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 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.9 and 3.0.10 [1 Aug 2023]
33 * Fix excessive time spent checking DH q parameter value.
35 The function DH_check() performs various checks on DH parameters. After
36 fixing CVE-2023-3446 it was discovered that a large q parameter value can
37 also trigger an overly long computation during some of these checks.
38 A correct q value, if present, cannot be larger than the modulus p
39 parameter, thus it is unnecessary to perform these checks if q is larger
42 If DH_check() is called with such q parameter value,
43 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
44 intensive checks are skipped.
50 * Fix DH_check() excessive time with over sized modulus.
52 The function DH_check() performs various checks on DH parameters. One of
53 those checks confirms that the modulus ("p" parameter) is not too large.
54 Trying to use a very large modulus is slow and OpenSSL will not normally use
55 a modulus which is over 10,000 bits in length.
57 However the DH_check() function checks numerous aspects of the key or
58 parameters that have been supplied. Some of those checks use the supplied
59 modulus value even if it has already been found to be too large.
61 A new limit has been added to DH_check of 32,768 bits. Supplying a
62 key/parameters with a modulus over this size will simply cause DH_check() to
69 * Do not ignore empty associated data entries with AES-SIV.
71 The AES-SIV algorithm allows for authentication of multiple associated
72 data entries along with the encryption. To authenticate empty data the
73 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
74 with NULL pointer as the output buffer and 0 as the input buffer length.
75 The AES-SIV implementation in OpenSSL just returns success for such call
76 instead of performing the associated data authentication operation.
77 The empty data thus will not be authenticated. ([CVE-2023-2975])
79 Thanks to Juerg Wullschleger (Google) for discovering the issue.
81 The fix changes the authentication tag value and the ciphertext for
82 applications that use empty associated data entries with AES-SIV.
83 To decrypt data encrypted with previous versions of OpenSSL the application
84 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
89 ### Changes between 3.0.8 and 3.0.9 [30 May 2023]
91 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
92 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
94 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
95 numeric text form. For gigantic sub-identifiers, this would take a very
96 long time, the time complexity being O(n^2) where n is the size of that
97 sub-identifier. ([CVE-2023-2650])
99 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
100 IDENTIFIER to canonical numeric text form if the size of that OBJECT
101 IDENTIFIER is 586 bytes or less, and fail otherwise.
103 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
104 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
105 most 128 sub-identifiers, and that the maximum value that each sub-
106 identifier may have is 2^32-1 (4294967295 decimal).
108 For each byte of every sub-identifier, only the 7 lower bits are part of
109 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
110 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
115 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
116 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
117 trigger a crash of an application using AES-XTS decryption if the memory
118 just after the buffer being decrypted is not mapped.
119 Thanks to Anton Romanov (Amazon) for discovering the issue.
124 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
125 The previous fix for this timing side channel turned out to cause
126 a severe 2-3x performance regression in the typical use case
127 compared to 3.0.7. The new fix uses existing constant time
128 code paths, and restores the previous performance level while
129 fully eliminating all existing timing side channels.
130 The fix was developed by Bernd Edlinger with testing support
135 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
136 that it does not enable policy checking. Thanks to David Benjamin for
137 discovering this issue.
142 * Fixed an issue where invalid certificate policies in leaf certificates are
143 silently ignored by OpenSSL and other certificate policy checks are skipped
144 for that certificate. A malicious CA could use this to deliberately assert
145 invalid certificate policies in order to circumvent policy checking on the
146 certificate altogether.
151 * Limited the number of nodes created in a policy tree to mitigate
152 against CVE-2023-0464. The default limit is set to 1000 nodes, which
153 should be sufficient for most installations. If required, the limit
154 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
155 time define to a desired maximum number of nodes or zero to allow
161 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
163 * Fixed NULL dereference during PKCS7 data verification.
165 A NULL pointer can be dereferenced when signatures are being
166 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
167 algorithm used for the signature is known to the OpenSSL library but
168 the implementation of the hash algorithm is not available the digest
169 initialization will fail. There is a missing check for the return
170 value from the initialization function which later leads to invalid
171 usage of the digest API most likely leading to a crash.
174 PKCS7 data is processed by the SMIME library calls and also by the
175 time stamp (TS) library calls. The TLS implementation in OpenSSL does
176 not call these functions however third party applications would be
177 affected if they call these functions to verify signatures on untrusted
182 * Fixed X.400 address type confusion in X.509 GeneralName.
184 There is a type confusion vulnerability relating to X.400 address processing
185 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
186 but the public structure definition for GENERAL_NAME incorrectly specified
187 the type of the x400Address field as ASN1_TYPE. This field is subsequently
188 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
191 When CRL checking is enabled (i.e. the application sets the
192 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
193 pass arbitrary pointers to a memcmp call, enabling them to read memory
194 contents or enact a denial of service.
199 * Fixed NULL dereference validating DSA public key.
201 An invalid pointer dereference on read can be triggered when an
202 application tries to check a malformed DSA public key by the
203 EVP_PKEY_public_check() function. This will most likely lead
204 to an application crash. This function can be called on public
205 keys supplied from untrusted sources which could allow an attacker
206 to cause a denial of service attack.
208 The TLS implementation in OpenSSL does not call this function
209 but applications might call the function if there are additional
210 security requirements imposed by standards such as FIPS 140-3.
213 *Shane Lontis, Tomáš Mráz*
215 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
217 An invalid pointer dereference on read can be triggered when an
218 application tries to load malformed PKCS7 data with the
219 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
221 The result of the dereference is an application crash which could
222 lead to a denial of service attack. The TLS implementation in OpenSSL
223 does not call this function however third party applications might
224 call these functions on untrusted data.
229 * Fixed Use-after-free following BIO_new_NDEF.
231 The public API function BIO_new_NDEF is a helper function used for
232 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
233 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
234 be called directly by end user applications.
236 The function receives a BIO from the caller, prepends a new BIO_f_asn1
237 filter BIO onto the front of it to form a BIO chain, and then returns
238 the new head of the BIO chain to the caller. Under certain conditions,
239 for example if a CMS recipient public key is invalid, the new filter BIO
240 is freed and the function returns a NULL result indicating a failure.
241 However, in this case, the BIO chain is not properly cleaned up and the
242 BIO passed by the caller still retains internal pointers to the previously
243 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
244 then a use-after-free will occur. This will most likely result in a crash.
247 *Viktor Dukhovni, Matt Caswell*
249 * Fixed Double free after calling PEM_read_bio_ex.
251 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
252 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
253 data. If the function succeeds then the "name_out", "header" and "data"
254 arguments are populated with pointers to buffers containing the relevant
255 decoded data. The caller is responsible for freeing those buffers. It is
256 possible to construct a PEM file that results in 0 bytes of payload data.
257 In this case PEM_read_bio_ex() will return a failure code but will populate
258 the header argument with a pointer to a buffer that has already been freed.
259 If the caller also frees this buffer then a double free will occur. This
260 will most likely lead to a crash.
262 The functions PEM_read_bio() and PEM_read() are simple wrappers around
263 PEM_read_bio_ex() and therefore these functions are also directly affected.
265 These functions are also called indirectly by a number of other OpenSSL
266 functions including PEM_X509_INFO_read_bio_ex() and
267 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
268 internal uses of these functions are not vulnerable because the caller does
269 not free the header argument if PEM_read_bio_ex() returns a failure code.
272 *Kurt Roeckx, Matt Caswell*
274 * Fixed Timing Oracle in RSA Decryption.
276 A timing based side channel exists in the OpenSSL RSA Decryption
277 implementation which could be sufficient to recover a plaintext across
278 a network in a Bleichenbacher style attack. To achieve a successful
279 decryption an attacker would have to be able to send a very large number
280 of trial messages for decryption. The vulnerability affects all RSA padding
281 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
284 *Dmitry Belyavsky, Hubert Kario*
286 * Fixed X.509 Name Constraints Read Buffer Overflow.
288 A read buffer overrun can be triggered in X.509 certificate verification,
289 specifically in name constraint checking. The read buffer overrun might
290 result in a crash which could lead to a denial of service attack.
291 In a TLS client, this can be triggered by connecting to a malicious
292 server. In a TLS server, this can be triggered if the server requests
293 client authentication and a malicious client connects.
298 * Fixed X.509 Policy Constraints Double Locking security issue.
300 If an X.509 certificate contains a malformed policy constraint and
301 policy processing is enabled, then a write lock will be taken twice
302 recursively. On some operating systems (most widely: Windows) this
303 results in a denial of service when the affected process hangs. Policy
304 processing being enabled on a publicly facing server is not considered
305 to be a common setup.
310 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
311 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
312 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
313 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
314 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
315 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
316 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
317 for legacy EC and SM2 keys is also changed similarly to honor the
318 equivalent conversion format flag as specified in the underlying
319 `EC_KEY` object being exported to a provider, when this function is
320 called through `EVP_PKEY_export()`.
324 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
326 * Fixed two buffer overflows in punycode decoding functions.
328 A buffer overrun can be triggered in X.509 certificate verification,
329 specifically in name constraint checking. Note that this occurs after
330 certificate chain signature verification and requires either a CA to
331 have signed the malicious certificate or for the application to continue
332 certificate verification despite failure to construct a path to a trusted
335 In a TLS client, this can be triggered by connecting to a malicious
336 server. In a TLS server, this can be triggered if the server requests
337 client authentication and a malicious client connects.
339 An attacker can craft a malicious email address to overflow
340 an arbitrary number of bytes containing the `.` character (decimal 46)
341 on the stack. This buffer overflow could result in a crash (causing a
345 An attacker can craft a malicious email address to overflow four
346 attacker-controlled bytes on the stack. This buffer overflow could
347 result in a crash (causing a denial of service) or potentially remote code
348 execution depending on stack layout for any given platform/compiler.
353 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
354 parameters in OpenSSL code.
355 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
356 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
357 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
358 Using these invalid names may cause algorithms to use slower methods
359 that ignore the CRT parameters.
363 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
368 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
369 data to be signed before signing the certificate.
373 * Added RIPEMD160 to the default provider.
377 * Ensured that the key share group sent or accepted for the key exchange
378 is allowed for the protocol version.
382 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
384 * OpenSSL supports creating a custom cipher via the legacy
385 EVP_CIPHER_meth_new() function and associated function calls. This function
386 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
387 to use the new provider mechanism in order to implement custom ciphers.
389 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
390 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
391 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
392 and decryption initialisation functions). Instead of using the custom cipher
393 directly it incorrectly tries to fetch an equivalent cipher from the
394 available providers. An equivalent cipher is found based on the NID passed to
395 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
396 given cipher. However it is possible for an application to incorrectly pass
397 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
398 is used in this way the OpenSSL encryption/decryption initialisation function
399 will match the NULL cipher as being equivalent and will fetch this from the
400 available providers. This will succeed if the default provider has been
401 loaded (or if a third party provider has been loaded that offers this
402 cipher). Using the NULL cipher means that the plaintext is emitted as the
405 Applications are only affected by this issue if they call
406 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
407 encryption/decryption initialisation function. Applications that only use
408 SSL/TLS are not impacted by this issue.
413 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
418 * Fixed the linux-mips64 Configure target which was missing the
419 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
424 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
429 * Correctly handle a retransmitted ClientHello in DTLS
433 * Fixed detection of ktls support in cross-compile environment on Linux
437 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
442 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
443 report correct results in some cases
447 * Fix UWP builds by defining VirtualLock
451 * For known safe primes use the minimum key length according to RFC 7919.
452 Longer private key sizes unnecessarily raise the cycles needed to compute the
453 shared secret without any increase of the real security. This fixes a
454 regression from 1.1.1 where these shorter keys were generated for the known
459 * Added the loongarch64 target
463 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
464 only passed to the FIPS provider and not to the default or legacy provider.
468 * Fixed reported performance degradation on aarch64. Restored the
469 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
470 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
471 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
472 The new algorithm is still used for 32 bit targets.
476 * Added a missing header for memcmp that caused compilation failure on some
481 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
483 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
484 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
485 This issue makes the RSA implementation with 2048 bit private keys
486 incorrect on such machines and memory corruption will happen during
487 the computation. As a consequence of the memory corruption an attacker
488 may be able to trigger a remote code execution on the machine performing
491 SSL/TLS servers or other servers using 2048 bit RSA private keys running
492 on machines supporting AVX512IFMA instructions of the X86_64 architecture
493 are affected by this issue.
498 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
499 implementation would not encrypt the entirety of the data under some
500 circumstances. This could reveal sixteen bytes of data that was
501 preexisting in the memory that wasn't written. In the special case of
502 "in place" encryption, sixteen bytes of the plaintext would be revealed.
504 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
505 they are both unaffected.
508 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
510 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
512 * In addition to the c_rehash shell command injection identified in
513 CVE-2022-1292, further bugs where the c_rehash script does not
514 properly sanitise shell metacharacters to prevent command injection have been
517 When the CVE-2022-1292 was fixed it was not discovered that there
518 are other places in the script where the file names of certificates
519 being hashed were possibly passed to a command executed through the shell.
521 This script is distributed by some operating systems in a manner where
522 it is automatically executed. On such operating systems, an attacker
523 could execute arbitrary commands with the privileges of the script.
525 Use of the c_rehash script is considered obsolete and should be replaced
526 by the OpenSSL rehash command line tool.
529 *Daniel Fiala, Tomáš Mráz*
531 * Case insensitive string comparison no longer uses locales. It has instead
532 been directly implemented.
536 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
538 * Case insensitive string comparison is reimplemented via new locale-agnostic
539 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
540 comparison. The previous implementation had problems when the Turkish locale
545 * Fixed a bug in the c_rehash script which was not properly sanitising shell
546 metacharacters to prevent command injection. This script is distributed by
547 some operating systems in a manner where it is automatically executed. On
548 such operating systems, an attacker could execute arbitrary commands with the
549 privileges of the script.
551 Use of the c_rehash script is considered obsolete and should be replaced
552 by the OpenSSL rehash command line tool.
557 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
558 certificate on an OCSP response. The bug caused the function in the case
559 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
560 response (meaning a successful verification) even in the case where the
561 response signing certificate fails to verify.
563 It is anticipated that most users of `OCSP_basic_verify` will not use the
564 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
565 a negative value (indicating a fatal error) in the case of a certificate
566 verification failure. The normal expected return value in this case would be
569 This issue also impacts the command line OpenSSL "ocsp" application. When
570 verifying an ocsp response with the "-no_cert_checks" option the command line
571 application will report that the verification is successful even though it
572 has in fact failed. In this case the incorrect successful response will also
573 be accompanied by error messages showing the failure and contradicting the
574 apparently successful result.
579 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
580 AAD data as the MAC key. This made the MAC key trivially predictable.
582 An attacker could exploit this issue by performing a man-in-the-middle attack
583 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
584 that the modified data would still pass the MAC integrity check.
586 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
587 endpoint will always be rejected by the recipient and the connection will
588 fail at that point. Many application protocols require data to be sent from
589 the client to the server first. Therefore, in such a case, only an OpenSSL
590 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
592 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
593 sent in both directions. In this case both clients and servers could be
594 affected, regardless of the application protocol.
596 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
597 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
598 the handshake when using this ciphersuite.
600 The confidentiality of data is not impacted by this issue, i.e. an attacker
601 cannot decrypt data that has been encrypted using this ciphersuite - they can
604 In order for this attack to work both endpoints must legitimately negotiate
605 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
606 OpenSSL 3.0, and is not available within the default provider or the default
607 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
608 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
609 following must have occurred:
611 1) OpenSSL must have been compiled with the (non-default) compile time option
612 enable-weak-ssl-ciphers
614 2) OpenSSL must have had the legacy provider explicitly loaded (either
615 through application code or via configuration)
617 3) The ciphersuite must have been explicitly added to the ciphersuite list
619 4) The libssl security level must have been set to 0 (default is 1)
621 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
623 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
624 others that both endpoints have in common
629 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
630 occuppied by the removed hash table entries.
632 This function is used when decoding certificates or keys. If a long lived
633 process periodically decodes certificates or keys its memory usage will
634 expand without bounds and the process might be terminated by the operating
635 system causing a denial of service. Also traversing the empty hash table
636 entries will take increasingly more time.
638 Typically such long lived processes might be TLS clients or TLS servers
639 configured to accept client certificate authentication.
642 *Hugo Landau, Aliaksei Levin*
644 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
645 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
646 statistics are no longer supported. For compatibility, these statistics are
647 still listed in the output but are now always reported as zero.
651 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
653 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
654 for non-prime moduli.
656 Internally this function is used when parsing certificates that contain
657 elliptic curve public keys in compressed form or explicit elliptic curve
658 parameters with a base point encoded in compressed form.
660 It is possible to trigger the infinite loop by crafting a certificate that
661 has invalid explicit curve parameters.
663 Since certificate parsing happens prior to verification of the certificate
664 signature, any process that parses an externally supplied certificate may thus
665 be subject to a denial of service attack. The infinite loop can also be
666 reached when parsing crafted private keys as they can contain explicit
667 elliptic curve parameters.
669 Thus vulnerable situations include:
671 - TLS clients consuming server certificates
672 - TLS servers consuming client certificates
673 - Hosting providers taking certificates or private keys from customers
674 - Certificate authorities parsing certification requests from subscribers
675 - Anything else which parses ASN.1 elliptic curve parameters
677 Also any other applications that use the BN_mod_sqrt() where the attacker
678 can control the parameter values are vulnerable to this DoS issue.
683 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
684 to the list of ciphersuites providing Perfect Forward Secrecy as
685 required by SECLEVEL >= 3.
687 *Dmitry Belyavskiy, Nicola Tuveri*
689 * Made the AES constant time code for no-asm configurations
690 optional due to the resulting 95% performance degradation.
691 The AES constant time code can be enabled, for no assembly
692 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
696 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
701 * The negative return value handling of the certificate verification callback
702 was reverted. The replacement is to set the verification retry state with
703 the SSL_set_retry_verify() function.
707 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
709 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
710 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
711 verify a certificate supplied by a server. That function may return a
712 negative return value to indicate an internal error (for example out of
713 memory). Such a negative return value is mishandled by OpenSSL and will cause
714 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
715 success and a subsequent call to SSL_get_error() to return the value
716 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
717 returned by OpenSSL if the application has previously called
718 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
719 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
720 totally unexpected and applications may not behave correctly as a result. The
721 exact behaviour will depend on the application but it could result in
722 crashes, infinite loops or other similar incorrect responses.
724 This issue is made more serious in combination with a separate bug in OpenSSL
725 3.0 that will cause X509_verify_cert() to indicate an internal error when
726 processing a certificate chain. This will occur where a certificate does not
727 include the Subject Alternative Name extension but where a Certificate
728 Authority has enforced name constraints. This issue can occur even with valid
734 * Corrected a few file name and file reference bugs in the build,
735 installation and setup scripts, which lead to installation verification
736 failures. Slightly enhanced the installation verification script.
740 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
745 * Fixed PVK encoder to properly query for the passphrase.
749 * Multiple fixes in the OSSL_HTTP API functions.
753 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
754 OSSL_PARAM_INTEGER data type and return error on negative numbers
755 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
756 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
760 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
764 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
768 * Multiple threading fixes.
772 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
776 * Allow fetching an operation from the provider that owns an unexportable key
777 as a fallback if that is still allowed by the property query.
781 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
783 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
788 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
789 S390X capability vector to zero. This simplifies testing of different code
790 paths on S390X architecture.
794 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
795 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
796 SP 800-38D". The communication will fail at this point.
800 * The EC_GROUP_clear_free() function is deprecated as there is nothing
801 confidential in EC_GROUP data.
805 * The byte order mark (BOM) character is ignored if encountered at the
806 beginning of a PEM-formatted file.
810 * Added CMS support for the Russian GOST algorithms.
814 * Due to move of the implementation of cryptographic operations
815 to the providers, validation of various operation parameters can
816 be postponed until the actual operation is executed where previously
817 it happened immediately when an operation parameter was set.
819 For example when setting an unsupported curve with
820 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
821 fail but later keygen operations with the EVP_PKEY_CTX will fail.
823 *OpenSSL team members and many third party contributors*
825 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
826 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
827 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
828 instead to retrieve these algorithms from a provider.
832 * On build targets where the multilib postfix is set in the build
833 configuration the libdir directory was changing based on whether
834 the lib directory with the multilib postfix exists on the system
835 or not. This unpredictable behavior was removed and eventual
836 multilib postfix is now always added to the default libdir. Use
837 `--libdir=lib` to override the libdir if adding the postfix is
842 * The triple DES key wrap functionality now conforms to RFC 3217 but is
843 no longer interoperable with OpenSSL 1.1.1.
847 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
848 function codes, this function can only cause problems for calling
853 * Add a configurable flag to output date formats as ISO 8601. Does not
854 change the default date format.
858 * Version of MSVC earlier than 1300 could get link warnings, which could
859 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
860 Support for this flag has been removed.
864 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
865 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
866 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
867 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
868 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
872 * The signatures of the functions to get and set options on SSL and
873 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
874 Some source code changes may be required.
878 * The public definitions of conf_method_st and conf_st have been
879 deprecated. They will be made opaque in a future release.
881 *Rich Salz and Tomáš Mráz*
883 * Client-initiated renegotiation is disabled by default. To allow it, use
884 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
885 flag, or the "ClientRenegotiation" config parameter as appropriate.
889 * Add "abspath" and "includedir" pragma's to config files, to prevent,
890 or modify relative pathname inclusion.
894 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
895 validated. Please consult the README-FIPS and
896 README-PROVIDERS files, as well as the migration guide.
898 *OpenSSL team members and many third party contributors*
900 * For the key types DH and DHX the allowed settable parameters are now different.
904 * The openssl commands that read keys, certificates, and CRLs now
905 automatically detect the PEM or DER format of the input files.
907 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
909 * Added enhanced PKCS#12 APIs which accept a library context.
913 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
917 * Added support for Kernel TLS (KTLS).
919 *Boris Pismenny, John Baldwin and Andrew Gallatin*
921 * Support for RFC 5746 secure renegotiation is now required by default for
922 SSL or TLS connections to succeed.
926 * The signature of the `copy` functional parameter of the
927 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
928 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
929 the signature of the `pub_decode` functional parameter of the
930 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
931 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
935 * The error return values from some control calls (ctrl) have changed.
939 * A public key check is now performed during EVP_PKEY_derive_set_peer().
943 * Many functions in the EVP_ namespace that are getters of values from
944 implementations or contexts were renamed to include get or get0 in their
945 names. Old names are provided as macro aliases for compatibility and
950 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
951 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
952 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
957 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
960 * The output from the command line applications may have minor
965 * The output from numerous "printing" may have minor changes.
969 * Windows thread synchronization uses read/write primitives (SRWLock) when
970 supported by the OS, otherwise CriticalSection continues to be used.
974 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
975 work on read only BIO source/sinks that do not support these functions.
976 This allows piping or redirection of a file BIO using stdin to be buffered
977 into memory. This is used internally in OSSL_DECODER_from_bio().
981 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
982 this function would return one of the values OSSL_STORE_INFO_NAME,
983 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
984 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
985 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
986 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
987 using this function should be amended to handle the changed return value.
991 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
992 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
993 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
994 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
995 contains more than one certificate identifier: This means that all
996 certificates referenced there MUST be part of the validation chain.
1000 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1001 RC5, DESX and DES have been moved to the legacy provider.
1005 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1006 RIPEMD-160 have been moved to the legacy provider.
1010 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1015 * The deprecated functions EVP_PKEY_get0_RSA(),
1016 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1017 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1018 well as the similarly named "get1" functions behave differently in
1023 * A number of functions handling low-level keys or engines were deprecated
1024 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1025 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1026 EVP_PKEY_get0_siphash().
1030 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1031 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1032 will need to load the legacy crypto provider. This includes these PBE
1033 algorithms which use this KDF:
1034 - NID_pbeWithMD2AndDES_CBC
1035 - NID_pbeWithMD5AndDES_CBC
1036 - NID_pbeWithSHA1AndRC2_CBC
1037 - NID_pbeWithMD2AndRC2_CBC
1038 - NID_pbeWithMD5AndRC2_CBC
1039 - NID_pbeWithSHA1AndDES_CBC
1043 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1044 BIO_debug_callback() functions.
1048 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1049 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1053 * The RAND_METHOD APIs have been deprecated.
1057 * The SRP APIs have been deprecated.
1061 * Add a compile time option to prevent the caching of provider fetched
1062 algorithms. This is enabled by including the no-cached-fetch option
1063 at configuration time.
1067 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1068 count of PKCS12_DEFAULT_ITER.
1070 *Tomáš Mráz and Sahana Prasad*
1072 * The openssl speed command does not use low-level API calls anymore.
1076 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1079 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1081 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1085 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1086 providers may supply their own group implementations (using either the "key
1087 exchange" or the "key encapsulation" methods) which will automatically be
1088 detected and used by libssl.
1090 *Matt Caswell, Nicola Tuveri*
1092 * The undocumented function X509_certificate_type() has been deprecated;
1096 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1100 * Removed RSA padding mode for SSLv23 (which was only used for
1101 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1102 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1107 * Deprecated the obsolete X9.31 RSA key generation related functions.
1109 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1110 is not allowed to return a value > 1, this is no more taken as failure.
1112 *Viktor Dukhovni and David von Oheimb*
1114 * Deprecated the obsolete X9.31 RSA key generation related functions
1115 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1116 BN_X931_generate_prime_ex().
1120 * The default key generation method for the regular 2-prime RSA keys was
1121 changed to the FIPS 186-4 B.3.6 method.
1125 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1129 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1133 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1134 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1136 *Rich Salz, Richard Levitte, and David von Oheimb*
1138 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1142 * Deprecated `OCSP_parse_url()`.
1146 * Validation of SM2 keys has been separated from the validation of regular EC
1151 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1152 switches: a validation failure triggers an early exit, returning a failure
1153 exit status to the parent process.
1157 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1158 to ignore unknown ciphers.
1162 * The `-cipher-commands` and `-digest-commands` options
1163 of the command line utility `list` have been deprecated.
1164 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1168 * Added convenience functions for generating asymmetric key pairs:
1169 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1170 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1174 * All of the low level EC_KEY functions have been deprecated.
1176 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1178 * Deprecated all the libcrypto and libssl error string loading
1183 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1184 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1189 * The `-crypt` option to the `passwd` command line tool has been removed.
1193 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1198 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1202 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1203 EVP_PKEY_get1_tls_encodedpoint().
1207 * The security callback, which can be customised by application code, supports
1208 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1209 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1213 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1214 interface. Their functionality remains unchanged.
1218 * Added new option for 'openssl list', '-providers', which will display the
1219 list of loaded providers, their names, version and status. It optionally
1220 displays their gettable parameters.
1224 * Removed EVP_PKEY_set_alias_type().
1228 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1229 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1233 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1234 parameters are still checked for type safety at compile time via helper
1239 * Remove the RAND_DRBG API
1241 *Paul Dale and Matthias St. Pierre*
1243 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1244 as well as actual hostnames.
1248 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1249 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1250 conversely, silently ignore DTLS protocol version bounds when configuring
1251 TLS-based contexts. The commands can be repeated to set bounds of both
1252 types. The same applies with the corresponding "min_protocol" and
1253 "max_protocol" command-line switches, in case some application uses both TLS
1256 SSL_CTX instances that are created for a fixed protocol version (e.g.
1257 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1258 attempts to apply bounds to these protocol versions would result in an
1259 error. Now only the "version-flexible" SSL_CTX instances are subject to
1260 limits in configuration files in command-line options.
1264 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1269 * Reworked the recorded ERR codes to make better space for system errors.
1270 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1271 given code is a system error (true) or an OpenSSL error (false).
1275 * Reworked the test perl framework to better allow parallel testing.
1277 *Nicola Tuveri and David von Oheimb*
1279 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1280 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1284 * 'Configure' has been changed to figure out the configuration target if
1285 none is given on the command line. Consequently, the 'config' script is
1286 now only a mere wrapper. All documentation is changed to only mention
1289 *Rich Salz and Richard Levitte*
1291 * Added a library context `OSSL_LIB_CTX` that applications as well as
1292 other libraries can use to form a separate context within which
1293 libcrypto operations are performed.
1297 * Added various `_ex` functions to the OpenSSL API that support using
1298 a non-default `OSSL_LIB_CTX`.
1302 * Handshake now fails if Extended Master Secret extension is dropped
1307 * Dropped interactive mode from the `openssl` program.
1311 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1313 *David von Oheimb and Shane Lontis*
1315 * Deprecated `EC_METHOD_get_field_type()`.
1319 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1320 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1321 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1325 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1329 * Add CAdES-BES signature verification support, mostly derived
1330 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1332 *Filipe Raimundo da Silva*
1334 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1338 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1339 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1343 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1347 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1348 EC_KEY_precompute_mult().
1352 * Deprecated EC_POINTs_mul().
1356 * Removed FIPS_mode() and FIPS_mode_set().
1360 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1364 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1365 EC_POINT_get_Jprojective_coordinates_GFp().
1369 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1370 arrays to be more easily constructed via a series of utility functions.
1371 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1372 the various push functions and finally convert to a passable OSSL_PARAM
1373 array using OSSL_PARAM_BLD_to_param().
1377 * The security strength of SHA1 and MD5 based signatures in TLS has been
1382 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1383 contain a provider side internal key.
1387 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1391 * Project text documents not yet having a proper file name extension
1392 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1393 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1397 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1398 have been converted to Markdown with the goal to produce documents
1399 which not only look pretty when viewed online in the browser, but
1400 remain well readable inside a plain text editor.
1402 To achieve this goal, a 'minimalistic' Markdown style has been applied
1403 which avoids formatting elements that interfere too much with the
1404 reading flow in the text file. For example, it
1406 * avoids [ATX headings][] and uses [setext headings][] instead
1407 (which works for `<h1>` and `<h2>` headings only).
1408 * avoids [inline links][] and uses [reference links][] instead.
1409 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1411 [ATX headings]: https://github.github.com/gfm/#atx-headings
1412 [setext headings]: https://github.github.com/gfm/#setext-headings
1413 [inline links]: https://github.github.com/gfm/#inline-link
1414 [reference links]: https://github.github.com/gfm/#reference-link
1415 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1416 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1418 *Matthias St. Pierre*
1420 * The test suite is changed to preserve results of each test recipe.
1421 A new directory test-runs/ with subdirectories named like the
1422 test recipes are created in the build tree for this purpose.
1426 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1427 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1428 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1430 *David von Oheimb, Martin Peylo*
1432 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1433 It supports arbitrary request and response content types, GET redirection,
1434 TLS, connections via HTTP(S) proxies, connections and exchange via
1435 user-defined BIOs (allowing implicit connections), persistent connections,
1436 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1437 The legacy OCSP-focused (and only partly documented) API
1438 is retained for backward compatibility, while most of it is deprecated.
1442 * Added `util/check-format.pl`, a tool for checking adherence to the
1443 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1444 The checks performed are incomplete and yield some false positives.
1445 Still the tool should be useful for detecting most typical glitches.
1449 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1450 If domain name resolution yields multiple IP addresses all of them are tried
1451 after `connect()` failures.
1455 * All of the low level RSA functions have been deprecated.
1459 * X509 certificates signed using SHA1 are no longer allowed at security
1464 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1465 modified to use PKEY APIs. These commands are now in maintenance mode
1466 and no new features will be added to them.
1470 * The command line utility rsautl has been deprecated.
1474 * The command line utilities genrsa and rsa have been modified to use PKEY
1475 APIs. They now write PKCS#8 keys by default. These commands are now in
1476 maintenance mode and no new features will be added to them.
1480 * All of the low level DH functions have been deprecated.
1482 *Paul Dale and Matt Caswell*
1484 * All of the low level DSA functions have been deprecated.
1488 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1489 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1493 * Deprecated low level ECDH and ECDSA functions.
1497 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1501 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1502 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1503 a new formulation to include all the things it can be used for,
1504 as well as words of caution.
1508 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1512 * All of the low level HMAC functions have been deprecated.
1514 *Paul Dale and David von Oheimb*
1516 * Over two thousand fixes were made to the documentation, including:
1517 - Common options (such as -rand/-writerand, TLS version control, etc)
1518 were refactored and point to newly-enhanced descriptions in openssl.pod.
1519 - Added style conformance for all options (with help from Richard Levitte),
1520 documented all reported missing options, added a CI build to check
1521 that all options are documented and that no unimplemented options
1523 - Documented some internals, such as all use of environment variables.
1524 - Addressed all internal broken L<> references.
1528 * All of the low level CMAC functions have been deprecated.
1532 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1533 functions have been deprecated.
1535 *Paul Dale and David von Oheimb*
1537 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1538 set of functions. The documentation mentioned negative values for some
1539 errors, but this was never the case, so the mention of negative values
1542 Code that followed the documentation and thereby check with something
1543 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1547 * All of the low level cipher functions have been deprecated.
1549 *Matt Caswell and Paul Dale*
1551 * Removed include/openssl/opensslconf.h.in and replaced it with
1552 include/openssl/configuration.h.in, which differs in not including
1553 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1554 was added to include both.
1556 This allows internal hacks where one might need to modify the set
1557 of configured macros, for example this if deprecated symbols are
1558 still supposed to be available internally:
1560 #include <openssl/configuration.h>
1562 #undef OPENSSL_NO_DEPRECATED
1563 #define OPENSSL_SUPPRESS_DEPRECATED
1565 #include <openssl/macros.h>
1567 This should not be used by applications that use the exported
1568 symbols, as that will lead to linking errors.
1572 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1573 used in exponentiation with 512-bit moduli. No EC algorithms are
1574 affected. Analysis suggests that attacks against 2-prime RSA1024,
1575 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1576 difficult to perform and are not believed likely. Attacks against DH512
1577 are considered just feasible. However, for an attack the target would
1578 have to re-use the DH512 private key, which is not recommended anyway.
1579 Also applications directly using the low-level API BN_mod_exp may be
1580 affected if they use BN_FLG_CONSTTIME.
1585 * Most memory-debug features have been deprecated, and the functionality
1586 replaced with no-ops.
1590 * Added documentation for the STACK API.
1594 * Introduced a new method type and API, OSSL_ENCODER, to represent
1595 generic encoders. These do the same sort of job that PEM writers
1596 and d2i functions do, but with support for methods supplied by
1597 providers, and the possibility for providers to support other
1602 * Introduced a new method type and API, OSSL_DECODER, to represent
1603 generic decoders. These do the same sort of job that PEM readers
1604 and i2d functions do, but with support for methods supplied by
1605 providers, and the possibility for providers to support other
1610 * Added a .pragma directive to the syntax of configuration files, to
1611 allow varying behavior in a supported and predictable manner.
1612 Currently added pragma:
1616 This allows dollar signs to be a keyword character unless it's
1617 followed by a opening brace or parenthesis. This is useful for
1618 platforms where dollar signs are commonly used in names, such as
1619 volume names and system directory names on VMS.
1623 * Added functionality to create an EVP_PKEY from user data.
1627 * Change the interpretation of the '--api' configuration option to
1628 mean that this is a desired API compatibility level with no
1629 further meaning. The previous interpretation, that this would
1630 also mean to remove all deprecated symbols up to and including
1631 the given version, no requires that 'no-deprecated' is also used
1632 in the configuration.
1634 When building applications, the desired API compatibility level
1635 can be set with the OPENSSL_API_COMPAT macro like before. For
1636 API compatibility version below 3.0, the old style numerical
1637 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1638 For version 3.0 and on, the value is expected to be the decimal
1639 value calculated from the major and minor version like this:
1641 MAJOR * 10000 + MINOR * 100
1645 -DOPENSSL_API_COMPAT=30000 For 3.0
1646 -DOPENSSL_API_COMPAT=30200 For 3.2
1648 To hide declarations that are deprecated up to and including the
1649 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1650 given when building the application as well.
1654 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1655 access to certificate and CRL stores via URIs and OSSL_STORE
1658 This adds the following functions:
1660 - X509_LOOKUP_store()
1661 - X509_STORE_load_file()
1662 - X509_STORE_load_path()
1663 - X509_STORE_load_store()
1664 - SSL_add_store_cert_subjects_to_stack()
1665 - SSL_CTX_set_default_verify_store()
1666 - SSL_CTX_load_verify_file()
1667 - SSL_CTX_load_verify_dir()
1668 - SSL_CTX_load_verify_store()
1672 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1673 The presence of this system service is determined at run-time.
1677 * Added functionality to create an EVP_PKEY context based on data
1678 for methods from providers. This takes an algorithm name and a
1679 property query string and simply stores them, with the intent
1680 that any operation that uses this context will use those strings
1681 to fetch the needed methods implicitly, thereby making the port
1682 of application written for pre-3.0 OpenSSL easier.
1686 * The undocumented function NCONF_WIN32() has been deprecated; for
1687 conversion details see the HISTORY section of doc/man5/config.pod
1691 * Introduced the new functions EVP_DigestSignInit_ex() and
1692 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1693 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1694 pages for further details.
1698 * Over two thousand fixes were made to the documentation, including:
1699 adding missing command flags, better style conformance, documentation
1702 *Rich Salz, Richard Levitte*
1704 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1705 X25519, X448, Ed25519 and Ed448.
1709 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1714 * Deprecated the public definition of `ERR_STATE` as well as the function
1715 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1720 * Added ERR functionality to give callers access to the stored function
1721 names that have replaced the older function code based functions.
1723 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1724 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1725 ERR_peek_error_all() and ERR_peek_last_error_all().
1727 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1728 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1729 ERR_func_error_string().
1733 * Extended testing to be verbose for failing tests only. The make variables
1734 VERBOSE_FAILURE or VF can be used to enable this:
1736 $ make VF=1 test # Unix
1737 $ mms /macro=(VF=1) test ! OpenVMS
1738 $ nmake VF=1 test # Windows
1742 * Added the `-copy_extensions` option to the `x509` command for use with
1743 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1744 all extensions in the request are copied to the certificate or vice versa.
1746 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1748 * Added the `-copy_extensions` option to the `req` command for use with
1749 `-x509`. When given with the `copy` or `copyall` argument,
1750 all extensions in the certification request are copied to the certificate.
1754 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1755 they generate are by default RFC 5280 compliant in the following sense:
1756 There is a subjectKeyIdentifier extension with a hash value of the public key
1757 and for not self-signed certs there is an authorityKeyIdentifier extension
1758 with a keyIdentifier field or issuer information identifying the signing key.
1759 This is done unless some configuration overrides the new default behavior,
1760 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1764 * Added several checks to `X509_verify_cert()` according to requirements in
1765 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1766 (which may be done by using the CLI option `-x509_strict`):
1767 * The basicConstraints of CA certificates must be marked critical.
1768 * CA certificates must explicitly include the keyUsage extension.
1769 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1770 * The issuer name of any certificate must not be empty.
1771 * The subject name of CA certs, certs with keyUsage crlSign,
1772 and certs without subjectAlternativeName must not be empty.
