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 3.1.0 and 3.1.1 [xx XXX xxxx]
27 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
28 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
29 trigger a crash of an application using AES-XTS decryption if the memory
30 just after the buffer being decrypted is not mapped.
31 Thanks to Anton Romanov (Amazon) for discovering the issue.
36 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
37 The previous fix for this timing side channel turned out to cause
38 a severe 2-3x performance regression in the typical use case
39 compared to 3.0.7. The new fix uses existing constant time
40 code paths, and restores the previous performance level while
41 fully eliminating all existing timing side channels.
42 The fix was developed by Bernd Edlinger with testing support
47 * Add FIPS provider configuration option to disallow the use of
48 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
49 The option '-no_drbg_truncated_digests' can optionally be
50 supplied to 'openssl fipsinstall'.
54 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
55 that it does not enable policy checking. Thanks to David Benjamin for
56 discovering this issue.
61 * Fixed an issue where invalid certificate policies in leaf certificates are
62 silently ignored by OpenSSL and other certificate policy checks are skipped
63 for that certificate. A malicious CA could use this to deliberately assert
64 invalid certificate policies in order to circumvent policy checking on the
65 certificate altogether.
70 * Limited the number of nodes created in a policy tree to mitigate
71 against CVE-2023-0464. The default limit is set to 1000 nodes, which
72 should be sufficient for most installations. If required, the limit
73 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
74 time define to a desired maximum number of nodes or zero to allow
80 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
82 * Add FIPS provider configuration option to enforce the
83 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
84 The option '-ems-check' can optionally be supplied to
85 'openssl fipsinstall'.
89 * The FIPS provider includes a few non-approved algorithms for
90 backward compatibility purposes and the "fips=yes" property query
91 must be used for all algorithm fetches to ensure FIPS compliance.
93 The algorithms that are included but not approved are Triple DES ECB,
94 Triple DES CBC and EdDSA.
98 * Added support for KMAC in KBKDF.
102 * RNDR and RNDRRS support in provider functions to provide
103 random number generation for Arm CPUs (aarch64).
107 * s_client and s_server apps now explicitly say when the TLS version
108 does not include the renegotiation mechanism. This avoids confusion
109 between that scenario versus when the TLS version includes secure
110 renegotiation but the peer lacks support for it.
114 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
116 *Tomasz Kantecki, Andrey Matyukov*
118 * The various OBJ_* functions have been made thread safe.
122 * Parallel dual-prime 1536/2048-bit modular exponentiation for
123 AVX512_IFMA capable processors.
125 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
127 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
128 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
129 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
130 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
131 `OPENSSL_NO_DEPRECATED_3_1`.
133 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
134 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
135 definitions for these functions regardless of whether
136 `OPENSSL_NO_DEPRECATED_3_1` is defined.
138 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
139 functions regardless of whether they are using them. It is recommended that
140 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
144 * When generating safe-prime DH parameters set the recommended private key
145 length equivalent to minimum key lengths as in RFC 7919.
149 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
150 maximum size that is smaller or equal to the digest length to comply with
151 FIPS 186-4 section 5. This is implemented by a new option
152 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
153 `rsa_pss_saltlen` parameter, which is now the default. Signature
154 verification is not affected by this change and continues to work as before.
161 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
162 listed here are only a brief description.
163 The migration guide contains more detailed information related to new features,
164 breaking changes, and mappings for the large list of deprecated functions.
166 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
168 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
170 * Fixed NULL dereference during PKCS7 data verification.
172 A NULL pointer can be dereferenced when signatures are being
173 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
174 algorithm used for the signature is known to the OpenSSL library but
175 the implementation of the hash algorithm is not available the digest
176 initialization will fail. There is a missing check for the return
177 value from the initialization function which later leads to invalid
178 usage of the digest API most likely leading to a crash.
181 PKCS7 data is processed by the SMIME library calls and also by the
182 time stamp (TS) library calls. The TLS implementation in OpenSSL does
183 not call these functions however third party applications would be
184 affected if they call these functions to verify signatures on untrusted
189 * Fixed X.400 address type confusion in X.509 GeneralName.
191 There is a type confusion vulnerability relating to X.400 address processing
192 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
193 but the public structure definition for GENERAL_NAME incorrectly specified
194 the type of the x400Address field as ASN1_TYPE. This field is subsequently
195 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
198 When CRL checking is enabled (i.e. the application sets the
199 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
200 pass arbitrary pointers to a memcmp call, enabling them to read memory
201 contents or enact a denial of service.
206 * Fixed NULL dereference validating DSA public key.
208 An invalid pointer dereference on read can be triggered when an
209 application tries to check a malformed DSA public key by the
210 EVP_PKEY_public_check() function. This will most likely lead
211 to an application crash. This function can be called on public
212 keys supplied from untrusted sources which could allow an attacker
213 to cause a denial of service attack.
215 The TLS implementation in OpenSSL does not call this function
216 but applications might call the function if there are additional
217 security requirements imposed by standards such as FIPS 140-3.
220 *Shane Lontis, Tomáš Mráz*
222 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
224 An invalid pointer dereference on read can be triggered when an
225 application tries to load malformed PKCS7 data with the
226 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
228 The result of the dereference is an application crash which could
229 lead to a denial of service attack. The TLS implementation in OpenSSL
230 does not call this function however third party applications might
231 call these functions on untrusted data.
236 * Fixed Use-after-free following BIO_new_NDEF.
238 The public API function BIO_new_NDEF is a helper function used for
239 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
240 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
241 be called directly by end user applications.
243 The function receives a BIO from the caller, prepends a new BIO_f_asn1
244 filter BIO onto the front of it to form a BIO chain, and then returns
245 the new head of the BIO chain to the caller. Under certain conditions,
246 for example if a CMS recipient public key is invalid, the new filter BIO
247 is freed and the function returns a NULL result indicating a failure.
248 However, in this case, the BIO chain is not properly cleaned up and the
249 BIO passed by the caller still retains internal pointers to the previously
250 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
251 then a use-after-free will occur. This will most likely result in a crash.
254 *Viktor Dukhovni, Matt Caswell*
256 * Fixed Double free after calling PEM_read_bio_ex.
258 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
259 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
260 data. If the function succeeds then the "name_out", "header" and "data"
261 arguments are populated with pointers to buffers containing the relevant
262 decoded data. The caller is responsible for freeing those buffers. It is
263 possible to construct a PEM file that results in 0 bytes of payload data.
264 In this case PEM_read_bio_ex() will return a failure code but will populate
265 the header argument with a pointer to a buffer that has already been freed.
266 If the caller also frees this buffer then a double free will occur. This
267 will most likely lead to a crash.
269 The functions PEM_read_bio() and PEM_read() are simple wrappers around
270 PEM_read_bio_ex() and therefore these functions are also directly affected.
272 These functions are also called indirectly by a number of other OpenSSL
273 functions including PEM_X509_INFO_read_bio_ex() and
274 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
275 internal uses of these functions are not vulnerable because the caller does
276 not free the header argument if PEM_read_bio_ex() returns a failure code.
279 *Kurt Roeckx, Matt Caswell*
281 * Fixed Timing Oracle in RSA Decryption.
283 A timing based side channel exists in the OpenSSL RSA Decryption
284 implementation which could be sufficient to recover a plaintext across
285 a network in a Bleichenbacher style attack. To achieve a successful
286 decryption an attacker would have to be able to send a very large number
287 of trial messages for decryption. The vulnerability affects all RSA padding
288 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
291 *Dmitry Belyavsky, Hubert Kario*
293 * Fixed X.509 Name Constraints Read Buffer Overflow.
295 A read buffer overrun can be triggered in X.509 certificate verification,
296 specifically in name constraint checking. The read buffer overrun might
297 result in a crash which could lead to a denial of service attack.
298 In a TLS client, this can be triggered by connecting to a malicious
299 server. In a TLS server, this can be triggered if the server requests
300 client authentication and a malicious client connects.
305 * Fixed X.509 Policy Constraints Double Locking security issue.
307 If an X.509 certificate contains a malformed policy constraint and
308 policy processing is enabled, then a write lock will be taken twice
309 recursively. On some operating systems (most widely: Windows) this
310 results in a denial of service when the affected process hangs. Policy
311 processing being enabled on a publicly facing server is not considered
312 to be a common setup.
317 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
318 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
319 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
320 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
321 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
322 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
323 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
324 for legacy EC and SM2 keys is also changed similarly to honor the
325 equivalent conversion format flag as specified in the underlying
326 `EC_KEY` object being exported to a provider, when this function is
327 called through `EVP_PKEY_export()`.
331 * Fixed a type confusion vulnerability relating to X.400 address processing
332 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
333 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
334 vulnerability may allow an attacker who can provide a certificate chain and
335 CRL (neither of which need have a valid signature) to pass arbitrary pointers
336 to a `memcmp` call, creating a possible read primitive, subject to some
337 constraints. Refer to the advisory for more information. Thanks to David
338 Benjamin for discovering this issue. ([CVE-2023-0286])
340 This issue has been fixed by changing the public header file definition of
341 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
342 possible for any existing application to successfully use the existing
343 definition; however, if any application references the `x400Address` field
344 (e.g. in dead code), note that the type of this field has changed. There is
349 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
351 * Fixed two buffer overflows in punycode decoding functions.
353 A buffer overrun can be triggered in X.509 certificate verification,
354 specifically in name constraint checking. Note that this occurs after
355 certificate chain signature verification and requires either a CA to
356 have signed the malicious certificate or for the application to continue
357 certificate verification despite failure to construct a path to a trusted
360 In a TLS client, this can be triggered by connecting to a malicious
361 server. In a TLS server, this can be triggered if the server requests
362 client authentication and a malicious client connects.
364 An attacker can craft a malicious email address to overflow
365 an arbitrary number of bytes containing the `.` character (decimal 46)
366 on the stack. This buffer overflow could result in a crash (causing a
370 An attacker can craft a malicious email address to overflow four
371 attacker-controlled bytes on the stack. This buffer overflow could
372 result in a crash (causing a denial of service) or potentially remote code
373 execution depending on stack layout for any given platform/compiler.
378 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
379 parameters in OpenSSL code.
380 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
381 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
382 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
383 Using these invalid names may cause algorithms to use slower methods
384 that ignore the CRT parameters.
388 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
393 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
394 data to be signed before signing the certificate.
398 * Added RIPEMD160 to the default provider.
402 * Ensured that the key share group sent or accepted for the key exchange
403 is allowed for the protocol version.
407 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
409 * OpenSSL supports creating a custom cipher via the legacy
410 EVP_CIPHER_meth_new() function and associated function calls. This function
411 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
412 to use the new provider mechanism in order to implement custom ciphers.
414 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
415 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
416 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
417 and decryption initialisation functions). Instead of using the custom cipher
418 directly it incorrectly tries to fetch an equivalent cipher from the
419 available providers. An equivalent cipher is found based on the NID passed to
420 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
421 given cipher. However it is possible for an application to incorrectly pass
422 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
423 is used in this way the OpenSSL encryption/decryption initialisation function
424 will match the NULL cipher as being equivalent and will fetch this from the
425 available providers. This will succeed if the default provider has been
426 loaded (or if a third party provider has been loaded that offers this
427 cipher). Using the NULL cipher means that the plaintext is emitted as the
430 Applications are only affected by this issue if they call
431 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
432 encryption/decryption initialisation function. Applications that only use
433 SSL/TLS are not impacted by this issue.
438 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
443 * Fixed the linux-mips64 Configure target which was missing the
444 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
449 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
454 * Correctly handle a retransmitted ClientHello in DTLS
458 * Fixed detection of ktls support in cross-compile environment on Linux
462 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
467 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
468 report correct results in some cases
472 * Fix UWP builds by defining VirtualLock
476 * For known safe primes use the minimum key length according to RFC 7919.
477 Longer private key sizes unnecessarily raise the cycles needed to compute the
478 shared secret without any increase of the real security. This fixes a
479 regression from 1.1.1 where these shorter keys were generated for the known
484 * Added the loongarch64 target
488 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
489 only passed to the FIPS provider and not to the default or legacy provider.
493 * Fixed reported performance degradation on aarch64. Restored the
494 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
495 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
496 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
497 The new algorithm is still used for 32 bit targets.
501 * Added a missing header for memcmp that caused compilation failure on some
506 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
508 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
509 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
510 This issue makes the RSA implementation with 2048 bit private keys
511 incorrect on such machines and memory corruption will happen during
512 the computation. As a consequence of the memory corruption an attacker
513 may be able to trigger a remote code execution on the machine performing
516 SSL/TLS servers or other servers using 2048 bit RSA private keys running
517 on machines supporting AVX512IFMA instructions of the X86_64 architecture
518 are affected by this issue.
523 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
524 implementation would not encrypt the entirety of the data under some
525 circumstances. This could reveal sixteen bytes of data that was
526 preexisting in the memory that wasn't written. In the special case of
527 "in place" encryption, sixteen bytes of the plaintext would be revealed.
529 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
530 they are both unaffected.
533 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
535 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
537 * In addition to the c_rehash shell command injection identified in
538 CVE-2022-1292, further bugs where the c_rehash script does not
539 properly sanitise shell metacharacters to prevent command injection have been
542 When the CVE-2022-1292 was fixed it was not discovered that there
543 are other places in the script where the file names of certificates
544 being hashed were possibly passed to a command executed through the shell.
546 This script is distributed by some operating systems in a manner where
547 it is automatically executed. On such operating systems, an attacker
548 could execute arbitrary commands with the privileges of the script.
550 Use of the c_rehash script is considered obsolete and should be replaced
551 by the OpenSSL rehash command line tool.
554 *Daniel Fiala, Tomáš Mráz*
556 * Case insensitive string comparison no longer uses locales. It has instead
557 been directly implemented.
561 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
563 * Case insensitive string comparison is reimplemented via new locale-agnostic
564 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
565 comparison. The previous implementation had problems when the Turkish locale
570 * Fixed a bug in the c_rehash script which was not properly sanitising shell
571 metacharacters to prevent command injection. This script is distributed by
572 some operating systems in a manner where it is automatically executed. On
573 such operating systems, an attacker could execute arbitrary commands with the
574 privileges of the script.
576 Use of the c_rehash script is considered obsolete and should be replaced
577 by the OpenSSL rehash command line tool.
582 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
583 certificate on an OCSP response. The bug caused the function in the case
584 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
585 response (meaning a successful verification) even in the case where the
586 response signing certificate fails to verify.
588 It is anticipated that most users of `OCSP_basic_verify` will not use the
589 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
590 a negative value (indicating a fatal error) in the case of a certificate
591 verification failure. The normal expected return value in this case would be
594 This issue also impacts the command line OpenSSL "ocsp" application. When
595 verifying an ocsp response with the "-no_cert_checks" option the command line
596 application will report that the verification is successful even though it
597 has in fact failed. In this case the incorrect successful response will also
598 be accompanied by error messages showing the failure and contradicting the
599 apparently successful result.
604 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
605 AAD data as the MAC key. This made the MAC key trivially predictable.
607 An attacker could exploit this issue by performing a man-in-the-middle attack
608 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
609 that the modified data would still pass the MAC integrity check.
611 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
612 endpoint will always be rejected by the recipient and the connection will
613 fail at that point. Many application protocols require data to be sent from
614 the client to the server first. Therefore, in such a case, only an OpenSSL
615 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
617 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
618 sent in both directions. In this case both clients and servers could be
619 affected, regardless of the application protocol.
621 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
622 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
623 the handshake when using this ciphersuite.
625 The confidentiality of data is not impacted by this issue, i.e. an attacker
626 cannot decrypt data that has been encrypted using this ciphersuite - they can
629 In order for this attack to work both endpoints must legitimately negotiate
630 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
631 OpenSSL 3.0, and is not available within the default provider or the default
632 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
633 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
634 following must have occurred:
636 1) OpenSSL must have been compiled with the (non-default) compile time option
637 enable-weak-ssl-ciphers
639 2) OpenSSL must have had the legacy provider explicitly loaded (either
640 through application code or via configuration)
642 3) The ciphersuite must have been explicitly added to the ciphersuite list
644 4) The libssl security level must have been set to 0 (default is 1)
646 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
648 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
649 others that both endpoints have in common
654 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
655 occupied by the removed hash table entries.
657 This function is used when decoding certificates or keys. If a long lived
658 process periodically decodes certificates or keys its memory usage will
659 expand without bounds and the process might be terminated by the operating
660 system causing a denial of service. Also traversing the empty hash table
661 entries will take increasingly more time.
663 Typically such long lived processes might be TLS clients or TLS servers
664 configured to accept client certificate authentication.
667 *Hugo Landau, Aliaksei Levin*
669 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
670 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
671 statistics are no longer supported. For compatibility, these statistics are
672 still listed in the output but are now always reported as zero.
676 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
678 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
679 for non-prime moduli.
681 Internally this function is used when parsing certificates that contain
682 elliptic curve public keys in compressed form or explicit elliptic curve
683 parameters with a base point encoded in compressed form.
685 It is possible to trigger the infinite loop by crafting a certificate that
686 has invalid explicit curve parameters.
688 Since certificate parsing happens prior to verification of the certificate
689 signature, any process that parses an externally supplied certificate may thus
690 be subject to a denial of service attack. The infinite loop can also be
691 reached when parsing crafted private keys as they can contain explicit
692 elliptic curve parameters.
694 Thus vulnerable situations include:
696 - TLS clients consuming server certificates
697 - TLS servers consuming client certificates
698 - Hosting providers taking certificates or private keys from customers
699 - Certificate authorities parsing certification requests from subscribers
700 - Anything else which parses ASN.1 elliptic curve parameters
702 Also any other applications that use the BN_mod_sqrt() where the attacker
703 can control the parameter values are vulnerable to this DoS issue.
708 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
709 to the list of ciphersuites providing Perfect Forward Secrecy as
710 required by SECLEVEL >= 3.
712 *Dmitry Belyavskiy, Nicola Tuveri*
714 * Made the AES constant time code for no-asm configurations
715 optional due to the resulting 95% performance degradation.
716 The AES constant time code can be enabled, for no assembly
717 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
721 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
726 * The negative return value handling of the certificate verification callback
727 was reverted. The replacement is to set the verification retry state with
728 the SSL_set_retry_verify() function.
732 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
734 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
735 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
736 verify a certificate supplied by a server. That function may return a
737 negative return value to indicate an internal error (for example out of
738 memory). Such a negative return value is mishandled by OpenSSL and will cause
739 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
740 success and a subsequent call to SSL_get_error() to return the value
741 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
742 returned by OpenSSL if the application has previously called
743 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
744 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
745 totally unexpected and applications may not behave correctly as a result. The
746 exact behaviour will depend on the application but it could result in
747 crashes, infinite loops or other similar incorrect responses.
749 This issue is made more serious in combination with a separate bug in OpenSSL
750 3.0 that will cause X509_verify_cert() to indicate an internal error when
751 processing a certificate chain. This will occur where a certificate does not
752 include the Subject Alternative Name extension but where a Certificate
753 Authority has enforced name constraints. This issue can occur even with valid
759 * Corrected a few file name and file reference bugs in the build,
760 installation and setup scripts, which lead to installation verification
761 failures. Slightly enhanced the installation verification script.
765 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
770 * Fixed PVK encoder to properly query for the passphrase.
774 * Multiple fixes in the OSSL_HTTP API functions.
778 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
779 OSSL_PARAM_INTEGER data type and return error on negative numbers
780 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
781 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
785 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
789 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
793 * Multiple threading fixes.
797 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
801 * Allow fetching an operation from the provider that owns an unexportable key
802 as a fallback if that is still allowed by the property query.
806 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
808 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
813 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
814 S390X capability vector to zero. This simplifies testing of different code
815 paths on S390X architecture.
819 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
820 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
821 SP 800-38D". The communication will fail at this point.
825 * The EC_GROUP_clear_free() function is deprecated as there is nothing
826 confidential in EC_GROUP data.
830 * The byte order mark (BOM) character is ignored if encountered at the
831 beginning of a PEM-formatted file.
835 * Added CMS support for the Russian GOST algorithms.
839 * Due to move of the implementation of cryptographic operations
840 to the providers, validation of various operation parameters can
841 be postponed until the actual operation is executed where previously
842 it happened immediately when an operation parameter was set.
844 For example when setting an unsupported curve with
845 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
846 fail but later keygen operations with the EVP_PKEY_CTX will fail.
848 *OpenSSL team members and many third party contributors*
850 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
851 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
852 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
853 instead to retrieve these algorithms from a provider.
857 * On build targets where the multilib postfix is set in the build
858 configuration the libdir directory was changing based on whether
859 the lib directory with the multilib postfix exists on the system
860 or not. This unpredictable behavior was removed and eventual
861 multilib postfix is now always added to the default libdir. Use
862 `--libdir=lib` to override the libdir if adding the postfix is
867 * The triple DES key wrap functionality now conforms to RFC 3217 but is
868 no longer interoperable with OpenSSL 1.1.1.
872 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
873 function codes, this function can only cause problems for calling
878 * Add a configurable flag to output date formats as ISO 8601. Does not
879 change the default date format.
883 * Version of MSVC earlier than 1300 could get link warnings, which could
884 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
885 Support for this flag has been removed.
889 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
890 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
891 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
892 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
893 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
897 * The signatures of the functions to get and set options on SSL and
898 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
899 Some source code changes may be required.
903 * The public definitions of conf_method_st and conf_st have been
904 deprecated. They will be made opaque in a future release.
906 *Rich Salz and Tomáš Mráz*
908 * Client-initiated renegotiation is disabled by default. To allow it, use
909 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
910 flag, or the "ClientRenegotiation" config parameter as appropriate.
914 * Add "abspath" and "includedir" pragma's to config files, to prevent,
915 or modify relative pathname inclusion.
919 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
920 validated. Please consult the README-FIPS and
921 README-PROVIDERS files, as well as the migration guide.
923 *OpenSSL team members and many third party contributors*
925 * For the key types DH and DHX the allowed settable parameters are now different.
929 * The openssl commands that read keys, certificates, and CRLs now
930 automatically detect the PEM or DER format of the input files.
932 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
934 * Added enhanced PKCS#12 APIs which accept a library context.
938 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
942 * Added support for Kernel TLS (KTLS).
944 *Boris Pismenny, John Baldwin and Andrew Gallatin*
946 * Support for RFC 5746 secure renegotiation is now required by default for
947 SSL or TLS connections to succeed.
951 * The signature of the `copy` functional parameter of the
952 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
953 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
954 the signature of the `pub_decode` functional parameter of the
955 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
956 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
960 * The error return values from some control calls (ctrl) have changed.
964 * A public key check is now performed during EVP_PKEY_derive_set_peer().
968 * Many functions in the EVP_ namespace that are getters of values from
969 implementations or contexts were renamed to include get or get0 in their
970 names. Old names are provided as macro aliases for compatibility and
975 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
976 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
977 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
982 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
985 * The output from the command line applications may have minor
990 * The output from numerous "printing" may have minor changes.
994 * Windows thread synchronization uses read/write primitives (SRWLock) when
995 supported by the OS, otherwise CriticalSection continues to be used.
999 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1000 work on read only BIO source/sinks that do not support these functions.
1001 This allows piping or redirection of a file BIO using stdin to be buffered
1002 into memory. This is used internally in OSSL_DECODER_from_bio().
1006 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1007 this function would return one of the values OSSL_STORE_INFO_NAME,
1008 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1009 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1010 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1011 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1012 using this function should be amended to handle the changed return value.
1016 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1017 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1018 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1019 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1020 contains more than one certificate identifier: This means that all
1021 certificates referenced there MUST be part of the validation chain.
1025 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1026 RC5, DESX and DES have been moved to the legacy provider.
1030 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1031 RIPEMD-160 have been moved to the legacy provider.
1035 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1040 * The deprecated functions EVP_PKEY_get0_RSA(),
1041 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1042 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1043 well as the similarly named "get1" functions behave differently in
1048 * A number of functions handling low-level keys or engines were deprecated
1049 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1050 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1051 EVP_PKEY_get0_siphash().
1055 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1056 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1057 will need to load the legacy crypto provider. This includes these PBE
1058 algorithms which use this KDF:
1059 - NID_pbeWithMD2AndDES_CBC
1060 - NID_pbeWithMD5AndDES_CBC
1061 - NID_pbeWithSHA1AndRC2_CBC
1062 - NID_pbeWithMD2AndRC2_CBC
1063 - NID_pbeWithMD5AndRC2_CBC
1064 - NID_pbeWithSHA1AndDES_CBC
1068 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1069 BIO_debug_callback() functions.
1073 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1074 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1078 * The RAND_METHOD APIs have been deprecated.
1082 * The SRP APIs have been deprecated.
1086 * Add a compile time option to prevent the caching of provider fetched
1087 algorithms. This is enabled by including the no-cached-fetch option
1088 at configuration time.
1092 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1093 count of PKCS12_DEFAULT_ITER.
1095 *Tomáš Mráz and Sahana Prasad*
1097 * The openssl speed command does not use low-level API calls anymore.
1101 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1104 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1106 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1110 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1111 providers may supply their own group implementations (using either the "key
1112 exchange" or the "key encapsulation" methods) which will automatically be
1113 detected and used by libssl.
1115 *Matt Caswell, Nicola Tuveri*
1117 * The undocumented function X509_certificate_type() has been deprecated;
1121 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1125 * Removed RSA padding mode for SSLv23 (which was only used for
1126 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1127 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1132 * Deprecated the obsolete X9.31 RSA key generation related functions.
1134 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1135 is not allowed to return a value > 1, this is no more taken as failure.
1137 *Viktor Dukhovni and David von Oheimb*
1139 * Deprecated the obsolete X9.31 RSA key generation related functions
1140 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1141 BN_X931_generate_prime_ex().
1145 * The default key generation method for the regular 2-prime RSA keys was
1146 changed to the FIPS 186-4 B.3.6 method.
1150 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1154 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1158 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1159 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1161 *Rich Salz, Richard Levitte, and David von Oheimb*
1163 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1167 * Deprecated `OCSP_parse_url()`.
1171 * Validation of SM2 keys has been separated from the validation of regular EC
1176 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1177 switches: a validation failure triggers an early exit, returning a failure
1178 exit status to the parent process.
1182 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1183 to ignore unknown ciphers.
1187 * The `-cipher-commands` and `-digest-commands` options
1188 of the command line utility `list` have been deprecated.
1189 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1193 * Added convenience functions for generating asymmetric key pairs:
1194 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1195 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1199 * All of the low-level EC_KEY functions have been deprecated.
1201 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1203 * Deprecated all the libcrypto and libssl error string loading
1208 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1209 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1214 * The `-crypt` option to the `passwd` command line tool has been removed.
1218 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1223 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1227 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1228 EVP_PKEY_get1_tls_encodedpoint().
1232 * The security callback, which can be customised by application code, supports
1233 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1234 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1238 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1239 interface. Their functionality remains unchanged.
1243 * Added new option for 'openssl list', '-providers', which will display the
1244 list of loaded providers, their names, version and status. It optionally
1245 displays their gettable parameters.
1249 * Removed EVP_PKEY_set_alias_type().
1253 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1254 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1258 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1259 parameters are still checked for type safety at compile time via helper
1264 * Remove the RAND_DRBG API
1266 *Paul Dale and Matthias St. Pierre*
1268 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1269 as well as actual hostnames.
1273 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1274 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1275 conversely, silently ignore DTLS protocol version bounds when configuring
1276 TLS-based contexts. The commands can be repeated to set bounds of both
1277 types. The same applies with the corresponding "min_protocol" and
1278 "max_protocol" command-line switches, in case some application uses both TLS
1281 SSL_CTX instances that are created for a fixed protocol version (e.g.
1282 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1283 attempts to apply bounds to these protocol versions would result in an
1284 error. Now only the "version-flexible" SSL_CTX instances are subject to
1285 limits in configuration files in command-line options.
1289 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1294 * Reworked the recorded ERR codes to make better space for system errors.
1295 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1296 given code is a system error (true) or an OpenSSL error (false).
1300 * Reworked the test perl framework to better allow parallel testing.
1302 *Nicola Tuveri and David von Oheimb*
1304 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1305 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1309 * 'Configure' has been changed to figure out the configuration target if
1310 none is given on the command line. Consequently, the 'config' script is
1311 now only a mere wrapper. All documentation is changed to only mention
1314 *Rich Salz and Richard Levitte*
1316 * Added a library context `OSSL_LIB_CTX` that applications as well as
1317 other libraries can use to form a separate context within which
1318 libcrypto operations are performed.
1322 * Added various `_ex` functions to the OpenSSL API that support using
1323 a non-default `OSSL_LIB_CTX`.
1327 * Handshake now fails if Extended Master Secret extension is dropped
1332 * Dropped interactive mode from the `openssl` program.
1336 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1338 *David von Oheimb and Shane Lontis*
1340 * Deprecated `EC_METHOD_get_field_type()`.
1344 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1345 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1346 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1350 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1354 * Add CAdES-BES signature verification support, mostly derived
1355 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1357 *Filipe Raimundo da Silva*
1359 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1363 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1364 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1368 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1372 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1373 EC_KEY_precompute_mult().
1377 * Deprecated EC_POINTs_mul().
1381 * Removed FIPS_mode() and FIPS_mode_set().
1385 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1389 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1390 EC_POINT_get_Jprojective_coordinates_GFp().
1394 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1395 arrays to be more easily constructed via a series of utility functions.
1396 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1397 the various push functions and finally convert to a passable OSSL_PARAM
1398 array using OSSL_PARAM_BLD_to_param().
1402 * The security strength of SHA1 and MD5 based signatures in TLS has been
1407 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1408 contain a provider side internal key.
1412 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1416 * Project text documents not yet having a proper file name extension
1417 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1418 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1422 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1423 have been converted to Markdown with the goal to produce documents
1424 which not only look pretty when viewed online in the browser, but
1425 remain well readable inside a plain text editor.
1427 To achieve this goal, a 'minimalistic' Markdown style has been applied
1428 which avoids formatting elements that interfere too much with the
1429 reading flow in the text file. For example, it
1431 * avoids [ATX headings][] and uses [setext headings][] instead
1432 (which works for `<h1>` and `<h2>` headings only).
1433 * avoids [inline links][] and uses [reference links][] instead.
1434 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1436 [ATX headings]: https://github.github.com/gfm/#atx-headings
1437 [setext headings]: https://github.github.com/gfm/#setext-headings
1438 [inline links]: https://github.github.com/gfm/#inline-link
1439 [reference links]: https://github.github.com/gfm/#reference-link
1440 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1441 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1443 *Matthias St. Pierre*
1445 * The test suite is changed to preserve results of each test recipe.
1446 A new directory test-runs/ with subdirectories named like the
1447 test recipes are created in the build tree for this purpose.
1451 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1452 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1453 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1455 *David von Oheimb, Martin Peylo*
1457 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1458 It supports arbitrary request and response content types, GET redirection,
1459 TLS, connections via HTTP(S) proxies, connections and exchange via
1460 user-defined BIOs (allowing implicit connections), persistent connections,
1461 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1462 The legacy OCSP-focused (and only partly documented) API
1463 is retained for backward compatibility, while most of it is deprecated.
1467 * Added `util/check-format.pl`, a tool for checking adherence to the
1468 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1469 The checks performed are incomplete and yield some false positives.
1470 Still the tool should be useful for detecting most typical glitches.
1474 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1475 If domain name resolution yields multiple IP addresses all of them are tried
1476 after `connect()` failures.
1480 * All of the low-level RSA functions have been deprecated.
1484 * X509 certificates signed using SHA1 are no longer allowed at security
1489 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1490 modified to use PKEY APIs. These commands are now in maintenance mode
1491 and no new features will be added to them.
1495 * The command line utility rsautl has been deprecated.
1499 * The command line utilities genrsa and rsa have been modified to use PKEY
1500 APIs. They now write PKCS#8 keys by default. These commands are now in
1501 maintenance mode and no new features will be added to them.
1505 * All of the low-level DH functions have been deprecated.
1507 *Paul Dale and Matt Caswell*
1509 * All of the low-level DSA functions have been deprecated.
1513 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1514 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1518 * Deprecated low-level ECDH and ECDSA functions.
1522 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1526 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1527 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1528 a new formulation to include all the things it can be used for,
1529 as well as words of caution.
1533 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1537 * All of the low-level HMAC functions have been deprecated.
1539 *Paul Dale and David von Oheimb*
1541 * Over two thousand fixes were made to the documentation, including:
1542 - Common options (such as -rand/-writerand, TLS version control, etc)
1543 were refactored and point to newly-enhanced descriptions in openssl.pod.
1544 - Added style conformance for all options (with help from Richard Levitte),
1545 documented all reported missing options, added a CI build to check
1546 that all options are documented and that no unimplemented options
1548 - Documented some internals, such as all use of environment variables.
1549 - Addressed all internal broken L<> references.
1553 * All of the low-level CMAC functions have been deprecated.
1557 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1558 functions have been deprecated.
1560 *Paul Dale and David von Oheimb*
1562 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1563 set of functions. The documentation mentioned negative values for some
1564 errors, but this was never the case, so the mention of negative values
1567 Code that followed the documentation and thereby check with something
1568 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1572 * All of the low-level cipher functions have been deprecated.
1574 *Matt Caswell and Paul Dale*
1576 * Removed include/openssl/opensslconf.h.in and replaced it with
1577 include/openssl/configuration.h.in, which differs in not including
1578 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1579 was added to include both.
1581 This allows internal hacks where one might need to modify the set
1582 of configured macros, for example this if deprecated symbols are
1583 still supposed to be available internally:
1585 #include <openssl/configuration.h>
1587 #undef OPENSSL_NO_DEPRECATED
1588 #define OPENSSL_SUPPRESS_DEPRECATED
1590 #include <openssl/macros.h>
1592 This should not be used by applications that use the exported
1593 symbols, as that will lead to linking errors.
1597 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1598 used in exponentiation with 512-bit moduli. No EC algorithms are
1599 affected. Analysis suggests that attacks against 2-prime RSA1024,
1600 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1601 difficult to perform and are not believed likely. Attacks against DH512
1602 are considered just feasible. However, for an attack the target would
1603 have to re-use the DH512 private key, which is not recommended anyway.
1604 Also applications directly using the low-level API BN_mod_exp may be
1605 affected if they use BN_FLG_CONSTTIME.
1610 * Most memory-debug features have been deprecated, and the functionality
1611 replaced with no-ops.
1615 * Added documentation for the STACK API.
1619 * Introduced a new method type and API, OSSL_ENCODER, to represent
1620 generic encoders. These do the same sort of job that PEM writers
1621 and d2i functions do, but with support for methods supplied by
1622 providers, and the possibility for providers to support other
1627 * Introduced a new method type and API, OSSL_DECODER, to represent
1628 generic decoders. These do the same sort of job that PEM readers
1629 and i2d functions do, but with support for methods supplied by
1630 providers, and the possibility for providers to support other
1635 * Added a .pragma directive to the syntax of configuration files, to
1636 allow varying behavior in a supported and predictable manner.
1637 Currently added pragma:
1641 This allows dollar signs to be a keyword character unless it's
1642 followed by a opening brace or parenthesis. This is useful for
1643 platforms where dollar signs are commonly used in names, such as
1644 volume names and system directory names on VMS.
1648 * Added functionality to create an EVP_PKEY from user data.
1652 * Change the interpretation of the '--api' configuration option to
1653 mean that this is a desired API compatibility level with no
1654 further meaning. The previous interpretation, that this would
1655 also mean to remove all deprecated symbols up to and including
1656 the given version, no requires that 'no-deprecated' is also used
1657 in the configuration.
1659 When building applications, the desired API compatibility level
1660 can be set with the OPENSSL_API_COMPAT macro like before. For
1661 API compatibility version below 3.0, the old style numerical
1662 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1663 For version 3.0 and on, the value is expected to be the decimal
1664 value calculated from the major and minor version like this:
1666 MAJOR * 10000 + MINOR * 100
1670 -DOPENSSL_API_COMPAT=30000 For 3.0
1671 -DOPENSSL_API_COMPAT=30200 For 3.2
1673 To hide declarations that are deprecated up to and including the
1674 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1675 given when building the application as well.
1679 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1680 access to certificate and CRL stores via URIs and OSSL_STORE
1683 This adds the following functions:
1685 - X509_LOOKUP_store()
1686 - X509_STORE_load_file()
1687 - X509_STORE_load_path()
1688 - X509_STORE_load_store()
1689 - SSL_add_store_cert_subjects_to_stack()
1690 - SSL_CTX_set_default_verify_store()
1691 - SSL_CTX_load_verify_file()
1692 - SSL_CTX_load_verify_dir()
1693 - SSL_CTX_load_verify_store()
1697 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1698 The presence of this system service is determined at run-time.
1702 * Added functionality to create an EVP_PKEY context based on data
1703 for methods from providers. This takes an algorithm name and a
1704 property query string and simply stores them, with the intent
1705 that any operation that uses this context will use those strings
1706 to fetch the needed methods implicitly, thereby making the port
1707 of application written for pre-3.0 OpenSSL easier.
1711 * The undocumented function NCONF_WIN32() has been deprecated; for
1712 conversion details see the HISTORY section of doc/man5/config.pod
1716 * Introduced the new functions EVP_DigestSignInit_ex() and
1717 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1718 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1719 pages for further details.
1723 * Over two thousand fixes were made to the documentation, including:
1724 adding missing command flags, better style conformance, documentation
1727 *Rich Salz, Richard Levitte*
1729 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1730 X25519, X448, Ed25519 and Ed448.
1734 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1739 * Deprecated the public definition of `ERR_STATE` as well as the function
1740 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1745 * Added ERR functionality to give callers access to the stored function
1746 names that have replaced the older function code based functions.
1748 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1749 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1750 ERR_peek_error_all() and ERR_peek_last_error_all().
1752 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1753 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1754 ERR_func_error_string().
1758 * Extended testing to be verbose for failing tests only. The make variables
1759 VERBOSE_FAILURE or VF can be used to enable this:
1761 $ make VF=1 test # Unix
1762 $ mms /macro=(VF=1) test ! OpenVMS
1763 $ nmake VF=1 test # Windows
1767 * Added the `-copy_extensions` option to the `x509` command for use with
1768 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1769 all extensions in the request are copied to the certificate or vice versa.
1771 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1773 * Added the `-copy_extensions` option to the `req` command for use with
1774 `-x509`. When given with the `copy` or `copyall` argument,
1775 all extensions in the certification request are copied to the certificate.
1779 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1780 they generate are by default RFC 5280 compliant in the following sense:
1781 There is a subjectKeyIdentifier extension with a hash value of the public key
1782 and for not self-signed certs there is an authorityKeyIdentifier extension
1783 with a keyIdentifier field or issuer information identifying the signing key.
1784 This is done unless some configuration overrides the new default behavior,
1785 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1789 * Added several checks to `X509_verify_cert()` according to requirements in
1790 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1791 (which may be done by using the CLI option `-x509_strict`):
1792 * The basicConstraints of CA certificates must be marked critical.
1793 * CA certificates must explicitly include the keyUsage extension.
1794 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1795 * The issuer name of any certificate must not be empty.
1796 * The subject name of CA certs, certs with keyUsage crlSign,
1797 and certs without subjectAlternativeName must not be empty.
1798 * If a subjectAlternativeName extension is given it must not be empty.
1799 * The signatureAlgorithm field and the cert signature must be consistent.
1800 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1801 must not be marked critical.
1802 * The authorityKeyIdentifier must be given for X.509v3 certs
1803 unless they are self-signed.
1804 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1808 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1809 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1813 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1814 used even when parsing explicit parameters, when loading a encoded key
1815 or calling `EC_GROUP_new_from_ecpkparameters()`/
1816 `EC_GROUP_new_from_ecparameters()`.
1817 This prevents bypass of security hardening and performance gains,
1818 especially for curves with specialized EC_METHODs.
1819 By default, if a key encoded with explicit parameters is loaded and later
1820 encoded, the output is still encoded with explicit parameters, even if
1821 internally a "named" EC_GROUP is used for computation.
1825 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1826 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1827 NULL. After this change, only the cofactor parameter can be NULL. It also
1828 does some minimal sanity checks on the passed order.
1833 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1834 An attack is simple, if the first CMS_recipientInfo is valid but the
1835 second CMS_recipientInfo is chosen ciphertext. If the second
1836 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1837 encryption key will be replaced by garbage, and the message cannot be
1838 decoded, but if the RSA decryption fails, the correct encryption key is
1839 used and the recipient will not notice the attack.
1840 As a work around for this potential attack the length of the decrypted
1841 key must be equal to the cipher default key length, in case the
1842 certificate is not given and all recipientInfo are tried out.
1843 The old behaviour can be re-enabled in the CMS code by setting the
1844 CMS_DEBUG_DECRYPT flag.
1848 * Early start up entropy quality from the DEVRANDOM seed source has been
1849 improved for older Linux systems. The RAND subsystem will wait for
1850 /dev/random to be producing output before seeding from /dev/urandom.
1851 The seeded state is stored for future library initialisations using
1852 a system global shared memory segment. The shared memory identifier
1853 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1854 the desired value. The default identifier is 114.
1858 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1859 when primes for RSA keys are computed.
1860 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1861 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1862 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
1863 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1864 This avoids possible fingerprinting of newly generated RSA modules.
1868 * Correct the extended master secret constant on EBCDIC systems. Without this
1869 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1870 negotiate EMS will fail. Unfortunately this also means that TLS connections
1871 between EBCDIC systems with this fix, and EBCDIC systems without this
1872 fix will fail if they negotiate EMS.
1876 * Changed the library initialisation so that the config file is now loaded
1877 by default. This was already the case for libssl. It now occurs for both
1878 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1879 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1883 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1884 where the former acts as a replacement for `ERR_put_error()`, and the
1885 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1886 `ERR_raise_data()` adds more flexibility by taking a format string and
1887 an arbitrary number of arguments following it, to be processed with
1892 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1893 to check if a named provider is loaded and available. When called, it
1894 will also activate all fallback providers if such are still present.
1898 * Enforce a minimum DH modulus size of 512 bits.
1902 * Changed DH parameters to generate the order q subgroup instead of 2q.
1903 Previously generated DH parameters are still accepted by DH_check
1904 but DH_generate_key works around that by clearing bit 0 of the
1905 private key for those. This avoids leaking bit 0 of the private key.
1909 * Significantly reduce secure memory usage by the randomness pools.
1913 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1918 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1919 algorithms. An implementation of a key exchange algorithm can be obtained
1920 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1921 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1922 the older EVP_PKEY_derive_init() function. See the man pages for the new
1923 functions for further details.
1927 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1931 * Removed the function names from error messages and deprecated the
1936 * Removed NextStep support and the macro OPENSSL_UNISTD
1940 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1941 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1942 Also removed "export var as function" capability; we do not export
1943 variables, only functions.
1947 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1948 an error and 1 indicating success. In previous versions of OpenSSL this
1949 was a void type. If a key was set longer than the maximum possible this
1954 * Support SM2 signing and verification schemes with X509 certificate.
1958 * Use SHA256 as the default digest for TS query in the `ts` app.
1962 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1966 * Default cipher lists/suites are now available via a function, the
1967 #defines are deprecated.
1971 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1972 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1973 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1977 * Join the directories crypto/x509 and crypto/x509v3
1981 * Added command 'openssl kdf' that uses the EVP_KDF API.
1985 * Added command 'openssl mac' that uses the EVP_MAC API.
1989 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1990 as default directories. Also added the command 'openssl info'
1991 for scripting purposes.
1995 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2000 * Add prediction resistance to the DRBG reseeding process.
2004 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2005 mandated by IEEE Std 1619-2018.
2009 * Added newline escaping functionality to a filename when using openssl dgst.
2010 This output format is to replicate the output format found in the `*sum`
2011 checksum programs. This aims to preserve backward compatibility.
2013 *Matt Eaton, Richard Levitte, and Paul Dale*
2015 * Removed the heartbeat message in DTLS feature, as it has very
2016 little usage and doesn't seem to fulfill a valuable purpose.
2017 The configuration option is now deprecated.
2021 * Changed the output of 'openssl {digestname} < file' to display the
2022 digest name in its output.
2026 * Added a new generic trace API which provides support for enabling
2027 instrumentation through trace output.
2029 *Richard Levitte & Matthias St. Pierre*
2031 * Added build tests for C++. These are generated files that only do one
2032 thing, to include one public OpenSSL head file each. This tests that
2033 the public header files can be usefully included in a C++ application.
2035 This test isn't enabled by default. It can be enabled with the option
2036 'enable-buildtest-c++'.
2040 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2044 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2048 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2052 * Added KMAC to EVP_MAC.
2056 * Added property based algorithm implementation selection framework to
2061 * Added SCA hardening for modular field inversion in EC_GROUP through
2062 a new dedicated field_inv() pointer in EC_METHOD.
2063 This also addresses a leakage affecting conversions from projective
2064 to affine coordinates.
2066 *Billy Bob Brumley, Nicola Tuveri*
2068 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2069 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2070 those algorithms that were already supported through the EVP_PKEY API
2071 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2072 and scrypt are now wrappers that call EVP_KDF.
2076 * Build devcrypto engine as a dynamic engine.
2078 *Eneas U de Queiroz*
2080 * Add keyed BLAKE2 to EVP_MAC.
2084 * Fix a bug in the computation of the endpoint-pair shared secret used
2085 by DTLS over SCTP. This breaks interoperability with older versions
2086 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2087 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2088 interoperability with such broken implementations. However, enabling
2089 this switch breaks interoperability with correct implementations.
2091 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2092 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2096 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2100 * Changed the license to the Apache License v2.0.
2104 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2106 - Major releases (indicated by incrementing the MAJOR release number)
2107 may introduce incompatible API/ABI changes.
2108 - Minor releases (indicated by incrementing the MINOR release number)
2109 may introduce new features but retain API/ABI compatibility.
2110 - Patch releases (indicated by incrementing the PATCH number)
2111 are intended for bug fixes and other improvements of existing
2112 features only (like improving performance or adding documentation)
2113 and retain API/ABI compatibility.
2117 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2121 * Remove the 'dist' target and add a tarball building script. The
2122 'dist' target has fallen out of use, and it shouldn't be
2123 necessary to configure just to create a source distribution.
2127 * Recreate the OS390-Unix config target. It no longer relies on a
2128 special script like it did for OpenSSL pre-1.1.0.
2132 * Instead of having the source directories listed in Configure, add
2133 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2138 * Add GMAC to EVP_MAC.
2142 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2146 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2147 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2148 to facilitate the continued use of MACs through raw private keys in
2149 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2153 * Deprecate ECDH_KDF_X9_62().
2157 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2158 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2159 are retained for backwards compatibility.
2163 * AES-XTS mode now enforces that its two keys are different to mitigate
2164 the attacked described in "Efficient Instantiations of Tweakable
2165 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2166 Details of this attack can be obtained from:
2167 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2171 * Rename the object files, i.e. give them other names than in previous
2172 versions. Their names now include the name of the final product, as
2173 well as its type mnemonic (bin, lib, shlib).
2177 * Added new option for 'openssl list', '-objects', which will display the
2178 list of built in objects, i.e. OIDs with names.
2182 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2183 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2188 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2189 improves application performance by removing data copies and providing
2190 applications with zero-copy system calls such as sendfile and splice.
2194 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2198 * `PKCS12_parse` now maintains the order of the parsed certificates
2199 when outputting them via `*ca` (rather than reversing it).
2203 * Deprecated pthread fork support methods.
2207 * Added support for FFDHE key exchange in TLS 1.3.
2211 * Added a new concept for OpenSSL plugability: providers. This
2212 functionality is designed to replace the ENGINE API and ENGINE
2213 implementations, and to be much more dynamic, allowing provider
2214 authors to introduce new algorithms among other things, as long as
2215 there's an API that supports the algorithm type.
2217 With this concept comes a new core API for interaction between
2218 libcrypto and provider implementations. Public libcrypto functions
2219 that want to use providers do so through this core API.
2221 The main documentation for this core API is found in
2222 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2223 refer to other manuals describing the API specific for supported
2224 algorithm types (also called operations).
2231 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2233 * Avoid loading of a dynamic engine twice.
2237 * Prioritise DANE TLSA issuer certs over peer certs
2241 * Fixed random API for MacOS prior to 10.12
2243 These MacOS versions don't support the CommonCrypto APIs
2247 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2249 * Fixed an SM2 Decryption Buffer Overflow.
2251 In order to decrypt SM2 encrypted data an application is expected to
2252 call the API function EVP_PKEY_decrypt(). Typically an application will
2253 call this function twice. The first time, on entry, the "out" parameter
2254 can be NULL and, on exit, the "outlen" parameter is populated with the
2255 buffer size required to hold the decrypted plaintext. The application
2256 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2257 again, but this time passing a non-NULL value for the "out" parameter.
2259 A bug in the implementation of the SM2 decryption code means that the
2260 calculation of the buffer size required to hold the plaintext returned
2261 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2262 size required by the second call. This can lead to a buffer overflow
2263 when EVP_PKEY_decrypt() is called by the application a second time with
2264 a buffer that is too small.
2266 A malicious attacker who is able present SM2 content for decryption to
2267 an application could cause attacker chosen data to overflow the buffer
2268 by up to a maximum of 62 bytes altering the contents of other data held
2269 after the buffer, possibly changing application behaviour or causing
2270 the application to crash. The location of the buffer is application
2271 dependent but is typically heap allocated.
2276 * Fixed various read buffer overruns processing ASN.1 strings
2278 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2279 structure which contains a buffer holding the string data and a field
2280 holding the buffer length. This contrasts with normal C strings which
2281 are represented as a buffer for the string data which is terminated
2282 with a NUL (0) byte.
2284 Although not a strict requirement, ASN.1 strings that are parsed using
2285 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2286 well as any string whose value has been set with the ASN1_STRING_set()
2287 function will additionally NUL terminate the byte array in the
2288 ASN1_STRING structure.
2290 However, it is possible for applications to directly construct valid
2291 ASN1_STRING structures which do not NUL terminate the byte array by
2292 directly setting the "data" and "length" fields in the ASN1_STRING
2293 array. This can also happen by using the ASN1_STRING_set0() function.
2295 Numerous OpenSSL functions that print ASN.1 data have been found to
2296 assume that the ASN1_STRING byte array will be NUL terminated, even
2297 though this is not guaranteed for strings that have been directly
2298 constructed. Where an application requests an ASN.1 structure to be
2299 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2300 been directly constructed by the application without NUL terminating
2301 the "data" field, then a read buffer overrun can occur.
2303 The same thing can also occur during name constraints processing
2304 of certificates (for example if a certificate has been directly
2305 constructed by the application instead of loading it via the OpenSSL
2306 parsing functions, and the certificate contains non NUL terminated
2307 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2308 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2310 If a malicious actor can cause an application to directly construct an
2311 ASN1_STRING and then process it through one of the affected OpenSSL
2312 functions then this issue could be hit. This might result in a crash
2313 (causing a Denial of Service attack). It could also result in the
2314 disclosure of private memory contents (such as private keys, or
2315 sensitive plaintext).
2320 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2322 * Fixed a problem with verifying a certificate chain when using the
2323 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2324 the certificates present in a certificate chain. It is not set by default.
2326 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2327 the chain that have explicitly encoded elliptic curve parameters was added
2328 as an additional strict check.
2330 An error in the implementation of this check meant that the result of a
2331 previous check to confirm that certificates in the chain are valid CA
2332 certificates was overwritten. This effectively bypasses the check
2333 that non-CA certificates must not be able to issue other certificates.
2335 If a "purpose" has been configured then there is a subsequent opportunity
2336 for checks that the certificate is a valid CA. All of the named "purpose"
2337 values implemented in libcrypto perform this check. Therefore, where
2338 a purpose is set the certificate chain will still be rejected even when the
2339 strict flag has been used. A purpose is set by default in libssl client and
2340 server certificate verification routines, but it can be overridden or
2341 removed by an application.
2343 In order to be affected, an application must explicitly set the
2344 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2345 for the certificate verification or, in the case of TLS client or server
2346 applications, override the default purpose.
2351 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2352 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2353 renegotiation ClientHello omits the signature_algorithms extension (where it
2354 was present in the initial ClientHello), but includes a
2355 signature_algorithms_cert extension then a NULL pointer dereference will
2356 result, leading to a crash and a denial of service attack.
2358 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2359 (which is the default configuration). OpenSSL TLS clients are not impacted by
2363 *Peter Kästle and Samuel Sapalski*
2365 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2367 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2368 create a unique hash value based on the issuer and serial number data
2369 contained within an X509 certificate. However, it was failing to correctly
2370 handle any errors that may occur while parsing the issuer field (which might
2371 occur if the issuer field is maliciously constructed). This may subsequently
2372 result in a NULL pointer deref and a crash leading to a potential denial of
2378 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2379 padding mode to correctly check for rollback attacks. This is considered a
2380 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2385 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2386 functions. Previously they could overflow the output length argument in some
2387 cases where the input length is close to the maximum permissible length for
2388 an integer on the platform. In such cases the return value from the function
2389 call would be 1 (indicating success), but the output length value would be
2390 negative. This could cause applications to behave incorrectly or crash.
2395 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2396 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2397 could be exploited in a side channel attack to recover the password. Since
2398 the attack is local host only this is outside of the current OpenSSL
2399 threat model and therefore no CVE is assigned.
2401 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2406 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2408 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2409 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2410 If an attacker can control both items being compared then this could lead
2411 to a possible denial of service attack. OpenSSL itself uses the
2412 GENERAL_NAME_cmp function for two purposes:
2413 1) Comparing CRL distribution point names between an available CRL and a
2414 CRL distribution point embedded in an X509 certificate
2415 2) When verifying that a timestamp response token signer matches the
2416 timestamp authority name (exposed via the API functions
2417 TS_RESP_verify_response and TS_RESP_verify_token)
2422 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2424 * Certificates with explicit curve parameters are now disallowed in
2425 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2429 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2430 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2431 conversely, silently ignore DTLS protocol version bounds when configuring
2432 TLS-based contexts. The commands can be repeated to set bounds of both
2433 types. The same applies with the corresponding "min_protocol" and
2434 "max_protocol" command-line switches, in case some application uses both TLS
2437 SSL_CTX instances that are created for a fixed protocol version (e.g.
2438 TLSv1_server_method()) also silently ignore version bounds. Previously
2439 attempts to apply bounds to these protocol versions would result in an
2440 error. Now only the "version-flexible" SSL_CTX instances are subject to
2441 limits in configuration files in command-line options.
2445 * Handshake now fails if Extended Master Secret extension is dropped
2450 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2452 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2454 * Fixed segmentation fault in SSL_check_chain()
2455 Server or client applications that call the SSL_check_chain() function
2456 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2457 dereference as a result of incorrect handling of the
2458 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2459 or unrecognised signature algorithm is received from the peer. This could
2460 be exploited by a malicious peer in a Denial of Service attack.
2465 * Added AES consttime code for no-asm configurations
2466 an optional constant time support for AES was added
2467 when building openssl for no-asm.
2468 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2469 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2470 At this time this feature is by default disabled.
2471 It will be enabled by default in 3.0.
2475 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2477 * Revert the change of EOF detection while reading in libssl to avoid
2478 regressions in applications depending on the current way of reporting
2479 the EOF. As the existing method is not fully accurate the change to
2480 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2481 branch and will be present in the 3.0 release.
2485 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2486 when primes for RSA keys are computed.
2487 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2488 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2489 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2490 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2491 This avoids possible fingerprinting of newly generated RSA modules.
2495 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2497 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2498 while reading in libssl then we would report an error back to the
2499 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2500 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2501 therefore give a hint as to what went wrong.
2505 * Check that ed25519 and ed448 are allowed by the security level. Previously
2506 signature algorithms not using an MD were not being checked that they were
2507 allowed by the security level.
2511 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2512 was not quite right. The behaviour was not consistent between resumption
2513 and normal handshakes, and also not quite consistent with historical
2514 behaviour. The behaviour in various scenarios has been clarified and
2515 it has been updated to make it match historical behaviour as closely as
2520 * *[VMS only]* The header files that the VMS compilers include automatically,
2521 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2522 that the C++ compiler doesn't understand. This is a shortcoming in the
2523 compiler, but can be worked around with `__cplusplus` guards.
2525 C++ applications that use OpenSSL libraries must be compiled using the
2526 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2527 functions. Otherwise, only functions with symbols of less than 31
2528 characters can be used, as the linker will not be able to successfully
2529 resolve symbols with longer names.
2533 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2534 The presence of this system service is determined at run-time.
2538 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2543 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2545 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2546 number generator (RNG). This was intended to include protection in the
2547 event of a fork() system call in order to ensure that the parent and child
2548 processes did not share the same RNG state. However, this protection was not
2549 being used in the default case.
2551 A partial mitigation for this issue is that the output from a high
2552 precision timer is mixed into the RNG state so the likelihood of a parent
2553 and child process sharing state is significantly reduced.
2555 If an application already calls OPENSSL_init_crypto() explicitly using
2556 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2559 *Matthias St. Pierre*
2561 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2562 used even when parsing explicit parameters, when loading a encoded key
2563 or calling `EC_GROUP_new_from_ecpkparameters()`/
2564 `EC_GROUP_new_from_ecparameters()`.
2565 This prevents bypass of security hardening and performance gains,
2566 especially for curves with specialized EC_METHODs.
2567 By default, if a key encoded with explicit parameters is loaded and later
2568 encoded, the output is still encoded with explicit parameters, even if
2569 internally a "named" EC_GROUP is used for computation.
2573 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2574 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2575 NULL. After this change, only the cofactor parameter can be NULL. It also
2576 does some minimal sanity checks on the passed order.
2581 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2582 An attack is simple, if the first CMS_recipientInfo is valid but the
2583 second CMS_recipientInfo is chosen ciphertext. If the second
2584 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2585 encryption key will be replaced by garbage, and the message cannot be
2586 decoded, but if the RSA decryption fails, the correct encryption key is
2587 used and the recipient will not notice the attack.
2588 As a work around for this potential attack the length of the decrypted
2589 key must be equal to the cipher default key length, in case the
2590 certificate is not given and all recipientInfo are tried out.
2591 The old behaviour can be re-enabled in the CMS code by setting the
2592 CMS_DEBUG_DECRYPT flag.
2597 * Early start up entropy quality from the DEVRANDOM seed source has been
2598 improved for older Linux systems. The RAND subsystem will wait for
2599 /dev/random to be producing output before seeding from /dev/urandom.
2600 The seeded state is stored for future library initialisations using
2601 a system global shared memory segment. The shared memory identifier
2602 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2603 the desired value. The default identifier is 114.
2607 * Correct the extended master secret constant on EBCDIC systems. Without this
2608 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2609 negotiate EMS will fail. Unfortunately this also means that TLS connections
2610 between EBCDIC systems with this fix, and EBCDIC systems without this
2611 fix will fail if they negotiate EMS.
2615 * Use Windows installation paths in the mingw builds
2617 Mingw isn't a POSIX environment per se, which means that Windows
2618 paths should be used for installation.
2623 * Changed DH_check to accept parameters with order q and 2q subgroups.
2624 With order 2q subgroups the bit 0 of the private key is not secret
2625 but DH_generate_key works around that by clearing bit 0 of the
2626 private key for those. This avoids leaking bit 0 of the private key.
2630 * Significantly reduce secure memory usage by the randomness pools.
2634 * Revert the DEVRANDOM_WAIT feature for Linux systems
2636 The DEVRANDOM_WAIT feature added a select() call to wait for the
2637 /dev/random device to become readable before reading from the
2638 /dev/urandom device.
2640 It turned out that this change had negative side effects on
2641 performance which were not acceptable. After some discussion it
2642 was decided to revert this feature and leave it up to the OS
2643 resp. the platform maintainer to ensure a proper initialization
2644 during early boot time.
2646 *Matthias St. Pierre*
2648 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2650 * Add build tests for C++. These are generated files that only do one
2651 thing, to include one public OpenSSL head file each. This tests that
2652 the public header files can be usefully included in a C++ application.
2654 This test isn't enabled by default. It can be enabled with the option
2655 'enable-buildtest-c++'.
