18724 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18725 modify error routines to work internally. Add error codes and PBE init
18726 to library startup routines.
18730 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18731 packing functions to asn1 and evp. Changed function names and error
18732 codes along the way.
18736 * PKCS12 integration: and so it begins... First of several patches to
18737 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18738 objects to objects.h
18742 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18743 and display support for Thawte strong extranet extension.
18747 * Add LinuxPPC support.
18749 *Jeff Dubrule <igor@pobox.org>*
18751 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18752 bn_div_words in alpha.s.
18754 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18756 * Make sure the RSA OAEP test is skipped under -DRSAref because
18757 OAEP isn't supported when OpenSSL is built with RSAref.
18759 *Ulf Moeller <ulf@fitug.de>*
18761 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18762 so they no longer are missing under -DNOPROTO.
18764 *Soren S. Jorvang <soren@t.dk>*
18766 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18768 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18769 doesn't work when the session is reused. Coming soon!
18773 * Fix a security hole, that allows sessions to be reused in the wrong
18774 context thus bypassing client cert protection! All software that uses
18775 client certs and session caches in multiple contexts NEEDS PATCHING to
18776 allow session reuse! A fuller solution is in the works.
18778 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18780 * Some more source tree cleanups (removed obsolete files
18781 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18782 permission on "config" script to be executable) and a fix for the INSTALL
18785 *Ulf Moeller <ulf@fitug.de>*
18787 * Remove some legacy and erroneous uses of malloc, free instead of
18790 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18792 * Make rsa_oaep_test return non-zero on error.
18794 *Ulf Moeller <ulf@fitug.de>*
18796 * Add support for native Solaris shared libraries. Configure
18797 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18798 if someone would make that last step automatic.
18800 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18802 * ctx_size was not built with the right compiler during "make links". Fixed.
18806 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18807 except NULL ciphers". This means the default cipher list will no longer
18808 enable NULL ciphers. They need to be specifically enabled e.g. with
18809 the string "DEFAULT:eNULL".
18813 * Fix to RSA private encryption routines: if p < q then it would
18814 occasionally produce an invalid result. This will only happen with
18815 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18819 * Be less restrictive and allow also `perl util/perlpath.pl
18820 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18821 because this way one can also use an interpreter named `perl5` (which is
18822 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18823 installed as `perl`).
18825 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18827 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18829 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18831 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18832 advapi32.lib to Win32 build and change the pem test comparison
18833 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18834 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18835 and crypto/des/ede_cbcm_enc.c.
18839 * DES quad checksum was broken on big-endian architectures. Fixed.
18843 * Comment out two functions in bio.h that aren't implemented. Fix up the
18844 Win32 test batch file so it (might) work again. The Win32 test batch file
18845 is horrible: I feel ill....
18849 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18850 in e_os.h. Audit of header files to check ANSI and non ANSI
18851 sections: 10 functions were absent from non ANSI section and not exported
18852 from Windows DLLs. Fixed up libeay.num for new functions.
18856 * Make `openssl version` output lines consistent.
18858 *Ralf S. Engelschall*
18860 * Fix Win32 symbol export lists for BIO functions: Added
18861 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18862 to ms/libeay{16,32}.def.
18864 *Ralf S. Engelschall*
18866 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18867 fine under Unix and passes some trivial tests I've now added. But the
18868 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18869 added to make sure no one expects that this stuff really works in the
18870 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18871 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18874 *Ralf S. Engelschall*
18876 * Fix the generation of two part addresses in perl.
18878 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18880 * Add config entry for Linux on MIPS.
18882 *John Tobey <jtobey@channel1.com>*
18884 * Make links whenever Configure is run, unless we are on Windoze.
18888 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18889 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18894 * Add a useful kludge to allow package maintainers to specify compiler and
18895 other platforms details on the command line without having to patch the
18896 Configure script every time: One now can use
18897 `perl Configure <id>:<details>`,
18898 i.e. platform ids are allowed to have details appended
18899 to them (separated by colons). This is treated as there would be a static
18900 pre-configured entry in Configure's %table under key `<id>` with value
18901 `<details>` and `perl Configure <id>` is called. So, when you want to
18902 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18903 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18904 now, which overrides the FreeBSD-elf entry on-the-fly.