1773 * If a subjectAlternativeName extension is given it must not be empty.
1774 * The signatureAlgorithm field and the cert signature must be consistent.
1775 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1776 must not be marked critical.
1777 * The authorityKeyIdentifier must be given for X.509v3 certs
1778 unless they are self-signed.
1779 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1783 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1784 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1788 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1789 used even when parsing explicit parameters, when loading a encoded key
1790 or calling `EC_GROUP_new_from_ecpkparameters()`/
1791 `EC_GROUP_new_from_ecparameters()`.
1792 This prevents bypass of security hardening and performance gains,
1793 especially for curves with specialized EC_METHODs.
1794 By default, if a key encoded with explicit parameters is loaded and later
1795 encoded, the output is still encoded with explicit parameters, even if
1796 internally a "named" EC_GROUP is used for computation.
1800 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1801 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1802 NULL. After this change, only the cofactor parameter can be NULL. It also
1803 does some minimal sanity checks on the passed order.
1808 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1809 An attack is simple, if the first CMS_recipientInfo is valid but the
1810 second CMS_recipientInfo is chosen ciphertext. If the second
1811 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1812 encryption key will be replaced by garbage, and the message cannot be
1813 decoded, but if the RSA decryption fails, the correct encryption key is
1814 used and the recipient will not notice the attack.
1815 As a work around for this potential attack the length of the decrypted
1816 key must be equal to the cipher default key length, in case the
1817 certifiate is not given and all recipientInfo are tried out.
1818 The old behaviour can be re-enabled in the CMS code by setting the
1819 CMS_DEBUG_DECRYPT flag.
1823 * Early start up entropy quality from the DEVRANDOM seed source has been
1824 improved for older Linux systems. The RAND subsystem will wait for
1825 /dev/random to be producing output before seeding from /dev/urandom.
1826 The seeded state is stored for future library initialisations using
1827 a system global shared memory segment. The shared memory identifier
1828 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1829 the desired value. The default identifier is 114.
1833 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1834 when primes for RSA keys are computed.
1835 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1836 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1837 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1838 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1839 This avoids possible fingerprinting of newly generated RSA modules.
1843 * Correct the extended master secret constant on EBCDIC systems. Without this
1844 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1845 negotiate EMS will fail. Unfortunately this also means that TLS connections
1846 between EBCDIC systems with this fix, and EBCDIC systems without this
1847 fix will fail if they negotiate EMS.
1851 * Changed the library initialisation so that the config file is now loaded
1852 by default. This was already the case for libssl. It now occurs for both
1853 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1854 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1858 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1859 where the former acts as a replacement for `ERR_put_error()`, and the
1860 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1861 `ERR_raise_data()` adds more flexibility by taking a format string and
1862 an arbitrary number of arguments following it, to be processed with
1867 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1868 to check if a named provider is loaded and available. When called, it
1869 will also activate all fallback providers if such are still present.
1873 * Enforce a minimum DH modulus size of 512 bits.
1877 * Changed DH parameters to generate the order q subgroup instead of 2q.
1878 Previously generated DH parameters are still accepted by DH_check
1879 but DH_generate_key works around that by clearing bit 0 of the
1880 private key for those. This avoids leaking bit 0 of the private key.
1884 * Significantly reduce secure memory usage by the randomness pools.
1888 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1893 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1894 algorithms. An implementation of a key exchange algorithm can be obtained
1895 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1896 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1897 the older EVP_PKEY_derive_init() function. See the man pages for the new
1898 functions for further details.
1902 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1906 * Removed the function names from error messages and deprecated the
1911 * Removed NextStep support and the macro OPENSSL_UNISTD
1915 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1916 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1917 Also removed "export var as function" capability; we do not export
1918 variables, only functions.
1922 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1923 an error and 1 indicating success. In previous versions of OpenSSL this
1924 was a void type. If a key was set longer than the maximum possible this
1929 * Support SM2 signing and verification schemes with X509 certificate.
1933 * Use SHA256 as the default digest for TS query in the `ts` app.
1937 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1941 * Default cipher lists/suites are now available via a function, the
1942 #defines are deprecated.
1946 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1947 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1948 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1952 * Join the directories crypto/x509 and crypto/x509v3
1956 * Added command 'openssl kdf' that uses the EVP_KDF API.
1960 * Added command 'openssl mac' that uses the EVP_MAC API.
1964 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1965 as default directories. Also added the command 'openssl info'
1966 for scripting purposes.
1970 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1975 * Add prediction resistance to the DRBG reseeding process.
1979 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1980 mandated by IEEE Std 1619-2018.
1984 * Added newline escaping functionality to a filename when using openssl dgst.
1985 This output format is to replicate the output format found in the `*sum`
1986 checksum programs. This aims to preserve backward compatibility.
1988 *Matt Eaton, Richard Levitte, and Paul Dale*
1990 * Removed the heartbeat message in DTLS feature, as it has very
1991 little usage and doesn't seem to fulfill a valuable purpose.
1992 The configuration option is now deprecated.
1996 * Changed the output of 'openssl {digestname} < file' to display the
1997 digest name in its output.
2001 * Added a new generic trace API which provides support for enabling
2002 instrumentation through trace output.
2004 *Richard Levitte & Matthias St. Pierre*
2006 * Added build tests for C++. These are generated files that only do one
2007 thing, to include one public OpenSSL head file each. This tests that
2008 the public header files can be usefully included in a C++ application.
2010 This test isn't enabled by default. It can be enabled with the option
2011 'enable-buildtest-c++'.
2015 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2019 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2023 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2027 * Added KMAC to EVP_MAC.
2031 * Added property based algorithm implementation selection framework to
2036 * Added SCA hardening for modular field inversion in EC_GROUP through
2037 a new dedicated field_inv() pointer in EC_METHOD.
2038 This also addresses a leakage affecting conversions from projective
2039 to affine coordinates.
2041 *Billy Bob Brumley, Nicola Tuveri*
2043 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2044 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2045 those algorithms that were already supported through the EVP_PKEY API
2046 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2047 and scrypt are now wrappers that call EVP_KDF.
2051 * Build devcrypto engine as a dynamic engine.
2053 *Eneas U de Queiroz*
2055 * Add keyed BLAKE2 to EVP_MAC.
2059 * Fix a bug in the computation of the endpoint-pair shared secret used
2060 by DTLS over SCTP. This breaks interoperability with older versions
2061 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2062 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2063 interoperability with such broken implementations. However, enabling
2064 this switch breaks interoperability with correct implementations.
2066 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2067 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2071 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2075 * Changed the license to the Apache License v2.0.
2079 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2081 - Major releases (indicated by incrementing the MAJOR release number)
2082 may introduce incompatible API/ABI changes.
2083 - Minor releases (indicated by incrementing the MINOR release number)
2084 may introduce new features but retain API/ABI compatibility.
2085 - Patch releases (indicated by incrementing the PATCH number)
2086 are intended for bug fixes and other improvements of existing
2087 features only (like improving performance or adding documentation)
2088 and retain API/ABI compatibility.
2092 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2096 * Remove the 'dist' target and add a tarball building script. The
2097 'dist' target has fallen out of use, and it shouldn't be
2098 necessary to configure just to create a source distribution.
2102 * Recreate the OS390-Unix config target. It no longer relies on a
2103 special script like it did for OpenSSL pre-1.1.0.
2107 * Instead of having the source directories listed in Configure, add
2108 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2113 * Add GMAC to EVP_MAC.
2117 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2121 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2122 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2123 to facilitate the continued use of MACs through raw private keys in
2124 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2128 * Deprecate ECDH_KDF_X9_62().
2132 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2133 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2134 are retained for backwards compatibility.
2138 * AES-XTS mode now enforces that its two keys are different to mitigate
2139 the attacked described in "Efficient Instantiations of Tweakable
2140 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2141 Details of this attack can be obtained from:
2142 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2146 * Rename the object files, i.e. give them other names than in previous
2147 versions. Their names now include the name of the final product, as
2148 well as its type mnemonic (bin, lib, shlib).
2152 * Added new option for 'openssl list', '-objects', which will display the
2153 list of built in objects, i.e. OIDs with names.
2157 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2158 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2163 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2164 improves application performance by removing data copies and providing
2165 applications with zero-copy system calls such as sendfile and splice.
2169 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2173 * `PKCS12_parse` now maintains the order of the parsed certificates
2174 when outputting them via `*ca` (rather than reversing it).
2178 * Deprecated pthread fork support methods.
2182 * Added support for FFDHE key exchange in TLS 1.3.
2186 * Added a new concept for OpenSSL plugability: providers. This
2187 functionality is designed to replace the ENGINE API and ENGINE
2188 implementations, and to be much more dynamic, allowing provider
2189 authors to introduce new algorithms among other things, as long as
2190 there's an API that supports the algorithm type.
2192 With this concept comes a new core API for interaction between
2193 libcrypto and provider implementations. Public libcrypto functions
2194 that want to use providers do so through this core API.
2196 The main documentation for this core API is found in
2197 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2198 refer to other manuals describing the API specific for supported
2199 algorithm types (also called operations).
2206 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2208 * Avoid loading of a dynamic engine twice.
2212 * Prioritise DANE TLSA issuer certs over peer certs
2216 * Fixed random API for MacOS prior to 10.12
2218 These MacOS versions don't support the CommonCrypto APIs
2222 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2224 * Fixed an SM2 Decryption Buffer Overflow.
2226 In order to decrypt SM2 encrypted data an application is expected to
2227 call the API function EVP_PKEY_decrypt(). Typically an application will
2228 call this function twice. The first time, on entry, the "out" parameter
2229 can be NULL and, on exit, the "outlen" parameter is populated with the
2230 buffer size required to hold the decrypted plaintext. The application
2231 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2232 again, but this time passing a non-NULL value for the "out" parameter.
2234 A bug in the implementation of the SM2 decryption code means that the
2235 calculation of the buffer size required to hold the plaintext returned
2236 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2237 size required by the second call. This can lead to a buffer overflow
2238 when EVP_PKEY_decrypt() is called by the application a second time with
2239 a buffer that is too small.
2241 A malicious attacker who is able present SM2 content for decryption to
2242 an application could cause attacker chosen data to overflow the buffer
2243 by up to a maximum of 62 bytes altering the contents of other data held
2244 after the buffer, possibly changing application behaviour or causing
2245 the application to crash. The location of the buffer is application
2246 dependent but is typically heap allocated.
2251 * Fixed various read buffer overruns processing ASN.1 strings
2253 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2254 structure which contains a buffer holding the string data and a field
2255 holding the buffer length. This contrasts with normal C strings which
2256 are repesented as a buffer for the string data which is terminated
2257 with a NUL (0) byte.
2259 Although not a strict requirement, ASN.1 strings that are parsed using
2260 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2261 well as any string whose value has been set with the ASN1_STRING_set()
2262 function will additionally NUL terminate the byte array in the
2263 ASN1_STRING structure.
2265 However, it is possible for applications to directly construct valid
2266 ASN1_STRING structures which do not NUL terminate the byte array by
2267 directly setting the "data" and "length" fields in the ASN1_STRING
2268 array. This can also happen by using the ASN1_STRING_set0() function.
2270 Numerous OpenSSL functions that print ASN.1 data have been found to
2271 assume that the ASN1_STRING byte array will be NUL terminated, even
2272 though this is not guaranteed for strings that have been directly
2273 constructed. Where an application requests an ASN.1 structure to be
2274 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2275 been directly constructed by the application without NUL terminating
2276 the "data" field, then a read buffer overrun can occur.
2278 The same thing can also occur during name constraints processing
2279 of certificates (for example if a certificate has been directly
2280 constructed by the application instead of loading it via the OpenSSL
2281 parsing functions, and the certificate contains non NUL terminated
2282 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2283 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2285 If a malicious actor can cause an application to directly construct an
2286 ASN1_STRING and then process it through one of the affected OpenSSL
2287 functions then this issue could be hit. This might result in a crash
2288 (causing a Denial of Service attack). It could also result in the
2289 disclosure of private memory contents (such as private keys, or
2290 sensitive plaintext).
2295 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2297 * Fixed a problem with verifying a certificate chain when using the
2298 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2299 the certificates present in a certificate chain. It is not set by default.
2301 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2302 the chain that have explicitly encoded elliptic curve parameters was added
2303 as an additional strict check.
2305 An error in the implementation of this check meant that the result of a
2306 previous check to confirm that certificates in the chain are valid CA
2307 certificates was overwritten. This effectively bypasses the check
2308 that non-CA certificates must not be able to issue other certificates.
2310 If a "purpose" has been configured then there is a subsequent opportunity
2311 for checks that the certificate is a valid CA. All of the named "purpose"
2312 values implemented in libcrypto perform this check. Therefore, where
2313 a purpose is set the certificate chain will still be rejected even when the
2314 strict flag has been used. A purpose is set by default in libssl client and
2315 server certificate verification routines, but it can be overridden or
2316 removed by an application.
2318 In order to be affected, an application must explicitly set the
2319 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2320 for the certificate verification or, in the case of TLS client or server
2321 applications, override the default purpose.
2326 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2327 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2328 renegotiation ClientHello omits the signature_algorithms extension (where it
2329 was present in the initial ClientHello), but includes a
2330 signature_algorithms_cert extension then a NULL pointer dereference will
2331 result, leading to a crash and a denial of service attack.
2333 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2334 (which is the default configuration). OpenSSL TLS clients are not impacted by
2338 *Peter Kästle and Samuel Sapalski*
2340 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2342 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2343 create a unique hash value based on the issuer and serial number data
2344 contained within an X509 certificate. However it was failing to correctly
2345 handle any errors that may occur while parsing the issuer field (which might
2346 occur if the issuer field is maliciously constructed). This may subsequently
2347 result in a NULL pointer deref and a crash leading to a potential denial of
2353 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2354 padding mode to correctly check for rollback attacks. This is considered a
2355 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2360 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2361 functions. Previously they could overflow the output length argument in some
2362 cases where the input length is close to the maximum permissable length for
2363 an integer on the platform. In such cases the return value from the function
2364 call would be 1 (indicating success), but the output length value would be
2365 negative. This could cause applications to behave incorrectly or crash.
2370 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2371 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2372 could be exploited in a side channel attack to recover the password. Since
2373 the attack is local host only this is outside of the current OpenSSL
2374 threat model and therefore no CVE is assigned.
2376 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2381 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2383 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2384 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2385 If an attacker can control both items being compared then this could lead
2386 to a possible denial of service attack. OpenSSL itself uses the
2387 GENERAL_NAME_cmp function for two purposes:
2388 1) Comparing CRL distribution point names between an available CRL and a
2389 CRL distribution point embedded in an X509 certificate
2390 2) When verifying that a timestamp response token signer matches the
2391 timestamp authority name (exposed via the API functions
2392 TS_RESP_verify_response and TS_RESP_verify_token)
2397 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2399 * Certificates with explicit curve parameters are now disallowed in
2400 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2404 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2405 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2406 conversely, silently ignore DTLS protocol version bounds when configuring
2407 TLS-based contexts. The commands can be repeated to set bounds of both
2408 types. The same applies with the corresponding "min_protocol" and
2409 "max_protocol" command-line switches, in case some application uses both TLS
2412 SSL_CTX instances that are created for a fixed protocol version (e.g.
2413 TLSv1_server_method()) also silently ignore version bounds. Previously
2414 attempts to apply bounds to these protocol versions would result in an
2415 error. Now only the "version-flexible" SSL_CTX instances are subject to
2416 limits in configuration files in command-line options.
2420 * Handshake now fails if Extended Master Secret extension is dropped
2425 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2427 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2429 * Fixed segmentation fault in SSL_check_chain()
2430 Server or client applications that call the SSL_check_chain() function
2431 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2432 dereference as a result of incorrect handling of the
2433 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2434 or unrecognised signature algorithm is received from the peer. This could
2435 be exploited by a malicious peer in a Denial of Service attack.
2440 * Added AES consttime code for no-asm configurations
2441 an optional constant time support for AES was added
2442 when building openssl for no-asm.
2443 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2444 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2445 At this time this feature is by default disabled.
2446 It will be enabled by default in 3.0.
2450 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2452 * Revert the change of EOF detection while reading in libssl to avoid
2453 regressions in applications depending on the current way of reporting
2454 the EOF. As the existing method is not fully accurate the change to
2455 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2456 branch and will be present in the 3.0 release.
2460 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2461 when primes for RSA keys are computed.
2462 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2463 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2464 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2465 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2466 This avoids possible fingerprinting of newly generated RSA modules.
2470 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2472 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2473 while reading in libssl then we would report an error back to the
2474 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2475 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2476 therefore give a hint as to what went wrong.
2480 * Check that ed25519 and ed448 are allowed by the security level. Previously
2481 signature algorithms not using an MD were not being checked that they were
2482 allowed by the security level.
2486 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2487 was not quite right. The behaviour was not consistent between resumption
2488 and normal handshakes, and also not quite consistent with historical
2489 behaviour. The behaviour in various scenarios has been clarified and
2490 it has been updated to make it match historical behaviour as closely as
2495 * *[VMS only]* The header files that the VMS compilers include automatically,
2496 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2497 that the C++ compiler doesn't understand. This is a shortcoming in the
2498 compiler, but can be worked around with `__cplusplus` guards.
2500 C++ applications that use OpenSSL libraries must be compiled using the
2501 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2502 functions. Otherwise, only functions with symbols of less than 31
2503 characters can be used, as the linker will not be able to successfully
2504 resolve symbols with longer names.
2508 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2509 The presence of this system service is determined at run-time.
2513 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2518 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2520 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2521 number generator (RNG). This was intended to include protection in the
2522 event of a fork() system call in order to ensure that the parent and child
2523 processes did not share the same RNG state. However this protection was not
2524 being used in the default case.
2526 A partial mitigation for this issue is that the output from a high
2527 precision timer is mixed into the RNG state so the likelihood of a parent
2528 and child process sharing state is significantly reduced.
2530 If an application already calls OPENSSL_init_crypto() explicitly using
2531 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2534 *Matthias St. Pierre*
2536 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2537 used even when parsing explicit parameters, when loading a encoded key
2538 or calling `EC_GROUP_new_from_ecpkparameters()`/
2539 `EC_GROUP_new_from_ecparameters()`.
2540 This prevents bypass of security hardening and performance gains,
2541 especially for curves with specialized EC_METHODs.
2542 By default, if a key encoded with explicit parameters is loaded and later
2543 encoded, the output is still encoded with explicit parameters, even if
2544 internally a "named" EC_GROUP is used for computation.
2548 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2549 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2550 NULL. After this change, only the cofactor parameter can be NULL. It also
2551 does some minimal sanity checks on the passed order.
2556 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2557 An attack is simple, if the first CMS_recipientInfo is valid but the
2558 second CMS_recipientInfo is chosen ciphertext. If the second
2559 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2560 encryption key will be replaced by garbage, and the message cannot be
2561 decoded, but if the RSA decryption fails, the correct encryption key is
2562 used and the recipient will not notice the attack.
2563 As a work around for this potential attack the length of the decrypted
2564 key must be equal to the cipher default key length, in case the
2565 certifiate is not given and all recipientInfo are tried out.
2566 The old behaviour can be re-enabled in the CMS code by setting the
2567 CMS_DEBUG_DECRYPT flag.
2572 * Early start up entropy quality from the DEVRANDOM seed source has been
2573 improved for older Linux systems. The RAND subsystem will wait for
2574 /dev/random to be producing output before seeding from /dev/urandom.
2575 The seeded state is stored for future library initialisations using
2576 a system global shared memory segment. The shared memory identifier
2577 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2578 the desired value. The default identifier is 114.
2582 * Correct the extended master secret constant on EBCDIC systems. Without this
2583 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2584 negotiate EMS will fail. Unfortunately this also means that TLS connections
2585 between EBCDIC systems with this fix, and EBCDIC systems without this
2586 fix will fail if they negotiate EMS.
2590 * Use Windows installation paths in the mingw builds
2592 Mingw isn't a POSIX environment per se, which means that Windows
2593 paths should be used for installation.
2598 * Changed DH_check to accept parameters with order q and 2q subgroups.
2599 With order 2q subgroups the bit 0 of the private key is not secret
2600 but DH_generate_key works around that by clearing bit 0 of the
2601 private key for those. This avoids leaking bit 0 of the private key.
2605 * Significantly reduce secure memory usage by the randomness pools.
2609 * Revert the DEVRANDOM_WAIT feature for Linux systems
2611 The DEVRANDOM_WAIT feature added a select() call to wait for the
2612 /dev/random device to become readable before reading from the
2613 /dev/urandom device.
2615 It turned out that this change had negative side effects on
2616 performance which were not acceptable. After some discussion it
2617 was decided to revert this feature and leave it up to the OS
2618 resp. the platform maintainer to ensure a proper initialization
2619 during early boot time.
2621 *Matthias St. Pierre*
2623 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2625 * Add build tests for C++. These are generated files that only do one
2626 thing, to include one public OpenSSL head file each. This tests that
2627 the public header files can be usefully included in a C++ application.
2629 This test isn't enabled by default. It can be enabled with the option
2630 'enable-buildtest-c++'.
2634 * Enable SHA3 pre-hashing for ECDSA and DSA.
2638 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2639 This changes the size when using the `genpkey` command when no size is given.
2640 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2641 generation commands to use 2048 bits by default.
2645 * Reorganize the manual pages to consistently have RETURN VALUES,
2646 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2647 util/fix-doc-nits accordingly.
2649 *Paul Yang, Joshua Lock*
2651 * Add the missing accessor EVP_PKEY_get0_engine()
2655 * Have commands like `s_client` and `s_server` output the signature scheme
2656 along with other cipher suite parameters when debugging.
2660 * Make OPENSSL_config() error agnostic again.
2664 * Do the error handling in RSA decryption constant time.
2668 * Prevent over long nonces in ChaCha20-Poly1305.
2670 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2671 for every encryption operation. RFC 7539 specifies that the nonce value
2672 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2673 and front pads the nonce with 0 bytes if it is less than 12
2674 bytes. However it also incorrectly allows a nonce to be set of up to 16
2675 bytes. In this case only the last 12 bytes are significant and any
2676 additional leading bytes are ignored.
2678 It is a requirement of using this cipher that nonce values are
2679 unique. Messages encrypted using a reused nonce value are susceptible to
2680 serious confidentiality and integrity attacks. If an application changes
2681 the default nonce length to be longer than 12 bytes and then makes a
2682 change to the leading bytes of the nonce expecting the new value to be a
2683 new unique nonce then such an application could inadvertently encrypt
2684 messages with a reused nonce.
2686 Additionally the ignored bytes in a long nonce are not covered by the
2687 integrity guarantee of this cipher. Any application that relies on the
2688 integrity of these ignored leading bytes of a long nonce may be further
2689 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2690 is safe because no such use sets such a long nonce value. However user
2691 applications that use this cipher directly and set a non-default nonce
2692 length to be longer than 12 bytes may be vulnerable.
2694 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2700 * Add DEVRANDOM_WAIT feature for Linux systems
2702 On older Linux systems where the getrandom() system call is not available,
2703 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2704 Contrary to getrandom(), the /dev/urandom device will not block during
2705 early boot when the kernel CSPRNG has not been seeded yet.
2707 To mitigate this known weakness, use select() to wait for /dev/random to
2708 become readable before reading from /dev/urandom.
2710 * Ensure that SM2 only uses SM3 as digest algorithm
2714 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2716 * Change the info callback signals for the start and end of a post-handshake
2717 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2718 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2719 confused by this and assume that a TLSv1.2 renegotiation has started. This
2720 can break KeyUpdate handling. Instead we no longer signal the start and end
2721 of a post handshake message exchange (although the messages themselves are
2722 still signalled). This could break some applications that were expecting
2723 the old signals. However without this KeyUpdate is not usable for many
2728 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2730 * Timing vulnerability in DSA signature generation
2732 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2733 timing side channel attack. An attacker could use variations in the signing
2734 algorithm to recover the private key.
2736 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2741 * Timing vulnerability in ECDSA signature generation
2743 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2744 timing side channel attack. An attacker could use variations in the signing
2745 algorithm to recover the private key.
2747 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2752 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2753 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2754 of two gigabytes and the error handling improved.
2756 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2757 categorized as a normal bug, not a security issue, because the DRBG reseeds
2758 automatically and is fully functional even without additional randomness
2759 provided by the application.
2761 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2763 * Add a new ClientHello callback. Provides a callback interface that gives
2764 the application the ability to adjust the nascent SSL object at the
2765 earliest stage of ClientHello processing, immediately after extensions have
2766 been collected but before they have been processed. In particular, this
2767 callback can adjust the supported TLS versions in response to the contents
2772 * Add SM2 base algorithm support.
2776 * s390x assembly pack: add (improved) hardware-support for the following
2777 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2778 aes-cfb/cfb8, aes-ecb.
2782 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2783 parameter is no longer accepted, as it leads to a corrupt table. NULL
2784 pem_str is reserved for alias entries only.
2788 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2789 step for prime curves. The new implementation is based on formulae from
2790 differential addition-and-doubling in homogeneous projective coordinates
2791 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2792 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2793 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2794 to work in projective coordinates.
2796 *Billy Bob Brumley, Nicola Tuveri*
2798 * Change generating and checking of primes so that the error rate of not
2799 being prime depends on the intended use based on the size of the input.
2800 For larger primes this will result in more rounds of Miller-Rabin.
2801 The maximal error rate for primes with more than 1080 bits is lowered
2804 *Kurt Roeckx, Annie Yousar*
2806 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2810 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2811 moving between systems, and to avoid confusion when a Windows build is
2812 done with mingw vs with MSVC. For POSIX installs, there's still a
2813 symlink or copy named 'tsget' to avoid that confusion as well.
2817 * Revert blinding in ECDSA sign and instead make problematic addition
2818 length-invariant. Switch even to fixed-length Montgomery multiplication.
2822 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2823 step for binary curves. The new implementation is based on formulae from
2824 differential addition-and-doubling in mixed Lopez-Dahab projective
2825 coordinates, modified to independently blind the operands.
2827 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2829 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2830 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2831 EC_METHODs to implement their own specialized "ladder step", to take
2832 advantage of more favorable coordinate systems or more efficient
2833 differential addition-and-doubling algorithms.
2835 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2837 * Modified the random device based seed sources to keep the relevant
2838 file descriptors open rather than reopening them on each access.
2839 This allows such sources to operate in a chroot() jail without
2840 the associated device nodes being available. This behaviour can be
2841 controlled using RAND_keep_random_devices_open().
2845 * Numerous side-channel attack mitigations have been applied. This may have
2846 performance impacts for some algorithms for the benefit of improved
2847 security. Specific changes are noted in this change log by their respective
2852 * AIX shared library support overhaul. Switch to AIX "natural" way of
2853 handling shared libraries, which means collecting shared objects of
2854 different versions and bitnesses in one common archive. This allows to
2855 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2856 doesn't affect the way 3rd party applications are linked, only how
2857 multi-version installation is managed.
2861 * Make ec_group_do_inverse_ord() more robust and available to other
2862 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2863 mitigations are applied to the fallback BN_mod_inverse().
2864 When using this function rather than BN_mod_inverse() directly, new
2865 EC cryptosystem implementations are then safer-by-default.
2869 * Add coordinate blinding for EC_POINT and implement projective
2870 coordinate blinding for generic prime curves as a countermeasure to
2871 chosen point SCA attacks.
2873 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2875 * Add blinding to ECDSA and DSA signatures to protect against side channel
2876 attacks discovered by Keegan Ryan (NCC Group).
2880 * Enforce checking in the `pkeyutl` command to ensure that the input
2881 length does not exceed the maximum supported digest length when performing
2882 a sign, verify or verifyrecover operation.
2886 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2887 I/O in combination with something like select() or poll() will hang. This
2888 can be turned off again using SSL_CTX_clear_mode().
2889 Many applications do not properly handle non-application data records, and
2890 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2891 around the problems in those applications, but can also break some.
2892 It's recommended to read the manpages about SSL_read(), SSL_write(),
2893 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2894 SSL_CTX_set_read_ahead() again.
2898 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2899 now allow empty (zero character) pass phrases.
2903 * Apply blinding to binary field modular inversion and remove patent
2904 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2908 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2909 binary and prime elliptic curves.
2913 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2914 constant time fixed point multiplication.
2918 * Revise elliptic curve scalar multiplication with timing attack
2919 defenses: ec_wNAF_mul redirects to a constant time implementation
2920 when computing fixed point and variable point multiplication (which
2921 in OpenSSL are mostly used with secret scalars in keygen, sign,
2922 ECDH derive operations).
2923 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2926 * Updated CONTRIBUTING
2930 * Updated DRBG / RAND to request nonce and additional low entropy
2931 randomness from the system.
2933 *Matthias St. Pierre*
2935 * Updated 'openssl rehash' to use OpenSSL consistent default.
2939 * Moved the load of the ssl_conf module to libcrypto, which helps
2940 loading engines that libssl uses before libssl is initialised.
2944 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2948 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2950 *Ingo Schwarze, Rich Salz*
2952 * Added output of accepting IP address and port for 'openssl s_server'
2956 * Added a new API for TLSv1.3 ciphersuites:
2957 SSL_CTX_set_ciphersuites()
2958 SSL_set_ciphersuites()
2962 * Memory allocation failures consistently add an error to the error
2967 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2968 in libcrypto when run as setuid/setgid.
2972 * Load any config file by default when libssl is used.
2976 * Added new public header file <openssl/rand_drbg.h> and documentation
2977 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2979 *Matthias St. Pierre*
2981 * QNX support removed (cannot find contributors to get their approval
2982 for the license change).
2986 * TLSv1.3 replay protection for early data has been implemented. See the
2987 SSL_read_early_data() man page for further details.
2991 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2992 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2993 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2994 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2995 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2996 configuration has been separated out. See the ciphers man page or the
2997 SSL_CTX_set_ciphersuites() man page for more information.
3001 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3002 in responder mode now supports the new "-multi" option, which
3003 spawns the specified number of child processes to handle OCSP
3004 requests. The "-timeout" option now also limits the OCSP
3005 responder's patience to wait to receive the full client request
3006 on a newly accepted connection. Child processes are respawned
3007 as needed, and the CA index file is automatically reloaded
3008 when changed. This makes it possible to run the "ocsp" responder
3009 as a long-running service, making the OpenSSL CA somewhat more
3010 feature-complete. In this mode, most diagnostic messages logged
3011 after entering the event loop are logged via syslog(3) rather than
3016 * Added support for X448 and Ed448. Heavily based on original work by
3021 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3022 objects loaded. This adds the functions OSSL_STORE_expect() and
3023 OSSL_STORE_find() as well as needed tools to construct searches and
3024 get the search data out of them.
3028 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3029 version of OpenSSL should review their configuration settings to ensure
3030 that they are still appropriate for TLSv1.3. For further information see:
3031 <https://wiki.openssl.org/index.php/TLS1.3>
3035 * Grand redesign of the OpenSSL random generator
3037 The default RAND method now utilizes an AES-CTR DRBG according to
3038 NIST standard SP 800-90Ar1. The new random generator is essentially
3039 a port of the default random generator from the OpenSSL FIPS 2.0
3040 object module. It is a hybrid deterministic random bit generator
3041 using an AES-CTR bit stream and which seeds and reseeds itself
3042 automatically using trusted system entropy sources.
3044 Some of its new features are:
3045 - Support for multiple DRBG instances with seed chaining.
3046 - The default RAND method makes use of a DRBG.
3047 - There is a public and private DRBG instance.
3048 - The DRBG instances are fork-safe.
3049 - Keep all global DRBG instances on the secure heap if it is enabled.
3050 - The public and private DRBG instance are per thread for lock free
3053 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3055 * Changed Configure so it only says what it does and doesn't dump
3056 so much data. Instead, ./configdata.pm should be used as a script
3057 to display all sorts of configuration data.
3061 * Added processing of "make variables" to Configure.
3065 * Added SHA512/224 and SHA512/256 algorithm support.
3069 * The last traces of Netware support, first removed in 1.1.0, have
3074 * Get rid of Makefile.shared, and in the process, make the processing
3075 of certain files (rc.obj, or the .def/.map/.opt files produced from
3076 the ordinal files) more visible and hopefully easier to trace and
3077 debug (or make silent).
3081 * Make it possible to have environment variable assignments as
3082 arguments to config / Configure.
3086 * Add multi-prime RSA (RFC 8017) support.
3090 * Add SM3 implemented according to GB/T 32905-2016
3091 *Jack Lloyd <jack.lloyd@ribose.com>,*
3092 *Ronald Tse <ronald.tse@ribose.com>,*
3093 *Erick Borsboom <erick.borsboom@ribose.com>*
3095 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3096 as documented in RFC6066.
3097 Based on a patch from Tomasz Moń
3099 *Filipe Raimundo da Silva*
3101 * Add SM4 implemented according to GB/T 32907-2016.
3102 *Jack Lloyd <jack.lloyd@ribose.com>,*
3103 *Ronald Tse <ronald.tse@ribose.com>,*
3104 *Erick Borsboom <erick.borsboom@ribose.com>*
3106 * Reimplement -newreq-nodes and ERR_error_string_n; the
3107 original author does not agree with the license change.
3111 * Add ARIA AEAD TLS support.
3115 * Some macro definitions to support VS6 have been removed. Visual
3116 Studio 6 has not worked since 1.1.0
3120 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3121 without clearing the errors.
3125 * Add "atfork" functions. If building on a system that without
3126 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3127 requirements. The RAND facility now uses/requires this.
3135 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3136 not possible to disable entirely. However, it's still possible to
3137 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3140 To disable, configure with 'no-ui-console'. 'no-ui' is still
3141 possible to use as an alias. Check at compile time with the
3142 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3143 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3147 * Add a STORE module, which implements a uniform and URI based reader of
3148 stores that can contain keys, certificates, CRLs and numerous other
3149 objects. The main API is loosely based on a few stdio functions,
3150 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3151 OSSL_STORE_error and OSSL_STORE_close.
3152 The implementation uses backends called "loaders" to implement arbitrary
3153 URI schemes. There is one built in "loader" for the 'file' scheme.
3157 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3158 then adjusted to work on FreeBSD 8.4 as well.
3159 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3160 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3164 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3165 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3166 error code calls like this:
3168 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3170 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3171 that can be encoded in C. For the foreseeable future, this will only
3174 *Richard Levitte and Tim Hudson*
3176 * Removed BSD cryptodev engine.
3180 * Add a build target 'build_all_generated', to build all generated files
3181 and only that. This can be used to prepare everything that requires
3182 things like perl for a system that lacks perl and then move everything
3183 to that system and do the rest of the build there.
3187 * In the UI interface, make it possible to duplicate the user data. This
3188 can be used by engines that need to retain the data for a longer time
3189 than just the call where this user data is passed.
3193 * Ignore the '-named_curve auto' value for compatibility of applications
3196 *Tomáš Mráz <tmraz@fedoraproject.org>*
3198 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3199 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3200 alerts across multiple records (some of which could be empty). In practice
3201 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3202 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3203 support this at all. Supporting it adds significant complexity to the
3204 record layer, and its removal is unlikely to cause interoperability
3209 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3210 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3211 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3216 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3217 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3219 *Richard Levitte, Andy Polyakov*
3221 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3226 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3227 platform rather than 'mingw'.
3231 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3232 success if they are asked to add an object which already exists
3233 in the store. This change cascades to other functions which load
3234 certificates and CRLs.
3238 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3239 facilitate stack unwinding even from assembly subroutines.
3243 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3244 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3248 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3249 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3250 which is the minimum version we support.
3254 * Certificate time validation (X509_cmp_time) enforces stricter
3255 compliance with RFC 5280. Fractional seconds and timezone offsets
3256 are no longer allowed.
3260 * Add support for ARIA
3264 * s_client will now send the Server Name Indication (SNI) extension by
3265 default unless the new "-noservername" option is used. The server name is
3266 based on the host provided to the "-connect" option unless overridden by
3267 using "-servername".
3271 * Add support for SipHash
3275 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3276 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3277 prevent issues where no progress is being made and the peer continually
3278 sends unrecognised record types, using up resources processing them.
3282 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3283 using the algorithm defined in
3284 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3288 * Heartbeat support has been removed; the ABI is changed for now.
3290 *Richard Levitte, Rich Salz*
3292 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3296 * The RSA "null" method, which was partially supported to avoid patent
3297 issues, has been replaced to always returns NULL.
3304 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3306 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3307 used even when parsing explicit parameters, when loading a encoded key
3308 or calling `EC_GROUP_new_from_ecpkparameters()`/
3309 `EC_GROUP_new_from_ecparameters()`.
3310 This prevents bypass of security hardening and performance gains,
3311 especially for curves with specialized EC_METHODs.
3312 By default, if a key encoded with explicit parameters is loaded and later
3313 encoded, the output is still encoded with explicit parameters, even if
3314 internally a "named" EC_GROUP is used for computation.
3318 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3319 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3320 NULL. After this change, only the cofactor parameter can be NULL. It also
3321 does some minimal sanity checks on the passed order.
3326 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3327 An attack is simple, if the first CMS_recipientInfo is valid but the
3328 second CMS_recipientInfo is chosen ciphertext. If the second
3329 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3330 encryption key will be replaced by garbage, and the message cannot be
3331 decoded, but if the RSA decryption fails, the correct encryption key is
3332 used and the recipient will not notice the attack.
3333 As a work around for this potential attack the length of the decrypted
3334 key must be equal to the cipher default key length, in case the
3335 certifiate is not given and all recipientInfo are tried out.
3336 The old behaviour can be re-enabled in the CMS code by setting the
3337 CMS_DEBUG_DECRYPT flag.
3342 * Use Windows installation paths in the mingw builds
3344 Mingw isn't a POSIX environment per se, which means that Windows
3345 paths should be used for installation.
3350 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3352 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3353 This changes the size when using the `genpkey` command when no size is given.
3354 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3355 generation commands to use 2048 bits by default.
3359 * Prevent over long nonces in ChaCha20-Poly1305.
3361 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3362 for every encryption operation. RFC 7539 specifies that the nonce value
3363 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3364 and front pads the nonce with 0 bytes if it is less than 12
3365 bytes. However it also incorrectly allows a nonce to be set of up to 16
3366 bytes. In this case only the last 12 bytes are significant and any
3367 additional leading bytes are ignored.
3369 It is a requirement of using this cipher that nonce values are
3370 unique. Messages encrypted using a reused nonce value are susceptible to
3371 serious confidentiality and integrity attacks. If an application changes
3372 the default nonce length to be longer than 12 bytes and then makes a
3373 change to the leading bytes of the nonce expecting the new value to be a
3374 new unique nonce then such an application could inadvertently encrypt
3375 messages with a reused nonce.
3377 Additionally the ignored bytes in a long nonce are not covered by the
3378 integrity guarantee of this cipher. Any application that relies on the
3379 integrity of these ignored leading bytes of a long nonce may be further
3380 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3381 is safe because no such use sets such a long nonce value. However user
3382 applications that use this cipher directly and set a non-default nonce
3383 length to be longer than 12 bytes may be vulnerable.