2659 * Enable SHA3 pre-hashing for ECDSA and DSA.
2663 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2664 This changes the size when using the `genpkey` command when no size is given.
2665 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2666 generation commands to use 2048 bits by default.
2670 * Reorganize the manual pages to consistently have RETURN VALUES,
2671 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2672 util/fix-doc-nits accordingly.
2674 *Paul Yang, Joshua Lock*
2676 * Add the missing accessor EVP_PKEY_get0_engine()
2680 * Have commands like `s_client` and `s_server` output the signature scheme
2681 along with other cipher suite parameters when debugging.
2685 * Make OPENSSL_config() error agnostic again.
2689 * Do the error handling in RSA decryption constant time.
2693 * Prevent over long nonces in ChaCha20-Poly1305.
2695 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2696 for every encryption operation. RFC 7539 specifies that the nonce value
2697 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2698 and front pads the nonce with 0 bytes if it is less than 12
2699 bytes. However it also incorrectly allows a nonce to be set of up to 16
2700 bytes. In this case only the last 12 bytes are significant and any
2701 additional leading bytes are ignored.
2703 It is a requirement of using this cipher that nonce values are
2704 unique. Messages encrypted using a reused nonce value are susceptible to
2705 serious confidentiality and integrity attacks. If an application changes
2706 the default nonce length to be longer than 12 bytes and then makes a
2707 change to the leading bytes of the nonce expecting the new value to be a
2708 new unique nonce then such an application could inadvertently encrypt
2709 messages with a reused nonce.
2711 Additionally the ignored bytes in a long nonce are not covered by the
2712 integrity guarantee of this cipher. Any application that relies on the
2713 integrity of these ignored leading bytes of a long nonce may be further
2714 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2715 is safe because no such use sets such a long nonce value. However user
2716 applications that use this cipher directly and set a non-default nonce
2717 length to be longer than 12 bytes may be vulnerable.
2719 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2725 * Add DEVRANDOM_WAIT feature for Linux systems
2727 On older Linux systems where the getrandom() system call is not available,
2728 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2729 Contrary to getrandom(), the /dev/urandom device will not block during
2730 early boot when the kernel CSPRNG has not been seeded yet.
2732 To mitigate this known weakness, use select() to wait for /dev/random to
2733 become readable before reading from /dev/urandom.
2735 * Ensure that SM2 only uses SM3 as digest algorithm
2739 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2741 * Change the info callback signals for the start and end of a post-handshake
2742 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2743 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2744 confused by this and assume that a TLSv1.2 renegotiation has started. This
2745 can break KeyUpdate handling. Instead we no longer signal the start and end
2746 of a post handshake message exchange (although the messages themselves are
2747 still signalled). This could break some applications that were expecting
2748 the old signals. However without this KeyUpdate is not usable for many
2753 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2755 * Timing vulnerability in DSA signature generation
2757 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2758 timing side channel attack. An attacker could use variations in the signing
2759 algorithm to recover the private key.
2761 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2766 * Timing vulnerability in ECDSA signature generation
2768 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2769 timing side channel attack. An attacker could use variations in the signing
2770 algorithm to recover the private key.
2772 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2777 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2778 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2779 of two gigabytes and the error handling improved.
2781 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2782 categorized as a normal bug, not a security issue, because the DRBG reseeds
2783 automatically and is fully functional even without additional randomness
2784 provided by the application.
2786 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2788 * Add a new ClientHello callback. Provides a callback interface that gives
2789 the application the ability to adjust the nascent SSL object at the
2790 earliest stage of ClientHello processing, immediately after extensions have
2791 been collected but before they have been processed. In particular, this
2792 callback can adjust the supported TLS versions in response to the contents
2797 * Add SM2 base algorithm support.
2801 * s390x assembly pack: add (improved) hardware-support for the following
2802 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2803 aes-cfb/cfb8, aes-ecb.
2807 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2808 parameter is no longer accepted, as it leads to a corrupt table. NULL
2809 pem_str is reserved for alias entries only.
2813 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2814 step for prime curves. The new implementation is based on formulae from
2815 differential addition-and-doubling in homogeneous projective coordinates
2816 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2817 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2818 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2819 to work in projective coordinates.
2821 *Billy Bob Brumley, Nicola Tuveri*
2823 * Change generating and checking of primes so that the error rate of not
2824 being prime depends on the intended use based on the size of the input.
2825 For larger primes this will result in more rounds of Miller-Rabin.
2826 The maximal error rate for primes with more than 1080 bits is lowered
2829 *Kurt Roeckx, Annie Yousar*
2831 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2835 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2836 moving between systems, and to avoid confusion when a Windows build is
2837 done with mingw vs with MSVC. For POSIX installs, there's still a
2838 symlink or copy named 'tsget' to avoid that confusion as well.
2842 * Revert blinding in ECDSA sign and instead make problematic addition
2843 length-invariant. Switch even to fixed-length Montgomery multiplication.
2847 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2848 step for binary curves. The new implementation is based on formulae from
2849 differential addition-and-doubling in mixed Lopez-Dahab projective
2850 coordinates, modified to independently blind the operands.
2852 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2854 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2855 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2856 EC_METHODs to implement their own specialized "ladder step", to take
2857 advantage of more favorable coordinate systems or more efficient
2858 differential addition-and-doubling algorithms.
2860 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2862 * Modified the random device based seed sources to keep the relevant
2863 file descriptors open rather than reopening them on each access.
2864 This allows such sources to operate in a chroot() jail without
2865 the associated device nodes being available. This behaviour can be
2866 controlled using RAND_keep_random_devices_open().
2870 * Numerous side-channel attack mitigations have been applied. This may have
2871 performance impacts for some algorithms for the benefit of improved
2872 security. Specific changes are noted in this change log by their respective
2877 * AIX shared library support overhaul. Switch to AIX "natural" way of
2878 handling shared libraries, which means collecting shared objects of
2879 different versions and bitnesses in one common archive. This allows to
2880 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2881 doesn't affect the way 3rd party applications are linked, only how
2882 multi-version installation is managed.
2886 * Make ec_group_do_inverse_ord() more robust and available to other
2887 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2888 mitigations are applied to the fallback BN_mod_inverse().
2889 When using this function rather than BN_mod_inverse() directly, new
2890 EC cryptosystem implementations are then safer-by-default.
2894 * Add coordinate blinding for EC_POINT and implement projective
2895 coordinate blinding for generic prime curves as a countermeasure to
2896 chosen point SCA attacks.
2898 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2900 * Add blinding to ECDSA and DSA signatures to protect against side channel
2901 attacks discovered by Keegan Ryan (NCC Group).
2905 * Enforce checking in the `pkeyutl` command to ensure that the input
2906 length does not exceed the maximum supported digest length when performing
2907 a sign, verify or verifyrecover operation.
2911 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2912 I/O in combination with something like select() or poll() will hang. This
2913 can be turned off again using SSL_CTX_clear_mode().
2914 Many applications do not properly handle non-application data records, and
2915 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2916 around the problems in those applications, but can also break some.
2917 It's recommended to read the manpages about SSL_read(), SSL_write(),
2918 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2919 SSL_CTX_set_read_ahead() again.
2923 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2924 now allow empty (zero character) pass phrases.
2928 * Apply blinding to binary field modular inversion and remove patent
2929 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2933 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2934 binary and prime elliptic curves.
2938 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2939 constant time fixed point multiplication.
2943 * Revise elliptic curve scalar multiplication with timing attack
2944 defenses: ec_wNAF_mul redirects to a constant time implementation
2945 when computing fixed point and variable point multiplication (which
2946 in OpenSSL are mostly used with secret scalars in keygen, sign,
2947 ECDH derive operations).
2948 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2951 * Updated CONTRIBUTING
2955 * Updated DRBG / RAND to request nonce and additional low entropy
2956 randomness from the system.
2958 *Matthias St. Pierre*
2960 * Updated 'openssl rehash' to use OpenSSL consistent default.
2964 * Moved the load of the ssl_conf module to libcrypto, which helps
2965 loading engines that libssl uses before libssl is initialised.
2969 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2973 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2975 *Ingo Schwarze, Rich Salz*
2977 * Added output of accepting IP address and port for 'openssl s_server'
2981 * Added a new API for TLSv1.3 ciphersuites:
2982 SSL_CTX_set_ciphersuites()
2983 SSL_set_ciphersuites()
2987 * Memory allocation failures consistently add an error to the error
2992 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2993 in libcrypto when run as setuid/setgid.
2997 * Load any config file by default when libssl is used.
3001 * Added new public header file <openssl/rand_drbg.h> and documentation
3002 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3004 *Matthias St. Pierre*
3006 * QNX support removed (cannot find contributors to get their approval
3007 for the license change).
3011 * TLSv1.3 replay protection for early data has been implemented. See the
3012 SSL_read_early_data() man page for further details.
3016 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3017 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3018 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3019 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3020 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3021 configuration has been separated out. See the ciphers man page or the
3022 SSL_CTX_set_ciphersuites() man page for more information.
3026 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3027 in responder mode now supports the new "-multi" option, which
3028 spawns the specified number of child processes to handle OCSP
3029 requests. The "-timeout" option now also limits the OCSP
3030 responder's patience to wait to receive the full client request
3031 on a newly accepted connection. Child processes are respawned
3032 as needed, and the CA index file is automatically reloaded
3033 when changed. This makes it possible to run the "ocsp" responder
3034 as a long-running service, making the OpenSSL CA somewhat more
3035 feature-complete. In this mode, most diagnostic messages logged
3036 after entering the event loop are logged via syslog(3) rather than
3041 * Added support for X448 and Ed448. Heavily based on original work by
3046 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3047 objects loaded. This adds the functions OSSL_STORE_expect() and
3048 OSSL_STORE_find() as well as needed tools to construct searches and
3049 get the search data out of them.
3053 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3054 version of OpenSSL should review their configuration settings to ensure
3055 that they are still appropriate for TLSv1.3. For further information see:
3056 <https://wiki.openssl.org/index.php/TLS1.3>
3060 * Grand redesign of the OpenSSL random generator
3062 The default RAND method now utilizes an AES-CTR DRBG according to
3063 NIST standard SP 800-90Ar1. The new random generator is essentially
3064 a port of the default random generator from the OpenSSL FIPS 2.0
3065 object module. It is a hybrid deterministic random bit generator
3066 using an AES-CTR bit stream and which seeds and reseeds itself
3067 automatically using trusted system entropy sources.
3069 Some of its new features are:
3070 - Support for multiple DRBG instances with seed chaining.
3071 - The default RAND method makes use of a DRBG.
3072 - There is a public and private DRBG instance.
3073 - The DRBG instances are fork-safe.
3074 - Keep all global DRBG instances on the secure heap if it is enabled.
3075 - The public and private DRBG instance are per thread for lock free
3078 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3080 * Changed Configure so it only says what it does and doesn't dump
3081 so much data. Instead, ./configdata.pm should be used as a script
3082 to display all sorts of configuration data.
3086 * Added processing of "make variables" to Configure.
3090 * Added SHA512/224 and SHA512/256 algorithm support.
3094 * The last traces of Netware support, first removed in 1.1.0, have
3099 * Get rid of Makefile.shared, and in the process, make the processing
3100 of certain files (rc.obj, or the .def/.map/.opt files produced from
3101 the ordinal files) more visible and hopefully easier to trace and
3102 debug (or make silent).
3106 * Make it possible to have environment variable assignments as
3107 arguments to config / Configure.
3111 * Add multi-prime RSA (RFC 8017) support.
3115 * Add SM3 implemented according to GB/T 32905-2016
3116 *Jack Lloyd <jack.lloyd@ribose.com>,*
3117 *Ronald Tse <ronald.tse@ribose.com>,*
3118 *Erick Borsboom <erick.borsboom@ribose.com>*
3120 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3121 as documented in RFC6066.
3122 Based on a patch from Tomasz Moń
3124 *Filipe Raimundo da Silva*
3126 * Add SM4 implemented according to GB/T 32907-2016.
3127 *Jack Lloyd <jack.lloyd@ribose.com>,*
3128 *Ronald Tse <ronald.tse@ribose.com>,*
3129 *Erick Borsboom <erick.borsboom@ribose.com>*
3131 * Reimplement -newreq-nodes and ERR_error_string_n; the
3132 original author does not agree with the license change.
3136 * Add ARIA AEAD TLS support.
3140 * Some macro definitions to support VS6 have been removed. Visual
3141 Studio 6 has not worked since 1.1.0
3145 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3146 without clearing the errors.
3150 * Add "atfork" functions. If building on a system that without
3151 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3152 requirements. The RAND facility now uses/requires this.
3160 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3161 not possible to disable entirely. However, it's still possible to
3162 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3165 To disable, configure with 'no-ui-console'. 'no-ui' is still
3166 possible to use as an alias. Check at compile time with the
3167 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3168 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3172 * Add a STORE module, which implements a uniform and URI based reader of
3173 stores that can contain keys, certificates, CRLs and numerous other
3174 objects. The main API is loosely based on a few stdio functions,
3175 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3176 OSSL_STORE_error and OSSL_STORE_close.
3177 The implementation uses backends called "loaders" to implement arbitrary
3178 URI schemes. There is one built in "loader" for the 'file' scheme.
3182 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3183 then adjusted to work on FreeBSD 8.4 as well.
3184 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3185 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3189 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3190 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3191 error code calls like this:
3193 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3195 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3196 that can be encoded in C. For the foreseeable future, this will only
3199 *Richard Levitte and Tim Hudson*
3201 * Removed BSD cryptodev engine.
3205 * Add a build target 'build_all_generated', to build all generated files
3206 and only that. This can be used to prepare everything that requires
3207 things like perl for a system that lacks perl and then move everything
3208 to that system and do the rest of the build there.
3212 * In the UI interface, make it possible to duplicate the user data. This
3213 can be used by engines that need to retain the data for a longer time
3214 than just the call where this user data is passed.
3218 * Ignore the '-named_curve auto' value for compatibility of applications
3221 *Tomáš Mráz <tmraz@fedoraproject.org>*
3223 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3224 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3225 alerts across multiple records (some of which could be empty). In practice
3226 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3227 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3228 support this at all. Supporting it adds significant complexity to the
3229 record layer, and its removal is unlikely to cause interoperability
3234 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3235 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3236 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3241 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3242 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3244 *Richard Levitte, Andy Polyakov*
3246 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3251 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3252 platform rather than 'mingw'.
3256 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3257 success if they are asked to add an object which already exists
3258 in the store. This change cascades to other functions which load
3259 certificates and CRLs.
3263 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3264 facilitate stack unwinding even from assembly subroutines.
3268 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3269 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3273 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3274 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3275 which is the minimum version we support.
3279 * Certificate time validation (X509_cmp_time) enforces stricter
3280 compliance with RFC 5280. Fractional seconds and timezone offsets
3281 are no longer allowed.
3285 * Add support for ARIA
3289 * s_client will now send the Server Name Indication (SNI) extension by
3290 default unless the new "-noservername" option is used. The server name is
3291 based on the host provided to the "-connect" option unless overridden by
3292 using "-servername".
3296 * Add support for SipHash
3300 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3301 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3302 prevent issues where no progress is being made and the peer continually
3303 sends unrecognised record types, using up resources processing them.
3307 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3308 using the algorithm defined in
3309 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3313 * Heartbeat support has been removed; the ABI is changed for now.
3315 *Richard Levitte, Rich Salz*
3317 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3321 * The RSA "null" method, which was partially supported to avoid patent
3322 issues, has been replaced to always returns NULL.
3329 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3331 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3332 used even when parsing explicit parameters, when loading a encoded key
3333 or calling `EC_GROUP_new_from_ecpkparameters()`/
3334 `EC_GROUP_new_from_ecparameters()`.
3335 This prevents bypass of security hardening and performance gains,
3336 especially for curves with specialized EC_METHODs.
3337 By default, if a key encoded with explicit parameters is loaded and later
3338 encoded, the output is still encoded with explicit parameters, even if
3339 internally a "named" EC_GROUP is used for computation.
3343 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3344 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3345 NULL. After this change, only the cofactor parameter can be NULL. It also
3346 does some minimal sanity checks on the passed order.
3351 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3352 An attack is simple, if the first CMS_recipientInfo is valid but the
3353 second CMS_recipientInfo is chosen ciphertext. If the second
3354 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3355 encryption key will be replaced by garbage, and the message cannot be
3356 decoded, but if the RSA decryption fails, the correct encryption key is
3357 used and the recipient will not notice the attack.
3358 As a work around for this potential attack the length of the decrypted
3359 key must be equal to the cipher default key length, in case the
3360 certificate is not given and all recipientInfo are tried out.
3361 The old behaviour can be re-enabled in the CMS code by setting the
3362 CMS_DEBUG_DECRYPT flag.
3367 * Use Windows installation paths in the mingw builds
3369 Mingw isn't a POSIX environment per se, which means that Windows
3370 paths should be used for installation.
3375 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3377 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3378 This changes the size when using the `genpkey` command when no size is given.
3379 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3380 generation commands to use 2048 bits by default.
3384 * Prevent over long nonces in ChaCha20-Poly1305.
3386 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3387 for every encryption operation. RFC 7539 specifies that the nonce value
3388 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3389 and front pads the nonce with 0 bytes if it is less than 12
3390 bytes. However it also incorrectly allows a nonce to be set of up to 16
3391 bytes. In this case only the last 12 bytes are significant and any
3392 additional leading bytes are ignored.
3394 It is a requirement of using this cipher that nonce values are
3395 unique. Messages encrypted using a reused nonce value are susceptible to
3396 serious confidentiality and integrity attacks. If an application changes
3397 the default nonce length to be longer than 12 bytes and then makes a
3398 change to the leading bytes of the nonce expecting the new value to be a
3399 new unique nonce then such an application could inadvertently encrypt
3400 messages with a reused nonce.
3402 Additionally the ignored bytes in a long nonce are not covered by the
3403 integrity guarantee of this cipher. Any application that relies on the
3404 integrity of these ignored leading bytes of a long nonce may be further
3405 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3406 is safe because no such use sets such a long nonce value. However user
3407 applications that use this cipher directly and set a non-default nonce
3408 length to be longer than 12 bytes may be vulnerable.
3410 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3416 * Added SCA hardening for modular field inversion in EC_GROUP through
3417 a new dedicated field_inv() pointer in EC_METHOD.
3418 This also addresses a leakage affecting conversions from projective
3419 to affine coordinates.
3421 *Billy Bob Brumley, Nicola Tuveri*
3423 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3424 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3428 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3432 * Remove the 'dist' target and add a tarball building script. The
3433 'dist' target has fallen out of use, and it shouldn't be
3434 necessary to configure just to create a source distribution.
3438 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3440 * Timing vulnerability in DSA signature generation
3442 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3443 timing side channel attack. An attacker could use variations in the signing
3444 algorithm to recover the private key.
3446 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3451 * Timing vulnerability in ECDSA signature generation
3453 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3454 timing side channel attack. An attacker could use variations in the signing
3455 algorithm to recover the private key.
3457 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3462 * Add coordinate blinding for EC_POINT and implement projective
3463 coordinate blinding for generic prime curves as a countermeasure to
3464 chosen point SCA attacks.
3466 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3468 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3470 * Client DoS due to large DH parameter
3472 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3473 malicious server can send a very large prime value to the client. This will
3474 cause the client to spend an unreasonably long period of time generating a
3475 key for this prime resulting in a hang until the client has finished. This
3476 could be exploited in a Denial Of Service attack.
3478 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3483 * Cache timing vulnerability in RSA Key Generation
3485 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3486 a cache timing side channel attack. An attacker with sufficient access to
3487 mount cache timing attacks during the RSA key generation process could
3488 recover the private key.
3490 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3491 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3496 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3497 parameter is no longer accepted, as it leads to a corrupt table. NULL
3498 pem_str is reserved for alias entries only.
3502 * Revert blinding in ECDSA sign and instead make problematic addition
3503 length-invariant. Switch even to fixed-length Montgomery multiplication.
3507 * Change generating and checking of primes so that the error rate of not
3508 being prime depends on the intended use based on the size of the input.
3509 For larger primes this will result in more rounds of Miller-Rabin.
3510 The maximal error rate for primes with more than 1080 bits is lowered
3513 *Kurt Roeckx, Annie Yousar*
3515 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3519 * Add blinding to ECDSA and DSA signatures to protect against side channel
3520 attacks discovered by Keegan Ryan (NCC Group).
3524 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3525 now allow empty (zero character) pass phrases.
3529 * Certificate time validation (X509_cmp_time) enforces stricter
3530 compliance with RFC 5280. Fractional seconds and timezone offsets
3531 are no longer allowed.
3535 * Fixed a text canonicalisation bug in CMS
3537 Where a CMS detached signature is used with text content the text goes
3538 through a canonicalisation process first prior to signing or verifying a
3539 signature. This process strips trailing space at the end of lines, converts
3540 line terminators to CRLF and removes additional trailing line terminators
3541 at the end of a file. A bug in the canonicalisation process meant that
3542 some characters, such as form-feed, were incorrectly treated as whitespace
3543 and removed. This is contrary to the specification (RFC5485). This fix
3544 could mean that detached text data signed with an earlier version of
3545 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3546 signed with a fixed OpenSSL may fail to verify with an earlier version of
3547 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3548 and use the "-binary" flag (for the "cms" command line application) or set
3549 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3553 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3555 * Constructed ASN.1 types with a recursive definition could exceed the stack
3557 Constructed ASN.1 types with a recursive definition (such as can be found
3558 in PKCS7) could eventually exceed the stack given malicious input with
3559 excessive recursion. This could result in a Denial Of Service attack. There
3560 are no such structures used within SSL/TLS that come from untrusted sources
3561 so this is considered safe.
3563 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3569 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3571 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3572 effectively reduced to only comparing the least significant bit of each
3573 byte. This allows an attacker to forge messages that would be considered as
3574 authenticated in an amount of tries lower than that guaranteed by the
3575 security claims of the scheme. The module can only be compiled by the
3576 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3578 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3584 * Add a build target 'build_all_generated', to build all generated files
3585 and only that. This can be used to prepare everything that requires
3586 things like perl for a system that lacks perl and then move everything
3587 to that system and do the rest of the build there.
3591 * Backport SSL_OP_NO_RENGOTIATION
3593 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3594 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3595 changes this is no longer possible in 1.1.0. Therefore, the new
3596 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3597 1.1.0 to provide equivalent functionality.
3599 Note that if an application built against 1.1.0h headers (or above) is run
3600 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3601 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3605 * Removed the OS390-Unix config target. It relied on a script that doesn't
3610 * rsaz_1024_mul_avx2 overflow bug on x86_64
3612 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3613 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3614 Analysis suggests that attacks against RSA and DSA as a result of this
3615 defect would be very difficult to perform and are not believed likely.
3616 Attacks against DH1024 are considered just feasible, because most of the
3617 work necessary to deduce information about a private key may be performed
3618 offline. The amount of resources required for such an attack would be
3619 significant. However, for an attack on TLS to be meaningful, the server
3620 would have to share the DH1024 private key among multiple clients, which is
3621 no longer an option since CVE-2016-0701.
3623 This only affects processors that support the AVX2 but not ADX extensions
3624 like Intel Haswell (4th generation).
3626 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3627 was originally found via the OSS-Fuzz project.
3632 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3634 * bn_sqrx8x_internal carry bug on x86_64
3636 There is a carry propagating bug in the x86_64 Montgomery squaring
3637 procedure. No EC algorithms are affected. Analysis suggests that attacks
3638 against RSA and DSA as a result of this defect would be very difficult to
3639 perform and are not believed likely. Attacks against DH are considered just
3640 feasible (although very difficult) because most of the work necessary to
3641 deduce information about a private key may be performed offline. The amount
3642 of resources required for such an attack would be very significant and
3643 likely only accessible to a limited number of attackers. An attacker would
3644 additionally need online access to an unpatched system using the target
3645 private key in a scenario with persistent DH parameters and a private
3646 key that is shared between multiple clients.
3648 This only affects processors that support the BMI1, BMI2 and ADX extensions
3649 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3651 This issue was reported to OpenSSL by the OSS-Fuzz project.
3656 * Malformed X.509 IPAddressFamily could cause OOB read
3658 If an X.509 certificate has a malformed IPAddressFamily extension,
3659 OpenSSL could do a one-byte buffer overread. The most likely result
3660 would be an erroneous display of the certificate in text format.
3662 This issue was reported to OpenSSL by the OSS-Fuzz project.
3667 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3669 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3670 platform rather than 'mingw'.
3674 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3675 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3676 which is the minimum version we support.
3680 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3682 * Encrypt-Then-Mac renegotiation crash
3684 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3685 negotiated where it was not in the original handshake (or vice-versa) then
3686 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3687 and servers are affected.
3689 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3694 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3696 * Truncated packet could crash via OOB read
3698 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3699 cipher is being used, then a truncated packet can cause that host to
3700 perform an out-of-bounds read, usually resulting in a crash.
3702 This issue was reported to OpenSSL by Robert Święcki of Google.
3707 * Bad (EC)DHE parameters cause a client crash
3709 If a malicious server supplies bad parameters for a DHE or ECDHE key
3710 exchange then this can result in the client attempting to dereference a
3711 NULL pointer leading to a client crash. This could be exploited in a Denial
3714 This issue was reported to OpenSSL by Guido Vranken.
3719 * BN_mod_exp may produce incorrect results on x86_64
3721 There is a carry propagating bug in the x86_64 Montgomery squaring
3722 procedure. No EC algorithms are affected. Analysis suggests that attacks
3723 against RSA and DSA as a result of this defect would be very difficult to
3724 perform and are not believed likely. Attacks against DH are considered just
3725 feasible (although very difficult) because most of the work necessary to
3726 deduce information about a private key may be performed offline. The amount
3727 of resources required for such an attack would be very significant and
3728 likely only accessible to a limited number of attackers. An attacker would
3729 additionally need online access to an unpatched system using the target
3730 private key in a scenario with persistent DH parameters and a private
3731 key that is shared between multiple clients. For example this can occur by
3732 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3733 similar to CVE-2015-3193 but must be treated as a separate problem.
3735 This issue was reported to OpenSSL by the OSS-Fuzz project.
3740 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3742 * ChaCha20/Poly1305 heap-buffer-overflow
3744 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3745 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3746 crash. This issue is not considered to be exploitable beyond a DoS.
3748 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3753 * CMS Null dereference
3755 Applications parsing invalid CMS structures can crash with a NULL pointer
3756 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3757 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3758 structure callback if an attempt is made to free certain invalid encodings.
3759 Only CHOICE structures using a callback which do not handle NULL value are
3762 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3767 * Montgomery multiplication may produce incorrect results
3769 There is a carry propagating bug in the Broadwell-specific Montgomery
3770 multiplication procedure that handles input lengths divisible by, but
3771 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3772 and DH private keys are impossible. This is because the subroutine in
3773 question is not used in operations with the private key itself and an input
3774 of the attacker's direct choice. Otherwise the bug can manifest itself as
3775 transient authentication and key negotiation failures or reproducible
3776 erroneous outcome of public-key operations with specially crafted input.
3777 Among EC algorithms only Brainpool P-512 curves are affected and one
3778 presumably can attack ECDH key negotiation. Impact was not analyzed in
3779 detail, because pre-requisites for attack are considered unlikely. Namely
3780 multiple clients have to choose the curve in question and the server has to
3781 share the private key among them, neither of which is default behaviour.
3782 Even then only clients that chose the curve will be affected.
3784 This issue was publicly reported as transient failures and was not
3785 initially recognized as a security issue. Thanks to Richard Morgan for
3786 providing reproducible case.
3791 * Removed automatic addition of RPATH in shared libraries and executables,
3792 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3796 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3798 * Fix Use After Free for large message sizes
3800 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3801 message larger than approx 16k is received then the underlying buffer to
3802 store the incoming message is reallocated and moved. Unfortunately a
3803 dangling pointer to the old location is left which results in an attempt to
3804 write to the previously freed location. This is likely to result in a
3805 crash, however it could potentially lead to execution of arbitrary code.
3807 This issue only affects OpenSSL 1.1.0a.
3809 This issue was reported to OpenSSL by Robert Święcki.
3814 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3816 * OCSP Status Request extension unbounded memory growth
3818 A malicious client can send an excessively large OCSP Status Request
3819 extension. If that client continually requests renegotiation, sending a
3820 large OCSP Status Request extension each time, then there will be unbounded
3821 memory growth on the server. This will eventually lead to a Denial Of
3822 Service attack through memory exhaustion. Servers with a default
3823 configuration are vulnerable even if they do not support OCSP. Builds using
3824 the "no-ocsp" build time option are not affected.
3826 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3831 * SSL_peek() hang on empty record
3833 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3834 sends an empty record. This could be exploited by a malicious peer in a
3835 Denial Of Service attack.
3837 This issue was reported to OpenSSL by Alex Gaynor.
3842 * Excessive allocation of memory in tls_get_message_header() and
3843 dtls1_preprocess_fragment()
3845 A (D)TLS message includes 3 bytes for its length in the header for the
3846 message. This would allow for messages up to 16Mb in length. Messages of
3847 this length are excessive and OpenSSL includes a check to ensure that a
3848 peer is sending reasonably sized messages in order to avoid too much memory
3849 being consumed to service a connection. A flaw in the logic of version
3850 1.1.0 means that memory for the message is allocated too early, prior to
3851 the excessive message length check. Due to way memory is allocated in
3852 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3853 to service a connection. This could lead to a Denial of Service through
3854 memory exhaustion. However, the excessive message length check still takes
3855 place, and this would cause the connection to immediately fail. Assuming
3856 that the application calls SSL_free() on the failed connection in a timely
3857 manner then the 21Mb of allocated memory will then be immediately freed
3858 again. Therefore, the excessive memory allocation will be transitory in
3859 nature. This then means that there is only a security impact if:
3861 1) The application does not call SSL_free() in a timely manner in the event
3862 that the connection fails
3864 2) The application is working in a constrained environment where there is
3865 very little free memory
3867 3) The attacker initiates multiple connection attempts such that there are
3868 multiple connections in a state where memory has been allocated for the
3869 connection; SSL_free() has not yet been called; and there is insufficient
3870 memory to service the multiple requests.
3872 Except in the instance of (1) above any Denial Of Service is likely to be
3873 transitory because as soon as the connection fails the memory is
3874 subsequently freed again in the SSL_free() call. However there is an
3875 increased risk during this period of application crashes due to the lack of
3876 memory - which would then mean a more serious Denial of Service.
3878 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3879 (CVE-2016-6307 and CVE-2016-6308)
3883 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3884 had to be removed. Primary reason is that vendor assembler can't
3885 assemble our modules with -KPIC flag. As result it, assembly
3886 support, was not even available as option. But its lack means
3887 lack of side-channel resistant code, which is incompatible with
3888 security by todays standards. Fortunately gcc is readily available
3889 prepackaged option, which we firmly point at...
3893 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3895 * Windows command-line tool supports UTF-8 opt-in option for arguments
3896 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3897 (to any value) allows Windows user to access PKCS#12 file generated
3898 with Windows CryptoAPI and protected with non-ASCII password, as well
3899 as files generated under UTF-8 locale on Linux also protected with
3904 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3905 have been disabled by default and removed from DEFAULT, just like RC4.
3906 See the RC4 item below to re-enable both.
3910 * The method for finding the storage location for the Windows RAND seed file
3911 has changed. First we check %RANDFILE%. If that is not set then we check
3912 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3913 all else fails we fall back to C:\.
3917 * The EVP_EncryptUpdate() function has had its return type changed from void
3918 to int. A return of 0 indicates and error while a return of 1 indicates
3923 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3924 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3925 off the constant time implementation for RSA, DSA and DH have been made
3926 no-ops and deprecated.
3930 * Windows RAND implementation was simplified to only get entropy by
3931 calling CryptGenRandom(). Various other RAND-related tickets
3934 *Joseph Wylie Yandle, Rich Salz*
3936 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3937 and `OPENSSL_LH_`, respectively. The old names are available
3938 with API compatibility. They new names are now completely documented.
3942 * Unify TYPE_up_ref(obj) methods signature.
3943 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3944 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3945 int (instead of void) like all others TYPE_up_ref() methods.
3946 So now these methods also check the return value of CRYPTO_atomic_add(),
3947 and the validity of object reference counter.
3949 *fdasilvayy@gmail.com*
3951 * With Windows Visual Studio builds, the .pdb files are installed
3952 alongside the installed libraries and executables. For a static
3953 library installation, ossl_static.pdb is the associate compiler
3954 generated .pdb file to be used when linking programs.
3958 * Remove openssl.spec. Packaging files belong with the packagers.
3962 * Automatic Darwin/OSX configuration has had a refresh, it will now
3963 recognise x86_64 architectures automatically. You can still decide
3964 to build for a different bitness with the environment variable
3965 KERNEL_BITS (can be 32 or 64), for example:
3967 KERNEL_BITS=32 ./config
3971 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3972 256 bit AES and HMAC with SHA256.
3976 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3980 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3984 * To enable users to have their own config files and build file templates,
3985 Configure looks in the directory indicated by the environment variable
3986 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3987 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3988 name and is used as is.
3992 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3993 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3994 X509_CERT_FILE_CTX was removed.
3998 * "shared" builds are now the default. To create only static libraries use
3999 the "no-shared" Configure option.
4003 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4004 All of these option have not worked for some while and are fundamental
4009 * Make various cleanup routines no-ops and mark them as deprecated. Most
4010 global cleanup functions are no longer required because they are handled
4011 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4012 Explicitly de-initing can cause problems (e.g. where a library that uses
4013 OpenSSL de-inits, but an application is still using it). The affected
4014 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4015 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4016 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4017 COMP_zlib_cleanup().
4021 * --strict-warnings no longer enables runtime debugging options
4022 such as REF_DEBUG. Instead, debug options are automatically
4023 enabled with '--debug' builds.
4025 *Andy Polyakov, Emilia Käsper*
4027 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4028 have been moved out of the public header files. New functions for managing
4029 these have been added.
4033 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4034 objects have been moved out of the public header files. New
4035 functions for managing these have been added.
4039 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4040 have been moved out of the public header files. New functions for managing
4041 these have been added.
4045 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4046 moved out of the public header files. New functions for managing these
4051 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4055 * Removed the mk1mf build scripts.
4059 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4060 it is always safe to #include a header now.
4064 * Removed the aged BC-32 config and all its supporting scripts
4068 * Removed support for Ultrix, Netware, and OS/2.
4072 * Add support for HKDF.
4074 *Alessandro Ghedini*
4076 * Add support for blake2b and blake2s
4080 * Added support for "pipelining". Ciphers that have the
4081 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4082 encryptions/decryptions simultaneously. There are currently no built-in
4083 ciphers with this property but the expectation is that engines will be able
4084 to offer it to significantly improve throughput. Support has been extended
4085 into libssl so that multiple records for a single connection can be
4086 processed in one go (for >=TLS 1.1).
4090 * Added the AFALG engine. This is an async capable engine which is able to
4091 offload work to the Linux kernel. In this initial version it only supports
4092 AES128-CBC. The kernel must be version 4.1.0 or greater.
4096 * OpenSSL now uses a new threading API. It is no longer necessary to
4097 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4098 are two supported threading models: pthreads and windows threads. It is
4099 also possible to configure OpenSSL at compile time for "no-threads". The
4100 old threading API should no longer be used. The functions have been
4101 replaced with "no-op" compatibility macros.
4103 *Alessandro Ghedini, Matt Caswell*
4105 * Modify behavior of ALPN to invoke callback after SNI/servername
4106 callback, such that updates to the SSL_CTX affect ALPN.
4110 * Add SSL_CIPHER queries for authentication and key-exchange.
4114 * Changes to the DEFAULT cipherlist:
4115 - Prefer (EC)DHE handshakes over plain RSA.
4116 - Prefer AEAD ciphers over legacy ciphers.
4117 - Prefer ECDSA over RSA when both certificates are available.
4118 - Prefer TLSv1.2 ciphers/PRF.
4119 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4124 * Change the ECC default curve list to be this, in order: x25519,
4125 secp256r1, secp521r1, secp384r1.
4129 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4130 disabled by default. They can be re-enabled using the
4131 enable-weak-ssl-ciphers option to Configure.
4135 * If the server has ALPN configured, but supports no protocols that the
4136 client advertises, send a fatal "no_application_protocol" alert.
4137 This behaviour is SHALL in RFC 7301, though it isn't universally
4138 implemented by other servers.
4142 * Add X25519 support.
4143 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4144 for public and private key encoding using the format documented in
4145 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4146 key generation and key derivation.
4148 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4153 * Deprecate SRP_VBASE_get_by_user.
4154 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4155 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4156 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4157 seed, even if the seed is configured.
4159 Users should use SRP_VBASE_get1_by_user instead. Note that in
4160 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4161 also that even though configuring the SRP seed attempts to hide
4162 invalid usernames by continuing the handshake with fake
4163 credentials, this behaviour is not constant time and no strong
4164 guarantees are made that the handshake is indistinguishable from
4165 that of a valid user.
4169 * Configuration change; it's now possible to build dynamic engines
4170 without having to build shared libraries and vice versa. This
4171 only applies to the engines in `engines/`, those in `crypto/engine/`
4172 will always be built into libcrypto (i.e. "static").
4174 Building dynamic engines is enabled by default; to disable, use
4175 the configuration option "disable-dynamic-engine".
4177 The only requirements for building dynamic engines are the
4178 presence of the DSO module and building with position independent
4179 code, so they will also automatically be disabled if configuring
4180 with "disable-dso" or "disable-pic".
4182 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4183 are also taken away from openssl/opensslconf.h, as they are
4188 * Configuration change; if there is a known flag to compile
4189 position independent code, it will always be applied on the
4190 libcrypto and libssl object files, and never on the application
4191 object files. This means other libraries that use routines from
4192 libcrypto / libssl can be made into shared libraries regardless
4193 of how OpenSSL was configured.
4195 If this isn't desirable, the configuration options "disable-pic"
4196 or "no-pic" can be used to disable the use of PIC. This will
4197 also disable building shared libraries and dynamic engines.
4201 * Removed JPAKE code. It was experimental and has no wide use.
4205 * The INSTALL_PREFIX Makefile variable has been renamed to
4206 DESTDIR. That makes for less confusion on what this variable
4207 is for. Also, the configuration option --install_prefix is
4212 * Heartbeat for TLS has been removed and is disabled by default
4213 for DTLS; configure with enable-heartbeats. Code that uses the
4214 old #define's might need to be updated.
4216 *Emilia Käsper, Rich Salz*
4218 * Rename REF_CHECK to REF_DEBUG.
4222 * New "unified" build system
4224 The "unified" build system is aimed to be a common system for all
4225 platforms we support. With it comes new support for VMS.
4227 This system builds supports building in a different directory tree
4228 than the source tree. It produces one Makefile (for unix family
4229 or lookalikes), or one descrip.mms (for VMS).
4231 The source of information to make the Makefile / descrip.mms is
4232 small files called 'build.info', holding the necessary
4233 information for each directory with source to compile, and a
4234 template in Configurations, like unix-Makefile.tmpl or
4237 With this change, the library names were also renamed on Windows
4238 and on VMS. They now have names that are closer to the standard
4239 on Unix, and include the major version number, and in certain
4240 cases, the architecture they are built for. See "Notes on shared
4241 libraries" in INSTALL.
4243 We rely heavily on the perl module Text::Template.
4247 * Added support for auto-initialisation and de-initialisation of the library.
4248 OpenSSL no longer requires explicit init or deinit routines to be called,
4249 except in certain circumstances. See the OPENSSL_init_crypto() and
4250 OPENSSL_init_ssl() man pages for further information.
4254 * The arguments to the DTLSv1_listen function have changed. Specifically the
4255 "peer" argument is now expected to be a BIO_ADDR object.
4257 * Rewrite of BIO networking library. The BIO library lacked consistent
4258 support of IPv6, and adding it required some more extensive
4259 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4260 which hold all types of addresses and chains of address information.
4261 It also introduces a new API, with functions like BIO_socket,
4262 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4263 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4264 have been adapted accordingly.
4268 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4273 * CRIME protection: disable compression by default, even if OpenSSL is
4274 compiled with zlib enabled. Applications can still enable compression
4275 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4276 using the SSL_CONF library to configure compression.
4280 * The signature of the session callback configured with
4281 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4282 was explicitly marked as `const unsigned char*` instead of
4287 * Always DPURIFY. Remove the use of uninitialized memory in the
4288 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4292 * Removed many obsolete configuration items, including
4293 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4294 MD2_CHAR, MD2_INT, MD2_LONG
4296 IDEA_SHORT, IDEA_LONG
4297 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4299 *Rich Salz, with advice from Andy Polyakov*
4301 * Many BN internals have been moved to an internal header file.
4303 *Rich Salz with help from Andy Polyakov*
4305 * Configuration and writing out the results from it has changed.
4306 Files such as Makefile include/openssl/opensslconf.h and are now
4307 produced through general templates, such as Makefile.in and
4308 crypto/opensslconf.h.in and some help from the perl module
4311 Also, the center of configuration information is no longer
4312 Makefile. Instead, Configure produces a perl module in
4313 configdata.pm which holds most of the config data (in the hash
4314 table %config), the target data that comes from the target
4315 configuration in one of the `Configurations/*.conf` files (in
4320 * To clarify their intended purposes, the Configure options
4321 --prefix and --openssldir change their semantics, and become more
4322 straightforward and less interdependent.
4324 --prefix shall be used exclusively to give the location INSTALLTOP
4325 where programs, scripts, libraries, include files and manuals are
4326 going to be installed. The default is now /usr/local.
4328 --openssldir shall be used exclusively to give the default
4329 location OPENSSLDIR where certificates, private keys, CRLs are
4330 managed. This is also where the default openssl.cnf gets
4332 If the directory given with this option is a relative path, the
4333 values of both the --prefix value and the --openssldir value will
4334 be combined to become OPENSSLDIR.
4335 The default for --openssldir is INSTALLTOP/ssl.
4337 Anyone who uses --openssldir to specify where OpenSSL is to be
4338 installed MUST change to use --prefix instead.
4342 * The GOST engine was out of date and therefore it has been removed. An up
4343 to date GOST engine is now being maintained in an external repository.
4344 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4345 support for GOST ciphersuites (these are only activated if a GOST engine
4350 * EGD is no longer supported by default; use enable-egd when
4353 *Ben Kaduk and Rich Salz*
4355 * The distribution now has Makefile.in files, which are used to
4356 create Makefile's when Configure is run. *Configure must be run
4357 before trying to build now.*
4361 * The return value for SSL_CIPHER_description() for error conditions
4366 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4368 Obtaining and performing DNSSEC validation of TLSA records is
4369 the application's responsibility. The application provides
4370 the TLSA records of its choice to OpenSSL, and these are then
4371 used to authenticate the peer.
4373 The TLSA records need not even come from DNS. They can, for
4374 example, be used to implement local end-entity certificate or
4375 trust-anchor "pinning", where the "pin" data takes the form
4376 of TLSA records, which can augment or replace verification
4377 based on the usual WebPKI public certification authorities.
4381 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4382 continues to support deprecated interfaces in default builds.
4383 However, applications are strongly advised to compile their
4384 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4385 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4386 or the 1.1.0 releases.
4388 In environments in which all applications have been ported to
4389 not use any deprecated interfaces OpenSSL's Configure script
4390 should be used with the --api=1.1.0 option to entirely remove
4391 support for the deprecated features from the library and
4392 unconditionally disable them in the installed headers.
4393 Essentially the same effect can be achieved with the "no-deprecated"
4394 argument to Configure, except that this will always restrict
4395 the build to just the latest API, rather than a fixed API
4398 As applications are ported to future revisions of the API,
4399 they should update their compile-time OPENSSL_API_COMPAT define
4400 accordingly, but in most cases should be able to continue to
4401 compile with later releases.
4403 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4404 0x10000000L and 0x00908000L, respectively. However those
4405 versions did not support the OPENSSL_API_COMPAT feature, and
4406 so applications are not typically tested for explicit support
4407 of just the undeprecated features of either release.
4411 * Add support for setting the minimum and maximum supported protocol.
4412 It can bet set via the SSL_set_min_proto_version() and
4413 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4414 MaxProtocol. It's recommended to use the new APIs to disable
4415 protocols instead of disabling individual protocols using
4416 SSL_set_options() or SSL_CONF's Protocol. This change also
4417 removes support for disabling TLS 1.2 in the OpenSSL TLS
4418 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4422 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4426 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4427 and integrates ECDSA and ECDH functionality into EC. Implementations can
4428 now redirect key generation and no longer need to convert to or from
4431 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4432 include the ec.h header file instead.
4436 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4437 ciphers who are no longer supported and drops support the ephemeral RSA key
4438 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4442 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4443 opaque. For HMAC_CTX, the following constructors and destructors
4446 HMAC_CTX *HMAC_CTX_new(void);
4447 void HMAC_CTX_free(HMAC_CTX *ctx);
4449 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4450 destroy such methods has been added. See EVP_MD_meth_new(3) and
4451 EVP_CIPHER_meth_new(3) for documentation.
4454 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4455 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4456 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4457 an already created structure.
4458 2) For consistency with the majority of our object creators and
4459 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4460 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4461 for deprecated builds.
4465 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4466 cryptographic operations to be performed asynchronously as long as an
4467 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4468 further details. Libssl has also had this capability integrated with the
4469 introduction of the new mode SSL_MODE_ASYNC and associated error
4470 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4471 pages. This work was developed in partnership with Intel Corp.
4475 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4476 always enabled now. If you want to disable the support you should
4477 exclude it using the list of supported ciphers. This also means that the
4478 "-no_ecdhe" option has been removed from s_server.
4482 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4483 SSL_{CTX_}set1_curves() which can set a list.
4487 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4488 curve you want to support using SSL_{CTX_}set1_curves().
4492 * State machine rewrite. The state machine code has been significantly
4493 refactored in order to remove much duplication of code and solve issues
4494 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4495 further details). This change does have some associated API changes.
4496 Notably the SSL_state() function has been removed and replaced by
4497 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4498 SSL_set_state() has been removed altogether. The previous handshake states
4499 defined in ssl.h and ssl3.h have also been removed.
4503 * All instances of the string "ssleay" in the public API were replaced
4504 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4505 Some error codes related to internal RSA_eay API's were renamed.
4509 * The demo files in crypto/threads were moved to demo/threads.
4513 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4516 *Matt Caswell, Rich Salz*
4518 * New ASN.1 embed macro.
4520 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4521 structure is not allocated: it is part of the parent. That is instead of
4529 This reduces memory fragmentation and make it impossible to accidentally
4530 set a mandatory field to NULL.
4532 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4533 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4534 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4539 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4543 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4544 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4545 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4546 DES and RC4 ciphersuites.
4550 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4551 This changes the decoding behaviour for some invalid messages,
4552 though the change is mostly in the more lenient direction, and
4553 legacy behaviour is preserved as much as possible.
4557 * Fix no-stdio build.
4558 *David Woodhouse <David.Woodhouse@intel.com> and also*
4559 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4561 * New testing framework
4562 The testing framework has been largely rewritten and is now using
4563 perl and the perl modules Test::Harness and an extended variant of
4564 Test::More called OpenSSL::Test to do its work. All test scripts in
4565 test/ have been rewritten into test recipes, and all direct calls to
4566 executables in test/Makefile have become individual recipes using the
4567 simplified testing OpenSSL::Test::Simple.
4569 For documentation on our testing modules, do:
4571 perldoc test/testlib/OpenSSL/Test/Simple.pm
4572 perldoc test/testlib/OpenSSL/Test.pm
4576 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4577 are used; the latter aborts on memory leaks (usually checked on exit).
4578 Some undocumented "set malloc, etc., hooks" functions were removed
4579 and others were changed. All are now documented.
4583 * In DSA_generate_parameters_ex, if the provided seed is too short,
4586 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4588 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4589 from RFC4279, RFC4785, RFC5487, RFC5489.
4591 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4592 original RSA_PSK patch.
4596 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4597 era flag was never set throughout the codebase (only read). Also removed
4598 SSL3_FLAGS_POP_BUFFER which was only used if
4599 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4603 * Changed the default name options in the "ca", "crl", "req" and "x509"
4604 to be "oneline" instead of "compat".
4608 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4609 not aware of clients that still exhibit this bug, and the workaround
4610 hasn't been working properly for a while.
4614 * The return type of BIO_number_read() and BIO_number_written() as well as
4615 the corresponding num_read and num_write members in the BIO structure has
4616 changed from unsigned long to uint64_t. On platforms where an unsigned
4617 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4622 * Given the pervasive nature of TLS extensions it is inadvisable to run
4623 OpenSSL without support for them. It also means that maintaining
4624 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4625 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4629 * Removed support for the two export grade static DH ciphersuites
4630 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4631 were newly added (along with a number of other static DH ciphersuites) to
4632 1.0.2. However the two export ones have *never* worked since they were
4633 introduced. It seems strange in any case to be adding new export
4634 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4638 * Version negotiation has been rewritten. In particular SSLv23_method(),
4639 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4640 and turned into macros which simply call the new preferred function names
4641 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4642 should use the new names instead. Also as part of this change the ssl23.h
4643 header file has been removed.
4647 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4648 code and the associated standard is no longer considered fit-for-purpose.
4652 * RT2547 was closed. When generating a private key, try to make the
4653 output file readable only by the owner. This behavior change might
4654 be noticeable when interacting with other software.
4656 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4661 * Added HTTP GET support to the ocsp command.
4665 * Changed default digest for the dgst and enc commands from MD5 to
4670 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4674 * Added support for TLS extended master secret from
4675 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4676 initial patch which was a great help during development.
4680 * All libssl internal structures have been removed from the public header
4681 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4682 now redundant). Users should not attempt to access internal structures
4683 directly. Instead they should use the provided API functions.
4687 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4688 Access to deprecated functions can be re-enabled by running config with
4689 "enable-deprecated". In addition applications wishing to use deprecated
4690 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4691 will, by default, disable some transitive includes that previously existed
4692 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4696 * Added support for OCB mode. OpenSSL has been granted a patent license
4697 compatible with the OpenSSL license for use of OCB. Details are available
4698 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4699 for OCB can be removed by calling config with no-ocb.
4703 * SSLv2 support has been removed. It still supports receiving an SSLv2
4704 compatible client hello.
4708 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4709 done while fixing the error code for the key-too-small case.
4711 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4713 * CA.sh has been removed; use CA.pl instead.
4717 * Removed old DES API.
4721 * Remove various unsupported platforms:
4727 Sinix/ReliantUNIX RM400
4732 16-bit platforms such as WIN16
4736 * Clean up OPENSSL_NO_xxx #define's
4737 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4738 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4739 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4740 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4741 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4742 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4743 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4744 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4745 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4746 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4750 * Cleaned up dead code
4751 Remove all but one '#ifdef undef' which is to be looked at.
4755 * Clean up calling of xxx_free routines.
4756 Just like free(), fix most of the xxx_free routines to accept
4757 NULL. Remove the non-null checks from callers. Save much code.
4761 * Add secure heap for storage of private keys (when possible).
4762 Add BIO_s_secmem(), CBIGNUM, etc.
4763 Contributed by Akamai Technologies under our Corporate CLA.
4767 * Experimental support for a new, fast, unbiased prime candidate generator,
4768 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4770 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4772 * New output format NSS in the sess_id command line tool. This allows
4773 exporting the session id and the master key in NSS keylog format.
4775 *Martin Kaiser <martin@kaiser.cx>*
4777 * Harmonize version and its documentation. -f flag is used to display
4780 *mancha <mancha1@zoho.com>*
4782 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4783 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4785 *mancha <mancha1@zoho.com>*
4787 * Fix some double frees. These are not thought to be exploitable.
4789 *mancha <mancha1@zoho.com>*
4791 * A missing bounds check in the handling of the TLS heartbeat extension
4792 can be used to reveal up to 64k of memory to a connected client or
4795 Thanks for Neel Mehta of Google Security for discovering this bug and to
4796 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4797 preparing the fix ([CVE-2014-0160])
4799 *Adam Langley, Bodo Moeller*
4801 * Fix for the attack described in the paper "Recovering OpenSSL
4802 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4803 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4804 <http://eprint.iacr.org/2014/140>
4806 Thanks to Yuval Yarom and Naomi Benger for discovering this
4807 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4809 *Yuval Yarom and Naomi Benger*
4811 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4812 this fixes a limitation in previous versions of OpenSSL.
4816 * Experimental encrypt-then-mac support.
4818 Experimental support for encrypt then mac from
4819 draft-gutmann-tls-encrypt-then-mac-02.txt
4821 To enable it set the appropriate extension number (0x42 for the test
4822 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4824 For non-compliant peers (i.e. just about everything) this should have no
4827 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4831 * Add EVP support for key wrapping algorithms, to avoid problems with
4832 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4833 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4834 algorithms and include tests cases.
4838 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4843 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4844 MGF1 digest and OAEP label.
4848 * Make openssl verify return errors.
4850 *Chris Palmer <palmer@google.com> and Ben Laurie*
4852 * New function ASN1_TIME_diff to calculate the difference between two
4853 ASN1_TIME structures or one structure and the current time.
4857 * Update fips_test_suite to support multiple command line options. New
4858 test to induce all self test errors in sequence and check expected
4863 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4864 sign or verify all in one operation.
4868 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4869 test programs and fips_test_suite. Includes functionality to parse
4870 the minimal script output of fipsalgest.pl directly.
4874 * Add authorisation parameter to FIPS_module_mode_set().
4878 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4882 * Use separate DRBG fields for internal and external flags. New function
4883 FIPS_drbg_health_check() to perform on demand health checking. Add
4884 generation tests to fips_test_suite with reduced health check interval to
4885 demonstrate periodic health checking. Add "nodh" option to
4886 fips_test_suite to skip very slow DH test.
4890 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4895 * More extensive health check for DRBG checking many more failure modes.
4896 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4897 combination: call this in fips_test_suite.
4901 * Add support for canonical generation of DSA parameter 'g'. See
4904 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4905 POST to handle HMAC cases.
4909 * Add functions FIPS_module_version() and FIPS_module_version_text()
4910 to return numerical and string versions of the FIPS module number.
4914 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4915 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4916 outside the validated module in the FIPS capable OpenSSL.
4920 * Minor change to DRBG entropy callback semantics. In some cases
4921 there is no multiple of the block length between min_len and
4922 max_len. Allow the callback to return more than max_len bytes
4923 of entropy but discard any extra: it is the callback's responsibility
4924 to ensure that the extra data discarded does not impact the
4925 requested amount of entropy.
4929 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4930 information in FIPS186-3, SP800-57 and SP800-131A.
4934 * CCM support via EVP. Interface is very similar to GCM case except we
4935 must supply all data in one chunk (i.e. no update, final) and the
4936 message length must be supplied if AAD is used. Add algorithm test
4941 * Initial version of POST overhaul. Add POST callback to allow the status
4942 of POST to be monitored and/or failures induced. Modify fips_test_suite
4943 to use callback. Always run all selftests even if one fails.
4947 * XTS support including algorithm test driver in the fips_gcmtest program.
4948 Note: this does increase the maximum key length from 32 to 64 bytes but
4949 there should be no binary compatibility issues as existing applications
4950 will never use XTS mode.
4954 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4955 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4956 performs algorithm blocking for unapproved PRNG types. Also do not
4957 set PRNG type in FIPS_mode_set(): leave this to the application.
4958 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4959 the standard OpenSSL PRNG: set additional data to a date time vector.
4963 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4964 This shouldn't present any incompatibility problems because applications
4965 shouldn't be using these directly and any that are will need to rethink
4966 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4970 * Extensive self tests and health checking required by SP800-90 DRBG.
4971 Remove strength parameter from FIPS_drbg_instantiate and always
4972 instantiate at maximum supported strength.
4976 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4980 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4984 * New function DH_compute_key_padded() to compute a DH key and pad with
4985 leading zeroes if needed: this complies with SP800-56A et al.
4989 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4990 anything, incomplete, subject to change and largely untested at present.
4994 * Modify fipscanisteronly build option to only build the necessary object
4995 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4999 * Add experimental option FIPSSYMS to give all symbols in
5000 fipscanister.o and FIPS or fips prefix. This will avoid
5001 conflicts with future versions of OpenSSL. Add perl script
5002 util/fipsas.pl to preprocess assembly language source files
5003 and rename any affected symbols.
5007 * Add selftest checks and algorithm block of non-fips algorithms in
5008 FIPS mode. Remove DES2 from selftests.
5012 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5013 return internal method without any ENGINE dependencies. Add new
5014 tiny fips sign and verify functions.
5018 * New build option no-ec2m to disable characteristic 2 code.
5022 * New build option "fipscanisteronly". This only builds fipscanister.o
5023 and (currently) associated fips utilities. Uses the file Makefile.fips
5024 instead of Makefile.org as the prototype.
5028 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5029 Update fips_gcmtest to use IV generator.
5033 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5034 setting output buffer to NULL. The `*Final` function must be
5035 called although it will not retrieve any additional data. The tag
5036 can be set or retrieved with a ctrl. The IV length is by default 12
5037 bytes (96 bits) but can be set to an alternative value. If the IV
5038 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5043 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5044 underlying do_cipher function handles all cipher semantics itself
5045 including padding and finalisation. This is useful if (for example)
5046 an ENGINE cipher handles block padding itself. The behaviour of
5047 do_cipher is subtly changed if this flag is set: the return value
5048 is the number of characters written to the output buffer (zero is
5049 no longer an error code) or a negative error code. Also if the
5050 input buffer is NULL and length 0 finalisation should be performed.
5054 * If a candidate issuer certificate is already part of the constructed
5055 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5059 * Improve forward-security support: add functions
5061 void SSL_CTX_set_not_resumable_session_callback(
5062 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5063 void SSL_set_not_resumable_session_callback(
5064 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5066 for use by SSL/TLS servers; the callback function will be called whenever a
5067 new session is created, and gets to decide whether the session may be
5068 cached to make it resumable (return 0) or not (return 1). (As by the
5069 SSL/TLS protocol specifications, the session_id sent by the server will be
5070 empty to indicate that the session is not resumable; also, the server will
5071 not generate RFC 4507 (RFC 5077) session tickets.)
5073 A simple reasonable callback implementation is to return is_forward_secure.
5074 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5075 by the SSL/TLS server library, indicating whether it can provide forward
5078 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5080 * New -verify_name option in command line utilities to set verification
5085 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5086 Add CMAC pkey methods.
5090 * Experimental renegotiation in s_server -www mode. If the client
5091 browses /reneg connection is renegotiated. If /renegcert it is
5092 renegotiated requesting a certificate.
5096 * Add an "external" session cache for debugging purposes to s_server. This
5097 should help trace issues which normally are only apparent in deployed
5098 multi-process servers.
5102 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5103 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5104 BIO_set_cipher() and some obscure PEM functions were changed so they
5105 can now return an error. The RAND changes required a change to the
5106 RAND_METHOD structure.
5110 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5111 a gcc attribute to warn if the result of a function is ignored. This
5112 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5113 whose return value is often ignored.
5117 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5118 These allow SCTs (signed certificate timestamps) to be requested and
5119 validated when establishing a connection.
5121 *Rob Percival <robpercival@google.com>*
5126 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5128 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5129 used even when parsing explicit parameters, when loading a encoded key
5130 or calling `EC_GROUP_new_from_ecpkparameters()`/
5131 `EC_GROUP_new_from_ecparameters()`.
5132 This prevents bypass of security hardening and performance gains,
5133 especially for curves with specialized EC_METHODs.
5134 By default, if a key encoded with explicit parameters is loaded and later
5135 encoded, the output is still encoded with explicit parameters, even if
5136 internally a "named" EC_GROUP is used for computation.
5140 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5141 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5142 NULL. After this change, only the cofactor parameter can be NULL. It also
5143 does some minimal sanity checks on the passed order.
5148 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5149 An attack is simple, if the first CMS_recipientInfo is valid but the
5150 second CMS_recipientInfo is chosen ciphertext. If the second
5151 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5152 encryption key will be replaced by garbage, and the message cannot be
5153 decoded, but if the RSA decryption fails, the correct encryption key is
5154 used and the recipient will not notice the attack.