18906 *Ralf S. Engelschall*
18908 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18912 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18913 on the `perl Configure ...` command line. This way one can compile
18914 OpenSSL libraries with Position Independent Code (PIC) which is needed
18915 for linking it into DSOs.
18917 *Ralf S. Engelschall*
18919 * Remarkably, export ciphers were totally broken and no-one had noticed!
18924 * Cleaned up the LICENSE document: The official contact for any license
18925 questions now is the OpenSSL core team under openssl-core@openssl.org.
18926 And add a paragraph about the dual-license situation to make sure people
18927 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18928 to the OpenSSL toolkit.
18930 *Ralf S. Engelschall*
18932 * General source tree makefile cleanups: Made `making xxx in yyy...`
18933 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18934 Additionally cleaned up the `make links` target: Remove unnecessary
18935 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18936 to speed processing and no longer clutter the display with confusing
18937 stuff. Instead only the actually done links are displayed.
18939 *Ralf S. Engelschall*
18941 * Permit null encryption ciphersuites, used for authentication only. It used
18942 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18943 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18948 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18949 signed attributes when verifying signatures (this would break them),
18950 the detached data encoding was wrong and public keys obtained using
18951 X509_get_pubkey() weren't freed.
18955 * Add text documentation for the BUFFER functions. Also added a work around
18956 to a Win95 console bug. This was triggered by the password read stuff: the
18957 last character typed gets carried over to the next fread(). If you were
18958 generating a new cert request using 'req' for example then the last
18959 character of the passphrase would be CR which would then enter the first
18964 * Added the new 'Includes OpenSSL Cryptography Software' button as
18965 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18966 button and can be used by applications based on OpenSSL to show the
18967 relationship to the OpenSSL project.
18969 *Ralf S. Engelschall*
18971 * Remove confusing variables in function signatures in files
18972 ssl/ssl_lib.c and ssl/ssl.h.
18974 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18976 * Don't install bss_file.c under PREFIX/include/
18978 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18980 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18981 functions that return function pointers and has support for NT specific
18982 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18983 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18984 unsigned to signed types: this was killing the Win32 compile.
18988 * Add new certificate file to stack functions,
18989 SSL_add_dir_cert_subjects_to_stack() and
18990 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18991 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18992 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18993 This means that Apache-SSL and similar packages don't have to mess around
18994 to add as many CAs as they want to the preferred list.
18998 * Experiment with doxygen documentation. Currently only partially applied to
19000 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19001 openssl.doxy as the configuration file.
19005 * Get rid of remaining C++-style comments which strict C compilers hate.
19007 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19009 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19010 compiled in by default: it has problems with large keys.
19014 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19015 DH private keys and/or callback functions which directly correspond to
19016 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19017 is needed for applications which have to configure certificates on a
19018 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19020 For the RSA certificate situation is makes no difference, but
19021 for the DSA certificate situation this fixes the "no shared cipher"
19022 problem where the OpenSSL cipher selection procedure failed because the
19023 temporary keys were not overtaken from the context and the API provided
19024 no way to reconfigure them.
19025 The new functions now let applications reconfigure the stuff and they
19026 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19027 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19028 non-public-API function ssl_cert_instantiate() is used as a helper
19029 function and also to reduce code redundancy inside ssl_rsa.c.
19031 *Ralf S. Engelschall*
19033 * Move s_server -dcert and -dkey options out of the undocumented feature
19034 area because they are useful for the DSA situation and should be
19035 recognized by the users.
19037 *Ralf S. Engelschall*
19039 * Fix the cipher decision scheme for export ciphers: the export bits are
19040 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19041 SSL_EXP_MASK. So, the original variable has to be used instead of the
19042 already masked variable.
19044 *Richard Levitte <levitte@stacken.kth.se>*
19046 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19048 *Richard Levitte <levitte@stacken.kth.se>*
19050 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19051 from `int` to `unsigned int` because it is a length and initialized by
19052 EVP_DigestFinal() which expects an `unsigned int *`.
19054 *Richard Levitte <levitte@stacken.kth.se>*
19056 * Don't hard-code path to Perl interpreter on shebang line of Configure
19057 script. Instead use the usual Shell->Perl transition trick.