3385 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3391 * Added SCA hardening for modular field inversion in EC_GROUP through
3392 a new dedicated field_inv() pointer in EC_METHOD.
3393 This also addresses a leakage affecting conversions from projective
3394 to affine coordinates.
3396 *Billy Bob Brumley, Nicola Tuveri*
3398 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3399 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3403 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3407 * Remove the 'dist' target and add a tarball building script. The
3408 'dist' target has fallen out of use, and it shouldn't be
3409 necessary to configure just to create a source distribution.
3413 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3415 * Timing vulnerability in DSA signature generation
3417 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3418 timing side channel attack. An attacker could use variations in the signing
3419 algorithm to recover the private key.
3421 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3426 * Timing vulnerability in ECDSA signature generation
3428 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3429 timing side channel attack. An attacker could use variations in the signing
3430 algorithm to recover the private key.
3432 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3437 * Add coordinate blinding for EC_POINT and implement projective
3438 coordinate blinding for generic prime curves as a countermeasure to
3439 chosen point SCA attacks.
3441 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3443 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3445 * Client DoS due to large DH parameter
3447 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3448 malicious server can send a very large prime value to the client. This will
3449 cause the client to spend an unreasonably long period of time generating a
3450 key for this prime resulting in a hang until the client has finished. This
3451 could be exploited in a Denial Of Service attack.
3453 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3458 * Cache timing vulnerability in RSA Key Generation
3460 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3461 a cache timing side channel attack. An attacker with sufficient access to
3462 mount cache timing attacks during the RSA key generation process could
3463 recover the private key.
3465 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3466 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3471 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3472 parameter is no longer accepted, as it leads to a corrupt table. NULL
3473 pem_str is reserved for alias entries only.
3477 * Revert blinding in ECDSA sign and instead make problematic addition
3478 length-invariant. Switch even to fixed-length Montgomery multiplication.
3482 * Change generating and checking of primes so that the error rate of not
3483 being prime depends on the intended use based on the size of the input.
3484 For larger primes this will result in more rounds of Miller-Rabin.
3485 The maximal error rate for primes with more than 1080 bits is lowered
3488 *Kurt Roeckx, Annie Yousar*
3490 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3494 * Add blinding to ECDSA and DSA signatures to protect against side channel
3495 attacks discovered by Keegan Ryan (NCC Group).
3499 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3500 now allow empty (zero character) pass phrases.
3504 * Certificate time validation (X509_cmp_time) enforces stricter
3505 compliance with RFC 5280. Fractional seconds and timezone offsets
3506 are no longer allowed.
3510 * Fixed a text canonicalisation bug in CMS
3512 Where a CMS detached signature is used with text content the text goes
3513 through a canonicalisation process first prior to signing or verifying a
3514 signature. This process strips trailing space at the end of lines, converts
3515 line terminators to CRLF and removes additional trailing line terminators
3516 at the end of a file. A bug in the canonicalisation process meant that
3517 some characters, such as form-feed, were incorrectly treated as whitespace
3518 and removed. This is contrary to the specification (RFC5485). This fix
3519 could mean that detached text data signed with an earlier version of
3520 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3521 signed with a fixed OpenSSL may fail to verify with an earlier version of
3522 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3523 and use the "-binary" flag (for the "cms" command line application) or set
3524 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3528 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3530 * Constructed ASN.1 types with a recursive definition could exceed the stack
3532 Constructed ASN.1 types with a recursive definition (such as can be found
3533 in PKCS7) could eventually exceed the stack given malicious input with
3534 excessive recursion. This could result in a Denial Of Service attack. There
3535 are no such structures used within SSL/TLS that come from untrusted sources
3536 so this is considered safe.
3538 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3544 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3546 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3547 effectively reduced to only comparing the least significant bit of each
3548 byte. This allows an attacker to forge messages that would be considered as
3549 authenticated in an amount of tries lower than that guaranteed by the
3550 security claims of the scheme. The module can only be compiled by the
3551 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3553 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3559 * Add a build target 'build_all_generated', to build all generated files
3560 and only that. This can be used to prepare everything that requires
3561 things like perl for a system that lacks perl and then move everything
3562 to that system and do the rest of the build there.
3566 * Backport SSL_OP_NO_RENGOTIATION
3568 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3569 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3570 changes this is no longer possible in 1.1.0. Therefore the new
3571 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3572 1.1.0 to provide equivalent functionality.
3574 Note that if an application built against 1.1.0h headers (or above) is run
3575 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3576 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3580 * Removed the OS390-Unix config target. It relied on a script that doesn't
3585 * rsaz_1024_mul_avx2 overflow bug on x86_64
3587 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3588 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3589 Analysis suggests that attacks against RSA and DSA as a result of this
3590 defect would be very difficult to perform and are not believed likely.
3591 Attacks against DH1024 are considered just feasible, because most of the
3592 work necessary to deduce information about a private key may be performed
3593 offline. The amount of resources required for such an attack would be
3594 significant. However, for an attack on TLS to be meaningful, the server
3595 would have to share the DH1024 private key among multiple clients, which is
3596 no longer an option since CVE-2016-0701.
3598 This only affects processors that support the AVX2 but not ADX extensions
3599 like Intel Haswell (4th generation).
3601 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3602 was originally found via the OSS-Fuzz project.
3607 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3609 * bn_sqrx8x_internal carry bug on x86_64
3611 There is a carry propagating bug in the x86_64 Montgomery squaring
3612 procedure. No EC algorithms are affected. Analysis suggests that attacks
3613 against RSA and DSA as a result of this defect would be very difficult to
3614 perform and are not believed likely. Attacks against DH are considered just
3615 feasible (although very difficult) because most of the work necessary to
3616 deduce information about a private key may be performed offline. The amount
3617 of resources required for such an attack would be very significant and
3618 likely only accessible to a limited number of attackers. An attacker would
3619 additionally need online access to an unpatched system using the target
3620 private key in a scenario with persistent DH parameters and a private
3621 key that is shared between multiple clients.
3623 This only affects processors that support the BMI1, BMI2 and ADX extensions
3624 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3626 This issue was reported to OpenSSL by the OSS-Fuzz project.
3631 * Malformed X.509 IPAddressFamily could cause OOB read
3633 If an X.509 certificate has a malformed IPAddressFamily extension,
3634 OpenSSL could do a one-byte buffer overread. The most likely result
3635 would be an erroneous display of the certificate in text format.
3637 This issue was reported to OpenSSL by the OSS-Fuzz project.
3642 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3644 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3645 platform rather than 'mingw'.
3649 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3650 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3651 which is the minimum version we support.
3655 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3657 * Encrypt-Then-Mac renegotiation crash
3659 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3660 negotiated where it was not in the original handshake (or vice-versa) then
3661 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3662 and servers are affected.
3664 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3669 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3671 * Truncated packet could crash via OOB read
3673 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3674 cipher is being used, then a truncated packet can cause that host to
3675 perform an out-of-bounds read, usually resulting in a crash.
3677 This issue was reported to OpenSSL by Robert Święcki of Google.
3682 * Bad (EC)DHE parameters cause a client crash
3684 If a malicious server supplies bad parameters for a DHE or ECDHE key
3685 exchange then this can result in the client attempting to dereference a
3686 NULL pointer leading to a client crash. This could be exploited in a Denial
3689 This issue was reported to OpenSSL by Guido Vranken.
3694 * BN_mod_exp may produce incorrect results on x86_64
3696 There is a carry propagating bug in the x86_64 Montgomery squaring
3697 procedure. No EC algorithms are affected. Analysis suggests that attacks
3698 against RSA and DSA as a result of this defect would be very difficult to
3699 perform and are not believed likely. Attacks against DH are considered just
3700 feasible (although very difficult) because most of the work necessary to
3701 deduce information about a private key may be performed offline. The amount
3702 of resources required for such an attack would be very significant and
3703 likely only accessible to a limited number of attackers. An attacker would
3704 additionally need online access to an unpatched system using the target
3705 private key in a scenario with persistent DH parameters and a private
3706 key that is shared between multiple clients. For example this can occur by
3707 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3708 similar to CVE-2015-3193 but must be treated as a separate problem.
3710 This issue was reported to OpenSSL by the OSS-Fuzz project.
3715 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3717 * ChaCha20/Poly1305 heap-buffer-overflow
3719 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3720 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3721 crash. This issue is not considered to be exploitable beyond a DoS.
3723 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3728 * CMS Null dereference
3730 Applications parsing invalid CMS structures can crash with a NULL pointer
3731 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3732 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3733 structure callback if an attempt is made to free certain invalid encodings.
3734 Only CHOICE structures using a callback which do not handle NULL value are
3737 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3742 * Montgomery multiplication may produce incorrect results
3744 There is a carry propagating bug in the Broadwell-specific Montgomery
3745 multiplication procedure that handles input lengths divisible by, but
3746 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3747 and DH private keys are impossible. This is because the subroutine in
3748 question is not used in operations with the private key itself and an input
3749 of the attacker's direct choice. Otherwise the bug can manifest itself as
3750 transient authentication and key negotiation failures or reproducible
3751 erroneous outcome of public-key operations with specially crafted input.
3752 Among EC algorithms only Brainpool P-512 curves are affected and one
3753 presumably can attack ECDH key negotiation. Impact was not analyzed in
3754 detail, because pre-requisites for attack are considered unlikely. Namely
3755 multiple clients have to choose the curve in question and the server has to
3756 share the private key among them, neither of which is default behaviour.
3757 Even then only clients that chose the curve will be affected.
3759 This issue was publicly reported as transient failures and was not
3760 initially recognized as a security issue. Thanks to Richard Morgan for
3761 providing reproducible case.
3766 * Removed automatic addition of RPATH in shared libraries and executables,
3767 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3771 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3773 * Fix Use After Free for large message sizes
3775 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3776 message larger than approx 16k is received then the underlying buffer to
3777 store the incoming message is reallocated and moved. Unfortunately a
3778 dangling pointer to the old location is left which results in an attempt to
3779 write to the previously freed location. This is likely to result in a
3780 crash, however it could potentially lead to execution of arbitrary code.
3782 This issue only affects OpenSSL 1.1.0a.
3784 This issue was reported to OpenSSL by Robert Święcki.
3789 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3791 * OCSP Status Request extension unbounded memory growth
3793 A malicious client can send an excessively large OCSP Status Request
3794 extension. If that client continually requests renegotiation, sending a
3795 large OCSP Status Request extension each time, then there will be unbounded
3796 memory growth on the server. This will eventually lead to a Denial Of
3797 Service attack through memory exhaustion. Servers with a default
3798 configuration are vulnerable even if they do not support OCSP. Builds using
3799 the "no-ocsp" build time option are not affected.
3801 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3806 * SSL_peek() hang on empty record
3808 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3809 sends an empty record. This could be exploited by a malicious peer in a
3810 Denial Of Service attack.
3812 This issue was reported to OpenSSL by Alex Gaynor.
3817 * Excessive allocation of memory in tls_get_message_header() and
3818 dtls1_preprocess_fragment()
3820 A (D)TLS message includes 3 bytes for its length in the header for the
3821 message. This would allow for messages up to 16Mb in length. Messages of
3822 this length are excessive and OpenSSL includes a check to ensure that a
3823 peer is sending reasonably sized messages in order to avoid too much memory
3824 being consumed to service a connection. A flaw in the logic of version
3825 1.1.0 means that memory for the message is allocated too early, prior to
3826 the excessive message length check. Due to way memory is allocated in
3827 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3828 to service a connection. This could lead to a Denial of Service through
3829 memory exhaustion. However, the excessive message length check still takes
3830 place, and this would cause the connection to immediately fail. Assuming
3831 that the application calls SSL_free() on the failed connection in a timely
3832 manner then the 21Mb of allocated memory will then be immediately freed
3833 again. Therefore the excessive memory allocation will be transitory in
3834 nature. This then means that there is only a security impact if:
3836 1) The application does not call SSL_free() in a timely manner in the event
3837 that the connection fails
3839 2) The application is working in a constrained environment where there is
3840 very little free memory
3842 3) The attacker initiates multiple connection attempts such that there are
3843 multiple connections in a state where memory has been allocated for the
3844 connection; SSL_free() has not yet been called; and there is insufficient
3845 memory to service the multiple requests.
3847 Except in the instance of (1) above any Denial Of Service is likely to be
3848 transitory because as soon as the connection fails the memory is
3849 subsequently freed again in the SSL_free() call. However there is an
3850 increased risk during this period of application crashes due to the lack of
3851 memory - which would then mean a more serious Denial of Service.
3853 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3854 (CVE-2016-6307 and CVE-2016-6308)
3858 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3859 had to be removed. Primary reason is that vendor assembler can't
3860 assemble our modules with -KPIC flag. As result it, assembly
3861 support, was not even available as option. But its lack means
3862 lack of side-channel resistant code, which is incompatible with
3863 security by todays standards. Fortunately gcc is readily available
3864 prepackaged option, which we firmly point at...
3868 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3870 * Windows command-line tool supports UTF-8 opt-in option for arguments
3871 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3872 (to any value) allows Windows user to access PKCS#12 file generated
3873 with Windows CryptoAPI and protected with non-ASCII password, as well
3874 as files generated under UTF-8 locale on Linux also protected with
3879 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3880 have been disabled by default and removed from DEFAULT, just like RC4.
3881 See the RC4 item below to re-enable both.
3885 * The method for finding the storage location for the Windows RAND seed file
3886 has changed. First we check %RANDFILE%. If that is not set then we check
3887 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3888 all else fails we fall back to C:\.
3892 * The EVP_EncryptUpdate() function has had its return type changed from void
3893 to int. A return of 0 indicates and error while a return of 1 indicates
3898 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3899 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3900 off the constant time implementation for RSA, DSA and DH have been made
3901 no-ops and deprecated.
3905 * Windows RAND implementation was simplified to only get entropy by
3906 calling CryptGenRandom(). Various other RAND-related tickets
3909 *Joseph Wylie Yandle, Rich Salz*
3911 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3912 and `OPENSSL_LH_`, respectively. The old names are available
3913 with API compatibility. They new names are now completely documented.
3917 * Unify TYPE_up_ref(obj) methods signature.
3918 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3919 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3920 int (instead of void) like all others TYPE_up_ref() methods.
3921 So now these methods also check the return value of CRYPTO_atomic_add(),
3922 and the validity of object reference counter.
3924 *fdasilvayy@gmail.com*
3926 * With Windows Visual Studio builds, the .pdb files are installed
3927 alongside the installed libraries and executables. For a static
3928 library installation, ossl_static.pdb is the associate compiler
3929 generated .pdb file to be used when linking programs.
3933 * Remove openssl.spec. Packaging files belong with the packagers.
3937 * Automatic Darwin/OSX configuration has had a refresh, it will now
3938 recognise x86_64 architectures automatically. You can still decide
3939 to build for a different bitness with the environment variable
3940 KERNEL_BITS (can be 32 or 64), for example:
3942 KERNEL_BITS=32 ./config
3946 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3947 256 bit AES and HMAC with SHA256.
3951 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3955 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3959 * To enable users to have their own config files and build file templates,
3960 Configure looks in the directory indicated by the environment variable
3961 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3962 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3963 name and is used as is.
3967 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3968 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3969 X509_CERT_FILE_CTX was removed.
3973 * "shared" builds are now the default. To create only static libraries use
3974 the "no-shared" Configure option.
3978 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3979 All of these option have not worked for some while and are fundamental
3984 * Make various cleanup routines no-ops and mark them as deprecated. Most
3985 global cleanup functions are no longer required because they are handled
3986 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3987 Explicitly de-initing can cause problems (e.g. where a library that uses
3988 OpenSSL de-inits, but an application is still using it). The affected
3989 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3990 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3991 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3992 COMP_zlib_cleanup().
3996 * --strict-warnings no longer enables runtime debugging options
3997 such as REF_DEBUG. Instead, debug options are automatically
3998 enabled with '--debug' builds.
4000 *Andy Polyakov, Emilia Käsper*
4002 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4003 have been moved out of the public header files. New functions for managing
4004 these have been added.
4008 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4009 objects have been moved out of the public header files. New
4010 functions for managing these have been added.
4014 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4015 have been moved out of the public header files. New functions for managing
4016 these have been added.
4020 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4021 moved out of the public header files. New functions for managing these
4026 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4030 * Removed the mk1mf build scripts.
4034 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4035 it is always safe to #include a header now.
4039 * Removed the aged BC-32 config and all its supporting scripts
4043 * Removed support for Ultrix, Netware, and OS/2.
4047 * Add support for HKDF.
4049 *Alessandro Ghedini*
4051 * Add support for blake2b and blake2s
4055 * Added support for "pipelining". Ciphers that have the
4056 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4057 encryptions/decryptions simultaneously. There are currently no built-in
4058 ciphers with this property but the expectation is that engines will be able
4059 to offer it to significantly improve throughput. Support has been extended
4060 into libssl so that multiple records for a single connection can be
4061 processed in one go (for >=TLS 1.1).
4065 * Added the AFALG engine. This is an async capable engine which is able to
4066 offload work to the Linux kernel. In this initial version it only supports
4067 AES128-CBC. The kernel must be version 4.1.0 or greater.
4071 * OpenSSL now uses a new threading API. It is no longer necessary to
4072 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4073 are two supported threading models: pthreads and windows threads. It is
4074 also possible to configure OpenSSL at compile time for "no-threads". The
4075 old threading API should no longer be used. The functions have been
4076 replaced with "no-op" compatibility macros.
4078 *Alessandro Ghedini, Matt Caswell*
4080 * Modify behavior of ALPN to invoke callback after SNI/servername
4081 callback, such that updates to the SSL_CTX affect ALPN.
4085 * Add SSL_CIPHER queries for authentication and key-exchange.
4089 * Changes to the DEFAULT cipherlist:
4090 - Prefer (EC)DHE handshakes over plain RSA.
4091 - Prefer AEAD ciphers over legacy ciphers.
4092 - Prefer ECDSA over RSA when both certificates are available.
4093 - Prefer TLSv1.2 ciphers/PRF.
4094 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4099 * Change the ECC default curve list to be this, in order: x25519,
4100 secp256r1, secp521r1, secp384r1.
4104 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4105 disabled by default. They can be re-enabled using the
4106 enable-weak-ssl-ciphers option to Configure.
4110 * If the server has ALPN configured, but supports no protocols that the
4111 client advertises, send a fatal "no_application_protocol" alert.
4112 This behaviour is SHALL in RFC 7301, though it isn't universally
4113 implemented by other servers.
4117 * Add X25519 support.
4118 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4119 for public and private key encoding using the format documented in
4120 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4121 key generation and key derivation.
4123 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4128 * Deprecate SRP_VBASE_get_by_user.
4129 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4130 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4131 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4132 seed, even if the seed is configured.
4134 Users should use SRP_VBASE_get1_by_user instead. Note that in
4135 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4136 also that even though configuring the SRP seed attempts to hide
4137 invalid usernames by continuing the handshake with fake
4138 credentials, this behaviour is not constant time and no strong
4139 guarantees are made that the handshake is indistinguishable from
4140 that of a valid user.
4144 * Configuration change; it's now possible to build dynamic engines
4145 without having to build shared libraries and vice versa. This
4146 only applies to the engines in `engines/`, those in `crypto/engine/`
4147 will always be built into libcrypto (i.e. "static").
4149 Building dynamic engines is enabled by default; to disable, use
4150 the configuration option "disable-dynamic-engine".
4152 The only requirements for building dynamic engines are the
4153 presence of the DSO module and building with position independent
4154 code, so they will also automatically be disabled if configuring
4155 with "disable-dso" or "disable-pic".
4157 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4158 are also taken away from openssl/opensslconf.h, as they are
4163 * Configuration change; if there is a known flag to compile
4164 position independent code, it will always be applied on the
4165 libcrypto and libssl object files, and never on the application
4166 object files. This means other libraries that use routines from
4167 libcrypto / libssl can be made into shared libraries regardless
4168 of how OpenSSL was configured.
4170 If this isn't desirable, the configuration options "disable-pic"
4171 or "no-pic" can be used to disable the use of PIC. This will
4172 also disable building shared libraries and dynamic engines.
4176 * Removed JPAKE code. It was experimental and has no wide use.
4180 * The INSTALL_PREFIX Makefile variable has been renamed to
4181 DESTDIR. That makes for less confusion on what this variable
4182 is for. Also, the configuration option --install_prefix is
4187 * Heartbeat for TLS has been removed and is disabled by default
4188 for DTLS; configure with enable-heartbeats. Code that uses the
4189 old #define's might need to be updated.
4191 *Emilia Käsper, Rich Salz*
4193 * Rename REF_CHECK to REF_DEBUG.
4197 * New "unified" build system
4199 The "unified" build system is aimed to be a common system for all
4200 platforms we support. With it comes new support for VMS.
4202 This system builds supports building in a different directory tree
4203 than the source tree. It produces one Makefile (for unix family
4204 or lookalikes), or one descrip.mms (for VMS).
4206 The source of information to make the Makefile / descrip.mms is
4207 small files called 'build.info', holding the necessary
4208 information for each directory with source to compile, and a
4209 template in Configurations, like unix-Makefile.tmpl or
4212 With this change, the library names were also renamed on Windows
4213 and on VMS. They now have names that are closer to the standard
4214 on Unix, and include the major version number, and in certain
4215 cases, the architecture they are built for. See "Notes on shared
4216 libraries" in INSTALL.
4218 We rely heavily on the perl module Text::Template.
4222 * Added support for auto-initialisation and de-initialisation of the library.
4223 OpenSSL no longer requires explicit init or deinit routines to be called,
4224 except in certain circumstances. See the OPENSSL_init_crypto() and
4225 OPENSSL_init_ssl() man pages for further information.
4229 * The arguments to the DTLSv1_listen function have changed. Specifically the
4230 "peer" argument is now expected to be a BIO_ADDR object.
4232 * Rewrite of BIO networking library. The BIO library lacked consistent
4233 support of IPv6, and adding it required some more extensive
4234 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4235 which hold all types of addresses and chains of address information.
4236 It also introduces a new API, with functions like BIO_socket,
4237 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4238 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4239 have been adapted accordingly.
4243 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4248 * CRIME protection: disable compression by default, even if OpenSSL is
4249 compiled with zlib enabled. Applications can still enable compression
4250 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4251 using the SSL_CONF library to configure compression.
4255 * The signature of the session callback configured with
4256 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4257 was explicitly marked as `const unsigned char*` instead of
4262 * Always DPURIFY. Remove the use of uninitialized memory in the
4263 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4267 * Removed many obsolete configuration items, including
4268 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4269 MD2_CHAR, MD2_INT, MD2_LONG
4271 IDEA_SHORT, IDEA_LONG
4272 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4274 *Rich Salz, with advice from Andy Polyakov*
4276 * Many BN internals have been moved to an internal header file.
4278 *Rich Salz with help from Andy Polyakov*
4280 * Configuration and writing out the results from it has changed.
4281 Files such as Makefile include/openssl/opensslconf.h and are now
4282 produced through general templates, such as Makefile.in and
4283 crypto/opensslconf.h.in and some help from the perl module
4286 Also, the center of configuration information is no longer
4287 Makefile. Instead, Configure produces a perl module in
4288 configdata.pm which holds most of the config data (in the hash
4289 table %config), the target data that comes from the target
4290 configuration in one of the `Configurations/*.conf` files (in
4295 * To clarify their intended purposes, the Configure options
4296 --prefix and --openssldir change their semantics, and become more
4297 straightforward and less interdependent.
4299 --prefix shall be used exclusively to give the location INSTALLTOP
4300 where programs, scripts, libraries, include files and manuals are
4301 going to be installed. The default is now /usr/local.
4303 --openssldir shall be used exclusively to give the default
4304 location OPENSSLDIR where certificates, private keys, CRLs are
4305 managed. This is also where the default openssl.cnf gets
4307 If the directory given with this option is a relative path, the
4308 values of both the --prefix value and the --openssldir value will
4309 be combined to become OPENSSLDIR.
4310 The default for --openssldir is INSTALLTOP/ssl.
4312 Anyone who uses --openssldir to specify where OpenSSL is to be
4313 installed MUST change to use --prefix instead.
4317 * The GOST engine was out of date and therefore it has been removed. An up
4318 to date GOST engine is now being maintained in an external repository.
4319 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4320 support for GOST ciphersuites (these are only activated if a GOST engine
4325 * EGD is no longer supported by default; use enable-egd when
4328 *Ben Kaduk and Rich Salz*
4330 * The distribution now has Makefile.in files, which are used to
4331 create Makefile's when Configure is run. *Configure must be run
4332 before trying to build now.*
4336 * The return value for SSL_CIPHER_description() for error conditions
4341 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4343 Obtaining and performing DNSSEC validation of TLSA records is
4344 the application's responsibility. The application provides
4345 the TLSA records of its choice to OpenSSL, and these are then
4346 used to authenticate the peer.
4348 The TLSA records need not even come from DNS. They can, for
4349 example, be used to implement local end-entity certificate or
4350 trust-anchor "pinning", where the "pin" data takes the form
4351 of TLSA records, which can augment or replace verification
4352 based on the usual WebPKI public certification authorities.
4356 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4357 continues to support deprecated interfaces in default builds.
4358 However, applications are strongly advised to compile their
4359 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4360 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4361 or the 1.1.0 releases.
4363 In environments in which all applications have been ported to
4364 not use any deprecated interfaces OpenSSL's Configure script
4365 should be used with the --api=1.1.0 option to entirely remove
4366 support for the deprecated features from the library and
4367 unconditionally disable them in the installed headers.
4368 Essentially the same effect can be achieved with the "no-deprecated"
4369 argument to Configure, except that this will always restrict
4370 the build to just the latest API, rather than a fixed API
4373 As applications are ported to future revisions of the API,
4374 they should update their compile-time OPENSSL_API_COMPAT define
4375 accordingly, but in most cases should be able to continue to
4376 compile with later releases.
4378 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4379 0x10000000L and 0x00908000L, respectively. However those
4380 versions did not support the OPENSSL_API_COMPAT feature, and
4381 so applications are not typically tested for explicit support
4382 of just the undeprecated features of either release.
4386 * Add support for setting the minimum and maximum supported protocol.
4387 It can bet set via the SSL_set_min_proto_version() and
4388 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4389 MaxProtocol. It's recommended to use the new APIs to disable
4390 protocols instead of disabling individual protocols using
4391 SSL_set_options() or SSL_CONF's Protocol. This change also
4392 removes support for disabling TLS 1.2 in the OpenSSL TLS
4393 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4397 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4401 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4402 and integrates ECDSA and ECDH functionality into EC. Implementations can
4403 now redirect key generation and no longer need to convert to or from
4406 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4407 include the ec.h header file instead.
4411 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4412 ciphers who are no longer supported and drops support the ephemeral RSA key
4413 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4417 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4418 opaque. For HMAC_CTX, the following constructors and destructors
4421 HMAC_CTX *HMAC_CTX_new(void);
4422 void HMAC_CTX_free(HMAC_CTX *ctx);
4424 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4425 destroy such methods has been added. See EVP_MD_meth_new(3) and
4426 EVP_CIPHER_meth_new(3) for documentation.
4429 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4430 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4431 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4432 an already created structure.
4433 2) For consistency with the majority of our object creators and
4434 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4435 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4436 for deprecated builds.
4440 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4441 cryptographic operations to be performed asynchronously as long as an
4442 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4443 further details. Libssl has also had this capability integrated with the
4444 introduction of the new mode SSL_MODE_ASYNC and associated error
4445 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4446 pages. This work was developed in partnership with Intel Corp.
4450 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4451 always enabled now. If you want to disable the support you should
4452 exclude it using the list of supported ciphers. This also means that the
4453 "-no_ecdhe" option has been removed from s_server.
4457 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4458 SSL_{CTX_}set1_curves() which can set a list.
4462 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4463 curve you want to support using SSL_{CTX_}set1_curves().
4467 * State machine rewrite. The state machine code has been significantly
4468 refactored in order to remove much duplication of code and solve issues
4469 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4470 further details). This change does have some associated API changes.
4471 Notably the SSL_state() function has been removed and replaced by
4472 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4473 SSL_set_state() has been removed altogether. The previous handshake states
4474 defined in ssl.h and ssl3.h have also been removed.
4478 * All instances of the string "ssleay" in the public API were replaced
4479 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4480 Some error codes related to internal RSA_eay API's were renamed.
4484 * The demo files in crypto/threads were moved to demo/threads.
4488 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4491 *Matt Caswell, Rich Salz*
4493 * New ASN.1 embed macro.
4495 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4496 structure is not allocated: it is part of the parent. That is instead of
4504 This reduces memory fragmentation and make it impossible to accidentally
4505 set a mandatory field to NULL.
4507 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4508 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4509 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4514 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4518 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4519 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4520 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4521 DES and RC4 ciphersuites.
4525 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4526 This changes the decoding behaviour for some invalid messages,
4527 though the change is mostly in the more lenient direction, and
4528 legacy behaviour is preserved as much as possible.
4532 * Fix no-stdio build.
4533 *David Woodhouse <David.Woodhouse@intel.com> and also*
4534 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4536 * New testing framework
4537 The testing framework has been largely rewritten and is now using
4538 perl and the perl modules Test::Harness and an extended variant of
4539 Test::More called OpenSSL::Test to do its work. All test scripts in
4540 test/ have been rewritten into test recipes, and all direct calls to
4541 executables in test/Makefile have become individual recipes using the
4542 simplified testing OpenSSL::Test::Simple.
4544 For documentation on our testing modules, do:
4546 perldoc test/testlib/OpenSSL/Test/Simple.pm
4547 perldoc test/testlib/OpenSSL/Test.pm
4551 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4552 are used; the latter aborts on memory leaks (usually checked on exit).
4553 Some undocumented "set malloc, etc., hooks" functions were removed
4554 and others were changed. All are now documented.
4558 * In DSA_generate_parameters_ex, if the provided seed is too short,
4561 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4563 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4564 from RFC4279, RFC4785, RFC5487, RFC5489.
4566 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4567 original RSA_PSK patch.
4571 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4572 era flag was never set throughout the codebase (only read). Also removed
4573 SSL3_FLAGS_POP_BUFFER which was only used if
4574 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4578 * Changed the default name options in the "ca", "crl", "req" and "x509"
4579 to be "oneline" instead of "compat".
4583 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4584 not aware of clients that still exhibit this bug, and the workaround
4585 hasn't been working properly for a while.
4589 * The return type of BIO_number_read() and BIO_number_written() as well as
4590 the corresponding num_read and num_write members in the BIO structure has
4591 changed from unsigned long to uint64_t. On platforms where an unsigned
4592 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4597 * Given the pervasive nature of TLS extensions it is inadvisable to run
4598 OpenSSL without support for them. It also means that maintaining
4599 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4600 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4604 * Removed support for the two export grade static DH ciphersuites
4605 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4606 were newly added (along with a number of other static DH ciphersuites) to
4607 1.0.2. However the two export ones have *never* worked since they were
4608 introduced. It seems strange in any case to be adding new export
4609 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4613 * Version negotiation has been rewritten. In particular SSLv23_method(),
4614 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4615 and turned into macros which simply call the new preferred function names
4616 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4617 should use the new names instead. Also as part of this change the ssl23.h
4618 header file has been removed.
4622 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4623 code and the associated standard is no longer considered fit-for-purpose.
4627 * RT2547 was closed. When generating a private key, try to make the
4628 output file readable only by the owner. This behavior change might
4629 be noticeable when interacting with other software.
4631 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4636 * Added HTTP GET support to the ocsp command.
4640 * Changed default digest for the dgst and enc commands from MD5 to
4645 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4649 * Added support for TLS extended master secret from
4650 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4651 initial patch which was a great help during development.
4655 * All libssl internal structures have been removed from the public header
4656 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4657 now redundant). Users should not attempt to access internal structures
4658 directly. Instead they should use the provided API functions.
4662 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4663 Access to deprecated functions can be re-enabled by running config with
4664 "enable-deprecated". In addition applications wishing to use deprecated
4665 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4666 will, by default, disable some transitive includes that previously existed
4667 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4671 * Added support for OCB mode. OpenSSL has been granted a patent license
4672 compatible with the OpenSSL license for use of OCB. Details are available
4673 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4674 for OCB can be removed by calling config with no-ocb.
4678 * SSLv2 support has been removed. It still supports receiving a SSLv2
4679 compatible client hello.
4683 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4684 done while fixing the error code for the key-too-small case.
4686 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4688 * CA.sh has been removed; use CA.pl instead.
4692 * Removed old DES API.
4696 * Remove various unsupported platforms:
4702 Sinix/ReliantUNIX RM400
4707 16-bit platforms such as WIN16
4711 * Clean up OPENSSL_NO_xxx #define's
4712 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4713 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4714 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4715 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4716 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4717 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4718 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4719 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4720 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4721 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4725 * Cleaned up dead code
4726 Remove all but one '#ifdef undef' which is to be looked at.
4730 * Clean up calling of xxx_free routines.
4731 Just like free(), fix most of the xxx_free routines to accept
4732 NULL. Remove the non-null checks from callers. Save much code.
4736 * Add secure heap for storage of private keys (when possible).
4737 Add BIO_s_secmem(), CBIGNUM, etc.
4738 Contributed by Akamai Technologies under our Corporate CLA.
4742 * Experimental support for a new, fast, unbiased prime candidate generator,
4743 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4745 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4747 * New output format NSS in the sess_id command line tool. This allows
4748 exporting the session id and the master key in NSS keylog format.
4750 *Martin Kaiser <martin@kaiser.cx>*
4752 * Harmonize version and its documentation. -f flag is used to display
4755 *mancha <mancha1@zoho.com>*
4757 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4758 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4760 *mancha <mancha1@zoho.com>*
4762 * Fix some double frees. These are not thought to be exploitable.
4764 *mancha <mancha1@zoho.com>*
4766 * A missing bounds check in the handling of the TLS heartbeat extension
4767 can be used to reveal up to 64k of memory to a connected client or
4770 Thanks for Neel Mehta of Google Security for discovering this bug and to
4771 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4772 preparing the fix ([CVE-2014-0160])
4774 *Adam Langley, Bodo Moeller*
4776 * Fix for the attack described in the paper "Recovering OpenSSL
4777 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4778 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4779 <http://eprint.iacr.org/2014/140>
4781 Thanks to Yuval Yarom and Naomi Benger for discovering this
4782 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4784 *Yuval Yarom and Naomi Benger*
4786 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4787 this fixes a limitation in previous versions of OpenSSL.
4791 * Experimental encrypt-then-mac support.
4793 Experimental support for encrypt then mac from
4794 draft-gutmann-tls-encrypt-then-mac-02.txt
4796 To enable it set the appropriate extension number (0x42 for the test
4797 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4799 For non-compliant peers (i.e. just about everything) this should have no
4802 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4806 * Add EVP support for key wrapping algorithms, to avoid problems with
4807 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4808 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4809 algorithms and include tests cases.
4813 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4818 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4819 MGF1 digest and OAEP label.
4823 * Make openssl verify return errors.
4825 *Chris Palmer <palmer@google.com> and Ben Laurie*
4827 * New function ASN1_TIME_diff to calculate the difference between two
4828 ASN1_TIME structures or one structure and the current time.
4832 * Update fips_test_suite to support multiple command line options. New
4833 test to induce all self test errors in sequence and check expected
4838 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4839 sign or verify all in one operation.
4843 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4844 test programs and fips_test_suite. Includes functionality to parse
4845 the minimal script output of fipsalgest.pl directly.
4849 * Add authorisation parameter to FIPS_module_mode_set().
4853 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4857 * Use separate DRBG fields for internal and external flags. New function
4858 FIPS_drbg_health_check() to perform on demand health checking. Add
4859 generation tests to fips_test_suite with reduced health check interval to
4860 demonstrate periodic health checking. Add "nodh" option to
4861 fips_test_suite to skip very slow DH test.
4865 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4870 * More extensive health check for DRBG checking many more failure modes.
4871 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4872 combination: call this in fips_test_suite.
4876 * Add support for canonical generation of DSA parameter 'g'. See
4879 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4880 POST to handle HMAC cases.
4884 * Add functions FIPS_module_version() and FIPS_module_version_text()
4885 to return numerical and string versions of the FIPS module number.
4889 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4890 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4891 outside the validated module in the FIPS capable OpenSSL.
4895 * Minor change to DRBG entropy callback semantics. In some cases
4896 there is no multiple of the block length between min_len and
4897 max_len. Allow the callback to return more than max_len bytes
4898 of entropy but discard any extra: it is the callback's responsibility
4899 to ensure that the extra data discarded does not impact the
4900 requested amount of entropy.
4904 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4905 information in FIPS186-3, SP800-57 and SP800-131A.
4909 * CCM support via EVP. Interface is very similar to GCM case except we
4910 must supply all data in one chunk (i.e. no update, final) and the
4911 message length must be supplied if AAD is used. Add algorithm test
4916 * Initial version of POST overhaul. Add POST callback to allow the status
4917 of POST to be monitored and/or failures induced. Modify fips_test_suite
4918 to use callback. Always run all selftests even if one fails.
4922 * XTS support including algorithm test driver in the fips_gcmtest program.
4923 Note: this does increase the maximum key length from 32 to 64 bytes but
4924 there should be no binary compatibility issues as existing applications
4925 will never use XTS mode.
4929 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4930 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4931 performs algorithm blocking for unapproved PRNG types. Also do not
4932 set PRNG type in FIPS_mode_set(): leave this to the application.
4933 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4934 the standard OpenSSL PRNG: set additional data to a date time vector.
4938 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4939 This shouldn't present any incompatibility problems because applications
4940 shouldn't be using these directly and any that are will need to rethink
4941 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4945 * Extensive self tests and health checking required by SP800-90 DRBG.
4946 Remove strength parameter from FIPS_drbg_instantiate and always
4947 instantiate at maximum supported strength.
4951 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4955 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4959 * New function DH_compute_key_padded() to compute a DH key and pad with
4960 leading zeroes if needed: this complies with SP800-56A et al.
4964 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4965 anything, incomplete, subject to change and largely untested at present.
4969 * Modify fipscanisteronly build option to only build the necessary object
4970 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4974 * Add experimental option FIPSSYMS to give all symbols in
4975 fipscanister.o and FIPS or fips prefix. This will avoid
4976 conflicts with future versions of OpenSSL. Add perl script
4977 util/fipsas.pl to preprocess assembly language source files
4978 and rename any affected symbols.
4982 * Add selftest checks and algorithm block of non-fips algorithms in
4983 FIPS mode. Remove DES2 from selftests.
4987 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4988 return internal method without any ENGINE dependencies. Add new
4989 tiny fips sign and verify functions.
4993 * New build option no-ec2m to disable characteristic 2 code.