5155 As a work around for this potential attack the length of the decrypted
5156 key must be equal to the cipher default key length, in case the
5157 certificate is not given and all recipientInfo are tried out.
5158 The old behaviour can be re-enabled in the CMS code by setting the
5159 CMS_DEBUG_DECRYPT flag.
5164 * Document issue with installation paths in diverse Windows builds
5166 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5167 binaries and run-time config file.
5172 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5174 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5175 This changes the size when using the `genpkey` command when no size is given.
5176 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5177 generation commands to use 2048 bits by default.
5181 * Add FIPS support for Android Arm 64-bit
5183 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5184 Module in Version 2.0.10. For some reason, the corresponding target
5185 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5186 built with FIPS support on Android Arm 64-bit. This omission has been
5189 *Matthias St. Pierre*
5191 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5193 * 0-byte record padding oracle
5195 If an application encounters a fatal protocol error and then calls
5196 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5197 then OpenSSL can respond differently to the calling application if a 0 byte
5198 record is received with invalid padding compared to if a 0 byte record is
5199 received with an invalid MAC. If the application then behaves differently
5200 based on that in a way that is detectable to the remote peer, then this
5201 amounts to a padding oracle that could be used to decrypt data.
5203 In order for this to be exploitable "non-stitched" ciphersuites must be in
5204 use. Stitched ciphersuites are optimised implementations of certain
5205 commonly used ciphersuites. Also the application must call SSL_shutdown()
5206 twice even if a protocol error has occurred (applications should not do
5207 this but some do anyway).
5209 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5210 Aviram, with additional investigation by Steven Collison and Andrew
5211 Hourselt. It was reported to OpenSSL on 10th December 2018.
5216 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5220 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5222 * Microarchitecture timing vulnerability in ECC scalar multiplication
5224 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5225 shown to be vulnerable to a microarchitecture timing side channel attack.
5226 An attacker with sufficient access to mount local timing attacks during
5227 ECDSA signature generation could recover the private key.
5229 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5230 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5236 * Timing vulnerability in DSA signature generation
5238 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5239 timing side channel attack. An attacker could use variations in the signing
5240 algorithm to recover the private key.
5242 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5247 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5248 Module, accidentally introduced while backporting security fixes from the
5249 development branch and hindering the use of ECC in FIPS mode.
5253 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5255 * Client DoS due to large DH parameter
5257 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5258 malicious server can send a very large prime value to the client. This will
5259 cause the client to spend an unreasonably long period of time generating a
5260 key for this prime resulting in a hang until the client has finished. This
5261 could be exploited in a Denial Of Service attack.
5263 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5268 * Cache timing vulnerability in RSA Key Generation
5270 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5271 a cache timing side channel attack. An attacker with sufficient access to
5272 mount cache timing attacks during the RSA key generation process could
5273 recover the private key.
5275 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5276 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5281 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5282 parameter is no longer accepted, as it leads to a corrupt table. NULL
5283 pem_str is reserved for alias entries only.
5287 * Revert blinding in ECDSA sign and instead make problematic addition
5288 length-invariant. Switch even to fixed-length Montgomery multiplication.
5292 * Change generating and checking of primes so that the error rate of not
5293 being prime depends on the intended use based on the size of the input.
5294 For larger primes this will result in more rounds of Miller-Rabin.
5295 The maximal error rate for primes with more than 1080 bits is lowered
5298 *Kurt Roeckx, Annie Yousar*
5300 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5304 * Add blinding to ECDSA and DSA signatures to protect against side channel
5305 attacks discovered by Keegan Ryan (NCC Group).
5309 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5310 now allow empty (zero character) pass phrases.
5314 * Certificate time validation (X509_cmp_time) enforces stricter
5315 compliance with RFC 5280. Fractional seconds and timezone offsets
5316 are no longer allowed.
5320 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5322 * Constructed ASN.1 types with a recursive definition could exceed the stack
5324 Constructed ASN.1 types with a recursive definition (such as can be found
5325 in PKCS7) could eventually exceed the stack given malicious input with
5326 excessive recursion. This could result in a Denial Of Service attack. There
5327 are no such structures used within SSL/TLS that come from untrusted sources
5328 so this is considered safe.
5330 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5336 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5338 * Read/write after SSL object in error state
5340 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5341 mechanism. The intent was that if a fatal error occurred during a handshake
5342 then OpenSSL would move into the error state and would immediately fail if
5343 you attempted to continue the handshake. This works as designed for the
5344 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5345 SSL_connect()), however due to a bug it does not work correctly if
5346 SSL_read() or SSL_write() is called directly. In that scenario, if the
5347 handshake fails then a fatal error will be returned in the initial function
5348 call. If SSL_read()/SSL_write() is subsequently called by the application
5349 for the same SSL object then it will succeed and the data is passed without
5350 being decrypted/encrypted directly from the SSL/TLS record layer.
5352 In order to exploit this issue an application bug would have to be present
5353 that resulted in a call to SSL_read()/SSL_write() being issued after having
5354 already received a fatal error.
5356 This issue was reported to OpenSSL by David Benjamin (Google).
5361 * rsaz_1024_mul_avx2 overflow bug on x86_64
5363 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5364 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5365 Analysis suggests that attacks against RSA and DSA as a result of this
5366 defect would be very difficult to perform and are not believed likely.
5367 Attacks against DH1024 are considered just feasible, because most of the
5368 work necessary to deduce information about a private key may be performed
5369 offline. The amount of resources required for such an attack would be
5370 significant. However, for an attack on TLS to be meaningful, the server
5371 would have to share the DH1024 private key among multiple clients, which is
5372 no longer an option since CVE-2016-0701.
5374 This only affects processors that support the AVX2 but not ADX extensions
5375 like Intel Haswell (4th generation).
5377 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5378 was originally found via the OSS-Fuzz project.
5383 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5385 * bn_sqrx8x_internal carry bug on x86_64
5387 There is a carry propagating bug in the x86_64 Montgomery squaring
5388 procedure. No EC algorithms are affected. Analysis suggests that attacks
5389 against RSA and DSA as a result of this defect would be very difficult to
5390 perform and are not believed likely. Attacks against DH are considered just
5391 feasible (although very difficult) because most of the work necessary to
5392 deduce information about a private key may be performed offline. The amount
5393 of resources required for such an attack would be very significant and
5394 likely only accessible to a limited number of attackers. An attacker would
5395 additionally need online access to an unpatched system using the target
5396 private key in a scenario with persistent DH parameters and a private
5397 key that is shared between multiple clients.
5399 This only affects processors that support the BMI1, BMI2 and ADX extensions
5400 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5402 This issue was reported to OpenSSL by the OSS-Fuzz project.
5407 * Malformed X.509 IPAddressFamily could cause OOB read
5409 If an X.509 certificate has a malformed IPAddressFamily extension,
5410 OpenSSL could do a one-byte buffer overread. The most likely result
5411 would be an erroneous display of the certificate in text format.
5413 This issue was reported to OpenSSL by the OSS-Fuzz project.
5417 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5419 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5420 platform rather than 'mingw'.
5424 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5426 * Truncated packet could crash via OOB read
5428 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5429 cipher is being used, then a truncated packet can cause that host to
5430 perform an out-of-bounds read, usually resulting in a crash.
5432 This issue was reported to OpenSSL by Robert Święcki of Google.
5437 * BN_mod_exp may produce incorrect results on x86_64
5439 There is a carry propagating bug in the x86_64 Montgomery squaring
5440 procedure. No EC algorithms are affected. Analysis suggests that attacks
5441 against RSA and DSA as a result of this defect would be very difficult to
5442 perform and are not believed likely. Attacks against DH are considered just
5443 feasible (although very difficult) because most of the work necessary to
5444 deduce information about a private key may be performed offline. The amount
5445 of resources required for such an attack would be very significant and
5446 likely only accessible to a limited number of attackers. An attacker would
5447 additionally need online access to an unpatched system using the target
5448 private key in a scenario with persistent DH parameters and a private
5449 key that is shared between multiple clients. For example this can occur by
5450 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5451 similar to CVE-2015-3193 but must be treated as a separate problem.
5453 This issue was reported to OpenSSL by the OSS-Fuzz project.
5458 * Montgomery multiplication may produce incorrect results
5460 There is a carry propagating bug in the Broadwell-specific Montgomery
5461 multiplication procedure that handles input lengths divisible by, but
5462 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5463 and DH private keys are impossible. This is because the subroutine in
5464 question is not used in operations with the private key itself and an input
5465 of the attacker's direct choice. Otherwise the bug can manifest itself as
5466 transient authentication and key negotiation failures or reproducible
5467 erroneous outcome of public-key operations with specially crafted input.
5468 Among EC algorithms only Brainpool P-512 curves are affected and one
5469 presumably can attack ECDH key negotiation. Impact was not analyzed in
5470 detail, because pre-requisites for attack are considered unlikely. Namely
5471 multiple clients have to choose the curve in question and the server has to
5472 share the private key among them, neither of which is default behaviour.
5473 Even then only clients that chose the curve will be affected.
5475 This issue was publicly reported as transient failures and was not
5476 initially recognized as a security issue. Thanks to Richard Morgan for
5477 providing reproducible case.
5482 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5483 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5484 prevent issues where no progress is being made and the peer continually
5485 sends unrecognised record types, using up resources processing them.
5489 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5491 * Missing CRL sanity check
5493 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5494 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5495 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5497 This issue only affects the OpenSSL 1.0.2i
5502 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5504 * OCSP Status Request extension unbounded memory growth
5506 A malicious client can send an excessively large OCSP Status Request
5507 extension. If that client continually requests renegotiation, sending a
5508 large OCSP Status Request extension each time, then there will be unbounded
5509 memory growth on the server. This will eventually lead to a Denial Of
5510 Service attack through memory exhaustion. Servers with a default
5511 configuration are vulnerable even if they do not support OCSP. Builds using
5512 the "no-ocsp" build time option are not affected.
5514 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5519 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5522 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5528 * OOB write in MDC2_Update()
5530 An overflow can occur in MDC2_Update() either if called directly or
5531 through the EVP_DigestUpdate() function using MDC2. If an attacker
5532 is able to supply very large amounts of input data after a previous
5533 call to EVP_EncryptUpdate() with a partial block then a length check
5534 can overflow resulting in a heap corruption.
5536 The amount of data needed is comparable to SIZE_MAX which is impractical
5539 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5544 * Malformed SHA512 ticket DoS
5546 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5547 DoS attack where a malformed ticket will result in an OOB read which will
5550 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5551 a custom server callback and ticket lookup mechanism.
5553 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5558 * OOB write in BN_bn2dec()
5560 The function BN_bn2dec() does not check the return value of BN_div_word().
5561 This can cause an OOB write if an application uses this function with an
5562 overly large BIGNUM. This could be a problem if an overly large certificate
5563 or CRL is printed out from an untrusted source. TLS is not affected because
5564 record limits will reject an oversized certificate before it is parsed.
5566 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5571 * OOB read in TS_OBJ_print_bio()
5573 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5574 the total length the OID text representation would use and not the amount
5575 of data written. This will result in OOB reads when large OIDs are
5578 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5583 * Pointer arithmetic undefined behaviour
5585 Avoid some undefined pointer arithmetic
5587 A common idiom in the codebase is to check limits in the following manner:
5590 Where "p" points to some malloc'd data of SIZE bytes and
5593 "len" here could be from some externally supplied data (e.g. from a TLS
5596 The rules of C pointer arithmetic are such that "p + len" is only well
5597 defined where len <= SIZE. Therefore the above idiom is actually
5598 undefined behaviour.
5600 For example this could cause problems if some malloc implementation
5601 provides an address for "p" such that "p + len" actually overflows for
5602 values of len that are too big and therefore p + len < limit.
5604 This issue was reported to OpenSSL by Guido Vranken
5609 * Constant time flag not preserved in DSA signing
5611 Operations in the DSA signing algorithm should run in constant time in
5612 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5613 implementation means that a non-constant time codepath is followed for
5614 certain operations. This has been demonstrated through a cache-timing
5615 attack to be sufficient for an attacker to recover the private DSA key.
5617 This issue was reported by César Pereida (Aalto University), Billy Brumley
5618 (Tampere University of Technology), and Yuval Yarom (The University of
5619 Adelaide and NICTA).
5624 * DTLS buffered message DoS
5626 In a DTLS connection where handshake messages are delivered out-of-order
5627 those messages that OpenSSL is not yet ready to process will be buffered
5628 for later use. Under certain circumstances, a flaw in the logic means that
5629 those messages do not get removed from the buffer even though the handshake
5630 has been completed. An attacker could force up to approx. 15 messages to
5631 remain in the buffer when they are no longer required. These messages will
5632 be cleared when the DTLS connection is closed. The default maximum size for
5633 a message is 100k. Therefore, the attacker could force an additional 1500k
5634 to be consumed per connection. By opening many simultaneous connections an
5635 attacker could cause a DoS attack through memory exhaustion.
5637 This issue was reported to OpenSSL by Quan Luo.
5642 * DTLS replay protection DoS
5644 A flaw in the DTLS replay attack protection mechanism means that records
5645 that arrive for future epochs update the replay protection "window" before
5646 the MAC for the record has been validated. This could be exploited by an
5647 attacker by sending a record for the next epoch (which does not have to
5648 decrypt or have a valid MAC), with a very large sequence number. This means
5649 that all subsequent legitimate packets are dropped causing a denial of
5650 service for a specific DTLS connection.
5652 This issue was reported to OpenSSL by the OCAP audit team.
5657 * Certificate message OOB reads
5659 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5660 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5661 theoretical DoS risk but this has not been observed in practice on common
5664 The messages affected are client certificate, client certificate request
5665 and server certificate. As a result the attack can only be performed
5666 against a client or a server which enables client authentication.
5668 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5673 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5675 * Prevent padding oracle in AES-NI CBC MAC check
5677 A MITM attacker can use a padding oracle attack to decrypt traffic
5678 when the connection uses an AES CBC cipher and the server support
5681 This issue was introduced as part of the fix for Lucky 13 padding
5682 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5683 constant time by making sure that always the same bytes are read and
5684 compared against either the MAC or padding bytes. But it no longer
5685 checked that there was enough data to have both the MAC and padding
5688 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5692 * Fix EVP_EncodeUpdate overflow
5694 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5695 Base64 encoding of binary data. If an attacker is able to supply very large
5696 amounts of input data then a length check can overflow resulting in a heap
5699 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5700 the `PEM_write_bio*` family of functions. These are mainly used within the
5701 OpenSSL command line applications, so any application which processes data
5702 from an untrusted source and outputs it as a PEM file should be considered
5703 vulnerable to this issue. User applications that call these APIs directly
5704 with large amounts of untrusted data may also be vulnerable.
5706 This issue was reported by Guido Vranken.
5711 * Fix EVP_EncryptUpdate overflow
5713 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5714 is able to supply very large amounts of input data after a previous call to
5715 EVP_EncryptUpdate() with a partial block then a length check can overflow
5716 resulting in a heap corruption. Following an analysis of all OpenSSL
5717 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5718 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5719 the first called function after an EVP_EncryptInit(), and therefore that
5720 specific call must be safe. The second form is where the length passed to
5721 EVP_EncryptUpdate() can be seen from the code to be some small value and
5722 therefore there is no possibility of an overflow. Since all instances are
5723 one of these two forms, it is believed that there can be no overflows in
5724 internal code due to this problem. It should be noted that
5725 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5726 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5727 of these calls have also been analysed too and it is believed there are no
5728 instances in internal usage where an overflow could occur.
5730 This issue was reported by Guido Vranken.
5735 * Prevent ASN.1 BIO excessive memory allocation
5737 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5738 a short invalid encoding can cause allocation of large amounts of memory
5739 potentially consuming excessive resources or exhausting memory.
5741 Any application parsing untrusted data through d2i BIO functions is
5742 affected. The memory based functions such as d2i_X509() are *not* affected.
5743 Since the memory based functions are used by the TLS library, TLS
5744 applications are not affected.
5746 This issue was reported by Brian Carpenter.
5753 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5754 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5755 in arbitrary stack data being returned in the buffer.
5757 This issue was reported by Guido Vranken.
5762 * Modify behavior of ALPN to invoke callback after SNI/servername
5763 callback, such that updates to the SSL_CTX affect ALPN.
5767 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5772 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5773 methods are enabled and ssl2 is disabled the methods return NULL.
5777 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5779 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5780 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5781 provide any "EXPORT" or "LOW" strength ciphers.
5785 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5786 is by default disabled at build-time. Builds that are not configured with
5787 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5788 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5789 will need to explicitly call either of:
5791 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5793 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5795 as appropriate. Even if either of those is used, or the application
5796 explicitly uses the version-specific SSLv2_method() or its client and
5797 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5798 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5799 ciphers, and SSLv2 56-bit DES are no longer available.
5804 * Fix a double-free in DSA code
5806 A double free bug was discovered when OpenSSL parses malformed DSA private
5807 keys and could lead to a DoS attack or memory corruption for applications
5808 that receive DSA private keys from untrusted sources. This scenario is
5811 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5817 * Disable SRP fake user seed to address a server memory leak.
5819 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5821 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5822 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5823 was changed to ignore the "fake user" SRP seed, even if the seed
5826 Users should use SRP_VBASE_get1_by_user instead. Note that in
5827 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5828 also that even though configuring the SRP seed attempts to hide
5829 invalid usernames by continuing the handshake with fake
5830 credentials, this behaviour is not constant time and no strong
5831 guarantees are made that the handshake is indistinguishable from
5832 that of a valid user.
5837 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5839 In the BN_hex2bn function the number of hex digits is calculated using an
5840 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5841 large values of `i` this can result in `bn_expand` not allocating any
5842 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5843 field as NULL leading to a subsequent NULL ptr deref. For very large values
5844 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5845 In this case memory is allocated to the internal BIGNUM data field, but it
5846 is insufficiently sized leading to heap corruption. A similar issue exists
5847 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5848 is ever called by user applications with very large untrusted hex/dec data.
5849 This is anticipated to be a rare occurrence.
5851 All OpenSSL internal usage of these functions use data that is not expected
5852 to be untrusted, e.g. config file data or application command line
5853 arguments. If user developed applications generate config file data based
5854 on untrusted data then it is possible that this could also lead to security
5855 consequences. This is also anticipated to be rare.
5857 This issue was reported to OpenSSL by Guido Vranken.
5862 * Fix memory issues in `BIO_*printf` functions
5864 The internal `fmtstr` function used in processing a "%s" format string in
5865 the `BIO_*printf` functions could overflow while calculating the length of a
5866 string and cause an OOB read when printing very long strings.
5868 Additionally the internal `doapr_outch` function can attempt to write to an
5869 OOB memory location (at an offset from the NULL pointer) in the event of a
5870 memory allocation failure. In 1.0.2 and below this could be caused where
5871 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5872 could be in processing a very long "%s" format string. Memory leaks can
5875 The first issue may mask the second issue dependent on compiler behaviour.
5876 These problems could enable attacks where large amounts of untrusted data
5877 is passed to the `BIO_*printf` functions. If applications use these functions
5878 in this way then they could be vulnerable. OpenSSL itself uses these
5879 functions when printing out human-readable dumps of ASN.1 data. Therefore
5880 applications that print this data could be vulnerable if the data is from
5881 untrusted sources. OpenSSL command line applications could also be
5882 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5883 as command line arguments.
5885 Libssl is not considered directly vulnerable. Additionally certificates etc
5886 received via remote connections via libssl are also unlikely to be able to
5887 trigger these issues because of message size limits enforced within libssl.
5889 This issue was reported to OpenSSL Guido Vranken.
5894 * Side channel attack on modular exponentiation
5896 A side-channel attack was found which makes use of cache-bank conflicts on
5897 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5898 of RSA keys. The ability to exploit this issue is limited as it relies on
5899 an attacker who has control of code in a thread running on the same
5900 hyper-threaded core as the victim thread which is performing decryptions.
5902 This issue was reported to OpenSSL by Yuval Yarom, The University of
5903 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5904 Nadia Heninger, University of Pennsylvania with more information at
5905 <http://cachebleed.info>.
5910 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5911 if no keysize is specified with default_bits. This fixes an
5912 omission in an earlier change that changed all RSA/DSA key generation
5913 commands to use 2048 bits by default.
5917 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5919 * DH small subgroups
5921 Historically OpenSSL only ever generated DH parameters based on "safe"
5922 primes. More recently (in version 1.0.2) support was provided for
5923 generating X9.42 style parameter files such as those required for RFC 5114
5924 support. The primes used in such files may not be "safe". Where an
5925 application is using DH configured with parameters based on primes that are
5926 not "safe" then an attacker could use this fact to find a peer's private
5927 DH exponent. This attack requires that the attacker complete multiple
5928 handshakes in which the peer uses the same private DH exponent. For example
5929 this could be used to discover a TLS server's private DH exponent if it's
5930 reusing the private DH exponent or it's using a static DH ciphersuite.
5932 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5933 TLS. It is not on by default. If the option is not set then the server
5934 reuses the same private DH exponent for the life of the server process and
5935 would be vulnerable to this attack. It is believed that many popular
5936 applications do set this option and would therefore not be at risk.
5938 The fix for this issue adds an additional check where a "q" parameter is
5939 available (as is the case in X9.42 based parameters). This detects the
5940 only known attack, and is the only possible defense for static DH
5941 ciphersuites. This could have some performance impact.
5943 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5944 default and cannot be disabled. This could have some performance impact.
5946 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5951 * SSLv2 doesn't block disabled ciphers
5953 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5954 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5955 been disabled, provided that the SSLv2 protocol was not also disabled via
5958 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5959 and Sebastian Schinzel.
5964 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5966 * BN_mod_exp may produce incorrect results on x86_64
5968 There is a carry propagating bug in the x86_64 Montgomery squaring
5969 procedure. No EC algorithms are affected. Analysis suggests that attacks
5970 against RSA and DSA as a result of this defect would be very difficult to
5971 perform and are not believed likely. Attacks against DH are considered just
5972 feasible (although very difficult) because most of the work necessary to
5973 deduce information about a private key may be performed offline. The amount
5974 of resources required for such an attack would be very significant and
5975 likely only accessible to a limited number of attackers. An attacker would
5976 additionally need online access to an unpatched system using the target
5977 private key in a scenario with persistent DH parameters and a private
5978 key that is shared between multiple clients. For example this can occur by
5979 default in OpenSSL DHE based SSL/TLS ciphersuites.
5981 This issue was reported to OpenSSL by Hanno Böck.
5986 * Certificate verify crash with missing PSS parameter
5988 The signature verification routines will crash with a NULL pointer
5989 dereference if presented with an ASN.1 signature using the RSA PSS
5990 algorithm and absent mask generation function parameter. Since these
5991 routines are used to verify certificate signature algorithms this can be
5992 used to crash any certificate verification operation and exploited in a
5993 DoS attack. Any application which performs certificate verification is
5994 vulnerable including OpenSSL clients and servers which enable client
5997 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6002 * X509_ATTRIBUTE memory leak
6004 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6005 memory. This structure is used by the PKCS#7 and CMS routines so any
6006 application which reads PKCS#7 or CMS data from untrusted sources is
6007 affected. SSL/TLS is not affected.
6009 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6015 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6016 This changes the decoding behaviour for some invalid messages,
6017 though the change is mostly in the more lenient direction, and
6018 legacy behaviour is preserved as much as possible.
6022 * In DSA_generate_parameters_ex, if the provided seed is too short,
6025 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6027 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6029 * Alternate chains certificate forgery
6031 During certificate verification, OpenSSL will attempt to find an
6032 alternative certificate chain if the first attempt to build such a chain
6033 fails. An error in the implementation of this logic can mean that an
6034 attacker could cause certain checks on untrusted certificates to be
6035 bypassed, such as the CA flag, enabling them to use a valid leaf
6036 certificate to act as a CA and "issue" an invalid certificate.
6038 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6043 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6045 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6046 incompatibility in the handling of HMAC. The previous ABI has now been
6051 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6053 * Malformed ECParameters causes infinite loop
6055 When processing an ECParameters structure OpenSSL enters an infinite loop
6056 if the curve specified is over a specially malformed binary polynomial
6059 This can be used to perform denial of service against any
6060 system which processes public keys, certificate requests or
6061 certificates. This includes TLS clients and TLS servers with
6062 client authentication enabled.
6064 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6069 * Exploitable out-of-bounds read in X509_cmp_time
6071 X509_cmp_time does not properly check the length of the ASN1_TIME
6072 string and can read a few bytes out of bounds. In addition,
6073 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6076 An attacker can use this to craft malformed certificates and CRLs of
6077 various sizes and potentially cause a segmentation fault, resulting in
6078 a DoS on applications that verify certificates or CRLs. TLS clients
6079 that verify CRLs are affected. TLS clients and servers with client
6080 authentication enabled may be affected if they use custom verification
6083 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6084 independently by Hanno Böck.
6089 * PKCS7 crash with missing EnvelopedContent
6091 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6092 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6093 with missing content and trigger a NULL pointer dereference on parsing.
6095 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6096 structures from untrusted sources are affected. OpenSSL clients and
6097 servers are not affected.
6099 This issue was reported to OpenSSL by Michal Zalewski (Google).
6104 * CMS verify infinite loop with unknown hash function
6106 When verifying a signedData message the CMS code can enter an infinite loop
6107 if presented with an unknown hash function OID. This can be used to perform
6108 denial of service against any system which verifies signedData messages using
6110 This issue was reported to OpenSSL by Johannes Bauer.
6115 * Race condition handling NewSessionTicket
6117 If a NewSessionTicket is received by a multi-threaded client when attempting to
6118 reuse a previous ticket then a race condition can occur potentially leading to
6119 a double free of the ticket data.
6124 * Only support 256-bit or stronger elliptic curves with the
6125 'ecdh_auto' setting (server) or by default (client). Of supported
6126 curves, prefer P-256 (both).
6130 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6132 * ClientHello sigalgs DoS fix
6134 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6135 invalid signature algorithms extension a NULL pointer dereference will
6136 occur. This can be exploited in a DoS attack against the server.
6138 This issue was was reported to OpenSSL by David Ramos of Stanford
6142 *Stephen Henson and Matt Caswell*
6144 * Multiblock corrupted pointer fix
6146 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6147 feature only applies on 64 bit x86 architecture platforms that support AES
6148 NI instructions. A defect in the implementation of "multiblock" can cause
6149 OpenSSL's internal write buffer to become incorrectly set to NULL when
6150 using non-blocking IO. Typically, when the user application is using a
6151 socket BIO for writing, this will only result in a failed connection.
6152 However if some other BIO is used then it is likely that a segmentation
6153 fault will be triggered, thus enabling a potential DoS attack.
6155 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6160 * Segmentation fault in DTLSv1_listen fix
6162 The DTLSv1_listen function is intended to be stateless and processes the
6163 initial ClientHello from many peers. It is common for user code to loop
6164 over the call to DTLSv1_listen until a valid ClientHello is received with
6165 an associated cookie. A defect in the implementation of DTLSv1_listen means
6166 that state is preserved in the SSL object from one invocation to the next
6167 that can lead to a segmentation fault. Errors processing the initial
6168 ClientHello can trigger this scenario. An example of such an error could be
6169 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6172 This issue was reported to OpenSSL by Per Allansson.
6177 * Segmentation fault in ASN1_TYPE_cmp fix
6179 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6180 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6181 certificate signature algorithm consistency this can be used to crash any
6182 certificate verification operation and exploited in a DoS attack. Any
6183 application which performs certificate verification is vulnerable including
6184 OpenSSL clients and servers which enable client authentication.
6189 * Segmentation fault for invalid PSS parameters fix
6191 The signature verification routines will crash with a NULL pointer
6192 dereference if presented with an ASN.1 signature using the RSA PSS
6193 algorithm and invalid parameters. Since these routines are used to verify
6194 certificate signature algorithms this can be used to crash any
6195 certificate verification operation and exploited in a DoS attack. Any
6196 application which performs certificate verification is vulnerable including
6197 OpenSSL clients and servers which enable client authentication.
6199 This issue was was reported to OpenSSL by Brian Carpenter.
6204 * ASN.1 structure reuse memory corruption fix
6206 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6207 memory corruption via an invalid write. Such reuse is and has been
6208 strongly discouraged and is believed to be rare.
6210 Applications that parse structures containing CHOICE or ANY DEFINED BY
6211 components may be affected. Certificate parsing (d2i_X509 and related
6212 functions) are however not affected. OpenSSL clients and servers are
6218 * PKCS7 NULL pointer dereferences fix
6220 The PKCS#7 parsing code does not handle missing outer ContentInfo
6221 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6222 missing content and trigger a NULL pointer dereference on parsing.
6224 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6225 otherwise parse PKCS#7 structures from untrusted sources are
6226 affected. OpenSSL clients and servers are not affected.
6228 This issue was reported to OpenSSL by Michal Zalewski (Google).
6233 * DoS via reachable assert in SSLv2 servers fix
6235 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6236 servers that both support SSLv2 and enable export cipher suites by sending
6237 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6239 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6240 (OpenSSL development team).
6245 * Empty CKE with client auth and DHE fix
6247 If client auth is used then a server can seg fault in the event of a DHE
6248 ciphersuite being selected and a zero length ClientKeyExchange message
6249 being sent by the client. This could be exploited in a DoS attack.
6254 * Handshake with unseeded PRNG fix
6256 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6257 with an unseeded PRNG. The conditions are:
6258 - The client is on a platform where the PRNG has not been seeded
6259 automatically, and the user has not seeded manually
6260 - A protocol specific client method version has been used (i.e. not
6261 SSL_client_methodv23)
6262 - A ciphersuite is used that does not require additional random data from
6263 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6265 If the handshake succeeds then the client random that has been used will
6266 have been generated from a PRNG with insufficient entropy and therefore the
6267 output may be predictable.
6269 For example using the following command with an unseeded openssl will
6270 succeed on an unpatched platform:
6272 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6277 * Use After Free following d2i_ECPrivatekey error fix
6279 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6280 could cause a use after free condition. This, in turn, could cause a double
6281 free in several private key parsing functions (such as d2i_PrivateKey
6282 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6283 for applications that receive EC private keys from untrusted
6284 sources. This scenario is considered rare.
6286 This issue was discovered by the BoringSSL project and fixed in their
6292 * X509_to_X509_REQ NULL pointer deref fix
6294 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6295 the certificate key is invalid. This function is rarely used in practice.
6297 This issue was discovered by Brian Carpenter.
6302 * Removed the export ciphers from the DEFAULT ciphers
6306 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6308 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6309 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6310 So far those who have to target multiple platforms would compromise
6311 and argue that binary targeting say ARMv5 would still execute on
6312 ARMv8. "Universal" build resolves this compromise by providing
6313 near-optimal performance even on newer platforms.
6317 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6318 (other platforms pending).
6320 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6322 * Add support for the SignedCertificateTimestampList certificate and
6323 OCSP response extensions from RFC6962.
6327 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6328 for corner cases. (Certain input points at infinity could lead to
6329 bogus results, with non-infinity inputs mapped to infinity too.)
6333 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6334 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6335 common cases are optimized and there still is room for further
6336 improvements. Vector Permutation AES for Altivec is also added.
6340 * Add support for little-endian ppc64 Linux target.
6342 *Marcelo Cerri (IBM)*
6344 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6345 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6346 are optimized and there still is room for further improvements.
6347 Both 32- and 64-bit modes are supported.
6349 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6351 * Improved ARMv7 NEON support.
6355 * Support for SPARC Architecture 2011 crypto extensions, first
6356 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6357 SHA256/512, MD5, GHASH and modular exponentiation.
6359 *Andy Polyakov, David Miller*
6361 * Accelerated modular exponentiation for Intel processors, a.k.a.
6364 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6366 * Support for new and upcoming Intel processors, including AVX2,
6367 BMI and SHA ISA extensions. This includes additional "stitched"
6368 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6371 This work was sponsored by Intel Corp.
6375 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6376 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6377 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6381 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6382 this fixes a limitation in previous versions of OpenSSL.
6386 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6387 MGF1 digest and OAEP label.
6391 * Add EVP support for key wrapping algorithms, to avoid problems with
6392 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6393 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6394 algorithms and include tests cases.
6398 * Add functions to allocate and set the fields of an ECDSA_METHOD
6401 *Douglas E. Engert, Steve Henson*
6403 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6404 difference in days and seconds between two tm or ASN1_TIME structures.
6408 * Add -rev test option to s_server to just reverse order of characters
6409 received by client and send back to server. Also prints an abbreviated
6410 summary of the connection parameters.
6414 * New option -brief for s_client and s_server to print out a brief summary
6415 of connection parameters.
6419 * Add callbacks for arbitrary TLS extensions.
6421 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6423 * New option -crl_download in several openssl utilities to download CRLs
6424 from CRLDP extension in certificates.
6428 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6432 * New function X509_CRL_diff to generate a delta CRL from the difference
6433 of two full CRLs. Add support to "crl" utility.
6437 * New functions to set lookup_crls function and to retrieve
6438 X509_STORE from X509_STORE_CTX.
6442 * Print out deprecated issuer and subject unique ID fields in
6447 * Extend OCSP I/O functions so they can be used for simple general purpose
6448 HTTP as well as OCSP. New wrapper function which can be used to download
6449 CRLs using the OCSP API.
6453 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6457 * `SSL_CONF*` functions. These provide a common framework for application
6458 configuration using configuration files or command lines.
6462 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6463 message callback and prints the results. Needs compile time option
6464 "enable-ssl-trace". New options to s_client and s_server to enable
6469 * New ctrl and macro to retrieve supported points extensions.
6470 Print out extension in s_server and s_client.
6474 * New functions to retrieve certificate signature and signature
6479 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6484 * New Suite B modes for TLS code. These use and enforce the requirements
6485 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6486 only use Suite B curves. The Suite B modes can be set by using the
6487 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6491 * New chain verification flags for Suite B levels of security. Check
6492 algorithms are acceptable when flags are set in X509_verify_cert.
6496 * Make tls1_check_chain return a set of flags indicating checks passed
6497 by a certificate chain. Add additional tests to handle client
6498 certificates: checks for matching certificate type and issuer name
6503 * If an attempt is made to use a signature algorithm not in the peer
6504 preference list abort the handshake. If client has no suitable
6505 signature algorithms in response to a certificate request do not
6506 use the certificate.
6510 * If server EC tmp key is not in client preference list abort handshake.
6514 * Add support for certificate stores in CERT structure. This makes it
6515 possible to have different stores per SSL structure or one store in
6516 the parent SSL_CTX. Include distinct stores for certificate chain
6517 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6518 to build and store a certificate chain in CERT structure: returning
6519 an error if the chain cannot be built: this will allow applications
6520 to test if a chain is correctly configured.
6522 Note: if the CERT based stores are not set then the parent SSL_CTX
6523 store is used to retain compatibility with existing behaviour.
6527 * New function ssl_set_client_disabled to set a ciphersuite disabled
6528 mask based on the current session, check mask when sending client
6529 hello and checking the requested ciphersuite.
6533 * New ctrls to retrieve and set certificate types in a certificate
6534 request message. Print out received values in s_client. If certificate
6535 types is not set with custom values set sensible values based on
6536 supported signature algorithms.
6540 * Support for distinct client and server supported signature algorithms.
6544 * Add certificate callback. If set this is called whenever a certificate
6545 is required by client or server. An application can decide which
6546 certificate chain to present based on arbitrary criteria: for example
6547 supported signature algorithms. Add very simple example to s_server.
6548 This fixes many of the problems and restrictions of the existing client
6549 certificate callback: for example you can now clear an existing
6550 certificate and specify the whole chain.
6554 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6555 the certificate can be used for (if anything). Set valid_flags field
6556 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6557 to have similar checks in it.
6559 Add new "cert_flags" field to CERT structure and include a "strict mode".
6560 This enforces some TLS certificate requirements (such as only permitting
6561 certificate signature algorithms contained in the supported algorithms
6562 extension) which some implementations ignore: this option should be used
6563 with caution as it could cause interoperability issues.
6567 * Update and tidy signature algorithm extension processing. Work out
6568 shared signature algorithms based on preferences and peer algorithms
6569 and print them out in s_client and s_server. Abort handshake if no
6570 shared signature algorithms.
6574 * Add new functions to allow customised supported signature algorithms
6575 for SSL and SSL_CTX structures. Add options to s_client and s_server
6580 * New function SSL_certs_clear() to delete all references to certificates
6581 from an SSL structure. Before this once a certificate had been added
6582 it couldn't be removed.
6586 * Integrate hostname, email address and IP address checking with certificate
6587 verification. New verify options supporting checking in openssl utility.
6591 * Fixes and wildcard matching support to hostname and email checking
6592 functions. Add manual page.
6594 *Florian Weimer (Red Hat Product Security Team)*
6596 * New functions to check a hostname email or IP address against a
6597 certificate. Add options x509 utility to print results of checks against
6602 * Fix OCSP checking.
6604 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6606 * Initial experimental support for explicitly trusted non-root CAs.
6607 OpenSSL still tries to build a complete chain to a root but if an
6608 intermediate CA has a trust setting included that is used. The first
6609 setting is used: whether to trust (e.g., -addtrust option to the x509
6614 * Add -trusted_first option which attempts to find certificates in the
6615 trusted store even if an untrusted chain is also supplied.
6619 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6620 platform support for Linux and Android.
6624 * Support for linux-x32, ILP32 environment in x86_64 framework.
6628 * Experimental multi-implementation support for FIPS capable OpenSSL.
6629 When in FIPS mode the approved implementations are used as normal,
6630 when not in FIPS mode the internal unapproved versions are used instead.
6631 This means that the FIPS capable OpenSSL isn't forced to use the
6632 (often lower performance) FIPS implementations outside FIPS mode.
6636 * Transparently support X9.42 DH parameters when calling
6637 PEM_read_bio_DHparameters. This means existing applications can handle
6638 the new parameter format automatically.
6642 * Initial experimental support for X9.42 DH parameter format: mainly
6643 to support use of 'q' parameter for RFC5114 parameters.
6647 * Add DH parameters from RFC5114 including test data to dhtest.
6651 * Support for automatic EC temporary key parameter selection. If enabled
6652 the most preferred EC parameters are automatically used instead of
6653 hardcoded fixed parameters. Now a server just has to call:
6654 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6655 support ECDH and use the most appropriate parameters.
6659 * Enhance and tidy EC curve and point format TLS extension code. Use
6660 static structures instead of allocation if default values are used.
6661 New ctrls to set curves we wish to support and to retrieve shared curves.
6662 Print out shared curves in s_server. New options to s_server and s_client
6663 to set list of supported curves.
6667 * New ctrls to retrieve supported signature algorithms and
6668 supported curve values as an array of NIDs. Extend openssl utility
6669 to print out received values.
6673 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6674 between NIDs and the more common NIST names such as "P-256". Enhance
6675 ecparam utility and ECC method to recognise the NIST names for curves.
6679 * Enhance SSL/TLS certificate chain handling to support different
6680 chains for each certificate instead of one chain in the parent SSL_CTX.
6684 * Support for fixed DH ciphersuite client authentication: where both
6685 server and client use DH certificates with common parameters.
6689 * Support for fixed DH ciphersuites: those requiring DH server
6694 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6696 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6697 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6698 X509_CINF_get_signature were reverted post internal team review.
6703 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6705 * OCSP Status Request extension unbounded memory growth
6707 A malicious client can send an excessively large OCSP Status Request
6708 extension. If that client continually requests renegotiation, sending a
6709 large OCSP Status Request extension each time, then there will be unbounded
6710 memory growth on the server. This will eventually lead to a Denial Of
6711 Service attack through memory exhaustion. Servers with a default
6712 configuration are vulnerable even if they do not support OCSP. Builds using
6713 the "no-ocsp" build time option are not affected.
6715 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6720 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6723 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6729 * OOB write in MDC2_Update()
6731 An overflow can occur in MDC2_Update() either if called directly or
6732 through the EVP_DigestUpdate() function using MDC2. If an attacker
6733 is able to supply very large amounts of input data after a previous
6734 call to EVP_EncryptUpdate() with a partial block then a length check
6735 can overflow resulting in a heap corruption.
6737 The amount of data needed is comparable to SIZE_MAX which is impractical
6740 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6745 * Malformed SHA512 ticket DoS
6747 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6748 DoS attack where a malformed ticket will result in an OOB read which will
6751 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6752 a custom server callback and ticket lookup mechanism.
6754 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6759 * OOB write in BN_bn2dec()
6761 The function BN_bn2dec() does not check the return value of BN_div_word().
6762 This can cause an OOB write if an application uses this function with an
6763 overly large BIGNUM. This could be a problem if an overly large certificate
6764 or CRL is printed out from an untrusted source. TLS is not affected because
6765 record limits will reject an oversized certificate before it is parsed.
6767 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6772 * OOB read in TS_OBJ_print_bio()
6774 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6775 the total length the OID text representation would use and not the amount
6776 of data written. This will result in OOB reads when large OIDs are
6779 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6784 * Pointer arithmetic undefined behaviour
6786 Avoid some undefined pointer arithmetic
6788 A common idiom in the codebase is to check limits in the following manner:
6791 Where "p" points to some malloc'd data of SIZE bytes and
6794 "len" here could be from some externally supplied data (e.g. from a TLS
6797 The rules of C pointer arithmetic are such that "p + len" is only well
6798 defined where len <= SIZE. Therefore, the above idiom is actually
6799 undefined behaviour.
6801 For example this could cause problems if some malloc implementation
6802 provides an address for "p" such that "p + len" actually overflows for
6803 values of len that are too big and therefore p + len < limit.
6805 This issue was reported to OpenSSL by Guido Vranken
6810 * Constant time flag not preserved in DSA signing
6812 Operations in the DSA signing algorithm should run in constant time in
6813 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6814 implementation means that a non-constant time codepath is followed for
6815 certain operations. This has been demonstrated through a cache-timing
6816 attack to be sufficient for an attacker to recover the private DSA key.
6818 This issue was reported by César Pereida (Aalto University), Billy Brumley
6819 (Tampere University of Technology), and Yuval Yarom (The University of
6820 Adelaide and NICTA).
6825 * DTLS buffered message DoS
6827 In a DTLS connection where handshake messages are delivered out-of-order
6828 those messages that OpenSSL is not yet ready to process will be buffered
6829 for later use. Under certain circumstances, a flaw in the logic means that
6830 those messages do not get removed from the buffer even though the handshake
6831 has been completed. An attacker could force up to approx. 15 messages to
6832 remain in the buffer when they are no longer required. These messages will
6833 be cleared when the DTLS connection is closed. The default maximum size for
6834 a message is 100k. Therefore, the attacker could force an additional 1500k
6835 to be consumed per connection. By opening many simultaneous connections an
6836 attacker could cause a DoS attack through memory exhaustion.
6838 This issue was reported to OpenSSL by Quan Luo.
6843 * DTLS replay protection DoS
6845 A flaw in the DTLS replay attack protection mechanism means that records
6846 that arrive for future epochs update the replay protection "window" before
6847 the MAC for the record has been validated. This could be exploited by an
6848 attacker by sending a record for the next epoch (which does not have to
6849 decrypt or have a valid MAC), with a very large sequence number. This means
6850 that all subsequent legitimate packets are dropped causing a denial of
6851 service for a specific DTLS connection.
6853 This issue was reported to OpenSSL by the OCAP audit team.
6858 * Certificate message OOB reads
6860 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6861 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6862 theoretical DoS risk but this has not been observed in practice on common
6865 The messages affected are client certificate, client certificate request
6866 and server certificate. As a result the attack can only be performed
6867 against a client or a server which enables client authentication.
6869 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6874 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6876 * Prevent padding oracle in AES-NI CBC MAC check
6878 A MITM attacker can use a padding oracle attack to decrypt traffic
6879 when the connection uses an AES CBC cipher and the server support
6882 This issue was introduced as part of the fix for Lucky 13 padding
6883 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6884 constant time by making sure that always the same bytes are read and
6885 compared against either the MAC or padding bytes. But it no longer
6886 checked that there was enough data to have both the MAC and padding
6889 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6894 * Fix EVP_EncodeUpdate overflow
6896 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6897 Base64 encoding of binary data. If an attacker is able to supply very large
6898 amounts of input data then a length check can overflow resulting in a heap
6901 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6902 the `PEM_write_bio*` family of functions. These are mainly used within the
6903 OpenSSL command line applications, so any application which processes data
6904 from an untrusted source and outputs it as a PEM file should be considered
6905 vulnerable to this issue. User applications that call these APIs directly
6906 with large amounts of untrusted data may also be vulnerable.
6908 This issue was reported by Guido Vranken.
6913 * Fix EVP_EncryptUpdate overflow
6915 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6916 is able to supply very large amounts of input data after a previous call to
6917 EVP_EncryptUpdate() with a partial block then a length check can overflow
6918 resulting in a heap corruption. Following an analysis of all OpenSSL
6919 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6920 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6921 the first called function after an EVP_EncryptInit(), and therefore that
6922 specific call must be safe. The second form is where the length passed to
6923 EVP_EncryptUpdate() can be seen from the code to be some small value and
6924 therefore there is no possibility of an overflow. Since all instances are
6925 one of these two forms, it is believed that there can be no overflows in
6926 internal code due to this problem. It should be noted that
6927 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6928 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6929 of these calls have also been analysed too and it is believed there are no
6930 instances in internal usage where an overflow could occur.
6932 This issue was reported by Guido Vranken.
6937 * Prevent ASN.1 BIO excessive memory allocation
6939 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6940 a short invalid encoding can casuse allocation of large amounts of memory
6941 potentially consuming excessive resources or exhausting memory.
6943 Any application parsing untrusted data through d2i BIO functions is
6944 affected. The memory based functions such as d2i_X509() are *not* affected.
6945 Since the memory based functions are used by the TLS library, TLS
6946 applications are not affected.
6948 This issue was reported by Brian Carpenter.
6955 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6956 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6957 in arbitrary stack data being returned in the buffer.
6959 This issue was reported by Guido Vranken.
6964 * Modify behavior of ALPN to invoke callback after SNI/servername
6965 callback, such that updates to the SSL_CTX affect ALPN.
6969 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6974 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6975 methods are enabled and ssl2 is disabled the methods return NULL.
6979 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6981 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6982 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6983 provide any "EXPORT" or "LOW" strength ciphers.
6987 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6988 is by default disabled at build-time. Builds that are not configured with
6989 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6990 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6991 will need to explicitly call either of:
6993 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6995 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6997 as appropriate. Even if either of those is used, or the application
6998 explicitly uses the version-specific SSLv2_method() or its client and
6999 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7000 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7001 ciphers, and SSLv2 56-bit DES are no longer available.
7006 * Fix a double-free in DSA code
7008 A double free bug was discovered when OpenSSL parses malformed DSA private
7009 keys and could lead to a DoS attack or memory corruption for applications
7010 that receive DSA private keys from untrusted sources. This scenario is
7013 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7019 * Disable SRP fake user seed to address a server memory leak.
7021 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7023 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7024 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7025 was changed to ignore the "fake user" SRP seed, even if the seed
7028 Users should use SRP_VBASE_get1_by_user instead. Note that in
7029 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7030 also that even though configuring the SRP seed attempts to hide
7031 invalid usernames by continuing the handshake with fake
7032 credentials, this behaviour is not constant time and no strong
7033 guarantees are made that the handshake is indistinguishable from
7034 that of a valid user.
7039 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7041 In the BN_hex2bn function the number of hex digits is calculated using an
7042 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7043 large values of `i` this can result in `bn_expand` not allocating any
7044 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7045 field as NULL leading to a subsequent NULL ptr deref. For very large values
7046 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7047 In this case memory is allocated to the internal BIGNUM data field, but it
7048 is insufficiently sized leading to heap corruption. A similar issue exists
7049 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7050 is ever called by user applications with very large untrusted hex/dec data.
7051 This is anticipated to be a rare occurrence.
7053 All OpenSSL internal usage of these functions use data that is not expected
7054 to be untrusted, e.g. config file data or application command line
7055 arguments. If user developed applications generate config file data based
7056 on untrusted data then it is possible that this could also lead to security
7057 consequences. This is also anticipated to be rare.
7059 This issue was reported to OpenSSL by Guido Vranken.
7064 * Fix memory issues in `BIO_*printf` functions
7066 The internal `fmtstr` function used in processing a "%s" format string in
7067 the `BIO_*printf` functions could overflow while calculating the length of a
7068 string and cause an OOB read when printing very long strings.
7070 Additionally the internal `doapr_outch` function can attempt to write to an
7071 OOB memory location (at an offset from the NULL pointer) in the event of a
7072 memory allocation failure. In 1.0.2 and below this could be caused where
7073 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7074 could be in processing a very long "%s" format string. Memory leaks can
7077 The first issue may mask the second issue dependent on compiler behaviour.
7078 These problems could enable attacks where large amounts of untrusted data
7079 is passed to the `BIO_*printf` functions. If applications use these functions
7080 in this way then they could be vulnerable. OpenSSL itself uses these
7081 functions when printing out human-readable dumps of ASN.1 data. Therefore
7082 applications that print this data could be vulnerable if the data is from
7083 untrusted sources. OpenSSL command line applications could also be
7084 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7085 as command line arguments.
7087 Libssl is not considered directly vulnerable. Additionally certificates etc
7088 received via remote connections via libssl are also unlikely to be able to
7089 trigger these issues because of message size limits enforced within libssl.
7091 This issue was reported to OpenSSL Guido Vranken.
7096 * Side channel attack on modular exponentiation
7098 A side-channel attack was found which makes use of cache-bank conflicts on
7099 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7100 of RSA keys. The ability to exploit this issue is limited as it relies on
7101 an attacker who has control of code in a thread running on the same
7102 hyper-threaded core as the victim thread which is performing decryptions.
7104 This issue was reported to OpenSSL by Yuval Yarom, The University of
7105 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7106 Nadia Heninger, University of Pennsylvania with more information at
7107 <http://cachebleed.info>.
7112 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7113 if no keysize is specified with default_bits. This fixes an
7114 omission in an earlier change that changed all RSA/DSA key generation
7115 commands to use 2048 bits by default.
7119 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7121 * Protection for DH small subgroup attacks
7123 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7124 switched on by default and cannot be disabled. This could have some
7129 * SSLv2 doesn't block disabled ciphers
7131 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7132 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7133 been disabled, provided that the SSLv2 protocol was not also disabled via
7136 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7137 and Sebastian Schinzel.
7142 * Reject DH handshakes with parameters shorter than 1024 bits.
7146 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7148 * Certificate verify crash with missing PSS parameter
7150 The signature verification routines will crash with a NULL pointer
7151 dereference if presented with an ASN.1 signature using the RSA PSS
7152 algorithm and absent mask generation function parameter. Since these
7153 routines are used to verify certificate signature algorithms this can be
7154 used to crash any certificate verification operation and exploited in a
7155 DoS attack. Any application which performs certificate verification is
7156 vulnerable including OpenSSL clients and servers which enable client
7159 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7164 * X509_ATTRIBUTE memory leak
7166 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7167 memory. This structure is used by the PKCS#7 and CMS routines so any
7168 application which reads PKCS#7 or CMS data from untrusted sources is
7169 affected. SSL/TLS is not affected.
7171 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7177 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7178 This changes the decoding behaviour for some invalid messages,
7179 though the change is mostly in the more lenient direction, and
7180 legacy behaviour is preserved as much as possible.
7184 * In DSA_generate_parameters_ex, if the provided seed is too short,
7185 use a random seed, as already documented.
7187 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7189 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7191 * Alternate chains certificate forgery
7193 During certificate verfification, OpenSSL will attempt to find an
7194 alternative certificate chain if the first attempt to build such a chain
7195 fails. An error in the implementation of this logic can mean that an
7196 attacker could cause certain checks on untrusted certificates to be
7197 bypassed, such as the CA flag, enabling them to use a valid leaf
7198 certificate to act as a CA and "issue" an invalid certificate.
7200 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7206 * Race condition handling PSK identify hint
7208 If PSK identity hints are received by a multi-threaded client then
7209 the values are wrongly updated in the parent SSL_CTX structure. This can
7210 result in a race condition potentially leading to a double free of the
7216 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7218 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7219 incompatibility in the handling of HMAC. The previous ABI has now been
7222 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7224 * Malformed ECParameters causes infinite loop
7226 When processing an ECParameters structure OpenSSL enters an infinite loop
7227 if the curve specified is over a specially malformed binary polynomial
7230 This can be used to perform denial of service against any
7231 system which processes public keys, certificate requests or
7232 certificates. This includes TLS clients and TLS servers with
7233 client authentication enabled.
7235 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7240 * Exploitable out-of-bounds read in X509_cmp_time
7242 X509_cmp_time does not properly check the length of the ASN1_TIME
7243 string and can read a few bytes out of bounds. In addition,
7244 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7247 An attacker can use this to craft malformed certificates and CRLs of
7248 various sizes and potentially cause a segmentation fault, resulting in
7249 a DoS on applications that verify certificates or CRLs. TLS clients
7250 that verify CRLs are affected. TLS clients and servers with client
7251 authentication enabled may be affected if they use custom verification
7254 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7255 independently by Hanno Böck.
7260 * PKCS7 crash with missing EnvelopedContent
7262 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7263 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7264 with missing content and trigger a NULL pointer dereference on parsing.
7266 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7267 structures from untrusted sources are affected. OpenSSL clients and
7268 servers are not affected.
7270 This issue was reported to OpenSSL by Michal Zalewski (Google).
7275 * CMS verify infinite loop with unknown hash function
7277 When verifying a signedData message the CMS code can enter an infinite loop
7278 if presented with an unknown hash function OID. This can be used to perform
7279 denial of service against any system which verifies signedData messages using
7281 This issue was reported to OpenSSL by Johannes Bauer.
7286 * Race condition handling NewSessionTicket
7288 If a NewSessionTicket is received by a multi-threaded client when attempting to
7289 reuse a previous ticket then a race condition can occur potentially leading to
7290 a double free of the ticket data.
7295 * Reject DH handshakes with parameters shorter than 768 bits.
7297 *Kurt Roeckx and Emilia Kasper*
7299 * dhparam: generate 2048-bit parameters by default.
7301 *Kurt Roeckx and Emilia Kasper*
7303 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7305 * Segmentation fault in ASN1_TYPE_cmp fix
7307 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7308 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7309 certificate signature algorithm consistency this can be used to crash any
7310 certificate verification operation and exploited in a DoS attack. Any
7311 application which performs certificate verification is vulnerable including
7312 OpenSSL clients and servers which enable client authentication.
7317 * ASN.1 structure reuse memory corruption fix
7319 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7320 memory corruption via an invalid write. Such reuse is and has been
7321 strongly discouraged and is believed to be rare.
7323 Applications that parse structures containing CHOICE or ANY DEFINED BY
7324 components may be affected. Certificate parsing (d2i_X509 and related
7325 functions) are however not affected. OpenSSL clients and servers are
7331 * PKCS7 NULL pointer dereferences fix
7333 The PKCS#7 parsing code does not handle missing outer ContentInfo
7334 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7335 missing content and trigger a NULL pointer dereference on parsing.
7337 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7338 otherwise parse PKCS#7 structures from untrusted sources are
7339 affected. OpenSSL clients and servers are not affected.
7341 This issue was reported to OpenSSL by Michal Zalewski (Google).
7346 * DoS via reachable assert in SSLv2 servers fix
7348 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7349 servers that both support SSLv2 and enable export cipher suites by sending
7350 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7352 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7353 (OpenSSL development team).
7358 * Use After Free following d2i_ECPrivatekey error fix
7360 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7361 could cause a use after free condition. This, in turn, could cause a double
7362 free in several private key parsing functions (such as d2i_PrivateKey
7363 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7364 for applications that receive EC private keys from untrusted
7365 sources. This scenario is considered rare.
7367 This issue was discovered by the BoringSSL project and fixed in their
7373 * X509_to_X509_REQ NULL pointer deref fix
7375 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7376 the certificate key is invalid. This function is rarely used in practice.
7378 This issue was discovered by Brian Carpenter.
7383 * Removed the export ciphers from the DEFAULT ciphers
7387 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7389 * Build fixes for the Windows and OpenVMS platforms
7391 *Matt Caswell and Richard Levitte*
7393 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7395 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7396 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7397 dereference. This could lead to a Denial Of Service attack. Thanks to
7398 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7403 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7404 dtls1_buffer_record function under certain conditions. In particular this
7405 could occur if an attacker sent repeated DTLS records with the same
7406 sequence number but for the next epoch. The memory leak could be exploited
7407 by an attacker in a Denial of Service attack through memory exhaustion.
7408 Thanks to Chris Mueller for reporting this issue.
7413 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7414 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7415 method would be set to NULL which could later result in a NULL pointer
7416 dereference. Thanks to Frank Schmirler for reporting this issue.
7421 * Abort handshake if server key exchange message is omitted for ephemeral
7424 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7425 reporting this issue.
7430 * Remove non-export ephemeral RSA code on client and server. This code
7431 violated the TLS standard by allowing the use of temporary RSA keys in
7432 non-export ciphersuites and could be used by a server to effectively
7433 downgrade the RSA key length used to a value smaller than the server
7434 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7435 INRIA or reporting this issue.
7440 * Fixed issue where DH client certificates are accepted without verification.
7441 An OpenSSL server will accept a DH certificate for client authentication
7442 without the certificate verify message. This effectively allows a client to
7443 authenticate without the use of a private key. This only affects servers
7444 which trust a client certificate authority which issues certificates
7445 containing DH keys: these are extremely rare and hardly ever encountered.
7446 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7452 * Ensure that the session ID context of an SSL is updated when its
7453 SSL_CTX is updated via SSL_set_SSL_CTX.
7455 The session ID context is typically set from the parent SSL_CTX,
7456 and can vary with the CTX.
7460 * Fix various certificate fingerprint issues.
7462 By using non-DER or invalid encodings outside the signed portion of a
7463 certificate the fingerprint can be changed without breaking the signature.
7464 Although no details of the signed portion of the certificate can be changed
7465 this can cause problems with some applications: e.g. those using the
7466 certificate fingerprint for blacklists.
7468 1. Reject signatures with non zero unused bits.
7470 If the BIT STRING containing the signature has non zero unused bits reject
7471 the signature. All current signature algorithms require zero unused bits.
7473 2. Check certificate algorithm consistency.
7475 Check the AlgorithmIdentifier inside TBS matches the one in the
7476 certificate signature. NB: this will result in signature failure
7477 errors for some broken certificates.
7479 Thanks to Konrad Kraszewski from Google for reporting this issue.
7481 3. Check DSA/ECDSA signatures use DER.
7483 Re-encode DSA/ECDSA signatures and compare with the original received
7484 signature. Return an error if there is a mismatch.
7486 This will reject various cases including garbage after signature
7487 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7488 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7489 (negative or with leading zeroes).
7491 Further analysis was conducted and fixes were developed by Stephen Henson
7492 of the OpenSSL core team.
7498 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7499 results on some platforms, including x86_64. This bug occurs at random
7500 with a very low probability, and is not known to be exploitable in any
7501 way, though its exact impact is difficult to determine. Thanks to Pieter
7502 Wuille (Blockstream) who reported this issue and also suggested an initial
7503 fix. Further analysis was conducted by the OpenSSL development team and
7504 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7505 the OpenSSL core team.
7510 * Do not resume sessions on the server if the negotiated protocol
7511 version does not match the session's version. Resuming with a different
7512 version, while not strictly forbidden by the RFC, is of questionable
7513 sanity and breaks all known clients.
7515 *David Benjamin, Emilia Käsper*
7517 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7518 early CCS messages during renegotiation. (Note that because
7519 renegotiation is encrypted, this early CCS was not exploitable.)
7523 * Tighten client-side session ticket handling during renegotiation:
7524 ensure that the client only accepts a session ticket if the server sends
7525 the extension anew in the ServerHello. Previously, a TLS client would
7526 reuse the old extension state and thus accept a session ticket if one was
7527 announced in the initial ServerHello.