19059 *Ralf S. Engelschall*
19061 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19062 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19063 -noout -modulus` as it's already the case for `openssl rsa -noout
19064 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19065 currently the public key is printed (a decision which was already done by
19066 `openssl dsa -modulus` in the past) which serves a similar purpose.
19067 Additionally the NO_RSA no longer completely removes the whole -modulus
19068 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19071 *Ralf S. Engelschall*
19073 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19074 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19076 *Arne Ansper <arne@ats.cyber.ee>*
19078 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19079 to be added. Now both 'req' and 'ca' can use new objects defined in the
19084 * Add cool BIO that does syslog (or event log on NT).
19086 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19088 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19089 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19090 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19091 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19095 * Add preliminary config info for new extension code.
19099 * Make RSA_NO_PADDING really use no padding.
19101 *Ulf Moeller <ulf@fitug.de>*
19103 * Generate errors when private/public key check is done.
19107 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19108 for some CRL extensions and new objects added.
19112 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19113 key usage extension and fuller support for authority key id.
19117 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19118 padding method for RSA, which is recommended for new applications in PKCS
19119 #1 v2.0 (RFC 2437, October 1998).
19120 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19121 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19122 against Bleichbacher's attack on RSA.
19123 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19126 * Updates to the new SSL compression code
19128 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19130 * Fix so that the version number in the master secret, when passed
19131 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19132 (because the server will not accept higher), that the version number
19133 is 0x03,0x01, not 0x03,0x00
19135 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19137 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19138 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19139 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19143 * Support for RAW extensions where an arbitrary extension can be
19144 created by including its DER encoding. See `apps/openssl.cnf` for
19149 * Make sure latest Perl versions don't interpret some generated C array
19150 code as Perl array code in the crypto/err/err_genc.pl script.
19152 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19154 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19155 not many people have the assembler. Various Win32 compilation fixes and
19156 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19157 build instructions.
19161 * Modify configure script 'Configure' to automatically create crypto/date.h
19162 file under Win32 and also build pem.h from pem.org. New script
19163 util/mkfiles.pl to create the MINFO file on environments that can't do a
19164 'make files': perl util/mkfiles.pl >MINFO should work.
19168 * Major rework of DES function declarations, in the pursuit of correctness
19169 and purity. As a result, many evil casts evaporated, and some weirdness,
19170 too. You may find this causes warnings in your code. Zapping your evil
19171 casts will probably fix them. Mostly.
19175 * Fix for a typo in asn1.h. Bug fix to object creation script
19176 obj_dat.pl. It considered a zero in an object definition to mean
19177 "end of object": none of the objects in objects.h have any zeros
19178 so it wasn't spotted.
19180 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19182 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19183 Masking (CBCM). In the absence of test vectors, the best I have been able
19184 to do is check that the decrypt undoes the encrypt, so far. Send me test
19185 vectors if you have them.
19189 * Correct calculation of key length for export ciphers (too much space was
19190 allocated for null ciphers). This has not been tested!
19194 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19195 message is now correct (it understands "crypto" and "ssl" on its
19196 command line). There is also now an "update" option. This will update
19197 the util/ssleay.num and util/libeay.num files with any new functions.
19199 perl util/mkdef.pl crypto ssl update
19200 it will update them.
19204 * Overhauled the Perl interface:
19205 - ported BN stuff to OpenSSL's different BN library
19206 - made the perl/ source tree CVS-aware
19207 - renamed the package from SSLeay to OpenSSL (the files still contain
19208 their history because I've copied them in the repository)
19209 - removed obsolete files (the test scripts will be replaced
19210 by better Test::Harness variants in the future)
19212 *Ralf S. Engelschall*
19214 * First cut for a very conservative source tree cleanup:
19215 1. merge various obsolete readme texts into doc/ssleay.txt
19216 where we collect the old documents and readme texts.
19217 2. remove the first part of files where I'm already sure that we no
19218 longer need them because of three reasons: either they are just temporary
19219 files which were left by Eric or they are preserved original files where
19220 I've verified that the diff is also available in the CVS via "cvs diff
19221 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19222 the crypto/md/ stuff).