4997 * New build option "fipscanisteronly". This only builds fipscanister.o
4998 and (currently) associated fips utilities. Uses the file Makefile.fips
4999 instead of Makefile.org as the prototype.
5003 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5004 Update fips_gcmtest to use IV generator.
5008 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5009 setting output buffer to NULL. The `*Final` function must be
5010 called although it will not retrieve any additional data. The tag
5011 can be set or retrieved with a ctrl. The IV length is by default 12
5012 bytes (96 bits) but can be set to an alternative value. If the IV
5013 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5018 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5019 underlying do_cipher function handles all cipher semantics itself
5020 including padding and finalisation. This is useful if (for example)
5021 an ENGINE cipher handles block padding itself. The behaviour of
5022 do_cipher is subtly changed if this flag is set: the return value
5023 is the number of characters written to the output buffer (zero is
5024 no longer an error code) or a negative error code. Also if the
5025 input buffer is NULL and length 0 finalisation should be performed.
5029 * If a candidate issuer certificate is already part of the constructed
5030 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5034 * Improve forward-security support: add functions
5036 void SSL_CTX_set_not_resumable_session_callback(
5037 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5038 void SSL_set_not_resumable_session_callback(
5039 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5041 for use by SSL/TLS servers; the callback function will be called whenever a
5042 new session is created, and gets to decide whether the session may be
5043 cached to make it resumable (return 0) or not (return 1). (As by the
5044 SSL/TLS protocol specifications, the session_id sent by the server will be
5045 empty to indicate that the session is not resumable; also, the server will
5046 not generate RFC 4507 (RFC 5077) session tickets.)
5048 A simple reasonable callback implementation is to return is_forward_secure.
5049 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5050 by the SSL/TLS server library, indicating whether it can provide forward
5053 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5055 * New -verify_name option in command line utilities to set verification
5060 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5061 Add CMAC pkey methods.
5065 * Experimental renegotiation in s_server -www mode. If the client
5066 browses /reneg connection is renegotiated. If /renegcert it is
5067 renegotiated requesting a certificate.
5071 * Add an "external" session cache for debugging purposes to s_server. This
5072 should help trace issues which normally are only apparent in deployed
5073 multi-process servers.
5077 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5078 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5079 BIO_set_cipher() and some obscure PEM functions were changed so they
5080 can now return an error. The RAND changes required a change to the
5081 RAND_METHOD structure.
5085 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5086 a gcc attribute to warn if the result of a function is ignored. This
5087 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5088 whose return value is often ignored.
5092 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5093 These allow SCTs (signed certificate timestamps) to be requested and
5094 validated when establishing a connection.
5096 *Rob Percival <robpercival@google.com>*
5101 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5103 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5104 used even when parsing explicit parameters, when loading a encoded key
5105 or calling `EC_GROUP_new_from_ecpkparameters()`/
5106 `EC_GROUP_new_from_ecparameters()`.
5107 This prevents bypass of security hardening and performance gains,
5108 especially for curves with specialized EC_METHODs.
5109 By default, if a key encoded with explicit parameters is loaded and later
5110 encoded, the output is still encoded with explicit parameters, even if
5111 internally a "named" EC_GROUP is used for computation.
5115 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5116 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5117 NULL. After this change, only the cofactor parameter can be NULL. It also
5118 does some minimal sanity checks on the passed order.
5123 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5124 An attack is simple, if the first CMS_recipientInfo is valid but the
5125 second CMS_recipientInfo is chosen ciphertext. If the second
5126 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5127 encryption key will be replaced by garbage, and the message cannot be
5128 decoded, but if the RSA decryption fails, the correct encryption key is
5129 used and the recipient will not notice the attack.
5130 As a work around for this potential attack the length of the decrypted
5131 key must be equal to the cipher default key length, in case the
5132 certifiate is not given and all recipientInfo are tried out.
5133 The old behaviour can be re-enabled in the CMS code by setting the
5134 CMS_DEBUG_DECRYPT flag.
5139 * Document issue with installation paths in diverse Windows builds
5141 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5142 binaries and run-time config file.
5147 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5149 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5150 This changes the size when using the `genpkey` command when no size is given.
5151 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5152 generation commands to use 2048 bits by default.
5156 * Add FIPS support for Android Arm 64-bit
5158 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5159 Module in Version 2.0.10. For some reason, the corresponding target
5160 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5161 built with FIPS support on Android Arm 64-bit. This omission has been
5164 *Matthias St. Pierre*
5166 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5168 * 0-byte record padding oracle
5170 If an application encounters a fatal protocol error and then calls
5171 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5172 then OpenSSL can respond differently to the calling application if a 0 byte
5173 record is received with invalid padding compared to if a 0 byte record is
5174 received with an invalid MAC. If the application then behaves differently
5175 based on that in a way that is detectable to the remote peer, then this
5176 amounts to a padding oracle that could be used to decrypt data.
5178 In order for this to be exploitable "non-stitched" ciphersuites must be in
5179 use. Stitched ciphersuites are optimised implementations of certain
5180 commonly used ciphersuites. Also the application must call SSL_shutdown()
5181 twice even if a protocol error has occurred (applications should not do
5182 this but some do anyway).
5184 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5185 Aviram, with additional investigation by Steven Collison and Andrew
5186 Hourselt. It was reported to OpenSSL on 10th December 2018.
5191 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5195 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5197 * Microarchitecture timing vulnerability in ECC scalar multiplication
5199 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5200 shown to be vulnerable to a microarchitecture timing side channel attack.
5201 An attacker with sufficient access to mount local timing attacks during
5202 ECDSA signature generation could recover the private key.
5204 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5205 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5211 * Timing vulnerability in DSA signature generation
5213 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5214 timing side channel attack. An attacker could use variations in the signing
5215 algorithm to recover the private key.
5217 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5222 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5223 Module, accidentally introduced while backporting security fixes from the
5224 development branch and hindering the use of ECC in FIPS mode.
5228 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5230 * Client DoS due to large DH parameter
5232 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5233 malicious server can send a very large prime value to the client. This will
5234 cause the client to spend an unreasonably long period of time generating a
5235 key for this prime resulting in a hang until the client has finished. This
5236 could be exploited in a Denial Of Service attack.
5238 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5243 * Cache timing vulnerability in RSA Key Generation
5245 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5246 a cache timing side channel attack. An attacker with sufficient access to
5247 mount cache timing attacks during the RSA key generation process could
5248 recover the private key.
5250 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5251 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5256 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5257 parameter is no longer accepted, as it leads to a corrupt table. NULL
5258 pem_str is reserved for alias entries only.
5262 * Revert blinding in ECDSA sign and instead make problematic addition
5263 length-invariant. Switch even to fixed-length Montgomery multiplication.
5267 * Change generating and checking of primes so that the error rate of not
5268 being prime depends on the intended use based on the size of the input.
5269 For larger primes this will result in more rounds of Miller-Rabin.
5270 The maximal error rate for primes with more than 1080 bits is lowered
5273 *Kurt Roeckx, Annie Yousar*
5275 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5279 * Add blinding to ECDSA and DSA signatures to protect against side channel
5280 attacks discovered by Keegan Ryan (NCC Group).
5284 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5285 now allow empty (zero character) pass phrases.
5289 * Certificate time validation (X509_cmp_time) enforces stricter
5290 compliance with RFC 5280. Fractional seconds and timezone offsets
5291 are no longer allowed.
5295 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5297 * Constructed ASN.1 types with a recursive definition could exceed the stack
5299 Constructed ASN.1 types with a recursive definition (such as can be found
5300 in PKCS7) could eventually exceed the stack given malicious input with
5301 excessive recursion. This could result in a Denial Of Service attack. There
5302 are no such structures used within SSL/TLS that come from untrusted sources
5303 so this is considered safe.
5305 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5311 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5313 * Read/write after SSL object in error state
5315 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5316 mechanism. The intent was that if a fatal error occurred during a handshake
5317 then OpenSSL would move into the error state and would immediately fail if
5318 you attempted to continue the handshake. This works as designed for the
5319 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5320 SSL_connect()), however due to a bug it does not work correctly if
5321 SSL_read() or SSL_write() is called directly. In that scenario, if the
5322 handshake fails then a fatal error will be returned in the initial function
5323 call. If SSL_read()/SSL_write() is subsequently called by the application
5324 for the same SSL object then it will succeed and the data is passed without
5325 being decrypted/encrypted directly from the SSL/TLS record layer.
5327 In order to exploit this issue an application bug would have to be present
5328 that resulted in a call to SSL_read()/SSL_write() being issued after having
5329 already received a fatal error.
5331 This issue was reported to OpenSSL by David Benjamin (Google).
5336 * rsaz_1024_mul_avx2 overflow bug on x86_64
5338 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5339 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5340 Analysis suggests that attacks against RSA and DSA as a result of this
5341 defect would be very difficult to perform and are not believed likely.
5342 Attacks against DH1024 are considered just feasible, because most of the
5343 work necessary to deduce information about a private key may be performed
5344 offline. The amount of resources required for such an attack would be
5345 significant. However, for an attack on TLS to be meaningful, the server
5346 would have to share the DH1024 private key among multiple clients, which is
5347 no longer an option since CVE-2016-0701.
5349 This only affects processors that support the AVX2 but not ADX extensions
5350 like Intel Haswell (4th generation).
5352 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5353 was originally found via the OSS-Fuzz project.
5358 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5360 * bn_sqrx8x_internal carry bug on x86_64
5362 There is a carry propagating bug in the x86_64 Montgomery squaring
5363 procedure. No EC algorithms are affected. Analysis suggests that attacks
5364 against RSA and DSA as a result of this defect would be very difficult to
5365 perform and are not believed likely. Attacks against DH are considered just
5366 feasible (although very difficult) because most of the work necessary to
5367 deduce information about a private key may be performed offline. The amount
5368 of resources required for such an attack would be very significant and
5369 likely only accessible to a limited number of attackers. An attacker would
5370 additionally need online access to an unpatched system using the target
5371 private key in a scenario with persistent DH parameters and a private
5372 key that is shared between multiple clients.
5374 This only affects processors that support the BMI1, BMI2 and ADX extensions
5375 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5377 This issue was reported to OpenSSL by the OSS-Fuzz project.
5382 * Malformed X.509 IPAddressFamily could cause OOB read
5384 If an X.509 certificate has a malformed IPAddressFamily extension,
5385 OpenSSL could do a one-byte buffer overread. The most likely result
5386 would be an erroneous display of the certificate in text format.
5388 This issue was reported to OpenSSL by the OSS-Fuzz project.
5392 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5394 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5395 platform rather than 'mingw'.
5399 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5401 * Truncated packet could crash via OOB read
5403 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5404 cipher is being used, then a truncated packet can cause that host to
5405 perform an out-of-bounds read, usually resulting in a crash.
5407 This issue was reported to OpenSSL by Robert Święcki of Google.
5412 * BN_mod_exp may produce incorrect results on x86_64
5414 There is a carry propagating bug in the x86_64 Montgomery squaring
5415 procedure. No EC algorithms are affected. Analysis suggests that attacks
5416 against RSA and DSA as a result of this defect would be very difficult to
5417 perform and are not believed likely. Attacks against DH are considered just
5418 feasible (although very difficult) because most of the work necessary to
5419 deduce information about a private key may be performed offline. The amount
5420 of resources required for such an attack would be very significant and
5421 likely only accessible to a limited number of attackers. An attacker would
5422 additionally need online access to an unpatched system using the target
5423 private key in a scenario with persistent DH parameters and a private
5424 key that is shared between multiple clients. For example this can occur by
5425 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5426 similar to CVE-2015-3193 but must be treated as a separate problem.
5428 This issue was reported to OpenSSL by the OSS-Fuzz project.
5433 * Montgomery multiplication may produce incorrect results
5435 There is a carry propagating bug in the Broadwell-specific Montgomery
5436 multiplication procedure that handles input lengths divisible by, but
5437 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5438 and DH private keys are impossible. This is because the subroutine in
5439 question is not used in operations with the private key itself and an input
5440 of the attacker's direct choice. Otherwise the bug can manifest itself as
5441 transient authentication and key negotiation failures or reproducible
5442 erroneous outcome of public-key operations with specially crafted input.
5443 Among EC algorithms only Brainpool P-512 curves are affected and one
5444 presumably can attack ECDH key negotiation. Impact was not analyzed in
5445 detail, because pre-requisites for attack are considered unlikely. Namely
5446 multiple clients have to choose the curve in question and the server has to
5447 share the private key among them, neither of which is default behaviour.
5448 Even then only clients that chose the curve will be affected.
5450 This issue was publicly reported as transient failures and was not
5451 initially recognized as a security issue. Thanks to Richard Morgan for
5452 providing reproducible case.
5457 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5458 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5459 prevent issues where no progress is being made and the peer continually
5460 sends unrecognised record types, using up resources processing them.
5464 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5466 * Missing CRL sanity check
5468 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5469 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5470 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5472 This issue only affects the OpenSSL 1.0.2i
5477 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5479 * OCSP Status Request extension unbounded memory growth
5481 A malicious client can send an excessively large OCSP Status Request
5482 extension. If that client continually requests renegotiation, sending a
5483 large OCSP Status Request extension each time, then there will be unbounded
5484 memory growth on the server. This will eventually lead to a Denial Of
5485 Service attack through memory exhaustion. Servers with a default
5486 configuration are vulnerable even if they do not support OCSP. Builds using
5487 the "no-ocsp" build time option are not affected.
5489 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5494 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5497 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5503 * OOB write in MDC2_Update()
5505 An overflow can occur in MDC2_Update() either if called directly or
5506 through the EVP_DigestUpdate() function using MDC2. If an attacker
5507 is able to supply very large amounts of input data after a previous
5508 call to EVP_EncryptUpdate() with a partial block then a length check
5509 can overflow resulting in a heap corruption.
5511 The amount of data needed is comparable to SIZE_MAX which is impractical
5514 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5519 * Malformed SHA512 ticket DoS
5521 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5522 DoS attack where a malformed ticket will result in an OOB read which will
5525 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5526 a custom server callback and ticket lookup mechanism.
5528 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5533 * OOB write in BN_bn2dec()
5535 The function BN_bn2dec() does not check the return value of BN_div_word().
5536 This can cause an OOB write if an application uses this function with an
5537 overly large BIGNUM. This could be a problem if an overly large certificate
5538 or CRL is printed out from an untrusted source. TLS is not affected because
5539 record limits will reject an oversized certificate before it is parsed.
5541 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5546 * OOB read in TS_OBJ_print_bio()
5548 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5549 the total length the OID text representation would use and not the amount
5550 of data written. This will result in OOB reads when large OIDs are
5553 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5558 * Pointer arithmetic undefined behaviour
5560 Avoid some undefined pointer arithmetic
5562 A common idiom in the codebase is to check limits in the following manner:
5565 Where "p" points to some malloc'd data of SIZE bytes and
5568 "len" here could be from some externally supplied data (e.g. from a TLS
5571 The rules of C pointer arithmetic are such that "p + len" is only well
5572 defined where len <= SIZE. Therefore the above idiom is actually
5573 undefined behaviour.
5575 For example this could cause problems if some malloc implementation
5576 provides an address for "p" such that "p + len" actually overflows for
5577 values of len that are too big and therefore p + len < limit.
5579 This issue was reported to OpenSSL by Guido Vranken
5584 * Constant time flag not preserved in DSA signing
5586 Operations in the DSA signing algorithm should run in constant time in
5587 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5588 implementation means that a non-constant time codepath is followed for
5589 certain operations. This has been demonstrated through a cache-timing
5590 attack to be sufficient for an attacker to recover the private DSA key.
5592 This issue was reported by César Pereida (Aalto University), Billy Brumley
5593 (Tampere University of Technology), and Yuval Yarom (The University of
5594 Adelaide and NICTA).
5599 * DTLS buffered message DoS
5601 In a DTLS connection where handshake messages are delivered out-of-order
5602 those messages that OpenSSL is not yet ready to process will be buffered
5603 for later use. Under certain circumstances, a flaw in the logic means that
5604 those messages do not get removed from the buffer even though the handshake
5605 has been completed. An attacker could force up to approx. 15 messages to
5606 remain in the buffer when they are no longer required. These messages will
5607 be cleared when the DTLS connection is closed. The default maximum size for
5608 a message is 100k. Therefore the attacker could force an additional 1500k
5609 to be consumed per connection. By opening many simulataneous connections an
5610 attacker could cause a DoS attack through memory exhaustion.
5612 This issue was reported to OpenSSL by Quan Luo.
5617 * DTLS replay protection DoS
5619 A flaw in the DTLS replay attack protection mechanism means that records
5620 that arrive for future epochs update the replay protection "window" before
5621 the MAC for the record has been validated. This could be exploited by an
5622 attacker by sending a record for the next epoch (which does not have to
5623 decrypt or have a valid MAC), with a very large sequence number. This means
5624 that all subsequent legitimate packets are dropped causing a denial of
5625 service for a specific DTLS connection.
5627 This issue was reported to OpenSSL by the OCAP audit team.
5632 * Certificate message OOB reads
5634 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5635 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5636 theoretical DoS risk but this has not been observed in practice on common
5639 The messages affected are client certificate, client certificate request
5640 and server certificate. As a result the attack can only be performed
5641 against a client or a server which enables client authentication.
5643 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5648 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5650 * Prevent padding oracle in AES-NI CBC MAC check
5652 A MITM attacker can use a padding oracle attack to decrypt traffic
5653 when the connection uses an AES CBC cipher and the server support
5656 This issue was introduced as part of the fix for Lucky 13 padding
5657 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5658 constant time by making sure that always the same bytes are read and
5659 compared against either the MAC or padding bytes. But it no longer
5660 checked that there was enough data to have both the MAC and padding
5663 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5667 * Fix EVP_EncodeUpdate overflow
5669 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5670 Base64 encoding of binary data. If an attacker is able to supply very large
5671 amounts of input data then a length check can overflow resulting in a heap
5674 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5675 the `PEM_write_bio*` family of functions. These are mainly used within the
5676 OpenSSL command line applications, so any application which processes data
5677 from an untrusted source and outputs it as a PEM file should be considered
5678 vulnerable to this issue. User applications that call these APIs directly
5679 with large amounts of untrusted data may also be vulnerable.
5681 This issue was reported by Guido Vranken.
5686 * Fix EVP_EncryptUpdate overflow
5688 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5689 is able to supply very large amounts of input data after a previous call to
5690 EVP_EncryptUpdate() with a partial block then a length check can overflow
5691 resulting in a heap corruption. Following an analysis of all OpenSSL
5692 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5693 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5694 the first called function after an EVP_EncryptInit(), and therefore that
5695 specific call must be safe. The second form is where the length passed to
5696 EVP_EncryptUpdate() can be seen from the code to be some small value and
5697 therefore there is no possibility of an overflow. Since all instances are
5698 one of these two forms, it is believed that there can be no overflows in
5699 internal code due to this problem. It should be noted that
5700 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5701 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5702 of these calls have also been analysed too and it is believed there are no
5703 instances in internal usage where an overflow could occur.
5705 This issue was reported by Guido Vranken.
5710 * Prevent ASN.1 BIO excessive memory allocation
5712 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5713 a short invalid encoding can cause allocation of large amounts of memory
5714 potentially consuming excessive resources or exhausting memory.
5716 Any application parsing untrusted data through d2i BIO functions is
5717 affected. The memory based functions such as d2i_X509() are *not* affected.
5718 Since the memory based functions are used by the TLS library, TLS
5719 applications are not affected.
5721 This issue was reported by Brian Carpenter.
5728 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5729 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5730 in arbitrary stack data being returned in the buffer.
5732 This issue was reported by Guido Vranken.
5737 * Modify behavior of ALPN to invoke callback after SNI/servername
5738 callback, such that updates to the SSL_CTX affect ALPN.
5742 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5747 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5748 methods are enabled and ssl2 is disabled the methods return NULL.
5752 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5754 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5755 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5756 provide any "EXPORT" or "LOW" strength ciphers.
5760 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5761 is by default disabled at build-time. Builds that are not configured with
5762 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5763 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5764 will need to explicitly call either of:
5766 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5768 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5770 as appropriate. Even if either of those is used, or the application
5771 explicitly uses the version-specific SSLv2_method() or its client and
5772 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5773 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5774 ciphers, and SSLv2 56-bit DES are no longer available.
5779 * Fix a double-free in DSA code
5781 A double free bug was discovered when OpenSSL parses malformed DSA private
5782 keys and could lead to a DoS attack or memory corruption for applications
5783 that receive DSA private keys from untrusted sources. This scenario is
5786 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5792 * Disable SRP fake user seed to address a server memory leak.
5794 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5796 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5797 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5798 was changed to ignore the "fake user" SRP seed, even if the seed
5801 Users should use SRP_VBASE_get1_by_user instead. Note that in
5802 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5803 also that even though configuring the SRP seed attempts to hide
5804 invalid usernames by continuing the handshake with fake
5805 credentials, this behaviour is not constant time and no strong
5806 guarantees are made that the handshake is indistinguishable from
5807 that of a valid user.
5812 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5814 In the BN_hex2bn function the number of hex digits is calculated using an
5815 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5816 large values of `i` this can result in `bn_expand` not allocating any
5817 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5818 field as NULL leading to a subsequent NULL ptr deref. For very large values
5819 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5820 In this case memory is allocated to the internal BIGNUM data field, but it
5821 is insufficiently sized leading to heap corruption. A similar issue exists
5822 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5823 is ever called by user applications with very large untrusted hex/dec data.
5824 This is anticipated to be a rare occurrence.
5826 All OpenSSL internal usage of these functions use data that is not expected
5827 to be untrusted, e.g. config file data or application command line
5828 arguments. If user developed applications generate config file data based
5829 on untrusted data then it is possible that this could also lead to security
5830 consequences. This is also anticipated to be rare.
5832 This issue was reported to OpenSSL by Guido Vranken.
5837 * Fix memory issues in `BIO_*printf` functions
5839 The internal `fmtstr` function used in processing a "%s" format string in
5840 the `BIO_*printf` functions could overflow while calculating the length of a
5841 string and cause an OOB read when printing very long strings.
5843 Additionally the internal `doapr_outch` function can attempt to write to an
5844 OOB memory location (at an offset from the NULL pointer) in the event of a
5845 memory allocation failure. In 1.0.2 and below this could be caused where
5846 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5847 could be in processing a very long "%s" format string. Memory leaks can
5850 The first issue may mask the second issue dependent on compiler behaviour.
5851 These problems could enable attacks where large amounts of untrusted data
5852 is passed to the `BIO_*printf` functions. If applications use these functions
5853 in this way then they could be vulnerable. OpenSSL itself uses these
5854 functions when printing out human-readable dumps of ASN.1 data. Therefore
5855 applications that print this data could be vulnerable if the data is from
5856 untrusted sources. OpenSSL command line applications could also be
5857 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5858 as command line arguments.
5860 Libssl is not considered directly vulnerable. Additionally certificates etc
5861 received via remote connections via libssl are also unlikely to be able to
5862 trigger these issues because of message size limits enforced within libssl.
5864 This issue was reported to OpenSSL Guido Vranken.
5869 * Side channel attack on modular exponentiation
5871 A side-channel attack was found which makes use of cache-bank conflicts on
5872 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5873 of RSA keys. The ability to exploit this issue is limited as it relies on
5874 an attacker who has control of code in a thread running on the same
5875 hyper-threaded core as the victim thread which is performing decryptions.
5877 This issue was reported to OpenSSL by Yuval Yarom, The University of
5878 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5879 Nadia Heninger, University of Pennsylvania with more information at
5880 <http://cachebleed.info>.
5885 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5886 if no keysize is specified with default_bits. This fixes an
5887 omission in an earlier change that changed all RSA/DSA key generation
5888 commands to use 2048 bits by default.
5892 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5894 * DH small subgroups
5896 Historically OpenSSL only ever generated DH parameters based on "safe"
5897 primes. More recently (in version 1.0.2) support was provided for
5898 generating X9.42 style parameter files such as those required for RFC 5114
5899 support. The primes used in such files may not be "safe". Where an
5900 application is using DH configured with parameters based on primes that are
5901 not "safe" then an attacker could use this fact to find a peer's private
5902 DH exponent. This attack requires that the attacker complete multiple
5903 handshakes in which the peer uses the same private DH exponent. For example
5904 this could be used to discover a TLS server's private DH exponent if it's
5905 reusing the private DH exponent or it's using a static DH ciphersuite.
5907 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5908 TLS. It is not on by default. If the option is not set then the server
5909 reuses the same private DH exponent for the life of the server process and
5910 would be vulnerable to this attack. It is believed that many popular
5911 applications do set this option and would therefore not be at risk.
5913 The fix for this issue adds an additional check where a "q" parameter is
5914 available (as is the case in X9.42 based parameters). This detects the
5915 only known attack, and is the only possible defense for static DH
5916 ciphersuites. This could have some performance impact.
5918 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5919 default and cannot be disabled. This could have some performance impact.
5921 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5926 * SSLv2 doesn't block disabled ciphers
5928 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5929 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5930 been disabled, provided that the SSLv2 protocol was not also disabled via
5933 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5934 and Sebastian Schinzel.
5939 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5941 * BN_mod_exp may produce incorrect results on x86_64
5943 There is a carry propagating bug in the x86_64 Montgomery squaring
5944 procedure. No EC algorithms are affected. Analysis suggests that attacks
5945 against RSA and DSA as a result of this defect would be very difficult to
5946 perform and are not believed likely. Attacks against DH are considered just
5947 feasible (although very difficult) because most of the work necessary to
5948 deduce information about a private key may be performed offline. The amount
5949 of resources required for such an attack would be very significant and
5950 likely only accessible to a limited number of attackers. An attacker would
5951 additionally need online access to an unpatched system using the target
5952 private key in a scenario with persistent DH parameters and a private
5953 key that is shared between multiple clients. For example this can occur by
5954 default in OpenSSL DHE based SSL/TLS ciphersuites.
5956 This issue was reported to OpenSSL by Hanno Böck.
5961 * Certificate verify crash with missing PSS parameter
5963 The signature verification routines will crash with a NULL pointer
5964 dereference if presented with an ASN.1 signature using the RSA PSS
5965 algorithm and absent mask generation function parameter. Since these
5966 routines are used to verify certificate signature algorithms this can be
5967 used to crash any certificate verification operation and exploited in a
5968 DoS attack. Any application which performs certificate verification is
5969 vulnerable including OpenSSL clients and servers which enable client
5972 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5977 * X509_ATTRIBUTE memory leak
5979 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5980 memory. This structure is used by the PKCS#7 and CMS routines so any
5981 application which reads PKCS#7 or CMS data from untrusted sources is
5982 affected. SSL/TLS is not affected.
5984 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5990 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5991 This changes the decoding behaviour for some invalid messages,
5992 though the change is mostly in the more lenient direction, and
5993 legacy behaviour is preserved as much as possible.
5997 * In DSA_generate_parameters_ex, if the provided seed is too short,
6000 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6002 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6004 * Alternate chains certificate forgery
6006 During certificate verification, OpenSSL will attempt to find an
6007 alternative certificate chain if the first attempt to build such a chain
6008 fails. An error in the implementation of this logic can mean that an
6009 attacker could cause certain checks on untrusted certificates to be
6010 bypassed, such as the CA flag, enabling them to use a valid leaf
6011 certificate to act as a CA and "issue" an invalid certificate.
6013 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6018 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6020 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6021 incompatibility in the handling of HMAC. The previous ABI has now been
6026 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6028 * Malformed ECParameters causes infinite loop
6030 When processing an ECParameters structure OpenSSL enters an infinite loop
6031 if the curve specified is over a specially malformed binary polynomial
6034 This can be used to perform denial of service against any
6035 system which processes public keys, certificate requests or
6036 certificates. This includes TLS clients and TLS servers with
6037 client authentication enabled.
6039 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6044 * Exploitable out-of-bounds read in X509_cmp_time
6046 X509_cmp_time does not properly check the length of the ASN1_TIME
6047 string and can read a few bytes out of bounds. In addition,
6048 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6051 An attacker can use this to craft malformed certificates and CRLs of
6052 various sizes and potentially cause a segmentation fault, resulting in
6053 a DoS on applications that verify certificates or CRLs. TLS clients
6054 that verify CRLs are affected. TLS clients and servers with client
6055 authentication enabled may be affected if they use custom verification
6058 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6059 independently by Hanno Böck.
6064 * PKCS7 crash with missing EnvelopedContent
6066 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6067 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6068 with missing content and trigger a NULL pointer dereference on parsing.
6070 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6071 structures from untrusted sources are affected. OpenSSL clients and
6072 servers are not affected.
6074 This issue was reported to OpenSSL by Michal Zalewski (Google).
6079 * CMS verify infinite loop with unknown hash function
6081 When verifying a signedData message the CMS code can enter an infinite loop
6082 if presented with an unknown hash function OID. This can be used to perform
6083 denial of service against any system which verifies signedData messages using
6085 This issue was reported to OpenSSL by Johannes Bauer.
6090 * Race condition handling NewSessionTicket
6092 If a NewSessionTicket is received by a multi-threaded client when attempting to
6093 reuse a previous ticket then a race condition can occur potentially leading to
6094 a double free of the ticket data.
6099 * Only support 256-bit or stronger elliptic curves with the
6100 'ecdh_auto' setting (server) or by default (client). Of supported
6101 curves, prefer P-256 (both).
6105 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6107 * ClientHello sigalgs DoS fix
6109 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6110 invalid signature algorithms extension a NULL pointer dereference will
6111 occur. This can be exploited in a DoS attack against the server.
6113 This issue was was reported to OpenSSL by David Ramos of Stanford
6117 *Stephen Henson and Matt Caswell*
6119 * Multiblock corrupted pointer fix
6121 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6122 feature only applies on 64 bit x86 architecture platforms that support AES
6123 NI instructions. A defect in the implementation of "multiblock" can cause
6124 OpenSSL's internal write buffer to become incorrectly set to NULL when
6125 using non-blocking IO. Typically, when the user application is using a
6126 socket BIO for writing, this will only result in a failed connection.
6127 However if some other BIO is used then it is likely that a segmentation
6128 fault will be triggered, thus enabling a potential DoS attack.
6130 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6135 * Segmentation fault in DTLSv1_listen fix
6137 The DTLSv1_listen function is intended to be stateless and processes the
6138 initial ClientHello from many peers. It is common for user code to loop
6139 over the call to DTLSv1_listen until a valid ClientHello is received with
6140 an associated cookie. A defect in the implementation of DTLSv1_listen means
6141 that state is preserved in the SSL object from one invocation to the next
6142 that can lead to a segmentation fault. Errors processing the initial
6143 ClientHello can trigger this scenario. An example of such an error could be
6144 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6147 This issue was reported to OpenSSL by Per Allansson.
6152 * Segmentation fault in ASN1_TYPE_cmp fix
6154 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6155 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6156 certificate signature algorithm consistency this can be used to crash any
6157 certificate verification operation and exploited in a DoS attack. Any
6158 application which performs certificate verification is vulnerable including
6159 OpenSSL clients and servers which enable client authentication.
6164 * Segmentation fault for invalid PSS parameters fix
6166 The signature verification routines will crash with a NULL pointer
6167 dereference if presented with an ASN.1 signature using the RSA PSS
6168 algorithm and invalid parameters. Since these routines are used to verify
6169 certificate signature algorithms this can be used to crash any
6170 certificate verification operation and exploited in a DoS attack. Any
6171 application which performs certificate verification is vulnerable including
6172 OpenSSL clients and servers which enable client authentication.
6174 This issue was was reported to OpenSSL by Brian Carpenter.
6179 * ASN.1 structure reuse memory corruption fix
6181 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6182 memory corruption via an invalid write. Such reuse is and has been
6183 strongly discouraged and is believed to be rare.
6185 Applications that parse structures containing CHOICE or ANY DEFINED BY
6186 components may be affected. Certificate parsing (d2i_X509 and related
6187 functions) are however not affected. OpenSSL clients and servers are
6193 * PKCS7 NULL pointer dereferences fix
6195 The PKCS#7 parsing code does not handle missing outer ContentInfo
6196 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6197 missing content and trigger a NULL pointer dereference on parsing.
6199 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6200 otherwise parse PKCS#7 structures from untrusted sources are
6201 affected. OpenSSL clients and servers are not affected.
6203 This issue was reported to OpenSSL by Michal Zalewski (Google).
6208 * DoS via reachable assert in SSLv2 servers fix
6210 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6211 servers that both support SSLv2 and enable export cipher suites by sending
6212 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6214 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6215 (OpenSSL development team).
6220 * Empty CKE with client auth and DHE fix
6222 If client auth is used then a server can seg fault in the event of a DHE
6223 ciphersuite being selected and a zero length ClientKeyExchange message
6224 being sent by the client. This could be exploited in a DoS attack.
6229 * Handshake with unseeded PRNG fix
6231 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6232 with an unseeded PRNG. The conditions are:
6233 - The client is on a platform where the PRNG has not been seeded
6234 automatically, and the user has not seeded manually
6235 - A protocol specific client method version has been used (i.e. not
6236 SSL_client_methodv23)
6237 - A ciphersuite is used that does not require additional random data from
6238 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6240 If the handshake succeeds then the client random that has been used will
6241 have been generated from a PRNG with insufficient entropy and therefore the
6242 output may be predictable.
6244 For example using the following command with an unseeded openssl will
6245 succeed on an unpatched platform:
6247 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6252 * Use After Free following d2i_ECPrivatekey error fix
6254 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6255 could cause a use after free condition. This, in turn, could cause a double
6256 free in several private key parsing functions (such as d2i_PrivateKey
6257 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6258 for applications that receive EC private keys from untrusted
6259 sources. This scenario is considered rare.
6261 This issue was discovered by the BoringSSL project and fixed in their
6267 * X509_to_X509_REQ NULL pointer deref fix
6269 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6270 the certificate key is invalid. This function is rarely used in practice.
6272 This issue was discovered by Brian Carpenter.
6277 * Removed the export ciphers from the DEFAULT ciphers
6281 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6283 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6284 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6285 So far those who have to target multiple platforms would compromise
6286 and argue that binary targeting say ARMv5 would still execute on
6287 ARMv8. "Universal" build resolves this compromise by providing
6288 near-optimal performance even on newer platforms.
6292 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6293 (other platforms pending).
6295 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6297 * Add support for the SignedCertificateTimestampList certificate and
6298 OCSP response extensions from RFC6962.
6302 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6303 for corner cases. (Certain input points at infinity could lead to
6304 bogus results, with non-infinity inputs mapped to infinity too.)
6308 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6309 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6310 common cases are optimized and there still is room for further
6311 improvements. Vector Permutation AES for Altivec is also added.
6315 * Add support for little-endian ppc64 Linux target.
6317 *Marcelo Cerri (IBM)*
6319 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6320 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6321 are optimized and there still is room for further improvements.
6322 Both 32- and 64-bit modes are supported.
6324 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6326 * Improved ARMv7 NEON support.
6330 * Support for SPARC Architecture 2011 crypto extensions, first
6331 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6332 SHA256/512, MD5, GHASH and modular exponentiation.
6334 *Andy Polyakov, David Miller*
6336 * Accelerated modular exponentiation for Intel processors, a.k.a.
6339 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6341 * Support for new and upcoming Intel processors, including AVX2,
6342 BMI and SHA ISA extensions. This includes additional "stitched"
6343 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6346 This work was sponsored by Intel Corp.
6350 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6351 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6352 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6356 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6357 this fixes a limitation in previous versions of OpenSSL.
6361 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6362 MGF1 digest and OAEP label.
6366 * Add EVP support for key wrapping algorithms, to avoid problems with
6367 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6368 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6369 algorithms and include tests cases.
6373 * Add functions to allocate and set the fields of an ECDSA_METHOD
6376 *Douglas E. Engert, Steve Henson*
6378 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6379 difference in days and seconds between two tm or ASN1_TIME structures.
6383 * Add -rev test option to s_server to just reverse order of characters
6384 received by client and send back to server. Also prints an abbreviated
6385 summary of the connection parameters.
6389 * New option -brief for s_client and s_server to print out a brief summary
6390 of connection parameters.
6394 * Add callbacks for arbitrary TLS extensions.
6396 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6398 * New option -crl_download in several openssl utilities to download CRLs
6399 from CRLDP extension in certificates.
6403 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6407 * New function X509_CRL_diff to generate a delta CRL from the difference
6408 of two full CRLs. Add support to "crl" utility.
6412 * New functions to set lookup_crls function and to retrieve
6413 X509_STORE from X509_STORE_CTX.
6417 * Print out deprecated issuer and subject unique ID fields in
6422 * Extend OCSP I/O functions so they can be used for simple general purpose
6423 HTTP as well as OCSP. New wrapper function which can be used to download
6424 CRLs using the OCSP API.
6428 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6432 * `SSL_CONF*` functions. These provide a common framework for application
6433 configuration using configuration files or command lines.
6437 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6438 message callback and prints the results. Needs compile time option
6439 "enable-ssl-trace". New options to s_client and s_server to enable
6444 * New ctrl and macro to retrieve supported points extensions.
6445 Print out extension in s_server and s_client.
6449 * New functions to retrieve certificate signature and signature
6454 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6459 * New Suite B modes for TLS code. These use and enforce the requirements
6460 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6461 only use Suite B curves. The Suite B modes can be set by using the
6462 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6466 * New chain verification flags for Suite B levels of security. Check
6467 algorithms are acceptable when flags are set in X509_verify_cert.
6471 * Make tls1_check_chain return a set of flags indicating checks passed
6472 by a certificate chain. Add additional tests to handle client
6473 certificates: checks for matching certificate type and issuer name
6478 * If an attempt is made to use a signature algorithm not in the peer
6479 preference list abort the handshake. If client has no suitable
6480 signature algorithms in response to a certificate request do not
6481 use the certificate.
6485 * If server EC tmp key is not in client preference list abort handshake.
6489 * Add support for certificate stores in CERT structure. This makes it
6490 possible to have different stores per SSL structure or one store in
6491 the parent SSL_CTX. Include distinct stores for certificate chain
6492 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6493 to build and store a certificate chain in CERT structure: returning
6494 an error if the chain cannot be built: this will allow applications
6495 to test if a chain is correctly configured.
6497 Note: if the CERT based stores are not set then the parent SSL_CTX
6498 store is used to retain compatibility with existing behaviour.
6502 * New function ssl_set_client_disabled to set a ciphersuite disabled
6503 mask based on the current session, check mask when sending client
6504 hello and checking the requested ciphersuite.
6508 * New ctrls to retrieve and set certificate types in a certificate
6509 request message. Print out received values in s_client. If certificate
6510 types is not set with custom values set sensible values based on
6511 supported signature algorithms.
6515 * Support for distinct client and server supported signature algorithms.
6519 * Add certificate callback. If set this is called whenever a certificate
6520 is required by client or server. An application can decide which
6521 certificate chain to present based on arbitrary criteria: for example
6522 supported signature algorithms. Add very simple example to s_server.