7529 Similarly, ensure that the client requires a session ticket if one
7530 was advertised in the ServerHello. Previously, a TLS client would
7531 ignore a missing NewSessionTicket message.
7535 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7539 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7540 sends a carefully crafted handshake message, to cause OpenSSL to fail
7541 to free up to 64k of memory causing a memory leak. This could be
7542 exploited in a Denial Of Service attack. This issue affects OpenSSL
7543 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7544 whether SRTP is used or configured. Implementations of OpenSSL that
7545 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7547 The fix was developed by the OpenSSL team.
7552 * Session Ticket Memory Leak.
7554 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7555 integrity of that ticket is first verified. In the event of a session
7556 ticket integrity check failing, OpenSSL will fail to free memory
7557 causing a memory leak. By sending a large number of invalid session
7558 tickets an attacker could exploit this issue in a Denial Of Service
7564 * Build option no-ssl3 is incomplete.
7566 When OpenSSL is configured with "no-ssl3" as a build option, servers
7567 could accept and complete an SSL 3.0 handshake, and clients could be
7568 configured to send them.
7571 *Akamai and the OpenSSL team*
7573 * Add support for TLS_FALLBACK_SCSV.
7574 Client applications doing fallback retries should call
7575 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7578 *Adam Langley, Bodo Moeller*
7580 * Add additional DigestInfo checks.
7582 Re-encode DigestInto in DER and check against the original when
7583 verifying RSA signature: this will reject any improperly encoded
7584 DigestInfo structures.
7586 Note: this is a precautionary measure and no attacks are currently known.
7590 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7592 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7593 SRP code can be overrun an internal buffer. Add sanity check that
7594 g, A, B < N to SRP code.
7596 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7597 Group for discovering this issue.
7602 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7603 TLS 1.0 instead of higher protocol versions when the ClientHello message
7604 is badly fragmented. This allows a man-in-the-middle attacker to force a
7605 downgrade to TLS 1.0 even if both the server and the client support a
7606 higher protocol version, by modifying the client's TLS records.
7608 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7609 researching this issue.
7614 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7615 to a denial of service attack. A malicious server can crash the client
7616 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7617 ciphersuite and sending carefully crafted handshake messages.
7619 Thanks to Felix Gröbert (Google) for discovering and researching this
7625 * By sending carefully crafted DTLS packets an attacker could cause openssl
7626 to leak memory. This can be exploited through a Denial of Service attack.
7627 Thanks to Adam Langley for discovering and researching this issue.
7632 * An attacker can force openssl to consume large amounts of memory whilst
7633 processing DTLS handshake messages. This can be exploited through a
7634 Denial of Service attack.
7635 Thanks to Adam Langley for discovering and researching this issue.
7640 * An attacker can force an error condition which causes openssl to crash
7641 whilst processing DTLS packets due to memory being freed twice. This
7642 can be exploited through a Denial of Service attack.
7643 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7649 * If a multithreaded client connects to a malicious server using a resumed
7650 session and the server sends an ec point format extension it could write
7651 up to 255 bytes to freed memory.
7653 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7659 * A malicious server can crash an OpenSSL client with a null pointer
7660 dereference (read) by specifying an SRP ciphersuite even though it was not
7661 properly negotiated with the client. This can be exploited through a
7662 Denial of Service attack.
7664 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7665 discovering and researching this issue.
7670 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7671 X509_name_oneline, X509_name_print_ex et al. to leak some information
7672 from the stack. Applications may be affected if they echo pretty printing
7673 output to the attacker.
7675 Thanks to Ivan Fratric (Google) for discovering this issue.
7678 *Emilia Käsper, and Steve Henson*
7680 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7681 for corner cases. (Certain input points at infinity could lead to
7682 bogus results, with non-infinity inputs mapped to infinity too.)
7686 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7688 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7689 handshake can force the use of weak keying material in OpenSSL
7690 SSL/TLS clients and servers.
7692 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7693 researching this issue. ([CVE-2014-0224])
7695 *KIKUCHI Masashi, Steve Henson*
7697 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7698 OpenSSL DTLS client the code can be made to recurse eventually crashing
7701 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7704 *Imre Rad, Steve Henson*
7706 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7707 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7708 client or server. This is potentially exploitable to run arbitrary
7709 code on a vulnerable client or server.
7711 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7713 *Jüri Aedla, Steve Henson*
7715 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7716 are subject to a denial of service attack.
7718 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7719 this issue. ([CVE-2014-3470])
7721 *Felix Gröbert, Ivan Fratric, Steve Henson*
7723 * Harmonize version and its documentation. -f flag is used to display
7726 *mancha <mancha1@zoho.com>*
7728 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7729 in i2d_ECPrivateKey.
7731 *mancha <mancha1@zoho.com>*
7733 * Fix some double frees. These are not thought to be exploitable.
7735 *mancha <mancha1@zoho.com>*
7737 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7739 * A missing bounds check in the handling of the TLS heartbeat extension
7740 can be used to reveal up to 64k of memory to a connected client or
7743 Thanks for Neel Mehta of Google Security for discovering this bug and to
7744 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7745 preparing the fix ([CVE-2014-0160])
7747 *Adam Langley, Bodo Moeller*
7749 * Fix for the attack described in the paper "Recovering OpenSSL
7750 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7751 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7752 <http://eprint.iacr.org/2014/140>
7754 Thanks to Yuval Yarom and Naomi Benger for discovering this
7755 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7757 *Yuval Yarom and Naomi Benger*
7759 * TLS pad extension: draft-agl-tls-padding-03
7761 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7762 TLS client Hello record length value would otherwise be > 255 and
7763 less that 512 pad with a dummy extension containing zeroes so it
7764 is at least 512 bytes long.
7766 *Adam Langley, Steve Henson*
7768 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7770 * Fix for TLS record tampering bug. A carefully crafted invalid
7771 handshake could crash OpenSSL with a NULL pointer exception.
7772 Thanks to Anton Johansson for reporting this issues.
7775 * Keep original DTLS digest and encryption contexts in retransmission
7776 structures so we can use the previous session parameters if they need
7777 to be resent. ([CVE-2013-6450])
7781 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7782 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7783 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7784 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7785 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7786 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7788 *Rob Stradling, Adam Langley*
7790 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7792 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7793 supporting platforms or when small records were transferred.
7795 *Andy Polyakov, Steve Henson*
7797 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7799 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7801 This addresses the flaw in CBC record processing discovered by
7802 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7803 at: <http://www.isg.rhul.ac.uk/tls/>
7805 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7806 Security Group at Royal Holloway, University of London
7807 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7808 Emilia Käsper for the initial patch.
7811 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7813 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7814 ciphersuites which can be exploited in a denial of service attack.
7815 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7816 and detecting this bug and to Wolfgang Ettlinger
7817 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7822 * Return an error when checking OCSP signatures when key is NULL.
7823 This fixes a DoS attack. ([CVE-2013-0166])
7827 * Make openssl verify return errors.
7829 *Chris Palmer <palmer@google.com> and Ben Laurie*
7831 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7832 the right response is stapled. Also change SSL_get_certificate()
7833 so it returns the certificate actually sent.
7834 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7836 *Rob Stradling <rob.stradling@comodo.com>*
7838 * Fix possible deadlock when decoding public keys.
7842 * Don't use TLS 1.0 record version number in initial client hello
7847 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7849 * Sanity check record length before skipping explicit IV in TLS
7850 1.2, 1.1 and DTLS to fix DoS attack.
7852 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7853 fuzzing as a service testing platform.
7858 * Initialise tkeylen properly when encrypting CMS messages.
7859 Thanks to Solar Designer of Openwall for reporting this issue.
7863 * In FIPS mode don't try to use composite ciphers as they are not
7868 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7870 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7871 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7872 mean any application compiled against OpenSSL 1.0.0 headers setting
7873 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7874 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7875 0x10000000L Any application which was previously compiled against
7876 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7877 will need to be recompiled as a result. Letting be results in
7878 inability to disable specifically TLS 1.1 and in client context,
7879 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7883 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7884 disable just protocol X, but all protocols above X *if* there are
7885 protocols *below* X still enabled. In more practical terms it means
7886 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7887 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7888 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7893 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7895 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7896 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7897 in CRYPTO_realloc_clean.
7899 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7900 issue and to Adam Langley <agl@chromium.org> for fixing it.
7903 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7905 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7909 * Workarounds for some broken servers that "hang" if a client hello
7910 record length exceeds 255 bytes.
7912 1. Do not use record version number > TLS 1.0 in initial client
7913 hello: some (but not all) hanging servers will now work.
7914 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7915 the number of ciphers sent in the client hello. This should be
7916 set to an even number, such as 50, for example by passing:
7917 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7918 Most broken servers should now work.
7919 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7920 TLS 1.2 client support entirely.
7924 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7928 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7930 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7931 STRING form instead of a DigestInfo.
7935 * The format used for MDC2 RSA signatures is inconsistent between EVP
7936 and the RSA_sign/RSA_verify functions. This was made more apparent when
7937 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7938 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7939 the correct format in RSA_verify so both forms transparently work.
7943 * Some servers which support TLS 1.0 can choke if we initially indicate
7944 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7945 encrypted premaster secret. As a workaround use the maximum permitted
7946 client version in client hello, this should keep such servers happy
7947 and still work with previous versions of OpenSSL.
7951 * Add support for TLS/DTLS heartbeats.
7953 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7955 * Add support for SCTP.
7957 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7959 * Improved PRNG seeding for VOS.
7961 *Paul Green <Paul.Green@stratus.com>*
7963 * Extensive assembler packs updates, most notably:
7965 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7966 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7967 - x86_64: bit-sliced AES implementation;
7968 - ARM: NEON support, contemporary platforms optimizations;
7969 - s390x: z196 support;
7970 - `*`: GHASH and GF(2^m) multiplication implementations;
7974 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7975 (removal of unnecessary code)
7977 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7979 * Add TLS key material exporter from RFC 5705.
7983 * Add DTLS-SRTP negotiation from RFC 5764.
7987 * Add Next Protocol Negotiation,
7988 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7989 disabled with a no-npn flag to config or Configure. Code donated
7992 *Adam Langley <agl@google.com> and Ben Laurie*
7994 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7995 NIST-P256, NIST-P521, with constant-time single point multiplication on
7996 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7997 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7998 Code made available under Apache License version 2.0.
8000 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8001 line to include this in your build of OpenSSL, and run "make depend" (or
8002 "make update"). This enables the following EC_METHODs:
8004 EC_GFp_nistp224_method()
8005 EC_GFp_nistp256_method()
8006 EC_GFp_nistp521_method()
8008 EC_GROUP_new_by_curve_name() will automatically use these (while
8009 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8012 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8014 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8015 all platforms. Move ssize_t definition from e_os.h to the public
8016 header file e_os2.h as it now appears in public header file cms.h
8020 * New -sigopt option to the ca, req and x509 utilities. Additional
8021 signature parameters can be passed using this option and in
8026 * Add RSA PSS signing function. This will generate and set the
8027 appropriate AlgorithmIdentifiers for PSS based on those in the
8028 corresponding EVP_MD_CTX structure. No application support yet.
8032 * Support for companion algorithm specific ASN1 signing routines.
8033 New function ASN1_item_sign_ctx() signs a pre-initialised
8034 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8035 the appropriate parameters.
8039 * Add new algorithm specific ASN1 verification initialisation function
8040 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8041 handling will be the same no matter what EVP_PKEY_METHOD is used.
8042 Add a PSS handler to support verification of PSS signatures: checked
8043 against a number of sample certificates.
8047 * Add signature printing for PSS. Add PSS OIDs.
8049 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8051 * Add algorithm specific signature printing. An individual ASN1 method
8052 can now print out signatures instead of the standard hex dump.
8054 More complex signatures (e.g. PSS) can print out more meaningful
8055 information. Include DSA version that prints out the signature
8060 * Password based recipient info support for CMS library: implementing
8065 * Split password based encryption into PBES2 and PBKDF2 functions. This
8066 neatly separates the code into cipher and PBE sections and is required
8067 for some algorithms that split PBES2 into separate pieces (such as
8068 password based CMS).
8072 * Session-handling fixes:
8073 - Fix handling of connections that are resuming with a session ID,
8074 but also support Session Tickets.
8075 - Fix a bug that suppressed issuing of a new ticket if the client
8076 presented a ticket with an expired session.
8077 - Try to set the ticket lifetime hint to something reasonable.
8078 - Make tickets shorter by excluding irrelevant information.
8079 - On the client side, don't ignore renewed tickets.
8081 *Adam Langley, Bodo Moeller (Google)*
8083 * Fix PSK session representation.
8087 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8089 This work was sponsored by Intel.
8093 * Add GCM support to TLS library. Some custom code is needed to split
8094 the IV between the fixed (from PRF) and explicit (from TLS record)
8095 portions. This adds all GCM ciphersuites supported by RFC5288 and
8096 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8097 add a special AESGCM string for GCM only.
8101 * Expand range of ctrls for AES GCM. Permit setting invocation
8102 field on decrypt and retrieval of invocation field only on encrypt.
8106 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8107 As required by RFC5289 these ciphersuites cannot be used if for
8108 versions of TLS earlier than 1.2.
8112 * For FIPS capable OpenSSL interpret a NULL default public key method
8113 as unset and return the appropriate default but do *not* set the default.
8114 This means we can return the appropriate method in applications that
8115 switch between FIPS and non-FIPS modes.
8119 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8120 ENGINE is used then we cannot handle that in the FIPS module so we
8121 keep original code iff non-FIPS operations are allowed.
8125 * Add -attime option to openssl utilities.
8127 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8129 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8133 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8134 FIPS EC methods unconditionally for now.
8138 * New build option no-ec2m to disable characteristic 2 code.
8142 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8143 all cases can be covered as some introduce binary incompatibilities.
8147 * Redirect RSA operations to FIPS module including keygen,
8148 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8152 * Add similar low-level API blocking to ciphers.
8156 * low-level digest APIs are not approved in FIPS mode: any attempt
8157 to use these will cause a fatal error. Applications that *really* want
8158 to use them can use the `private_*` version instead.
8162 * Redirect cipher operations to FIPS module for FIPS builds.
8166 * Redirect digest operations to FIPS module for FIPS builds.
8170 * Update build system to add "fips" flag which will link in fipscanister.o
8171 for static and shared library builds embedding a signature if needed.
8175 * Output TLS supported curves in preference order instead of numerical
8176 order. This is currently hardcoded for the highest order curves first.
8177 This should be configurable so applications can judge speed vs strength.
8181 * Add TLS v1.2 server support for client authentication.
8185 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8190 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8191 FIPS modules versions.
8195 * Add TLS v1.2 client side support for client authentication. Keep cache
8196 of handshake records longer as we don't know the hash algorithm to use
8197 until after the certificate request message is received.
8201 * Initial TLS v1.2 client support. Add a default signature algorithms
8202 extension including all the algorithms we support. Parse new signature
8203 format in client key exchange. Relax some ECC signing restrictions for
8204 TLS v1.2 as indicated in RFC5246.
8208 * Add server support for TLS v1.2 signature algorithms extension. Switch
8209 to new signature format when needed using client digest preference.
8210 All server ciphersuites should now work correctly in TLS v1.2. No client
8211 support yet and no support for client certificates.
8215 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8216 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8217 ciphersuites. At present only RSA key exchange ciphersuites work with
8218 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8219 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8220 and version checking.
8224 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8225 with this defined it will not be affected by any changes to ssl internal
8226 structures. Add several utility functions to allow openssl application
8227 to work with OPENSSL_NO_SSL_INTERN defined.
8231 * A long standing patch to add support for SRP from EdelWeb (Peter
8232 Sylvester and Christophe Renou) was integrated.
8233 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8234 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8237 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8241 * Permit abbreviated handshakes when renegotiating using the function
8242 SSL_renegotiate_abbreviated().
8244 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8246 * Add call to ENGINE_register_all_complete() to
8247 ENGINE_load_builtin_engines(), so some implementations get used
8248 automatically instead of needing explicit application support.
8252 * Add support for TLS key exporter as described in RFC5705.
8254 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8256 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8257 a few changes are required:
8259 Add SSL_OP_NO_TLSv1_1 flag.
8260 Add TLSv1_1 methods.
8261 Update version checking logic to handle version 1.1.
8262 Add explicit IV handling (ported from DTLS code).
8263 Add command line options to s_client/s_server.
8270 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8272 * X509_ATTRIBUTE memory leak
8274 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8275 memory. This structure is used by the PKCS#7 and CMS routines so any
8276 application which reads PKCS#7 or CMS data from untrusted sources is
8277 affected. SSL/TLS is not affected.
8279 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8285 * Race condition handling PSK identify hint
8287 If PSK identity hints are received by a multi-threaded client then
8288 the values are wrongly updated in the parent SSL_CTX structure. This can
8289 result in a race condition potentially leading to a double free of the
8295 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8297 * Malformed ECParameters causes infinite loop
8299 When processing an ECParameters structure OpenSSL enters an infinite loop
8300 if the curve specified is over a specially malformed binary polynomial
8303 This can be used to perform denial of service against any
8304 system which processes public keys, certificate requests or
8305 certificates. This includes TLS clients and TLS servers with
8306 client authentication enabled.
8308 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8313 * Exploitable out-of-bounds read in X509_cmp_time
8315 X509_cmp_time does not properly check the length of the ASN1_TIME
8316 string and can read a few bytes out of bounds. In addition,
8317 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8320 An attacker can use this to craft malformed certificates and CRLs of
8321 various sizes and potentially cause a segmentation fault, resulting in
8322 a DoS on applications that verify certificates or CRLs. TLS clients
8323 that verify CRLs are affected. TLS clients and servers with client
8324 authentication enabled may be affected if they use custom verification
8327 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8328 independently by Hanno Böck.
8333 * PKCS7 crash with missing EnvelopedContent
8335 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8336 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8337 with missing content and trigger a NULL pointer dereference on parsing.
8339 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8340 structures from untrusted sources are affected. OpenSSL clients and
8341 servers are not affected.
8343 This issue was reported to OpenSSL by Michal Zalewski (Google).
8348 * CMS verify infinite loop with unknown hash function
8350 When verifying a signedData message the CMS code can enter an infinite loop
8351 if presented with an unknown hash function OID. This can be used to perform
8352 denial of service against any system which verifies signedData messages using
8354 This issue was reported to OpenSSL by Johannes Bauer.
8359 * Race condition handling NewSessionTicket
8361 If a NewSessionTicket is received by a multi-threaded client when attempting to
8362 reuse a previous ticket then a race condition can occur potentially leading to
8363 a double free of the ticket data.
8368 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8370 * Segmentation fault in ASN1_TYPE_cmp fix
8372 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8373 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8374 certificate signature algorithm consistency this can be used to crash any
8375 certificate verification operation and exploited in a DoS attack. Any
8376 application which performs certificate verification is vulnerable including
8377 OpenSSL clients and servers which enable client authentication.
8382 * ASN.1 structure reuse memory corruption fix
8384 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8385 memory corruption via an invalid write. Such reuse is and has been
8386 strongly discouraged and is believed to be rare.
8388 Applications that parse structures containing CHOICE or ANY DEFINED BY
8389 components may be affected. Certificate parsing (d2i_X509 and related
8390 functions) are however not affected. OpenSSL clients and servers are
8396 * PKCS7 NULL pointer dereferences fix
8398 The PKCS#7 parsing code does not handle missing outer ContentInfo
8399 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8400 missing content and trigger a NULL pointer dereference on parsing.
8402 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8403 otherwise parse PKCS#7 structures from untrusted sources are
8404 affected. OpenSSL clients and servers are not affected.
8406 This issue was reported to OpenSSL by Michal Zalewski (Google).
8411 * DoS via reachable assert in SSLv2 servers fix
8413 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8414 servers that both support SSLv2 and enable export cipher suites by sending
8415 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8417 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8418 (OpenSSL development team).
8423 * Use After Free following d2i_ECPrivatekey error fix
8425 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8426 could cause a use after free condition. This, in turn, could cause a double
8427 free in several private key parsing functions (such as d2i_PrivateKey
8428 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8429 for applications that receive EC private keys from untrusted
8430 sources. This scenario is considered rare.
8432 This issue was discovered by the BoringSSL project and fixed in their
8438 * X509_to_X509_REQ NULL pointer deref fix
8440 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8441 the certificate key is invalid. This function is rarely used in practice.
8443 This issue was discovered by Brian Carpenter.
8448 * Removed the export ciphers from the DEFAULT ciphers
8452 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8454 * Build fixes for the Windows and OpenVMS platforms
8456 *Matt Caswell and Richard Levitte*
8458 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8460 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8461 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8462 dereference. This could lead to a Denial Of Service attack. Thanks to
8463 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8468 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8469 dtls1_buffer_record function under certain conditions. In particular this
8470 could occur if an attacker sent repeated DTLS records with the same
8471 sequence number but for the next epoch. The memory leak could be exploited
8472 by an attacker in a Denial of Service attack through memory exhaustion.
8473 Thanks to Chris Mueller for reporting this issue.
8478 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8479 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8480 method would be set to NULL which could later result in a NULL pointer
8481 dereference. Thanks to Frank Schmirler for reporting this issue.
8486 * Abort handshake if server key exchange message is omitted for ephemeral
8489 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8490 reporting this issue.
8495 * Remove non-export ephemeral RSA code on client and server. This code
8496 violated the TLS standard by allowing the use of temporary RSA keys in
8497 non-export ciphersuites and could be used by a server to effectively
8498 downgrade the RSA key length used to a value smaller than the server
8499 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8500 INRIA or reporting this issue.
8505 * Fixed issue where DH client certificates are accepted without verification.
8506 An OpenSSL server will accept a DH certificate for client authentication
8507 without the certificate verify message. This effectively allows a client to
8508 authenticate without the use of a private key. This only affects servers
8509 which trust a client certificate authority which issues certificates
8510 containing DH keys: these are extremely rare and hardly ever encountered.
8511 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8517 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8518 results on some platforms, including x86_64. This bug occurs at random
8519 with a very low probability, and is not known to be exploitable in any
8520 way, though its exact impact is difficult to determine. Thanks to Pieter
8521 Wuille (Blockstream) who reported this issue and also suggested an initial
8522 fix. Further analysis was conducted by the OpenSSL development team and
8523 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8524 the OpenSSL core team.
8529 * Fix various certificate fingerprint issues.
8531 By using non-DER or invalid encodings outside the signed portion of a
8532 certificate the fingerprint can be changed without breaking the signature.
8533 Although no details of the signed portion of the certificate can be changed
8534 this can cause problems with some applications: e.g. those using the
8535 certificate fingerprint for blacklists.
8537 1. Reject signatures with non zero unused bits.
8539 If the BIT STRING containing the signature has non zero unused bits reject
8540 the signature. All current signature algorithms require zero unused bits.
8542 2. Check certificate algorithm consistency.
8544 Check the AlgorithmIdentifier inside TBS matches the one in the
8545 certificate signature. NB: this will result in signature failure
8546 errors for some broken certificates.
8548 Thanks to Konrad Kraszewski from Google for reporting this issue.
8550 3. Check DSA/ECDSA signatures use DER.
8552 Reencode DSA/ECDSA signatures and compare with the original received
8553 signature. Return an error if there is a mismatch.
8555 This will reject various cases including garbage after signature
8556 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8557 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8558 (negative or with leading zeroes).
8560 Further analysis was conducted and fixes were developed by Stephen Henson
8561 of the OpenSSL core team.
8567 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8569 * Session Ticket Memory Leak.
8571 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8572 integrity of that ticket is first verified. In the event of a session
8573 ticket integrity check failing, OpenSSL will fail to free memory
8574 causing a memory leak. By sending a large number of invalid session
8575 tickets an attacker could exploit this issue in a Denial Of Service
8581 * Build option no-ssl3 is incomplete.
8583 When OpenSSL is configured with "no-ssl3" as a build option, servers
8584 could accept and complete an SSL 3.0 handshake, and clients could be
8585 configured to send them.
8588 *Akamai and the OpenSSL team*
8590 * Add support for TLS_FALLBACK_SCSV.
8591 Client applications doing fallback retries should call
8592 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8595 *Adam Langley, Bodo Moeller*
8597 * Add additional DigestInfo checks.
8599 Reencode DigestInto in DER and check against the original when
8600 verifying RSA signature: this will reject any improperly encoded
8601 DigestInfo structures.
8603 Note: this is a precautionary measure and no attacks are currently known.
8607 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8609 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8610 to a denial of service attack. A malicious server can crash the client
8611 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8612 ciphersuite and sending carefully crafted handshake messages.
8614 Thanks to Felix Gröbert (Google) for discovering and researching this
8620 * By sending carefully crafted DTLS packets an attacker could cause openssl
8621 to leak memory. This can be exploited through a Denial of Service attack.
8622 Thanks to Adam Langley for discovering and researching this issue.
8627 * An attacker can force openssl to consume large amounts of memory whilst
8628 processing DTLS handshake messages. This can be exploited through a
8629 Denial of Service attack.
8630 Thanks to Adam Langley for discovering and researching this issue.
8635 * An attacker can force an error condition which causes openssl to crash
8636 whilst processing DTLS packets due to memory being freed twice. This
8637 can be exploited through a Denial of Service attack.
8638 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8644 * If a multithreaded client connects to a malicious server using a resumed
8645 session and the server sends an ec point format extension it could write
8646 up to 255 bytes to freed memory.
8648 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8654 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8655 X509_name_oneline, X509_name_print_ex et al. to leak some information
8656 from the stack. Applications may be affected if they echo pretty printing
8657 output to the attacker.
8659 Thanks to Ivan Fratric (Google) for discovering this issue.
8662 *Emilia Käsper, and Steve Henson*
8664 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8665 for corner cases. (Certain input points at infinity could lead to
8666 bogus results, with non-infinity inputs mapped to infinity too.)
8670 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8672 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8673 handshake can force the use of weak keying material in OpenSSL
8674 SSL/TLS clients and servers.
8676 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8677 researching this issue. ([CVE-2014-0224])
8679 *KIKUCHI Masashi, Steve Henson*
8681 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8682 OpenSSL DTLS client the code can be made to recurse eventually crashing
8685 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8688 *Imre Rad, Steve Henson*
8690 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8691 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8692 client or server. This is potentially exploitable to run arbitrary
8693 code on a vulnerable client or server.
8695 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8697 *Jüri Aedla, Steve Henson*
8699 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8700 are subject to a denial of service attack.
8702 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8703 this issue. ([CVE-2014-3470])
8705 *Felix Gröbert, Ivan Fratric, Steve Henson*
8707 * Harmonize version and its documentation. -f flag is used to display
8710 *mancha <mancha1@zoho.com>*
8712 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8713 in i2d_ECPrivateKey.
8715 *mancha <mancha1@zoho.com>*
8717 * Fix some double frees. These are not thought to be exploitable.
8719 *mancha <mancha1@zoho.com>*
8721 * Fix for the attack described in the paper "Recovering OpenSSL
8722 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8723 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8724 <http://eprint.iacr.org/2014/140>
8726 Thanks to Yuval Yarom and Naomi Benger for discovering this
8727 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8729 *Yuval Yarom and Naomi Benger*
8731 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8733 * Keep original DTLS digest and encryption contexts in retransmission
8734 structures so we can use the previous session parameters if they need
8735 to be resent. ([CVE-2013-6450])
8739 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8740 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8741 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8742 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8743 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8744 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8746 *Rob Stradling, Adam Langley*
8748 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8750 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8752 This addresses the flaw in CBC record processing discovered by
8753 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8754 at: <http://www.isg.rhul.ac.uk/tls/>
8756 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8757 Security Group at Royal Holloway, University of London
8758 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8759 Emilia Käsper for the initial patch.
8762 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8764 * Return an error when checking OCSP signatures when key is NULL.
8765 This fixes a DoS attack. ([CVE-2013-0166])
8769 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8770 the right response is stapled. Also change SSL_get_certificate()
8771 so it returns the certificate actually sent.
8772 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8773 (This is a backport)
8775 *Rob Stradling <rob.stradling@comodo.com>*
8777 * Fix possible deadlock when decoding public keys.
8781 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8783 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8786 * Sanity check record length before skipping explicit IV in DTLS
8789 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8790 fuzzing as a service testing platform.
8795 * Initialise tkeylen properly when encrypting CMS messages.
8796 Thanks to Solar Designer of Openwall for reporting this issue.
8800 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8802 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8803 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8804 in CRYPTO_realloc_clean.
8806 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8807 issue and to Adam Langley <agl@chromium.org> for fixing it.
8810 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8812 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8814 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8815 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8816 content decryption and always return the same error. Note: this attack
8817 needs on average 2^20 messages so it only affects automated senders. The
8818 old behaviour can be re-enabled in the CMS code by setting the
8819 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8820 an MMA defence is not necessary.
8821 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8822 this issue. ([CVE-2012-0884])
8826 * Fix CVE-2011-4619: make sure we really are receiving a
8827 client hello before rejecting multiple SGC restarts. Thanks to
8828 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8832 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8834 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8835 Thanks to Antonio Martin, Enterprise Secure Access Research and
8836 Development, Cisco Systems, Inc. for discovering this bug and
8837 preparing a fix. ([CVE-2012-0050])
8841 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8843 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8844 of the Vaudenay padding oracle attack on CBC mode encryption
8845 which enables an efficient plaintext recovery attack against
8846 the OpenSSL implementation of DTLS. Their attack exploits timing
8847 differences arising during decryption processing. A research
8848 paper describing this attack can be found at:
8849 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8850 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8851 Security Group at Royal Holloway, University of London
8852 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8853 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8854 for preparing the fix. ([CVE-2011-4108])
8856 *Robin Seggelmann, Michael Tuexen*
8858 * Clear bytes used for block padding of SSL 3.0 records.
8861 *Adam Langley (Google)*
8863 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8864 Kadianakis <desnacked@gmail.com> for discovering this issue and
8865 Adam Langley for preparing the fix. ([CVE-2011-4619])
8867 *Adam Langley (Google)*
8869 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8871 *Andrey Kulikov <amdeich@gmail.com>*
8873 * Prevent malformed RFC3779 data triggering an assertion failure.
8874 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8875 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8877 *Rob Austein <sra@hactrn.net>*
8879 * Improved PRNG seeding for VOS.
8881 *Paul Green <Paul.Green@stratus.com>*
8883 * Fix ssl_ciph.c set-up race.
8885 *Adam Langley (Google)*
8887 * Fix spurious failures in ecdsatest.c.
8889 *Emilia Käsper (Google)*
8891 * Fix the BIO_f_buffer() implementation (which was mixing different
8892 interpretations of the `..._len` fields).
8894 *Adam Langley (Google)*
8896 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8897 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8898 threads won't reuse the same blinding coefficients.
8900 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8901 lock to call BN_BLINDING_invert_ex, and avoids one use of
8902 BN_BLINDING_update for each BN_BLINDING structure (previously,
8903 the last update always remained unused).
8905 *Emilia Käsper (Google)*
8907 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8909 *Bob Buckholz (Google)*
8911 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8913 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8914 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8916 *Kaspar Brand <ossl@velox.ch>*
8918 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8919 for multi-threaded use of ECDH. ([CVE-2011-3210])
8921 *Adam Langley (Google)*
8923 * Fix x509_name_ex_d2i memory leak on bad inputs.
8927 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8928 signature public key algorithm by using OID xref utilities instead.
8929 Before this you could only use some ECC ciphersuites with SHA1 only.
8933 * Add protection against ECDSA timing attacks as mentioned in the paper
8934 by Billy Bob Brumley and Nicola Tuveri, see:
8935 <http://eprint.iacr.org/2011/232.pdf>
8937 *Billy Bob Brumley and Nicola Tuveri*
8939 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8941 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8943 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8945 * Fix bug in string printing code: if *any* escaping is enabled we must
8946 escape the escape character (backslash) or the resulting string is
8951 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8953 * Disable code workaround for ancient and obsolete Netscape browsers
8954 and servers: an attacker can use it in a ciphersuite downgrade attack.
8955 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8959 * Fixed J-PAKE implementation error, originally discovered by
8960 Sebastien Martini, further info and confirmation from Stefan
8961 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8965 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8967 * Fix extension code to avoid race conditions which can result in a buffer
8968 overrun vulnerability: resumed sessions must not be modified as they can
8969 be shared by multiple threads. CVE-2010-3864
8973 * Fix WIN32 build system to correctly link an ENGINE directory into
8978 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8980 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8983 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8985 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8987 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8988 context. The operation can be customised via the ctrl mechanism in
8989 case ENGINEs want to include additional functionality.
8993 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8997 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8998 output hashes compatible with older versions of OpenSSL.
9000 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9002 * Fix compression algorithm handling: if resuming a session use the
9003 compression algorithm of the resumed session instead of determining
9004 it from client hello again. Don't allow server to change algorithm.
9008 * Add load_crls() function to commands tidying load_certs() too. Add option
9009 to verify utility to allow additional CRLs to be included.
9013 * Update OCSP request code to permit adding custom headers to the request:
9014 some responders need this.
9018 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9021 *Julia Lawall <julia@diku.dk>*
9023 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9024 needlessly dereferenced structures, used obsolete functions and
9025 didn't handle all updated verify codes correctly.
9029 * Disable MD2 in the default configuration.
9033 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9034 indicate the initial BIO being pushed or popped. This makes it possible
9035 to determine whether the BIO is the one explicitly called or as a result
9036 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9037 it handles reference counts correctly and doesn't zero out the I/O bio
9038 when it is not being explicitly popped. WARNING: applications which
9039 included workarounds for the old buggy behaviour will need to be modified
9040 or they could free up already freed BIOs.
9044 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9045 renaming to all platforms (within the 0.9.8 branch, this was
9046 done conditionally on Netware platforms to avoid a name clash).
9048 *Guenter <lists@gknw.net>*
9050 * Add ECDHE and PSK support to DTLS.
9052 *Michael Tuexen <tuexen@fh-muenster.de>*
9054 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9059 * Add "missing" function EVP_MD_flags() (without this the only way to
9060 retrieve a digest flags is by accessing the structure directly. Update
9061 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9062 or cipher is registered as in the "from" argument. Print out all
9063 registered digests in the dgst usage message instead of manually
9064 attempting to work them out.
9068 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9069 this allows the use of compression and extensions. Change default cipher
9070 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9071 by default unless an application cipher string requests it.
9075 * Alter match criteria in PKCS12_parse(). It used to try to use local
9076 key ids to find matching certificates and keys but some PKCS#12 files
9077 don't follow the (somewhat unwritten) rules and this strategy fails.
9078 Now just gather all certificates together and the first private key
9079 then look for the first certificate that matches the key.
9083 * Support use of registered digest and cipher names for dgst and cipher
9084 commands instead of having to add each one as a special case. So now
9091 openssl dgst -sha256 foo
9093 and this works for ENGINE based algorithms too.
9097 * Update Gost ENGINE to support parameter files.
9099 *Victor B. Wagner <vitus@cryptocom.ru>*
9101 * Support GeneralizedTime in ca utility.
9103 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9105 * Enhance the hash format used for certificate directory links. The new
9106 form uses the canonical encoding (meaning equivalent names will work
9107 even if they aren't identical) and uses SHA1 instead of MD5. This form
9108 is incompatible with the older format and as a result c_rehash should
9109 be used to rebuild symbolic links.
9113 * Make PKCS#8 the default write format for private keys, replacing the
9114 traditional format. This form is standardised, more secure and doesn't
9115 include an implicit MD5 dependency.
9119 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9120 committed to OpenSSL should pass this lot as a minimum.
9124 * Add session ticket override functionality for use by EAP-FAST.
9126 *Jouni Malinen <j@w1.fi>*
9128 * Modify HMAC functions to return a value. Since these can be implemented
9129 in an ENGINE errors can occur.
9133 * Type-checked OBJ_bsearch_ex.
9137 * Type-checked OBJ_bsearch. Also some constification necessitated
9138 by type-checking. Still to come: TXT_DB, bsearch(?),
9139 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9144 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9145 seconds to a tm structure directly, instead of going through OS
9146 specific date routines. This avoids any issues with OS routines such
9147 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9148 and X509_time_adj_ex() to cover the extended range. The existing
9149 X509_time_adj() is still usable and will no longer have any date issues.
9153 * Delta CRL support. New use deltas option which will attempt to locate
9154 and search any appropriate delta CRLs available.
9156 This work was sponsored by Google.
9160 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9161 code and add additional score elements. Validate alternate CRL paths
9162 as part of the CRL checking and indicate a new error "CRL path validation
9163 error" in this case. Applications wanting additional details can use
9164 the verify callback and check the new "parent" field. If this is not
9165 NULL CRL path validation is taking place. Existing applications won't
9166 see this because it requires extended CRL support which is off by
9169 This work was sponsored by Google.
9173 * Support for freshest CRL extension.
9175 This work was sponsored by Google.
9179 * Initial indirect CRL support. Currently only supported in the CRLs
9180 passed directly and not via lookup. Process certificate issuer
9181 CRL entry extension and lookup CRL entries by bother issuer name
9182 and serial number. Check and process CRL issuer entry in IDP extension.
9184 This work was sponsored by Google.
9188 * Add support for distinct certificate and CRL paths. The CRL issuer
9189 certificate is validated separately in this case. Only enabled if
9190 an extended CRL support flag is set: this flag will enable additional
9191 CRL functionality in future.
9193 This work was sponsored by Google.
9197 * Add support for policy mappings extension.
9199 This work was sponsored by Google.
9203 * Fixes to pathlength constraint, self issued certificate handling,
9204 policy processing to align with RFC3280 and PKITS tests.
9206 This work was sponsored by Google.
9210 * Support for name constraints certificate extension. DN, email, DNS
9211 and URI types are currently supported.
9213 This work was sponsored by Google.
9217 * To cater for systems that provide a pointer-based thread ID rather
9218 than numeric, deprecate the current numeric thread ID mechanism and
9219 replace it with a structure and associated callback type. This
9220 mechanism allows a numeric "hash" to be extracted from a thread ID in
9221 either case, and on platforms where pointers are larger than 'long',
9222 mixing is done to help ensure the numeric 'hash' is usable even if it
9223 can't be guaranteed unique. The default mechanism is to use "&errno"
9224 as a pointer-based thread ID to distinguish between threads.
9226 Applications that want to provide their own thread IDs should now use
9227 CRYPTO_THREADID_set_callback() to register a callback that will call
9228 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9230 Note that ERR_remove_state() is now deprecated, because it is tied
9231 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9232 to free the current thread's error state should be replaced by
9233 ERR_remove_thread_state(NULL).
9235 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9236 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9237 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9238 application was previously providing a numeric thread callback that
9239 was inappropriate for distinguishing threads, then uniqueness might
9240 have been obtained with &errno that happened immediately in the
9241 intermediate development versions of OpenSSL; this is no longer the
9242 case, the numeric thread callback will now override the automatic use
9245 *Geoff Thorpe, with help from Bodo Moeller*
9247 * Initial support for different CRL issuing certificates. This covers a
9248 simple case where the self issued certificates in the chain exist and
9249 the real CRL issuer is higher in the existing chain.
9251 This work was sponsored by Google.
9255 * Removed effectively defunct crypto/store from the build.
9259 * Revamp of STACK to provide stronger type-checking. Still to come:
9260 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9261 ASN1_STRING, CONF_VALUE.
9265 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9266 RAM on SSL connections. This option can save about 34k per idle SSL.
9270 * Revamp of LHASH to provide stronger type-checking. Still to come:
9271 STACK, TXT_DB, bsearch, qsort.
9275 * Initial support for Cryptographic Message Syntax (aka CMS) based
9276 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9277 support for data, signedData, compressedData, digestedData and
9278 encryptedData, envelopedData types included. Scripts to check against
9279 RFC4134 examples draft and interop and consistency checks of many
9280 content types and variants.
9284 * Add options to enc utility to support use of zlib compression BIO.
9288 * Extend mk1mf to support importing of options and assembly language
9289 files from Configure script, currently only included in VC-WIN32.
9290 The assembly language rules can now optionally generate the source
9291 files from the associated perl scripts.
9295 * Implement remaining functionality needed to support GOST ciphersuites.
9296 Interop testing has been performed using CryptoPro implementations.
9298 *Victor B. Wagner <vitus@cryptocom.ru>*
9300 * s390x assembler pack.
9304 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9309 * Implement Opaque PRF Input TLS extension as specified in
9310 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9311 official specification yet and no extension type assignment by
9312 IANA exists, this extension (for now) will have to be explicitly
9313 enabled when building OpenSSL by providing the extension number
9314 to use. For example, specify an option
9316 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9318 to the "config" or "Configure" script to enable the extension,
9319 assuming extension number 0x9527 (which is a completely arbitrary
9320 and unofficial assignment based on the MD5 hash of the Internet
9321 Draft). Note that by doing so, you potentially lose
9322 interoperability with other TLS implementations since these might
9323 be using the same extension number for other purposes.
9325 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9326 opaque PRF input value to use in the handshake. This will create
9327 an internal copy of the length-'len' string at 'src', and will
9328 return non-zero for success.
9330 To get more control and flexibility, provide a callback function
9333 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9334 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9338 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9341 Callback function 'cb' will be called in handshakes, and is
9342 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9343 Argument 'arg' is for application purposes (the value as given to
9344 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9345 be provided to the callback function). The callback function
9346 has to return non-zero to report success: usually 1 to use opaque
9347 PRF input just if possible, or 2 to enforce use of the opaque PRF
9348 input. In the latter case, the library will abort the handshake
9349 if opaque PRF input is not successfully negotiated.
9351 Arguments 'peerinput' and 'len' given to the callback function
9352 will always be NULL and 0 in the case of a client. A server will
9353 see the client's opaque PRF input through these variables if
9354 available (NULL and 0 otherwise). Note that if the server
9355 provides an opaque PRF input, the length must be the same as the
9356 length of the client's opaque PRF input.
9358 Note that the callback function will only be called when creating
9359 a new session (session resumption can resume whatever was
9360 previously negotiated), and will not be called in SSL 2.0
9361 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9362 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9363 for applications that need to enforce opaque PRF input.
9367 * Update ssl code to support digests other than SHA1+MD5 for handshake
9370 *Victor B. Wagner <vitus@cryptocom.ru>*
9372 * Add RFC4507 support to OpenSSL. This includes the corrections in
9373 RFC4507bis. The encrypted ticket format is an encrypted encoded
9374 SSL_SESSION structure, that way new session features are automatically
9377 If a client application caches session in an SSL_SESSION structure
9378 support is transparent because tickets are now stored in the encoded
9381 The SSL_CTX structure automatically generates keys for ticket
9382 protection in servers so again support should be possible
9383 with no application modification.
9385 If a client or server wishes to disable RFC4507 support then the option
9386 SSL_OP_NO_TICKET can be set.
9388 Add a TLS extension debugging callback to allow the contents of any client
9389 or server extensions to be examined.
9391 This work was sponsored by Google.
9395 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9396 OpenSSL should now compile cleanly on gcc 4.2
9398 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9400 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9401 support including streaming MAC support: this is required for GOST
9402 ciphersuite support.
9404 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9406 * Add option -stream to use PKCS#7 streaming in smime utility. New
9407 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9408 to output in BER and PEM format.
9412 * Experimental support for use of HMAC via EVP_PKEY interface. This
9413 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9414 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9415 ENGINE support for HMAC keys which are unextractable. New -mac and
9416 -macopt options to dgst utility.
9420 * New option -sigopt to dgst utility. Update dgst to use
9421 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9422 alternative signing parameters such as X9.31 or PSS in the dgst
9427 * Change ssl_cipher_apply_rule(), the internal function that does
9428 the work each time a ciphersuite string requests enabling
9429 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9430 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9431 the order of disabled ciphersuites such that those ciphersuites
9432 that most recently went from enabled to disabled not only stay
9433 in order with respect to each other, but also have higher priority
9434 than other disabled ciphersuites the next time ciphersuites are
9437 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9438 the same ciphersuites as with "HIGH" alone, but in a specific
9439 order where the PSK ciphersuites come first (since they are the
9440 most recently disabled ciphersuites when "HIGH" is parsed).
9442 Also, change ssl_create_cipher_list() (using this new
9443 functionality) such that between otherwise identical
9444 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9449 * Change ssl_create_cipher_list() so that it automatically
9450 arranges the ciphersuites in reasonable order before starting
9451 to process the rule string. Thus, the definition for "DEFAULT"
9452 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9453 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9454 This makes it much easier to arrive at a reasonable default order
9455 in applications for which anonymous ciphers are OK (meaning
9456 that you can't actually use DEFAULT).
9458 *Bodo Moeller; suggested by Victor Duchovni*
9460 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9461 processing) into multiple integers instead of setting
9462 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9463 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9464 (These masks as well as the individual bit definitions are hidden
9465 away into the non-exported interface ssl/ssl_locl.h, so this
9466 change to the definition of the SSL_CIPHER structure shouldn't
9467 affect applications.) This give us more bits for each of these
9468 categories, so there is no longer a need to coagulate AES128 and
9469 AES256 into a single algorithm bit, and to coagulate Camellia128
9470 and Camellia256 into a single algorithm bit, which has led to all
9473 Thus, among other things, the kludge introduced in 0.9.7m and
9474 0.9.8e for masking out AES256 independently of AES128 or masking
9475 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9477 With the change, we also introduce new ciphersuite aliases that
9478 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9483 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9484 Use the leftmost N bytes of the signature input if the input is
9485 larger than the prime q (with N being the size in bytes of q).
9489 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9490 it yet and it is largely untested.
9494 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9498 * Initial incomplete changes to avoid need for function casts in OpenSSL
9499 some compilers (gcc 4.2 and later) reject their use. Safestack is
9500 reimplemented. Update ASN1 to avoid use of legacy functions.
9504 * Win32/64 targets are linked with Winsock2.
9508 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9509 to external functions. This can be used to increase CRL handling
9510 efficiency especially when CRLs are very large by (for example) storing
9511 the CRL revoked certificates in a database.
9515 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9516 new CRLs added to a directory can be used. New command line option
9517 -verify_return_error to s_client and s_server. This causes real errors
9518 to be returned by the verify callback instead of carrying on no matter
9519 what. This reflects the way a "real world" verify callback would behave.
9523 * GOST engine, supporting several GOST algorithms and public key formats.
9524 Kindly donated by Cryptocom.
9528 * Partial support for Issuing Distribution Point CRL extension. CRLs
9529 partitioned by DP are handled but no indirect CRL or reason partitioning
9530 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9531 selected via a scoring technique which handles IDP and AKID in CRLs.
9535 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9536 will ultimately be used for all verify operations: this will remove the
9537 X509_STORE dependency on certificate verification and allow alternative
9538 lookup methods. X509_STORE based implementations of these two callbacks.
9542 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9543 Modify get_crl() to find a valid (unexpired) CRL if possible.
9547 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9548 this would be called X509_CRL_cmp() but that name is already used by
9549 a function that just compares CRL issuer names. Cache several CRL
9550 extensions in X509_CRL structure and cache CRLDP in X509.
9554 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9555 this maps equivalent X509_NAME structures into a consistent structure.
9556 Name comparison can then be performed rapidly using memcmp().
9560 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9565 * Allow digests to supply their own micalg string for S/MIME type using
9566 the ctrl EVP_MD_CTRL_MICALG.
9570 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9571 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9572 ctrl. It can then customise the structure before and/or after signing
9577 * New function OBJ_add_sigid() to allow application defined signature OIDs
9578 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9579 to free up any added signature OIDs.
9583 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9584 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9585 digest and cipher tables. New options added to openssl utility:
9586 list-message-digest-algorithms and list-cipher-algorithms.
9590 * Change the array representation of binary polynomials: the list
9591 of degrees of non-zero coefficients is now terminated with -1.
9592 Previously it was terminated with 0, which was also part of the
9593 value; thus, the array representation was not applicable to
9594 polynomials where t^0 has coefficient zero. This change makes
9595 the array representation useful in a more general context.
9599 * Various modifications and fixes to SSL/TLS cipher string
9600 handling. For ECC, the code now distinguishes between fixed ECDH
9601 with RSA certificates on the one hand and with ECDSA certificates
9602 on the other hand, since these are separate ciphersuites. The
9603 unused code for Fortezza ciphersuites has been removed.
9605 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9606 (not "ECDHE"). For consistency with the code for DH
9607 certificates, use of ECDH certificates is now considered ECDH
9608 authentication, not RSA or ECDSA authentication (the latter is
9609 merely the CA's signing algorithm and not actively used in the
9612 The temporary ciphersuite alias "ECCdraft" is no longer
9613 available, and ECC ciphersuites are no longer excluded from "ALL"
9614 and "DEFAULT". The following aliases now exist for RFC 4492
9615 ciphersuites, most of these by analogy with the DH case:
9617 kECDHr - ECDH cert, signed with RSA
9618 kECDHe - ECDH cert, signed with ECDSA
9619 kECDH - ECDH cert (signed with either RSA or ECDSA)
9620 kEECDH - ephemeral ECDH
9621 ECDH - ECDH cert or ephemeral ECDH
9627 AECDH - anonymous ECDH
9628 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9632 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9633 Use correct micalg parameters depending on digest(s) in signed message.
9637 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9638 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9642 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9643 an engine to register a method. Add ENGINE lookups for methods and
9644 functional reference processing.
9648 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9649 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9654 * New -resign option to smime utility. This adds one or more signers
9655 to an existing PKCS#7 signedData structure. Also -md option to use an
9656 alternative message digest algorithm for signing.
9660 * Tidy up PKCS#7 routines and add new functions to make it easier to
9661 create PKCS7 structures containing multiple signers. Update smime
9662 application to support multiple signers.
9666 * New -macalg option to pkcs12 utility to allow setting of an alternative
9671 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9672 Reorganize PBE internals to lookup from a static table using NIDs,
9673 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9674 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9675 PRF which will be automatically used with PBES2.
9679 * Replace the algorithm specific calls to generate keys in "req" with the
9684 * Update PKCS#7 enveloped data routines to use new API. This is now
9685 supported by any public key method supporting the encrypt operation. A
9686 ctrl is added to allow the public key algorithm to examine or modify
9687 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9692 * Add a ctrl to asn1 method to allow a public key algorithm to express
9693 a default digest type to use. In most cases this will be SHA1 but some
9694 algorithms (such as GOST) need to specify an alternative digest. The
9695 return value indicates how strong the preference is 1 means optional and
9696 2 is mandatory (that is it is the only supported type). Modify
9697 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9698 use the default md. Update openssl utilities to use the default digest
9699 type for signing if it is not explicitly indicated.
9703 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9704 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9705 signing method from the key type. This effectively removes the link
9706 between digests and public key types.
9710 * Add an OID cross reference table and utility functions. Its purpose is to
9711 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9712 rsaEncryption. This will allow some of the algorithm specific hackery
9713 needed to use the correct OID to be removed.
9717 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9718 structures for PKCS7_sign(). They are now set up by the relevant public
9723 * Add provisional EC pkey method with support for ECDSA and ECDH.
9727 * Add support for key derivation (agreement) in the API, DH method and
9732 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9733 public and private key formats. As a side effect these add additional
9734 command line functionality not previously available: DSA signatures can be
9735 generated and verified using pkeyutl and DH key support and generation in
9742 *Oliver Tappe <zooey@hirschkaefer.de>*
9744 * New make target "install_html_docs" installs HTML renditions of the
9747 *Oliver Tappe <zooey@hirschkaefer.de>*
9749 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9750 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9751 support key and parameter generation and add initial key generation
9752 functionality for RSA.
9756 * Add functions for main EVP_PKEY_method operations. The undocumented
9757 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9758 `EVP_PKEY_{encrypt,decrypt}_old`.
9762 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9763 key API, doesn't do much yet.
9767 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9768 public key algorithms. New option to openssl utility:
9769 "list-public-key-algorithms" to print out info.
9773 * Implement the Supported Elliptic Curves Extension for
9774 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9778 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9779 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9783 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9784 utilities such as rsa, dsa, dsaparam etc except they process any key
9789 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9790 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9791 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9796 * Initial support for pluggable public key ASN1.
9797 De-spaghettify the public key ASN1 handling. Move public and private
9798 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9799 algorithm specific handling to a single module within the relevant
9800 algorithm directory. Add functions to allow (near) opaque processing
9801 of public and private key structures.
9805 * Implement the Supported Point Formats Extension for
9806 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9810 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9811 for the psk identity [hint] and the psk callback functions to the
9812 SSL_SESSION, SSL and SSL_CTX structure.
9815 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9819 SSL_CTX_use_psk_identity_hint
9820 SSL_get_psk_identity_hint
9821 SSL_get_psk_identity
9822 SSL_use_psk_identity_hint
9824 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9826 * Add RFC 3161 compliant time stamp request creation, response generation
9827 and response verification functionality.
9829 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9831 * Add initial support for TLS extensions, specifically for the server_name
9832 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9833 have new members for a hostname. The SSL data structure has an
9834 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9835 stored in that context to allow for session resumption, even after the
9836 SSL has been switched to a new SSL_CTX in reaction to a client's
9837 server_name extension.
9839 New functions (subject to change):
9841 SSL_get_servername()
9842 SSL_get_servername_type()
9845 New CTRL codes and macros (subject to change):
9847 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9848 - SSL_CTX_set_tlsext_servername_callback()
9849 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9850 - SSL_CTX_set_tlsext_servername_arg()
9851 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9853 openssl s_client has a new '-servername ...' option.
9855 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9856 '-key2 ...', '-servername_fatal' (subject to change). This allows
9857 testing the HostName extension for a specific single hostname ('-cert'
9858 and '-key' remain fallbacks for handshakes without HostName
9859 negotiation). If the unrecognized_name alert has to be sent, this by
9860 default is a warning; it becomes fatal with the '-servername_fatal'
9863 *Peter Sylvester, Remy Allais, Christophe Renou*
9865 * Whirlpool hash implementation is added.
9869 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9870 bn(64,32). Because of instruction set limitations it doesn't have
9871 any negative impact on performance. This was done mostly in order
9872 to make it possible to share assembler modules, such as bn_mul_mont
9873 implementations, between 32- and 64-bit builds without hassle.
9877 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9878 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9883 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9884 dedicated Montgomery multiplication procedure, is introduced.
9885 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9886 "64-bit" performance on certain 32-bit targets.
9890 * New option SSL_OP_NO_COMP to disable use of compression selectively
9891 in SSL structures. New SSL ctrl to set maximum send fragment size.
9892 Save memory by setting the I/O buffer sizes dynamically instead of
9893 using the maximum available value.
9897 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9898 in addition to the text details.
9902 * Very, very preliminary EXPERIMENTAL support for printing of general
9903 ASN1 structures. This currently produces rather ugly output and doesn't
9904 handle several customised structures at all.
9908 * Integrated support for PVK file format and some related formats such
9909 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9910 these in the 'rsa' and 'dsa' utilities.
9914 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9918 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9919 place for the (very old) "NETSCAPE" format certificates which are now
9920 handled using new ASN1 code equivalents.
9924 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9925 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9926 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9930 * Modify CRL distribution points extension code to print out previously
9931 unsupported fields. Enhance extension setting code to allow setting of
9936 * Add print and set support for Issuing Distribution Point CRL extension.
9940 * Change 'Configure' script to enable Camellia by default.
9947 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9949 * When rejecting SSL/TLS records due to an incorrect version number, never
9950 update s->server with a new major version number. As of
9951 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9952 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9953 the previous behavior could result in a read attempt at NULL when
9954 receiving specific incorrect SSL/TLS records once record payload
9955 protection is active. ([CVE-2010-0740])
9957 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9959 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9960 could be crashed if the relevant tables were not present (e.g. chrooted).
9962 *Tomas Hoger <thoger@redhat.com>*
9964 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9966 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9968 *Martin Olsson, Neel Mehta*
9970 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9971 accommodate for stack sorting, always a write lock!).
9975 * On some versions of WIN32 Heap32Next is very slow. This can cause
9976 excessive delays in the RAND_poll(): over a minute. As a workaround
9977 include a time check in the inner Heap32Next loop too.
9981 * The code that handled flushing of data in SSL/TLS originally used the
9982 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9983 the problem outlined in PR#1949. The fix suggested there however can
9984 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9985 of Apache). So instead simplify the code to flush unconditionally.
9986 This should be fine since flushing with no data to flush is a no op.
9990 * Handle TLS versions 2.0 and later properly and correctly use the
9991 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9992 off ancient servers have a habit of sticking around for a while...
9996 * Modify compression code so it frees up structures without using the
9997 ex_data callbacks. This works around a problem where some applications
9998 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9999 restarting) then use compression (e.g. SSL with compression) later.
10000 This results in significant per-connection memory leaks and
10001 has caused some security issues including CVE-2008-1678 and
10006 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10007 change when encrypting or decrypting.
10011 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10012 connect and renegotiate with servers which do not support RI.
10013 Until RI is more widely deployed this option is enabled by default.
10017 * Add "missing" ssl ctrls to clear options and mode.
10021 * If client attempts to renegotiate and doesn't support RI respond with
10022 a no_renegotiation alert as required by RFC5746. Some renegotiating
10023 TLS clients will continue a connection gracefully when they receive
10024 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10025 waiting for a server hello which it will never receive. Now we treat a
10026 received no_renegotiation alert as a fatal error. This is because
10027 applications requesting a renegotiation might well expect it to succeed
10028 and would have no code in place to handle the server denying it so the
10029 only safe thing to do is to terminate the connection.
10033 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10034 peer supports secure renegotiation and 0 otherwise. Print out peer
10035 renegotiation support in s_client/s_server.
10039 * Replace the highly broken and deprecated SPKAC certification method with
10040 the updated NID creation version. This should correctly handle UTF8.
10044 * Implement RFC5746. Re-enable renegotiation but require the extension
10045 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10046 turns out to be a bad idea. It has been replaced by
10047 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10048 SSL_CTX_set_options(). This is really not recommended unless you
10049 know what you are doing.
10051 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10053 * Fixes to stateless session resumption handling. Use initial_ctx when
10054 issuing and attempting to decrypt tickets in case it has changed during
10055 servername handling. Use a non-zero length session ID when attempting
10056 stateless session resumption: this makes it possible to determine if
10057 a resumption has occurred immediately after receiving server hello
10058 (several places in OpenSSL subtly assume this) instead of later in
10063 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10064 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10065 fixes for a few places where the return code is not checked
10068 *Julia Lawall <julia@diku.dk>*
10070 * Add --strict-warnings option to Configure script to include devteam
10071 warnings in other configurations.
10075 * Add support for --libdir option and LIBDIR variable in makefiles. This
10076 makes it possible to install openssl libraries in locations which
10077 have names other than "lib", for example "/usr/lib64" which some
10080 *Steve Henson, based on patch from Jeremy Utley*
10082 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10083 X690 8.9.12 and can produce some misleading textual output of OIDs.
10085 *Steve Henson, reported by Dan Kaminsky*
10087 * Delete MD2 from algorithm tables. This follows the recommendation in
10088 several standards that it is not used in new applications due to
10089 several cryptographic weaknesses. For binary compatibility reasons
10090 the MD2 API is still compiled in by default.
10094 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10099 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10100 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10103 *Guenter <lists@gknw.net>*
10105 * Fix the server certificate chain building code to use X509_verify_cert(),
10106 it used to have an ad-hoc builder which was unable to cope with anything
10107 other than a simple chain.
10109 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10111 * Don't check self signed certificate signatures in X509_verify_cert()
10112 by default (a flag can override this): it just wastes time without
10113 adding any security. As a useful side effect self signed root CAs
10114 with non-FIPS digests are now usable in FIPS mode.
10118 * In dtls1_process_out_of_seq_message() the check if the current message
10119 is already buffered was missing. For every new message was memory
10120 allocated, allowing an attacker to perform an denial of service attack
10121 with sending out of seq handshake messages until there is no memory
10122 left. Additionally every future message was buffered, even if the
10123 sequence number made no sense and would be part of another handshake.