19224 *Ralf S. Engelschall*
19226 * More extension code. Incomplete support for subject and issuer alt
19227 name, issuer and authority key id. Change the i2v function parameters
19228 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19229 what that's for :-) Fix to ASN1 macro which messed up
19230 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19234 * Preliminary support for ENUMERATED type. This is largely copied from the
19239 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19241 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19243 * Make sure `make rehash` target really finds the `openssl` program.
19245 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19247 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19248 like to hear about it if this slows down other processors.
19252 * Add CygWin32 platform information to Configure script.
19254 *Alan Batie <batie@aahz.jf.intel.com>*
19256 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19258 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19260 * New program nseq to manipulate netscape certificate sequences
19264 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19269 * Fixes to BN code. Previously the default was to define BN_RECURSION
19270 but the BN code had some problems that would cause failures when
19271 doing certificate verification and some other functions.
19273 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19275 * Add ASN1 and PEM code to support netscape certificate sequences.
19279 * Add ASN1 and PEM code to support netscape certificate sequences.
19283 * Add several PKIX and private extended key usage OIDs.
19287 * Modify the 'ca' program to handle the new extension code. Modify
19288 openssl.cnf for new extension format, add comments.
19292 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19293 and add a sample to openssl.cnf so req -x509 now adds appropriate
19298 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19299 error code, add initial support to X509_print() and x509 application.
19303 * Takes a deep breath and start adding X509 V3 extension support code. Add
19304 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19305 stuff is currently isolated and isn't even compiled yet.
19309 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19310 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19311 Removed the versions check from X509 routines when loading extensions:
19312 this allows certain broken certificates that don't set the version
19313 properly to be processed.
19317 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19318 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19319 can still be regenerated with "make depend".
19323 * Spelling mistake in C version of CAST-128.
19325 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19327 * Changes to the error generation code. The perl script err-code.pl
19328 now reads in the old error codes and retains the old numbers, only
19329 adding new ones if necessary. It also only changes the .err files if new
19330 codes are added. The makefiles have been modified to only insert errors
19331 when needed (to avoid needlessly modifying header files). This is done
19332 by only inserting errors if the .err file is newer than the auto generated
19333 C file. To rebuild all the error codes from scratch (the old behaviour)
19334 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19335 or delete all the .err files.
19339 * CAST-128 was incorrectly implemented for short keys. The C version has
19340 been fixed, but is untested. The assembler versions are also fixed, but
19341 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19342 to regenerate it if needed.
19343 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19344 Hagino <itojun@kame.net>*
19346 * File was opened incorrectly in randfile.c.
19348 *Ulf Möller <ulf@fitug.de>*
19350 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19351 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19352 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19353 al: it's just almost always a UTCTime. Note this patch adds new error
19354 codes so do a "make errors" if there are problems.
19358 * Correct Linux 1 recognition in config.
19360 *Ulf Möller <ulf@fitug.de>*
19362 * Remove pointless MD5 hash when using DSA keys in ca.
19364 *Anonymous <nobody@replay.com>*
19366 * Generate an error if given an empty string as a cert directory. Also
19367 generate an error if handed NULL (previously returned 0 to indicate an
19368 error, but didn't set one).
19370 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19372 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19376 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19377 parameters. This was causing a warning which killed off the Win32 compile.
19381 * Remove C++ style comments from crypto/bn/bn_local.h.
19383 *Neil Costigan <neil.costigan@celocom.com>*
19385 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19386 based on a text string, looking up short and long names and finally
19387 "dot" format. The "dot" format stuff didn't work. Added new function
19388 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19389 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19390 OID is not part of the table.
19394 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19395 X509_LOOKUP_by_alias().
19399 * Sort openssl functions by name.
19403 * Get the `gendsa` command working and add it to the `list` command. Remove
19404 encryption from sample DSA keys (in case anyone is interested the password
19409 * Make *all* `*_free` functions accept a NULL pointer.
19411 *Frans Heymans <fheymans@isaserver.be>*
19413 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19416 *Anonymous <nobody@replay.com>*
19418 * s_server should send the CAfile as acceptable CAs, not its own cert.