6523 This fixes many of the problems and restrictions of the existing client
6524 certificate callback: for example you can now clear an existing
6525 certificate and specify the whole chain.
6529 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6530 the certificate can be used for (if anything). Set valid_flags field
6531 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6532 to have similar checks in it.
6534 Add new "cert_flags" field to CERT structure and include a "strict mode".
6535 This enforces some TLS certificate requirements (such as only permitting
6536 certificate signature algorithms contained in the supported algorithms
6537 extension) which some implementations ignore: this option should be used
6538 with caution as it could cause interoperability issues.
6542 * Update and tidy signature algorithm extension processing. Work out
6543 shared signature algorithms based on preferences and peer algorithms
6544 and print them out in s_client and s_server. Abort handshake if no
6545 shared signature algorithms.
6549 * Add new functions to allow customised supported signature algorithms
6550 for SSL and SSL_CTX structures. Add options to s_client and s_server
6555 * New function SSL_certs_clear() to delete all references to certificates
6556 from an SSL structure. Before this once a certificate had been added
6557 it couldn't be removed.
6561 * Integrate hostname, email address and IP address checking with certificate
6562 verification. New verify options supporting checking in openssl utility.
6566 * Fixes and wildcard matching support to hostname and email checking
6567 functions. Add manual page.
6569 *Florian Weimer (Red Hat Product Security Team)*
6571 * New functions to check a hostname email or IP address against a
6572 certificate. Add options x509 utility to print results of checks against
6577 * Fix OCSP checking.
6579 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6581 * Initial experimental support for explicitly trusted non-root CAs.
6582 OpenSSL still tries to build a complete chain to a root but if an
6583 intermediate CA has a trust setting included that is used. The first
6584 setting is used: whether to trust (e.g., -addtrust option to the x509
6589 * Add -trusted_first option which attempts to find certificates in the
6590 trusted store even if an untrusted chain is also supplied.
6594 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6595 platform support for Linux and Android.
6599 * Support for linux-x32, ILP32 environment in x86_64 framework.
6603 * Experimental multi-implementation support for FIPS capable OpenSSL.
6604 When in FIPS mode the approved implementations are used as normal,
6605 when not in FIPS mode the internal unapproved versions are used instead.
6606 This means that the FIPS capable OpenSSL isn't forced to use the
6607 (often lower performance) FIPS implementations outside FIPS mode.
6611 * Transparently support X9.42 DH parameters when calling
6612 PEM_read_bio_DHparameters. This means existing applications can handle
6613 the new parameter format automatically.
6617 * Initial experimental support for X9.42 DH parameter format: mainly
6618 to support use of 'q' parameter for RFC5114 parameters.
6622 * Add DH parameters from RFC5114 including test data to dhtest.
6626 * Support for automatic EC temporary key parameter selection. If enabled
6627 the most preferred EC parameters are automatically used instead of
6628 hardcoded fixed parameters. Now a server just has to call:
6629 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6630 support ECDH and use the most appropriate parameters.
6634 * Enhance and tidy EC curve and point format TLS extension code. Use
6635 static structures instead of allocation if default values are used.
6636 New ctrls to set curves we wish to support and to retrieve shared curves.
6637 Print out shared curves in s_server. New options to s_server and s_client
6638 to set list of supported curves.
6642 * New ctrls to retrieve supported signature algorithms and
6643 supported curve values as an array of NIDs. Extend openssl utility
6644 to print out received values.
6648 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6649 between NIDs and the more common NIST names such as "P-256". Enhance
6650 ecparam utility and ECC method to recognise the NIST names for curves.
6654 * Enhance SSL/TLS certificate chain handling to support different
6655 chains for each certificate instead of one chain in the parent SSL_CTX.
6659 * Support for fixed DH ciphersuite client authentication: where both
6660 server and client use DH certificates with common parameters.
6664 * Support for fixed DH ciphersuites: those requiring DH server
6669 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6671 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6672 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6673 X509_CINF_get_signature were reverted post internal team review.
6678 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6680 * OCSP Status Request extension unbounded memory growth
6682 A malicious client can send an excessively large OCSP Status Request
6683 extension. If that client continually requests renegotiation, sending a
6684 large OCSP Status Request extension each time, then there will be unbounded
6685 memory growth on the server. This will eventually lead to a Denial Of
6686 Service attack through memory exhaustion. Servers with a default
6687 configuration are vulnerable even if they do not support OCSP. Builds using
6688 the "no-ocsp" build time option are not affected.
6690 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6695 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6698 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6704 * OOB write in MDC2_Update()
6706 An overflow can occur in MDC2_Update() either if called directly or
6707 through the EVP_DigestUpdate() function using MDC2. If an attacker
6708 is able to supply very large amounts of input data after a previous
6709 call to EVP_EncryptUpdate() with a partial block then a length check
6710 can overflow resulting in a heap corruption.
6712 The amount of data needed is comparable to SIZE_MAX which is impractical
6715 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6720 * Malformed SHA512 ticket DoS
6722 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6723 DoS attack where a malformed ticket will result in an OOB read which will
6726 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6727 a custom server callback and ticket lookup mechanism.
6729 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6734 * OOB write in BN_bn2dec()
6736 The function BN_bn2dec() does not check the return value of BN_div_word().
6737 This can cause an OOB write if an application uses this function with an
6738 overly large BIGNUM. This could be a problem if an overly large certificate
6739 or CRL is printed out from an untrusted source. TLS is not affected because
6740 record limits will reject an oversized certificate before it is parsed.
6742 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6747 * OOB read in TS_OBJ_print_bio()
6749 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6750 the total length the OID text representation would use and not the amount
6751 of data written. This will result in OOB reads when large OIDs are
6754 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6759 * Pointer arithmetic undefined behaviour
6761 Avoid some undefined pointer arithmetic
6763 A common idiom in the codebase is to check limits in the following manner:
6766 Where "p" points to some malloc'd data of SIZE bytes and
6769 "len" here could be from some externally supplied data (e.g. from a TLS
6772 The rules of C pointer arithmetic are such that "p + len" is only well
6773 defined where len <= SIZE. Therefore the above idiom is actually
6774 undefined behaviour.
6776 For example this could cause problems if some malloc implementation
6777 provides an address for "p" such that "p + len" actually overflows for
6778 values of len that are too big and therefore p + len < limit.
6780 This issue was reported to OpenSSL by Guido Vranken
6785 * Constant time flag not preserved in DSA signing
6787 Operations in the DSA signing algorithm should run in constant time in
6788 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6789 implementation means that a non-constant time codepath is followed for
6790 certain operations. This has been demonstrated through a cache-timing
6791 attack to be sufficient for an attacker to recover the private DSA key.
6793 This issue was reported by César Pereida (Aalto University), Billy Brumley
6794 (Tampere University of Technology), and Yuval Yarom (The University of
6795 Adelaide and NICTA).
6800 * DTLS buffered message DoS
6802 In a DTLS connection where handshake messages are delivered out-of-order
6803 those messages that OpenSSL is not yet ready to process will be buffered
6804 for later use. Under certain circumstances, a flaw in the logic means that
6805 those messages do not get removed from the buffer even though the handshake
6806 has been completed. An attacker could force up to approx. 15 messages to
6807 remain in the buffer when they are no longer required. These messages will
6808 be cleared when the DTLS connection is closed. The default maximum size for
6809 a message is 100k. Therefore the attacker could force an additional 1500k
6810 to be consumed per connection. By opening many simulataneous connections an
6811 attacker could cause a DoS attack through memory exhaustion.
6813 This issue was reported to OpenSSL by Quan Luo.
6818 * DTLS replay protection DoS
6820 A flaw in the DTLS replay attack protection mechanism means that records
6821 that arrive for future epochs update the replay protection "window" before
6822 the MAC for the record has been validated. This could be exploited by an
6823 attacker by sending a record for the next epoch (which does not have to
6824 decrypt or have a valid MAC), with a very large sequence number. This means
6825 that all subsequent legitimate packets are dropped causing a denial of
6826 service for a specific DTLS connection.
6828 This issue was reported to OpenSSL by the OCAP audit team.
6833 * Certificate message OOB reads
6835 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6836 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6837 theoretical DoS risk but this has not been observed in practice on common
6840 The messages affected are client certificate, client certificate request
6841 and server certificate. As a result the attack can only be performed
6842 against a client or a server which enables client authentication.
6844 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6849 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6851 * Prevent padding oracle in AES-NI CBC MAC check
6853 A MITM attacker can use a padding oracle attack to decrypt traffic
6854 when the connection uses an AES CBC cipher and the server support
6857 This issue was introduced as part of the fix for Lucky 13 padding
6858 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6859 constant time by making sure that always the same bytes are read and
6860 compared against either the MAC or padding bytes. But it no longer
6861 checked that there was enough data to have both the MAC and padding
6864 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6869 * Fix EVP_EncodeUpdate overflow
6871 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6872 Base64 encoding of binary data. If an attacker is able to supply very large
6873 amounts of input data then a length check can overflow resulting in a heap
6876 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6877 the `PEM_write_bio*` family of functions. These are mainly used within the
6878 OpenSSL command line applications, so any application which processes data
6879 from an untrusted source and outputs it as a PEM file should be considered
6880 vulnerable to this issue. User applications that call these APIs directly
6881 with large amounts of untrusted data may also be vulnerable.
6883 This issue was reported by Guido Vranken.
6888 * Fix EVP_EncryptUpdate overflow
6890 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6891 is able to supply very large amounts of input data after a previous call to
6892 EVP_EncryptUpdate() with a partial block then a length check can overflow
6893 resulting in a heap corruption. Following an analysis of all OpenSSL
6894 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6895 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6896 the first called function after an EVP_EncryptInit(), and therefore that
6897 specific call must be safe. The second form is where the length passed to
6898 EVP_EncryptUpdate() can be seen from the code to be some small value and
6899 therefore there is no possibility of an overflow. Since all instances are
6900 one of these two forms, it is believed that there can be no overflows in
6901 internal code due to this problem. It should be noted that
6902 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6903 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6904 of these calls have also been analysed too and it is believed there are no
6905 instances in internal usage where an overflow could occur.
6907 This issue was reported by Guido Vranken.
6912 * Prevent ASN.1 BIO excessive memory allocation
6914 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6915 a short invalid encoding can casuse allocation of large amounts of memory
6916 potentially consuming excessive resources or exhausting memory.
6918 Any application parsing untrusted data through d2i BIO functions is
6919 affected. The memory based functions such as d2i_X509() are *not* affected.
6920 Since the memory based functions are used by the TLS library, TLS
6921 applications are not affected.
6923 This issue was reported by Brian Carpenter.
6930 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6931 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6932 in arbitrary stack data being returned in the buffer.
6934 This issue was reported by Guido Vranken.
6939 * Modify behavior of ALPN to invoke callback after SNI/servername
6940 callback, such that updates to the SSL_CTX affect ALPN.
6944 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6949 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6950 methods are enabled and ssl2 is disabled the methods return NULL.
6954 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6956 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6957 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6958 provide any "EXPORT" or "LOW" strength ciphers.
6962 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6963 is by default disabled at build-time. Builds that are not configured with
6964 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6965 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6966 will need to explicitly call either of:
6968 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6970 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6972 as appropriate. Even if either of those is used, or the application
6973 explicitly uses the version-specific SSLv2_method() or its client and
6974 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6975 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6976 ciphers, and SSLv2 56-bit DES are no longer available.
6981 * Fix a double-free in DSA code
6983 A double free bug was discovered when OpenSSL parses malformed DSA private
6984 keys and could lead to a DoS attack or memory corruption for applications
6985 that receive DSA private keys from untrusted sources. This scenario is
6988 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6994 * Disable SRP fake user seed to address a server memory leak.
6996 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6998 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6999 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7000 was changed to ignore the "fake user" SRP seed, even if the seed
7003 Users should use SRP_VBASE_get1_by_user instead. Note that in
7004 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7005 also that even though configuring the SRP seed attempts to hide
7006 invalid usernames by continuing the handshake with fake
7007 credentials, this behaviour is not constant time and no strong
7008 guarantees are made that the handshake is indistinguishable from
7009 that of a valid user.
7014 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7016 In the BN_hex2bn function the number of hex digits is calculated using an
7017 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7018 large values of `i` this can result in `bn_expand` not allocating any
7019 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7020 field as NULL leading to a subsequent NULL ptr deref. For very large values
7021 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7022 In this case memory is allocated to the internal BIGNUM data field, but it
7023 is insufficiently sized leading to heap corruption. A similar issue exists
7024 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7025 is ever called by user applications with very large untrusted hex/dec data.
7026 This is anticipated to be a rare occurrence.
7028 All OpenSSL internal usage of these functions use data that is not expected
7029 to be untrusted, e.g. config file data or application command line
7030 arguments. If user developed applications generate config file data based
7031 on untrusted data then it is possible that this could also lead to security
7032 consequences. This is also anticipated to be rare.
7034 This issue was reported to OpenSSL by Guido Vranken.
7039 * Fix memory issues in `BIO_*printf` functions
7041 The internal `fmtstr` function used in processing a "%s" format string in
7042 the `BIO_*printf` functions could overflow while calculating the length of a
7043 string and cause an OOB read when printing very long strings.
7045 Additionally the internal `doapr_outch` function can attempt to write to an
7046 OOB memory location (at an offset from the NULL pointer) in the event of a
7047 memory allocation failure. In 1.0.2 and below this could be caused where
7048 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7049 could be in processing a very long "%s" format string. Memory leaks can
7052 The first issue may mask the second issue dependent on compiler behaviour.
7053 These problems could enable attacks where large amounts of untrusted data
7054 is passed to the `BIO_*printf` functions. If applications use these functions
7055 in this way then they could be vulnerable. OpenSSL itself uses these
7056 functions when printing out human-readable dumps of ASN.1 data. Therefore
7057 applications that print this data could be vulnerable if the data is from
7058 untrusted sources. OpenSSL command line applications could also be
7059 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7060 as command line arguments.
7062 Libssl is not considered directly vulnerable. Additionally certificates etc
7063 received via remote connections via libssl are also unlikely to be able to
7064 trigger these issues because of message size limits enforced within libssl.
7066 This issue was reported to OpenSSL Guido Vranken.
7071 * Side channel attack on modular exponentiation
7073 A side-channel attack was found which makes use of cache-bank conflicts on
7074 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7075 of RSA keys. The ability to exploit this issue is limited as it relies on
7076 an attacker who has control of code in a thread running on the same
7077 hyper-threaded core as the victim thread which is performing decryptions.
7079 This issue was reported to OpenSSL by Yuval Yarom, The University of
7080 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7081 Nadia Heninger, University of Pennsylvania with more information at
7082 <http://cachebleed.info>.
7087 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7088 if no keysize is specified with default_bits. This fixes an
7089 omission in an earlier change that changed all RSA/DSA key generation
7090 commands to use 2048 bits by default.
7094 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7096 * Protection for DH small subgroup attacks
7098 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7099 switched on by default and cannot be disabled. This could have some
7104 * SSLv2 doesn't block disabled ciphers
7106 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7107 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7108 been disabled, provided that the SSLv2 protocol was not also disabled via
7111 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7112 and Sebastian Schinzel.
7117 * Reject DH handshakes with parameters shorter than 1024 bits.
7121 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7123 * Certificate verify crash with missing PSS parameter
7125 The signature verification routines will crash with a NULL pointer
7126 dereference if presented with an ASN.1 signature using the RSA PSS
7127 algorithm and absent mask generation function parameter. Since these
7128 routines are used to verify certificate signature algorithms this can be
7129 used to crash any certificate verification operation and exploited in a
7130 DoS attack. Any application which performs certificate verification is
7131 vulnerable including OpenSSL clients and servers which enable client
7134 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7139 * X509_ATTRIBUTE memory leak
7141 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7142 memory. This structure is used by the PKCS#7 and CMS routines so any
7143 application which reads PKCS#7 or CMS data from untrusted sources is
7144 affected. SSL/TLS is not affected.
7146 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7152 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7153 This changes the decoding behaviour for some invalid messages,
7154 though the change is mostly in the more lenient direction, and
7155 legacy behaviour is preserved as much as possible.
7159 * In DSA_generate_parameters_ex, if the provided seed is too short,
7160 use a random seed, as already documented.
7162 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7164 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7166 * Alternate chains certificate forgery
7168 During certificate verfification, OpenSSL will attempt to find an
7169 alternative certificate chain if the first attempt to build such a chain
7170 fails. An error in the implementation of this logic can mean that an
7171 attacker could cause certain checks on untrusted certificates to be
7172 bypassed, such as the CA flag, enabling them to use a valid leaf
7173 certificate to act as a CA and "issue" an invalid certificate.
7175 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7181 * Race condition handling PSK identify hint
7183 If PSK identity hints are received by a multi-threaded client then
7184 the values are wrongly updated in the parent SSL_CTX structure. This can
7185 result in a race condition potentially leading to a double free of the
7191 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7193 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7194 incompatibility in the handling of HMAC. The previous ABI has now been
7197 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7199 * Malformed ECParameters causes infinite loop
7201 When processing an ECParameters structure OpenSSL enters an infinite loop
7202 if the curve specified is over a specially malformed binary polynomial
7205 This can be used to perform denial of service against any
7206 system which processes public keys, certificate requests or
7207 certificates. This includes TLS clients and TLS servers with
7208 client authentication enabled.
7210 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7215 * Exploitable out-of-bounds read in X509_cmp_time
7217 X509_cmp_time does not properly check the length of the ASN1_TIME
7218 string and can read a few bytes out of bounds. In addition,
7219 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7222 An attacker can use this to craft malformed certificates and CRLs of
7223 various sizes and potentially cause a segmentation fault, resulting in
7224 a DoS on applications that verify certificates or CRLs. TLS clients
7225 that verify CRLs are affected. TLS clients and servers with client
7226 authentication enabled may be affected if they use custom verification
7229 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7230 independently by Hanno Böck.
7235 * PKCS7 crash with missing EnvelopedContent
7237 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7238 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7239 with missing content and trigger a NULL pointer dereference on parsing.
7241 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7242 structures from untrusted sources are affected. OpenSSL clients and
7243 servers are not affected.
7245 This issue was reported to OpenSSL by Michal Zalewski (Google).
7250 * CMS verify infinite loop with unknown hash function
7252 When verifying a signedData message the CMS code can enter an infinite loop
7253 if presented with an unknown hash function OID. This can be used to perform
7254 denial of service against any system which verifies signedData messages using
7256 This issue was reported to OpenSSL by Johannes Bauer.
7261 * Race condition handling NewSessionTicket
7263 If a NewSessionTicket is received by a multi-threaded client when attempting to
7264 reuse a previous ticket then a race condition can occur potentially leading to
7265 a double free of the ticket data.
7270 * Reject DH handshakes with parameters shorter than 768 bits.
7272 *Kurt Roeckx and Emilia Kasper*
7274 * dhparam: generate 2048-bit parameters by default.
7276 *Kurt Roeckx and Emilia Kasper*
7278 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7280 * Segmentation fault in ASN1_TYPE_cmp fix
7282 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7283 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7284 certificate signature algorithm consistency this can be used to crash any
7285 certificate verification operation and exploited in a DoS attack. Any
7286 application which performs certificate verification is vulnerable including
7287 OpenSSL clients and servers which enable client authentication.
7292 * ASN.1 structure reuse memory corruption fix
7294 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7295 memory corruption via an invalid write. Such reuse is and has been
7296 strongly discouraged and is believed to be rare.
7298 Applications that parse structures containing CHOICE or ANY DEFINED BY
7299 components may be affected. Certificate parsing (d2i_X509 and related
7300 functions) are however not affected. OpenSSL clients and servers are
7306 * PKCS7 NULL pointer dereferences fix
7308 The PKCS#7 parsing code does not handle missing outer ContentInfo
7309 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7310 missing content and trigger a NULL pointer dereference on parsing.
7312 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7313 otherwise parse PKCS#7 structures from untrusted sources are
7314 affected. OpenSSL clients and servers are not affected.
7316 This issue was reported to OpenSSL by Michal Zalewski (Google).
7321 * DoS via reachable assert in SSLv2 servers fix
7323 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7324 servers that both support SSLv2 and enable export cipher suites by sending
7325 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7327 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7328 (OpenSSL development team).
7333 * Use After Free following d2i_ECPrivatekey error fix
7335 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7336 could cause a use after free condition. This, in turn, could cause a double
7337 free in several private key parsing functions (such as d2i_PrivateKey
7338 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7339 for applications that receive EC private keys from untrusted
7340 sources. This scenario is considered rare.
7342 This issue was discovered by the BoringSSL project and fixed in their
7348 * X509_to_X509_REQ NULL pointer deref fix
7350 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7351 the certificate key is invalid. This function is rarely used in practice.
7353 This issue was discovered by Brian Carpenter.
7358 * Removed the export ciphers from the DEFAULT ciphers
7362 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7364 * Build fixes for the Windows and OpenVMS platforms
7366 *Matt Caswell and Richard Levitte*
7368 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7370 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7371 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7372 dereference. This could lead to a Denial Of Service attack. Thanks to
7373 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7378 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7379 dtls1_buffer_record function under certain conditions. In particular this
7380 could occur if an attacker sent repeated DTLS records with the same
7381 sequence number but for the next epoch. The memory leak could be exploited
7382 by an attacker in a Denial of Service attack through memory exhaustion.
7383 Thanks to Chris Mueller for reporting this issue.
7388 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7389 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7390 method would be set to NULL which could later result in a NULL pointer
7391 dereference. Thanks to Frank Schmirler for reporting this issue.
7396 * Abort handshake if server key exchange message is omitted for ephemeral
7399 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7400 reporting this issue.
7405 * Remove non-export ephemeral RSA code on client and server. This code
7406 violated the TLS standard by allowing the use of temporary RSA keys in
7407 non-export ciphersuites and could be used by a server to effectively
7408 downgrade the RSA key length used to a value smaller than the server
7409 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7410 INRIA or reporting this issue.
7415 * Fixed issue where DH client certificates are accepted without verification.
7416 An OpenSSL server will accept a DH certificate for client authentication
7417 without the certificate verify message. This effectively allows a client to
7418 authenticate without the use of a private key. This only affects servers
7419 which trust a client certificate authority which issues certificates
7420 containing DH keys: these are extremely rare and hardly ever encountered.
7421 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7427 * Ensure that the session ID context of an SSL is updated when its
7428 SSL_CTX is updated via SSL_set_SSL_CTX.
7430 The session ID context is typically set from the parent SSL_CTX,
7431 and can vary with the CTX.
7435 * Fix various certificate fingerprint issues.
7437 By using non-DER or invalid encodings outside the signed portion of a
7438 certificate the fingerprint can be changed without breaking the signature.
7439 Although no details of the signed portion of the certificate can be changed
7440 this can cause problems with some applications: e.g. those using the
7441 certificate fingerprint for blacklists.
7443 1. Reject signatures with non zero unused bits.
7445 If the BIT STRING containing the signature has non zero unused bits reject
7446 the signature. All current signature algorithms require zero unused bits.
7448 2. Check certificate algorithm consistency.
7450 Check the AlgorithmIdentifier inside TBS matches the one in the
7451 certificate signature. NB: this will result in signature failure
7452 errors for some broken certificates.
7454 Thanks to Konrad Kraszewski from Google for reporting this issue.
7456 3. Check DSA/ECDSA signatures use DER.
7458 Re-encode DSA/ECDSA signatures and compare with the original received
7459 signature. Return an error if there is a mismatch.
7461 This will reject various cases including garbage after signature
7462 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7463 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7464 (negative or with leading zeroes).
7466 Further analysis was conducted and fixes were developed by Stephen Henson
7467 of the OpenSSL core team.
7473 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7474 results on some platforms, including x86_64. This bug occurs at random
7475 with a very low probability, and is not known to be exploitable in any
7476 way, though its exact impact is difficult to determine. Thanks to Pieter
7477 Wuille (Blockstream) who reported this issue and also suggested an initial
7478 fix. Further analysis was conducted by the OpenSSL development team and
7479 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7480 the OpenSSL core team.
7485 * Do not resume sessions on the server if the negotiated protocol
7486 version does not match the session's version. Resuming with a different
7487 version, while not strictly forbidden by the RFC, is of questionable
7488 sanity and breaks all known clients.
7490 *David Benjamin, Emilia Käsper*
7492 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7493 early CCS messages during renegotiation. (Note that because
7494 renegotiation is encrypted, this early CCS was not exploitable.)
7498 * Tighten client-side session ticket handling during renegotiation:
7499 ensure that the client only accepts a session ticket if the server sends
7500 the extension anew in the ServerHello. Previously, a TLS client would
7501 reuse the old extension state and thus accept a session ticket if one was
7502 announced in the initial ServerHello.
7504 Similarly, ensure that the client requires a session ticket if one
7505 was advertised in the ServerHello. Previously, a TLS client would
7506 ignore a missing NewSessionTicket message.
7510 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7514 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7515 sends a carefully crafted handshake message, to cause OpenSSL to fail
7516 to free up to 64k of memory causing a memory leak. This could be
7517 exploited in a Denial Of Service attack. This issue affects OpenSSL
7518 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7519 whether SRTP is used or configured. Implementations of OpenSSL that
7520 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7522 The fix was developed by the OpenSSL team.
7527 * Session Ticket Memory Leak.
7529 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7530 integrity of that ticket is first verified. In the event of a session
7531 ticket integrity check failing, OpenSSL will fail to free memory
7532 causing a memory leak. By sending a large number of invalid session
7533 tickets an attacker could exploit this issue in a Denial Of Service
7539 * Build option no-ssl3 is incomplete.
7541 When OpenSSL is configured with "no-ssl3" as a build option, servers
7542 could accept and complete a SSL 3.0 handshake, and clients could be
7543 configured to send them.
7546 *Akamai and the OpenSSL team*
7548 * Add support for TLS_FALLBACK_SCSV.
7549 Client applications doing fallback retries should call
7550 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7553 *Adam Langley, Bodo Moeller*
7555 * Add additional DigestInfo checks.
7557 Re-encode DigestInto in DER and check against the original when
7558 verifying RSA signature: this will reject any improperly encoded
7559 DigestInfo structures.
7561 Note: this is a precautionary measure and no attacks are currently known.
7565 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7567 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7568 SRP code can be overrun an internal buffer. Add sanity check that
7569 g, A, B < N to SRP code.
7571 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7572 Group for discovering this issue.
7577 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7578 TLS 1.0 instead of higher protocol versions when the ClientHello message
7579 is badly fragmented. This allows a man-in-the-middle attacker to force a
7580 downgrade to TLS 1.0 even if both the server and the client support a
7581 higher protocol version, by modifying the client's TLS records.
7583 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7584 researching this issue.
7589 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7590 to a denial of service attack. A malicious server can crash the client
7591 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7592 ciphersuite and sending carefully crafted handshake messages.
7594 Thanks to Felix Gröbert (Google) for discovering and researching this
7600 * By sending carefully crafted DTLS packets an attacker could cause openssl
7601 to leak memory. This can be exploited through a Denial of Service attack.
7602 Thanks to Adam Langley for discovering and researching this issue.
7607 * An attacker can force openssl to consume large amounts of memory whilst
7608 processing DTLS handshake messages. This can be exploited through a
7609 Denial of Service attack.
7610 Thanks to Adam Langley for discovering and researching this issue.
7615 * An attacker can force an error condition which causes openssl to crash
7616 whilst processing DTLS packets due to memory being freed twice. This
7617 can be exploited through a Denial of Service attack.
7618 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7624 * If a multithreaded client connects to a malicious server using a resumed
7625 session and the server sends an ec point format extension it could write
7626 up to 255 bytes to freed memory.
7628 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7634 * A malicious server can crash an OpenSSL client with a null pointer
7635 dereference (read) by specifying an SRP ciphersuite even though it was not
7636 properly negotiated with the client. This can be exploited through a
7637 Denial of Service attack.
7639 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7640 discovering and researching this issue.
7645 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7646 X509_name_oneline, X509_name_print_ex et al. to leak some information
7647 from the stack. Applications may be affected if they echo pretty printing
7648 output to the attacker.
7650 Thanks to Ivan Fratric (Google) for discovering this issue.
7653 *Emilia Käsper, and Steve Henson*
7655 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7656 for corner cases. (Certain input points at infinity could lead to
7657 bogus results, with non-infinity inputs mapped to infinity too.)
7661 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7663 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7664 handshake can force the use of weak keying material in OpenSSL
7665 SSL/TLS clients and servers.
7667 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7668 researching this issue. ([CVE-2014-0224])
7670 *KIKUCHI Masashi, Steve Henson*
7672 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7673 OpenSSL DTLS client the code can be made to recurse eventually crashing
7676 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7679 *Imre Rad, Steve Henson*
7681 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7682 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7683 client or server. This is potentially exploitable to run arbitrary
7684 code on a vulnerable client or server.
7686 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7688 *Jüri Aedla, Steve Henson*
7690 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7691 are subject to a denial of service attack.
7693 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7694 this issue. ([CVE-2014-3470])
7696 *Felix Gröbert, Ivan Fratric, Steve Henson*
7698 * Harmonize version and its documentation. -f flag is used to display
7701 *mancha <mancha1@zoho.com>*
7703 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7704 in i2d_ECPrivateKey.
7706 *mancha <mancha1@zoho.com>*
7708 * Fix some double frees. These are not thought to be exploitable.
7710 *mancha <mancha1@zoho.com>*
7712 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7714 * A missing bounds check in the handling of the TLS heartbeat extension
7715 can be used to reveal up to 64k of memory to a connected client or
7718 Thanks for Neel Mehta of Google Security for discovering this bug and to
7719 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7720 preparing the fix ([CVE-2014-0160])
7722 *Adam Langley, Bodo Moeller*
7724 * Fix for the attack described in the paper "Recovering OpenSSL
7725 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7726 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7727 <http://eprint.iacr.org/2014/140>
7729 Thanks to Yuval Yarom and Naomi Benger for discovering this
7730 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7732 *Yuval Yarom and Naomi Benger*
7734 * TLS pad extension: draft-agl-tls-padding-03
7736 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7737 TLS client Hello record length value would otherwise be > 255 and
7738 less that 512 pad with a dummy extension containing zeroes so it
7739 is at least 512 bytes long.
7741 *Adam Langley, Steve Henson*
7743 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7745 * Fix for TLS record tampering bug. A carefully crafted invalid
7746 handshake could crash OpenSSL with a NULL pointer exception.
7747 Thanks to Anton Johansson for reporting this issues.
7750 * Keep original DTLS digest and encryption contexts in retransmission
7751 structures so we can use the previous session parameters if they need
7752 to be resent. ([CVE-2013-6450])
7756 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7757 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7758 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7759 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7760 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7761 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7763 *Rob Stradling, Adam Langley*
7765 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7767 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7768 supporting platforms or when small records were transferred.
7770 *Andy Polyakov, Steve Henson*
7772 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7774 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7776 This addresses the flaw in CBC record processing discovered by
7777 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7778 at: <http://www.isg.rhul.ac.uk/tls/>
7780 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7781 Security Group at Royal Holloway, University of London
7782 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7783 Emilia Käsper for the initial patch.
7786 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7788 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7789 ciphersuites which can be exploited in a denial of service attack.
7790 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7791 and detecting this bug and to Wolfgang Ettlinger
7792 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7797 * Return an error when checking OCSP signatures when key is NULL.
7798 This fixes a DoS attack. ([CVE-2013-0166])
7802 * Make openssl verify return errors.
7804 *Chris Palmer <palmer@google.com> and Ben Laurie*
7806 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7807 the right response is stapled. Also change SSL_get_certificate()
7808 so it returns the certificate actually sent.
7809 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7811 *Rob Stradling <rob.stradling@comodo.com>*
7813 * Fix possible deadlock when decoding public keys.
7817 * Don't use TLS 1.0 record version number in initial client hello
7822 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7824 * Sanity check record length before skipping explicit IV in TLS
7825 1.2, 1.1 and DTLS to fix DoS attack.
7827 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7828 fuzzing as a service testing platform.
7833 * Initialise tkeylen properly when encrypting CMS messages.
7834 Thanks to Solar Designer of Openwall for reporting this issue.
7838 * In FIPS mode don't try to use composite ciphers as they are not
7843 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7845 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7846 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7847 mean any application compiled against OpenSSL 1.0.0 headers setting
7848 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7849 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7850 0x10000000L Any application which was previously compiled against
7851 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7852 will need to be recompiled as a result. Letting be results in
7853 inability to disable specifically TLS 1.1 and in client context,
7854 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7858 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7859 disable just protocol X, but all protocols above X *if* there are
7860 protocols *below* X still enabled. In more practical terms it means
7861 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7862 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7863 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7868 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7870 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7871 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7872 in CRYPTO_realloc_clean.
7874 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7875 issue and to Adam Langley <agl@chromium.org> for fixing it.
7878 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7880 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7884 * Workarounds for some broken servers that "hang" if a client hello
7885 record length exceeds 255 bytes.
7887 1. Do not use record version number > TLS 1.0 in initial client
7888 hello: some (but not all) hanging servers will now work.
7889 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7890 the number of ciphers sent in the client hello. This should be
7891 set to an even number, such as 50, for example by passing:
7892 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7893 Most broken servers should now work.
7894 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7895 TLS 1.2 client support entirely.
7899 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7903 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7905 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7906 STRING form instead of a DigestInfo.
7910 * The format used for MDC2 RSA signatures is inconsistent between EVP
7911 and the RSA_sign/RSA_verify functions. This was made more apparent when
7912 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7913 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7914 the correct format in RSA_verify so both forms transparently work.
7918 * Some servers which support TLS 1.0 can choke if we initially indicate
7919 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7920 encrypted premaster secret. As a workaround use the maximum permitted
7921 client version in client hello, this should keep such servers happy
7922 and still work with previous versions of OpenSSL.
7926 * Add support for TLS/DTLS heartbeats.
7928 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7930 * Add support for SCTP.
7932 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7934 * Improved PRNG seeding for VOS.
7936 *Paul Green <Paul.Green@stratus.com>*
7938 * Extensive assembler packs updates, most notably:
7940 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7941 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7942 - x86_64: bit-sliced AES implementation;
7943 - ARM: NEON support, contemporary platforms optimizations;
7944 - s390x: z196 support;
7945 - `*`: GHASH and GF(2^m) multiplication implementations;
7949 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7950 (removal of unnecessary code)
7952 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7954 * Add TLS key material exporter from RFC 5705.
7958 * Add DTLS-SRTP negotiation from RFC 5764.
7962 * Add Next Protocol Negotiation,
7963 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7964 disabled with a no-npn flag to config or Configure. Code donated
7967 *Adam Langley <agl@google.com> and Ben Laurie*
7969 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7970 NIST-P256, NIST-P521, with constant-time single point multiplication on
7971 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7972 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7973 Code made available under Apache License version 2.0.
7975 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7976 line to include this in your build of OpenSSL, and run "make depend" (or
7977 "make update"). This enables the following EC_METHODs:
7979 EC_GFp_nistp224_method()
7980 EC_GFp_nistp256_method()
7981 EC_GFp_nistp521_method()
7983 EC_GROUP_new_by_curve_name() will automatically use these (while
7984 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7987 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7989 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7990 all platforms. Move ssize_t definition from e_os.h to the public
7991 header file e_os2.h as it now appears in public header file cms.h
7995 * New -sigopt option to the ca, req and x509 utilities. Additional
7996 signature parameters can be passed using this option and in
8001 * Add RSA PSS signing function. This will generate and set the
8002 appropriate AlgorithmIdentifiers for PSS based on those in the
8003 corresponding EVP_MD_CTX structure. No application support yet.
8007 * Support for companion algorithm specific ASN1 signing routines.
8008 New function ASN1_item_sign_ctx() signs a pre-initialised
8009 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8010 the appropriate parameters.
8014 * Add new algorithm specific ASN1 verification initialisation function
8015 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8016 handling will be the same no matter what EVP_PKEY_METHOD is used.
8017 Add a PSS handler to support verification of PSS signatures: checked
8018 against a number of sample certificates.
8022 * Add signature printing for PSS. Add PSS OIDs.
8024 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8026 * Add algorithm specific signature printing. An individual ASN1 method
8027 can now print out signatures instead of the standard hex dump.
8029 More complex signatures (e.g. PSS) can print out more meaningful
8030 information. Include DSA version that prints out the signature
8035 * Password based recipient info support for CMS library: implementing
8040 * Split password based encryption into PBES2 and PBKDF2 functions. This
8041 neatly separates the code into cipher and PBE sections and is required
8042 for some algorithms that split PBES2 into separate pieces (such as
8043 password based CMS).
8047 * Session-handling fixes:
8048 - Fix handling of connections that are resuming with a session ID,
8049 but also support Session Tickets.
8050 - Fix a bug that suppressed issuing of a new ticket if the client
8051 presented a ticket with an expired session.
8052 - Try to set the ticket lifetime hint to something reasonable.
8053 - Make tickets shorter by excluding irrelevant information.
8054 - On the client side, don't ignore renewed tickets.
8056 *Adam Langley, Bodo Moeller (Google)*
8058 * Fix PSK session representation.
8062 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8064 This work was sponsored by Intel.
8068 * Add GCM support to TLS library. Some custom code is needed to split
8069 the IV between the fixed (from PRF) and explicit (from TLS record)
8070 portions. This adds all GCM ciphersuites supported by RFC5288 and
8071 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8072 add a special AESGCM string for GCM only.
8076 * Expand range of ctrls for AES GCM. Permit setting invocation
8077 field on decrypt and retrieval of invocation field only on encrypt.
8081 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8082 As required by RFC5289 these ciphersuites cannot be used if for
8083 versions of TLS earlier than 1.2.
8087 * For FIPS capable OpenSSL interpret a NULL default public key method
8088 as unset and return the appropriate default but do *not* set the default.
8089 This means we can return the appropriate method in applications that
8090 switch between FIPS and non-FIPS modes.
8094 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8095 ENGINE is used then we cannot handle that in the FIPS module so we
8096 keep original code iff non-FIPS operations are allowed.
8100 * Add -attime option to openssl utilities.
8102 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8104 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8108 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8109 FIPS EC methods unconditionally for now.
8113 * New build option no-ec2m to disable characteristic 2 code.
8117 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8118 all cases can be covered as some introduce binary incompatibilities.
8122 * Redirect RSA operations to FIPS module including keygen,
8123 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8127 * Add similar low-level API blocking to ciphers.
8131 * low-level digest APIs are not approved in FIPS mode: any attempt
8132 to use these will cause a fatal error. Applications that *really* want
8133 to use them can use the `private_*` version instead.
8137 * Redirect cipher operations to FIPS module for FIPS builds.
8141 * Redirect digest operations to FIPS module for FIPS builds.
8145 * Update build system to add "fips" flag which will link in fipscanister.o
8146 for static and shared library builds embedding a signature if needed.