10124 So only messages with sequence numbers less than 10 in advance will be
10125 buffered. ([CVE-2009-1378])
10127 *Robin Seggelmann, discovered by Daniel Mentz*
10129 * Records are buffered if they arrive with a future epoch to be
10130 processed after finishing the corresponding handshake. There is
10131 currently no limitation to this buffer allowing an attacker to perform
10132 a DOS attack with sending records with future epochs until there is no
10133 memory left. This patch adds the pqueue_size() function to determine
10134 the size of a buffer and limits the record buffer to 100 entries.
10137 *Robin Seggelmann, discovered by Daniel Mentz*
10139 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10140 parent structure is freed. ([CVE-2009-1379])
10144 * Handle non-blocking I/O properly in SSL_shutdown() call.
10146 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10148 * Add `2.5.4.*` OIDs
10150 *Ilya O. <vrghost@gmail.com>*
10152 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10154 * Disable renegotiation completely - this fixes a severe security
10155 problem ([CVE-2009-3555]) at the cost of breaking all
10156 renegotiation. Renegotiation can be re-enabled by setting
10157 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10158 run-time. This is really not recommended unless you know what
10163 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10165 * Don't set val to NULL when freeing up structures, it is freed up by
10166 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10167 zeroing past the valid field. ([CVE-2009-0789])
10169 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10171 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10172 checked correctly. This would allow some invalid signed attributes to
10173 appear to verify correctly. ([CVE-2009-0591])
10175 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10177 * Reject UniversalString and BMPString types with invalid lengths. This
10178 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10179 a legal length. ([CVE-2009-0590])
10183 * Set S/MIME signing as the default purpose rather than setting it
10184 unconditionally. This allows applications to override it at the store
10189 * Permit restricted recursion of ASN1 strings. This is needed in practice
10190 to handle some structures.
10194 * Improve efficiency of mem_gets: don't search whole buffer each time
10197 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10199 * New -hex option for openssl rand.
10203 * Print out UTF8String and NumericString when parsing ASN1.
10207 * Support NumericString type for name components.
10211 * Allow CC in the environment to override the automatically chosen
10212 compiler. Note that nothing is done to ensure flags work with the
10217 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10219 * Properly check EVP_VerifyFinal() and similar return values
10222 *Ben Laurie, Bodo Moeller, Google Security Team*
10224 * Enable TLS extensions by default.
10228 * Allow the CHIL engine to be loaded, whether the application is
10229 multithreaded or not. (This does not release the developer from the
10230 obligation to set up the dynamic locking callbacks.)
10232 *Sander Temme <sander@temme.net>*
10234 * Use correct exit code if there is an error in dgst command.
10236 *Steve Henson; problem pointed out by Roland Dirlewanger*
10238 * Tweak Configure so that you need to say "experimental-jpake" to enable
10239 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10243 * Add experimental JPAKE support, including demo authentication in
10244 s_client and s_server.
10248 * Set the comparison function in v3_addr_canonize().
10250 *Rob Austein <sra@hactrn.net>*
10252 * Add support for XMPP STARTTLS in s_client.
10254 *Philip Paeps <philip@freebsd.org>*
10256 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10257 to ensure that even with this option, only ciphersuites in the
10258 server's preference list will be accepted. (Note that the option
10259 applies only when resuming a session, so the earlier behavior was
10260 just about the algorithm choice for symmetric cryptography.)
10264 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10266 * Fix NULL pointer dereference if a DTLS server received
10267 ChangeCipherSpec as first record ([CVE-2009-1386]).
10271 * Fix a state transition in s3_srvr.c and d1_srvr.c
10272 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10274 *Nagendra Modadugu*
10276 * The fix in 0.9.8c that supposedly got rid of unsafe
10277 double-checked locking was incomplete for RSA blinding,
10278 addressing just one layer of what turns out to have been
10279 doubly unsafe triple-checked locking.
10281 So now fix this for real by retiring the MONT_HELPER macro
10282 in crypto/rsa/rsa_eay.c.
10284 *Bodo Moeller; problem pointed out by Marius Schilder*
10286 * Various precautionary measures:
10288 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10290 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10291 (NB: This would require knowledge of the secret session ticket key
10292 to exploit, in which case you'd be SOL either way.)
10294 - Change bn_nist.c so that it will properly handle input BIGNUMs
10295 outside the expected range.
10297 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10300 *Neel Mehta, Bodo Moeller*
10302 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10303 the load fails. Useful for distros.
10305 *Ben Laurie and the FreeBSD team*
10307 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10311 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10315 * Expand ENGINE to support engine supplied SSL client certificate functions.
10317 This work was sponsored by Logica.
10321 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10322 keystores. Support for SSL/TLS client authentication too.
10323 Not compiled unless enable-capieng specified to Configure.
10325 This work was sponsored by Logica.
10329 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10330 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10331 attribute creation routines such as certificate requests and PKCS#12
10336 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10338 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10339 handshake which could lead to a client crash as found using the
10340 Codenomicon TLS test suite ([CVE-2008-1672])
10342 *Steve Henson, Mark Cox*
10344 * Fix double free in TLS server name extensions which could lead to
10345 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10349 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10351 Clear the error queue to ensure that error entries left from
10352 older function calls do not interfere with the correct operation.
10354 *Lutz Jaenicke, Erik de Castro Lopo*
10356 * Remove root CA certificates of commercial CAs:
10358 The OpenSSL project does not recommend any specific CA and does not
10359 have any policy with respect to including or excluding any CA.
10360 Therefore, it does not make any sense to ship an arbitrary selection
10361 of root CA certificates with the OpenSSL software.
10365 * RSA OAEP patches to fix two separate invalid memory reads.
10366 The first one involves inputs when 'lzero' is greater than
10367 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10368 before the beginning of from). The second one involves inputs where
10369 the 'db' section contains nothing but zeroes (there is a one-byte
10370 invalid read after the end of 'db').
10372 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10374 * Partial backport from 0.9.9-dev:
10376 Introduce bn_mul_mont (dedicated Montgomery multiplication
10377 procedure) as a candidate for BIGNUM assembler implementation.
10378 While 0.9.9-dev uses assembler for various architectures, only
10379 x86_64 is available by default here in the 0.9.8 branch, and
10380 32-bit x86 is available through a compile-time setting.
10382 To try the 32-bit x86 assembler implementation, use Configure
10383 option "enable-montasm" (which exists only for this backport).
10385 As "enable-montasm" for 32-bit x86 disclaims code stability
10386 anyway, in this constellation we activate additional code
10387 backported from 0.9.9-dev for further performance improvements,
10388 namely BN_from_montgomery_word. (To enable this otherwise,
10389 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10391 *Andy Polyakov (backport partially by Bodo Moeller)*
10393 * Add TLS session ticket callback. This allows an application to set
10394 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10395 values. This is useful for key rollover for example where several key
10396 sets may exist with different names.
10400 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10401 This was broken until now in 0.9.8 releases, such that the only way
10402 a registered ENGINE could be used (assuming it initialises
10403 successfully on the host) was to explicitly set it as the default
10404 for the relevant algorithms. This is in contradiction with 0.9.7
10405 behaviour and the documentation. With this fix, when an ENGINE is
10406 registered into a given algorithm's table of implementations, the
10407 'uptodate' flag is reset so that auto-discovery will be used next
10408 time a new context for that algorithm attempts to select an
10411 *Ian Lister (tweaked by Geoff Thorpe)*
10413 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10414 implementation in the following ways:
10416 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10419 Lack of BER streaming support means one pass streaming processing is
10420 only supported if data is detached: setting the streaming flag is
10421 ignored for embedded content.
10423 CMS support is disabled by default and must be explicitly enabled
10424 with the enable-cms configuration option.
10428 * Update the GMP engine glue to do direct copies between BIGNUM and
10429 mpz_t when openssl and GMP use the same limb size. Otherwise the
10430 existing "conversion via a text string export" trick is still used.
10432 *Paul Sheer <paulsheer@gmail.com>*
10434 * Zlib compression BIO. This is a filter BIO which compressed and
10435 uncompresses any data passed through it.
10439 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10440 RFC3394 compatible AES key wrapping.
10444 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10445 sets string data without copying. X509_ALGOR_set0() and
10446 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10447 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10448 from an X509_ATTRIBUTE structure optionally checking it occurs only
10449 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10454 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10455 to get the expected BN_FLG_CONSTTIME behavior.
10457 *Bodo Moeller (Google)*
10461 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10462 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10463 - added some more tests to do_tests.pl
10464 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10465 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10466 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10467 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10468 - various changes to netware.pl to enable gcc-cross builds on Win32
10470 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10471 - various changes to fix missing prototype warnings
10472 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10473 - added AES, WHIRLPOOL and CPUID assembler code to build files
10474 - added missing AES assembler make rules to mk1mf.pl
10475 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10477 *Guenter Knauf <eflash@gmx.net>*
10479 * Implement certificate status request TLS extension defined in RFC3546.
10480 A client can set the appropriate parameters and receive the encoded
10481 OCSP response via a callback. A server can query the supplied parameters
10482 and set the encoded OCSP response in the callback. Add simplified examples
10483 to s_client and s_server.
10487 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10489 * Fix various bugs:
10490 + Binary incompatibility of ssl_ctx_st structure
10491 + DTLS interoperation with non-compliant servers
10492 + Don't call get_session_cb() without proposed session
10493 + Fix ia64 assembler code
10495 *Andy Polyakov, Steve Henson*
10497 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10499 * DTLS Handshake overhaul. There were longstanding issues with
10500 OpenSSL DTLS implementation, which were making it impossible for
10501 RFC 4347 compliant client to communicate with OpenSSL server.
10502 Unfortunately just fixing these incompatibilities would "cut off"
10503 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10504 server keeps tolerating non RFC compliant syntax. The opposite is
10505 not true, 0.9.8f client can not communicate with earlier server.
10506 This update even addresses CVE-2007-4995.
10510 * Changes to avoid need for function casts in OpenSSL: some compilers
10511 (gcc 4.2 and later) reject their use.
10512 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10515 * Add RFC4507 support to OpenSSL. This includes the corrections in
10516 RFC4507bis. The encrypted ticket format is an encrypted encoded
10517 SSL_SESSION structure, that way new session features are automatically
10520 If a client application caches session in an SSL_SESSION structure
10521 support is transparent because tickets are now stored in the encoded
10524 The SSL_CTX structure automatically generates keys for ticket
10525 protection in servers so again support should be possible
10526 with no application modification.
10528 If a client or server wishes to disable RFC4507 support then the option
10529 SSL_OP_NO_TICKET can be set.
10531 Add a TLS extension debugging callback to allow the contents of any client
10532 or server extensions to be examined.
10534 This work was sponsored by Google.
10538 * Add initial support for TLS extensions, specifically for the server_name
10539 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10540 have new members for a hostname. The SSL data structure has an
10541 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10542 stored in that context to allow for session resumption, even after the
10543 SSL has been switched to a new SSL_CTX in reaction to a client's
10544 server_name extension.
10546 New functions (subject to change):
10548 SSL_get_servername()
10549 SSL_get_servername_type()
10552 New CTRL codes and macros (subject to change):
10554 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10555 - SSL_CTX_set_tlsext_servername_callback()
10556 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10557 - SSL_CTX_set_tlsext_servername_arg()
10558 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10560 openssl s_client has a new '-servername ...' option.
10562 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10563 '-key2 ...', '-servername_fatal' (subject to change). This allows
10564 testing the HostName extension for a specific single hostname ('-cert'
10565 and '-key' remain fallbacks for handshakes without HostName
10566 negotiation). If the unrecognized_name alert has to be sent, this by
10567 default is a warning; it becomes fatal with the '-servername_fatal'
10570 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10572 * Add AES and SSE2 assembly language support to VC++ build.
10576 * Mitigate attack on final subtraction in Montgomery reduction.
10580 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10581 (which previously caused an internal error).
10585 * Squeeze another 10% out of IGE mode when in != out.
10589 * AES IGE mode speedup.
10591 *Dean Gaudet (Google)*
10593 * Add the Korean symmetric 128-bit cipher SEED (see
10594 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10595 add SEED ciphersuites from RFC 4162:
10597 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10598 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10599 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10600 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10602 To minimize changes between patchlevels in the OpenSSL 0.9.8
10603 series, SEED remains excluded from compilation unless OpenSSL
10604 is configured with 'enable-seed'.
10606 *KISA, Bodo Moeller*
10608 * Mitigate branch prediction attacks, which can be practical if a
10609 single processor is shared, allowing a spy process to extract
10610 information. For detailed background information, see
10611 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10612 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10613 and Necessary Software Countermeasures"). The core of the change
10614 are new versions BN_div_no_branch() and
10615 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10616 respectively, which are slower, but avoid the security-relevant
10617 conditional branches. These are automatically called by BN_div()
10618 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10619 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10620 remove a conditional branch.
10622 BN_FLG_CONSTTIME is the new name for the previous
10623 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10624 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10625 in the exponent causes BN_mod_exp_mont() to use the alternative
10626 implementation in BN_mod_exp_mont_consttime().) The old name
10627 remains as a deprecated alias.
10629 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10630 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10631 constant-time implementations for more than just exponentiation.
10632 Here too the old name is kept as a deprecated alias.
10634 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10635 the BN_BLINDING structure gets an independent copy of the
10636 modulus. This means that the previous `BIGNUM *m` argument to
10637 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10638 essentially becomes `const BIGNUM *m`, although we can't actually
10639 change this in the header file before 0.9.9. It allows
10640 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10641 enable BN_FLG_CONSTTIME.
10643 *Matthew D Wood (Intel Corp)*
10645 * In the SSL/TLS server implementation, be strict about session ID
10646 context matching (which matters if an application uses a single
10647 external cache for different purposes). Previously,
10648 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10649 set. This did ensure strict client verification, but meant that,
10650 with applications using a single external cache for quite
10651 different requirements, clients could circumvent ciphersuite
10652 restrictions for a given session ID context by starting a session
10653 in a different context.
10657 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10658 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10659 authentication-only ciphersuites.
10663 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10664 not complete and could lead to a possible single byte overflow
10665 ([CVE-2007-5135]) [Ben Laurie]
10667 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10669 * Since AES128 and AES256 (and similarly Camellia128 and
10670 Camellia256) share a single mask bit in the logic of
10671 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10672 kludge to work properly if AES128 is available and AES256 isn't
10673 (or if Camellia128 is available and Camellia256 isn't).
10677 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10678 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10679 When a point or a seed is encoded in a BIT STRING, we need to
10680 prevent the removal of trailing zero bits to get the proper DER
10681 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10682 of a NamedBitList, for which trailing 0 bits need to be removed.)
10686 * Have SSL/TLS server implementation tolerate "mismatched" record
10687 protocol version while receiving ClientHello even if the
10688 ClientHello is fragmented. (The server can't insist on the
10689 particular protocol version it has chosen before the ServerHello
10690 message has informed the client about his choice.)
10694 * Add RFC 3779 support.
10696 *Rob Austein for ARIN, Ben Laurie*
10698 * Load error codes if they are not already present instead of using a
10699 static variable. This allows them to be cleanly unloaded and reloaded.
10700 Improve header file function name parsing.
10704 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10705 or CAPABILITY handshake as required by RFCs.
10707 *Goetz Babin-Ebell*
10709 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10711 * Introduce limits to prevent malicious keys being able to
10712 cause a denial of service. ([CVE-2006-2940])
10714 *Steve Henson, Bodo Moeller*
10716 * Fix ASN.1 parsing of certain invalid structures that can result
10717 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10719 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10720 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10722 * Fix SSL client code which could crash if connecting to a
10723 malicious SSLv2 server. ([CVE-2006-4343])
10725 *Tavis Ormandy and Will Drewry, Google Security Team*
10727 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10728 match only those. Before that, "AES256-SHA" would be interpreted
10729 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10730 the same strength classification in 0.9.7h) as we currently only
10731 have a single AES bit in the ciphersuite description bitmap.
10732 That change, however, also applied to ciphersuite strings such as
10733 "RC4-MD5" that intentionally matched multiple ciphersuites --
10734 namely, SSL 2.0 ciphersuites in addition to the more common ones
10735 from SSL 3.0/TLS 1.0.
10737 So we change the selection algorithm again: Naming an explicit
10738 ciphersuite selects this one ciphersuite, and any other similar
10739 ciphersuite (same bitmap) from *other* protocol versions.
10740 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10741 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10743 Since SSL 2.0 does not have any ciphersuites for which the
10744 128/256 bit distinction would be relevant, this works for now.
10745 The proper fix will be to use different bits for AES128 and
10746 AES256, which would have avoided the problems from the beginning;
10747 however, bits are scarce, so we can only do this in a new release
10748 (not just a patchlevel) when we can change the SSL_CIPHER
10749 definition to split the single 'unsigned long mask' bitmap into
10750 multiple values to extend the available space.
10754 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10756 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10757 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10759 * Add AES IGE and biIGE modes.
10763 * Change the Unix randomness entropy gathering to use poll() when
10764 possible instead of select(), since the latter has some
10765 undesirable limitations.
10767 *Darryl Miles via Richard Levitte and Bodo Moeller*
10769 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10770 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10771 cannot be implicitly activated as part of, e.g., the "AES" alias.
10772 However, please upgrade to OpenSSL 0.9.9[-dev] for
10773 non-experimental use of the ECC ciphersuites to get TLS extension
10774 support, which is required for curve and point format negotiation
10775 to avoid potential handshake problems.
10779 * Disable rogue ciphersuites:
10781 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10782 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10783 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10785 The latter two were purportedly from
10786 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10789 Also deactivate the remaining ciphersuites from
10790 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10791 unofficial, and the ID has long expired.
10795 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10796 dual-core machines) and other potential thread-safety issues.
10800 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10801 versions), which is now available for royalty-free use
10802 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10803 Also, add Camellia TLS ciphersuites from RFC 4132.
10805 To minimize changes between patchlevels in the OpenSSL 0.9.8
10806 series, Camellia remains excluded from compilation unless OpenSSL
10807 is configured with 'enable-camellia'.
10811 * Disable the padding bug check when compression is in use. The padding
10812 bug check assumes the first packet is of even length, this is not
10813 necessarily true if compression is enabled and can result in false
10814 positives causing handshake failure. The actual bug test is ancient
10815 code so it is hoped that implementations will either have fixed it by
10816 now or any which still have the bug do not support compression.
10820 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10822 * When applying a cipher rule check to see if string match is an explicit
10823 cipher suite and only match that one cipher suite if it is.
10827 * Link in manifests for VC++ if needed.
10829 *Austin Ziegler <halostatue@gmail.com>*
10831 * Update support for ECC-based TLS ciphersuites according to
10832 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10833 TLS extensions, which are supported starting with the 0.9.9
10834 branch, not in the OpenSSL 0.9.8 branch).
10838 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10839 opaque EVP_CIPHER_CTX handling.
10843 * Fixes and enhancements to zlib compression code. We now only use
10844 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10845 to conform with the standards mentioned here:
10846 <http://www.zlib.net/DLL_FAQ.txt>
10847 Static zlib linking now works on Windows and the new --with-zlib-include
10848 --with-zlib-lib options to Configure can be used to supply the location
10849 of the headers and library. Gracefully handle case where zlib library
10854 * Several fixes and enhancements to the OID generation code. The old code
10855 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10856 handle numbers larger than ULONG_MAX, truncated printing and had a
10857 non standard OBJ_obj2txt() behaviour.
10861 * Add support for building of engines under engine/ as shared libraries
10862 under VC++ build system.
10866 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10867 Hopefully, we will not see any false combination of paths any more.
10871 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10873 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10874 (part of SSL_OP_ALL). This option used to disable the
10875 countermeasure against man-in-the-middle protocol-version
10876 rollback in the SSL 2.0 server implementation, which is a bad
10877 idea. ([CVE-2005-2969])
10879 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10880 for Information Security, National Institute of Advanced Industrial
10881 Science and Technology [AIST], Japan)*
10883 * Add two function to clear and return the verify parameter flags.
10887 * Keep cipherlists sorted in the source instead of sorting them at
10888 runtime, thus removing the need for a lock.
10892 * Avoid some small subgroup attacks in Diffie-Hellman.
10894 *Nick Mathewson and Ben Laurie*
10896 * Add functions for well-known primes.
10900 * Extended Windows CE support.
10902 *Satoshi Nakamura and Andy Polyakov*
10904 * Initialize SSL_METHOD structures at compile time instead of during
10905 runtime, thus removing the need for a lock.
10909 * Make PKCS7_decrypt() work even if no certificate is supplied by
10910 attempting to decrypt each encrypted key in turn. Add support to
10915 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10917 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10920 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10924 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10925 key into the same file any more.
10929 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10933 * Add -utf8 command line and config file option to 'ca'.
10935 *Stefan <stf@udoma.org*
10937 * Removed the macro des_crypt(), as it seems to conflict with some
10938 libraries. Use DES_crypt().
10942 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10943 involves renaming the source and generated shared-libs for
10944 both. The engines will accept the corrected or legacy ids
10945 ('ncipher' and '4758_cca' respectively) when binding. NB,
10946 this only applies when building 'shared'.
10948 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10950 * Add attribute functions to EVP_PKEY structure. Modify
10951 PKCS12_create() to recognize a CSP name attribute and
10952 use it. Make -CSP option work again in pkcs12 utility.
10956 * Add new functionality to the bn blinding code:
10957 - automatic re-creation of the BN_BLINDING parameters after
10958 a fixed number of uses (currently 32)
10959 - add new function for parameter creation
10960 - introduce flags to control the update behaviour of the
10961 BN_BLINDING parameters
10962 - hide BN_BLINDING structure
10963 Add a second BN_BLINDING slot to the RSA structure to improve
10964 performance when a single RSA object is shared among several
10969 * Add support for DTLS.
10971 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10973 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10974 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10978 * Remove buggy and incomplete DH cert support from
10979 ssl/ssl_rsa.c and ssl/s3_both.c
10983 * Use SHA-1 instead of MD5 as the default digest algorithm for
10984 the `apps/openssl` commands.
10988 * Compile clean with "-Wall -Wmissing-prototypes
10989 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10990 DEBUG_SAFESTACK must also be set.
10994 * Change ./Configure so that certain algorithms can be disabled by default.
10995 The new counterpiece to "no-xxx" is "enable-xxx".
10997 The patented RC5 and MDC2 algorithms will now be disabled unless
10998 "enable-rc5" and "enable-mdc2", respectively, are specified.
11000 (IDEA remains enabled despite being patented. This is because IDEA
11001 is frequently required for interoperability, and there is no license
11002 fee for non-commercial use. As before, "no-idea" can be used to
11003 avoid this algorithm.)
11007 * Add processing of proxy certificates (see RFC 3820). This work was
11008 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11009 EGEE (Enabling Grids for E-science in Europe).
11013 * RC4 performance overhaul on modern architectures/implementations, such
11014 as Intel P4, IA-64 and AMD64.
11018 * New utility extract-section.pl. This can be used specify an alternative
11019 section number in a pod file instead of having to treat each file as
11020 a separate case in Makefile. This can be done by adding two lines to the
11023 =for comment openssl_section:XXX
11025 The blank line is mandatory.
11029 * New arguments -certform, -keyform and -pass for s_client and s_server
11030 to allow alternative format key and certificate files and passphrase
11035 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11036 update associated structures and add various utility functions.
11038 Add new policy related verify parameters, include policy checking in
11039 standard verify code. Enhance 'smime' application with extra parameters
11040 to support policy checking and print out.
11044 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11045 Nehemiah processors. These extensions support AES encryption in hardware
11046 as well as RNG (though RNG support is currently disabled).
11048 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11050 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11054 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11056 *Andy Polyakov and a number of other people*
11058 * Improved PowerPC platform support. Most notably BIGNUM assembler
11059 implementation contributed by IBM.
11061 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11063 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11064 exponent rather than 'unsigned long'. There is a corresponding change to
11065 the new 'rsa_keygen' element of the RSA_METHOD structure.
11067 *Jelte Jansen, Geoff Thorpe*
11069 * Functionality for creating the initial serial number file is now
11070 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11072 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11073 number file to 1, which is bound to cause problems. To avoid
11074 the problems while respecting compatibility between different 0.9.7
11075 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11076 CA.pl for serial number initialization. With the new release 0.9.8,
11077 we can fix the problem directly in the 'ca' utility.)
11081 * Reduced header interdependencies by declaring more opaque objects in
11082 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11083 give fewer recursive includes, which could break lazy source code - so
11084 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11085 developers should define this symbol when building and using openssl to
11086 ensure they track the recommended behaviour, interfaces, [etc], but
11087 backwards-compatible behaviour prevails when this isn't defined.
11091 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11095 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11096 This will generate a random key of the appropriate length based on the
11097 cipher context. The EVP_CIPHER can provide its own random key generation
11098 routine to support keys of a specific form. This is used in the des and
11099 3des routines to generate a key of the correct parity. Update S/MIME
11100 code to use new functions and hence generate correct parity DES keys.
11101 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11102 valid (weak or incorrect parity).
11106 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11107 as looking them up. This is useful when the verified structure may contain
11108 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11109 present unless the new PKCS7_NO_CRL flag is asserted.
11113 * Extend ASN1 oid configuration module. It now additionally accepts the
11116 shortName = some long name, 1.2.3.4
11120 * Reimplemented the BN_CTX implementation. There is now no more static
11121 limitation on the number of variables it can handle nor the depth of the
11122 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11123 information can now expand as required, and rather than having a single
11124 static array of bignums, BN_CTX now uses a linked-list of such arrays
11125 allowing it to expand on demand whilst maintaining the usefulness of
11126 BN_CTX's "bundling".
11130 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11131 to allow all RSA operations to function using a single BN_CTX.
11135 * Preliminary support for certificate policy evaluation and checking. This
11136 is initially intended to pass the tests outlined in "Conformance Testing
11137 of Relying Party Client Certificate Path Processing Logic" v1.07.
11141 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11142 remained unused and not that useful. A variety of other little bignum
11143 tweaks and fixes have also been made continuing on from the audit (see
11148 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11149 associated ASN1, EVP and SSL functions and old ASN1 macros.
11153 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11154 and this should never fail. So the return value from the use of
11155 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11156 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11160 * BN_CTX_get() should return zero-valued bignums, providing the same
11161 initialised value as BN_new().
11163 *Geoff Thorpe, suggested by Ulf Möller*
11165 * Support for inhibitAnyPolicy certificate extension.
11169 * An audit of the BIGNUM code is underway, for which debugging code is
11170 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11171 is considered valid when processing BIGNUMs, and causes execution to
11172 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11173 further steps are taken to deliberately pollute unused data in BIGNUM
11174 structures to try and expose faulty code further on. For now, openssl will
11175 (in its default mode of operation) continue to tolerate the inconsistent
11176 forms that it has tolerated in the past, but authors and packagers should
11177 consider trying openssl and their own applications when compiled with
11178 these debugging symbols defined. It will help highlight potential bugs in
11179 their own code, and will improve the test coverage for OpenSSL itself. At
11180 some point, these tighter rules will become openssl's default to improve
11181 maintainability, though the assert()s and other overheads will remain only
11182 in debugging configurations. See bn.h for more details.
11184 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11186 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11187 that can only be obtained through BN_CTX_new() (which implicitly
11188 initialises it). The presence of this function only made it possible
11189 to overwrite an existing structure (and cause memory leaks).
11193 * Because of the callback-based approach for implementing LHASH as a
11194 template type, lh_insert() adds opaque objects to hash-tables and
11195 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11196 to clean up those corresponding objects before destroying the hash table
11197 (and losing the object pointers). So some over-zealous constifications in
11198 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11199 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11200 prototyped to have "const" restrictions on the object pointers they are
11201 given (and so aren't required to cast them away any more).
11205 * The tmdiff.h API was so ugly and minimal that our own timing utility
11206 (speed) prefers to use its own implementation. The two implementations
11207 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11208 its object type properly exposed (MS_TM) instead of casting to/from
11209 `char *`. This may still change yet if someone realises MS_TM and
11211 aren't necessarily the greatest nomenclatures - but this is what was used
11212 internally to the implementation so I've used that for now.
11216 * Ensure that deprecated functions do not get compiled when
11217 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11218 the self-tests were still using deprecated key-generation functions so
11219 these have been updated also.
11223 * Reorganise PKCS#7 code to separate the digest location functionality
11224 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11225 New function PKCS7_set_digest() to set the digest type for PKCS#7
11226 digestedData type. Add additional code to correctly generate the
11227 digestedData type and add support for this type in PKCS7 initialization
11232 * New function PKCS7_set0_type_other() this initializes a PKCS7
11233 structure of type "other".
11237 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11238 sure the loop does correctly stop and breaking ("division by zero")
11239 modulus operations are not performed. The (pre-generated) prime
11240 table crypto/bn/bn_prime.h was already correct, but it could not be
11241 re-generated on some platforms because of the "division by zero"
11242 situation in the script.
11244 *Ralf S. Engelschall*
11246 * Update support for ECC-based TLS ciphersuites according to
11247 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11248 SHA-1 now is only used for "small" curves (where the
11249 representation of a field element takes up to 24 bytes); for
11250 larger curves, the field element resulting from ECDH is directly
11251 used as premaster secret.
11253 *Douglas Stebila (Sun Microsystems Laboratories)*
11255 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11256 curve secp160r1 to the tests.
11258 *Douglas Stebila (Sun Microsystems Laboratories)*
11260 * Add the possibility to load symbols globally with DSO.
11262 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11264 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11265 control of the error stack.
11269 * Add support for STORE in ENGINE.
11273 * Add the STORE type. The intention is to provide a common interface
11274 to certificate and key stores, be they simple file-based stores, or
11275 HSM-type store, or LDAP stores, or...
11276 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11280 * Add a generic structure called OPENSSL_ITEM. This can be used to
11281 pass a list of arguments to any function as well as provide a way
11282 for a function to pass data back to the caller.
11286 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11287 works like BUF_strdup() but can be used to duplicate a portion of
11288 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11293 * Add the function sk_find_ex() which works like sk_find(), but will
11294 return an index to an element even if an exact match couldn't be
11295 found. The index is guaranteed to point at the element where the
11296 searched-for key would be inserted to preserve sorting order.
11300 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11301 takes an extra flags argument for optional functionality. Currently,
11302 the following flags are defined:
11304 OBJ_BSEARCH_VALUE_ON_NOMATCH
11305 This one gets OBJ_bsearch_ex() to return a pointer to the first
11306 element where the comparing function returns a negative or zero
11309 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11310 This one gets OBJ_bsearch_ex() to return a pointer to the first
11311 element where the comparing function returns zero. This is useful
11312 if there are more than one element where the comparing function
11317 * Make it possible to create self-signed certificates with 'openssl ca'
11318 in such a way that the self-signed certificate becomes part of the
11319 CA database and uses the same mechanisms for serial number generation
11320 as all other certificate signing. The new flag '-selfsign' enables
11321 this functionality. Adapt CA.sh and CA.pl.in.
11325 * Add functionality to check the public key of a certificate request
11326 against a given private. This is useful to check that a certificate
11327 request can be signed by that key (self-signing).
11331 * Make it possible to have multiple active certificates with the same
11332 subject in the CA index file. This is done only if the keyword
11333 'unique_subject' is set to 'no' in the main CA section (default
11334 if 'CA_default') of the configuration file. The value is saved
11335 with the database itself in a separate index attribute file,
11336 named like the index file with '.attr' appended to the name.
11340 * Generate multi-valued AVAs using '+' notation in config files for
11345 * Support for nameConstraints certificate extension.
11349 * Support for policyConstraints certificate extension.
11353 * Support for policyMappings certificate extension.
11357 * Make sure the default DSA_METHOD implementation only uses its
11358 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11359 and change its own handlers to be NULL so as to remove unnecessary
11360 indirection. This lets alternative implementations fallback to the
11361 default implementation more easily.
11365 * Support for directoryName in GeneralName related extensions
11370 * Make it possible to link applications using Makefile.shared.
11371 Make that possible even when linking against static libraries!
11375 * Support for single pass processing for S/MIME signing. This now
11376 means that S/MIME signing can be done from a pipe, in addition
11377 cleartext signing (multipart/signed type) is effectively streaming
11378 and the signed data does not need to be all held in memory.
11380 This is done with a new flag PKCS7_STREAM. When this flag is set
11381 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11382 is done after the data is output (and digests calculated) in
11383 SMIME_write_PKCS7().
11387 * Add full support for -rpath/-R, both in shared libraries and
11388 applications, at least on the platforms where it's known how
11393 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11394 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11395 will now compute a table of multiples of the generator that
11396 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11397 faster (notably in the case of a single point multiplication,
11398 scalar * generator).
11400 *Nils Larsch, Bodo Moeller*
11402 * IPv6 support for certificate extensions. The various extensions
11403 which use the IP:a.b.c.d can now take IPv6 addresses using the
11404 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11409 * Added an ENGINE that implements RSA by performing private key
11410 exponentiations with the GMP library. The conversions to and from
11411 GMP's mpz_t format aren't optimised nor are any montgomery forms
11412 cached, and on x86 it appears OpenSSL's own performance has caught up.
11413 However there are likely to be other architectures where GMP could
11414 provide a boost. This ENGINE is not built in by default, but it can be
11415 specified at Configure time and should be accompanied by the necessary
11416 linker additions, eg;
11417 ./config -DOPENSSL_USE_GMP -lgmp
11421 * "openssl engine" will not display ENGINE/DSO load failure errors when
11422 testing availability of engines with "-t" - the old behaviour is
11423 produced by increasing the feature's verbosity with "-tt".
11427 * ECDSA routines: under certain error conditions uninitialized BN objects
11428 could be freed. Solution: make sure initialization is performed early
11429 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11434 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11435 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11436 software implementations. For DSA and DH, parameter generation can
11437 also be overridden by providing the appropriate method callbacks.
11441 * Change the "progress" mechanism used in key-generation and
11442 primality testing to functions that take a new BN_GENCB pointer in
11443 place of callback/argument pairs. The new API functions have `_ex`
11444 postfixes and the older functions are reimplemented as wrappers for
11445 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11446 declarations of the old functions to help (graceful) attempts to
11447 migrate to the new functions. Also, the new key-generation API
11448 functions operate on a caller-supplied key-structure and return
11449 success/failure rather than returning a key or NULL - this is to
11450 help make "keygen" another member function of RSA_METHOD etc.
11452 Example for using the new callback interface:
11454 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11455 void *my_arg = ...;
11458 BN_GENCB_set(&my_cb, my_callback, my_arg);
11460 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11461 /* For the meaning of a, b in calls to my_callback(), see the
11462 * documentation of the function that calls the callback.
11463 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11464 * my_callback should return 1 if it wants BN_is_prime_ex()
11465 * to continue, or 0 to stop.
11470 * Change the ZLIB compression method to be stateful, and make it
11471 available to TLS with the number defined in
11472 draft-ietf-tls-compression-04.txt.
11476 * Add the ASN.1 structures and functions for CertificatePair, which
11477 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11479 CertificatePair ::= SEQUENCE {
11480 forward [0] Certificate OPTIONAL,
11481 reverse [1] Certificate OPTIONAL,
11482 -- at least one of the pair shall be present -- }
11484 Also implement the PEM functions to read and write certificate
11485 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11487 This needed to be defined, mostly for the sake of the LDAP
11488 attribute crossCertificatePair, but may prove useful elsewhere as
11493 * Make it possible to inhibit symlinking of shared libraries in
11494 Makefile.shared, for Cygwin's sake.
11498 * Extend the BIGNUM API by creating a function
11499 void BN_set_negative(BIGNUM *a, int neg);
11500 and a macro that behave like
11501 int BN_is_negative(const BIGNUM *a);
11503 to avoid the need to access 'a->neg' directly in applications.
11507 * Implement fast modular reduction for pseudo-Mersenne primes
11508 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11509 EC_GROUP_new_curve_GFp() will now automatically use this
11512 *Nils Larsch <nla@trustcenter.de>*
11514 * Add new lock type (CRYPTO_LOCK_BN).
11518 * Change the ENGINE framework to automatically load engines
11519 dynamically from specific directories unless they could be
11520 found to already be built in or loaded. Move all the
11521 current engines except for the cryptodev one to a new
11522 directory engines/.
11523 The engines in engines/ are built as shared libraries if
11524 the "shared" options was given to ./Configure or ./config.
11525 Otherwise, they are inserted in libcrypto.a.
11526 /usr/local/ssl/engines is the default directory for dynamic
11527 engines, but that can be overridden at configure time through
11528 the usual use of --prefix and/or --openssldir, and at run
11529 time with the environment variable OPENSSL_ENGINES.
11531 *Geoff Thorpe and Richard Levitte*
11533 * Add Makefile.shared, a helper makefile to build shared
11534 libraries. Adapt Makefile.org.
11538 * Add version info to Win32 DLLs.
11540 *Peter 'Luna' Runestig" <peter@runestig.com>*
11542 * Add new 'medium level' PKCS#12 API. Certificates and keys
11543 can be added using this API to created arbitrary PKCS#12
11544 files while avoiding the low-level API.
11546 New options to PKCS12_create(), key or cert can be NULL and
11547 will then be omitted from the output file. The encryption
11548 algorithm NIDs can be set to -1 for no encryption, the mac
11549 iteration count can be set to 0 to omit the mac.
11551 Enhance pkcs12 utility by making the -nokeys and -nocerts
11552 options work when creating a PKCS#12 file. New option -nomac
11553 to omit the mac, NONE can be set for an encryption algorithm.
11554 New code is modified to use the enhanced PKCS12_create()
11555 instead of the low-level API.
11559 * Extend ASN1 encoder to support indefinite length constructed
11560 encoding. This can output sequences tags and octet strings in
11561 this form. Modify pk7_asn1.c to support indefinite length
11562 encoding. This is experimental and needs additional code to
11563 be useful, such as an ASN1 bio and some enhanced streaming
11566 Extend template encode functionality so that tagging is passed
11567 down to the template encoder.
11571 * Let 'openssl req' fail if an argument to '-newkey' is not
11572 recognized instead of using RSA as a default.
11576 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11577 As these are not official, they are not included in "ALL";
11578 the "ECCdraft" ciphersuite group alias can be used to select them.
11580 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11582 * Add ECDH engine support.
11584 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11586 * Add ECDH in new directory crypto/ecdh/.
11588 *Douglas Stebila (Sun Microsystems Laboratories)*
11590 * Let BN_rand_range() abort with an error after 100 iterations
11591 without success (which indicates a broken PRNG).
11595 * Change BN_mod_sqrt() so that it verifies that the input value
11596 is really the square of the return value. (Previously,
11597 BN_mod_sqrt would show GIGO behaviour.)
11601 * Add named elliptic curves over binary fields from X9.62, SECG,
11602 and WAP/WTLS; add OIDs that were still missing.
11604 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11606 * Extend the EC library for elliptic curves over binary fields
11607 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11610 EC_GF2m_simple_method
11614 EC_GROUP_new_curve_GF2m
11615 EC_GROUP_set_curve_GF2m
11616 EC_GROUP_get_curve_GF2m
11617 EC_POINT_set_affine_coordinates_GF2m
11618 EC_POINT_get_affine_coordinates_GF2m
11619 EC_POINT_set_compressed_coordinates_GF2m
11621 Point compression for binary fields is disabled by default for
11622 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11625 As binary polynomials are represented as BIGNUMs, various members
11626 of the EC_GROUP and EC_POINT data structures can be shared
11627 between the implementations for prime fields and binary fields;
11628 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11629 are essentially identical to their `..._GFp` counterparts.
11630 (For simplicity, the `..._GFp` prefix has been dropped from
11631 various internal method names.)
11633 An internal 'field_div' method (similar to 'field_mul' and
11634 'field_sqr') has been added; this is used only for binary fields.
11636 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11638 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11639 through methods ('mul', 'precompute_mult').
11641 The generic implementations (now internally called 'ec_wNAF_mul'
11642 and 'ec_wNAF_precomputed_mult') remain the default if these
11643 methods are undefined.
11645 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11647 * New function EC_GROUP_get_degree, which is defined through
11648 EC_METHOD. For curves over prime fields, this returns the bit
11649 length of the modulus.
11651 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11653 * New functions EC_GROUP_dup, EC_POINT_dup.
11654 (These simply call ..._new and ..._copy).
11656 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11658 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11659 Polynomials are represented as BIGNUMs (where the sign bit is not
11660 used) in the following functions [macros]:
11663 BN_GF2m_sub [= BN_GF2m_add]
11664 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11665 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11666 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11668 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11669 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11670 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11671 BN_GF2m_cmp [= BN_ucmp]
11673 (Note that only the 'mod' functions are actually for fields GF(2^m).
11674 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11676 For some functions, an the irreducible polynomial defining a
11677 field can be given as an 'unsigned int[]' with strictly
11678 decreasing elements giving the indices of those bits that are set;
11679 i.e., p[] represents the polynomial
11680 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11682 p[0] > p[1] > ... > p[k] = 0.
11683 This applies to the following functions:
11686 BN_GF2m_mod_mul_arr
11687 BN_GF2m_mod_sqr_arr
11688 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11689 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11690 BN_GF2m_mod_exp_arr
11691 BN_GF2m_mod_sqrt_arr
11692 BN_GF2m_mod_solve_quad_arr
11696 Conversion can be performed by the following functions:
11701 bntest.c has additional tests for binary polynomial arithmetic.
11703 Two implementations for BN_GF2m_mod_div() are available.
11704 The default algorithm simply uses BN_GF2m_mod_inv() and
11705 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11706 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11707 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11709 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11711 * Add new error code 'ERR_R_DISABLED' that can be used when some
11712 functionality is disabled at compile-time.
11714 *Douglas Stebila <douglas.stebila@sun.com>*
11716 * Change default behaviour of 'openssl asn1parse' so that more
11717 information is visible when viewing, e.g., a certificate:
11719 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11720 mode the content of non-printable OCTET STRINGs is output in a
11721 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11722 avoid the appearance of a printable string.
11724 *Nils Larsch <nla@trustcenter.de>*
11726 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11728 EC_GROUP_set_asn1_flag()
11729 EC_GROUP_get_asn1_flag()
11730 EC_GROUP_set_point_conversion_form()
11731 EC_GROUP_get_point_conversion_form()
11732 These control ASN1 encoding details:
11733 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11734 has been set to OPENSSL_EC_NAMED_CURVE.
11735 - Points are encoded in uncompressed form by default; options for
11736 asn1_for are as for point2oct, namely
11737 POINT_CONVERSION_COMPRESSED
11738 POINT_CONVERSION_UNCOMPRESSED
11739 POINT_CONVERSION_HYBRID
11741 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11743 EC_GROUP_set_seed()
11744 EC_GROUP_get0_seed()
11745 EC_GROUP_get_seed_len()
11746 This is used only for ASN1 purposes (so far).
11748 *Nils Larsch <nla@trustcenter.de>*
11750 * Add 'field_type' member to EC_METHOD, which holds the NID
11751 of the appropriate field type OID. The new function
11752 EC_METHOD_get_field_type() returns this value.
11754 *Nils Larsch <nla@trustcenter.de>*
11757 EC_POINT_point2bn()
11758 EC_POINT_bn2point()
11759 EC_POINT_point2hex()
11760 EC_POINT_hex2point()
11761 providing useful interfaces to EC_POINT_point2oct() and
11762 EC_POINT_oct2point().
11764 *Nils Larsch <nla@trustcenter.de>*
11766 * Change internals of the EC library so that the functions
11767 EC_GROUP_set_generator()
11768 EC_GROUP_get_generator()
11769 EC_GROUP_get_order()
11770 EC_GROUP_get_cofactor()
11771 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11772 to methods, which would lead to unnecessary code duplication when
11773 adding different types of curves.
11775 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11777 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11778 arithmetic, and such that modified wNAFs are generated
11779 (which avoid length expansion in many cases).
11783 * Add a function EC_GROUP_check_discriminant() (defined via
11784 EC_METHOD) that verifies that the curve discriminant is non-zero.
11786 Add a function EC_GROUP_check() that makes some sanity tests
11787 on a EC_GROUP, its generator and order. This includes
11788 EC_GROUP_check_discriminant().
11790 *Nils Larsch <nla@trustcenter.de>*
11792 * Add ECDSA in new directory crypto/ecdsa/.
11794 Add applications 'openssl ecparam' and 'openssl ecdsa'
11795 (these are based on 'openssl dsaparam' and 'openssl dsa').
11797 ECDSA support is also included in various other files across the
11798 library. Most notably,
11799 - 'openssl req' now has a '-newkey ecdsa:file' option;
11800 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11801 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11802 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11803 them suitable for ECDSA where domain parameters must be
11804 extracted before the specific public key;
11805 - ECDSA engine support has been added.
11807 *Nils Larsch <nla@trustcenter.de>*
11809 * Include some named elliptic curves, and add OIDs from X9.62,
11810 SECG, and WAP/WTLS. Each curve can be obtained from the new
11812 EC_GROUP_new_by_curve_name(),
11813 and the list of available named curves can be obtained with
11814 EC_get_builtin_curves().
11815 Also add a 'curve_name' member to EC_GROUP objects, which can be
11817 EC_GROUP_set_curve_name()
11818 EC_GROUP_get_curve_name()
11820 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11822 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11823 was actually never needed) and in BN_mul(). The removal in BN_mul()
11824 required a small change in bn_mul_part_recursive() and the addition
11825 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11826 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11827 bn_sub_words() and bn_add_words() except they take arrays with
11832 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11834 * Cleanse PEM buffers before freeing them since they may contain
11837 *Benjamin Bennett <ben@psc.edu>*
11839 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11840 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11841 authentication-only ciphersuites.
11845 * Since AES128 and AES256 share a single mask bit in the logic of
11846 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11847 kludge to work properly if AES128 is available and AES256 isn't.
11851 * Expand security boundary to match 1.1.1 module.
11855 * Remove redundant features: hash file source, editing of test vectors
11856 modify fipsld to use external fips_premain.c signature.
11860 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11861 run algorithm test programs.
11865 * Make algorithm test programs more tolerant of whitespace.
11869 * Have SSL/TLS server implementation tolerate "mismatched" record
11870 protocol version while receiving ClientHello even if the
11871 ClientHello is fragmented. (The server can't insist on the
11872 particular protocol version it has chosen before the ServerHello
11873 message has informed the client about his choice.)
11877 * Load error codes if they are not already present instead of using a
11878 static variable. This allows them to be cleanly unloaded and reloaded.
11882 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11884 * Introduce limits to prevent malicious keys being able to
11885 cause a denial of service. ([CVE-2006-2940])
11887 *Steve Henson, Bodo Moeller*
11889 * Fix ASN.1 parsing of certain invalid structures that can result
11890 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11892 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11893 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11895 * Fix SSL client code which could crash if connecting to a
11896 malicious SSLv2 server. ([CVE-2006-4343])
11898 *Tavis Ormandy and Will Drewry, Google Security Team*
11900 * Change ciphersuite string processing so that an explicit
11901 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11902 will no longer include "AES128-SHA"), and any other similar
11903 ciphersuite (same bitmap) from *other* protocol versions (so that
11904 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11905 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11906 changes from 0.9.8b and 0.9.8d.
11910 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11912 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11913 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11915 * Change the Unix randomness entropy gathering to use poll() when
11916 possible instead of select(), since the latter has some
11917 undesirable limitations.
11919 *Darryl Miles via Richard Levitte and Bodo Moeller*
11921 * Disable rogue ciphersuites:
11923 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11924 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11925 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11927 The latter two were purportedly from
11928 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11931 Also deactivate the remaining ciphersuites from
11932 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11933 unofficial, and the ID has long expired.
11937 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11938 dual-core machines) and other potential thread-safety issues.
11942 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11944 * Adapt fipsld and the build system to link against the validated FIPS
11945 module in FIPS mode.
11949 * Fixes for VC++ 2005 build under Windows.
11953 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11954 from a Windows bash shell such as MSYS. It is autodetected from the
11955 "config" script when run from a VC++ environment. Modify standard VC++
11956 build to use fipscanister.o from the GNU make build.
11960 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11962 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11963 The value now differs depending on if you build for FIPS or not.
11964 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11965 safely run with a non-FIPSed libcrypto, as it may crash because of
11966 the difference induced by this change.
11970 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11972 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11973 (part of SSL_OP_ALL). This option used to disable the
11974 countermeasure against man-in-the-middle protocol-version
11975 rollback in the SSL 2.0 server implementation, which is a bad
11976 idea. ([CVE-2005-2969])
11978 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11979 for Information Security, National Institute of Advanced Industrial
11980 Science and Technology [AIST, Japan)]*
11982 * Minimal support for X9.31 signatures and PSS padding modes. This is
11983 mainly for FIPS compliance and not fully integrated at this stage.
11987 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11988 the exponentiation using a fixed-length exponent. (Otherwise,
11989 the information leaked through timing could expose the secret key
11990 after many signatures; cf. Bleichenbacher's attack on DSA with
11995 * Make a new fixed-window mod_exp implementation the default for
11996 RSA, DSA, and DH private-key operations so that the sequence of
11997 squares and multiplies and the memory access pattern are
11998 independent of the particular secret key. This will mitigate
11999 cache-timing and potential related attacks.
12001 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12002 and this is automatically used by BN_mod_exp_mont() if the new flag
12003 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12004 will use this BN flag for private exponents unless the flag
12005 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12006 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12008 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12010 * Change the client implementation for SSLv23_method() and
12011 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12012 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12013 (Previously, the SSL 2.0 backwards compatible Client Hello
12014 message format would be used even with SSL_OP_NO_SSLv2.)
12018 * Add support for smime-type MIME parameter in S/MIME messages which some
12023 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12024 a threadsafe manner. Modify rsa code to use new function and add calls
12025 to dsa and dh code (which had race conditions before).
12029 * Include the fixed error library code in the C error file definitions
12030 instead of fixing them up at runtime. This keeps the error code
12031 structures constant.
12035 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12037 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12040 * Fixes for newer kerberos headers. NB: the casts are needed because
12041 the 'length' field is signed on one version and unsigned on another
12042 with no (?) obvious way to tell the difference, without these VC++
12043 complains. Also the "definition" of FAR (blank) is no longer included
12044 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12045 some needed definitions.
12049 * Undo Cygwin change.
12053 * Added support for proxy certificates according to RFC 3820.
12054 Because they may be a security thread to unaware applications,
12055 they must be explicitly allowed in run-time. See
12056 docs/HOWTO/proxy_certificates.txt for further information.
12060 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12062 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12063 server and client random values. Previously
12064 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12065 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12067 This change has negligible security impact because:
12069 1. Server and client random values still have 24 bytes of pseudo random
12072 2. Server and client random values are sent in the clear in the initial
12075 3. The master secret is derived using the premaster secret (48 bytes in
12076 size for static RSA ciphersuites) as well as client server and random
12079 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12082 *Stephen Henson, reported by UK NISCC*
12084 * Use Windows randomness collection on Cygwin.
12088 * Fix hang in EGD/PRNGD query when communication socket is closed
12089 prematurely by EGD/PRNGD.
12091 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12093 * Prompt for pass phrases when appropriate for PKCS12 input format.
12097 * Back-port of selected performance improvements from development
12098 branch, as well as improved support for PowerPC platforms.
12102 * Add lots of checks for memory allocation failure, error codes to indicate
12103 failure and freeing up memory if a failure occurs.
12105 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12107 * Add new -passin argument to dgst.
12111 * Perform some character comparisons of different types in X509_NAME_cmp:
12112 this is needed for some certificates that re-encode DNs into UTF8Strings
12113 (in violation of RFC3280) and can't or won't issue name rollover
12118 * Make an explicit check during certificate validation to see that
12119 the CA setting in each certificate on the chain is correct. As a
12120 side effect always do the following basic checks on extensions,
12121 not just when there's an associated purpose to the check:
12123 - if there is an unhandled critical extension (unless the user
12124 has chosen to ignore this fault)
12125 - if the path length has been exceeded (if one is set at all)
12126 - that certain extensions fit the associated purpose (if one has
12131 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12133 * Avoid a race condition when CRLs are checked in a multi threaded
12134 environment. This would happen due to the reordering of the revoked
12135 entries during signature checking and serial number lookup. Now the
12136 encoding is cached and the serial number sort performed under a lock.
12137 Add new STACK function sk_is_sorted().
12141 * Add Delta CRL to the extension code.
12145 * Various fixes to s3_pkt.c so alerts are sent properly.
12147 *David Holmes <d.holmes@f5.com>*
12149 * Reduce the chances of duplicate issuer name and serial numbers (in
12150 violation of RFC3280) using the OpenSSL certificate creation utilities.
12151 This is done by creating a random 64 bit value for the initial serial
12152 number when a serial number file is created or when a self signed
12153 certificate is created using 'openssl req -x509'. The initial serial
12154 number file is created using 'openssl x509 -next_serial' in CA.pl
12155 rather than being initialized to 1.
12159 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12161 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12162 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12164 *Joe Orton, Steve Henson*
12166 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12169 *Joe Orton, Steve Henson*
12171 * Make it possible to have multiple active certificates with the same
12172 subject in the CA index file. This is done only if the keyword
12173 'unique_subject' is set to 'no' in the main CA section (default
12174 if 'CA_default') of the configuration file. The value is saved
12175 with the database itself in a separate index attribute file,
12176 named like the index file with '.attr' appended to the name.
12180 * X509 verify fixes. Disable broken certificate workarounds when
12181 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12182 keyUsage extension present. Don't accept CRLs with unhandled critical
12183 extensions: since verify currently doesn't process CRL extensions this
12184 rejects a CRL with *any* critical extensions. Add new verify error codes
12189 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12190 A clarification of RFC2560 will require the use of OCTET STRINGs and
12191 some implementations cannot handle the current raw format. Since OpenSSL
12192 copies and compares OCSP nonces as opaque blobs without any attempt at
12193 parsing them this should not create any compatibility issues.
12197 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12198 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12199 this HMAC (and other) operations are several times slower than OpenSSL
12204 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12206 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12208 * Use the correct content when signing type "other".
12212 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12214 * Fix various bugs revealed by running the NISCC test suite:
12216 Stop out of bounds reads in the ASN1 code when presented with
12217 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12219 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12221 If verify callback ignores invalid public key errors don't try to check
12222 certificate signature with the NULL public key.
12226 * New -ignore_err option in ocsp application to stop the server
12227 exiting on the first error in a request.
12231 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12232 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12237 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12238 extra data after the compression methods not only for TLS 1.0
12239 but also for SSL 3.0 (as required by the specification).
12241 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12243 * Change X509_certificate_type() to mark the key as exported/exportable
12244 when it's 512 *bits* long, not 512 bytes.
12248 * Change AES_cbc_encrypt() so it outputs exact multiple of
12249 blocks during encryption.
12253 * Various fixes to base64 BIO and non blocking I/O. On write
12254 flushes were not handled properly if the BIO retried. On read
12255 data was not being buffered properly and had various logic bugs.
12256 This also affects blocking I/O when the data being decoded is a
12261 * Various S/MIME bugfixes and compatibility changes:
12262 output correct application/pkcs7 MIME type if
12263 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12264 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12265 of files as .eml work). Correctly handle very long lines in MIME
12270 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12272 * Countermeasure against the Klima-Pokorny-Rosa extension of
12273 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12274 a protocol version number mismatch like a decryption error
12275 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12279 * Turn on RSA blinding by default in the default implementation
12280 to avoid a timing attack. Applications that don't want it can call
12281 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12282 They would be ill-advised to do so in most cases.
12284 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12286 * Change RSA blinding code so that it works when the PRNG is not
12287 seeded (in this case, the secret RSA exponent is abused as
12288 an unpredictable seed -- if it is not unpredictable, there
12289 is no point in blinding anyway). Make RSA blinding thread-safe
12290 by remembering the creator's thread ID in rsa->blinding and
12291 having all other threads use local one-time blinding factors
12292 (this requires more computation than sharing rsa->blinding, but
12293 avoids excessive locking; and if an RSA object is not shared
12294 between threads, blinding will still be very fast).
12298 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12299 ENGINE as defaults for all supported algorithms irrespective of
12300 the 'flags' parameter. 'flags' is now honoured, so applications
12301 should make sure they are passing it correctly.
12305 * Target "mingw" now allows native Windows code to be generated in
12306 the Cygwin environment as well as with the MinGW compiler.
12310 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12312 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12313 via timing by performing a MAC computation even if incorrect
12314 block cipher padding has been found. This is a countermeasure
12315 against active attacks where the attacker has to distinguish
12316 between bad padding and a MAC verification error. ([CVE-2003-0078])
12318 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12319 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12320 Martin Vuagnoux (EPFL, Ilion)*
12322 * Make the no-err option work as intended. The intention with no-err
12323 is not to have the whole error stack handling routines removed from
12324 libcrypto, it's only intended to remove all the function name and
12325 reason texts, thereby removing some of the footprint that may not
12326 be interesting if those errors aren't displayed anyway.
12328 NOTE: it's still possible for any application or module to have its
12329 own set of error texts inserted. The routines are there, just not
12330 used by default when no-err is given.
12334 * Add support for FreeBSD on IA64.
12336 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12338 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12339 Kerberos function mit_des_cbc_cksum(). Before this change,
12340 the value returned by DES_cbc_cksum() was like the one from
12341 mit_des_cbc_cksum(), except the bytes were swapped.
12343 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12345 * Allow an application to disable the automatic SSL chain building.
12346 Before this a rather primitive chain build was always performed in
12347 ssl3_output_cert_chain(): an application had no way to send the
12348 correct chain if the automatic operation produced an incorrect result.
12350 Now the chain builder is disabled if either:
12352 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12354 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12356 The reasoning behind this is that an application would not want the
12357 auto chain building to take place if extra chain certificates are
12358 present and it might also want a means of sending no additional
12359 certificates (for example the chain has two certificates and the
12364 * Add the possibility to build without the ENGINE framework.
12366 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12368 * Under Win32 gmtime() can return NULL: check return value in
12369 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12373 * DSA routines: under certain error conditions uninitialized BN objects
12374 could be freed. Solution: make sure initialization is performed early
12375 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12376 Nils Larsch <nla@trustcenter.de> via PR#459)
12380 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12381 checked on reconnect on the client side, therefore session resumption
12382 could still fail with a "ssl session id is different" error. This
12383 behaviour is masked when SSL_OP_ALL is used due to
12384 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12385 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12386 followup to PR #377.
12390 * IA-32 assembler support enhancements: unified ELF targets, support
12391 for SCO/Caldera platforms, fix for Cygwin shared build.
12395 * Add support for FreeBSD on sparc64. As a consequence, support for
12396 FreeBSD on non-x86 processors is separate from x86 processors on
12397 the config script, much like the NetBSD support.
12399 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12401 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12403 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12406 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12407 code (06) was taken as the first octet of the session ID and the last
12408 octet was ignored consequently. As a result SSLv2 client side session
12409 caching could not have worked due to the session ID mismatch between
12411 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12416 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12417 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12422 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12423 seems that in spite of existing for more than a year, many application
12424 author have done nothing to provide the necessary callbacks, which
12425 means that this particular engine will not work properly anywhere.
12426 This is a very unfortunate situation which forces us, in the name
12427 of usability, to give the hw_ncipher.c a static lock, which is part
12429 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12430 appear in 0.9.8 or later. We EXPECT application authors to have
12431 dealt properly with this when 0.9.8 is released (unless we actually
12432 make such changes in the libcrypto locking code that changes will
12433 have to be made anyway).
12437 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12438 octets have been read, EOF or an error occurs. Without this change
12439 some truncated ASN1 structures will not produce an error.
12443 * Disable Heimdal support, since it hasn't been fully implemented.
12444 Still give the possibility to force the use of Heimdal, but with
12445 warnings and a request that patches get sent to openssl-dev.
12449 * Add the VC-CE target, introduce the WINCE sysname, and add
12450 INSTALL.WCE and appropriate conditionals to make it build.
12452 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12454 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12455 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12456 edit numbers of the version.
12458 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12460 * Introduce safe string copy and catenation functions
12461 (BUF_strlcpy() and BUF_strlcat()).
12463 *Ben Laurie (CHATS) and Richard Levitte*
12465 * Avoid using fixed-size buffers for one-line DNs.