19420 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19422 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19424 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19426 * Temp key "for export" tests were wrong in s3_srvr.c.
19428 *Anonymous <nobody@replay.com>*
19430 * Add prototype for temp key callback functions
19431 SSL_CTX_set_tmp_{rsa,dh}_callback().
19435 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19436 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19440 * X509_name_add_entry() freed the wrong thing after an error.
19442 *Arne Ansper <arne@ats.cyber.ee>*
19444 * rsa_eay.c would attempt to free a NULL context.
19446 *Arne Ansper <arne@ats.cyber.ee>*
19448 * BIO_s_socket() had a broken should_retry() on Windoze.
19450 *Arne Ansper <arne@ats.cyber.ee>*
19452 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19454 *Arne Ansper <arne@ats.cyber.ee>*
19456 * Make sure the already existing X509_STORE->depth variable is initialized
19457 in X509_STORE_new(), but document the fact that this variable is still
19458 unused in the certificate verification process.
19460 *Ralf S. Engelschall*
19462 * Fix the various library and `apps/` files to free up pkeys obtained from
19463 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19467 * Fix reference counting in X509_PUBKEY_get(). This makes
19468 demos/maurice/example2.c work, amongst others, probably.
19470 *Steve Henson and Ben Laurie*
19472 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19473 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19474 are no longer created. This way we have a single and consistent command
19475 line interface `openssl <command>`, similar to `cvs <command>`.
19477 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19479 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19480 BIT STRING wrapper always have zero unused bits.
19484 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19488 * Make the top-level INSTALL documentation easier to understand.
19492 * Makefiles updated to exit if an error occurs in a sub-directory
19493 make (including if user presses ^C) [Paul Sutton]
19495 * Make Montgomery context stuff explicit in RSA data structure.
19499 * Fix build order of pem and err to allow for generated pem.h.
19503 * Fix renumbering bug in X509_NAME_delete_entry().
19507 * Enhanced the err-ins.pl script so it makes the error library number
19508 global and can add a library name. This is needed for external ASN1 and
19509 other error libraries.
19513 * Fixed sk_insert which never worked properly.
19517 * Fix ASN1 macros so they can handle indefinite length constructed
19518 EXPLICIT tags. Some non standard certificates use these: they can now
19523 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19524 into a single doc/ssleay.txt bundle. This way the information is still
19525 preserved but no longer messes up this directory. Now it's new room for
19526 the new set of documentation files.
19528 *Ralf S. Engelschall*
19530 * SETs were incorrectly DER encoded. This was a major pain, because they
19531 shared code with SEQUENCEs, which aren't coded the same. This means that
19532 almost everything to do with SETs or SEQUENCEs has either changed name or
19533 number of arguments.
19535 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19537 * Fix test data to work with the above.
19541 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19542 was already fixed by Eric for 0.9.1 it seems.
19544 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19546 * Autodetect FreeBSD3.
19550 * Fix various bugs in Configure. This affects the following platforms:
19554 unixware-2.0-pentium
19559 * Eliminate generated files from CVS. Reorder tests to regenerate files
19560 before they are needed.
19564 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19568 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19570 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19571 changed SSLeay to OpenSSL in version strings.
19573 *Ralf S. Engelschall*
19575 * Some fixups to the top-level documents.
19579 * Fixed the nasty bug where rsaref.h was not found under compile-time
19580 because the symlink to include/ was missing.
19582 *Ralf S. Engelschall*
19584 * Incorporated the popular no-RSA/DSA-only patches
19585 which allow to compile an RSA-free SSLeay.
19587 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19589 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19590 when "ssleay" is still not found.
19592 *Ralf S. Engelschall*
19594 * Added more platforms to Configure: Cray T3E, HPUX 11,
19596 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19598 * Updated the README file.
19600 *Ralf S. Engelschall*
19602 * Added various .cvsignore files in the CVS repository subdirs
19603 to make a "cvs update" really silent.
19605 *Ralf S. Engelschall*
19607 * Recompiled the error-definition header files and added
19608 missing symbols to the Win32 linker tables.