8150 * Output TLS supported curves in preference order instead of numerical
8151 order. This is currently hardcoded for the highest order curves first.
8152 This should be configurable so applications can judge speed vs strength.
8156 * Add TLS v1.2 server support for client authentication.
8160 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8165 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8166 FIPS modules versions.
8170 * Add TLS v1.2 client side support for client authentication. Keep cache
8171 of handshake records longer as we don't know the hash algorithm to use
8172 until after the certificate request message is received.
8176 * Initial TLS v1.2 client support. Add a default signature algorithms
8177 extension including all the algorithms we support. Parse new signature
8178 format in client key exchange. Relax some ECC signing restrictions for
8179 TLS v1.2 as indicated in RFC5246.
8183 * Add server support for TLS v1.2 signature algorithms extension. Switch
8184 to new signature format when needed using client digest preference.
8185 All server ciphersuites should now work correctly in TLS v1.2. No client
8186 support yet and no support for client certificates.
8190 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8191 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8192 ciphersuites. At present only RSA key exchange ciphersuites work with
8193 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8194 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8195 and version checking.
8199 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8200 with this defined it will not be affected by any changes to ssl internal
8201 structures. Add several utility functions to allow openssl application
8202 to work with OPENSSL_NO_SSL_INTERN defined.
8206 * A long standing patch to add support for SRP from EdelWeb (Peter
8207 Sylvester and Christophe Renou) was integrated.
8208 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8209 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8212 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8216 * Permit abbreviated handshakes when renegotiating using the function
8217 SSL_renegotiate_abbreviated().
8219 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8221 * Add call to ENGINE_register_all_complete() to
8222 ENGINE_load_builtin_engines(), so some implementations get used
8223 automatically instead of needing explicit application support.
8227 * Add support for TLS key exporter as described in RFC5705.
8229 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8231 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8232 a few changes are required:
8234 Add SSL_OP_NO_TLSv1_1 flag.
8235 Add TLSv1_1 methods.
8236 Update version checking logic to handle version 1.1.
8237 Add explicit IV handling (ported from DTLS code).
8238 Add command line options to s_client/s_server.
8245 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8247 * X509_ATTRIBUTE memory leak
8249 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8250 memory. This structure is used by the PKCS#7 and CMS routines so any
8251 application which reads PKCS#7 or CMS data from untrusted sources is
8252 affected. SSL/TLS is not affected.
8254 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8260 * Race condition handling PSK identify hint
8262 If PSK identity hints are received by a multi-threaded client then
8263 the values are wrongly updated in the parent SSL_CTX structure. This can
8264 result in a race condition potentially leading to a double free of the
8270 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8272 * Malformed ECParameters causes infinite loop
8274 When processing an ECParameters structure OpenSSL enters an infinite loop
8275 if the curve specified is over a specially malformed binary polynomial
8278 This can be used to perform denial of service against any
8279 system which processes public keys, certificate requests or
8280 certificates. This includes TLS clients and TLS servers with
8281 client authentication enabled.
8283 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8288 * Exploitable out-of-bounds read in X509_cmp_time
8290 X509_cmp_time does not properly check the length of the ASN1_TIME
8291 string and can read a few bytes out of bounds. In addition,
8292 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8295 An attacker can use this to craft malformed certificates and CRLs of
8296 various sizes and potentially cause a segmentation fault, resulting in
8297 a DoS on applications that verify certificates or CRLs. TLS clients
8298 that verify CRLs are affected. TLS clients and servers with client
8299 authentication enabled may be affected if they use custom verification
8302 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8303 independently by Hanno Böck.
8308 * PKCS7 crash with missing EnvelopedContent
8310 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8311 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8312 with missing content and trigger a NULL pointer dereference on parsing.
8314 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8315 structures from untrusted sources are affected. OpenSSL clients and
8316 servers are not affected.
8318 This issue was reported to OpenSSL by Michal Zalewski (Google).
8323 * CMS verify infinite loop with unknown hash function
8325 When verifying a signedData message the CMS code can enter an infinite loop
8326 if presented with an unknown hash function OID. This can be used to perform
8327 denial of service against any system which verifies signedData messages using
8329 This issue was reported to OpenSSL by Johannes Bauer.
8334 * Race condition handling NewSessionTicket
8336 If a NewSessionTicket is received by a multi-threaded client when attempting to
8337 reuse a previous ticket then a race condition can occur potentially leading to
8338 a double free of the ticket data.
8343 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8345 * Segmentation fault in ASN1_TYPE_cmp fix
8347 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8348 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8349 certificate signature algorithm consistency this can be used to crash any
8350 certificate verification operation and exploited in a DoS attack. Any
8351 application which performs certificate verification is vulnerable including
8352 OpenSSL clients and servers which enable client authentication.
8357 * ASN.1 structure reuse memory corruption fix
8359 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8360 memory corruption via an invalid write. Such reuse is and has been
8361 strongly discouraged and is believed to be rare.
8363 Applications that parse structures containing CHOICE or ANY DEFINED BY
8364 components may be affected. Certificate parsing (d2i_X509 and related
8365 functions) are however not affected. OpenSSL clients and servers are
8371 * PKCS7 NULL pointer dereferences fix
8373 The PKCS#7 parsing code does not handle missing outer ContentInfo
8374 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8375 missing content and trigger a NULL pointer dereference on parsing.
8377 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8378 otherwise parse PKCS#7 structures from untrusted sources are
8379 affected. OpenSSL clients and servers are not affected.
8381 This issue was reported to OpenSSL by Michal Zalewski (Google).
8386 * DoS via reachable assert in SSLv2 servers fix
8388 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8389 servers that both support SSLv2 and enable export cipher suites by sending
8390 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8392 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8393 (OpenSSL development team).
8398 * Use After Free following d2i_ECPrivatekey error fix
8400 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8401 could cause a use after free condition. This, in turn, could cause a double
8402 free in several private key parsing functions (such as d2i_PrivateKey
8403 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8404 for applications that receive EC private keys from untrusted
8405 sources. This scenario is considered rare.
8407 This issue was discovered by the BoringSSL project and fixed in their
8413 * X509_to_X509_REQ NULL pointer deref fix
8415 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8416 the certificate key is invalid. This function is rarely used in practice.
8418 This issue was discovered by Brian Carpenter.
8423 * Removed the export ciphers from the DEFAULT ciphers
8427 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8429 * Build fixes for the Windows and OpenVMS platforms
8431 *Matt Caswell and Richard Levitte*
8433 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8435 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8436 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8437 dereference. This could lead to a Denial Of Service attack. Thanks to
8438 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8443 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8444 dtls1_buffer_record function under certain conditions. In particular this
8445 could occur if an attacker sent repeated DTLS records with the same
8446 sequence number but for the next epoch. The memory leak could be exploited
8447 by an attacker in a Denial of Service attack through memory exhaustion.
8448 Thanks to Chris Mueller for reporting this issue.
8453 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8454 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8455 method would be set to NULL which could later result in a NULL pointer
8456 dereference. Thanks to Frank Schmirler for reporting this issue.
8461 * Abort handshake if server key exchange message is omitted for ephemeral
8464 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8465 reporting this issue.
8470 * Remove non-export ephemeral RSA code on client and server. This code
8471 violated the TLS standard by allowing the use of temporary RSA keys in
8472 non-export ciphersuites and could be used by a server to effectively
8473 downgrade the RSA key length used to a value smaller than the server
8474 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8475 INRIA or reporting this issue.
8480 * Fixed issue where DH client certificates are accepted without verification.
8481 An OpenSSL server will accept a DH certificate for client authentication
8482 without the certificate verify message. This effectively allows a client to
8483 authenticate without the use of a private key. This only affects servers
8484 which trust a client certificate authority which issues certificates
8485 containing DH keys: these are extremely rare and hardly ever encountered.
8486 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8492 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8493 results on some platforms, including x86_64. This bug occurs at random
8494 with a very low probability, and is not known to be exploitable in any
8495 way, though its exact impact is difficult to determine. Thanks to Pieter
8496 Wuille (Blockstream) who reported this issue and also suggested an initial
8497 fix. Further analysis was conducted by the OpenSSL development team and
8498 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8499 the OpenSSL core team.
8504 * Fix various certificate fingerprint issues.
8506 By using non-DER or invalid encodings outside the signed portion of a
8507 certificate the fingerprint can be changed without breaking the signature.
8508 Although no details of the signed portion of the certificate can be changed
8509 this can cause problems with some applications: e.g. those using the
8510 certificate fingerprint for blacklists.
8512 1. Reject signatures with non zero unused bits.
8514 If the BIT STRING containing the signature has non zero unused bits reject
8515 the signature. All current signature algorithms require zero unused bits.
8517 2. Check certificate algorithm consistency.
8519 Check the AlgorithmIdentifier inside TBS matches the one in the
8520 certificate signature. NB: this will result in signature failure
8521 errors for some broken certificates.
8523 Thanks to Konrad Kraszewski from Google for reporting this issue.
8525 3. Check DSA/ECDSA signatures use DER.
8527 Reencode DSA/ECDSA signatures and compare with the original received
8528 signature. Return an error if there is a mismatch.
8530 This will reject various cases including garbage after signature
8531 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8532 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8533 (negative or with leading zeroes).
8535 Further analysis was conducted and fixes were developed by Stephen Henson
8536 of the OpenSSL core team.
8542 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8544 * Session Ticket Memory Leak.
8546 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8547 integrity of that ticket is first verified. In the event of a session
8548 ticket integrity check failing, OpenSSL will fail to free memory
8549 causing a memory leak. By sending a large number of invalid session
8550 tickets an attacker could exploit this issue in a Denial Of Service
8556 * Build option no-ssl3 is incomplete.
8558 When OpenSSL is configured with "no-ssl3" as a build option, servers
8559 could accept and complete a SSL 3.0 handshake, and clients could be
8560 configured to send them.
8563 *Akamai and the OpenSSL team*
8565 * Add support for TLS_FALLBACK_SCSV.
8566 Client applications doing fallback retries should call
8567 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8570 *Adam Langley, Bodo Moeller*
8572 * Add additional DigestInfo checks.
8574 Reencode DigestInto in DER and check against the original when
8575 verifying RSA signature: this will reject any improperly encoded
8576 DigestInfo structures.
8578 Note: this is a precautionary measure and no attacks are currently known.
8582 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8584 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8585 to a denial of service attack. A malicious server can crash the client
8586 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8587 ciphersuite and sending carefully crafted handshake messages.
8589 Thanks to Felix Gröbert (Google) for discovering and researching this
8595 * By sending carefully crafted DTLS packets an attacker could cause openssl
8596 to leak memory. This can be exploited through a Denial of Service attack.
8597 Thanks to Adam Langley for discovering and researching this issue.
8602 * An attacker can force openssl to consume large amounts of memory whilst
8603 processing DTLS handshake messages. This can be exploited through a
8604 Denial of Service attack.
8605 Thanks to Adam Langley for discovering and researching this issue.
8610 * An attacker can force an error condition which causes openssl to crash
8611 whilst processing DTLS packets due to memory being freed twice. This
8612 can be exploited through a Denial of Service attack.
8613 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8619 * If a multithreaded client connects to a malicious server using a resumed
8620 session and the server sends an ec point format extension it could write
8621 up to 255 bytes to freed memory.
8623 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8629 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8630 X509_name_oneline, X509_name_print_ex et al. to leak some information
8631 from the stack. Applications may be affected if they echo pretty printing
8632 output to the attacker.
8634 Thanks to Ivan Fratric (Google) for discovering this issue.
8637 *Emilia Käsper, and Steve Henson*
8639 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8640 for corner cases. (Certain input points at infinity could lead to
8641 bogus results, with non-infinity inputs mapped to infinity too.)
8645 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8647 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8648 handshake can force the use of weak keying material in OpenSSL
8649 SSL/TLS clients and servers.
8651 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8652 researching this issue. ([CVE-2014-0224])
8654 *KIKUCHI Masashi, Steve Henson*
8656 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8657 OpenSSL DTLS client the code can be made to recurse eventually crashing
8660 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8663 *Imre Rad, Steve Henson*
8665 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8666 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8667 client or server. This is potentially exploitable to run arbitrary
8668 code on a vulnerable client or server.
8670 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8672 *Jüri Aedla, Steve Henson*
8674 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8675 are subject to a denial of service attack.
8677 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8678 this issue. ([CVE-2014-3470])
8680 *Felix Gröbert, Ivan Fratric, Steve Henson*
8682 * Harmonize version and its documentation. -f flag is used to display
8685 *mancha <mancha1@zoho.com>*
8687 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8688 in i2d_ECPrivateKey.
8690 *mancha <mancha1@zoho.com>*
8692 * Fix some double frees. These are not thought to be exploitable.
8694 *mancha <mancha1@zoho.com>*
8696 * Fix for the attack described in the paper "Recovering OpenSSL
8697 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8698 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8699 <http://eprint.iacr.org/2014/140>
8701 Thanks to Yuval Yarom and Naomi Benger for discovering this
8702 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8704 *Yuval Yarom and Naomi Benger*
8706 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8708 * Keep original DTLS digest and encryption contexts in retransmission
8709 structures so we can use the previous session parameters if they need
8710 to be resent. ([CVE-2013-6450])
8714 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8715 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8716 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8717 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8718 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8719 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8721 *Rob Stradling, Adam Langley*
8723 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8725 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8727 This addresses the flaw in CBC record processing discovered by
8728 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8729 at: <http://www.isg.rhul.ac.uk/tls/>
8731 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8732 Security Group at Royal Holloway, University of London
8733 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8734 Emilia Käsper for the initial patch.
8737 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8739 * Return an error when checking OCSP signatures when key is NULL.
8740 This fixes a DoS attack. ([CVE-2013-0166])
8744 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8745 the right response is stapled. Also change SSL_get_certificate()
8746 so it returns the certificate actually sent.
8747 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8748 (This is a backport)
8750 *Rob Stradling <rob.stradling@comodo.com>*
8752 * Fix possible deadlock when decoding public keys.
8756 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8758 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8761 * Sanity check record length before skipping explicit IV in DTLS
8764 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8765 fuzzing as a service testing platform.
8770 * Initialise tkeylen properly when encrypting CMS messages.
8771 Thanks to Solar Designer of Openwall for reporting this issue.
8775 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8777 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8778 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8779 in CRYPTO_realloc_clean.
8781 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8782 issue and to Adam Langley <agl@chromium.org> for fixing it.
8785 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8787 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8789 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8790 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8791 content decryption and always return the same error. Note: this attack
8792 needs on average 2^20 messages so it only affects automated senders. The
8793 old behaviour can be re-enabled in the CMS code by setting the
8794 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8795 an MMA defence is not necessary.
8796 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8797 this issue. ([CVE-2012-0884])
8801 * Fix CVE-2011-4619: make sure we really are receiving a
8802 client hello before rejecting multiple SGC restarts. Thanks to
8803 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8807 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8809 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8810 Thanks to Antonio Martin, Enterprise Secure Access Research and
8811 Development, Cisco Systems, Inc. for discovering this bug and
8812 preparing a fix. ([CVE-2012-0050])
8816 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8818 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8819 of the Vaudenay padding oracle attack on CBC mode encryption
8820 which enables an efficient plaintext recovery attack against
8821 the OpenSSL implementation of DTLS. Their attack exploits timing
8822 differences arising during decryption processing. A research
8823 paper describing this attack can be found at:
8824 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8825 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8826 Security Group at Royal Holloway, University of London
8827 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8828 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8829 for preparing the fix. ([CVE-2011-4108])
8831 *Robin Seggelmann, Michael Tuexen*
8833 * Clear bytes used for block padding of SSL 3.0 records.
8836 *Adam Langley (Google)*
8838 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8839 Kadianakis <desnacked@gmail.com> for discovering this issue and
8840 Adam Langley for preparing the fix. ([CVE-2011-4619])
8842 *Adam Langley (Google)*
8844 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8846 *Andrey Kulikov <amdeich@gmail.com>*
8848 * Prevent malformed RFC3779 data triggering an assertion failure.
8849 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8850 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8852 *Rob Austein <sra@hactrn.net>*
8854 * Improved PRNG seeding for VOS.
8856 *Paul Green <Paul.Green@stratus.com>*
8858 * Fix ssl_ciph.c set-up race.
8860 *Adam Langley (Google)*
8862 * Fix spurious failures in ecdsatest.c.
8864 *Emilia Käsper (Google)*
8866 * Fix the BIO_f_buffer() implementation (which was mixing different
8867 interpretations of the `..._len` fields).
8869 *Adam Langley (Google)*
8871 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8872 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8873 threads won't reuse the same blinding coefficients.
8875 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8876 lock to call BN_BLINDING_invert_ex, and avoids one use of
8877 BN_BLINDING_update for each BN_BLINDING structure (previously,
8878 the last update always remained unused).
8880 *Emilia Käsper (Google)*
8882 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8884 *Bob Buckholz (Google)*
8886 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8888 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8889 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8891 *Kaspar Brand <ossl@velox.ch>*
8893 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8894 for multi-threaded use of ECDH. ([CVE-2011-3210])
8896 *Adam Langley (Google)*
8898 * Fix x509_name_ex_d2i memory leak on bad inputs.
8902 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8903 signature public key algorithm by using OID xref utilities instead.
8904 Before this you could only use some ECC ciphersuites with SHA1 only.
8908 * Add protection against ECDSA timing attacks as mentioned in the paper
8909 by Billy Bob Brumley and Nicola Tuveri, see:
8910 <http://eprint.iacr.org/2011/232.pdf>
8912 *Billy Bob Brumley and Nicola Tuveri*
8914 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8916 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8918 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8920 * Fix bug in string printing code: if *any* escaping is enabled we must
8921 escape the escape character (backslash) or the resulting string is
8926 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8928 * Disable code workaround for ancient and obsolete Netscape browsers
8929 and servers: an attacker can use it in a ciphersuite downgrade attack.
8930 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8934 * Fixed J-PAKE implementation error, originally discovered by
8935 Sebastien Martini, further info and confirmation from Stefan
8936 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8940 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8942 * Fix extension code to avoid race conditions which can result in a buffer
8943 overrun vulnerability: resumed sessions must not be modified as they can
8944 be shared by multiple threads. CVE-2010-3864
8948 * Fix WIN32 build system to correctly link an ENGINE directory into
8953 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8955 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8958 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8960 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8962 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8963 context. The operation can be customised via the ctrl mechanism in
8964 case ENGINEs want to include additional functionality.
8968 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8972 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8973 output hashes compatible with older versions of OpenSSL.
8975 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8977 * Fix compression algorithm handling: if resuming a session use the
8978 compression algorithm of the resumed session instead of determining
8979 it from client hello again. Don't allow server to change algorithm.
8983 * Add load_crls() function to commands tidying load_certs() too. Add option
8984 to verify utility to allow additional CRLs to be included.
8988 * Update OCSP request code to permit adding custom headers to the request:
8989 some responders need this.
8993 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8996 *Julia Lawall <julia@diku.dk>*
8998 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8999 needlessly dereferenced structures, used obsolete functions and
9000 didn't handle all updated verify codes correctly.
9004 * Disable MD2 in the default configuration.
9008 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9009 indicate the initial BIO being pushed or popped. This makes it possible
9010 to determine whether the BIO is the one explicitly called or as a result
9011 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9012 it handles reference counts correctly and doesn't zero out the I/O bio
9013 when it is not being explicitly popped. WARNING: applications which
9014 included workarounds for the old buggy behaviour will need to be modified
9015 or they could free up already freed BIOs.
9019 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9020 renaming to all platforms (within the 0.9.8 branch, this was
9021 done conditionally on Netware platforms to avoid a name clash).
9023 *Guenter <lists@gknw.net>*
9025 * Add ECDHE and PSK support to DTLS.
9027 *Michael Tuexen <tuexen@fh-muenster.de>*
9029 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9034 * Add "missing" function EVP_MD_flags() (without this the only way to
9035 retrieve a digest flags is by accessing the structure directly. Update
9036 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9037 or cipher is registered as in the "from" argument. Print out all
9038 registered digests in the dgst usage message instead of manually
9039 attempting to work them out.
9043 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9044 this allows the use of compression and extensions. Change default cipher
9045 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9046 by default unless an application cipher string requests it.
9050 * Alter match criteria in PKCS12_parse(). It used to try to use local
9051 key ids to find matching certificates and keys but some PKCS#12 files
9052 don't follow the (somewhat unwritten) rules and this strategy fails.
9053 Now just gather all certificates together and the first private key
9054 then look for the first certificate that matches the key.
9058 * Support use of registered digest and cipher names for dgst and cipher
9059 commands instead of having to add each one as a special case. So now
9066 openssl dgst -sha256 foo
9068 and this works for ENGINE based algorithms too.
9072 * Update Gost ENGINE to support parameter files.
9074 *Victor B. Wagner <vitus@cryptocom.ru>*
9076 * Support GeneralizedTime in ca utility.
9078 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9080 * Enhance the hash format used for certificate directory links. The new
9081 form uses the canonical encoding (meaning equivalent names will work
9082 even if they aren't identical) and uses SHA1 instead of MD5. This form
9083 is incompatible with the older format and as a result c_rehash should
9084 be used to rebuild symbolic links.
9088 * Make PKCS#8 the default write format for private keys, replacing the
9089 traditional format. This form is standardised, more secure and doesn't
9090 include an implicit MD5 dependency.
9094 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9095 committed to OpenSSL should pass this lot as a minimum.
9099 * Add session ticket override functionality for use by EAP-FAST.
9101 *Jouni Malinen <j@w1.fi>*
9103 * Modify HMAC functions to return a value. Since these can be implemented
9104 in an ENGINE errors can occur.
9108 * Type-checked OBJ_bsearch_ex.
9112 * Type-checked OBJ_bsearch. Also some constification necessitated
9113 by type-checking. Still to come: TXT_DB, bsearch(?),
9114 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9119 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9120 seconds to a tm structure directly, instead of going through OS
9121 specific date routines. This avoids any issues with OS routines such
9122 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9123 and X509_time_adj_ex() to cover the extended range. The existing
9124 X509_time_adj() is still usable and will no longer have any date issues.
9128 * Delta CRL support. New use deltas option which will attempt to locate
9129 and search any appropriate delta CRLs available.
9131 This work was sponsored by Google.
9135 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9136 code and add additional score elements. Validate alternate CRL paths
9137 as part of the CRL checking and indicate a new error "CRL path validation
9138 error" in this case. Applications wanting additional details can use
9139 the verify callback and check the new "parent" field. If this is not
9140 NULL CRL path validation is taking place. Existing applications won't
9141 see this because it requires extended CRL support which is off by
9144 This work was sponsored by Google.
9148 * Support for freshest CRL extension.
9150 This work was sponsored by Google.
9154 * Initial indirect CRL support. Currently only supported in the CRLs
9155 passed directly and not via lookup. Process certificate issuer
9156 CRL entry extension and lookup CRL entries by bother issuer name
9157 and serial number. Check and process CRL issuer entry in IDP extension.
9159 This work was sponsored by Google.
9163 * Add support for distinct certificate and CRL paths. The CRL issuer
9164 certificate is validated separately in this case. Only enabled if
9165 an extended CRL support flag is set: this flag will enable additional
9166 CRL functionality in future.
9168 This work was sponsored by Google.
9172 * Add support for policy mappings extension.
9174 This work was sponsored by Google.
9178 * Fixes to pathlength constraint, self issued certificate handling,
9179 policy processing to align with RFC3280 and PKITS tests.
9181 This work was sponsored by Google.
9185 * Support for name constraints certificate extension. DN, email, DNS
9186 and URI types are currently supported.
9188 This work was sponsored by Google.
9192 * To cater for systems that provide a pointer-based thread ID rather
9193 than numeric, deprecate the current numeric thread ID mechanism and
9194 replace it with a structure and associated callback type. This
9195 mechanism allows a numeric "hash" to be extracted from a thread ID in
9196 either case, and on platforms where pointers are larger than 'long',
9197 mixing is done to help ensure the numeric 'hash' is usable even if it
9198 can't be guaranteed unique. The default mechanism is to use "&errno"
9199 as a pointer-based thread ID to distinguish between threads.
9201 Applications that want to provide their own thread IDs should now use
9202 CRYPTO_THREADID_set_callback() to register a callback that will call
9203 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9205 Note that ERR_remove_state() is now deprecated, because it is tied
9206 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9207 to free the current thread's error state should be replaced by
9208 ERR_remove_thread_state(NULL).
9210 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9211 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9212 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9213 application was previously providing a numeric thread callback that
9214 was inappropriate for distinguishing threads, then uniqueness might
9215 have been obtained with &errno that happened immediately in the
9216 intermediate development versions of OpenSSL; this is no longer the
9217 case, the numeric thread callback will now override the automatic use
9220 *Geoff Thorpe, with help from Bodo Moeller*
9222 * Initial support for different CRL issuing certificates. This covers a
9223 simple case where the self issued certificates in the chain exist and
9224 the real CRL issuer is higher in the existing chain.
9226 This work was sponsored by Google.
9230 * Removed effectively defunct crypto/store from the build.
9234 * Revamp of STACK to provide stronger type-checking. Still to come:
9235 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9236 ASN1_STRING, CONF_VALUE.
9240 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9241 RAM on SSL connections. This option can save about 34k per idle SSL.
9245 * Revamp of LHASH to provide stronger type-checking. Still to come:
9246 STACK, TXT_DB, bsearch, qsort.
9250 * Initial support for Cryptographic Message Syntax (aka CMS) based
9251 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9252 support for data, signedData, compressedData, digestedData and
9253 encryptedData, envelopedData types included. Scripts to check against
9254 RFC4134 examples draft and interop and consistency checks of many
9255 content types and variants.
9259 * Add options to enc utility to support use of zlib compression BIO.
9263 * Extend mk1mf to support importing of options and assembly language
9264 files from Configure script, currently only included in VC-WIN32.
9265 The assembly language rules can now optionally generate the source
9266 files from the associated perl scripts.
9270 * Implement remaining functionality needed to support GOST ciphersuites.
9271 Interop testing has been performed using CryptoPro implementations.
9273 *Victor B. Wagner <vitus@cryptocom.ru>*
9275 * s390x assembler pack.
9279 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9284 * Implement Opaque PRF Input TLS extension as specified in
9285 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9286 official specification yet and no extension type assignment by
9287 IANA exists, this extension (for now) will have to be explicitly
9288 enabled when building OpenSSL by providing the extension number
9289 to use. For example, specify an option
9291 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9293 to the "config" or "Configure" script to enable the extension,
9294 assuming extension number 0x9527 (which is a completely arbitrary
9295 and unofficial assignment based on the MD5 hash of the Internet
9296 Draft). Note that by doing so, you potentially lose
9297 interoperability with other TLS implementations since these might
9298 be using the same extension number for other purposes.
9300 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9301 opaque PRF input value to use in the handshake. This will create
9302 an internal copy of the length-'len' string at 'src', and will
9303 return non-zero for success.
9305 To get more control and flexibility, provide a callback function
9308 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9309 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9313 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9316 Callback function 'cb' will be called in handshakes, and is
9317 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9318 Argument 'arg' is for application purposes (the value as given to
9319 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9320 be provided to the callback function). The callback function
9321 has to return non-zero to report success: usually 1 to use opaque
9322 PRF input just if possible, or 2 to enforce use of the opaque PRF
9323 input. In the latter case, the library will abort the handshake
9324 if opaque PRF input is not successfully negotiated.
9326 Arguments 'peerinput' and 'len' given to the callback function
9327 will always be NULL and 0 in the case of a client. A server will
9328 see the client's opaque PRF input through these variables if
9329 available (NULL and 0 otherwise). Note that if the server
9330 provides an opaque PRF input, the length must be the same as the
9331 length of the client's opaque PRF input.
9333 Note that the callback function will only be called when creating
9334 a new session (session resumption can resume whatever was
9335 previously negotiated), and will not be called in SSL 2.0
9336 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9337 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9338 for applications that need to enforce opaque PRF input.
9342 * Update ssl code to support digests other than SHA1+MD5 for handshake
9345 *Victor B. Wagner <vitus@cryptocom.ru>*
9347 * Add RFC4507 support to OpenSSL. This includes the corrections in
9348 RFC4507bis. The encrypted ticket format is an encrypted encoded
9349 SSL_SESSION structure, that way new session features are automatically
9352 If a client application caches session in an SSL_SESSION structure
9353 support is transparent because tickets are now stored in the encoded
9356 The SSL_CTX structure automatically generates keys for ticket
9357 protection in servers so again support should be possible
9358 with no application modification.
9360 If a client or server wishes to disable RFC4507 support then the option
9361 SSL_OP_NO_TICKET can be set.
9363 Add a TLS extension debugging callback to allow the contents of any client
9364 or server extensions to be examined.
9366 This work was sponsored by Google.
9370 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9371 OpenSSL should now compile cleanly on gcc 4.2
9373 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9375 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9376 support including streaming MAC support: this is required for GOST
9377 ciphersuite support.
9379 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9381 * Add option -stream to use PKCS#7 streaming in smime utility. New
9382 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9383 to output in BER and PEM format.
9387 * Experimental support for use of HMAC via EVP_PKEY interface. This
9388 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9389 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9390 ENGINE support for HMAC keys which are unextractable. New -mac and
9391 -macopt options to dgst utility.
9395 * New option -sigopt to dgst utility. Update dgst to use
9396 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9397 alternative signing parameters such as X9.31 or PSS in the dgst
9402 * Change ssl_cipher_apply_rule(), the internal function that does
9403 the work each time a ciphersuite string requests enabling
9404 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9405 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9406 the order of disabled ciphersuites such that those ciphersuites
9407 that most recently went from enabled to disabled not only stay
9408 in order with respect to each other, but also have higher priority
9409 than other disabled ciphersuites the next time ciphersuites are
9412 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9413 the same ciphersuites as with "HIGH" alone, but in a specific
9414 order where the PSK ciphersuites come first (since they are the
9415 most recently disabled ciphersuites when "HIGH" is parsed).
9417 Also, change ssl_create_cipher_list() (using this new
9418 functionality) such that between otherwise identical
9419 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9424 * Change ssl_create_cipher_list() so that it automatically
9425 arranges the ciphersuites in reasonable order before starting
9426 to process the rule string. Thus, the definition for "DEFAULT"
9427 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9428 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9429 This makes it much easier to arrive at a reasonable default order
9430 in applications for which anonymous ciphers are OK (meaning
9431 that you can't actually use DEFAULT).
9433 *Bodo Moeller; suggested by Victor Duchovni*
9435 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9436 processing) into multiple integers instead of setting
9437 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9438 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9439 (These masks as well as the individual bit definitions are hidden
9440 away into the non-exported interface ssl/ssl_locl.h, so this
9441 change to the definition of the SSL_CIPHER structure shouldn't
9442 affect applications.) This give us more bits for each of these
9443 categories, so there is no longer a need to coagulate AES128 and
9444 AES256 into a single algorithm bit, and to coagulate Camellia128
9445 and Camellia256 into a single algorithm bit, which has led to all
9448 Thus, among other things, the kludge introduced in 0.9.7m and
9449 0.9.8e for masking out AES256 independently of AES128 or masking
9450 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9452 With the change, we also introduce new ciphersuite aliases that
9453 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9458 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9459 Use the leftmost N bytes of the signature input if the input is
9460 larger than the prime q (with N being the size in bytes of q).
9464 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9465 it yet and it is largely untested.
9469 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9473 * Initial incomplete changes to avoid need for function casts in OpenSSL
9474 some compilers (gcc 4.2 and later) reject their use. Safestack is
9475 reimplemented. Update ASN1 to avoid use of legacy functions.
9479 * Win32/64 targets are linked with Winsock2.
9483 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9484 to external functions. This can be used to increase CRL handling
9485 efficiency especially when CRLs are very large by (for example) storing
9486 the CRL revoked certificates in a database.
9490 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9491 new CRLs added to a directory can be used. New command line option
9492 -verify_return_error to s_client and s_server. This causes real errors
9493 to be returned by the verify callback instead of carrying on no matter
9494 what. This reflects the way a "real world" verify callback would behave.
9498 * GOST engine, supporting several GOST algorithms and public key formats.
9499 Kindly donated by Cryptocom.
9503 * Partial support for Issuing Distribution Point CRL extension. CRLs
9504 partitioned by DP are handled but no indirect CRL or reason partitioning
9505 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9506 selected via a scoring technique which handles IDP and AKID in CRLs.
9510 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9511 will ultimately be used for all verify operations: this will remove the
9512 X509_STORE dependency on certificate verification and allow alternative
9513 lookup methods. X509_STORE based implementations of these two callbacks.
9517 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9518 Modify get_crl() to find a valid (unexpired) CRL if possible.
9522 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9523 this would be called X509_CRL_cmp() but that name is already used by
9524 a function that just compares CRL issuer names. Cache several CRL
9525 extensions in X509_CRL structure and cache CRLDP in X509.
9529 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9530 this maps equivalent X509_NAME structures into a consistent structure.
9531 Name comparison can then be performed rapidly using memcmp().
9535 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9540 * Allow digests to supply their own micalg string for S/MIME type using
9541 the ctrl EVP_MD_CTRL_MICALG.
9545 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9546 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9547 ctrl. It can then customise the structure before and/or after signing
9552 * New function OBJ_add_sigid() to allow application defined signature OIDs
9553 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9554 to free up any added signature OIDs.
9558 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9559 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9560 digest and cipher tables. New options added to openssl utility:
9561 list-message-digest-algorithms and list-cipher-algorithms.
9565 * Change the array representation of binary polynomials: the list
9566 of degrees of non-zero coefficients is now terminated with -1.
9567 Previously it was terminated with 0, which was also part of the
9568 value; thus, the array representation was not applicable to
9569 polynomials where t^0 has coefficient zero. This change makes
9570 the array representation useful in a more general context.
9574 * Various modifications and fixes to SSL/TLS cipher string
9575 handling. For ECC, the code now distinguishes between fixed ECDH
9576 with RSA certificates on the one hand and with ECDSA certificates
9577 on the other hand, since these are separate ciphersuites. The
9578 unused code for Fortezza ciphersuites has been removed.
9580 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9581 (not "ECDHE"). For consistency with the code for DH
9582 certificates, use of ECDH certificates is now considered ECDH
9583 authentication, not RSA or ECDSA authentication (the latter is
9584 merely the CA's signing algorithm and not actively used in the
9587 The temporary ciphersuite alias "ECCdraft" is no longer
9588 available, and ECC ciphersuites are no longer excluded from "ALL"
9589 and "DEFAULT". The following aliases now exist for RFC 4492
9590 ciphersuites, most of these by analogy with the DH case:
9592 kECDHr - ECDH cert, signed with RSA
9593 kECDHe - ECDH cert, signed with ECDSA
9594 kECDH - ECDH cert (signed with either RSA or ECDSA)
9595 kEECDH - ephemeral ECDH
9596 ECDH - ECDH cert or ephemeral ECDH
9602 AECDH - anonymous ECDH
9603 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9607 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9608 Use correct micalg parameters depending on digest(s) in signed message.
9612 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9613 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9617 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9618 an engine to register a method. Add ENGINE lookups for methods and
9619 functional reference processing.
9623 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9624 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9629 * New -resign option to smime utility. This adds one or more signers
9630 to an existing PKCS#7 signedData structure. Also -md option to use an
9631 alternative message digest algorithm for signing.
9635 * Tidy up PKCS#7 routines and add new functions to make it easier to
9636 create PKCS7 structures containing multiple signers. Update smime
9637 application to support multiple signers.
9641 * New -macalg option to pkcs12 utility to allow setting of an alternative
9646 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9647 Reorganize PBE internals to lookup from a static table using NIDs,
9648 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9649 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9650 PRF which will be automatically used with PBES2.
9654 * Replace the algorithm specific calls to generate keys in "req" with the
9659 * Update PKCS#7 enveloped data routines to use new API. This is now
9660 supported by any public key method supporting the encrypt operation. A
9661 ctrl is added to allow the public key algorithm to examine or modify
9662 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9667 * Add a ctrl to asn1 method to allow a public key algorithm to express
9668 a default digest type to use. In most cases this will be SHA1 but some
9669 algorithms (such as GOST) need to specify an alternative digest. The
9670 return value indicates how strong the preference is 1 means optional and
9671 2 is mandatory (that is it is the only supported type). Modify
9672 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9673 use the default md. Update openssl utilities to use the default digest
9674 type for signing if it is not explicitly indicated.
9678 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9679 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9680 signing method from the key type. This effectively removes the link
9681 between digests and public key types.
9685 * Add an OID cross reference table and utility functions. Its purpose is to
9686 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9687 rsaEncryption. This will allow some of the algorithm specific hackery
9688 needed to use the correct OID to be removed.
9692 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9693 structures for PKCS7_sign(). They are now set up by the relevant public
9698 * Add provisional EC pkey method with support for ECDSA and ECDH.
9702 * Add support for key derivation (agreement) in the API, DH method and
9707 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9708 public and private key formats. As a side effect these add additional
9709 command line functionality not previously available: DSA signatures can be
9710 generated and verified using pkeyutl and DH key support and generation in
9717 *Oliver Tappe <zooey@hirschkaefer.de>*
9719 * New make target "install_html_docs" installs HTML renditions of the
9722 *Oliver Tappe <zooey@hirschkaefer.de>*
9724 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9725 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9726 support key and parameter generation and add initial key generation
9727 functionality for RSA.
9731 * Add functions for main EVP_PKEY_method operations. The undocumented
9732 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9733 `EVP_PKEY_{encrypt,decrypt}_old`.
9737 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9738 key API, doesn't do much yet.
9742 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9743 public key algorithms. New option to openssl utility:
9744 "list-public-key-algorithms" to print out info.
9748 * Implement the Supported Elliptic Curves Extension for
9749 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9753 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9754 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9758 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9759 utilities such as rsa, dsa, dsaparam etc except they process any key
9764 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9765 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9766 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9771 * Initial support for pluggable public key ASN1.
9772 De-spaghettify the public key ASN1 handling. Move public and private
9773 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9774 algorithm specific handling to a single module within the relevant
9775 algorithm directory. Add functions to allow (near) opaque processing
9776 of public and private key structures.
9780 * Implement the Supported Point Formats Extension for
9781 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9785 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9786 for the psk identity [hint] and the psk callback functions to the
9787 SSL_SESSION, SSL and SSL_CTX structure.
9790 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9794 SSL_CTX_use_psk_identity_hint
9795 SSL_get_psk_identity_hint
9796 SSL_get_psk_identity
9797 SSL_use_psk_identity_hint
9799 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9801 * Add RFC 3161 compliant time stamp request creation, response generation
9802 and response verification functionality.
9804 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9806 * Add initial support for TLS extensions, specifically for the server_name
9807 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9808 have new members for a host name. The SSL data structure has an
9809 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9810 stored in that context to allow for session resumption, even after the
9811 SSL has been switched to a new SSL_CTX in reaction to a client's
9812 server_name extension.