12467 *Ben Laurie (CHATS)*
12469 * Add BUF_MEM_grow_clean() to avoid information leakage when
12470 resizing buffers containing secrets, and use where appropriate.
12472 *Ben Laurie (CHATS)*
12474 * Avoid using fixed size buffers for configuration file location.
12476 *Ben Laurie (CHATS)*
12478 * Avoid filename truncation for various CA files.
12480 *Ben Laurie (CHATS)*
12482 * Use sizeof in preference to magic numbers.
12484 *Ben Laurie (CHATS)*
12486 * Avoid filename truncation in cert requests.
12488 *Ben Laurie (CHATS)*
12490 * Add assertions to check for (supposedly impossible) buffer
12493 *Ben Laurie (CHATS)*
12495 * Don't cache truncated DNS entries in the local cache (this could
12496 potentially lead to a spoofing attack).
12498 *Ben Laurie (CHATS)*
12500 * Fix various buffers to be large enough for hex/decimal
12501 representations in a platform independent manner.
12503 *Ben Laurie (CHATS)*
12505 * Add CRYPTO_realloc_clean() to avoid information leakage when
12506 resizing buffers containing secrets, and use where appropriate.
12508 *Ben Laurie (CHATS)*
12510 * Add BIO_indent() to avoid much slightly worrying code to do
12513 *Ben Laurie (CHATS)*
12515 * Convert sprintf()/BIO_puts() to BIO_printf().
12517 *Ben Laurie (CHATS)*
12519 * buffer_gets() could terminate with the buffer only half
12522 *Ben Laurie (CHATS)*
12524 * Add assertions to prevent user-supplied crypto functions from
12525 overflowing internal buffers by having large block sizes, etc.
12527 *Ben Laurie (CHATS)*
12529 * New OPENSSL_assert() macro (similar to assert(), but enabled
12532 *Ben Laurie (CHATS)*
12534 * Eliminate unused copy of key in RC4.
12536 *Ben Laurie (CHATS)*
12538 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12540 *Ben Laurie (CHATS)*
12542 * Fix off-by-one error in EGD path.
12544 *Ben Laurie (CHATS)*
12546 * If RANDFILE path is too long, ignore instead of truncating.
12548 *Ben Laurie (CHATS)*
12550 * Eliminate unused and incorrectly sized X.509 structure
12553 *Ben Laurie (CHATS)*
12555 * Eliminate unused and dangerous function knumber().
12557 *Ben Laurie (CHATS)*
12559 * Eliminate unused and dangerous structure, KSSL_ERR.
12561 *Ben Laurie (CHATS)*
12563 * Protect against overlong session ID context length in an encoded
12564 session object. Since these are local, this does not appear to be
12567 *Ben Laurie (CHATS)*
12569 * Change from security patch (see 0.9.6e below) that did not affect
12570 the 0.9.6 release series:
12572 Remote buffer overflow in SSL3 protocol - an attacker could
12573 supply an oversized master key in Kerberos-enabled versions.
12576 *Ben Laurie (CHATS)*
12578 * Change the SSL kerb5 codes to match RFC 2712.
12582 * Make -nameopt work fully for req and add -reqopt switch.
12584 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12586 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12588 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12590 * Make sure tests can be performed even if the corresponding algorithms
12591 have been removed entirely. This was also the last step to make
12592 OpenSSL compilable with DJGPP under all reasonable conditions.
12594 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12596 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12597 to allow version independent disabling of normally unselected ciphers,
12598 which may be activated as a side-effect of selecting a single cipher.
12600 (E.g., cipher list string "RSA" enables ciphersuites that are left
12601 out of "ALL" because they do not provide symmetric encryption.
12602 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12604 *Lutz Jaenicke, Bodo Moeller*
12606 * Add appropriate support for separate platform-dependent build
12607 directories. The recommended way to make a platform-dependent
12608 build directory is the following (tested on Linux), maybe with
12611 # Place yourself outside of the OpenSSL source tree. In
12612 # this example, the environment variable OPENSSL_SOURCE
12613 # is assumed to contain the absolute OpenSSL source directory.
12614 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12615 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12616 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12617 mkdir -p `dirname $F`
12618 ln -s $OPENSSL_SOURCE/$F $F
12621 To be absolutely sure not to disturb the source tree, a "make clean"
12622 is a good thing. If it isn't successful, don't worry about it,
12623 it probably means the source directory is very clean.
12627 * Make sure any ENGINE control commands make local copies of string
12628 pointers passed to them whenever necessary. Otherwise it is possible
12629 the caller may have overwritten (or deallocated) the original string
12630 data when a later ENGINE operation tries to use the stored values.
12632 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12634 * Improve diagnostics in file reading and command-line digests.
12636 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12638 * Add AES modes CFB and OFB to the object database. Correct an
12639 error in AES-CFB decryption.
12643 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12644 allows existing EVP_CIPHER_CTX structures to be reused after
12645 calling `EVP_*Final()`. This behaviour is used by encryption
12646 BIOs and some applications. This has the side effect that
12647 applications must explicitly clean up cipher contexts with
12648 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12652 * Check the values of dna and dnb in bn_mul_recursive before calling
12653 bn_mul_comba (a non zero value means the a or b arrays do not contain
12654 n2 elements) and fallback to bn_mul_normal if either is not zero.
12658 * Fix escaping of non-ASCII characters when using the -subj option
12659 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12663 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12664 form for "surname", serialNumber has no short form.
12665 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12666 therefore remove "mail" short name for "internet 7".
12667 The OID for unique identifiers in X509 certificates is
12668 x500UniqueIdentifier, not uniqueIdentifier.
12669 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12673 * Add an "init" command to the ENGINE config module and auto initialize
12674 ENGINEs. Without any "init" command the ENGINE will be initialized
12675 after all ctrl commands have been executed on it. If init=1 the
12676 ENGINE is initialized at that point (ctrls before that point are run
12677 on the uninitialized ENGINE and after on the initialized one). If
12678 init=0 then the ENGINE will not be initialized at all.
12682 * Fix the 'app_verify_callback' interface so that the user-defined
12683 argument is actually passed to the callback: In the
12684 SSL_CTX_set_cert_verify_callback() prototype, the callback
12685 declaration has been changed from
12688 int (*cb)(X509_STORE_CTX *,void *);
12689 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12690 i=s->ctx->app_verify_callback(&ctx)
12691 has been changed into
12692 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12694 To update applications using SSL_CTX_set_cert_verify_callback(),
12695 a dummy argument can be added to their callback functions.
12697 *D. K. Smetters <smetters@parc.xerox.com>*
12699 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12701 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12703 * Add and OPENSSL_LOAD_CONF define which will cause
12704 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12705 This allows older applications to transparently support certain
12706 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12707 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12708 load the config file and OPENSSL_add_all_algorithms_conf() which will
12709 always load it have also been added.
12713 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12714 Adjust NIDs and EVP layer.
12716 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12718 * Config modules support in openssl utility.
12720 Most commands now load modules from the config file,
12721 though in a few (such as version) this isn't done
12722 because it couldn't be used for anything.
12724 In the case of ca and req the config file used is
12725 the same as the utility itself: that is the -config
12726 command line option can be used to specify an
12731 * Move default behaviour from OPENSSL_config(). If appname is NULL
12732 use "openssl_conf" if filename is NULL use default openssl config file.
12736 * Add an argument to OPENSSL_config() to allow the use of an alternative
12737 config section name. Add a new flag to tolerate a missing config file
12738 and move code to CONF_modules_load_file().
12742 * Support for crypto accelerator cards from Accelerated Encryption
12743 Processing, www.aep.ie. (Use engine 'aep')
12744 The support was copied from 0.9.6c [engine] and adapted/corrected
12745 to work with the new engine framework.
12747 *AEP Inc. and Richard Levitte*
12749 * Support for SureWare crypto accelerator cards from Baltimore
12750 Technologies. (Use engine 'sureware')
12751 The support was copied from 0.9.6c [engine] and adapted
12752 to work with the new engine framework.
12756 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12757 make the newer ENGINE framework commands for the CHIL engine work.
12759 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12761 * Make it possible to produce shared libraries on ReliantUNIX.
12763 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12765 * Add the configuration target debug-linux-ppro.
12766 Make 'openssl rsa' use the general key loading routines
12767 implemented in `apps.c`, and make those routines able to
12768 handle the key format FORMAT_NETSCAPE and the variant
12771 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12773 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12775 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12777 * Add -keyform to rsautl, and document -engine.
12779 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12781 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12782 BIO_R_NO_SUCH_FILE error code rather than the generic
12783 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12787 * Add new functions
12788 ERR_peek_last_error
12789 ERR_peek_last_error_line
12790 ERR_peek_last_error_line_data.
12791 These are similar to
12793 ERR_peek_error_line
12794 ERR_peek_error_line_data,
12795 but report on the latest error recorded rather than the first one
12796 still in the error queue.
12798 *Ben Laurie, Bodo Moeller*
12800 * default_algorithms option in ENGINE config module. This allows things
12802 default_algorithms = ALL
12803 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12807 * Preliminary ENGINE config module.
12811 * New experimental application configuration code.
12815 * Change the AES code to follow the same name structure as all other
12816 symmetric ciphers, and behave the same way. Move everything to
12817 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12819 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12821 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12823 *Ben Laurie and Theo de Raadt*
12825 * Add option to output public keys in req command.
12827 *Massimiliano Pala madwolf@openca.org*
12829 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12830 (up to about 10% better than before for P-192 and P-224).
12834 * New functions/macros
12836 SSL_CTX_set_msg_callback(ctx, cb)
12837 SSL_CTX_set_msg_callback_arg(ctx, arg)
12838 SSL_set_msg_callback(ssl, cb)
12839 SSL_set_msg_callback_arg(ssl, arg)
12841 to request calling a callback function
12843 void cb(int write_p, int version, int content_type,
12844 const void *buf, size_t len, SSL *ssl, void *arg)
12846 whenever a protocol message has been completely received
12847 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12848 protocol version according to which the SSL library interprets
12849 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12850 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12851 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12852 specification (change_cipher_spec(20), alert(21), handshake(22)).
12853 'buf' and 'len' point to the actual message, 'ssl' to the
12854 SSL object, and 'arg' is the application-defined value set by
12855 SSL[_CTX]_set_msg_callback_arg().
12857 'openssl s_client' and 'openssl s_server' have new '-msg' options
12858 to enable a callback that displays all protocol messages.
12862 * Change the shared library support so shared libraries are built as
12863 soon as the corresponding static library is finished, and thereby get
12864 openssl and the test programs linked against the shared library.
12865 This still only happens when the keyword "shard" has been given to
12866 the configuration scripts.
12868 NOTE: shared library support is still an experimental thing, and
12869 backward binary compatibility is still not guaranteed.
12871 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12873 * Add support for Subject Information Access extension.
12875 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12877 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12878 additional bytes when new memory had to be allocated, not just
12879 when reusing an existing buffer.
12883 * New command line and configuration option 'utf8' for the req command.
12884 This allows field values to be specified as UTF8 strings.
12888 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12889 runs for the former and machine-readable output for the latter.
12893 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12894 of the e-mail address in the DN (i.e., it will go into a certificate
12895 extension only). The new configuration file option 'email_in_dn = no'
12896 has the same effect.
12898 *Massimiliano Pala madwolf@openca.org*
12900 * Change all functions with names starting with `des_` to be starting
12901 with `DES_` instead. Add wrappers that are compatible with libdes,
12902 but are named `_ossl_old_des_*`. Finally, add macros that map the
12903 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12904 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12905 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12908 Since we provide two compatibility mappings, the user needs to
12909 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12910 compatibility is desired. The default (i.e., when that macro
12911 isn't defined) is OpenSSL 0.9.6c compatibility.
12913 There are also macros that enable and disable the support of old
12914 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12915 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12916 are defined, the default will apply: to support the old des routines.
12918 In either case, one must include openssl/des.h to get the correct
12919 definitions. Do not try to just include openssl/des_old.h, that
12922 NOTE: This is a major break of an old API into a new one. Software
12923 authors are encouraged to switch to the `DES_` style functions. Some
12924 time in the future, des_old.h and the libdes compatibility functions
12925 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12926 default), and then completely removed.
12930 * Test for certificates which contain unsupported critical extensions.
12931 If such a certificate is found during a verify operation it is
12932 rejected by default: this behaviour can be overridden by either
12933 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12934 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12935 X509_supported_extension() has also been added which returns 1 if a
12936 particular extension is supported.
12940 * Modify the behaviour of EVP cipher functions in similar way to digests
12941 to retain compatibility with existing code.
12945 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12946 compatibility with existing code. In particular the 'ctx' parameter does
12947 not have to be to be initialized before the call to EVP_DigestInit() and
12948 it is tidied up after a call to EVP_DigestFinal(). New function
12949 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12950 EVP_MD_CTX_copy() changed to not require the destination to be
12951 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12952 requires the destination to be valid.
12954 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12955 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12959 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12960 so that complete 'Handshake' protocol structures are kept in memory
12961 instead of overwriting 'msg_type' and 'length' with 'body' data.
12965 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12967 *Massimo Santin via Richard Levitte*
12969 * Major restructuring to the underlying ENGINE code. This includes
12970 reduction of linker bloat, separation of pure "ENGINE" manipulation
12971 (initialisation, etc) from functionality dealing with implementations
12972 of specific crypto interfaces. This change also introduces integrated
12973 support for symmetric ciphers and digest implementations - so ENGINEs
12974 can now accelerate these by providing EVP_CIPHER and EVP_MD
12975 implementations of their own. This is detailed in
12976 [crypto/engine/README.md](crypto/engine/README.md)
12977 as it couldn't be adequately described here. However, there are a few
12978 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12979 were changed in the original introduction of ENGINE code have now
12980 reverted back - the hooking from this code to ENGINE is now a good
12981 deal more passive and at run-time, operations deal directly with
12982 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12983 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12984 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12985 they were not being used by the framework as there is no concept of a
12986 BIGNUM_METHOD and they could not be generalised to the new
12987 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12988 ENGINE_cpy() has been removed as it cannot be consistently defined in
12993 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12997 * Change mkdef.pl to sort symbols that get the same entry number,
12998 and make sure the automatically generated functions `ERR_load_*`
12999 become part of libeay.num as well.
13003 * New function SSL_renegotiate_pending(). This returns true once
13004 renegotiation has been requested (either SSL_renegotiate() call
13005 or HelloRequest/ClientHello received from the peer) and becomes
13006 false once a handshake has been completed.
13007 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13008 sends a HelloRequest, but does not ensure that a handshake takes
13009 place. SSL_renegotiate_pending() is useful for checking if the
13010 client has followed the request.)
13014 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13015 By default, clients may request session resumption even during
13016 renegotiation (if session ID contexts permit); with this option,
13017 session resumption is possible only in the first handshake.
13019 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13020 more bits available for options that should not be part of
13021 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13025 * Add some demos for certificate and certificate request creation.
13029 * Make maximum certificate chain size accepted from the peer application
13030 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13031 "Douglas E. Engert" <deengert@anl.gov>.
13035 * Add support for shared libraries for Unixware-7
13036 (Boyd Lynn Gerber <gerberb@zenez.com>).
13040 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13041 be done prior to destruction. Use this to unload error strings from
13042 ENGINEs that load their own error strings. NB: This adds two new API
13043 functions to "get" and "set" this destroy handler in an ENGINE.
13047 * Alter all existing ENGINE implementations (except "openssl" and
13048 "openbsd") to dynamically instantiate their own error strings. This
13049 makes them more flexible to be built both as statically-linked ENGINEs
13050 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13051 Also, add stub code to each that makes building them as self-contained
13052 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13056 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13057 implementations into applications that are completely implemented in
13058 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13059 commands that can be used to configure what shared-library to load and
13060 to control aspects of the way it is handled. Also, made an update to
13061 the [README-Engine.md](README-Engine.md) file
13062 that brings its information up-to-date and
13063 provides some information and instructions on the "dynamic" ENGINE
13064 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13068 * Make it possible to unload ranges of ERR strings with a new
13069 "ERR_unload_strings" function.
13073 * Add a copy() function to EVP_MD.
13077 * Make EVP_MD routines take a context pointer instead of just the
13078 md_data void pointer.
13082 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13083 that the digest can only process a single chunk of data
13084 (typically because it is provided by a piece of
13085 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13086 is only going to provide a single chunk of data, and hence the
13087 framework needn't accumulate the data for oneshot drivers.
13091 * As with "ERR", make it possible to replace the underlying "ex_data"
13092 functions. This change also alters the storage and management of global
13093 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13094 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13095 index counters. The API functions that use this state have been changed
13096 to take a "class_index" rather than pointers to the class's local STACK
13097 and counter, and there is now an API function to dynamically create new
13098 classes. This centralisation allows us to (a) plug a lot of the
13099 thread-safety problems that existed, and (b) makes it possible to clean
13100 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13101 such data would previously have always leaked in application code and
13102 workarounds were in place to make the memory debugging turn a blind eye
13103 to it. Application code that doesn't use this new function will still
13104 leak as before, but their memory debugging output will announce it now
13105 rather than letting it slide.
13107 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13108 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13109 has a return value to indicate success or failure.
13113 * Make it possible to replace the underlying "ERR" functions such that the
13114 global state (2 LHASH tables and 2 locks) is only used by the "default"
13115 implementation. This change also adds two functions to "get" and "set"
13116 the implementation prior to it being automatically set the first time
13117 any other ERR function takes place. Ie. an application can call "get",
13118 pass the return value to a module it has just loaded, and that module
13119 can call its own "set" function using that value. This means the
13120 module's "ERR" operations will use (and modify) the error state in the
13121 application and not in its own statically linked copy of OpenSSL code.
13125 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13126 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13127 the operation, and provides a more encapsulated way for external code
13128 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13129 to use these functions rather than manually incrementing the counts.
13131 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13135 * Add EVP test program.
13139 * Add symmetric cipher support to ENGINE. Expect the API to change!
13143 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13144 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13145 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13146 These allow a CRL to be built without having to access X509_CRL fields
13147 directly. Modify 'ca' application to use new functions.
13151 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13152 bug workarounds. Rollback attack detection is a security feature.
13153 The problem will only arise on OpenSSL servers when TLSv1 is not
13154 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13155 Software authors not wanting to support TLSv1 will have special reasons
13156 for their choice and can explicitly enable this option.
13158 *Bodo Moeller, Lutz Jaenicke*
13160 * Rationalise EVP so it can be extended: don't include a union of
13161 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13162 (similar to those existing for EVP_CIPHER_CTX).
13167 EVP_MD_CTX_init(&md); /* new function call */
13168 EVP_DigestInit(&md, EVP_sha1());
13169 EVP_DigestUpdate(&md, in, len);
13170 EVP_DigestFinal(&md, out, NULL);
13171 EVP_MD_CTX_cleanup(&md); /* new function call */
13175 * Make DES key schedule conform to the usual scheme, as well as
13176 correcting its structure. This means that calls to DES functions
13177 now have to pass a pointer to a des_key_schedule instead of a
13178 plain des_key_schedule (which was actually always a pointer
13181 des_key_schedule ks;
13183 des_set_key_checked(..., &ks);
13184 des_ncbc_encrypt(..., &ks, ...);
13186 (Note that a later change renames 'des_...' into 'DES_...'.)
13190 * Initial reduction of linker bloat: the use of some functions, such as
13191 PEM causes large amounts of unused functions to be linked in due to
13192 poor organisation. For example pem_all.c contains every PEM function
13193 which has a knock on effect of linking in large amounts of (unused)
13194 ASN1 code. Grouping together similar functions and splitting unrelated
13195 functions prevents this.
13199 * Cleanup of EVP macros.
13203 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13204 correct `_ecb suffix`.
13208 * Add initial OCSP responder support to ocsp application. The
13209 revocation information is handled using the text based index
13210 use by the ca application. The responder can either handle
13211 requests generated internally, supplied in files (for example
13212 via a CGI script) or using an internal minimal server.
13216 * Add configuration choices to get zlib compression for TLS.
13220 * Changes to Kerberos SSL for RFC 2712 compliance:
13221 1. Implemented real KerberosWrapper, instead of just using
13222 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13223 2. Implemented optional authenticator field of KerberosWrapper.
13225 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13226 and authenticator structs; see crypto/krb5/.
13228 Generalized Kerberos calls to support multiple Kerberos libraries.
13229 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13230 via Richard Levitte*
13232 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13233 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13234 values for each of the key sizes rather than having just
13235 parameters (and 'speed' generating keys each time).
13239 * Speed up EVP routines.
13242 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13243 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13244 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13245 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13247 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13248 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13249 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13252 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13254 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13258 * Added the OS2-EMX target.
13260 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13262 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13263 New functions to support `NCONF` routines in extension code.
13264 New function `CONF_set_nconf()`
13265 to allow functions which take an `NCONF` to also handle the old `LHASH`
13266 structure: this means that the old `CONF` compatible routines can be
13267 retained (in particular w.rt. extensions) without having to duplicate the
13268 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13272 * Enhance the general user interface with mechanisms for inner control
13273 and with possibilities to have yes/no kind of prompts.
13277 * Change all calls to low-level digest routines in the library and
13278 applications to use EVP. Add missing calls to HMAC_cleanup() and
13279 don't assume HMAC_CTX can be copied using memcpy().
13281 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13283 * Add the possibility to control engines through control names but with
13284 arbitrary arguments instead of just a string.
13285 Change the key loaders to take a UI_METHOD instead of a callback
13286 function pointer. NOTE: this breaks binary compatibility with earlier
13287 versions of OpenSSL [engine].
13288 Adapt the nCipher code for these new conditions and add a card insertion
13293 * Enhance the general user interface with mechanisms to better support
13294 dialog box interfaces, application-defined prompts, the possibility
13295 to use defaults (for example default passwords from somewhere else)
13296 and interrupts/cancellations.
13300 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13301 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13305 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13306 tidy up some unnecessarily weird code in 'sk_new()').
13308 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13310 * Change the key loading routines for ENGINEs to use the same kind
13311 callback (pem_password_cb) as all other routines that need this
13316 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13317 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13318 than this minimum value is recommended.
13322 * New random seeder for OpenVMS, using the system process statistics
13323 that are easily reachable.
13327 * Windows apparently can't transparently handle global
13328 variables defined in DLLs. Initialisations such as:
13330 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13332 won't compile. This is used by the any applications that need to
13333 declare their own ASN1 modules. This was fixed by adding the option
13334 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13335 needed for static libraries under Win32.
13339 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13340 setting of purpose and trust fields. New X509_STORE trust and
13341 purpose functions and tidy up setting in other SSL functions.
13345 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13346 structure. These are inherited by X509_STORE_CTX when it is
13347 initialised. This allows various defaults to be set in the
13348 X509_STORE structure (such as flags for CRL checking and custom
13349 purpose or trust settings) for functions which only use X509_STORE_CTX
13350 internally such as S/MIME.
13352 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13353 trust settings if they are not set in X509_STORE. This allows X509_STORE
13354 purposes and trust (in S/MIME for example) to override any set by default.
13356 Add command line options for CRL checking to smime, s_client and s_server
13361 * Initial CRL based revocation checking. If the CRL checking flag(s)
13362 are set then the CRL is looked up in the X509_STORE structure and
13363 its validity and signature checked, then if the certificate is found
13364 in the CRL the verify fails with a revoked error.
13366 Various new CRL related callbacks added to X509_STORE_CTX structure.
13368 Command line options added to 'verify' application to support this.
13370 This needs some additional work, such as being able to handle multiple
13371 CRLs with different times, extension based lookup (rather than just
13372 by subject name) and ultimately more complete V2 CRL extension
13377 * Add a general user interface API (crypto/ui/). This is designed
13378 to replace things like des_read_password and friends (backward
13379 compatibility functions using this new API are provided).
13380 The purpose is to remove prompting functions from the DES code
13381 section as well as provide for prompting through dialog boxes in
13382 a window system and the like.
13386 * Add "ex_data" support to ENGINE so implementations can add state at a
13387 per-structure level rather than having to store it globally.
13391 * Make it possible for ENGINE structures to be copied when retrieved by
13392 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13393 This causes the "original" ENGINE structure to act like a template,
13394 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13395 operational state can be localised to each ENGINE structure, despite the
13396 fact they all share the same "methods". New ENGINE structures returned in
13397 this case have no functional references and the return value is the single
13398 structural reference. This matches the single structural reference returned
13399 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13404 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13405 needs to match any other type at all we need to manually clear the
13410 * Changes to the "openssl engine" utility to include;
13411 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13412 about an ENGINE's available control commands.
13413 - executing control commands from command line arguments using the
13414 '-pre' and '-post' switches. '-post' is only used if '-t' is
13415 specified and the ENGINE is successfully initialised. The syntax for
13416 the individual commands are colon-separated, for example;
13417 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13421 * New dynamic control command support for ENGINEs. ENGINEs can now
13422 declare their own commands (numbers), names (strings), descriptions,
13423 and input types for run-time discovery by calling applications. A
13424 subset of these commands are implicitly classed as "executable"
13425 depending on their input type, and only these can be invoked through
13426 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13427 can be based on user input, config files, etc). The distinction is
13428 that "executable" commands cannot return anything other than a boolean
13429 result and can only support numeric or string input, whereas some
13430 discoverable commands may only be for direct use through
13431 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13432 pointers, or other custom uses. The "executable" commands are to
13433 support parameterisations of ENGINE behaviour that can be
13434 unambiguously defined by ENGINEs and used consistently across any
13435 OpenSSL-based application. Commands have been added to all the
13436 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13437 control over shared-library paths without source code alterations.
13441 * Changed all ENGINE implementations to dynamically allocate their
13442 ENGINEs rather than declaring them statically. Apart from this being
13443 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13444 this also allows the implementations to compile without using the
13445 internal engine_int.h header.
13449 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13450 'const' value. Any code that should be able to modify a RAND_METHOD
13451 should already have non-const pointers to it (ie. they should only
13452 modify their own ones).
13456 * Made a variety of little tweaks to the ENGINE code.
13457 - "atalla" and "ubsec" string definitions were moved from header files
13458 to C code. "nuron" string definitions were placed in variables
13459 rather than hard-coded - allowing parameterisation of these values
13460 later on via ctrl() commands.
13461 - Removed unused "#if 0"'d code.
13462 - Fixed engine list iteration code so it uses ENGINE_free() to release
13463 structural references.
13464 - Constified the RAND_METHOD element of ENGINE structures.
13465 - Constified various get/set functions as appropriate and added
13466 missing functions (including a catch-all ENGINE_cpy that duplicates
13467 all ENGINE values onto a new ENGINE except reference counts/state).
13468 - Removed NULL parameter checks in get/set functions. Setting a method
13469 or function to NULL is a way of cancelling out a previously set
13470 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13471 and doesn't justify the extra error symbols and code.
13472 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13473 flags from engine_int.h to engine.h.
13474 - Changed prototypes for ENGINE handler functions (init(), finish(),
13475 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13479 * Implement binary inversion algorithm for BN_mod_inverse in addition
13480 to the algorithm using long division. The binary algorithm can be
13481 used only if the modulus is odd. On 32-bit systems, it is faster
13482 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13483 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13484 up to 450 bits. In 64-bit environments, the binary algorithm
13485 appears to be advantageous for much longer moduli; here we use it
13486 for moduli up to 2048 bits.
13490 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13491 could not support the combine flag in choice fields.
13495 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13496 extensions from a certificate request to the certificate.
13500 * Allow multiple 'certopt' and 'nameopt' options to be separated
13501 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13502 file: this allows the display of the certificate about to be
13503 signed to be customised, to allow certain fields to be included
13504 or excluded and extension details. The old system didn't display
13505 multicharacter strings properly, omitted fields not in the policy
13506 and couldn't display additional details such as extensions.
13510 * Function EC_POINTs_mul for multiple scalar multiplication
13511 of an arbitrary number of elliptic curve points
13512 \sum scalars[i]*points[i],
13513 optionally including the generator defined for the EC_GROUP:
13514 scalar*generator + \sum scalars[i]*points[i].
13516 EC_POINT_mul is a simple wrapper function for the typical case
13517 that the point list has just one item (besides the optional
13522 * First EC_METHODs for curves over GF(p):
13524 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13525 operations and provides various method functions that can also
13526 operate with faster implementations of modular arithmetic.
13528 EC_GFp_mont_method() reuses most functions that are part of
13529 EC_GFp_simple_method, but uses Montgomery arithmetic.
13531 *Bodo Moeller; point addition and point doubling
13532 implementation directly derived from source code provided by
13533 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13535 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13536 crypto/ec/ec_lib.c):
13538 Curves are EC_GROUP objects (with an optional group generator)
13539 based on EC_METHODs that are built into the library.
13541 Points are EC_POINT objects based on EC_GROUP objects.
13543 Most of the framework would be able to handle curves over arbitrary
13544 finite fields, but as there are no obvious types for fields other
13545 than GF(p), some functions are limited to that for now.
13549 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13550 that the file contains a complete HTTP response.
13554 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13555 change the def and num file printf format specifier from "%-40sXXX"
13556 to "%-39s XXX". The latter will always guarantee a space after the
13557 field while the former will cause them to run together if the field
13558 is 40 of more characters long.
13562 * Constify the cipher and digest 'method' functions and structures
13563 and modify related functions to take constant EVP_MD and EVP_CIPHER
13568 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13569 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13573 * Modify `EVP_Digest*()` routines so they now return values. Although the
13574 internal software routines can never fail additional hardware versions
13579 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13581 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13582 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13585 ERR_R_NESTED_ASN1_ERROR
13587 ERR_R_MISSING_ASN1_EOS
13588 were 4 .. 9, conflicting with
13589 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13591 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13592 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13594 Add new error code 'ERR_R_INTERNAL_ERROR'.
13598 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13603 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13604 sets the subject name for a new request or supersedes the
13605 subject name in a given request. Formats that can be parsed are
13606 'CN=Some Name, OU=myOU, C=IT'
13608 'CN=Some Name/OU=myOU/C=IT'.
13610 Add options '-batch' and '-verbose' to 'openssl req'.
13612 *Massimiliano Pala <madwolf@hackmasters.net>*
13614 * Introduce the possibility to access global variables through
13615 functions on platform were that's the best way to handle exporting
13616 global variables in shared libraries. To enable this functionality,
13617 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13618 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13619 is normally done by Configure or something similar).
13621 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13622 in the source file (foo.c) like this:
13624 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13625 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13627 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13628 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13630 OPENSSL_DECLARE_GLOBAL(int,foo);
13631 #define foo OPENSSL_GLOBAL_REF(foo)
13632 OPENSSL_DECLARE_GLOBAL(double,bar);
13633 #define bar OPENSSL_GLOBAL_REF(bar)
13635 The #defines are very important, and therefore so is including the
13636 header file everywhere where the defined globals are used.
13638 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13639 of ASN.1 items, but that structure is a bit different.
13641 The largest change is in util/mkdef.pl which has been enhanced with
13642 better and easier to understand logic to choose which symbols should
13643 go into the Windows .def files as well as a number of fixes and code
13644 cleanup (among others, algorithm keywords are now sorted
13645 lexicographically to avoid constant rewrites).
13649 * In BN_div() keep a copy of the sign of 'num' before writing the
13650 result to 'rm' because if rm==num the value will be overwritten
13651 and produce the wrong result if 'num' is negative: this caused
13652 problems with BN_mod() and BN_nnmod().
13656 * Function OCSP_request_verify(). This checks the signature on an
13657 OCSP request and verifies the signer certificate. The signer
13658 certificate is just checked for a generic purpose and OCSP request
13663 * Add OCSP_check_validity() function to check the validity of OCSP
13664 responses. OCSP responses are prepared in real time and may only
13665 be a few seconds old. Simply checking that the current time lies
13666 between thisUpdate and nextUpdate max reject otherwise valid responses
13667 caused by either OCSP responder or client clock inaccuracy. Instead
13668 we allow thisUpdate and nextUpdate to fall within a certain period of
13669 the current time. The age of the response can also optionally be
13670 checked. Two new options -validity_period and -status_age added to
13675 * If signature or public key algorithm is unrecognized print out its
13676 OID rather that just UNKNOWN.
13680 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13681 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13682 ID to be generated from the issuer certificate alone which can then be
13683 passed to OCSP_id_issuer_cmp().
13687 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13688 ASN1 modules to export functions returning ASN1_ITEM pointers
13689 instead of the ASN1_ITEM structures themselves. This adds several
13690 new macros which allow the underlying ASN1 function/structure to
13691 be accessed transparently. As a result code should not use ASN1_ITEM
13692 references directly (such as &X509_it) but instead use the relevant
13693 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13694 use of the new ASN1 code on platforms where exporting structures
13695 is problematical (for example in shared libraries) but exporting
13696 functions returning pointers to structures is not.
13700 * Add support for overriding the generation of SSL/TLS session IDs.
13701 These callbacks can be registered either in an SSL_CTX or per SSL.
13702 The purpose of this is to allow applications to control, if they wish,
13703 the arbitrary values chosen for use as session IDs, particularly as it
13704 can be useful for session caching in multiple-server environments. A
13705 command-line switch for testing this (and any client code that wishes
13706 to use such a feature) has been added to "s_server".
13708 *Geoff Thorpe, Lutz Jaenicke*
13710 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13711 of the form `#if defined(...) || defined(...) || ...` and
13712 `#if !defined(...) && !defined(...) && ...`. This also avoids
13713 the growing number of special cases it was previously handling.
13717 * Make all configuration macros available for application by making
13718 sure they are available in opensslconf.h, by giving them names starting
13719 with `OPENSSL_` to avoid conflicts with other packages and by making
13720 sure e_os2.h will cover all platform-specific cases together with
13722 Additionally, it is now possible to define configuration/platform-
13723 specific names (called "system identities"). In the C code, these
13724 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13725 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13726 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13731 * New option -set_serial to 'req' and 'x509' this allows the serial
13732 number to use to be specified on the command line. Previously self
13733 signed certificates were hard coded with serial number 0 and the
13734 CA options of 'x509' had to use a serial number in a file which was
13739 * New options to 'ca' utility to support V2 CRL entry extensions.
13740 Currently CRL reason, invalidity date and hold instruction are
13741 supported. Add new CRL extensions to V3 code and some new objects.
13745 * New function EVP_CIPHER_CTX_set_padding() this is used to
13746 disable standard block padding (aka PKCS#5 padding) in the EVP
13747 API, which was previously mandatory. This means that the data is
13748 not padded in any way and so the total length much be a multiple
13749 of the block size, otherwise an error occurs.
13753 * Initial (incomplete) OCSP SSL support.
13757 * New function OCSP_parse_url(). This splits up a URL into its host,
13758 port and path components: primarily to parse OCSP URLs. New -url
13759 option to ocsp utility.
13763 * New nonce behavior. The return value of OCSP_check_nonce() now
13764 reflects the various checks performed. Applications can decide
13765 whether to tolerate certain situations such as an absent nonce
13766 in a response when one was present in a request: the ocsp application
13767 just prints out a warning. New function OCSP_add1_basic_nonce()
13768 this is to allow responders to include a nonce in a response even if
13769 the request is nonce-less.
13773 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13774 skipped when using openssl x509 multiple times on a single input file,
13775 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13779 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13780 set string type: to handle setting ASN1_TIME structures. Fix ca
13781 utility to correctly initialize revocation date of CRLs.
13785 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13786 the clients preferred ciphersuites and rather use its own preferences.
13787 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13788 Internet Explorer by ensuring unchanged hash method during stepup.
13789 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13793 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13794 to aes and add a new 'exist' option to print out symbols that don't
13799 * Additional options to ocsp utility to allow flags to be set and
13800 additional certificates supplied.
13804 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13805 OCSP client a number of certificate to only verify the response
13810 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13811 handle the new API. Currently only ECB, CBC modes supported. Add new
13814 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13815 Encryption Standard (AES) Ciphersuites for Transport Layer
13816 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13817 not enabled by default and were not part of the "ALL" ciphersuite
13818 alias because they were not yet official; they could be
13819 explicitly requested by specifying the "AESdraft" ciphersuite
13820 group alias. In the final release of OpenSSL 0.9.7, the group
13821 alias is called "AES" and is part of "ALL".)
13823 *Ben Laurie, Steve Henson, Bodo Moeller*
13825 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13826 request to response.
13830 * Functions for OCSP responders. OCSP_request_onereq_count(),
13831 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13832 extract information from a certificate request. OCSP_response_create()
13833 creates a response and optionally adds a basic response structure.
13834 OCSP_basic_add1_status() adds a complete single response to a basic
13835 response and returns the OCSP_SINGLERESP structure just added (to allow
13836 extensions to be included for example). OCSP_basic_add1_cert() adds a
13837 certificate to a basic response and OCSP_basic_sign() signs a basic
13838 response with various flags. New helper functions ASN1_TIME_check()
13839 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13840 (converts ASN1_TIME to GeneralizedTime).
13844 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13845 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13846 structure from a certificate. X509_pubkey_digest() digests the public_key
13847 contents: this is used in various key identifiers.
13851 * Make sk_sort() tolerate a NULL argument.
13853 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13855 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13856 passed by the function are trusted implicitly. If any of them signed the
13857 response then it is assumed to be valid and is not verified.
13861 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13862 to data. This was previously part of the PKCS7 ASN1 code. This
13863 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13864 *Steve Henson, reported by Kenneth R. Robinette
13865 <support@securenetterm.com>*
13867 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13868 routines: without these tracing memory leaks is very painful.
13869 Fix leaks in PKCS12 and PKCS7 routines.
13873 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13874 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13875 effectively meant GeneralizedTime would never be used. Now it
13876 is initialised to -1 but X509_time_adj() now has to check the value
13877 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13878 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13879 *Steve Henson, reported by Kenneth R. Robinette
13880 <support@securenetterm.com>*
13882 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13883 result in a zero length in the ASN1_INTEGER structure which was
13884 not consistent with the structure when d2i_ASN1_INTEGER() was used
13885 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13886 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13887 where it did not print out a minus for negative ASN1_INTEGER.
13891 * Add summary printout to ocsp utility. The various functions which
13892 convert status values to strings have been renamed to:
13893 OCSP_response_status_str(), OCSP_cert_status_str() and
13894 OCSP_crl_reason_str() and are no longer static. New options
13895 to verify nonce values and to disable verification. OCSP response
13896 printout format cleaned up.
13900 * Add additional OCSP certificate checks. These are those specified
13901 in RFC2560. This consists of two separate checks: the CA of the
13902 certificate being checked must either be the OCSP signer certificate
13903 or the issuer of the OCSP signer certificate. In the latter case the
13904 OCSP signer certificate must contain the OCSP signing extended key
13905 usage. This check is performed by attempting to match the OCSP
13906 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13907 in the OCSP_CERTID structures of the response.
13911 * Initial OCSP certificate verification added to OCSP_basic_verify()
13912 and related routines. This uses the standard OpenSSL certificate
13913 verify routines to perform initial checks (just CA validity) and
13914 to obtain the certificate chain. Then additional checks will be
13915 performed on the chain. Currently the root CA is checked to see
13916 if it is explicitly trusted for OCSP signing. This is used to set
13917 a root CA as a global signing root: that is any certificate that
13918 chains to that CA is an acceptable OCSP signing certificate.
13922 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13923 extensions from a separate configuration file.
13924 As when reading extensions from the main configuration file,
13925 the '-extensions ...' option may be used for specifying the
13928 *Massimiliano Pala <madwolf@comune.modena.it>*
13930 * New OCSP utility. Allows OCSP requests to be generated or
13931 read. The request can be sent to a responder and the output
13932 parsed, outputted or printed in text form. Not complete yet:
13933 still needs to check the OCSP response validity.
13937 * New subcommands for 'openssl ca':
13938 `openssl ca -status <serial>` prints the status of the cert with
13939 the given serial number (according to the index file).
13940 `openssl ca -updatedb` updates the expiry status of certificates
13943 *Massimiliano Pala <madwolf@comune.modena.it>*
13945 * New '-newreq-nodes' command option to CA.pl. This is like
13946 '-newreq', but calls 'openssl req' with the '-nodes' option
13947 so that the resulting key is not encrypted.
13949 *Damien Miller <djm@mindrot.org>*
13951 * New configuration for the GNU Hurd.
13953 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13955 * Initial code to implement OCSP basic response verify. This
13956 is currently incomplete. Currently just finds the signer's
13957 certificate and verifies the signature on the response.
13961 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13962 value of OPENSSLDIR. This is available via the new '-d' option
13963 to 'openssl version', and is also included in 'openssl version -a'.
13967 * Allowing defining memory allocation callbacks that will be given
13968 file name and line number information in additional arguments
13969 (a `const char*` and an int). The basic functionality remains, as
13970 well as the original possibility to just replace malloc(),
13971 realloc() and free() by functions that do not know about these
13972 additional arguments. To register and find out the current
13973 settings for extended allocation functions, the following
13974 functions are provided:
13976 CRYPTO_set_mem_ex_functions
13977 CRYPTO_set_locked_mem_ex_functions
13978 CRYPTO_get_mem_ex_functions
13979 CRYPTO_get_locked_mem_ex_functions
13981 These work the same way as CRYPTO_set_mem_functions and friends.
13982 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13983 extended allocation function is enabled.
13984 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13985 a conventional allocation function is enabled.
13987 *Richard Levitte, Bodo Moeller*
13989 * Finish off removing the remaining LHASH function pointer casts.
13990 There should no longer be any prototype-casting required when using
13991 the LHASH abstraction, and any casts that remain are "bugs". See
13992 the callback types and macros at the head of lhash.h for details
13993 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13997 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13998 If /dev/[u]random devices are not available or do not return enough
13999 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14001 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14002 /etc/entropy will be queried once each in this sequence, querying stops
14003 when enough entropy was collected without querying more sockets.
14007 * Change the Unix RAND_poll() variant to be able to poll several
14008 random devices, as specified by DEVRANDOM, until a sufficient amount
14009 of data has been collected. We spend at most 10 ms on each file
14010 (select timeout) and read in non-blocking mode. DEVRANDOM now
14011 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14012 (previously it was just the string "/dev/urandom"), so on typical
14013 platforms the 10 ms delay will never occur.
14014 Also separate out the Unix variant to its own file, rand_unix.c.
14015 For VMS, there's a currently-empty rand_vms.c.
14019 * Move OCSP client related routines to ocsp_cl.c. These
14020 provide utility functions which an application needing
14021 to issue a request to an OCSP responder and analyse the
14022 response will typically need: as opposed to those which an
14023 OCSP responder itself would need which will be added later.
14025 OCSP_request_sign() signs an OCSP request with an API similar
14026 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14027 response. OCSP_response_get1_basic() extracts basic response
14028 from response. OCSP_resp_find_status(): finds and extracts status
14029 information from an OCSP_CERTID structure (which will be created
14030 when the request structure is built). These are built from lower
14031 level functions which work on OCSP_SINGLERESP structures but
14032 won't normally be used unless the application wishes to examine
14033 extensions in the OCSP response for example.
14035 Replace nonce routines with a pair of functions.
14036 OCSP_request_add1_nonce() adds a nonce value and optionally
14037 generates a random value. OCSP_check_nonce() checks the
14038 validity of the nonce in an OCSP response.
14042 * Change function OCSP_request_add() to OCSP_request_add0_id().
14043 This doesn't copy the supplied OCSP_CERTID and avoids the
14044 need to free up the newly created id. Change return type
14045 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14046 This can then be used to add extensions to the request.
14047 Deleted OCSP_request_new(), since most of its functionality
14048 is now in OCSP_REQUEST_new() (and the case insensitive name
14049 clash) apart from the ability to set the request name which
14050 will be added elsewhere.
14054 * Update OCSP API. Remove obsolete extensions argument from
14055 various functions. Extensions are now handled using the new
14056 OCSP extension code. New simple OCSP HTTP function which
14057 can be used to send requests and parse the response.
14061 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14062 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14063 uses the special reorder version of SET OF to sort the attributes
14064 and reorder them to match the encoded order. This resolves a long
14065 standing problem: a verify on a PKCS7 structure just after signing
14066 it used to fail because the attribute order did not match the
14067 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14068 it uses the received order. This is necessary to tolerate some broken
14069 software that does not order SET OF. This is handled by encoding
14070 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14071 to produce the required SET OF.
14075 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14076 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14077 files to get correct declarations of the ASN.1 item variables.
14081 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14082 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14083 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14084 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14085 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14086 ASN1_ITEM and no wrapper functions.
14090 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14091 replace the old function pointer based I/O routines. Change most of
14092 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14096 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14097 lines, recognize more "algorithms" that can be deselected, and make
14098 it complain about algorithm deselection that isn't recognised.
14102 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14103 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14104 to use new functions. Add NO_ASN1_OLD which can be set to remove
14105 some old style ASN1 functions: this can be used to determine if old
14106 code will still work when these eventually go away.
14110 * New extension functions for OCSP structures, these follow the
14111 same conventions as certificates and CRLs.
14115 * New function X509V3_add1_i2d(). This automatically encodes and
14116 adds an extension. Its behaviour can be customised with various
14117 flags to append, replace or delete. Various wrappers added for
14118 certificates and CRLs.
14122 * Fix to avoid calling the underlying ASN1 print routine when
14123 an extension cannot be parsed. Correct a typo in the
14124 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14128 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14129 entries for variables.
14133 * Add functionality to `apps/openssl.c` for detecting locking
14134 problems: As the program is single-threaded, all we have
14135 to do is register a locking callback using an array for
14136 storing which locks are currently held by the program.
14140 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14141 SSL_get_ex_data_X509_STORE_idx(), which is used in
14142 ssl_verify_cert_chain() and thus can be called at any time
14143 during TLS/SSL handshakes so that thread-safety is essential.
14144 Unfortunately, the ex_data design is not at all suited
14145 for multi-threaded use, so it probably should be abolished.
14149 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14151 *Broadcom, tweaked and integrated by Geoff Thorpe*
14153 * Move common extension printing code to new function
14154 X509V3_print_extensions(). Reorganise OCSP print routines and
14155 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14159 * New function X509_signature_print() to remove duplication in some
14164 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14165 set (this was treated exactly the same as SET OF previously). This
14166 is used to reorder the STACK representing the structure to match the
14167 encoding. This will be used to get round a problem where a PKCS7
14168 structure which was signed could not be verified because the STACK
14169 order did not reflect the encoded order.
14173 * Reimplement the OCSP ASN1 module using the new code.
14177 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14178 for its ASN1 operations. The old style function pointers still exist
14179 for now but they will eventually go away.
14183 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14184 completely replaces the old ASN1 functionality with a table driven
14185 encoder and decoder which interprets an ASN1_ITEM structure describing
14186 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14187 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14188 has also been converted to the new form.
14192 * Change BN_mod_exp_recp so that negative moduli are tolerated
14193 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14194 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14195 for negative moduli.
14199 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14200 of not touching the result's sign bit.
14204 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14209 * Changed the LHASH code to use prototypes for callbacks, and created
14210 macros to declare and implement thin (optionally static) functions
14211 that provide type-safety and avoid function pointer casting for the
14212 type-specific callbacks.
14216 * Added Kerberos Cipher Suites to be used with TLS, as written in
14218 *Veers Staats <staatsvr@asc.hpc.mil>,
14219 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14221 * Reformat the FAQ so the different questions and answers can be divided
14222 in sections depending on the subject.
14226 * Have the zlib compression code load ZLIB.DLL dynamically under
14231 * New function BN_mod_sqrt for computing square roots modulo a prime
14232 (using the probabilistic Tonelli-Shanks algorithm unless
14233 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14234 be handled deterministically).
14236 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14238 * Make BN_mod_inverse faster by explicitly handling small quotients
14239 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14240 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14244 * New function BN_kronecker.
14248 * Fix BN_gcd so that it works on negative inputs; the result is
14249 positive unless both parameters are zero.
14250 Previously something reasonably close to an infinite loop was
14251 possible because numbers could be growing instead of shrinking
14252 in the implementation of Euclid's algorithm.
14256 * Fix BN_is_word() and BN_is_one() macros to take into account the
14257 sign of the number in question.
14259 Fix BN_is_word(a,w) to work correctly for w == 0.
14261 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14262 because its test if the absolute value of 'a' equals 'w'.
14263 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14264 it exists mostly for use in the implementations of BN_is_zero(),
14265 BN_is_one(), and BN_is_word().
14269 * New function BN_swap.
14273 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14274 the exponentiation functions are more likely to produce reasonable
14275 results on negative inputs.
14279 * Change BN_mod_mul so that the result is always non-negative.
14280 Previously, it could be negative if one of the factors was negative;
14281 I don't think anyone really wanted that behaviour.
14285 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14286 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14287 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14288 and add new functions:
14297 BN_mod_lshift1_quick
14299 BN_mod_lshift_quick
14301 These functions always generate non-negative results.
14303 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14304 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14306 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14307 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14308 be reduced modulo `m`.
14310 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14313 The following entry accidentally appeared in the CHANGES file
14314 distributed with OpenSSL 0.9.7. The modifications described in
14315 it do *not* apply to OpenSSL 0.9.7.
14317 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14318 was actually never needed) and in BN_mul(). The removal in BN_mul()
14319 required a small change in bn_mul_part_recursive() and the addition
14320 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14321 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14322 bn_sub_words() and bn_add_words() except they take arrays with
14328 * In 'openssl passwd', verify passwords read from the terminal
14329 unless the '-salt' option is used (which usually means that
14330 verification would just waste user's time since the resulting
14331 hash is going to be compared with some given password hash)
14332 or the new '-noverify' option is used.
14334 This is an incompatible change, but it does not affect
14335 non-interactive use of 'openssl passwd' (passwords on the command
14336 line, '-stdin' option, '-in ...' option) and thus should not
14337 cause any problems.
14341 * Remove all references to RSAref, since there's no more need for it.
14345 * Make DSO load along a path given through an environment variable
14346 (SHLIB_PATH) with shl_load().
14350 * Constify the ENGINE code as a result of BIGNUM constification.
14351 Also constify the RSA code and most things related to it. In a
14352 few places, most notable in the depth of the ASN.1 code, ugly
14353 casts back to non-const were required (to be solved at a later
14358 * Make it so the openssl application has all engines loaded by default.
14362 * Constify the BIGNUM routines a little more.
14366 * Add the following functions:
14368 ENGINE_load_cswift()
14370 ENGINE_load_atalla()
14371 ENGINE_load_nuron()
14372 ENGINE_load_builtin_engines()
14374 That way, an application can itself choose if external engines that
14375 are built-in in OpenSSL shall ever be used or not. The benefit is
14376 that applications won't have to be linked with libdl or other dso
14377 libraries unless it's really needed.
14379 Changed 'openssl engine' to load all engines on demand.
14380 Changed the engine header files to avoid the duplication of some
14381 declarations (they differed!).
14385 * 'openssl engine' can now list capabilities.
14389 * Better error reporting in 'openssl engine'.
14393 * Never call load_dh_param(NULL) in s_server.
14397 * Add engine application. It can currently list engines by name and
14398 identity, and test if they are actually available.
14402 * Improve RPM specification file by forcing symbolic linking and making
14403 sure the installed documentation is also owned by root.root.
14405 *Damien Miller <djm@mindrot.org>*
14407 * Give the OpenSSL applications more possibilities to make use of
14408 keys (public as well as private) handled by engines.
14412 * Add OCSP code that comes from CertCo.
14416 * Add VMS support for the Rijndael code.
14420 * Added untested support for Nuron crypto accelerator.
14424 * Add support for external cryptographic devices. This code was
14425 previously distributed separately as the "engine" branch.
14427 *Geoff Thorpe, Richard Levitte*
14429 * Rework the filename-translation in the DSO code. It is now possible to
14430 have far greater control over how a "name" is turned into a filename
14431 depending on the operating environment and any oddities about the
14432 different shared library filenames on each system.
14436 * Support threads on FreeBSD-elf in Configure.
14440 * Fix for SHA1 assembly problem with MASM: it produces
14441 warnings about corrupt line number information when assembling
14442 with debugging information. This is caused by the overlapping
14445 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14448 NCONF_get_number() has no error checking at all. As a replacement,
14449 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14450 promoted strongly. The old NCONF_get_number is kept around for
14451 binary backward compatibility.
14452 Make it possible for methods to load from something other than a BIO,
14453 by providing a function pointer that is given a name instead of a BIO.
14454 For example, this could be used to load configuration data from an
14459 * Fix for non blocking accept BIOs. Added new I/O special reason
14460 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14461 with non blocking I/O was not possible because no retry code was
14462 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14467 * Added the beginnings of Rijndael support.
14471 * Fix for bug in DirectoryString mask setting. Add support for
14472 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14473 to allow certificate printing to more controllable, additional
14474 'certopt' option to 'x509' to allow new printing options to be
14479 * Clean old EAY MD5 hack from e_os.h.
14483 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14485 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14486 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14488 *Joe Orton, Steve Henson*
14490 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14492 * Fix additional bug revealed by the NISCC test suite:
14494 Stop bug triggering large recursion when presented with
14495 certain ASN.1 tags ([CVE-2003-0851])
14499 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14501 * Fix various bugs revealed by running the NISCC test suite:
14503 Stop out of bounds reads in the ASN1 code when presented with
14504 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14506 If verify callback ignores invalid public key errors don't try to check
14507 certificate signature with the NULL public key.
14511 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14512 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14517 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14518 extra data after the compression methods not only for TLS 1.0
14519 but also for SSL 3.0 (as required by the specification).
14521 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14523 * Change X509_certificate_type() to mark the key as exported/exportable
14524 when it's 512 *bits* long, not 512 bytes.
14528 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14530 * Countermeasure against the Klima-Pokorny-Rosa extension of
14531 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14532 a protocol version number mismatch like a decryption error
14533 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14537 * Turn on RSA blinding by default in the default implementation
14538 to avoid a timing attack. Applications that don't want it can call
14539 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14540 They would be ill-advised to do so in most cases.
14542 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14544 * Change RSA blinding code so that it works when the PRNG is not
14545 seeded (in this case, the secret RSA exponent is abused as
14546 an unpredictable seed -- if it is not unpredictable, there
14547 is no point in blinding anyway). Make RSA blinding thread-safe
14548 by remembering the creator's thread ID in rsa->blinding and
14549 having all other threads use local one-time blinding factors
14550 (this requires more computation than sharing rsa->blinding, but
14551 avoids excessive locking; and if an RSA object is not shared
14552 between threads, blinding will still be very fast).
14556 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14558 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14559 via timing by performing a MAC computation even if incorrect
14560 block cipher padding has been found. This is a countermeasure
14561 against active attacks where the attacker has to distinguish
14562 between bad padding and a MAC verification error. ([CVE-2003-0078])
14564 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14565 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14566 Martin Vuagnoux (EPFL, Ilion)*
14568 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14570 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14571 memory from its contents. This is done with a counter that will
14572 place alternating values in each byte. This can be used to solve
14573 two issues: 1) the removal of calls to memset() by highly optimizing
14574 compilers, and 2) cleansing with other values than 0, since those can
14575 be read through on certain media, for example a swap space on disk.
14579 * Bugfix: client side session caching did not work with external caching,
14580 because the session->cipher setting was not restored when reloading
14581 from the external cache. This problem was masked, when
14582 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14583 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14587 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14588 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14590 *Zeev Lieber <zeev-l@yahoo.com>*
14592 * Undo an undocumented change introduced in 0.9.6e which caused
14593 repeated calls to OpenSSL_add_all_ciphers() and
14594 OpenSSL_add_all_digests() to be ignored, even after calling
14599 * Change the default configuration reader to deal with last line not
14600 being properly terminated.
14604 * Change X509_NAME_cmp() so it applies the special rules on handling
14605 DN values that are of type PrintableString, as well as RDNs of type
14606 emailAddress where the value has the type ia5String.
14608 *stefank@valicert.com via Richard Levitte*
14610 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14611 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14612 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14613 the bitwise-OR of the two for use by the majority of applications
14614 wanting this behaviour, and update the docs. The documented
14615 behaviour and actual behaviour were inconsistent and had been
14616 changing anyway, so this is more a bug-fix than a behavioural
14619 *Geoff Thorpe, diagnosed by Nadav Har'El*
14621 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14622 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14626 * Fix initialization code race conditions in
14627 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14628 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14629 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14630 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14631 ssl2_get_cipher_by_char(),
14632 ssl3_get_cipher_by_char().
14634 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14636 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14637 the cached sessions are flushed, as the remove_cb() might use ex_data
14638 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14639 (see [openssl.org #212]).
14641 *Geoff Thorpe, Lutz Jaenicke*
14643 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14644 length, instead of the encoding length to d2i_ASN1_OBJECT.
14648 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14650 * [In 0.9.6g-engine release:]
14651 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14653 *Lynn Gazis <lgazis@rainbow.com>*
14655 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14657 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14658 and get fix the header length calculation.
14659 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14660 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14662 * Use proper error handling instead of 'assertions' in buffer
14663 overflow checks added in 0.9.6e. This prevents DoS (the
14664 assertions could call abort()).
14666 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14668 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14670 * Add various sanity checks to asn1_get_length() to reject
14671 the ASN1 length bytes if they exceed sizeof(long), will appear
14672 negative or the content length exceeds the length of the
14675 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14677 * Fix cipher selection routines: ciphers without encryption had no flags
14678 for the cipher strength set and where therefore not handled correctly
14679 by the selection routines (PR #130).
14683 * Fix EVP_dsa_sha macro.
14688 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14689 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14690 that was added in OpenSSL 0.9.6d.
14692 As the countermeasure turned out to be incompatible with some
14693 broken SSL implementations, the new option is part of SSL_OP_ALL.
14694 SSL_OP_ALL is usually employed when compatibility with weird SSL
14695 implementations is desired (e.g. '-bugs' option to 's_client' and
14696 's_server'), so the new option is automatically set in many
14701 * Changes in security patch:
14703 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14704 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14705 Air Force Materiel Command, USAF, under agreement number
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
14711 supplied buffer. ([CVE-2002-0659])
14713 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14715 * Assertions for various potential buffer overflows, not known to
14716 happen in practice.
14718 *Ben Laurie (CHATS)*
14720 * Various temporary buffers to hold ASCII versions of integers were
14721 too small for 64 bit platforms. ([CVE-2002-0655])
14722 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14724 * Remote buffer overflow in SSL3 protocol - an attacker could
14725 supply an oversized session ID to a client. ([CVE-2002-0656])
14727 *Ben Laurie (CHATS)*
14729 * Remote buffer overflow in SSL2 protocol - an attacker could
14730 supply an oversized client master key. ([CVE-2002-0656])
14732 *Ben Laurie (CHATS)*
14734 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14736 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14737 encoded as NULL) with id-dsa-with-sha1.
14739 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14741 * Check various `X509_...()` return values in `apps/req.c`.
14743 *Nils Larsch <nla@trustcenter.de>*
14745 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14746 an end-of-file condition would erroneously be flagged, when the CRLF
14747 was just at the end of a processed block. The bug was discovered when
14748 processing data through a buffering memory BIO handing the data to a
14749 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14750 <ptsekov@syntrex.com> and Nedelcho Stanev.
14754 * Implement a countermeasure against a vulnerability recently found
14755 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14756 before application data chunks to avoid the use of known IVs
14757 with data potentially chosen by the attacker.
14761 * Fix length checks in ssl3_get_client_hello().
14765 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14766 to prevent ssl3_read_internal() from incorrectly assuming that
14767 ssl3_read_bytes() found application data while handshake
14768 processing was enabled when in fact s->s3->in_read_app_data was
14769 merely automatically cleared during the initial handshake.
14771 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14773 * Fix object definitions for Private and Enterprise: they were not
14774 recognized in their shortname (=lowercase) representation. Extend
14775 obj_dat.pl to issue an error when using undefined keywords instead
14776 of silently ignoring the problem (Svenning Sorensen
14777 <sss@sss.dnsalias.net>).