19610 *Ralf S. Engelschall*
19612 * Cleaned up the top-level documents;
19613 o new files: CHANGES and LICENSE
19614 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19615 o merged COPYRIGHT into LICENSE
19616 o removed obsolete TODO file
19617 o renamed MICROSOFT to INSTALL.W32
19619 *Ralf S. Engelschall*
19621 * Removed dummy files from the 0.9.1b source tree:
19622 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19623 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19624 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19625 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19626 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19628 *Ralf S. Engelschall*
19630 * Added various platform portability fixes.
19634 * The Genesis of the OpenSSL rpject:
19635 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19636 Young and Tim J. Hudson created while they were working for C2Net until
19639 *The OpenSSL Project*
19641 ### Changes between 0.9.0b and 0.9.1b [not released]
19643 * Updated a few CA certificates under certs/
19647 * Changed some BIGNUM api stuff.
19651 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19652 DGUX x86, Linux Alpha, etc.
19656 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19657 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19662 * Add -strparse option to asn1pars program which parses nested
19665 *Dr Stephen Henson <shenson@bigfoot.com>*
19667 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19671 * DSA fix for "ca" program.
19675 * Added "-genkey" option to "dsaparam" program.
19679 * Added RIPE MD160 (rmd160) message digest.
19683 * Added -a (all) option to "ssleay version" command.
19687 * Added PLATFORM define which is the id given to Configure.
19691 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19695 * Extended the ASN.1 parser routines.
19699 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19703 * Added a BN_CTX to the BN library.
19707 * Fixed the weak key values in DES library
19711 * Changed API in EVP library for cipher aliases.
19715 * Added support for RC2/64bit cipher.
19719 * Converted the lhash library to the crypto/mem.c functions.
19723 * Added more recognized ASN.1 object ids.
19727 * Added more RSA padding checks for SSL/TLS.
19731 * Added BIO proxy/filter functionality.
19735 * Added extra_certs to SSL_CTX which can be used
19736 send extra CA certificates to the client in the CA cert chain sending
19737 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19741 * Now Fortezza is denied in the authentication phase because
19742 this is key exchange mechanism is not supported by SSLeay at all.
19746 * Additional PKCS1 checks.
19750 * Support the string "TLSv1" for all TLS v1 ciphers.
19754 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19755 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19759 * Fixed a few memory leaks.
19763 * Fixed various code and comment typos.
19767 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19768 bytes sent in the client random.
19770 *Edward Bishop <ebishop@spyglass.com>*
19774 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19775 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19776 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19777 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19778 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19779 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19780 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19781 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19782 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19783 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19784 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19785 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19786 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19787 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19788 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19789 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19790 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19791 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19792 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19793 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19794 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19795 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19796 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19797 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19798 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19799 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19800 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19801 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19802 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19803 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19804 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19805 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19806 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19807 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19808 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19809 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19810 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19811 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19812 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19813 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19814 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19815 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19816 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19817 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19818 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19819 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19820 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19821 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19822 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19823 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19824 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19825 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19826 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19827 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19828 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19829 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19830 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19831 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19832 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19833 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19834 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19835 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19836 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19837 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19838 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19839 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19840 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19841 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19842 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19843 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19844 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19845 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19846 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19847 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19848 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19849 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19850 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19851 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19852 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19853 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19854 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19855 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19856 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19857 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19858 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19859 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19860 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19861 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19862 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19863 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19864 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19865 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19866 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19867 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19868 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19869 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19870 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19871 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19872 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19873 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19874 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19875 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19876 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19877 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19878 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19879 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19880 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19881 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19882 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19883 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19884 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19885 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19886 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19887 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19888 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19889 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19890 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19891 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19892 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19893 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19894 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19895 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19896 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19897 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19898 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19899 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19900 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19901 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19902 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19903 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19904 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19905 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19906 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19907 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19908 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19909 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19910 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19911 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19912 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19913 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19914 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19915 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19916 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19917 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19918 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19919 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19920 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19921 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19922 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19923 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19924 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19925 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19926 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19927 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19928 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19929 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19930 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19931 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19932 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19933 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19934 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19935 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19936 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19937 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19938 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19939 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19940 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19941 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19942 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19943 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19944 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19945 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19946 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19947 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19948 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19949 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19950 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19951 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19952 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655