9814 New functions (subject to change):
9816 SSL_get_servername()
9817 SSL_get_servername_type()
9820 New CTRL codes and macros (subject to change):
9822 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9823 - SSL_CTX_set_tlsext_servername_callback()
9824 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9825 - SSL_CTX_set_tlsext_servername_arg()
9826 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9828 openssl s_client has a new '-servername ...' option.
9830 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9831 '-key2 ...', '-servername_fatal' (subject to change). This allows
9832 testing the HostName extension for a specific single host name ('-cert'
9833 and '-key' remain fallbacks for handshakes without HostName
9834 negotiation). If the unrecognized_name alert has to be sent, this by
9835 default is a warning; it becomes fatal with the '-servername_fatal'
9838 *Peter Sylvester, Remy Allais, Christophe Renou*
9840 * Whirlpool hash implementation is added.
9844 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9845 bn(64,32). Because of instruction set limitations it doesn't have
9846 any negative impact on performance. This was done mostly in order
9847 to make it possible to share assembler modules, such as bn_mul_mont
9848 implementations, between 32- and 64-bit builds without hassle.
9852 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9853 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9858 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9859 dedicated Montgomery multiplication procedure, is introduced.
9860 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9861 "64-bit" performance on certain 32-bit targets.
9865 * New option SSL_OP_NO_COMP to disable use of compression selectively
9866 in SSL structures. New SSL ctrl to set maximum send fragment size.
9867 Save memory by setting the I/O buffer sizes dynamically instead of
9868 using the maximum available value.
9872 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9873 in addition to the text details.
9877 * Very, very preliminary EXPERIMENTAL support for printing of general
9878 ASN1 structures. This currently produces rather ugly output and doesn't
9879 handle several customised structures at all.
9883 * Integrated support for PVK file format and some related formats such
9884 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9885 these in the 'rsa' and 'dsa' utilities.
9889 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9893 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9894 place for the (very old) "NETSCAPE" format certificates which are now
9895 handled using new ASN1 code equivalents.
9899 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9900 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9901 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9905 * Modify CRL distribution points extension code to print out previously
9906 unsupported fields. Enhance extension setting code to allow setting of
9911 * Add print and set support for Issuing Distribution Point CRL extension.
9915 * Change 'Configure' script to enable Camellia by default.
9922 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9924 * When rejecting SSL/TLS records due to an incorrect version number, never
9925 update s->server with a new major version number. As of
9926 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9927 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9928 the previous behavior could result in a read attempt at NULL when
9929 receiving specific incorrect SSL/TLS records once record payload
9930 protection is active. ([CVE-2010-0740])
9932 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9934 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9935 could be crashed if the relevant tables were not present (e.g. chrooted).
9937 *Tomas Hoger <thoger@redhat.com>*
9939 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9941 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9943 *Martin Olsson, Neel Mehta*
9945 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9946 accommodate for stack sorting, always a write lock!).
9950 * On some versions of WIN32 Heap32Next is very slow. This can cause
9951 excessive delays in the RAND_poll(): over a minute. As a workaround
9952 include a time check in the inner Heap32Next loop too.
9956 * The code that handled flushing of data in SSL/TLS originally used the
9957 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9958 the problem outlined in PR#1949. The fix suggested there however can
9959 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9960 of Apache). So instead simplify the code to flush unconditionally.
9961 This should be fine since flushing with no data to flush is a no op.
9965 * Handle TLS versions 2.0 and later properly and correctly use the
9966 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9967 off ancient servers have a habit of sticking around for a while...
9971 * Modify compression code so it frees up structures without using the
9972 ex_data callbacks. This works around a problem where some applications
9973 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9974 restarting) then use compression (e.g. SSL with compression) later.
9975 This results in significant per-connection memory leaks and
9976 has caused some security issues including CVE-2008-1678 and
9981 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9982 change when encrypting or decrypting.
9986 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9987 connect and renegotiate with servers which do not support RI.
9988 Until RI is more widely deployed this option is enabled by default.
9992 * Add "missing" ssl ctrls to clear options and mode.
9996 * If client attempts to renegotiate and doesn't support RI respond with
9997 a no_renegotiation alert as required by RFC5746. Some renegotiating
9998 TLS clients will continue a connection gracefully when they receive
9999 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10000 waiting for a server hello which it will never receive. Now we treat a
10001 received no_renegotiation alert as a fatal error. This is because
10002 applications requesting a renegotiation might well expect it to succeed
10003 and would have no code in place to handle the server denying it so the
10004 only safe thing to do is to terminate the connection.
10008 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10009 peer supports secure renegotiation and 0 otherwise. Print out peer
10010 renegotiation support in s_client/s_server.
10014 * Replace the highly broken and deprecated SPKAC certification method with
10015 the updated NID creation version. This should correctly handle UTF8.
10019 * Implement RFC5746. Re-enable renegotiation but require the extension
10020 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10021 turns out to be a bad idea. It has been replaced by
10022 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10023 SSL_CTX_set_options(). This is really not recommended unless you
10024 know what you are doing.
10026 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10028 * Fixes to stateless session resumption handling. Use initial_ctx when
10029 issuing and attempting to decrypt tickets in case it has changed during
10030 servername handling. Use a non-zero length session ID when attempting
10031 stateless session resumption: this makes it possible to determine if
10032 a resumption has occurred immediately after receiving server hello
10033 (several places in OpenSSL subtly assume this) instead of later in
10038 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10039 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10040 fixes for a few places where the return code is not checked
10043 *Julia Lawall <julia@diku.dk>*
10045 * Add --strict-warnings option to Configure script to include devteam
10046 warnings in other configurations.
10050 * Add support for --libdir option and LIBDIR variable in makefiles. This
10051 makes it possible to install openssl libraries in locations which
10052 have names other than "lib", for example "/usr/lib64" which some
10055 *Steve Henson, based on patch from Jeremy Utley*
10057 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10058 X690 8.9.12 and can produce some misleading textual output of OIDs.
10060 *Steve Henson, reported by Dan Kaminsky*
10062 * Delete MD2 from algorithm tables. This follows the recommendation in
10063 several standards that it is not used in new applications due to
10064 several cryptographic weaknesses. For binary compatibility reasons
10065 the MD2 API is still compiled in by default.
10069 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10074 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10075 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10078 *Guenter <lists@gknw.net>*
10080 * Fix the server certificate chain building code to use X509_verify_cert(),
10081 it used to have an ad-hoc builder which was unable to cope with anything
10082 other than a simple chain.
10084 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10086 * Don't check self signed certificate signatures in X509_verify_cert()
10087 by default (a flag can override this): it just wastes time without
10088 adding any security. As a useful side effect self signed root CAs
10089 with non-FIPS digests are now usable in FIPS mode.
10093 * In dtls1_process_out_of_seq_message() the check if the current message
10094 is already buffered was missing. For every new message was memory
10095 allocated, allowing an attacker to perform an denial of service attack
10096 with sending out of seq handshake messages until there is no memory
10097 left. Additionally every future message was buffered, even if the
10098 sequence number made no sense and would be part of another handshake.
10099 So only messages with sequence numbers less than 10 in advance will be
10100 buffered. ([CVE-2009-1378])
10102 *Robin Seggelmann, discovered by Daniel Mentz*
10104 * Records are buffered if they arrive with a future epoch to be
10105 processed after finishing the corresponding handshake. There is
10106 currently no limitation to this buffer allowing an attacker to perform
10107 a DOS attack with sending records with future epochs until there is no
10108 memory left. This patch adds the pqueue_size() function to determine
10109 the size of a buffer and limits the record buffer to 100 entries.
10112 *Robin Seggelmann, discovered by Daniel Mentz*
10114 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10115 parent structure is freed. ([CVE-2009-1379])
10119 * Handle non-blocking I/O properly in SSL_shutdown() call.
10121 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10123 * Add `2.5.4.*` OIDs
10125 *Ilya O. <vrghost@gmail.com>*
10127 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10129 * Disable renegotiation completely - this fixes a severe security
10130 problem ([CVE-2009-3555]) at the cost of breaking all
10131 renegotiation. Renegotiation can be re-enabled by setting
10132 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10133 run-time. This is really not recommended unless you know what
10138 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10140 * Don't set val to NULL when freeing up structures, it is freed up by
10141 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10142 zeroing past the valid field. ([CVE-2009-0789])
10144 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10146 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10147 checked correctly. This would allow some invalid signed attributes to
10148 appear to verify correctly. ([CVE-2009-0591])
10150 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10152 * Reject UniversalString and BMPString types with invalid lengths. This
10153 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10154 a legal length. ([CVE-2009-0590])
10158 * Set S/MIME signing as the default purpose rather than setting it
10159 unconditionally. This allows applications to override it at the store
10164 * Permit restricted recursion of ASN1 strings. This is needed in practice
10165 to handle some structures.
10169 * Improve efficiency of mem_gets: don't search whole buffer each time
10172 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10174 * New -hex option for openssl rand.
10178 * Print out UTF8String and NumericString when parsing ASN1.
10182 * Support NumericString type for name components.
10186 * Allow CC in the environment to override the automatically chosen
10187 compiler. Note that nothing is done to ensure flags work with the
10192 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10194 * Properly check EVP_VerifyFinal() and similar return values
10197 *Ben Laurie, Bodo Moeller, Google Security Team*
10199 * Enable TLS extensions by default.
10203 * Allow the CHIL engine to be loaded, whether the application is
10204 multithreaded or not. (This does not release the developer from the
10205 obligation to set up the dynamic locking callbacks.)
10207 *Sander Temme <sander@temme.net>*
10209 * Use correct exit code if there is an error in dgst command.
10211 *Steve Henson; problem pointed out by Roland Dirlewanger*
10213 * Tweak Configure so that you need to say "experimental-jpake" to enable
10214 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10218 * Add experimental JPAKE support, including demo authentication in
10219 s_client and s_server.
10223 * Set the comparison function in v3_addr_canonize().
10225 *Rob Austein <sra@hactrn.net>*
10227 * Add support for XMPP STARTTLS in s_client.
10229 *Philip Paeps <philip@freebsd.org>*
10231 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10232 to ensure that even with this option, only ciphersuites in the
10233 server's preference list will be accepted. (Note that the option
10234 applies only when resuming a session, so the earlier behavior was
10235 just about the algorithm choice for symmetric cryptography.)
10239 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10241 * Fix NULL pointer dereference if a DTLS server received
10242 ChangeCipherSpec as first record ([CVE-2009-1386]).
10246 * Fix a state transition in s3_srvr.c and d1_srvr.c
10247 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10249 *Nagendra Modadugu*
10251 * The fix in 0.9.8c that supposedly got rid of unsafe
10252 double-checked locking was incomplete for RSA blinding,
10253 addressing just one layer of what turns out to have been
10254 doubly unsafe triple-checked locking.
10256 So now fix this for real by retiring the MONT_HELPER macro
10257 in crypto/rsa/rsa_eay.c.
10259 *Bodo Moeller; problem pointed out by Marius Schilder*
10261 * Various precautionary measures:
10263 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10265 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10266 (NB: This would require knowledge of the secret session ticket key
10267 to exploit, in which case you'd be SOL either way.)
10269 - Change bn_nist.c so that it will properly handle input BIGNUMs
10270 outside the expected range.
10272 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10275 *Neel Mehta, Bodo Moeller*
10277 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10278 the load fails. Useful for distros.
10280 *Ben Laurie and the FreeBSD team*
10282 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10286 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10290 * Expand ENGINE to support engine supplied SSL client certificate functions.
10292 This work was sponsored by Logica.
10296 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10297 keystores. Support for SSL/TLS client authentication too.
10298 Not compiled unless enable-capieng specified to Configure.
10300 This work was sponsored by Logica.
10304 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10305 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10306 attribute creation routines such as certificate requests and PKCS#12
10311 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10313 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10314 handshake which could lead to a client crash as found using the
10315 Codenomicon TLS test suite ([CVE-2008-1672])
10317 *Steve Henson, Mark Cox*
10319 * Fix double free in TLS server name extensions which could lead to
10320 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10324 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10326 Clear the error queue to ensure that error entries left from
10327 older function calls do not interfere with the correct operation.
10329 *Lutz Jaenicke, Erik de Castro Lopo*
10331 * Remove root CA certificates of commercial CAs:
10333 The OpenSSL project does not recommend any specific CA and does not
10334 have any policy with respect to including or excluding any CA.
10335 Therefore it does not make any sense to ship an arbitrary selection
10336 of root CA certificates with the OpenSSL software.
10340 * RSA OAEP patches to fix two separate invalid memory reads.
10341 The first one involves inputs when 'lzero' is greater than
10342 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10343 before the beginning of from). The second one involves inputs where
10344 the 'db' section contains nothing but zeroes (there is a one-byte
10345 invalid read after the end of 'db').
10347 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10349 * Partial backport from 0.9.9-dev:
10351 Introduce bn_mul_mont (dedicated Montgomery multiplication
10352 procedure) as a candidate for BIGNUM assembler implementation.
10353 While 0.9.9-dev uses assembler for various architectures, only
10354 x86_64 is available by default here in the 0.9.8 branch, and
10355 32-bit x86 is available through a compile-time setting.
10357 To try the 32-bit x86 assembler implementation, use Configure
10358 option "enable-montasm" (which exists only for this backport).
10360 As "enable-montasm" for 32-bit x86 disclaims code stability
10361 anyway, in this constellation we activate additional code
10362 backported from 0.9.9-dev for further performance improvements,
10363 namely BN_from_montgomery_word. (To enable this otherwise,
10364 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10366 *Andy Polyakov (backport partially by Bodo Moeller)*
10368 * Add TLS session ticket callback. This allows an application to set
10369 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10370 values. This is useful for key rollover for example where several key
10371 sets may exist with different names.
10375 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10376 This was broken until now in 0.9.8 releases, such that the only way
10377 a registered ENGINE could be used (assuming it initialises
10378 successfully on the host) was to explicitly set it as the default
10379 for the relevant algorithms. This is in contradiction with 0.9.7
10380 behaviour and the documentation. With this fix, when an ENGINE is
10381 registered into a given algorithm's table of implementations, the
10382 'uptodate' flag is reset so that auto-discovery will be used next
10383 time a new context for that algorithm attempts to select an
10386 *Ian Lister (tweaked by Geoff Thorpe)*
10388 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10389 implementation in the following ways:
10391 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10394 Lack of BER streaming support means one pass streaming processing is
10395 only supported if data is detached: setting the streaming flag is
10396 ignored for embedded content.
10398 CMS support is disabled by default and must be explicitly enabled
10399 with the enable-cms configuration option.
10403 * Update the GMP engine glue to do direct copies between BIGNUM and
10404 mpz_t when openssl and GMP use the same limb size. Otherwise the
10405 existing "conversion via a text string export" trick is still used.
10407 *Paul Sheer <paulsheer@gmail.com>*
10409 * Zlib compression BIO. This is a filter BIO which compressed and
10410 uncompresses any data passed through it.
10414 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10415 RFC3394 compatible AES key wrapping.
10419 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10420 sets string data without copying. X509_ALGOR_set0() and
10421 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10422 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10423 from an X509_ATTRIBUTE structure optionally checking it occurs only
10424 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10429 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10430 to get the expected BN_FLG_CONSTTIME behavior.
10432 *Bodo Moeller (Google)*
10436 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10437 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10438 - added some more tests to do_tests.pl
10439 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10440 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10441 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10442 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10443 - various changes to netware.pl to enable gcc-cross builds on Win32
10445 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10446 - various changes to fix missing prototype warnings
10447 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10448 - added AES, WHIRLPOOL and CPUID assembler code to build files
10449 - added missing AES assembler make rules to mk1mf.pl
10450 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10452 *Guenter Knauf <eflash@gmx.net>*
10454 * Implement certificate status request TLS extension defined in RFC3546.
10455 A client can set the appropriate parameters and receive the encoded
10456 OCSP response via a callback. A server can query the supplied parameters
10457 and set the encoded OCSP response in the callback. Add simplified examples
10458 to s_client and s_server.
10462 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10464 * Fix various bugs:
10465 + Binary incompatibility of ssl_ctx_st structure
10466 + DTLS interoperation with non-compliant servers
10467 + Don't call get_session_cb() without proposed session
10468 + Fix ia64 assembler code
10470 *Andy Polyakov, Steve Henson*
10472 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10474 * DTLS Handshake overhaul. There were longstanding issues with
10475 OpenSSL DTLS implementation, which were making it impossible for
10476 RFC 4347 compliant client to communicate with OpenSSL server.
10477 Unfortunately just fixing these incompatibilities would "cut off"
10478 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10479 server keeps tolerating non RFC compliant syntax. The opposite is
10480 not true, 0.9.8f client can not communicate with earlier server.
10481 This update even addresses CVE-2007-4995.
10485 * Changes to avoid need for function casts in OpenSSL: some compilers
10486 (gcc 4.2 and later) reject their use.
10487 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10490 * Add RFC4507 support to OpenSSL. This includes the corrections in
10491 RFC4507bis. The encrypted ticket format is an encrypted encoded
10492 SSL_SESSION structure, that way new session features are automatically
10495 If a client application caches session in an SSL_SESSION structure
10496 support is transparent because tickets are now stored in the encoded
10499 The SSL_CTX structure automatically generates keys for ticket
10500 protection in servers so again support should be possible
10501 with no application modification.
10503 If a client or server wishes to disable RFC4507 support then the option
10504 SSL_OP_NO_TICKET can be set.
10506 Add a TLS extension debugging callback to allow the contents of any client
10507 or server extensions to be examined.
10509 This work was sponsored by Google.
10513 * Add initial support for TLS extensions, specifically for the server_name
10514 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10515 have new members for a host name. The SSL data structure has an
10516 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10517 stored in that context to allow for session resumption, even after the
10518 SSL has been switched to a new SSL_CTX in reaction to a client's
10519 server_name extension.
10521 New functions (subject to change):
10523 SSL_get_servername()
10524 SSL_get_servername_type()
10527 New CTRL codes and macros (subject to change):
10529 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10530 - SSL_CTX_set_tlsext_servername_callback()
10531 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10532 - SSL_CTX_set_tlsext_servername_arg()
10533 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10535 openssl s_client has a new '-servername ...' option.
10537 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10538 '-key2 ...', '-servername_fatal' (subject to change). This allows
10539 testing the HostName extension for a specific single host name ('-cert'
10540 and '-key' remain fallbacks for handshakes without HostName
10541 negotiation). If the unrecognized_name alert has to be sent, this by
10542 default is a warning; it becomes fatal with the '-servername_fatal'
10545 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10547 * Add AES and SSE2 assembly language support to VC++ build.
10551 * Mitigate attack on final subtraction in Montgomery reduction.
10555 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10556 (which previously caused an internal error).
10560 * Squeeze another 10% out of IGE mode when in != out.
10564 * AES IGE mode speedup.
10566 *Dean Gaudet (Google)*
10568 * Add the Korean symmetric 128-bit cipher SEED (see
10569 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10570 add SEED ciphersuites from RFC 4162:
10572 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10573 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10574 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10575 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10577 To minimize changes between patchlevels in the OpenSSL 0.9.8
10578 series, SEED remains excluded from compilation unless OpenSSL
10579 is configured with 'enable-seed'.
10581 *KISA, Bodo Moeller*
10583 * Mitigate branch prediction attacks, which can be practical if a
10584 single processor is shared, allowing a spy process to extract
10585 information. For detailed background information, see
10586 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10587 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10588 and Necessary Software Countermeasures"). The core of the change
10589 are new versions BN_div_no_branch() and
10590 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10591 respectively, which are slower, but avoid the security-relevant
10592 conditional branches. These are automatically called by BN_div()
10593 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10594 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10595 remove a conditional branch.
10597 BN_FLG_CONSTTIME is the new name for the previous
10598 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10599 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10600 in the exponent causes BN_mod_exp_mont() to use the alternative
10601 implementation in BN_mod_exp_mont_consttime().) The old name
10602 remains as a deprecated alias.
10604 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10605 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10606 constant-time implementations for more than just exponentiation.
10607 Here too the old name is kept as a deprecated alias.
10609 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10610 the BN_BLINDING structure gets an independent copy of the
10611 modulus. This means that the previous `BIGNUM *m` argument to
10612 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10613 essentially becomes `const BIGNUM *m`, although we can't actually
10614 change this in the header file before 0.9.9. It allows
10615 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10616 enable BN_FLG_CONSTTIME.
10618 *Matthew D Wood (Intel Corp)*
10620 * In the SSL/TLS server implementation, be strict about session ID
10621 context matching (which matters if an application uses a single
10622 external cache for different purposes). Previously,
10623 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10624 set. This did ensure strict client verification, but meant that,
10625 with applications using a single external cache for quite
10626 different requirements, clients could circumvent ciphersuite
10627 restrictions for a given session ID context by starting a session
10628 in a different context.
10632 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10633 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10634 authentication-only ciphersuites.
10638 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10639 not complete and could lead to a possible single byte overflow
10640 ([CVE-2007-5135]) [Ben Laurie]
10642 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10644 * Since AES128 and AES256 (and similarly Camellia128 and
10645 Camellia256) share a single mask bit in the logic of
10646 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10647 kludge to work properly if AES128 is available and AES256 isn't
10648 (or if Camellia128 is available and Camellia256 isn't).
10652 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10653 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10654 When a point or a seed is encoded in a BIT STRING, we need to
10655 prevent the removal of trailing zero bits to get the proper DER
10656 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10657 of a NamedBitList, for which trailing 0 bits need to be removed.)
10661 * Have SSL/TLS server implementation tolerate "mismatched" record
10662 protocol version while receiving ClientHello even if the
10663 ClientHello is fragmented. (The server can't insist on the
10664 particular protocol version it has chosen before the ServerHello
10665 message has informed the client about his choice.)
10669 * Add RFC 3779 support.
10671 *Rob Austein for ARIN, Ben Laurie*
10673 * Load error codes if they are not already present instead of using a
10674 static variable. This allows them to be cleanly unloaded and reloaded.
10675 Improve header file function name parsing.
10679 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10680 or CAPABILITY handshake as required by RFCs.
10682 *Goetz Babin-Ebell*
10684 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10686 * Introduce limits to prevent malicious keys being able to
10687 cause a denial of service. ([CVE-2006-2940])
10689 *Steve Henson, Bodo Moeller*
10691 * Fix ASN.1 parsing of certain invalid structures that can result
10692 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10694 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10695 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10697 * Fix SSL client code which could crash if connecting to a
10698 malicious SSLv2 server. ([CVE-2006-4343])
10700 *Tavis Ormandy and Will Drewry, Google Security Team*
10702 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10703 match only those. Before that, "AES256-SHA" would be interpreted
10704 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10705 the same strength classification in 0.9.7h) as we currently only
10706 have a single AES bit in the ciphersuite description bitmap.
10707 That change, however, also applied to ciphersuite strings such as
10708 "RC4-MD5" that intentionally matched multiple ciphersuites --
10709 namely, SSL 2.0 ciphersuites in addition to the more common ones
10710 from SSL 3.0/TLS 1.0.
10712 So we change the selection algorithm again: Naming an explicit
10713 ciphersuite selects this one ciphersuite, and any other similar
10714 ciphersuite (same bitmap) from *other* protocol versions.
10715 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10716 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10718 Since SSL 2.0 does not have any ciphersuites for which the
10719 128/256 bit distinction would be relevant, this works for now.
10720 The proper fix will be to use different bits for AES128 and
10721 AES256, which would have avoided the problems from the beginning;
10722 however, bits are scarce, so we can only do this in a new release
10723 (not just a patchlevel) when we can change the SSL_CIPHER
10724 definition to split the single 'unsigned long mask' bitmap into
10725 multiple values to extend the available space.
10729 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10731 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10732 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10734 * Add AES IGE and biIGE modes.
10738 * Change the Unix randomness entropy gathering to use poll() when
10739 possible instead of select(), since the latter has some
10740 undesirable limitations.
10742 *Darryl Miles via Richard Levitte and Bodo Moeller*
10744 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10745 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10746 cannot be implicitly activated as part of, e.g., the "AES" alias.
10747 However, please upgrade to OpenSSL 0.9.9[-dev] for
10748 non-experimental use of the ECC ciphersuites to get TLS extension
10749 support, which is required for curve and point format negotiation
10750 to avoid potential handshake problems.
10754 * Disable rogue ciphersuites:
10756 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10757 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10758 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10760 The latter two were purportedly from
10761 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10764 Also deactivate the remaining ciphersuites from
10765 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10766 unofficial, and the ID has long expired.
10770 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10771 dual-core machines) and other potential thread-safety issues.
10775 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10776 versions), which is now available for royalty-free use
10777 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10778 Also, add Camellia TLS ciphersuites from RFC 4132.
10780 To minimize changes between patchlevels in the OpenSSL 0.9.8
10781 series, Camellia remains excluded from compilation unless OpenSSL
10782 is configured with 'enable-camellia'.
10786 * Disable the padding bug check when compression is in use. The padding
10787 bug check assumes the first packet is of even length, this is not
10788 necessarily true if compression is enabled and can result in false
10789 positives causing handshake failure. The actual bug test is ancient
10790 code so it is hoped that implementations will either have fixed it by
10791 now or any which still have the bug do not support compression.
10795 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10797 * When applying a cipher rule check to see if string match is an explicit
10798 cipher suite and only match that one cipher suite if it is.
10802 * Link in manifests for VC++ if needed.
10804 *Austin Ziegler <halostatue@gmail.com>*
10806 * Update support for ECC-based TLS ciphersuites according to
10807 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10808 TLS extensions, which are supported starting with the 0.9.9
10809 branch, not in the OpenSSL 0.9.8 branch).
10813 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10814 opaque EVP_CIPHER_CTX handling.
10818 * Fixes and enhancements to zlib compression code. We now only use
10819 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10820 to conform with the standards mentioned here:
10821 <http://www.zlib.net/DLL_FAQ.txt>
10822 Static zlib linking now works on Windows and the new --with-zlib-include
10823 --with-zlib-lib options to Configure can be used to supply the location
10824 of the headers and library. Gracefully handle case where zlib library
10829 * Several fixes and enhancements to the OID generation code. The old code
10830 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10831 handle numbers larger than ULONG_MAX, truncated printing and had a
10832 non standard OBJ_obj2txt() behaviour.
10836 * Add support for building of engines under engine/ as shared libraries
10837 under VC++ build system.
10841 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10842 Hopefully, we will not see any false combination of paths any more.
10846 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10848 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10849 (part of SSL_OP_ALL). This option used to disable the
10850 countermeasure against man-in-the-middle protocol-version
10851 rollback in the SSL 2.0 server implementation, which is a bad
10852 idea. ([CVE-2005-2969])
10854 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10855 for Information Security, National Institute of Advanced Industrial
10856 Science and Technology [AIST], Japan)*
10858 * Add two function to clear and return the verify parameter flags.
10862 * Keep cipherlists sorted in the source instead of sorting them at
10863 runtime, thus removing the need for a lock.
10867 * Avoid some small subgroup attacks in Diffie-Hellman.
10869 *Nick Mathewson and Ben Laurie*
10871 * Add functions for well-known primes.
10875 * Extended Windows CE support.
10877 *Satoshi Nakamura and Andy Polyakov*
10879 * Initialize SSL_METHOD structures at compile time instead of during
10880 runtime, thus removing the need for a lock.
10884 * Make PKCS7_decrypt() work even if no certificate is supplied by
10885 attempting to decrypt each encrypted key in turn. Add support to
10890 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10892 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10895 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10899 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10900 key into the same file any more.
10904 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10908 * Add -utf8 command line and config file option to 'ca'.
10910 *Stefan <stf@udoma.org*
10912 * Removed the macro des_crypt(), as it seems to conflict with some
10913 libraries. Use DES_crypt().
10917 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10918 involves renaming the source and generated shared-libs for
10919 both. The engines will accept the corrected or legacy ids
10920 ('ncipher' and '4758_cca' respectively) when binding. NB,
10921 this only applies when building 'shared'.
10923 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10925 * Add attribute functions to EVP_PKEY structure. Modify
10926 PKCS12_create() to recognize a CSP name attribute and
10927 use it. Make -CSP option work again in pkcs12 utility.
10931 * Add new functionality to the bn blinding code:
10932 - automatic re-creation of the BN_BLINDING parameters after
10933 a fixed number of uses (currently 32)
10934 - add new function for parameter creation
10935 - introduce flags to control the update behaviour of the
10936 BN_BLINDING parameters
10937 - hide BN_BLINDING structure
10938 Add a second BN_BLINDING slot to the RSA structure to improve
10939 performance when a single RSA object is shared among several
10944 * Add support for DTLS.
10946 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10948 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10949 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10953 * Remove buggy and incomplete DH cert support from
10954 ssl/ssl_rsa.c and ssl/s3_both.c
10958 * Use SHA-1 instead of MD5 as the default digest algorithm for
10959 the `apps/openssl` commands.
10963 * Compile clean with "-Wall -Wmissing-prototypes
10964 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10965 DEBUG_SAFESTACK must also be set.
10969 * Change ./Configure so that certain algorithms can be disabled by default.
10970 The new counterpiece to "no-xxx" is "enable-xxx".
10972 The patented RC5 and MDC2 algorithms will now be disabled unless
10973 "enable-rc5" and "enable-mdc2", respectively, are specified.
10975 (IDEA remains enabled despite being patented. This is because IDEA
10976 is frequently required for interoperability, and there is no license
10977 fee for non-commercial use. As before, "no-idea" can be used to
10978 avoid this algorithm.)
10982 * Add processing of proxy certificates (see RFC 3820). This work was
10983 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10984 EGEE (Enabling Grids for E-science in Europe).
10988 * RC4 performance overhaul on modern architectures/implementations, such
10989 as Intel P4, IA-64 and AMD64.
10993 * New utility extract-section.pl. This can be used specify an alternative
10994 section number in a pod file instead of having to treat each file as
10995 a separate case in Makefile. This can be done by adding two lines to the
10998 =for comment openssl_section:XXX
11000 The blank line is mandatory.
11004 * New arguments -certform, -keyform and -pass for s_client and s_server
11005 to allow alternative format key and certificate files and passphrase
11010 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11011 update associated structures and add various utility functions.
11013 Add new policy related verify parameters, include policy checking in
11014 standard verify code. Enhance 'smime' application with extra parameters
11015 to support policy checking and print out.
11019 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11020 Nehemiah processors. These extensions support AES encryption in hardware
11021 as well as RNG (though RNG support is currently disabled).
11023 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11025 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11029 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11031 *Andy Polyakov and a number of other people*
11033 * Improved PowerPC platform support. Most notably BIGNUM assembler
11034 implementation contributed by IBM.
11036 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11038 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11039 exponent rather than 'unsigned long'. There is a corresponding change to
11040 the new 'rsa_keygen' element of the RSA_METHOD structure.
11042 *Jelte Jansen, Geoff Thorpe*
11044 * Functionality for creating the initial serial number file is now
11045 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11047 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11048 number file to 1, which is bound to cause problems. To avoid
11049 the problems while respecting compatibility between different 0.9.7
11050 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11051 CA.pl for serial number initialization. With the new release 0.9.8,
11052 we can fix the problem directly in the 'ca' utility.)
11056 * Reduced header interdependencies by declaring more opaque objects in
11057 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11058 give fewer recursive includes, which could break lazy source code - so
11059 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11060 developers should define this symbol when building and using openssl to
11061 ensure they track the recommended behaviour, interfaces, [etc], but
11062 backwards-compatible behaviour prevails when this isn't defined.
11066 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11070 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11071 This will generate a random key of the appropriate length based on the
11072 cipher context. The EVP_CIPHER can provide its own random key generation
11073 routine to support keys of a specific form. This is used in the des and
11074 3des routines to generate a key of the correct parity. Update S/MIME
11075 code to use new functions and hence generate correct parity DES keys.
11076 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11077 valid (weak or incorrect parity).
11081 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11082 as looking them up. This is useful when the verified structure may contain
11083 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11084 present unless the new PKCS7_NO_CRL flag is asserted.
11088 * Extend ASN1 oid configuration module. It now additionally accepts the
11091 shortName = some long name, 1.2.3.4
11095 * Reimplemented the BN_CTX implementation. There is now no more static
11096 limitation on the number of variables it can handle nor the depth of the
11097 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11098 information can now expand as required, and rather than having a single
11099 static array of bignums, BN_CTX now uses a linked-list of such arrays
11100 allowing it to expand on demand whilst maintaining the usefulness of
11101 BN_CTX's "bundling".
11105 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11106 to allow all RSA operations to function using a single BN_CTX.
11110 * Preliminary support for certificate policy evaluation and checking. This
11111 is initially intended to pass the tests outlined in "Conformance Testing
11112 of Relying Party Client Certificate Path Processing Logic" v1.07.
11116 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11117 remained unused and not that useful. A variety of other little bignum
11118 tweaks and fixes have also been made continuing on from the audit (see
11123 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11124 associated ASN1, EVP and SSL functions and old ASN1 macros.
11128 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11129 and this should never fail. So the return value from the use of
11130 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11131 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11135 * BN_CTX_get() should return zero-valued bignums, providing the same
11136 initialised value as BN_new().
11138 *Geoff Thorpe, suggested by Ulf Möller*
11140 * Support for inhibitAnyPolicy certificate extension.
11144 * An audit of the BIGNUM code is underway, for which debugging code is
11145 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11146 is considered valid when processing BIGNUMs, and causes execution to
11147 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11148 further steps are taken to deliberately pollute unused data in BIGNUM
11149 structures to try and expose faulty code further on. For now, openssl will
11150 (in its default mode of operation) continue to tolerate the inconsistent
11151 forms that it has tolerated in the past, but authors and packagers should
11152 consider trying openssl and their own applications when compiled with
11153 these debugging symbols defined. It will help highlight potential bugs in
11154 their own code, and will improve the test coverage for OpenSSL itself. At
11155 some point, these tighter rules will become openssl's default to improve
11156 maintainability, though the assert()s and other overheads will remain only
11157 in debugging configurations. See bn.h for more details.
11159 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11161 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11162 that can only be obtained through BN_CTX_new() (which implicitly
11163 initialises it). The presence of this function only made it possible
11164 to overwrite an existing structure (and cause memory leaks).
11168 * Because of the callback-based approach for implementing LHASH as a
11169 template type, lh_insert() adds opaque objects to hash-tables and
11170 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11171 to clean up those corresponding objects before destroying the hash table
11172 (and losing the object pointers). So some over-zealous constifications in
11173 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11174 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11175 prototyped to have "const" restrictions on the object pointers they are
11176 given (and so aren't required to cast them away any more).
11180 * The tmdiff.h API was so ugly and minimal that our own timing utility
11181 (speed) prefers to use its own implementation. The two implementations
11182 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11183 its object type properly exposed (MS_TM) instead of casting to/from
11184 `char *`. This may still change yet if someone realises MS_TM and
11186 aren't necessarily the greatest nomenclatures - but this is what was used
11187 internally to the implementation so I've used that for now.
11191 * Ensure that deprecated functions do not get compiled when
11192 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11193 the self-tests were still using deprecated key-generation functions so
11194 these have been updated also.
11198 * Reorganise PKCS#7 code to separate the digest location functionality
11199 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11200 New function PKCS7_set_digest() to set the digest type for PKCS#7
11201 digestedData type. Add additional code to correctly generate the
11202 digestedData type and add support for this type in PKCS7 initialization
11207 * New function PKCS7_set0_type_other() this initializes a PKCS7
11208 structure of type "other".
11212 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11213 sure the loop does correctly stop and breaking ("division by zero")
11214 modulus operations are not performed. The (pre-generated) prime
11215 table crypto/bn/bn_prime.h was already correct, but it could not be
11216 re-generated on some platforms because of the "division by zero"
11217 situation in the script.
11219 *Ralf S. Engelschall*
11221 * Update support for ECC-based TLS ciphersuites according to
11222 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11223 SHA-1 now is only used for "small" curves (where the
11224 representation of a field element takes up to 24 bytes); for
11225 larger curves, the field element resulting from ECDH is directly
11226 used as premaster secret.
11228 *Douglas Stebila (Sun Microsystems Laboratories)*
11230 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11231 curve secp160r1 to the tests.
11233 *Douglas Stebila (Sun Microsystems Laboratories)*
11235 * Add the possibility to load symbols globally with DSO.
11237 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11239 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11240 control of the error stack.
11244 * Add support for STORE in ENGINE.
11248 * Add the STORE type. The intention is to provide a common interface
11249 to certificate and key stores, be they simple file-based stores, or
11250 HSM-type store, or LDAP stores, or...
11251 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11255 * Add a generic structure called OPENSSL_ITEM. This can be used to
11256 pass a list of arguments to any function as well as provide a way
11257 for a function to pass data back to the caller.
11261 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11262 works like BUF_strdup() but can be used to duplicate a portion of
11263 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11268 * Add the function sk_find_ex() which works like sk_find(), but will
11269 return an index to an element even if an exact match couldn't be
11270 found. The index is guaranteed to point at the element where the
11271 searched-for key would be inserted to preserve sorting order.
11275 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11276 takes an extra flags argument for optional functionality. Currently,
11277 the following flags are defined:
11279 OBJ_BSEARCH_VALUE_ON_NOMATCH
11280 This one gets OBJ_bsearch_ex() to return a pointer to the first
11281 element where the comparing function returns a negative or zero
11284 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11285 This one gets OBJ_bsearch_ex() to return a pointer to the first
11286 element where the comparing function returns zero. This is useful
11287 if there are more than one element where the comparing function
11292 * Make it possible to create self-signed certificates with 'openssl ca'
11293 in such a way that the self-signed certificate becomes part of the
11294 CA database and uses the same mechanisms for serial number generation
11295 as all other certificate signing. The new flag '-selfsign' enables
11296 this functionality. Adapt CA.sh and CA.pl.in.
11300 * Add functionality to check the public key of a certificate request
11301 against a given private. This is useful to check that a certificate
11302 request can be signed by that key (self-signing).
11306 * Make it possible to have multiple active certificates with the same
11307 subject in the CA index file. This is done only if the keyword
11308 'unique_subject' is set to 'no' in the main CA section (default
11309 if 'CA_default') of the configuration file. The value is saved
11310 with the database itself in a separate index attribute file,
11311 named like the index file with '.attr' appended to the name.
11315 * Generate multi-valued AVAs using '+' notation in config files for
11320 * Support for nameConstraints certificate extension.
11324 * Support for policyConstraints certificate extension.
11328 * Support for policyMappings certificate extension.
11332 * Make sure the default DSA_METHOD implementation only uses its
11333 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11334 and change its own handlers to be NULL so as to remove unnecessary
11335 indirection. This lets alternative implementations fallback to the
11336 default implementation more easily.
11340 * Support for directoryName in GeneralName related extensions
11345 * Make it possible to link applications using Makefile.shared.