14781 * Fix DH_generate_parameters() so that it works for 'non-standard'
14782 generators, i.e. generators other than 2 and 5. (Previously, the
14783 code did not properly initialise the 'add' and 'rem' values to
14784 BN_generate_prime().)
14786 In the new general case, we do not insist that 'generator' is
14787 actually a primitive root: This requirement is rather pointless;
14788 a generator of the order-q subgroup is just as good, if not
14793 * Map new X509 verification errors to alerts. Discovered and submitted by
14794 Tom Wu <tom@arcot.com>.
14798 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14799 returning non-zero before the data has been completely received
14800 when using non-blocking I/O.
14802 *Bodo Moeller; problem pointed out by John Hughes*
14804 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14806 *Ben Laurie, Lutz Jaenicke*
14808 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14809 Yoram Zahavi <YoramZ@gilian.com>).
14813 * Add information about CygWin 1.3 and on, and preserve proper
14814 configuration for the versions before that.
14816 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14818 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14819 check whether we deal with a copy of a session and do not delete from
14820 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14821 <izhar@checkpoint.com>.
14825 * Do not store session data into the internal session cache, if it
14826 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14827 flag is set). Proposed by Aslam <aslam@funk.com>.
14831 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14836 * [In 0.9.6d-engine release:]
14837 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14839 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14841 * Add the configuration target linux-s390x.
14843 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14845 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14846 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14847 variable as an indication that a ClientHello message has been
14848 received. As the flag value will be lost between multiple
14849 invocations of ssl3_accept when using non-blocking I/O, the
14850 function may not be aware that a handshake has actually taken
14851 place, thus preventing a new session from being added to the
14854 To avoid this problem, we now set s->new_session to 2 instead of
14855 using a local variable.
14857 *Lutz Jaenicke, Bodo Moeller*
14859 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14860 if the SSL_R_LENGTH_MISMATCH error is detected.
14862 *Geoff Thorpe, Bodo Moeller*
14864 * New 'shared_ldflag' column in Configure platform table.
14868 * Fix EVP_CIPHER_mode macro.
14870 *"Dan S. Camper" <dan@bti.net>*
14872 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14873 type, we must throw them away by setting rr->length to 0.
14875 *D P Chang <dpc@qualys.com>*
14877 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14879 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14880 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14881 worked incorrectly for those cases where range = `10..._2` and
14882 `3*range` is two bits longer than range.)
14886 * Only add signing time to PKCS7 structures if it is not already
14891 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14892 OBJ_ld_ce should be OBJ_id_ce.
14893 Also some ip-pda OIDs in crypto/objects/objects.txt were
14894 incorrect (cf. RFC 3039).
14896 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14898 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14899 returns early because it has nothing to do.
14901 *Andy Schneider <andy.schneider@bjss.co.uk>*
14903 * [In 0.9.6c-engine release:]
14904 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14906 *Andy Schneider <andy.schneider@bjss.co.uk>*
14908 * [In 0.9.6c-engine release:]
14909 Add support for Cryptographic Appliance's keyserver technology.
14910 (Use engine 'keyclient')
14912 *Cryptographic Appliances and Geoff Thorpe*
14914 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14915 is called via tools/c89.sh because arguments have to be
14916 rearranged (all '-L' options must appear before the first object
14919 *Richard Shapiro <rshapiro@abinitio.com>*
14921 * [In 0.9.6c-engine release:]
14922 Add support for Broadcom crypto accelerator cards, backported
14925 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14927 * [In 0.9.6c-engine release:]
14928 Add support for SureWare crypto accelerator cards from
14929 Baltimore Technologies. (Use engine 'sureware')
14931 *Baltimore Technologies and Mark Cox*
14933 * [In 0.9.6c-engine release:]
14934 Add support for crypto accelerator cards from Accelerated
14935 Encryption Processing, www.aep.ie. (Use engine 'aep')
14937 *AEP Inc. and Mark Cox*
14939 * Add a configuration entry for gcc on UnixWare.
14941 *Gary Benson <gbenson@redhat.com>*
14943 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14944 messages are stored in a single piece (fixed-length part and
14945 variable-length part combined) and fix various bugs found on the way.
14949 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14950 instead. BIO_gethostbyname() does not know what timeouts are
14951 appropriate, so entries would stay in cache even when they have
14953 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14955 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14956 faced with a pathologically small ClientHello fragment that does
14957 not contain client_version: Instead of aborting with an error,
14958 simply choose the highest available protocol version (i.e.,
14959 TLS 1.0 unless it is disabled). In practice, ClientHello
14960 messages are never sent like this, but this change gives us
14961 strictly correct behaviour at least for TLS.
14965 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14966 never resets s->method to s->ctx->method when called from within
14967 one of the SSL handshake functions.
14969 *Bodo Moeller; problem pointed out by Niko Baric*
14971 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14972 (sent using the client's version number) if client_version is
14973 smaller than the protocol version in use. Also change
14974 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14975 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14976 the client will at least see that alert.
14980 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14985 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14986 client receives HelloRequest while in a handshake.
14988 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14990 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14991 should end in 'break', not 'goto end' which circumvents various
14992 cleanups done in state SSL_ST_OK. But session related stuff
14993 must be disabled for SSL_ST_OK in the case that we just sent a
14996 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14997 before just sending a HelloRequest.
14999 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15001 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15002 reveal whether illegal block cipher padding was found or a MAC
15003 verification error occurred. (Neither SSLerr() codes nor alerts
15004 are directly visible to potential attackers, but the information
15005 may leak via logfiles.)
15007 Similar changes are not required for the SSL 2.0 implementation
15008 because the number of padding bytes is sent in clear for SSL 2.0,
15009 and the extra bytes are just ignored. However ssl/s2_pkt.c
15010 failed to verify that the purported number of padding bytes is in
15015 * Add OpenUNIX-8 support including shared libraries
15016 (Boyd Lynn Gerber <gerberb@zenez.com>).
15020 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15021 'wristwatch attack' using huge encoding parameters (cf.
15022 James H. Manger's CRYPTO 2001 paper). Note that the
15023 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15024 encoding parameters and hence was not vulnerable.
15028 * BN_sqr() bug fix.
15030 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15032 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15033 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15034 followed by modular reduction.
15036 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15038 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15039 equivalent based on BN_pseudo_rand() instead of BN_rand().
15043 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15044 This function was broken, as the check for a new client hello message
15045 to handle SGC did not allow these large messages.
15046 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15050 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15054 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15055 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15059 * Rework the configuration and shared library support for Tru64 Unix.
15060 The configuration part makes use of modern compiler features and
15061 still retains old compiler behavior for those that run older versions
15062 of the OS. The shared library support part includes a variant that
15063 uses the RPATH feature, and is available through the special
15064 configuration target "alpha-cc-rpath", which will never be selected
15067 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15069 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15070 with the same message size as in ssl3_get_certificate_request().
15071 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15072 messages might inadvertently be reject as too long.
15074 *Petr Lampa <lampa@fee.vutbr.cz>*
15076 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15080 * Modified SSL library such that the verify_callback that has been set
15081 specifically for an SSL object with SSL_set_verify() is actually being
15082 used. Before the change, a verify_callback set with this function was
15083 ignored and the verify_callback() set in the SSL_CTX at the time of
15084 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15085 to allow the necessary settings.
15089 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15090 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15091 done automatically (in contradiction to the requirements of the C
15092 standard). This made problems when used from OpenSSH.
15096 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15097 dh->length and always used
15099 BN_rand_range(priv_key, dh->p).
15101 BN_rand_range() is not necessary for Diffie-Hellman, and this
15102 specific range makes Diffie-Hellman unnecessarily inefficient if
15103 dh->length (recommended exponent length) is much smaller than the
15104 length of dh->p. We could use BN_rand_range() if the order of
15105 the subgroup was stored in the DH structure, but we only have
15110 BN_rand(priv_key, l, ...)
15112 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15119 RSA_eay_public_encrypt
15120 RSA_eay_private_decrypt
15121 RSA_eay_private_encrypt (signing)
15122 RSA_eay_public_decrypt (signature verification)
15124 (default implementations for RSA_public_encrypt,
15125 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15126 always reject numbers >= n.
15130 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15131 to synchronize access to 'locking_thread'. This is necessary on
15132 systems where access to 'locking_thread' (an 'unsigned long'
15133 variable) is not atomic.
15137 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15138 *before* setting the 'crypto_lock_rand' flag. The previous code had
15139 a race condition if 0 is a valid thread ID.
15141 *Travis Vitek <vitek@roguewave.com>*
15143 * Add support for shared libraries under Irix.
15145 *Albert Chin-A-Young <china@thewrittenword.com>*
15147 * Add configuration option to build on Linux on both big-endian and
15148 little-endian MIPS.
15150 *Ralf Baechle <ralf@uni-koblenz.de>*
15152 * Add the possibility to create shared libraries on HP-UX.
15156 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15158 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15159 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15160 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15161 PRNG state recovery was possible based on the output of
15162 one PRNG request appropriately sized to gain knowledge on
15163 'md' followed by enough consecutive 1-byte PRNG requests
15164 to traverse all of 'state'.
15166 1. When updating 'md_local' (the current thread's copy of 'md')
15167 during PRNG output generation, hash all of the previous
15168 'md_local' value, not just the half used for PRNG output.
15170 2. Make the number of bytes from 'state' included into the hash
15171 independent from the number of PRNG bytes requested.
15173 The first measure alone would be sufficient to avoid
15174 Markku-Juhani's attack. (Actually it had never occurred
15175 to me that the half of 'md_local' used for chaining was the
15176 half from which PRNG output bytes were taken -- I had always
15177 assumed that the secret half would be used.) The second
15178 measure makes sure that additional data from 'state' is never
15179 mixed into 'md_local' in small portions; this heuristically
15180 further strengthens the PRNG.
15184 * Fix crypto/bn/asm/mips3.s.
15188 * When only the key is given to "enc", the IV is undefined. Print out
15189 an error message in this case.
15193 * Handle special case when X509_NAME is empty in X509 printing routines.
15197 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15198 positive and less than q.
15202 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15203 used: it isn't thread safe and the add_lock_callback should handle
15206 *Paul Rose <Paul.Rose@bridge.com>*
15208 * Verify that incoming data obeys the block size in
15209 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15215 *Ulf Möller, Bodo Möller*
15217 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15218 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15219 when fixing the server behaviour for backwards-compatible 'client
15220 hello' messages. (Note that the attack is impractical against
15221 SSL 3.0 and TLS 1.0 anyway because length and version checking
15222 means that the probability of guessing a valid ciphertext is
15223 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15226 Before 0.9.5, the countermeasure (hide the error by generating a
15227 random 'decryption result') did not work properly because
15228 ERR_clear_error() was missing, meaning that SSL_get_error() would
15229 detect the supposedly ignored error.
15231 Both problems are now fixed.
15235 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15236 (previously it was 1024).
15240 * Fix for compatibility mode trust settings: ignore trust settings
15241 unless some valid trust or reject settings are present.
15245 * Fix for blowfish EVP: its a variable length cipher.
15249 * Fix various bugs related to DSA S/MIME verification. Handle missing
15250 parameters in DSA public key structures and return an error in the
15251 DSA routines if parameters are absent.
15255 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15256 in the current directory if neither $RANDFILE nor $HOME was set.
15257 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15258 caused some confusion to Windows users who haven't defined $HOME.
15259 Thus RAND_file_name() is changed again: e_os.h can define a
15260 DEFAULT_HOME, which will be used if $HOME is not set.
15261 For Windows, we use "C:"; on other platforms, we still require
15262 environment variables.
15264 * Move 'if (!initialized) RAND_poll()' into regions protected by
15265 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15266 having multiple threads call RAND_poll() concurrently.
15270 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15271 combination of a flag and a thread ID variable.
15272 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15273 flag), *other* threads can enter ssleay_add_bytes without obeying
15274 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15275 that they do not hold after the first thread unsets add_do_not_lock).
15279 * Change bctest again: '-x' expressions are not available in all
15280 versions of 'test'.
15284 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15286 * Fix a couple of memory leaks in PKCS7_dataDecode()
15288 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15290 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15291 the default extension for executables, if any. Also, make the perl
15292 scripts that use symlink() to test if it really exists and use "cp"
15293 if it doesn't. All this made OpenSSL compilable and installable in
15298 * Fix for asn1_GetSequence() for indefinite length constructed data.
15299 If SEQUENCE is length is indefinite just set c->slen to the total
15300 amount of data available.
15302 *Steve Henson, reported by shige@FreeBSD.org*
15304 *This change does not apply to 0.9.7.*
15306 * Change bctest to avoid here-documents inside command substitution
15307 (workaround for FreeBSD /bin/sh bug).
15308 For compatibility with Ultrix, avoid shell functions (introduced
15309 in the bctest version that searches along $PATH).
15313 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15314 with des_encrypt() defined on some operating systems, like Solaris
15319 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15320 On the Importance of Eliminating Errors in Cryptographic
15321 Computations, J. Cryptology 14 (2001) 2, 101-119,
15322 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15326 * MIPS assembler BIGNUM division bug fix.
15330 * Disabled incorrect Alpha assembler code.
15334 * Fix PKCS#7 decode routines so they correctly update the length
15335 after reading an EOC for the EXPLICIT tag.
15339 *This change does not apply to 0.9.7.*
15341 * Fix bug in PKCS#12 key generation routines. This was triggered
15342 if a 3DES key was generated with a 0 initial byte. Include
15343 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15344 (but broken) behaviour.
15348 * Enhance bctest to search for a working bc along $PATH and print
15351 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15353 * Fix memory leaks in err.c: free err_data string if necessary;
15354 don't write to the wrong index in ERR_set_error_data.
15358 * Implement ssl23_peek (analogous to ssl23_read), which previously
15363 * Replace rdtsc with `_emit` statements for VC++ version 5.
15365 *Jeremy Cooper <jeremy@baymoo.org>*
15367 * Make it possible to reuse SSLv2 sessions.
15371 * In copy_email() check for >= 0 as a return value for
15372 X509_NAME_get_index_by_NID() since 0 is a valid index.
15374 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15376 * Avoid coredump with unsupported or invalid public keys by checking if
15377 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15378 PKCS7_verify() fails with non detached data.
15382 * Don't use getenv in library functions when run as setuid/setgid.
15383 New function OPENSSL_issetugid().
15387 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15388 due to incorrect handling of multi-threading:
15390 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15392 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15394 3. Count how many times MemCheck_off() has been called so that
15395 nested use can be treated correctly. This also avoids
15396 inband-signalling in the previous code (which relied on the
15397 assumption that thread ID 0 is impossible).
15401 * Add "-rand" option also to s_client and s_server.
15405 * Fix CPU detection on Irix 6.x.
15406 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15407 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15409 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15414 *This change does not apply to 0.9.7.*
15416 * Use the cached encoding of an X509_NAME structure rather than
15417 copying it. This is apparently the reason for the libsafe "errors"
15418 but the code is actually correct.
15422 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15423 Bleichenbacher's DSA attack.
15424 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15425 to be set and top=0 forces the highest bit to be set; top=-1 is new
15426 and leaves the highest bit random.
15428 *Ulf Moeller, Bodo Moeller*
15430 * In the `NCONF_...`-based implementations for `CONF_...` queries
15431 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15432 a temporary CONF structure with the data component set to NULL
15433 (which gives segmentation faults in lh_retrieve).
15434 Instead, use NULL for the CONF pointer in CONF_get_string and
15435 CONF_get_number (which may use environment variables) and directly
15436 return NULL from CONF_get_section.
15440 * Fix potential buffer overrun for EBCDIC.
15444 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15445 keyUsage if basicConstraints absent for a CA.
15449 * Make SMIME_write_PKCS7() write mail header values with a format that
15450 is more generally accepted (no spaces before the semicolon), since
15451 some programs can't parse those values properly otherwise. Also make
15452 sure BIO's that break lines after each write do not create invalid
15457 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15458 macros previously used would not encode an empty SEQUENCE OF
15459 and break the signature.
15463 *This change does not apply to 0.9.7.*
15465 * Zero the premaster secret after deriving the master secret in
15470 * Add some EVP_add_digest_alias registrations (as found in
15471 OpenSSL_add_all_digests()) to SSL_library_init()
15472 aka OpenSSL_add_ssl_algorithms(). This provides improved
15473 compatibility with peers using X.509 certificates
15474 with unconventional AlgorithmIdentifier OIDs.
15478 * Fix for Irix with NO_ASM.
15480 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15482 * ./config script fixes.
15484 *Ulf Moeller, Richard Levitte*
15486 * Fix 'openssl passwd -1'.
15490 * Change PKCS12_key_gen_asc() so it can cope with non null
15491 terminated strings whose length is passed in the passlen
15492 parameter, for example from PEM callbacks. This was done
15493 by adding an extra length parameter to asc2uni().
15495 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15497 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15498 call failed, free the DSA structure.
15502 * Fix to uni2asc() to cope with zero length Unicode strings.
15503 These are present in some PKCS#12 files.
15507 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15508 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15509 when writing a 32767 byte record.
15511 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15513 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15514 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15516 (RSA objects have a reference count access to which is protected
15517 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15518 so they are meant to be shared between threads.)
15519 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15520 "Reddie, Steven" <Steven.Reddie@ca.com>*
15522 * Fix a deadlock in CRYPTO_mem_leaks().
15526 * Use better test patterns in bntest.
15530 * rand_win.c fix for Borland C.
15534 * BN_rshift bugfix for n == 0.
15538 * Add a 'bctest' script that checks for some known 'bc' bugs
15539 so that 'make test' does not abort just because 'bc' is broken.
15543 * Store verify_result within SSL_SESSION also for client side to
15544 avoid potential security hole. (Re-used sessions on the client side
15545 always resulted in verify_result==X509_V_OK, not using the original
15546 result of the server certificate verification.)
15550 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15551 SSL3_RT_APPLICATION_DATA, return 0.
15552 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15557 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15558 releases, have been re-implemented by renaming the previous
15559 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15560 and ssl3_read_internal, respectively, and adding 'peek' parameters
15561 to them. The new ssl[23]_{read,peek} functions are calls to
15562 ssl[23]_read_internal with the 'peek' flag set appropriately.
15563 A 'peek' parameter has also been added to ssl3_read_bytes, which
15564 does the actual work for ssl3_read_internal.
15568 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15569 the method-specific "init()" handler. Also clean up ex_data after
15570 calling the method-specific "finish()" handler. Previously, this was
15571 happening the other way round.
15575 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15576 The previous value, 12, was not always sufficient for BN_mod_exp().
15580 * Make sure that shared libraries get the internal name engine with
15581 the full version number and not just 0. This should mark the
15582 shared libraries as not backward compatible. Of course, this should
15583 be changed again when we can guarantee backward binary compatibility.
15587 * Fix typo in get_cert_by_subject() in by_dir.c
15589 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15591 * Rework the system to generate shared libraries:
15593 - Make note of the expected extension for the shared libraries and
15594 if there is a need for symbolic links from for example libcrypto.so.0
15595 to libcrypto.so.0.9.7. There is extended info in Configure for
15598 - Make as few rebuilds of the shared libraries as possible.
15600 - Still avoid linking the OpenSSL programs with the shared libraries.
15602 - When installing, install the shared libraries separately from the
15607 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15609 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15610 and not in SSL_clear because the latter is also used by the
15611 accept/connect functions; previously, the settings made by
15612 SSL_set_read_ahead would be lost during the handshake.
15614 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15616 * Correct util/mkdef.pl to be selective about disabled algorithms.
15617 Previously, it would create entries for disabled algorithms no
15622 * Added several new manual pages for SSL_* function.
15626 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15628 * In ssl23_get_client_hello, generate an error message when faced
15629 with an initial SSL 3.0/TLS record that is too small to contain the
15630 first two bytes of the ClientHello message, i.e. client_version.
15631 (Note that this is a pathologic case that probably has never happened
15632 in real life.) The previous approach was to use the version number
15633 from the record header as a substitute; but our protocol choice
15634 should not depend on that one because it is not authenticated
15635 by the Finished messages.
15639 * More robust randomness gathering functions for Windows.
15641 *Jeffrey Altman <jaltman@columbia.edu>*
15643 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15644 not set then we don't setup the error code for issuer check errors
15645 to avoid possibly overwriting other errors which the callback does
15646 handle. If an application does set the flag then we assume it knows
15647 what it is doing and can handle the new informational codes
15652 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15653 a general "ANY" type, as such it should be able to decode anything
15654 including tagged types. However it didn't check the class so it would
15655 wrongly interpret tagged types in the same way as their universal
15656 counterpart and unknown types were just rejected. Changed so that the
15657 tagged and unknown types are handled in the same way as a SEQUENCE:
15658 that is the encoding is stored intact. There is also a new type
15659 "V_ASN1_OTHER" which is used when the class is not universal, in this
15660 case we have no idea what the actual type is so we just lump them all
15665 * On VMS, stdout may very well lead to a file that is written to
15666 in a record-oriented fashion. That means that every write() will
15667 write a separate record, which will be read separately by the
15668 programs trying to read from it. This can be very confusing.
15670 The solution is to put a BIO filter in the way that will buffer
15671 text until a linefeed is reached, and then write everything a
15672 line at a time, so every record written will be an actual line,
15673 not chunks of lines and not (usually doesn't happen, but I've
15674 seen it once) several lines in one record. BIO_f_linebuffer() is
15677 Currently, it's a VMS-only method, because that's where it has
15678 been tested well enough.
15682 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15683 it can return incorrect results.
15684 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15685 but it was in 0.9.6-beta[12].)
15689 * Disable the check for content being present when verifying detached
15690 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15691 include zero length content when signing messages.
15695 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15696 BIO_ctrl (for BIO pairs).
15700 * Add DSO method for VMS.
15704 * Bug fix: Montgomery multiplication could produce results with the
15709 * Add RPM specification openssl.spec and modify it to build three
15710 packages. The default package contains applications, application
15711 documentation and run-time libraries. The devel package contains
15712 include files, static libraries and function documentation. The
15713 doc package contains the contents of the doc directory. The original
15714 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15718 * Add a large number of documentation files for many SSL routines.
15720 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15722 * Add a configuration entry for Sony News 4.
15724 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15726 * Don't set the two most significant bits to one when generating a
15727 random number < q in the DSA library.
15731 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15732 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15733 the underlying transport is blocking) if a handshake took place.
15734 (The default behaviour is needed by applications such as s_client
15735 and s_server that use select() to determine when to use SSL_read;
15736 but for applications that know in advance when to expect data, it
15737 just makes things more complicated.)
15741 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15746 * Add a few more EBCDIC conditionals that make `req` and `x509`
15747 work better on such systems.
15749 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15751 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15752 Update PKCS12_parse() so it copies the friendlyName and the
15753 keyid to the certificates aux info.
15757 * Fix bug in PKCS7_verify() which caused an infinite loop
15758 if there was more than one signature.
15760 *Sven Uszpelkat <su@celocom.de>*
15762 * Major change in util/mkdef.pl to include extra information
15763 about each symbol, as well as presenting variables as well
15764 as functions. This change means that there's n more need
15765 to rebuild the .num files when some algorithms are excluded.
15769 * Allow the verify time to be set by an application,
15770 rather than always using the current time.
15774 * Phase 2 verify code reorganisation. The certificate
15775 verify code now looks up an issuer certificate by a
15776 number of criteria: subject name, authority key id
15777 and key usage. It also verifies self signed certificates
15778 by the same criteria. The main comparison function is
15779 X509_check_issued() which performs these checks.
15781 Lot of changes were necessary in order to support this
15782 without completely rewriting the lookup code.
15784 Authority and subject key identifier are now cached.
15786 The LHASH 'certs' is X509_STORE has now been replaced
15787 by a STACK_OF(X509_OBJECT). This is mainly because an
15788 LHASH can't store or retrieve multiple objects with
15789 the same hash value.
15791 As a result various functions (which were all internal
15792 use only) have changed to handle the new X509_STORE
15793 structure. This will break anything that messed round
15794 with X509_STORE internally.
15796 The functions X509_STORE_add_cert() now checks for an
15797 exact match, rather than just subject name.
15799 The X509_STORE API doesn't directly support the retrieval
15800 of multiple certificates matching a given criteria, however
15801 this can be worked round by performing a lookup first
15802 (which will fill the cache with candidate certificates)
15803 and then examining the cache for matches. This is probably
15804 the best we can do without throwing out X509_LOOKUP
15805 entirely (maybe later...).
15807 The X509_VERIFY_CTX structure has been enhanced considerably.
15809 All certificate lookup operations now go via a get_issuer()
15810 callback. Although this currently uses an X509_STORE it
15811 can be replaced by custom lookups. This is a simple way
15812 to bypass the X509_STORE hackery necessary to make this
15813 work and makes it possible to use more efficient techniques
15814 in future. A very simple version which uses a simple
15815 STACK for its trusted certificate store is also provided
15816 using X509_STORE_CTX_trusted_stack().
15818 The verify_cb() and verify() callbacks now have equivalents
15819 in the X509_STORE_CTX structure.
15821 X509_STORE_CTX also has a 'flags' field which can be used
15822 to customise the verify behaviour.
15826 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15827 excludes S/MIME capabilities.
15831 * When a certificate request is read in keep a copy of the
15832 original encoding of the signed data and use it when outputting
15833 again. Signatures then use the original encoding rather than
15834 a decoded, encoded version which may cause problems if the
15835 request is improperly encoded.
15839 * For consistency with other BIO_puts implementations, call
15840 buffer_write(b, ...) directly in buffer_puts instead of calling
15843 In BIO_puts, increment b->num_write as in BIO_write.
15845 *Peter.Sylvester@EdelWeb.fr*
15847 * Fix BN_mul_word for the case where the word is 0. (We have to use
15848 BN_zero, we may not return a BIGNUM with an array consisting of
15849 words set to zero.)
15853 * Avoid calling abort() from within the library when problems are
15854 detected, except if preprocessor symbols have been defined
15855 (such as REF_CHECK, BN_DEBUG etc.).
15859 * New openssl application 'rsautl'. This utility can be
15860 used for low-level RSA operations. DER public key
15861 BIO/fp routines also added.
15865 * New Configure entry and patches for compiling on QNX 4.
15867 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15869 * A demo state-machine implementation was sponsored by
15870 Nuron (<http://www.nuron.com/>) and is now available in
15871 demos/state_machine.
15875 * New options added to the 'dgst' utility for signature
15876 generation and verification.
15880 * Unrecognized PKCS#7 content types are now handled via a
15881 catch all ASN1_TYPE structure. This allows unsupported
15882 types to be stored as a "blob" and an application can
15883 encode and decode it manually.
15887 * Fix various signed/unsigned issues to make a_strex.c
15888 compile under VC++.
15890 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15892 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15893 length if passed a buffer. ASN1_INTEGER_to_BN failed
15894 if passed a NULL BN and its argument was negative.
15896 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15898 * Modification to PKCS#7 encoding routines to output definite
15899 length encoding. Since currently the whole structures are in
15900 memory there's not real point in using indefinite length
15901 constructed encoding. However if OpenSSL is compiled with
15902 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15906 * Added BIO_vprintf() and BIO_vsnprintf().
15910 * Added more prefixes to parse for in the strings written
15911 through a logging bio, to cover all the levels that are available
15912 through syslog. The prefixes are now:
15914 PANIC, EMERG, EMR => LOG_EMERG
15915 ALERT, ALR => LOG_ALERT
15916 CRIT, CRI => LOG_CRIT
15917 ERROR, ERR => LOG_ERR
15918 WARNING, WARN, WAR => LOG_WARNING
15919 NOTICE, NOTE, NOT => LOG_NOTICE
15920 INFO, INF => LOG_INFO
15921 DEBUG, DBG => LOG_DEBUG
15923 and as before, if none of those prefixes are present at the
15924 beginning of the string, LOG_ERR is chosen.
15926 On Win32, the `LOG_*` levels are mapped according to this:
15928 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15929 LOG_WARNING => EVENTLOG_WARNING_TYPE
15930 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15934 * Made it possible to reconfigure with just the configuration
15935 argument "reconf" or "reconfigure". The command line arguments
15936 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15937 and are retrieved from there when reconfiguring.
15943 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15945 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15949 * The obj_dat.pl script was messing up the sorting of object
15950 names. The reason was that it compared the quoted version
15951 of strings as a result "OCSP" > "OCSP Signing" because
15952 " > SPACE. Changed script to store unquoted versions of
15953 names and add quotes on output. It was also omitting some
15954 names from the lookup table if they were given a default
15955 value (that is if SN is missing it is given the same
15956 value as LN and vice versa), these are now added on the
15957 grounds that if an object has a name we should be able to
15958 look it up. Finally added warning output when duplicate
15959 short or long names are found.
15963 * Changes needed for Tandem NSK.
15965 *Scott Uroff <scott@xypro.com>*
15967 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15968 RSA_padding_check_SSLv23(), special padding was never detected
15969 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15970 version rollback attacks was not effective.
15972 In s23_clnt.c, don't use special rollback-attack detection padding
15973 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15974 client; similarly, in s23_srvr.c, don't do the rollback check if
15975 SSL 2.0 is the only protocol enabled in the server.
15979 * Make it possible to get hexdumps of unprintable data with 'openssl
15980 asn1parse'. By implication, the functions ASN1_parse_dump() and
15981 BIO_dump_indent() are added.
15985 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15986 these print out strings and name structures based on various
15987 flags including RFC2253 support and proper handling of
15988 multibyte characters. Added options to the 'x509' utility
15989 to allow the various flags to be set.
15993 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15994 Also change the functions X509_cmp_current_time() and
15995 X509_gmtime_adj() work with an ASN1_TIME structure,
15996 this will enable certificates using GeneralizedTime in validity
15997 dates to be checked.
16001 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16002 negative public key encodings) on by default,
16003 NO_NEG_PUBKEY_BUG can be set to disable it.
16007 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16008 content octets. An i2c_ASN1_OBJECT is unnecessary because
16009 the encoding can be trivially obtained from the structure.
16013 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16014 not read locks (`CRYPTO_r_[un]lock`).
16018 * A first attempt at creating official support for shared
16019 libraries through configuration. I've kept it so the
16020 default is static libraries only, and the OpenSSL programs
16021 are always statically linked for now, but there are
16022 preparations for dynamic linking in place.
16023 This has been tested on Linux and Tru64.
16027 * Randomness polling function for Win9x, as described in:
16028 Peter Gutmann, Software Generation of Practically Strong
16033 * Fix so PRNG is seeded in req if using an already existing
16038 * New options to smime application. -inform and -outform
16039 allow alternative formats for the S/MIME message including
16040 PEM and DER. The -content option allows the content to be
16041 specified separately. This should allow things like Netscape
16042 form signing output easier to verify.
16046 * Fix the ASN1 encoding of tags using the 'long form'.
16050 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16051 STRING types. These convert content octets to and from the
16052 underlying type. The actual tag and length octets are
16053 already assumed to have been read in and checked. These
16054 are needed because all other string types have virtually
16055 identical handling apart from the tag. By having versions
16056 of the ASN1 functions that just operate on content octets
16057 IMPLICIT tagging can be handled properly. It also allows
16058 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16059 and ASN1_INTEGER are identical apart from the tag.
16063 * Change the handling of OID objects as follows:
16065 - New object identifiers are inserted in objects.txt, following
16066 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16067 - objects.pl is used to process obj_mac.num and create a new
16069 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16072 This is currently kind of a hack, and the perl code in objects.pl
16073 isn't very elegant, but it works as I intended. The simplest way
16074 to check that it worked correctly is to look in obj_dat.h and
16075 check the array nid_objs and make sure the objects haven't moved
16076 around (this is important!). Additions are OK, as well as
16077 consistent name changes.
16081 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16085 * Addition of the command line parameter '-rand file' to 'openssl req'.
16086 The given file adds to whatever has already been seeded into the
16087 random pool through the RANDFILE configuration file option or
16088 environment variable, or the default random state file.
16092 * mkstack.pl now sorts each macro group into lexical order.
16093 Previously the output order depended on the order the files
16094 appeared in the directory, resulting in needless rewriting
16099 * Patches to make OpenSSL compile under Win32 again. Mostly
16100 work arounds for the VC++ problem that it treats func() as
16101 func(void). Also stripped out the parts of mkdef.pl that
16102 added extra typesafe functions: these no longer exist.
16106 * Reorganisation of the stack code. The macros are now all
16107 collected in safestack.h . Each macro is defined in terms of
16108 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16109 DEBUG_SAFESTACK is now handled in terms of function casts,
16110 this has the advantage of retaining type safety without the
16111 use of additional functions. If DEBUG_SAFESTACK is not defined
16112 then the non typesafe macros are used instead. Also modified the
16113 mkstack.pl script to handle the new form. Needs testing to see
16114 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16115 the default if no major problems. Similar behaviour for ASN1_SET_OF
16116 and PKCS12_STACK_OF.
16120 * When some versions of IIS use the 'NET' form of private key the
16121 key derivation algorithm is different. Normally MD5(password) is
16122 used as a 128 bit RC4 key. In the modified case
16123 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16124 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16125 as the old Netscape_RSA functions except they have an additional
16126 'sgckey' parameter which uses the modified algorithm. Also added
16127 an -sgckey command line option to the rsa utility. Thanks to
16128 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16129 algorithm to openssl-dev.
16133 * The evp_local.h macros were using 'c.##kname' which resulted in
16134 invalid expansion on some systems (SCO 5.0.5 for example).
16135 Corrected to 'c.kname'.
16137 *Phillip Porch <root@theporch.com>*
16139 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16140 a STACK of email addresses from a certificate or request, these look
16141 in the subject name and the subject alternative name extensions and
16142 omit any duplicate addresses.
16146 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16147 This makes DSA verification about 2 % faster.
16151 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16152 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16153 plus overhead for 1024 bit moduli).
16154 This makes exponentiations about 0.5 % faster for 1024 bit
16155 exponents (as measured by "openssl speed rsa2048").
16159 * Rename memory handling macros to avoid conflicts with other
16161 Malloc => OPENSSL_malloc
16162 Malloc_locked => OPENSSL_malloc_locked
16163 Realloc => OPENSSL_realloc
16164 Free => OPENSSL_free
16168 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16169 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16173 * CygWin32 support.
16175 *John Jarvie <jjarvie@newsguy.com>*
16177 * The type-safe stack code has been rejigged. It is now only compiled
16178 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16179 by default all type-specific stack functions are "#define"d back to
16180 standard stack functions. This results in more streamlined output
16181 but retains the type-safety checking possibilities of the original
16186 * The STACK code has been cleaned up, and certain type declarations
16187 that didn't make a lot of sense have been brought in line. This has
16188 also involved a cleanup of sorts in safestack.h to more correctly
16189 map type-safe stack functions onto their plain stack counterparts.
16190 This work has also resulted in a variety of "const"ifications of
16191 lots of the code, especially `_cmp` operations which should normally
16192 be prototyped with "const" parameters anyway.
16196 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16197 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16198 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16199 where all of 'md' is used each time the PRNG is used, but 'state'
16200 is used only indexed by a cyclic counter. As entropy may not be
16201 well distributed from the beginning, 'md' is important as a
16202 chaining variable. However, the output function chains only half
16203 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16204 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16205 in all of 'state' being rewritten, with the new values depending
16206 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16210 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16211 the handshake is continued after ssl_verify_cert_chain();
16212 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16213 can lead to 'unexplainable' connection aborts later.
16215 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16217 * Major EVP API cipher revision.
16218 Add hooks for extra EVP features. This allows various cipher
16219 parameters to be set in the EVP interface. Support added for variable
16220 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16221 setting of RC2 and RC5 parameters.
16223 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16226 Remove lots of duplicated code from the EVP library. For example *every*
16227 cipher init() function handles the 'iv' in the same way according to the
16228 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16229 for CFB and OFB modes they zero ctx->num.
16231 New functionality allows removal of S/MIME code RC2 hack.
16233 Most of the routines have the same form and so can be declared in terms
16236 By shifting this to the top level EVP_CipherInit() it can be removed from
16237 all individual ciphers. If the cipher wants to handle IVs or keys
16238 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16241 Change lots of functions like EVP_EncryptUpdate() to now return a
16242 value: although software versions of the algorithms cannot fail
16243 any installed hardware versions can.
16247 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16248 this option is set, tolerate broken clients that send the negotiated
16249 protocol version number instead of the requested protocol version
16254 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16255 i.e. non-zero for export ciphersuites, zero otherwise.
16256 Previous versions had this flag inverted, inconsistent with
16257 rsa_tmp_cb (..._TMP_RSA_CB).
16259 *Bodo Moeller; problem reported by Amit Chopra*
16261 * Add missing DSA library text string. Work around for some IIS
16262 key files with invalid SEQUENCE encoding.
16266 * Add a document (doc/standards.txt) that list all kinds of standards
16267 and so on that are implemented in OpenSSL.
16271 * Enhance c_rehash script. Old version would mishandle certificates
16272 with the same subject name hash and wouldn't handle CRLs at all.
16273 Added -fingerprint option to crl utility, to support new c_rehash
16278 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16282 * Fix for SSL server purpose checking. Server checking was
16283 rejecting certificates which had extended key usage present
16284 but no ssl client purpose.
16286 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16288 * Make PKCS#12 code work with no password. The PKCS#12 spec
16289 is a little unclear about how a blank password is handled.
16290 Since the password in encoded as a BMPString with terminating
16291 double NULL a zero length password would end up as just the
16292 double NULL. However no password at all is different and is
16293 handled differently in the PKCS#12 key generation code. NS
16294 treats a blank password as zero length. MSIE treats it as no
16295 password on export: but it will try both on import. We now do
16296 the same: PKCS12_parse() tries zero length and no password if
16297 the password is set to "" or NULL (NULL is now a valid password:
16298 it wasn't before) as does the pkcs12 application.
16302 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16303 perror when PEM_read_bio_X509_REQ fails, the error message must
16304 be obtained from the error queue.
16308 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16309 it in ERR_remove_state if appropriate, and change ERR_get_state
16310 accordingly to avoid race conditions (this is necessary because
16311 thread_hash is no longer constant once set).
16315 * Bugfix for linux-elf makefile.one.
16319 * RSA_get_default_method() will now cause a default
16320 RSA_METHOD to be chosen if one doesn't exist already.
16321 Previously this was only set during a call to RSA_new()
16322 or RSA_new_method(NULL) meaning it was possible for
16323 RSA_get_default_method() to return NULL.
16327 * Added native name translation to the existing DSO code
16328 that will convert (if the flag to do so is set) filenames
16329 that are sufficiently small and have no path information
16330 into a canonical native form. Eg. "blah" converted to
16331 "libblah.so" or "blah.dll" etc.
16335 * New function ERR_error_string_n(e, buf, len) which is like
16336 ERR_error_string(e, buf), but writes at most 'len' bytes
16337 including the 0 terminator. For ERR_error_string_n, 'buf'
16340 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16342 * CONF library reworked to become more general. A new CONF
16343 configuration file reader "class" is implemented as well as a
16344 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16345 old `CONF_*` functions are still there, but are reimplemented to
16346 work in terms of the new functions. Also, a set of functions
16347 to handle the internal storage of the configuration data is
16348 provided to make it easier to write new configuration file
16349 reader "classes" (I can definitely see something reading a
16350 configuration file in XML format, for example), called `_CONF_*`,
16351 or "the configuration storage API"...
16353 The new configuration file reading functions are:
16355 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16356 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16358 NCONF_default, NCONF_WIN32
16360 NCONF_dump_fp, NCONF_dump_bio
16362 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16363 NCONF_new creates a new CONF object. This works in the same way
16364 as other interfaces in OpenSSL, like the BIO interface.
16365 `NCONF_dump_*` dump the internal storage of the configuration file,
16366 which is useful for debugging. All other functions take the same
16367 arguments as the old `CONF_*` functions with the exception of the
16368 first that must be a `CONF *` instead of a `LHASH *`.
16370 To make it easier to use the new classes with the old `CONF_*` functions,
16371 the function CONF_set_default_method is provided.
16375 * Add '-tls1' option to 'openssl ciphers', which was already
16376 mentioned in the documentation but had not been implemented.
16377 (This option is not yet really useful because even the additional
16378 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16382 * Initial DSO code added into libcrypto for letting OpenSSL (and
16383 OpenSSL-based applications) load shared libraries and bind to
16384 them in a portable way.
16386 *Geoff Thorpe, with contributions from Richard Levitte*
16388 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16390 * Make sure _lrotl and _lrotr are only used with MSVC.
16392 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16393 (the default implementation of RAND_status).
16395 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16396 to '-clrext' (= clear extensions), as intended and documented.
16397 *Bodo Moeller; inconsistency pointed out by Michael Attili
16398 <attili@amaxo.com>*
16400 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16401 was larger than the MD block size.
16403 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16405 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16406 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16407 using the passed key: if the passed key was a private key the result
16408 of X509_print(), for example, would be to print out all the private key
16413 * des_quad_cksum() byte order bug fix.
16414 *Ulf Möller, using the problem description in krb4-0.9.7, where
16415 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16417 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16420 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16422 * For easily testing in shell scripts whether some command
16423 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16424 returns with exit code 0 iff no command of the given name is available.
16425 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16426 the output goes to stdout and nothing is printed to stderr.
16427 Additional arguments are always ignored.
16429 Since for each cipher there is a command of the same name,
16430 the 'no-cipher' compilation switches can be tested this way.
16432 ('openssl no-XXX' is not able to detect pseudo-commands such
16433 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16437 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16441 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16442 is set; it will be thrown away anyway because each handshake creates
16444 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16445 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16446 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16447 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16451 * New s_client option -ign_eof: EOF at stdin is ignored, and
16452 'Q' and 'R' lose their special meanings (quit/renegotiate).
16453 This is part of what -quiet does; unlike -quiet, -ign_eof
16454 does not suppress any output.
16458 * Add compatibility options to the purpose and trust code. The
16459 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16460 accepts a certificate or CA, this was the previous behaviour,
16461 with all the associated security issues.
16463 X509_TRUST_COMPAT is the old trust behaviour: only and
16464 automatically trust self signed roots in certificate store. A
16465 new trust setting X509_TRUST_DEFAULT is used to specify that
16466 a purpose has no associated trust setting and it should instead
16467 use the value in the default purpose.
16471 * Fix the PKCS#8 DSA private key code so it decodes keys again
16472 and fix a memory leak.
16476 * In util/mkerr.pl (which implements 'make errors'), preserve
16477 reason strings from the previous version of the .c file, as
16478 the default to have only downcase letters (and digits) in
16479 automatically generated reasons codes is not always appropriate.
16483 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16484 using strerror. Previously, ERR_reason_error_string() returned
16485 library names as reason strings for SYSerr; but SYSerr is a special
16486 case where small numbers are errno values, not library numbers.
16490 * Add '-dsaparam' option to 'openssl dhparam' application. This
16491 converts DSA parameters into DH parameters. (When creating parameters,
16492 DSA_generate_parameters is used.)
16496 * Include 'length' (recommended exponent length) in C code generated
16497 by 'openssl dhparam -C'.
16501 * The second argument to set_label in perlasm was already being used
16502 so couldn't be used as a "file scope" flag. Moved to third argument
16507 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16508 instead of RAND_bytes for encryption IVs and salts.
16512 * Include RAND_status() into RAND_METHOD instead of implementing
16513 it only for md_rand.c Otherwise replacing the PRNG by calling
16514 RAND_set_rand_method would be impossible.
16518 * Don't let DSA_generate_key() enter an infinite loop if the random
16519 number generation fails.
16523 * New 'rand' application for creating pseudo-random output.
16527 * Added configuration support for Linux/IA64
16529 *Rolf Haberrecker <rolf@suse.de>*
16531 * Assembler module support for Mingw32.
16535 * Shared library support for HPUX (in shlib/).
16537 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16539 * Shared library support for Solaris gcc.
16541 *Lutz Behnke <behnke@trustcenter.de>*
16543 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16545 * PKCS7_encrypt() was adding text MIME headers twice because they
16546 were added manually and by SMIME_crlf_copy().
16550 * In bntest.c don't call BN_rand with zero bits argument.
16552 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16554 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16555 case was implemented. This caused BN_div_recp() to fail occasionally.
16559 * Add an optional second argument to the set_label() in the perl
16560 assembly language builder. If this argument exists and is set
16561 to 1 it signals that the assembler should use a symbol whose
16562 scope is the entire file, not just the current function. This
16563 is needed with MASM which uses the format label:: for this scope.
16565 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16567 * Change the ASN1 types so they are typedefs by default. Before
16568 almost all types were #define'd to ASN1_STRING which was causing
16569 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16574 * Change names of new functions to the new get1/get0 naming
16575 convention: After 'get1', the caller owns a reference count
16576 and has to call `..._free`; 'get0' returns a pointer to some
16577 data structure without incrementing reference counters.
16578 (Some of the existing 'get' functions increment a reference
16579 counter, some don't.)
16580 Similarly, 'set1' and 'add1' functions increase reference
16581 counters or duplicate objects.
16585 * Allow for the possibility of temp RSA key generation failure:
16586 the code used to assume it always worked and crashed on failure.
16590 * Fix potential buffer overrun problem in BIO_printf().
16591 *Ulf Möller, using public domain code by Patrick Powell; problem
16592 pointed out by David Sacerdote <das33@cornell.edu>*
16594 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16595 RAND_egd() and RAND_status(). In the command line application,
16596 the EGD socket can be specified like a seed file using RANDFILE
16601 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16602 Some CAs (e.g. Verisign) distribute certificates in this form.
16606 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16607 list to exclude them. This means that no special compilation option
16608 is needed to use anonymous DH: it just needs to be included in the
16613 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16614 EVP_MD_type. The old functionality is available in a new macro called
16615 EVP_MD_md(). Change code that uses it and update docs.
16619 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16620 where the `void *` argument is replaced by a function pointer argument.
16621 Previously `void *` was abused to point to functions, which works on
16622 many platforms, but is not correct. As these functions are usually
16623 called by macros defined in OpenSSL header files, most source code
16624 should work without changes.
16628 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16629 sections with information on -D... compiler switches used for
16630 compiling the library so that applications can see them. To enable
16631 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16632 must be defined. E.g.,
16633 #define OPENSSL_ALGORITHM_DEFINES
16634 #include <openssl/opensslconf.h>
16635 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16637 *Richard Levitte, Ulf and Bodo Möller*
16639 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16644 * Change the 'other' type in certificate aux info to a STACK_OF
16645 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16646 the required ASN1 format: arbitrary types determined by an OID.
16650 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16651 argument to 'req'. This is not because the function is newer or
16652 better than others it just uses the work 'NEW' in the certificate
16653 request header lines. Some software needs this.
16657 * Reorganise password command line arguments: now passwords can be
16658 obtained from various sources. Delete the PEM_cb function and make
16659 it the default behaviour: i.e. if the callback is NULL and the
16660 usrdata argument is not NULL interpret it as a null terminated pass
16661 phrase. If usrdata and the callback are NULL then the pass phrase
16662 is prompted for as usual.
16666 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16667 the support is automatically enabled. The resulting binaries will
16668 autodetect the card and use it if present.
16670 *Ben Laurie and Compaq Inc.*
16672 * Work around for Netscape hang bug. This sends certificate request
16673 and server done in one record. Since this is perfectly legal in the
16674 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16675 the bugs/SSLv3 entry for more info.
16679 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16683 * Add -rand argument to smime and pkcs12 applications and read/write
16688 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16692 * Add command line password options to the remaining applications.
16696 * Bug fix for BN_div_recp() for numerators with an even number of
16701 * More tests in bntest.c, and changed test_bn output.
16705 * ./config recognizes MacOS X now.
16709 * Bug fix for BN_div() when the first words of num and divisor are
16710 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16714 * Add support for various broken PKCS#8 formats, and command line
16715 options to produce them.
16719 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16720 get temporary BIGNUMs from a BN_CTX.
16724 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16729 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16730 include a #define from the old name to the new. The original intent
16731 was that statically linked binaries could for example just call
16732 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16733 link with digests. This never worked because SSLeay_add_all_digests()
16734 and SSLeay_add_all_ciphers() were in the same source file so calling
16735 one would link with the other. They are now in separate source files.
16739 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16743 * Use a less unusual form of the Miller-Rabin primality test (it used
16744 a binary algorithm for exponentiation integrated into the Miller-Rabin
16745 loop, our standard modexp algorithms are faster).
16749 * Support for the EBCDIC character set completed.
16751 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16753 * Source code cleanups: use const where appropriate, eliminate casts,
16754 use `void *` instead of `char *` in lhash.
16758 * Bugfix: ssl3_send_server_key_exchange was not restartable
16759 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16760 this the server could overwrite ephemeral keys that the client
16765 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16766 using 50 iterations of the Rabin-Miller test.
16768 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16769 iterations of the Rabin-Miller test as required by the appendix
16770 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16771 As BN_is_prime_fasttest includes trial division, DSA parameter
16772 generation becomes much faster.
16774 This implies a change for the callback functions in DSA_is_prime
16775 and DSA_generate_parameters: The callback function is called once
16776 for each positive witness in the Rabin-Miller test, not just
16777 occasionally in the inner loop; and the parameters to the
16778 callback function now provide an iteration count for the outer
16779 loop rather than for the current invocation of the inner loop.
16780 DSA_generate_parameters additionally can call the callback
16781 function with an 'iteration count' of -1, meaning that a
16782 candidate has passed the trial division test (when q is generated
16783 from an application-provided seed, trial division is skipped).
16787 * New function BN_is_prime_fasttest that optionally does trial
16788 division before starting the Rabin-Miller test and has
16789 an additional BN_CTX * argument (whereas BN_is_prime always
16790 has to allocate at least one BN_CTX).
16791 'callback(1, -1, cb_arg)' is called when a number has passed the
16792 trial division stage.
16796 * Fix for bug in CRL encoding. The validity dates weren't being handled
16801 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16805 * New function BN_pseudo_rand().
16809 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16810 bignum version of BN_from_montgomery() with the working code from
16811 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16816 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16817 made it impossible to use the same SSL_SESSION data structure in
16818 SSL2 clients in multiple threads.
16822 * The return value of RAND_load_file() no longer counts bytes obtained
16823 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16824 to seed the PRNG (previously an explicit byte count was required).
16826 *Ulf Möller, Bodo Möller*
16828 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16829 used `char *` instead of `void *` and had casts all over the place.
16833 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16837 * Retain source code compatibility for BN_prime_checks macro:
16838 BN_is_prime(..., BN_prime_checks, ...) now uses
16839 BN_prime_checks_for_size to determine the appropriate number of
16840 Rabin-Miller iterations.
16844 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16845 DH_CHECK_P_NOT_SAFE_PRIME.
16846 (Check if this is true? OpenPGP calls them "strong".)
16850 * Merge the functionality of "dh" and "gendh" programs into a new program
16851 "dhparam". The old programs are retained for now but will handle DH keys
16852 (instead of parameters) in future.
16856 * Make the ciphers, s_server and s_client programs check the return values
16857 when a new cipher list is set.
16861 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16862 ciphers. Before when the 56bit ciphers were enabled the sorting was
16865 The syntax for the cipher sorting has been extended to support sorting by
16866 cipher-strength (using the strength_bits hard coded in the tables).
16867 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16869 Fix a bug in the cipher-command parser: when supplying a cipher command
16870 string with an "undefined" symbol (neither command nor alphanumeric
16871 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16872 an error is flagged.
16874 Due to the strength-sorting extension, the code of the
16875 ssl_create_cipher_list() function was completely rearranged. I hope that
16876 the readability was also increased :-)
16878 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16880 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16881 for the first serial number and places 2 in the serial number file. This
16882 avoids problems when the root CA is created with serial number zero and
16883 the first user certificate has the same issuer name and serial number
16888 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16889 the new code. Add documentation for this stuff.
16893 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16894 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16895 structures and behave in an analogous way to the X509v3 functions:
16896 they shouldn't be called directly but wrapper functions should be used
16899 So we also now have some wrapper functions that call the X509at functions
16900 when passed certificate requests. (TO DO: similar things can be done with
16901 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16902 things. Some of these need some d2i or i2d and print functionality
16903 because they handle more complex structures.)
16907 * Add missing #ifndefs that caused missing symbols when building libssl
16908 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16909 NO_RSA in `ssl/s2*.c`.
16911 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16913 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16914 has a return value which indicates the quality of the random data
16915 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16916 error queue. New function RAND_pseudo_bytes() generates output that is
16917 guaranteed to be unique but not unpredictable. RAND_add is like
16918 RAND_seed, but takes an extra argument for an entropy estimate
16919 (RAND_seed always assumes full entropy).
16923 * Do more iterations of Rabin-Miller probable prime test (specifically,
16924 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16925 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16926 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16927 false-positive rate of at most 2^-80 for random input.
16931 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16935 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16936 in the 0.9.5 release), this returns the chain
16937 from an X509_CTX structure with a dup of the stack and all
16938 the X509 reference counts upped: so the stack will exist
16939 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16942 Also make SSL_SESSION_print() print out the verify return
16947 * Add manpage for the pkcs12 command. Also change the default
16948 behaviour so MAC iteration counts are used unless the new
16949 -nomaciter option is used. This improves file security and
16950 only older versions of MSIE (4.0 for example) need it.
16954 * Honor the no-xxx Configure options when creating .DEF files.
16958 * Add PKCS#10 attributes to field table: challengePassword,
16959 unstructuredName and unstructuredAddress. These are taken from
16960 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16961 international characters are used.
16963 More changes to X509_ATTRIBUTE code: allow the setting of types
16964 based on strings. Remove the 'loc' parameter when adding
16965 attributes because these will be a SET OF encoding which is sorted
16970 * Initial changes to the 'req' utility to allow request generation
16971 automation. This will allow an application to just generate a template
16972 file containing all the field values and have req construct the
16975 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16976 used all over the place including certificate requests and PKCS#7
16977 structures. They are currently handled manually where necessary with
16978 some primitive wrappers for PKCS#7. The new functions behave in a
16979 manner analogous to the X509 extension functions: they allow
16980 attributes to be looked up by NID and added.
16982 Later something similar to the X509V3 code would be desirable to
16983 automatically handle the encoding, decoding and printing of the
16984 more complex types. The string types like challengePassword can
16985 be handled by the string table functions.
16987 Also modified the multi byte string table handling. Now there is
16988 a 'global mask' which masks out certain types. The table itself
16989 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16990 is useful when for example there is only one permissible type
16991 (as in countryName) and using the mask might result in no valid
16996 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16997 SSL_get_peer_finished to allow applications to obtain the latest
16998 Finished messages sent to the peer or expected from the peer,
16999 respectively. (SSL_get_peer_finished is usually the Finished message
17000 actually received from the peer, otherwise the protocol will be aborted.)
17002 As the Finished message are message digests of the complete handshake
17003 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17004 be used for external authentication procedures when the authentication
17005 provided by SSL/TLS is not desired or is not enough.
17009 * Enhanced support for Alpha Linux is added. Now ./config checks if
17010 the host supports BWX extension and if Compaq C is present on the
17011 $PATH. Just exploiting of the BWX extension results in 20-30%
17012 performance kick for some algorithms, e.g. DES and RC4 to mention
17013 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17018 * Add support for MS "fast SGC". This is arguably a violation of the
17019 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17020 weak crypto and after checking the certificate is SGC a second one
17021 with strong crypto. MS SGC stops the first handshake after receiving
17022 the server certificate message and sends a second client hello. Since
17023 a server will typically do all the time consuming operations before
17024 expecting any further messages from the client (server key exchange
17025 is the most expensive) there is little difference between the two.
17027 To get OpenSSL to support MS SGC we have to permit a second client
17028 hello message after we have sent server done. In addition we have to
17029 reset the MAC if we do get this second client hello.
17033 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17034 if a DER encoded private key is RSA or DSA traditional format. Changed
17035 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17036 format DER encoded private key. Newer code should use PKCS#8 format which
17037 has the key type encoded in the ASN1 structure. Added DER private key
17038 support to pkcs8 application.
17042 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17043 ciphersuites has been selected (as required by the SSL 3/TLS 1
17044 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17045 is set, we interpret this as a request to violate the specification
17046 (the worst that can happen is a handshake failure, and 'correct'
17047 behaviour would result in a handshake failure anyway).
17051 * In SSL_CTX_add_session, take into account that there might be multiple
17052 SSL_SESSION structures with the same session ID (e.g. when two threads
17053 concurrently obtain them from an external cache).
17054 The internal cache can handle only one SSL_SESSION with a given ID,
17055 so if there's a conflict, we now throw out the old one to achieve
17060 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17061 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17062 some routines that use cipher OIDs: some ciphers do not have OIDs
17063 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17068 * Simplify the trust setting structure and code. Now we just have
17069 two sequences of OIDs for trusted and rejected settings. These will
17070 typically have values the same as the extended key usage extension
17071 and any application specific purposes.
17073 The trust checking code now has a default behaviour: it will just
17074 check for an object with the same NID as the passed id. Functions can
17075 be provided to override either the default behaviour or the behaviour
17076 for a given id. SSL client, server and email already have functions
17077 in place for compatibility: they check the NID and also return "trusted"
17078 if the certificate is self signed.
17082 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17083 traditional format into an EVP_PKEY structure.
17087 * Add a password callback function PEM_cb() which either prompts for
17088 a password if usr_data is NULL or otherwise assumes it is a null
17089 terminated password. Allow passwords to be passed on command line
17090 environment or config files in a few more utilities.
17094 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17095 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17096 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17097 Update documentation.
17101 * Support for ASN1 "NULL" type. This could be handled before by using
17102 ASN1_TYPE but there wasn't any function that would try to read a NULL
17103 and produce an error if it couldn't. For compatibility we also have
17104 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17105 don't allocate anything because they don't need to.
17109 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17112 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17114 * Rebuild of the memory allocation routines used by OpenSSL code and
17115 possibly others as well. The purpose is to make an interface that
17116 provide hooks so anyone can build a separate set of allocation and
17117 deallocation routines to be used by OpenSSL, for example memory
17118 pool implementations, or something else, which was previously hard
17119 since Malloc(), Realloc() and Free() were defined as macros having
17120 the values malloc, realloc and free, respectively (except for Win32
17121 compilations). The same is provided for memory debugging code.
17122 OpenSSL already comes with functionality to find memory leaks, but
17123 this gives people a chance to debug other memory problems.
17125 With these changes, a new set of functions and macros have appeared:
17127 CRYPTO_set_mem_debug_functions() [F]
17128 CRYPTO_get_mem_debug_functions() [F]
17129 CRYPTO_dbg_set_options() [F]
17130 CRYPTO_dbg_get_options() [F]
17131 CRYPTO_malloc_debug_init() [M]
17133 The memory debug functions are NULL by default, unless the library
17134 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17135 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17136 gives the standard debugging functions that come with OpenSSL) or
17137 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17138 provided by the library user) must be used. When the standard
17139 debugging functions are used, CRYPTO_dbg_set_options can be used to
17140 request additional information:
17141 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17142 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17144 Also, things like CRYPTO_set_mem_functions will always give the
17145 expected result (the new set of functions is used for allocation
17146 and deallocation) at all times, regardless of platform and compiler
17149 To finish it up, some functions that were never use in any other
17150 way than through macros have a new API and new semantic:
17152 CRYPTO_dbg_malloc()
17153 CRYPTO_dbg_realloc()
17156 All macros of value have retained their old syntax.
17158 *Richard Levitte and Bodo Moeller*
17160 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17161 ordering of SMIMECapabilities wasn't in "strength order" and there
17162 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17167 * Some ASN1 types with illegal zero length encoding (INTEGER,
17168 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17170 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17172 * Merge in my S/MIME library for OpenSSL. This provides a simple
17173 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17174 functionality to handle multipart/signed properly) and a utility
17175 called 'smime' to call all this stuff. This is based on code I
17176 originally wrote for Celo who have kindly allowed it to be
17177 included in OpenSSL.
17181 * Add variants des_set_key_checked and des_set_key_unchecked of
17182 des_set_key (aka des_key_sched). Global variable des_check_key
17183 decides which of these is called by des_set_key; this way
17184 des_check_key behaves as it always did, but applications and
17185 the library itself, which was buggy for des_check_key == 1,
17186 have a cleaner way to pick the version they need.
17190 * New function PKCS12_newpass() which changes the password of a
17195 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17196 dynamic mix. In both cases the ids can be used as an index into the
17197 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17198 functions so they accept a list of the field values and the
17199 application doesn't need to directly manipulate the X509_TRUST
17204 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17209 * Modify the way the V3 extension code looks up extensions. This now
17210 works in a similar way to the object code: we have some "standard"
17211 extensions in a static table which is searched with OBJ_bsearch()
17212 and the application can add dynamic ones if needed. The file
17213 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17214 updated whenever a new extension is added to the core code and kept
17215 in ext_nid order. There is a simple program 'tabtest.c' which checks
17216 this. New extensions are not added too often so this file can readily
17217 be maintained manually.
17219 There are two big advantages in doing things this way. The extensions
17220 can be looked up immediately and no longer need to be "added" using
17221 X509V3_add_standard_extensions(): this function now does nothing.
17222 Side note: I get *lots* of email saying the extension code doesn't
17223 work because people forget to call this function.
17224 Also no dynamic allocation is done unless new extensions are added:
17225 so if we don't add custom extensions there is no need to call
17226 X509V3_EXT_cleanup().
17230 * Modify enc utility's salting as follows: make salting the default. Add a
17231 magic header, so unsalted files fail gracefully instead of just decrypting
17232 to garbage. This is because not salting is a big security hole, so people
17233 should be discouraged from doing it.
17237 * Fixes and enhancements to the 'x509' utility. It allowed a message
17238 digest to be passed on the command line but it only used this
17239 parameter when signing a certificate. Modified so all relevant
17240 operations are affected by the digest parameter including the
17241 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17242 DSA key was used because it didn't fix the digest.
17246 * Initial certificate chain verify code. Currently tests the untrusted
17247 certificates for consistency with the verify purpose (which is set
17248 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17250 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17251 this is because it will reject chains with invalid extensions whereas
17252 every previous version of OpenSSL and SSLeay made no checks at all.
17254 Trust code: checks the root CA for the relevant trust settings. Trust
17255 settings have an initial value consistent with the verify purpose: e.g.
17256 if the verify purpose is for SSL client use it expects the CA to be
17257 trusted for SSL client use. However the default value can be changed to
17258 permit custom trust settings: one example of this would be to only trust
17259 certificates from a specific "secure" set of CAs.
17261 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17262 which should be used for version portability: especially since the
17263 verify structure is likely to change more often now.
17265 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17266 to set them. If not set then assume SSL clients will verify SSL servers
17269 Two new options to the verify program: -untrusted allows a set of
17270 untrusted certificates to be passed in and -purpose which sets the
17271 intended purpose of the certificate. If a purpose is set then the
17272 new chain verify code is used to check extension consistency.
17276 * Support for the authority information access extension.
17280 * Modify RSA and DSA PEM read routines to transparently handle
17281 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17282 public keys in a format compatible with certificate
17283 SubjectPublicKeyInfo structures. Unfortunately there were already
17284 functions called *_PublicKey_* which used various odd formats so
17285 these are retained for compatibility: however the DSA variants were
17286 never in a public release so they have been deleted. Changed dsa/rsa
17287 utilities to handle the new format: note no releases ever handled public
17288 keys so we should be OK.
17290 The primary motivation for this change is to avoid the same fiasco
17291 that dogs private keys: there are several incompatible private key
17292 formats some of which are standard and some OpenSSL specific and
17293 require various evil hacks to allow partial transparent handling and
17294 even then it doesn't work with DER formats. Given the option anything
17295 other than PKCS#8 should be dumped: but the other formats have to
17296 stay in the name of compatibility.
17298 With public keys and the benefit of hindsight one standard format
17299 is used which works with EVP_PKEY, RSA or DSA structures: though
17300 it clearly returns an error if you try to read the wrong kind of key.
17302 Added a -pubkey option to the 'x509' utility to output the public key.
17303 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17304 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17305 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17306 that do the same as the `EVP_PKEY_assign_*()` except they up the
17307 reference count of the added key (they don't "swallow" the
17312 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17313 CRLs would fail if the file contained no certificates or no CRLs:
17314 added a new function to read in both types and return the number
17315 read: this means that if none are read it will be an error. The
17316 DER versions of the certificate and CRL reader would always fail
17317 because it isn't possible to mix certificates and CRLs in DER format
17318 without choking one or the other routine. Changed this to just read
17319 a certificate: this is the best we can do. Also modified the code
17320 in `apps/verify.c` to take notice of return codes: it was previously
17321 attempting to read in certificates from NULL pointers and ignoring
17322 any errors: this is one reason why the cert and CRL reader seemed
17323 to work. It doesn't check return codes from the default certificate
17324 routines: these may well fail if the certificates aren't installed.
17328 * Code to support otherName option in GeneralName.
17332 * First update to verify code. Change the verify utility
17333 so it warns if it is passed a self signed certificate:
17334 for consistency with the normal behaviour. X509_verify
17335 has been modified to it will now verify a self signed
17336 certificate if *exactly* the same certificate appears
17337 in the store: it was previously impossible to trust a
17338 single self signed certificate. This means that:
17339 openssl verify ss.pem
17340 now gives a warning about a self signed certificate but
17341 openssl verify -CAfile ss.pem ss.pem
17346 * For servers, store verify_result in SSL_SESSION data structure
17347 (and add it to external session representation).
17348 This is needed when client certificate verifications fails,
17349 but an application-provided verification callback (set by
17350 SSL_CTX_set_cert_verify_callback) allows accepting the session
17351 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17352 but returns 1): When the session is reused, we have to set
17353 ssl->verify_result to the appropriate error code to avoid
17356 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17358 * Fix a bug in the new PKCS#7 code: it didn't consider the
17359 case in PKCS7_dataInit() where the signed PKCS7 structure
17360 didn't contain any existing data because it was being created.
17362 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17364 * Add a salt to the key derivation routines in enc.c. This
17365 forms the first 8 bytes of the encrypted file. Also add a
17366 -S option to allow a salt to be input on the command line.
17370 * New function X509_cmp(). Oddly enough there wasn't a function
17371 to compare two certificates. We do this by working out the SHA1
17372 hash and comparing that. X509_cmp() will be needed by the trust
17377 * SSL_get1_session() is like SSL_get_session(), but increments
17378 the reference count in the SSL_SESSION returned.
17380 *Geoff Thorpe <geoff@eu.c2.net>*
17382 * Fix for 'req': it was adding a null to request attributes.
17383 Also change the X509_LOOKUP and X509_INFO code to handle
17384 certificate auxiliary information.
17388 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17393 * Add the possibility to add extra information to the memory leak
17394 detecting output, to form tracebacks, showing from where each
17395 allocation was originated: CRYPTO_push_info("constant string") adds
17396 the string plus current file name and line number to a per-thread
17397 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17398 is like calling CYRPTO_pop_info() until the stack is empty.
17399 Also updated memory leak detection code to be multi-thread-safe.
17403 * Add options -text and -noout to pkcs7 utility and delete the
17404 encryption options which never did anything. Update docs.
17408 * Add options to some of the utilities to allow the pass phrase
17409 to be included on either the command line (not recommended on
17410 OSes like Unix) or read from the environment. Update the
17411 manpages and fix a few bugs.
17415 * Add a few manpages for some of the openssl commands.
17419 * Fix the -revoke option in ca. It was freeing up memory twice,
17420 leaking and not finding already revoked certificates.
17424 * Extensive changes to support certificate auxiliary information.
17425 This involves the use of X509_CERT_AUX structure and X509_AUX
17426 functions. An X509_AUX function such as PEM_read_X509_AUX()
17427 can still read in a certificate file in the usual way but it
17428 will also read in any additional "auxiliary information". By
17429 doing things this way a fair degree of compatibility can be
17430 retained: existing certificates can have this information added
17431 using the new 'x509' options.
17433 Current auxiliary information includes an "alias" and some trust
17434 settings. The trust settings will ultimately be used in enhanced
17435 certificate chain verification routines: currently a certificate
17436 can only be trusted if it is self signed and then it is trusted
17441 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17442 The problem was that one of the replacement routines had not been working
17443 since SSLeay releases. For now the offending routine has been replaced
17444 with non-optimised assembler. Even so, this now gives around 95%
17445 performance improvement for 1024 bit RSA signs.
17449 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17450 handling. Most clients have the effective key size in bits equal to
17451 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17452 A few however don't do this and instead use the size of the decrypted key
17453 to determine the RC2 key length and the AlgorithmIdentifier to determine
17454 the effective key length. In this case the effective key length can still
17455 be 40 bits but the key length can be 168 bits for example. This is fixed
17456 by manually forcing an RC2 key into the EVP_PKEY structure because the
17457 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17458 the key length and effective key length are equal.
17462 * Add a bunch of functions that should simplify the creation of
17463 X509_NAME structures. Now you should be able to do:
17464 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17465 and have it automatically work out the correct field type and fill in
17466 the structures. The more adventurous can try:
17467 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17468 and it will (hopefully) work out the correct multibyte encoding.
17472 * Change the 'req' utility to use the new field handling and multibyte
17473 copy routines. Before the DN field creation was handled in an ad hoc
17474 way in req, ca, and x509 which was rather broken and didn't support
17475 BMPStrings or UTF8Strings. Since some software doesn't implement
17476 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17477 using the dirstring_type option. See the new comment in the default
17478 openssl.cnf for more info.
17482 * Make crypto/rand/md_rand.c more robust:
17483 - Assure unique random numbers after fork().
17484 - Make sure that concurrent threads access the global counter and
17485 md serializably so that we never lose entropy in them
17486 or use exactly the same state in multiple threads.
17487 Access to the large state is not always serializable because
17488 the additional locking could be a performance killer, and
17489 md should be large enough anyway.
17493 * New file `apps/app_rand.c` with commonly needed functionality
17494 for handling the random seed file.
17496 Use the random seed file in some applications that previously did not:
17498 dsaparam -genkey (which also ignored its '-rand' option),
17501 x509 (when signing).
17502 Except on systems with /dev/urandom, it is crucial to have a random
17503 seed file at least for key creation, DSA signing, and for DH exchanges;
17504 for RSA signatures we could do without one.
17506 gendh and gendsa (unlike genrsa) used to read only the first byte
17507 of each file listed in the '-rand' option. The function as previously
17508 found in genrsa is now in app_rand.c and is used by all programs
17509 that support '-rand'.
17513 * In RAND_write_file, use mode 0600 for creating files;
17514 don't just chmod when it may be too late.
17518 * Report an error from X509_STORE_load_locations
17519 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17523 * New function ASN1_mbstring_copy() this copies a string in either
17524 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17525 into an ASN1_STRING type. A mask of permissible types is passed
17526 and it chooses the "minimal" type to use or an error if not type
17531 * Add function equivalents to the various macros in asn1.h. The old
17532 macros are retained with an `M_` prefix. Code inside the library can
17533 use the `M_` macros. External code (including the openssl utility)
17534 should *NOT* in order to be "shared library friendly".
17538 * Add various functions that can check a certificate's extensions
17539 to see if it usable for various purposes such as SSL client,
17540 server or S/MIME and CAs of these types. This is currently
17541 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17542 verification. Also added a -purpose flag to x509 utility to
17543 print out all the purposes.
17547 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17552 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17553 for, obtain and decode and extension and obtain its critical flag.
17554 This allows all the necessary extension code to be handled in a
17555 single function call.
17559 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17560 platforms. See crypto/rc4/rc4_enc.c for further details.
17564 * New -noout option to asn1parse. This causes no output to be produced
17565 its main use is when combined with -strparse and -out to extract data
17566 from a file (which may not be in ASN.1 format).
17570 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17571 when producing the local key id.
17573 *Richard Levitte <levitte@stacken.kth.se>*
17575 * New option -dhparam in s_server. This allows a DH parameter file to be
17576 stated explicitly. If it is not stated then it tries the first server
17577 certificate file. The previous behaviour hard coded the filename
17582 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17583 a public key to be input or output. For example:
17584 openssl rsa -in key.pem -pubout -out pubkey.pem
17585 Also added necessary DSA public key functions to handle this.
17589 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17590 in the message. This was handled by allowing
17591 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17593 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17595 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17596 to the end of the strings whereas this didn't. This would cause problems
17597 if strings read with d2i_ASN1_bytes() were later modified.
17599 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17601 * Fix for base64 decode bug. When a base64 bio reads only one line of
17602 data and it contains EOF it will end up returning an error. This is
17603 caused by input 46 bytes long. The cause is due to the way base64
17604 BIOs find the start of base64 encoded data. They do this by trying a
17605 trial decode on each line until they find one that works. When they
17606 do a flag is set and it starts again knowing it can pass all the
17607 data directly through the decoder. Unfortunately it doesn't reset
17608 the context it uses. This means that if EOF is reached an attempt
17609 is made to pass two EOFs through the context and this causes the
17610 resulting error. This can also cause other problems as well. As is
17611 usual with these problems it takes *ages* to find and the fix is
17612 trivial: move one line.
17614 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17616 * Ugly workaround to get s_client and s_server working under Windows. The
17617 old code wouldn't work because it needed to select() on sockets and the
17618 tty (for keypresses and to see if data could be written). Win32 only
17619 supports select() on sockets so we select() with a 1s timeout on the
17620 sockets and then see if any characters are waiting to be read, if none
17621 are present then we retry, we also assume we can always write data to
17622 the tty. This isn't nice because the code then blocks until we've
17623 received a complete line of data and it is effectively polling the
17624 keyboard at 1s intervals: however it's quite a bit better than not
17625 working at all :-) A dedicated Windows application might handle this
17626 with an event loop for example.
17630 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17631 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17632 will be called when RSA_sign() and RSA_verify() are used. This is useful
17633 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17634 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17635 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17636 This necessitated the support of an extra signature type NID_md5_sha1
17637 for SSL signatures and modifications to the SSL library to use it instead
17638 of calling RSA_public_decrypt() and RSA_private_encrypt().
17642 * Add new -verify -CAfile and -CApath options to the crl program, these
17643 will lookup a CRL issuers certificate and verify the signature in a
17644 similar way to the verify program. Tidy up the crl program so it
17645 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17646 less strict. It will now permit CRL extensions even if it is not
17647 a V2 CRL: this will allow it to tolerate some broken CRLs.
17651 * Initialize all non-automatic variables each time one of the openssl
17652 sub-programs is started (this is necessary as they may be started
17653 multiple times from the "OpenSSL>" prompt).
17655 *Lennart Bang, Bodo Moeller*
17657 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17658 removing all other RSA functionality (this is what NO_RSA does). This
17659 is so (for example) those in the US can disable those operations covered
17660 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17665 * Non-copying interface to BIO pairs.
17666 (still largely untested)
17670 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17671 ASCII string. This was handled independently in various places before.
17675 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17676 UTF8 strings a character at a time.
17680 * Use client_version from client hello to select the protocol
17681 (s23_srvr.c) and for RSA client key exchange verification
17682 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17686 * Add various utility functions to handle SPKACs, these were previously
17687 handled by poking round in the structure internals. Added new function
17688 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17689 print, verify and generate SPKACs. Based on an original idea from
17690 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17694 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17698 * Allow the config file extension section to be overwritten on the
17699 command line. Based on an original idea from Massimiliano Pala
17700 <madwolf@comune.modena.it>. The new option is called -extensions
17701 and can be applied to ca, req and x509. Also -reqexts to override
17702 the request extensions in req and -crlexts to override the crl extensions
17707 * Add new feature to the SPKAC handling in ca. Now you can include
17708 the same field multiple times by preceding it by "XXXX." for example:
17711 this is the same syntax as used in the req config file.
17715 * Allow certificate extensions to be added to certificate requests. These
17716 are specified in a 'req_extensions' option of the req section of the
17717 config file. They can be printed out with the -text option to req but
17718 are otherwise ignored at present.
17722 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17723 data read consists of only the final block it would not decrypted because
17724 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17725 A misplaced 'break' also meant the decrypted final block might not be
17726 copied until the next read.
17730 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17731 a few extra parameters to the DH structure: these will be useful if
17732 for example we want the value of 'q' or implement X9.42 DH.
17736 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17737 provides hooks that allow the default DSA functions or functions on a
17738 "per key" basis to be replaced. This allows hardware acceleration and
17739 hardware key storage to be handled without major modification to the
17740 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17741 associated functions.
17745 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17746 as "read only": it can't be written to and the buffer it points to will
17747 not be freed. Reading from a read only BIO is much more efficient than
17748 a normal memory BIO. This was added because there are several times when
17749 an area of memory needs to be read from a BIO. The previous method was
17750 to create a memory BIO and write the data to it, this results in two
17751 copies of the data and an O(n^2) reading algorithm. There is a new
17752 function BIO_new_mem_buf() which creates a read only memory BIO from
17753 an area of memory. Also modified the PKCS#7 routines to use read only
17758 * Bugfix: ssl23_get_client_hello did not work properly when called in
17759 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17760 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17761 but a retry condition occurred while trying to read the rest.
17765 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17766 NID_pkcs7_encrypted by default: this was wrong since this should almost
17767 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17768 the encrypted data type: this is a more sensible place to put it and it
17769 allows the PKCS#12 code to be tidied up that duplicated this
17774 * Changed obj_dat.pl script so it takes its input and output files on
17775 the command line. This should avoid shell escape redirection problems
17780 * Initial support for certificate extension requests, these are included
17781 in things like Xenroll certificate requests. Included functions to allow
17782 extensions to be obtained and added.
17786 * -crlf option to s_client and s_server for sending newlines as
17787 CRLF (as required by many protocols).
17791 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17793 * Install libRSAglue.a when OpenSSL is built with RSAref.
17795 *Ralf S. Engelschall*
17797 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17799 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17801 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17806 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17807 DH parameters/keys (q is lost during that conversion, but the resulting
17808 DH parameters contain its length).
17810 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17811 much faster than DH_generate_parameters (which creates parameters
17812 where `p = 2*q + 1`), and also the smaller q makes DH computations
17813 much more efficient (160-bit exponentiation instead of 1024-bit
17814 exponentiation); so this provides a convenient way to support DHE
17815 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17816 utter importance to use
17817 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17819 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17820 when such DH parameters are used, because otherwise small subgroup
17821 attacks may become possible!
17825 * Avoid memory leak in i2d_DHparams.
17829 * Allow the -k option to be used more than once in the enc program:
17830 this allows the same encrypted message to be read by multiple recipients.
17834 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17835 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17836 it will always use the numerical form of the OID, even if it has a short
17841 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17842 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17843 otherwise bn_mod_exp was called. In the case of hardware keys for example
17844 no private key components need be present and it might store extra data
17845 in the RSA structure, which cannot be accessed from bn_mod_exp.
17846 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17847 private key operations.
17851 * Added support for SPARC Linux.
17855 * pem_password_cb function type incompatibly changed from
17856 typedef int pem_password_cb(char *buf, int size, int rwflag);
17858 ....(char *buf, int size, int rwflag, void *userdata);
17859 so that applications can pass data to their callbacks:
17860 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17861 additional void * argument, which is just handed through whenever
17862 the password callback is called.
17864 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17866 New function SSL_CTX_set_default_passwd_cb_userdata.
17868 Compatibility note: As many C implementations push function arguments
17869 onto the stack in reverse order, the new library version is likely to
17870 interoperate with programs that have been compiled with the old
17871 pem_password_cb definition (PEM_whatever takes some data that
17872 happens to be on the stack as its last argument, and the callback
17873 just ignores this garbage); but there is no guarantee whatsoever that
17876 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17877 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17878 problems not only on Windows, but also on some Unix platforms.
17879 To avoid problematic command lines, these definitions are now in an
17880 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17881 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17885 * MIPS III/IV assembler module is reimplemented.
17889 * More DES library cleanups: remove references to srand/rand and
17890 delete an unused file.
17894 * Add support for the free Netwide assembler (NASM) under Win32,
17895 since not many people have MASM (ml) and it can be hard to obtain.
17896 This is currently experimental but it seems to work OK and pass all
17897 the tests. Check out INSTALL.W32 for info.
17901 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17902 without temporary keys kept an extra copy of the server key,
17903 and connections with temporary keys did not free everything in case
17908 * New function RSA_check_key and new openssl rsa option -check
17909 for verifying the consistency of RSA keys.
17911 *Ulf Moeller, Bodo Moeller*
17913 * Various changes to make Win32 compile work:
17914 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17915 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17916 comparison" warnings.
17917 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17921 * Add a debugging option to PKCS#5 v2 key generation function: when
17922 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17923 derived keys are printed to stderr.
17927 * Copy the flags in ASN1_STRING_dup().
17929 *Roman E. Pavlov <pre@mo.msk.ru>*
17931 * The x509 application mishandled signing requests containing DSA
17932 keys when the signing key was also DSA and the parameters didn't match.
17934 It was supposed to omit the parameters when they matched the signing key:
17935 the verifying software was then supposed to automatically use the CA's
17936 parameters if they were absent from the end user certificate.
17938 Omitting parameters is no longer recommended. The test was also
17939 the wrong way round! This was probably due to unusual behaviour in
17940 EVP_cmp_parameters() which returns 1 if the parameters match.
17941 This meant that parameters were omitted when they *didn't* match and
17942 the certificate was useless. Certificates signed with 'ca' didn't have
17945 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17947 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17948 The interface is as follows:
17949 Applications can use
17950 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17951 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17952 "off" is now the default.
17953 The library internally uses
17954 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17955 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17956 to disable memory-checking temporarily.
17958 Some inconsistent states that previously were possible (and were
17959 even the default) are now avoided.
17961 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17962 with each memory chunk allocated; this is occasionally more helpful
17963 than just having a counter.
17965 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17967 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17972 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17973 which largely parallels "options", but is for changing API behaviour,
17974 whereas "options" are about protocol behaviour.
17975 Initial "mode" flags are:
17977 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17978 a single record has been written.
17979 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17980 retries use the same buffer location.
17981 (But all of the contents must be
17986 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17989 * Fix problems with no-hmac etc.
17991 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17993 * New functions RSA_get_default_method(), RSA_set_method() and
17994 RSA_get_method(). These allows replacement of RSA_METHODs without having
17995 to mess around with the internals of an RSA structure.
17999 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18000 Also really enable memory leak checks in openssl.c and in some
18003 *Chad C. Mulligan, Bodo Moeller*
18005 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18006 up the length of negative integers. This has now been simplified to just
18007 store the length when it is first determined and use it later, rather
18008 than trying to keep track of where data is copied and updating it to
18010 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18012 * Add a new function PKCS7_signatureVerify. This allows the verification
18013 of a PKCS#7 signature but with the signing certificate passed to the
18014 function itself. This contrasts with PKCS7_dataVerify which assumes the
18015 certificate is present in the PKCS#7 structure. This isn't always the
18016 case: certificates can be omitted from a PKCS#7 structure and be
18017 distributed by "out of band" means (such as a certificate database).
18021 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18022 function prototypes in pem.h, also change util/mkdef.pl to add the
18023 necessary function names.
18027 * mk1mf.pl (used by Windows builds) did not properly read the
18028 options set by Configure in the top level Makefile, and Configure
18029 was not even able to write more than one option correctly.
18030 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18034 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18035 file to be loaded from a BIO or FILE pointer. The BIO version will
18036 for example allow memory BIOs to contain config info.
18040 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18041 Whoever hopes to achieve shared-library compatibility across versions
18042 must use this, not the compile-time macro.
18043 (Exercise 0.9.4: Which is the minimum library version required by
18045 Note: All this applies only to multi-threaded programs, others don't
18050 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18051 through a BIO pair triggered the default case, i.e.
18052 SSLerr(...,SSL_R_UNKNOWN_STATE).
18056 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18057 can use the SSL library even if none of the specific BIOs is
18062 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18063 for the encoded length.
18065 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18067 * Add initial documentation of the X509V3 functions.
18071 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18072 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18073 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18074 secure PKCS#8 private key format with a high iteration count.
18078 * Fix determination of Perl interpreter: A perl or perl5
18079 *directory* in $PATH was also accepted as the interpreter.
18081 *Ralf S. Engelschall*
18083 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18084 wrong with it but it was very old and did things like calling
18085 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18086 unusual formatting.
18090 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18091 to use the new extension code.
18095 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18096 with macros. This should make it easier to change their form, add extra
18097 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18102 * Add to configuration table a new entry that can specify an alternative
18103 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18104 according to Mark Crispin <MRC@Panda.COM>.
18108 * DES CBC did not update the IV. Weird.
18112 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18113 Changing the behaviour of the former might break existing programs --
18114 where IV updating is needed, des_ncbc_encrypt can be used.
18117 * When bntest is run from "make test" it drives bc to check its
18118 calculations, as well as internally checking them. If an internal check
18119 fails, it needs to cause bc to give a non-zero result or make test carries
18120 on without noticing the failure. Fixed.
18124 * DES library cleanups.
18128 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18129 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18130 ciphers. NOTE: although the key derivation function has been verified
18131 against some published test vectors it has not been extensively tested
18132 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18137 * Instead of "mkdir -p", which is not fully portable, use new
18138 Perl script "util/mkdir-p.pl".
18142 * Rewrite the way password based encryption (PBE) is handled. It used to
18143 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18144 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18145 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18146 the 'parameter' field of the AlgorithmIdentifier is passed to the
18147 underlying key generation function so it must do its own ASN1 parsing.
18148 This has also changed the EVP_PBE_CipherInit() function which now has a
18149 'parameter' argument instead of literal salt and iteration count values
18150 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18154 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18155 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18156 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18157 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18158 value was just used as a "magic string" and not used directly its
18159 value doesn't matter.
18163 * Introduce some semblance of const correctness to BN. Shame C doesn't
18168 * "linux-sparc64" configuration (ultrapenguin).
18170 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18171 "linux-sparc" configuration.
18173 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18175 * config now generates no-xxx options for missing ciphers.
18179 * Support the EBCDIC character set (work in progress).
18180 File ebcdic.c not yet included because it has a different license.
18182 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18184 * Support BS2000/OSD-POSIX.
18186 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18188 * Make callbacks for key generation use `void *` instead of `char *`.
18192 * Make S/MIME samples compile (not yet tested).
18196 * Additional typesafe stacks.
18200 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18204 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18206 * New configuration variant "sco5-gcc".
18208 * Updated some demos.
18210 *Sean O Riordain, Wade Scholine*
18212 * Add missing BIO_free at exit of pkcs12 application.
18216 * Fix memory leak in conf.c.
18220 * Updates for Win32 to assembler version of MD5.
18224 * Set #! path to perl in `apps/der_chop` to where we found it
18225 instead of using a fixed path.
18229 * SHA library changes for irix64-mips4-cc.
18233 * Improvements for VMS support.
18237 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18239 * Bignum library bug fix. IRIX 6 passes "make test" now!
18240 This also avoids the problems with SC4.2 and unpatched SC5.
18242 *Andy Polyakov <appro@fy.chalmers.se>*
18244 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18245 These are required because of the typesafe stack would otherwise break
18246 existing code. If old code used a structure member which used to be STACK
18247 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18248 sk_num or sk_value it would produce an error because the num, data members
18249 are not present in STACK_OF. Now it just produces a warning. sk_set
18250 replaces the old method of assigning a value to sk_value
18251 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18252 that does this will no longer work (and should use sk_set instead) but
18253 this could be regarded as a "questionable" behaviour anyway.
18257 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18258 correctly handle encrypted S/MIME data.
18262 * Change type of various DES function arguments from des_cblock
18263 (which means, in function argument declarations, pointer to char)
18264 to des_cblock * (meaning pointer to array with 8 char elements),
18265 which allows the compiler to do more typechecking; it was like
18266 that back in SSLeay, but with lots of ugly casts.
18268 Introduce new type const_des_cblock.
18272 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18273 problems: find RecipientInfo structure that matches recipient certificate
18274 and initialise the ASN1 structures properly based on passed cipher.
18278 * Belatedly make the BN tests actually check the results.
18282 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18283 to and from BNs: it was completely broken. New compilation option
18284 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18285 key elements as negative integers.
18289 * Reorganize and speed up MD5.
18291 *Andy Polyakov <appro@fy.chalmers.se>*
18295 *Richard Levitte <richard@levitte.org>*
18297 * New option -out to asn1parse to allow the parsed structure to be
18298 output to a file. This is most useful when combined with the -strparse
18299 option to examine the output of things like OCTET STRINGS.
18303 * Make SSL library a little more fool-proof by not requiring any longer
18304 that `SSL_set_{accept,connect}_state` be called before
18305 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18306 in many applications because usually everything *appeared* to work as
18307 intended anyway -- now it really works as intended).
18311 * Move openssl.cnf out of lib/.
18315 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18316 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18317 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18319 *Ralf S. Engelschall*
18321 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18322 handle PKCS#7 enveloped data properly.
18324 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18326 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18327 copying pointers. The cert_st handling is changed by this in
18328 various ways (and thus what used to be known as ctx->default_cert
18329 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18330 any longer when s->cert does not give us what we need).
18331 ssl_cert_instantiate becomes obsolete by this change.
18332 As soon as we've got the new code right (possibly it already is?),
18333 we have solved a couple of bugs of the earlier code where s->cert
18334 was used as if it could not have been shared with other SSL structures.
18336 Note that using the SSL API in certain dirty ways now will result
18337 in different behaviour than observed with earlier library versions:
18338 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18339 does not influence s as it used to.
18341 In order to clean up things more thoroughly, inside SSL_SESSION
18342 we don't use CERT any longer, but a new structure SESS_CERT
18343 that holds per-session data (if available); currently, this is
18344 the peer's certificate chain and, for clients, the server's certificate
18345 and temporary key. CERT holds only those values that can have
18346 meaningful defaults in an SSL_CTX.
18350 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18351 from the internal representation. Various PKCS#7 fixes: remove some
18352 evil casts and set the enc_dig_alg field properly based on the signing
18357 * Allow PKCS#12 password to be set from the command line or the
18358 environment. Let 'ca' get its config file name from the environment
18359 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18364 * Allow certificate policies extension to use an IA5STRING for the
18365 organization field. This is contrary to the PKIX definition but
18366 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18371 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18372 without disallowing inline assembler and the like for non-pedantic builds.
18376 * Support Borland C++ builder.
18378 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18384 * SHA-1 cleanups and performance enhancements.
18386 *Andy Polyakov <appro@fy.chalmers.se>*
18388 * Sparc v8plus assembler for the bignum library.
18390 *Andy Polyakov <appro@fy.chalmers.se>*
18392 * Accept any -xxx and +xxx compiler options in Configure.
18396 * Update HPUX configuration.
18400 * Add missing `sk_<type>_unshift()` function to safestack.h
18402 *Ralf S. Engelschall*
18404 * New function SSL_CTX_use_certificate_chain_file that sets the
18405 "extra_cert"s in addition to the certificate. (This makes sense
18406 only for "PEM" format files, as chains as a whole are not
18411 * Support verify_depth from the SSL API.
18412 x509_vfy.c had what can be considered an off-by-one-error:
18413 Its depth (which was not part of the external interface)
18414 was actually counting the number of certificates in a chain;
18415 now it really counts the depth.
18419 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18420 instead of X509err, which often resulted in confusing error
18421 messages since the error codes are not globally unique
18422 (e.g. an alleged error in ssl3_accept when a certificate
18423 didn't match the private key).
18425 * New function SSL_CTX_set_session_id_context that allows to set a default
18426 value (so that you don't need SSL_set_session_id_context for each
18427 connection using the SSL_CTX).
18431 * OAEP decoding bug fix.
18435 * Support INSTALL_PREFIX for package builders, as proposed by
18440 * New Configure options "threads" and "no-threads". For systems
18441 where the proper compiler options are known (currently Solaris
18442 and Linux), "threads" is the default.
18446 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18450 * Install various scripts to $(OPENSSLDIR)/misc, not to
18451 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18452 such as /usr/local/bin.
18456 * "make linux-shared" to build shared libraries.
18458 *Niels Poppe <niels@netbox.org>*
18460 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18464 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18465 extension adding in x509 utility.
18469 * Remove NOPROTO sections and error code comments.
18473 * Partial rewrite of the DEF file generator to now parse the ANSI
18478 * New Configure options --prefix=DIR and --openssldir=DIR.
18482 * Complete rewrite of the error code script(s). It is all now handled
18483 by one script at the top level which handles error code gathering,
18484 header rewriting and C source file generation. It should be much better
18485 than the old method: it now uses a modified version of Ulf's parser to
18486 read the ANSI prototypes in all header files (thus the old K&R definitions
18487 aren't needed for error creation any more) and do a better job of
18488 translating function codes into names. The old 'ASN1 error code embedded
18489 in a comment' is no longer necessary and it doesn't use .err files which
18490 have now been deleted. Also the error code call doesn't have to appear all
18491 on one line (which resulted in some large lines...).
18495 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18499 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18500 0 (which usually indicates a closed connection), but continue reading.
18504 * Fix some race conditions.
18508 * Add support for CRL distribution points extension. Add Certificate
18509 Policies and CRL distribution points documentation.
18513 * Move the autogenerated header file parts to crypto/opensslconf.h.
18517 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18518 8 of keying material. Merlin has also confirmed interop with this fix
18519 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18521 *Merlin Hughes <merlin@baltimore.ie>*
18523 * Fix lots of warnings.
18525 *Richard Levitte <levitte@stacken.kth.se>*
18527 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18528 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18530 *Richard Levitte <levitte@stacken.kth.se>*
18532 * Fix problems with sizeof(long) == 8.
18534 *Andy Polyakov <appro@fy.chalmers.se>*
18536 * Change functions to ANSI C.
18540 * Fix typos in error codes.
18542 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18544 * Remove defunct assembler files from Configure.
18548 * SPARC v8 assembler BIGNUM implementation.
18550 *Andy Polyakov <appro@fy.chalmers.se>*
18552 * Support for Certificate Policies extension: both print and set.
18553 Various additions to support the r2i method this uses.
18557 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18558 return a const string when you are expecting an allocated buffer.
18562 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18563 types DirectoryString and DisplayText.
18567 * Add code to allow r2i extensions to access the configuration database,
18568 add an LHASH database driver and add several ctx helper functions.
18572 * Fix an evil bug in bn_expand2() which caused various BN functions to
18573 fail when they extended the size of a BIGNUM.
18577 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18578 support typesafe stack.
18582 * Fix typo in SSL_[gs]et_options().
18584 *Nils Frostberg <nils@medcom.se>*
18586 * Delete various functions and files that belonged to the (now obsolete)
18587 old X509V3 handling code.
18591 * New Configure option "rsaref".
18595 * Don't auto-generate pem.h.
18599 * Introduce type-safe ASN.1 SETs.
18603 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18605 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18607 * Introduce type-safe STACKs. This will almost certainly break lots of code
18608 that links with OpenSSL (well at least cause lots of warnings), but fear
18609 not: the conversion is trivial, and it eliminates loads of evil casts. A
18610 few STACKed things have been converted already. Feel free to convert more.
18611 In the fullness of time, I'll do away with the STACK type altogether.
18615 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18616 specified in `<certfile>` by updating the entry in the index.txt file.
18617 This way one no longer has to edit the index.txt file manually for
18618 revoking a certificate. The -revoke option does the gory details now.
18620 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18622 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18623 `-text` option at all and this way the `-noout -text` combination was
18624 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18626 *Ralf S. Engelschall*
18628 * Make sure a corresponding plain text error message exists for the
18629 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18630 verify callback function determined that a certificate was revoked.
18632 *Ralf S. Engelschall*
18634 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18635 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18636 all available ciphers including rc5, which was forgotten until now.
18637 In order to let the testing shell script know which algorithms
18638 are available, a new (up to now undocumented) command
18639 `openssl list-cipher-commands` is used.
18643 * Bugfix: s_client occasionally would sleep in select() when
18644 it should have checked SSL_pending() first.
18648 * New functions DSA_do_sign and DSA_do_verify to provide access to
18649 the raw DSA values prior to ASN.1 encoding.
18653 * Tweaks to Configure
18655 *Niels Poppe <niels@netbox.org>*
18657 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18662 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18666 * New config option to avoid instructions that are illegal on the 80386.
18667 The default code is faster, but requires at least a 486.
18671 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18672 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18673 same as SSL2_VERSION anyway.
18677 * New "-showcerts" option for s_client.
18681 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18682 application. Various cleanups and fixes.
18686 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18687 modify error routines to work internally. Add error codes and PBE init
18688 to library startup routines.
18692 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18693 packing functions to asn1 and evp. Changed function names and error
18694 codes along the way.
18698 * PKCS12 integration: and so it begins... First of several patches to
18699 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18700 objects to objects.h
18704 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18705 and display support for Thawte strong extranet extension.
18709 * Add LinuxPPC support.
18711 *Jeff Dubrule <igor@pobox.org>*
18713 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18714 bn_div_words in alpha.s.
18716 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18718 * Make sure the RSA OAEP test is skipped under -DRSAref because
18719 OAEP isn't supported when OpenSSL is built with RSAref.
18721 *Ulf Moeller <ulf@fitug.de>*
18723 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18724 so they no longer are missing under -DNOPROTO.
18726 *Soren S. Jorvang <soren@t.dk>*
18728 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18730 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18731 doesn't work when the session is reused. Coming soon!
18735 * Fix a security hole, that allows sessions to be reused in the wrong
18736 context thus bypassing client cert protection! All software that uses
18737 client certs and session caches in multiple contexts NEEDS PATCHING to
18738 allow session reuse! A fuller solution is in the works.
18740 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18742 * Some more source tree cleanups (removed obsolete files
18743 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18744 permission on "config" script to be executable) and a fix for the INSTALL
18747 *Ulf Moeller <ulf@fitug.de>*
18749 * Remove some legacy and erroneous uses of malloc, free instead of
18752 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18754 * Make rsa_oaep_test return non-zero on error.
18756 *Ulf Moeller <ulf@fitug.de>*
18758 * Add support for native Solaris shared libraries. Configure
18759 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18760 if someone would make that last step automatic.
18762 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18764 * ctx_size was not built with the right compiler during "make links". Fixed.
18768 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18769 except NULL ciphers". This means the default cipher list will no longer
18770 enable NULL ciphers. They need to be specifically enabled e.g. with
18771 the string "DEFAULT:eNULL".
18775 * Fix to RSA private encryption routines: if p < q then it would
18776 occasionally produce an invalid result. This will only happen with
18777 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18781 * Be less restrictive and allow also `perl util/perlpath.pl
18782 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18783 because this way one can also use an interpreter named `perl5` (which is
18784 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18785 installed as `perl`).
18787 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18789 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18791 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18793 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18794 advapi32.lib to Win32 build and change the pem test comparison
18795 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18796 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18797 and crypto/des/ede_cbcm_enc.c.
18801 * DES quad checksum was broken on big-endian architectures. Fixed.
18805 * Comment out two functions in bio.h that aren't implemented. Fix up the
18806 Win32 test batch file so it (might) work again. The Win32 test batch file
18807 is horrible: I feel ill....
18811 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18812 in e_os.h. Audit of header files to check ANSI and non ANSI
18813 sections: 10 functions were absent from non ANSI section and not exported
18814 from Windows DLLs. Fixed up libeay.num for new functions.
18818 * Make `openssl version` output lines consistent.
18820 *Ralf S. Engelschall*
18822 * Fix Win32 symbol export lists for BIO functions: Added
18823 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18824 to ms/libeay{16,32}.def.
18826 *Ralf S. Engelschall*
18828 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18829 fine under Unix and passes some trivial tests I've now added. But the
18830 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18831 added to make sure no one expects that this stuff really works in the
18832 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18833 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18836 *Ralf S. Engelschall*
18838 * Fix the generation of two part addresses in perl.
18840 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18842 * Add config entry for Linux on MIPS.
18844 *John Tobey <jtobey@channel1.com>*
18846 * Make links whenever Configure is run, unless we are on Windoze.
18850 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18851 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18856 * Add a useful kludge to allow package maintainers to specify compiler and
18857 other platforms details on the command line without having to patch the
18858 Configure script every time: One now can use
18859 `perl Configure <id>:<details>`,
18860 i.e. platform ids are allowed to have details appended
18861 to them (separated by colons). This is treated as there would be a static
18862 pre-configured entry in Configure's %table under key `<id>` with value
18863 `<details>` and `perl Configure <id>` is called. So, when you want to
18864 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18865 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18866 now, which overrides the FreeBSD-elf entry on-the-fly.
18868 *Ralf S. Engelschall*
18870 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18874 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18875 on the `perl Configure ...` command line. This way one can compile
18876 OpenSSL libraries with Position Independent Code (PIC) which is needed
18877 for linking it into DSOs.
18879 *Ralf S. Engelschall*
18881 * Remarkably, export ciphers were totally broken and no-one had noticed!
18886 * Cleaned up the LICENSE document: The official contact for any license
18887 questions now is the OpenSSL core team under openssl-core@openssl.org.
18888 And add a paragraph about the dual-license situation to make sure people
18889 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18890 to the OpenSSL toolkit.
18892 *Ralf S. Engelschall*
18894 * General source tree makefile cleanups: Made `making xxx in yyy...`
18895 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18896 Additionally cleaned up the `make links` target: Remove unnecessary
18897 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18898 to speed processing and no longer clutter the display with confusing
18899 stuff. Instead only the actually done links are displayed.
18901 *Ralf S. Engelschall*
18903 * Permit null encryption ciphersuites, used for authentication only. It used
18904 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18905 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18910 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18911 signed attributes when verifying signatures (this would break them),
18912 the detached data encoding was wrong and public keys obtained using
18913 X509_get_pubkey() weren't freed.
18917 * Add text documentation for the BUFFER functions. Also added a work around
18918 to a Win95 console bug. This was triggered by the password read stuff: the
18919 last character typed gets carried over to the next fread(). If you were
18920 generating a new cert request using 'req' for example then the last
18921 character of the passphrase would be CR which would then enter the first
18926 * Added the new 'Includes OpenSSL Cryptography Software' button as
18927 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18928 button and can be used by applications based on OpenSSL to show the
18929 relationship to the OpenSSL project.
18931 *Ralf S. Engelschall*
18933 * Remove confusing variables in function signatures in files
18934 ssl/ssl_lib.c and ssl/ssl.h.
18936 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18938 * Don't install bss_file.c under PREFIX/include/
18940 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18942 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18943 functions that return function pointers and has support for NT specific
18944 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18945 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18946 unsigned to signed types: this was killing the Win32 compile.
18950 * Add new certificate file to stack functions,
18951 SSL_add_dir_cert_subjects_to_stack() and
18952 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18953 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18954 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18955 This means that Apache-SSL and similar packages don't have to mess around
18956 to add as many CAs as they want to the preferred list.
18960 * Experiment with doxygen documentation. Currently only partially applied to
18962 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18963 openssl.doxy as the configuration file.
18967 * Get rid of remaining C++-style comments which strict C compilers hate.
18969 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18971 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18972 compiled in by default: it has problems with large keys.
18976 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18977 DH private keys and/or callback functions which directly correspond to
18978 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18979 is needed for applications which have to configure certificates on a
18980 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18982 For the RSA certificate situation is makes no difference, but
18983 for the DSA certificate situation this fixes the "no shared cipher"
18984 problem where the OpenSSL cipher selection procedure failed because the
18985 temporary keys were not overtaken from the context and the API provided
18986 no way to reconfigure them.
18987 The new functions now let applications reconfigure the stuff and they
18988 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18989 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18990 non-public-API function ssl_cert_instantiate() is used as a helper
18991 function and also to reduce code redundancy inside ssl_rsa.c.
18993 *Ralf S. Engelschall*
18995 * Move s_server -dcert and -dkey options out of the undocumented feature
18996 area because they are useful for the DSA situation and should be
18997 recognized by the users.
18999 *Ralf S. Engelschall*
19001 * Fix the cipher decision scheme for export ciphers: the export bits are
19002 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19003 SSL_EXP_MASK. So, the original variable has to be used instead of the
19004 already masked variable.
19006 *Richard Levitte <levitte@stacken.kth.se>*
19008 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19010 *Richard Levitte <levitte@stacken.kth.se>*
19012 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19013 from `int` to `unsigned int` because it is a length and initialized by
19014 EVP_DigestFinal() which expects an `unsigned int *`.
19016 *Richard Levitte <levitte@stacken.kth.se>*
19018 * Don't hard-code path to Perl interpreter on shebang line of Configure
19019 script. Instead use the usual Shell->Perl transition trick.
19021 *Ralf S. Engelschall*
19023 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19024 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19025 -noout -modulus` as it's already the case for `openssl rsa -noout
19026 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19027 currently the public key is printed (a decision which was already done by
19028 `openssl dsa -modulus` in the past) which serves a similar purpose.
19029 Additionally the NO_RSA no longer completely removes the whole -modulus
19030 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19033 *Ralf S. Engelschall*
19035 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19036 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19038 *Arne Ansper <arne@ats.cyber.ee>*
19040 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19041 to be added. Now both 'req' and 'ca' can use new objects defined in the
19046 * Add cool BIO that does syslog (or event log on NT).
19048 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19050 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19051 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19052 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19053 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19057 * Add preliminary config info for new extension code.
19061 * Make RSA_NO_PADDING really use no padding.
19063 *Ulf Moeller <ulf@fitug.de>*
19065 * Generate errors when private/public key check is done.
19069 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19070 for some CRL extensions and new objects added.
19074 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19075 key usage extension and fuller support for authority key id.
19079 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19080 padding method for RSA, which is recommended for new applications in PKCS
19081 #1 v2.0 (RFC 2437, October 1998).
19082 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19083 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19084 against Bleichbacher's attack on RSA.
19085 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19088 * Updates to the new SSL compression code
19090 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19092 * Fix so that the version number in the master secret, when passed
19093 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19094 (because the server will not accept higher), that the version number
19095 is 0x03,0x01, not 0x03,0x00
19097 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19099 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19100 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19101 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19105 * Support for RAW extensions where an arbitrary extension can be
19106 created by including its DER encoding. See `apps/openssl.cnf` for
19111 * Make sure latest Perl versions don't interpret some generated C array
19112 code as Perl array code in the crypto/err/err_genc.pl script.
19114 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19116 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19117 not many people have the assembler. Various Win32 compilation fixes and
19118 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19119 build instructions.
19123 * Modify configure script 'Configure' to automatically create crypto/date.h
19124 file under Win32 and also build pem.h from pem.org. New script
19125 util/mkfiles.pl to create the MINFO file on environments that can't do a
19126 'make files': perl util/mkfiles.pl >MINFO should work.
19130 * Major rework of DES function declarations, in the pursuit of correctness
19131 and purity. As a result, many evil casts evaporated, and some weirdness,
19132 too. You may find this causes warnings in your code. Zapping your evil
19133 casts will probably fix them. Mostly.
19137 * Fix for a typo in asn1.h. Bug fix to object creation script
19138 obj_dat.pl. It considered a zero in an object definition to mean
19139 "end of object": none of the objects in objects.h have any zeros
19140 so it wasn't spotted.
19142 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19144 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19145 Masking (CBCM). In the absence of test vectors, the best I have been able
19146 to do is check that the decrypt undoes the encrypt, so far. Send me test
19147 vectors if you have them.
19151 * Correct calculation of key length for export ciphers (too much space was
19152 allocated for null ciphers). This has not been tested!
19156 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19157 message is now correct (it understands "crypto" and "ssl" on its
19158 command line). There is also now an "update" option. This will update
19159 the util/ssleay.num and util/libeay.num files with any new functions.
19161 perl util/mkdef.pl crypto ssl update
19162 it will update them.
19166 * Overhauled the Perl interface:
19167 - ported BN stuff to OpenSSL's different BN library
19168 - made the perl/ source tree CVS-aware
19169 - renamed the package from SSLeay to OpenSSL (the files still contain
19170 their history because I've copied them in the repository)
19171 - removed obsolete files (the test scripts will be replaced
19172 by better Test::Harness variants in the future)
19174 *Ralf S. Engelschall*
19176 * First cut for a very conservative source tree cleanup:
19177 1. merge various obsolete readme texts into doc/ssleay.txt
19178 where we collect the old documents and readme texts.
19179 2. remove the first part of files where I'm already sure that we no
19180 longer need them because of three reasons: either they are just temporary
19181 files which were left by Eric or they are preserved original files where
19182 I've verified that the diff is also available in the CVS via "cvs diff
19183 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19184 the crypto/md/ stuff).
19186 *Ralf S. Engelschall*
19188 * More extension code. Incomplete support for subject and issuer alt
19189 name, issuer and authority key id. Change the i2v function parameters
19190 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19191 what that's for :-) Fix to ASN1 macro which messed up
19192 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19196 * Preliminary support for ENUMERATED type. This is largely copied from the
19201 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19203 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19205 * Make sure `make rehash` target really finds the `openssl` program.
19207 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19209 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19210 like to hear about it if this slows down other processors.
19214 * Add CygWin32 platform information to Configure script.
19216 *Alan Batie <batie@aahz.jf.intel.com>*
19218 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19220 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19222 * New program nseq to manipulate netscape certificate sequences
19226 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19231 * Fixes to BN code. Previously the default was to define BN_RECURSION
19232 but the BN code had some problems that would cause failures when
19233 doing certificate verification and some other functions.
19235 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19237 * Add ASN1 and PEM code to support netscape certificate sequences.
19241 * Add ASN1 and PEM code to support netscape certificate sequences.
19245 * Add several PKIX and private extended key usage OIDs.
19249 * Modify the 'ca' program to handle the new extension code. Modify
19250 openssl.cnf for new extension format, add comments.
19254 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19255 and add a sample to openssl.cnf so req -x509 now adds appropriate
19260 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19261 error code, add initial support to X509_print() and x509 application.
19265 * Takes a deep breath and start adding X509 V3 extension support code. Add
19266 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19267 stuff is currently isolated and isn't even compiled yet.
19271 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19272 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19273 Removed the versions check from X509 routines when loading extensions:
19274 this allows certain broken certificates that don't set the version
19275 properly to be processed.
19279 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19280 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19281 can still be regenerated with "make depend".
19285 * Spelling mistake in C version of CAST-128.
19287 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19289 * Changes to the error generation code. The perl script err-code.pl
19290 now reads in the old error codes and retains the old numbers, only
19291 adding new ones if necessary. It also only changes the .err files if new
19292 codes are added. The makefiles have been modified to only insert errors
19293 when needed (to avoid needlessly modifying header files). This is done
19294 by only inserting errors if the .err file is newer than the auto generated
19295 C file. To rebuild all the error codes from scratch (the old behaviour)
19296 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19297 or delete all the .err files.
19301 * CAST-128 was incorrectly implemented for short keys. The C version has
19302 been fixed, but is untested. The assembler versions are also fixed, but
19303 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19304 to regenerate it if needed.
19305 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19306 Hagino <itojun@kame.net>*
19308 * File was opened incorrectly in randfile.c.
19310 *Ulf Möller <ulf@fitug.de>*
19312 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19313 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19314 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19315 al: it's just almost always a UTCTime. Note this patch adds new error
19316 codes so do a "make errors" if there are problems.
19320 * Correct Linux 1 recognition in config.
19322 *Ulf Möller <ulf@fitug.de>*
19324 * Remove pointless MD5 hash when using DSA keys in ca.
19326 *Anonymous <nobody@replay.com>*
19328 * Generate an error if given an empty string as a cert directory. Also
19329 generate an error if handed NULL (previously returned 0 to indicate an
19330 error, but didn't set one).
19332 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19334 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19338 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19339 parameters. This was causing a warning which killed off the Win32 compile.
19343 * Remove C++ style comments from crypto/bn/bn_local.h.
19345 *Neil Costigan <neil.costigan@celocom.com>*
19347 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19348 based on a text string, looking up short and long names and finally
19349 "dot" format. The "dot" format stuff didn't work. Added new function
19350 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19351 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19352 OID is not part of the table.
19356 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19357 X509_LOOKUP_by_alias().
19361 * Sort openssl functions by name.
19365 * Get the `gendsa` command working and add it to the `list` command. Remove
19366 encryption from sample DSA keys (in case anyone is interested the password
19371 * Make *all* `*_free` functions accept a NULL pointer.
19373 *Frans Heymans <fheymans@isaserver.be>*
19375 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19378 *Anonymous <nobody@replay.com>*
19380 * s_server should send the CAfile as acceptable CAs, not its own cert.
19382 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19384 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19386 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19388 * Temp key "for export" tests were wrong in s3_srvr.c.
19390 *Anonymous <nobody@replay.com>*
19392 * Add prototype for temp key callback functions
19393 SSL_CTX_set_tmp_{rsa,dh}_callback().
19397 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19398 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19402 * X509_name_add_entry() freed the wrong thing after an error.
19404 *Arne Ansper <arne@ats.cyber.ee>*
19406 * rsa_eay.c would attempt to free a NULL context.
19408 *Arne Ansper <arne@ats.cyber.ee>*
19410 * BIO_s_socket() had a broken should_retry() on Windoze.
19412 *Arne Ansper <arne@ats.cyber.ee>*
19414 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19416 *Arne Ansper <arne@ats.cyber.ee>*
19418 * Make sure the already existing X509_STORE->depth variable is initialized
19419 in X509_STORE_new(), but document the fact that this variable is still
19420 unused in the certificate verification process.
19422 *Ralf S. Engelschall*
19424 * Fix the various library and `apps/` files to free up pkeys obtained from
19425 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19429 * Fix reference counting in X509_PUBKEY_get(). This makes
19430 demos/maurice/example2.c work, amongst others, probably.
19432 *Steve Henson and Ben Laurie*
19434 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19435 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19436 are no longer created. This way we have a single and consistent command
19437 line interface `openssl <command>`, similar to `cvs <command>`.
19439 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19441 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19442 BIT STRING wrapper always have zero unused bits.
19446 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19450 * Make the top-level INSTALL documentation easier to understand.
19454 * Makefiles updated to exit if an error occurs in a sub-directory
19455 make (including if user presses ^C) [Paul Sutton]
19457 * Make Montgomery context stuff explicit in RSA data structure.
19461 * Fix build order of pem and err to allow for generated pem.h.
19465 * Fix renumbering bug in X509_NAME_delete_entry().
19469 * Enhanced the err-ins.pl script so it makes the error library number
19470 global and can add a library name. This is needed for external ASN1 and
19471 other error libraries.
19475 * Fixed sk_insert which never worked properly.
19479 * Fix ASN1 macros so they can handle indefinite length constructed
19480 EXPLICIT tags. Some non standard certificates use these: they can now
19485 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19486 into a single doc/ssleay.txt bundle. This way the information is still
19487 preserved but no longer messes up this directory. Now it's new room for
19488 the new set of documentation files.
19490 *Ralf S. Engelschall*
19492 * SETs were incorrectly DER encoded. This was a major pain, because they
19493 shared code with SEQUENCEs, which aren't coded the same. This means that
19494 almost everything to do with SETs or SEQUENCEs has either changed name or
19495 number of arguments.
19497 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19499 * Fix test data to work with the above.
19503 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19504 was already fixed by Eric for 0.9.1 it seems.
19506 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19508 * Autodetect FreeBSD3.
19512 * Fix various bugs in Configure. This affects the following platforms:
19516 unixware-2.0-pentium
19521 * Eliminate generated files from CVS. Reorder tests to regenerate files
19522 before they are needed.
19526 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19530 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19532 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19533 changed SSLeay to OpenSSL in version strings.
19535 *Ralf S. Engelschall*
19537 * Some fixups to the top-level documents.
19541 * Fixed the nasty bug where rsaref.h was not found under compile-time
19542 because the symlink to include/ was missing.
19544 *Ralf S. Engelschall*
19546 * Incorporated the popular no-RSA/DSA-only patches
19547 which allow to compile an RSA-free SSLeay.
19549 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19551 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19552 when "ssleay" is still not found.
19554 *Ralf S. Engelschall*
19556 * Added more platforms to Configure: Cray T3E, HPUX 11,
19558 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19560 * Updated the README file.
19562 *Ralf S. Engelschall*
19564 * Added various .cvsignore files in the CVS repository subdirs
19565 to make a "cvs update" really silent.
19567 *Ralf S. Engelschall*
19569 * Recompiled the error-definition header files and added
19570 missing symbols to the Win32 linker tables.
19572 *Ralf S. Engelschall*
19574 * Cleaned up the top-level documents;
19575 o new files: CHANGES and LICENSE
19576 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19577 o merged COPYRIGHT into LICENSE
19578 o removed obsolete TODO file
19579 o renamed MICROSOFT to INSTALL.W32
19581 *Ralf S. Engelschall*
19583 * Removed dummy files from the 0.9.1b source tree:
19584 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19585 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19586 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19587 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19588 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19590 *Ralf S. Engelschall*
19592 * Added various platform portability fixes.
19596 * The Genesis of the OpenSSL rpject:
19597 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19598 Young and Tim J. Hudson created while they were working for C2Net until
19601 *The OpenSSL Project*
19603 ### Changes between 0.9.0b and 0.9.1b [not released]
19605 * Updated a few CA certificates under certs/
19609 * Changed some BIGNUM api stuff.
19613 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19614 DGUX x86, Linux Alpha, etc.
19618 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19619 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19624 * Add -strparse option to asn1pars program which parses nested
19627 *Dr Stephen Henson <shenson@bigfoot.com>*
19629 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19633 * DSA fix for "ca" program.
19637 * Added "-genkey" option to "dsaparam" program.
19641 * Added RIPE MD160 (rmd160) message digest.
19645 * Added -a (all) option to "ssleay version" command.
19649 * Added PLATFORM define which is the id given to Configure.
19653 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19657 * Extended the ASN.1 parser routines.
19661 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19665 * Added a BN_CTX to the BN library.
19669 * Fixed the weak key values in DES library
19673 * Changed API in EVP library for cipher aliases.
19677 * Added support for RC2/64bit cipher.
19681 * Converted the lhash library to the crypto/mem.c functions.
19685 * Added more recognized ASN.1 object ids.
19689 * Added more RSA padding checks for SSL/TLS.
19693 * Added BIO proxy/filter functionality.
19697 * Added extra_certs to SSL_CTX which can be used
19698 send extra CA certificates to the client in the CA cert chain sending
19699 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19703 * Now Fortezza is denied in the authentication phase because
19704 this is key exchange mechanism is not supported by SSLeay at all.
19708 * Additional PKCS1 checks.
19712 * Support the string "TLSv1" for all TLS v1 ciphers.
19716 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19717 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19721 * Fixed a few memory leaks.
19725 * Fixed various code and comment typos.
19729 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19730 bytes sent in the client random.
19732 *Edward Bishop <ebishop@spyglass.com>*
19736 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19737 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19738 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19739 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19740 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19741 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19742 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19743 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19744 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19745 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19746 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19747 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19748 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19749 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19750 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19751 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19752 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19753 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19754 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19755 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19756 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19757 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19758 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19759 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19760 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19761 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19762 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19763 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19764 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19765 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19766 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19767 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19768 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19769 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19770 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19771 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19772 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19773 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19774 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19775 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19776 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19777 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19778 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19779 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19780 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19781 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19782 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19783 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19784 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19785 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19786 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19787 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19788 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19789 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19790 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19791 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19792 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19793 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19794 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19795 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19796 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19797 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19798 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19799 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19800 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19801 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19802 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19803 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19804 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19805 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19806 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19807 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19808 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19809 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19810 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19811 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19812 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19813 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19814 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19815 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19816 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19817 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19818 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19819 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19820 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19821 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19822 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19823 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19824 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19825 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19826 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19827 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19828 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19829 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19830 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19831 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19832 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19833 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19834 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19835 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19836 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19837 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19838 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19839 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19840 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19841 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19842 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19843 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19844 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19845 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19846 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19847 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19848 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19849 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19850 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19851 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19852 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19853 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19854 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19855 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19856 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19857 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19858 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19859 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19860 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19861 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19862 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19863 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19864 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19865 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19866 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19867 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19868 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19869 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19870 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19871 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19872 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19873 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19874 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19875 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19876 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19877 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19878 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19879 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19880 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19881 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19882 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19883 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19884 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19885 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19886 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19887 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19888 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19889 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19890 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19891 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19892 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19893 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19894 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19895 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19896 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19897 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19898 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19899 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19900 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19901 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19902 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19903 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19904 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19905 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19906 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19907 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19908 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19909 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19910 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19911 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19912 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655