11346 Make that possible even when linking against static libraries!
11350 * Support for single pass processing for S/MIME signing. This now
11351 means that S/MIME signing can be done from a pipe, in addition
11352 cleartext signing (multipart/signed type) is effectively streaming
11353 and the signed data does not need to be all held in memory.
11355 This is done with a new flag PKCS7_STREAM. When this flag is set
11356 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11357 is done after the data is output (and digests calculated) in
11358 SMIME_write_PKCS7().
11362 * Add full support for -rpath/-R, both in shared libraries and
11363 applications, at least on the platforms where it's known how
11368 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11369 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11370 will now compute a table of multiples of the generator that
11371 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11372 faster (notably in the case of a single point multiplication,
11373 scalar * generator).
11375 *Nils Larsch, Bodo Moeller*
11377 * IPv6 support for certificate extensions. The various extensions
11378 which use the IP:a.b.c.d can now take IPv6 addresses using the
11379 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11384 * Added an ENGINE that implements RSA by performing private key
11385 exponentiations with the GMP library. The conversions to and from
11386 GMP's mpz_t format aren't optimised nor are any montgomery forms
11387 cached, and on x86 it appears OpenSSL's own performance has caught up.
11388 However there are likely to be other architectures where GMP could
11389 provide a boost. This ENGINE is not built in by default, but it can be
11390 specified at Configure time and should be accompanied by the necessary
11391 linker additions, eg;
11392 ./config -DOPENSSL_USE_GMP -lgmp
11396 * "openssl engine" will not display ENGINE/DSO load failure errors when
11397 testing availability of engines with "-t" - the old behaviour is
11398 produced by increasing the feature's verbosity with "-tt".
11402 * ECDSA routines: under certain error conditions uninitialized BN objects
11403 could be freed. Solution: make sure initialization is performed early
11404 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11409 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11410 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11411 software implementations. For DSA and DH, parameter generation can
11412 also be overridden by providing the appropriate method callbacks.
11416 * Change the "progress" mechanism used in key-generation and
11417 primality testing to functions that take a new BN_GENCB pointer in
11418 place of callback/argument pairs. The new API functions have `_ex`
11419 postfixes and the older functions are reimplemented as wrappers for
11420 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11421 declarations of the old functions to help (graceful) attempts to
11422 migrate to the new functions. Also, the new key-generation API
11423 functions operate on a caller-supplied key-structure and return
11424 success/failure rather than returning a key or NULL - this is to
11425 help make "keygen" another member function of RSA_METHOD etc.
11427 Example for using the new callback interface:
11429 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11430 void *my_arg = ...;
11433 BN_GENCB_set(&my_cb, my_callback, my_arg);
11435 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11436 /* For the meaning of a, b in calls to my_callback(), see the
11437 * documentation of the function that calls the callback.
11438 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11439 * my_callback should return 1 if it wants BN_is_prime_ex()
11440 * to continue, or 0 to stop.
11445 * Change the ZLIB compression method to be stateful, and make it
11446 available to TLS with the number defined in
11447 draft-ietf-tls-compression-04.txt.
11451 * Add the ASN.1 structures and functions for CertificatePair, which
11452 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11454 CertificatePair ::= SEQUENCE {
11455 forward [0] Certificate OPTIONAL,
11456 reverse [1] Certificate OPTIONAL,
11457 -- at least one of the pair shall be present -- }
11459 Also implement the PEM functions to read and write certificate
11460 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11462 This needed to be defined, mostly for the sake of the LDAP
11463 attribute crossCertificatePair, but may prove useful elsewhere as
11468 * Make it possible to inhibit symlinking of shared libraries in
11469 Makefile.shared, for Cygwin's sake.
11473 * Extend the BIGNUM API by creating a function
11474 void BN_set_negative(BIGNUM *a, int neg);
11475 and a macro that behave like
11476 int BN_is_negative(const BIGNUM *a);
11478 to avoid the need to access 'a->neg' directly in applications.
11482 * Implement fast modular reduction for pseudo-Mersenne primes
11483 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11484 EC_GROUP_new_curve_GFp() will now automatically use this
11487 *Nils Larsch <nla@trustcenter.de>*
11489 * Add new lock type (CRYPTO_LOCK_BN).
11493 * Change the ENGINE framework to automatically load engines
11494 dynamically from specific directories unless they could be
11495 found to already be built in or loaded. Move all the
11496 current engines except for the cryptodev one to a new
11497 directory engines/.
11498 The engines in engines/ are built as shared libraries if
11499 the "shared" options was given to ./Configure or ./config.
11500 Otherwise, they are inserted in libcrypto.a.
11501 /usr/local/ssl/engines is the default directory for dynamic
11502 engines, but that can be overridden at configure time through
11503 the usual use of --prefix and/or --openssldir, and at run
11504 time with the environment variable OPENSSL_ENGINES.
11506 *Geoff Thorpe and Richard Levitte*
11508 * Add Makefile.shared, a helper makefile to build shared
11509 libraries. Adapt Makefile.org.
11513 * Add version info to Win32 DLLs.
11515 *Peter 'Luna' Runestig" <peter@runestig.com>*
11517 * Add new 'medium level' PKCS#12 API. Certificates and keys
11518 can be added using this API to created arbitrary PKCS#12
11519 files while avoiding the low-level API.
11521 New options to PKCS12_create(), key or cert can be NULL and
11522 will then be omitted from the output file. The encryption
11523 algorithm NIDs can be set to -1 for no encryption, the mac
11524 iteration count can be set to 0 to omit the mac.
11526 Enhance pkcs12 utility by making the -nokeys and -nocerts
11527 options work when creating a PKCS#12 file. New option -nomac
11528 to omit the mac, NONE can be set for an encryption algorithm.
11529 New code is modified to use the enhanced PKCS12_create()
11530 instead of the low-level API.
11534 * Extend ASN1 encoder to support indefinite length constructed
11535 encoding. This can output sequences tags and octet strings in
11536 this form. Modify pk7_asn1.c to support indefinite length
11537 encoding. This is experimental and needs additional code to
11538 be useful, such as an ASN1 bio and some enhanced streaming
11541 Extend template encode functionality so that tagging is passed
11542 down to the template encoder.
11546 * Let 'openssl req' fail if an argument to '-newkey' is not
11547 recognized instead of using RSA as a default.
11551 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11552 As these are not official, they are not included in "ALL";
11553 the "ECCdraft" ciphersuite group alias can be used to select them.
11555 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11557 * Add ECDH engine support.
11559 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11561 * Add ECDH in new directory crypto/ecdh/.
11563 *Douglas Stebila (Sun Microsystems Laboratories)*
11565 * Let BN_rand_range() abort with an error after 100 iterations
11566 without success (which indicates a broken PRNG).
11570 * Change BN_mod_sqrt() so that it verifies that the input value
11571 is really the square of the return value. (Previously,
11572 BN_mod_sqrt would show GIGO behaviour.)
11576 * Add named elliptic curves over binary fields from X9.62, SECG,
11577 and WAP/WTLS; add OIDs that were still missing.
11579 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11581 * Extend the EC library for elliptic curves over binary fields
11582 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11585 EC_GF2m_simple_method
11589 EC_GROUP_new_curve_GF2m
11590 EC_GROUP_set_curve_GF2m
11591 EC_GROUP_get_curve_GF2m
11592 EC_POINT_set_affine_coordinates_GF2m
11593 EC_POINT_get_affine_coordinates_GF2m
11594 EC_POINT_set_compressed_coordinates_GF2m
11596 Point compression for binary fields is disabled by default for
11597 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11600 As binary polynomials are represented as BIGNUMs, various members
11601 of the EC_GROUP and EC_POINT data structures can be shared
11602 between the implementations for prime fields and binary fields;
11603 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11604 are essentially identical to their `..._GFp` counterparts.
11605 (For simplicity, the `..._GFp` prefix has been dropped from
11606 various internal method names.)
11608 An internal 'field_div' method (similar to 'field_mul' and
11609 'field_sqr') has been added; this is used only for binary fields.
11611 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11613 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11614 through methods ('mul', 'precompute_mult').
11616 The generic implementations (now internally called 'ec_wNAF_mul'
11617 and 'ec_wNAF_precomputed_mult') remain the default if these
11618 methods are undefined.
11620 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11622 * New function EC_GROUP_get_degree, which is defined through
11623 EC_METHOD. For curves over prime fields, this returns the bit
11624 length of the modulus.
11626 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11628 * New functions EC_GROUP_dup, EC_POINT_dup.
11629 (These simply call ..._new and ..._copy).
11631 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11633 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11634 Polynomials are represented as BIGNUMs (where the sign bit is not
11635 used) in the following functions [macros]:
11638 BN_GF2m_sub [= BN_GF2m_add]
11639 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11640 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11641 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11643 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11644 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11645 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11646 BN_GF2m_cmp [= BN_ucmp]
11648 (Note that only the 'mod' functions are actually for fields GF(2^m).
11649 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11651 For some functions, an the irreducible polynomial defining a
11652 field can be given as an 'unsigned int[]' with strictly
11653 decreasing elements giving the indices of those bits that are set;
11654 i.e., p[] represents the polynomial
11655 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11657 p[0] > p[1] > ... > p[k] = 0.
11658 This applies to the following functions:
11661 BN_GF2m_mod_mul_arr
11662 BN_GF2m_mod_sqr_arr
11663 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11664 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11665 BN_GF2m_mod_exp_arr
11666 BN_GF2m_mod_sqrt_arr
11667 BN_GF2m_mod_solve_quad_arr
11671 Conversion can be performed by the following functions:
11676 bntest.c has additional tests for binary polynomial arithmetic.
11678 Two implementations for BN_GF2m_mod_div() are available.
11679 The default algorithm simply uses BN_GF2m_mod_inv() and
11680 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11681 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11682 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11684 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11686 * Add new error code 'ERR_R_DISABLED' that can be used when some
11687 functionality is disabled at compile-time.
11689 *Douglas Stebila <douglas.stebila@sun.com>*
11691 * Change default behaviour of 'openssl asn1parse' so that more
11692 information is visible when viewing, e.g., a certificate:
11694 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11695 mode the content of non-printable OCTET STRINGs is output in a
11696 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11697 avoid the appearance of a printable string.
11699 *Nils Larsch <nla@trustcenter.de>*
11701 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11703 EC_GROUP_set_asn1_flag()
11704 EC_GROUP_get_asn1_flag()
11705 EC_GROUP_set_point_conversion_form()
11706 EC_GROUP_get_point_conversion_form()
11707 These control ASN1 encoding details:
11708 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11709 has been set to OPENSSL_EC_NAMED_CURVE.
11710 - Points are encoded in uncompressed form by default; options for
11711 asn1_for are as for point2oct, namely
11712 POINT_CONVERSION_COMPRESSED
11713 POINT_CONVERSION_UNCOMPRESSED
11714 POINT_CONVERSION_HYBRID
11716 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11718 EC_GROUP_set_seed()
11719 EC_GROUP_get0_seed()
11720 EC_GROUP_get_seed_len()
11721 This is used only for ASN1 purposes (so far).
11723 *Nils Larsch <nla@trustcenter.de>*
11725 * Add 'field_type' member to EC_METHOD, which holds the NID
11726 of the appropriate field type OID. The new function
11727 EC_METHOD_get_field_type() returns this value.
11729 *Nils Larsch <nla@trustcenter.de>*
11732 EC_POINT_point2bn()
11733 EC_POINT_bn2point()
11734 EC_POINT_point2hex()
11735 EC_POINT_hex2point()
11736 providing useful interfaces to EC_POINT_point2oct() and
11737 EC_POINT_oct2point().
11739 *Nils Larsch <nla@trustcenter.de>*
11741 * Change internals of the EC library so that the functions
11742 EC_GROUP_set_generator()
11743 EC_GROUP_get_generator()
11744 EC_GROUP_get_order()
11745 EC_GROUP_get_cofactor()
11746 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11747 to methods, which would lead to unnecessary code duplication when
11748 adding different types of curves.
11750 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11752 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11753 arithmetic, and such that modified wNAFs are generated
11754 (which avoid length expansion in many cases).
11758 * Add a function EC_GROUP_check_discriminant() (defined via
11759 EC_METHOD) that verifies that the curve discriminant is non-zero.
11761 Add a function EC_GROUP_check() that makes some sanity tests
11762 on a EC_GROUP, its generator and order. This includes
11763 EC_GROUP_check_discriminant().
11765 *Nils Larsch <nla@trustcenter.de>*
11767 * Add ECDSA in new directory crypto/ecdsa/.
11769 Add applications 'openssl ecparam' and 'openssl ecdsa'
11770 (these are based on 'openssl dsaparam' and 'openssl dsa').
11772 ECDSA support is also included in various other files across the
11773 library. Most notably,
11774 - 'openssl req' now has a '-newkey ecdsa:file' option;
11775 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11776 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11777 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11778 them suitable for ECDSA where domain parameters must be
11779 extracted before the specific public key;
11780 - ECDSA engine support has been added.
11782 *Nils Larsch <nla@trustcenter.de>*
11784 * Include some named elliptic curves, and add OIDs from X9.62,
11785 SECG, and WAP/WTLS. Each curve can be obtained from the new
11787 EC_GROUP_new_by_curve_name(),
11788 and the list of available named curves can be obtained with
11789 EC_get_builtin_curves().
11790 Also add a 'curve_name' member to EC_GROUP objects, which can be
11792 EC_GROUP_set_curve_name()
11793 EC_GROUP_get_curve_name()
11795 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11797 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11798 was actually never needed) and in BN_mul(). The removal in BN_mul()
11799 required a small change in bn_mul_part_recursive() and the addition
11800 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11801 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11802 bn_sub_words() and bn_add_words() except they take arrays with
11807 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11809 * Cleanse PEM buffers before freeing them since they may contain
11812 *Benjamin Bennett <ben@psc.edu>*
11814 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11815 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11816 authentication-only ciphersuites.
11820 * Since AES128 and AES256 share a single mask bit in the logic of
11821 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11822 kludge to work properly if AES128 is available and AES256 isn't.
11826 * Expand security boundary to match 1.1.1 module.
11830 * Remove redundant features: hash file source, editing of test vectors
11831 modify fipsld to use external fips_premain.c signature.
11835 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11836 run algorithm test programs.
11840 * Make algorithm test programs more tolerant of whitespace.
11844 * Have SSL/TLS server implementation tolerate "mismatched" record
11845 protocol version while receiving ClientHello even if the
11846 ClientHello is fragmented. (The server can't insist on the
11847 particular protocol version it has chosen before the ServerHello
11848 message has informed the client about his choice.)
11852 * Load error codes if they are not already present instead of using a
11853 static variable. This allows them to be cleanly unloaded and reloaded.
11857 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11859 * Introduce limits to prevent malicious keys being able to
11860 cause a denial of service. ([CVE-2006-2940])
11862 *Steve Henson, Bodo Moeller*
11864 * Fix ASN.1 parsing of certain invalid structures that can result
11865 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11867 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11868 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11870 * Fix SSL client code which could crash if connecting to a
11871 malicious SSLv2 server. ([CVE-2006-4343])
11873 *Tavis Ormandy and Will Drewry, Google Security Team*
11875 * Change ciphersuite string processing so that an explicit
11876 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11877 will no longer include "AES128-SHA"), and any other similar
11878 ciphersuite (same bitmap) from *other* protocol versions (so that
11879 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11880 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11881 changes from 0.9.8b and 0.9.8d.
11885 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11887 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11888 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11890 * Change the Unix randomness entropy gathering to use poll() when
11891 possible instead of select(), since the latter has some
11892 undesirable limitations.
11894 *Darryl Miles via Richard Levitte and Bodo Moeller*
11896 * Disable rogue ciphersuites:
11898 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11899 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11900 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11902 The latter two were purportedly from
11903 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11906 Also deactivate the remaining ciphersuites from
11907 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11908 unofficial, and the ID has long expired.
11912 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11913 dual-core machines) and other potential thread-safety issues.
11917 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11919 * Adapt fipsld and the build system to link against the validated FIPS
11920 module in FIPS mode.
11924 * Fixes for VC++ 2005 build under Windows.
11928 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11929 from a Windows bash shell such as MSYS. It is autodetected from the
11930 "config" script when run from a VC++ environment. Modify standard VC++
11931 build to use fipscanister.o from the GNU make build.
11935 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11937 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11938 The value now differs depending on if you build for FIPS or not.
11939 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11940 safely run with a non-FIPSed libcrypto, as it may crash because of
11941 the difference induced by this change.
11945 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11947 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11948 (part of SSL_OP_ALL). This option used to disable the
11949 countermeasure against man-in-the-middle protocol-version
11950 rollback in the SSL 2.0 server implementation, which is a bad
11951 idea. ([CVE-2005-2969])
11953 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11954 for Information Security, National Institute of Advanced Industrial
11955 Science and Technology [AIST, Japan)]*
11957 * Minimal support for X9.31 signatures and PSS padding modes. This is
11958 mainly for FIPS compliance and not fully integrated at this stage.
11962 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11963 the exponentiation using a fixed-length exponent. (Otherwise,
11964 the information leaked through timing could expose the secret key
11965 after many signatures; cf. Bleichenbacher's attack on DSA with
11970 * Make a new fixed-window mod_exp implementation the default for
11971 RSA, DSA, and DH private-key operations so that the sequence of
11972 squares and multiplies and the memory access pattern are
11973 independent of the particular secret key. This will mitigate
11974 cache-timing and potential related attacks.
11976 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11977 and this is automatically used by BN_mod_exp_mont() if the new flag
11978 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11979 will use this BN flag for private exponents unless the flag
11980 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11981 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11983 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11985 * Change the client implementation for SSLv23_method() and
11986 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11987 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11988 (Previously, the SSL 2.0 backwards compatible Client Hello
11989 message format would be used even with SSL_OP_NO_SSLv2.)
11993 * Add support for smime-type MIME parameter in S/MIME messages which some
11998 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11999 a threadsafe manner. Modify rsa code to use new function and add calls
12000 to dsa and dh code (which had race conditions before).
12004 * Include the fixed error library code in the C error file definitions
12005 instead of fixing them up at runtime. This keeps the error code
12006 structures constant.
12010 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12012 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12015 * Fixes for newer kerberos headers. NB: the casts are needed because
12016 the 'length' field is signed on one version and unsigned on another
12017 with no (?) obvious way to tell the difference, without these VC++
12018 complains. Also the "definition" of FAR (blank) is no longer included
12019 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12020 some needed definitions.
12024 * Undo Cygwin change.
12028 * Added support for proxy certificates according to RFC 3820.
12029 Because they may be a security thread to unaware applications,
12030 they must be explicitly allowed in run-time. See
12031 docs/HOWTO/proxy_certificates.txt for further information.
12035 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12037 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12038 server and client random values. Previously
12039 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12040 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12042 This change has negligible security impact because:
12044 1. Server and client random values still have 24 bytes of pseudo random
12047 2. Server and client random values are sent in the clear in the initial
12050 3. The master secret is derived using the premaster secret (48 bytes in
12051 size for static RSA ciphersuites) as well as client server and random
12054 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12057 *Stephen Henson, reported by UK NISCC*
12059 * Use Windows randomness collection on Cygwin.
12063 * Fix hang in EGD/PRNGD query when communication socket is closed
12064 prematurely by EGD/PRNGD.
12066 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12068 * Prompt for pass phrases when appropriate for PKCS12 input format.
12072 * Back-port of selected performance improvements from development
12073 branch, as well as improved support for PowerPC platforms.
12077 * Add lots of checks for memory allocation failure, error codes to indicate
12078 failure and freeing up memory if a failure occurs.
12080 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12082 * Add new -passin argument to dgst.
12086 * Perform some character comparisons of different types in X509_NAME_cmp:
12087 this is needed for some certificates that re-encode DNs into UTF8Strings
12088 (in violation of RFC3280) and can't or won't issue name rollover
12093 * Make an explicit check during certificate validation to see that
12094 the CA setting in each certificate on the chain is correct. As a
12095 side effect always do the following basic checks on extensions,
12096 not just when there's an associated purpose to the check:
12098 - if there is an unhandled critical extension (unless the user
12099 has chosen to ignore this fault)
12100 - if the path length has been exceeded (if one is set at all)
12101 - that certain extensions fit the associated purpose (if one has
12106 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12108 * Avoid a race condition when CRLs are checked in a multi threaded
12109 environment. This would happen due to the reordering of the revoked
12110 entries during signature checking and serial number lookup. Now the
12111 encoding is cached and the serial number sort performed under a lock.
12112 Add new STACK function sk_is_sorted().
12116 * Add Delta CRL to the extension code.
12120 * Various fixes to s3_pkt.c so alerts are sent properly.
12122 *David Holmes <d.holmes@f5.com>*
12124 * Reduce the chances of duplicate issuer name and serial numbers (in
12125 violation of RFC3280) using the OpenSSL certificate creation utilities.
12126 This is done by creating a random 64 bit value for the initial serial
12127 number when a serial number file is created or when a self signed
12128 certificate is created using 'openssl req -x509'. The initial serial
12129 number file is created using 'openssl x509 -next_serial' in CA.pl
12130 rather than being initialized to 1.
12134 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12136 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12137 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12139 *Joe Orton, Steve Henson*
12141 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12144 *Joe Orton, Steve Henson*
12146 * Make it possible to have multiple active certificates with the same
12147 subject in the CA index file. This is done only if the keyword
12148 'unique_subject' is set to 'no' in the main CA section (default
12149 if 'CA_default') of the configuration file. The value is saved
12150 with the database itself in a separate index attribute file,
12151 named like the index file with '.attr' appended to the name.
12155 * X509 verify fixes. Disable broken certificate workarounds when
12156 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12157 keyUsage extension present. Don't accept CRLs with unhandled critical
12158 extensions: since verify currently doesn't process CRL extensions this
12159 rejects a CRL with *any* critical extensions. Add new verify error codes
12164 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12165 A clarification of RFC2560 will require the use of OCTET STRINGs and
12166 some implementations cannot handle the current raw format. Since OpenSSL
12167 copies and compares OCSP nonces as opaque blobs without any attempt at
12168 parsing them this should not create any compatibility issues.
12172 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12173 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12174 this HMAC (and other) operations are several times slower than OpenSSL
12179 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12181 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12183 * Use the correct content when signing type "other".
12187 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12189 * Fix various bugs revealed by running the NISCC test suite:
12191 Stop out of bounds reads in the ASN1 code when presented with
12192 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12194 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12196 If verify callback ignores invalid public key errors don't try to check
12197 certificate signature with the NULL public key.
12201 * New -ignore_err option in ocsp application to stop the server
12202 exiting on the first error in a request.
12206 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12207 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12212 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12213 extra data after the compression methods not only for TLS 1.0
12214 but also for SSL 3.0 (as required by the specification).
12216 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12218 * Change X509_certificate_type() to mark the key as exported/exportable
12219 when it's 512 *bits* long, not 512 bytes.
12223 * Change AES_cbc_encrypt() so it outputs exact multiple of
12224 blocks during encryption.
12228 * Various fixes to base64 BIO and non blocking I/O. On write
12229 flushes were not handled properly if the BIO retried. On read
12230 data was not being buffered properly and had various logic bugs.
12231 This also affects blocking I/O when the data being decoded is a
12236 * Various S/MIME bugfixes and compatibility changes:
12237 output correct application/pkcs7 MIME type if
12238 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12239 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12240 of files as .eml work). Correctly handle very long lines in MIME
12245 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12247 * Countermeasure against the Klima-Pokorny-Rosa extension of
12248 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12249 a protocol version number mismatch like a decryption error
12250 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12254 * Turn on RSA blinding by default in the default implementation
12255 to avoid a timing attack. Applications that don't want it can call
12256 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12257 They would be ill-advised to do so in most cases.
12259 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12261 * Change RSA blinding code so that it works when the PRNG is not
12262 seeded (in this case, the secret RSA exponent is abused as
12263 an unpredictable seed -- if it is not unpredictable, there
12264 is no point in blinding anyway). Make RSA blinding thread-safe
12265 by remembering the creator's thread ID in rsa->blinding and
12266 having all other threads use local one-time blinding factors
12267 (this requires more computation than sharing rsa->blinding, but
12268 avoids excessive locking; and if an RSA object is not shared
12269 between threads, blinding will still be very fast).
12273 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12274 ENGINE as defaults for all supported algorithms irrespective of
12275 the 'flags' parameter. 'flags' is now honoured, so applications
12276 should make sure they are passing it correctly.
12280 * Target "mingw" now allows native Windows code to be generated in
12281 the Cygwin environment as well as with the MinGW compiler.
12285 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12287 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12288 via timing by performing a MAC computation even if incorrect
12289 block cipher padding has been found. This is a countermeasure
12290 against active attacks where the attacker has to distinguish
12291 between bad padding and a MAC verification error. ([CVE-2003-0078])
12293 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12294 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12295 Martin Vuagnoux (EPFL, Ilion)*
12297 * Make the no-err option work as intended. The intention with no-err
12298 is not to have the whole error stack handling routines removed from
12299 libcrypto, it's only intended to remove all the function name and
12300 reason texts, thereby removing some of the footprint that may not
12301 be interesting if those errors aren't displayed anyway.
12303 NOTE: it's still possible for any application or module to have its
12304 own set of error texts inserted. The routines are there, just not
12305 used by default when no-err is given.
12309 * Add support for FreeBSD on IA64.
12311 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12313 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12314 Kerberos function mit_des_cbc_cksum(). Before this change,
12315 the value returned by DES_cbc_cksum() was like the one from
12316 mit_des_cbc_cksum(), except the bytes were swapped.
12318 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12320 * Allow an application to disable the automatic SSL chain building.
12321 Before this a rather primitive chain build was always performed in
12322 ssl3_output_cert_chain(): an application had no way to send the
12323 correct chain if the automatic operation produced an incorrect result.
12325 Now the chain builder is disabled if either:
12327 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12329 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12331 The reasoning behind this is that an application would not want the
12332 auto chain building to take place if extra chain certificates are
12333 present and it might also want a means of sending no additional
12334 certificates (for example the chain has two certificates and the
12339 * Add the possibility to build without the ENGINE framework.
12341 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12343 * Under Win32 gmtime() can return NULL: check return value in
12344 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12348 * DSA routines: under certain error conditions uninitialized BN objects
12349 could be freed. Solution: make sure initialization is performed early
12350 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12351 Nils Larsch <nla@trustcenter.de> via PR#459)
12355 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12356 checked on reconnect on the client side, therefore session resumption
12357 could still fail with a "ssl session id is different" error. This
12358 behaviour is masked when SSL_OP_ALL is used due to
12359 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12360 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12361 followup to PR #377.
12365 * IA-32 assembler support enhancements: unified ELF targets, support
12366 for SCO/Caldera platforms, fix for Cygwin shared build.
12370 * Add support for FreeBSD on sparc64. As a consequence, support for
12371 FreeBSD on non-x86 processors is separate from x86 processors on
12372 the config script, much like the NetBSD support.
12374 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12376 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12378 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12381 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12382 code (06) was taken as the first octet of the session ID and the last
12383 octet was ignored consequently. As a result SSLv2 client side session
12384 caching could not have worked due to the session ID mismatch between
12386 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12391 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12392 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12397 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12398 seems that in spite of existing for more than a year, many application
12399 author have done nothing to provide the necessary callbacks, which
12400 means that this particular engine will not work properly anywhere.
12401 This is a very unfortunate situation which forces us, in the name
12402 of usability, to give the hw_ncipher.c a static lock, which is part
12404 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12405 appear in 0.9.8 or later. We EXPECT application authors to have
12406 dealt properly with this when 0.9.8 is released (unless we actually
12407 make such changes in the libcrypto locking code that changes will
12408 have to be made anyway).
12412 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12413 octets have been read, EOF or an error occurs. Without this change
12414 some truncated ASN1 structures will not produce an error.
12418 * Disable Heimdal support, since it hasn't been fully implemented.
12419 Still give the possibility to force the use of Heimdal, but with
12420 warnings and a request that patches get sent to openssl-dev.
12424 * Add the VC-CE target, introduce the WINCE sysname, and add
12425 INSTALL.WCE and appropriate conditionals to make it build.
12427 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12429 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12430 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12431 edit numbers of the version.
12433 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12435 * Introduce safe string copy and catenation functions
12436 (BUF_strlcpy() and BUF_strlcat()).
12438 *Ben Laurie (CHATS) and Richard Levitte*
12440 * Avoid using fixed-size buffers for one-line DNs.
12442 *Ben Laurie (CHATS)*
12444 * Add BUF_MEM_grow_clean() to avoid information leakage when
12445 resizing buffers containing secrets, and use where appropriate.
12447 *Ben Laurie (CHATS)*
12449 * Avoid using fixed size buffers for configuration file location.
12451 *Ben Laurie (CHATS)*
12453 * Avoid filename truncation for various CA files.
12455 *Ben Laurie (CHATS)*
12457 * Use sizeof in preference to magic numbers.
12459 *Ben Laurie (CHATS)*
12461 * Avoid filename truncation in cert requests.
12463 *Ben Laurie (CHATS)*
12465 * Add assertions to check for (supposedly impossible) buffer
12468 *Ben Laurie (CHATS)*
12470 * Don't cache truncated DNS entries in the local cache (this could
12471 potentially lead to a spoofing attack).
12473 *Ben Laurie (CHATS)*
12475 * Fix various buffers to be large enough for hex/decimal
12476 representations in a platform independent manner.
12478 *Ben Laurie (CHATS)*
12480 * Add CRYPTO_realloc_clean() to avoid information leakage when
12481 resizing buffers containing secrets, and use where appropriate.
12483 *Ben Laurie (CHATS)*
12485 * Add BIO_indent() to avoid much slightly worrying code to do
12488 *Ben Laurie (CHATS)*
12490 * Convert sprintf()/BIO_puts() to BIO_printf().
12492 *Ben Laurie (CHATS)*
12494 * buffer_gets() could terminate with the buffer only half
12497 *Ben Laurie (CHATS)*
12499 * Add assertions to prevent user-supplied crypto functions from
12500 overflowing internal buffers by having large block sizes, etc.
12502 *Ben Laurie (CHATS)*
12504 * New OPENSSL_assert() macro (similar to assert(), but enabled
12507 *Ben Laurie (CHATS)*
12509 * Eliminate unused copy of key in RC4.
12511 *Ben Laurie (CHATS)*
12513 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12515 *Ben Laurie (CHATS)*
12517 * Fix off-by-one error in EGD path.
12519 *Ben Laurie (CHATS)*
12521 * If RANDFILE path is too long, ignore instead of truncating.
12523 *Ben Laurie (CHATS)*
12525 * Eliminate unused and incorrectly sized X.509 structure
12528 *Ben Laurie (CHATS)*
12530 * Eliminate unused and dangerous function knumber().
12532 *Ben Laurie (CHATS)*
12534 * Eliminate unused and dangerous structure, KSSL_ERR.
12536 *Ben Laurie (CHATS)*
12538 * Protect against overlong session ID context length in an encoded
12539 session object. Since these are local, this does not appear to be
12542 *Ben Laurie (CHATS)*
12544 * Change from security patch (see 0.9.6e below) that did not affect
12545 the 0.9.6 release series:
12547 Remote buffer overflow in SSL3 protocol - an attacker could
12548 supply an oversized master key in Kerberos-enabled versions.
12551 *Ben Laurie (CHATS)*
12553 * Change the SSL kerb5 codes to match RFC 2712.
12557 * Make -nameopt work fully for req and add -reqopt switch.
12559 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12561 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12563 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12565 * Make sure tests can be performed even if the corresponding algorithms
12566 have been removed entirely. This was also the last step to make
12567 OpenSSL compilable with DJGPP under all reasonable conditions.
12569 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12571 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12572 to allow version independent disabling of normally unselected ciphers,
12573 which may be activated as a side-effect of selecting a single cipher.
12575 (E.g., cipher list string "RSA" enables ciphersuites that are left
12576 out of "ALL" because they do not provide symmetric encryption.
12577 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12579 *Lutz Jaenicke, Bodo Moeller*
12581 * Add appropriate support for separate platform-dependent build
12582 directories. The recommended way to make a platform-dependent
12583 build directory is the following (tested on Linux), maybe with
12586 # Place yourself outside of the OpenSSL source tree. In
12587 # this example, the environment variable OPENSSL_SOURCE
12588 # is assumed to contain the absolute OpenSSL source directory.
12589 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12590 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12591 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12592 mkdir -p `dirname $F`
12593 ln -s $OPENSSL_SOURCE/$F $F
12596 To be absolutely sure not to disturb the source tree, a "make clean"
12597 is a good thing. If it isn't successful, don't worry about it,
12598 it probably means the source directory is very clean.
12602 * Make sure any ENGINE control commands make local copies of string
12603 pointers passed to them whenever necessary. Otherwise it is possible
12604 the caller may have overwritten (or deallocated) the original string
12605 data when a later ENGINE operation tries to use the stored values.
12607 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12609 * Improve diagnostics in file reading and command-line digests.
12611 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12613 * Add AES modes CFB and OFB to the object database. Correct an
12614 error in AES-CFB decryption.
12618 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12619 allows existing EVP_CIPHER_CTX structures to be reused after
12620 calling `EVP_*Final()`. This behaviour is used by encryption
12621 BIOs and some applications. This has the side effect that
12622 applications must explicitly clean up cipher contexts with
12623 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12627 * Check the values of dna and dnb in bn_mul_recursive before calling
12628 bn_mul_comba (a non zero value means the a or b arrays do not contain
12629 n2 elements) and fallback to bn_mul_normal if either is not zero.
12633 * Fix escaping of non-ASCII characters when using the -subj option
12634 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12638 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12639 form for "surname", serialNumber has no short form.
12640 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12641 therefore remove "mail" short name for "internet 7".
12642 The OID for unique identifiers in X509 certificates is
12643 x500UniqueIdentifier, not uniqueIdentifier.
12644 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12648 * Add an "init" command to the ENGINE config module and auto initialize
12649 ENGINEs. Without any "init" command the ENGINE will be initialized
12650 after all ctrl commands have been executed on it. If init=1 the
12651 ENGINE is initialized at that point (ctrls before that point are run
12652 on the uninitialized ENGINE and after on the initialized one). If
12653 init=0 then the ENGINE will not be initialized at all.
12657 * Fix the 'app_verify_callback' interface so that the user-defined
12658 argument is actually passed to the callback: In the
12659 SSL_CTX_set_cert_verify_callback() prototype, the callback
12660 declaration has been changed from
12663 int (*cb)(X509_STORE_CTX *,void *);
12664 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12665 i=s->ctx->app_verify_callback(&ctx)
12666 has been changed into
12667 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12669 To update applications using SSL_CTX_set_cert_verify_callback(),
12670 a dummy argument can be added to their callback functions.
12672 *D. K. Smetters <smetters@parc.xerox.com>*
12674 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12676 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12678 * Add and OPENSSL_LOAD_CONF define which will cause
12679 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12680 This allows older applications to transparently support certain
12681 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12682 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12683 load the config file and OPENSSL_add_all_algorithms_conf() which will
12684 always load it have also been added.
12688 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12689 Adjust NIDs and EVP layer.
12691 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12693 * Config modules support in openssl utility.
12695 Most commands now load modules from the config file,
12696 though in a few (such as version) this isn't done
12697 because it couldn't be used for anything.
12699 In the case of ca and req the config file used is
12700 the same as the utility itself: that is the -config
12701 command line option can be used to specify an
12706 * Move default behaviour from OPENSSL_config(). If appname is NULL
12707 use "openssl_conf" if filename is NULL use default openssl config file.
12711 * Add an argument to OPENSSL_config() to allow the use of an alternative
12712 config section name. Add a new flag to tolerate a missing config file
12713 and move code to CONF_modules_load_file().
12717 * Support for crypto accelerator cards from Accelerated Encryption
12718 Processing, www.aep.ie. (Use engine 'aep')
12719 The support was copied from 0.9.6c [engine] and adapted/corrected
12720 to work with the new engine framework.
12722 *AEP Inc. and Richard Levitte*
12724 * Support for SureWare crypto accelerator cards from Baltimore
12725 Technologies. (Use engine 'sureware')
12726 The support was copied from 0.9.6c [engine] and adapted
12727 to work with the new engine framework.
12731 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12732 make the newer ENGINE framework commands for the CHIL engine work.
12734 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12736 * Make it possible to produce shared libraries on ReliantUNIX.
12738 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12740 * Add the configuration target debug-linux-ppro.
12741 Make 'openssl rsa' use the general key loading routines
12742 implemented in `apps.c`, and make those routines able to
12743 handle the key format FORMAT_NETSCAPE and the variant
12746 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12748 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12750 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12752 * Add -keyform to rsautl, and document -engine.
12754 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12756 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12757 BIO_R_NO_SUCH_FILE error code rather than the generic
12758 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12762 * Add new functions
12763 ERR_peek_last_error
12764 ERR_peek_last_error_line
12765 ERR_peek_last_error_line_data.
12766 These are similar to
12768 ERR_peek_error_line
12769 ERR_peek_error_line_data,
12770 but report on the latest error recorded rather than the first one
12771 still in the error queue.
12773 *Ben Laurie, Bodo Moeller*
12775 * default_algorithms option in ENGINE config module. This allows things
12777 default_algorithms = ALL
12778 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12782 * Preliminary ENGINE config module.
12786 * New experimental application configuration code.
12790 * Change the AES code to follow the same name structure as all other
12791 symmetric ciphers, and behave the same way. Move everything to
12792 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12794 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12796 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12798 *Ben Laurie and Theo de Raadt*
12800 * Add option to output public keys in req command.
12802 *Massimiliano Pala madwolf@openca.org*
12804 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12805 (up to about 10% better than before for P-192 and P-224).
12809 * New functions/macros
12811 SSL_CTX_set_msg_callback(ctx, cb)
12812 SSL_CTX_set_msg_callback_arg(ctx, arg)
12813 SSL_set_msg_callback(ssl, cb)
12814 SSL_set_msg_callback_arg(ssl, arg)
12816 to request calling a callback function
12818 void cb(int write_p, int version, int content_type,
12819 const void *buf, size_t len, SSL *ssl, void *arg)
12821 whenever a protocol message has been completely received
12822 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12823 protocol version according to which the SSL library interprets
12824 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12825 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12826 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12827 specification (change_cipher_spec(20), alert(21), handshake(22)).
12828 'buf' and 'len' point to the actual message, 'ssl' to the
12829 SSL object, and 'arg' is the application-defined value set by
12830 SSL[_CTX]_set_msg_callback_arg().
12832 'openssl s_client' and 'openssl s_server' have new '-msg' options
12833 to enable a callback that displays all protocol messages.
12837 * Change the shared library support so shared libraries are built as
12838 soon as the corresponding static library is finished, and thereby get
12839 openssl and the test programs linked against the shared library.
12840 This still only happens when the keyword "shard" has been given to
12841 the configuration scripts.
12843 NOTE: shared library support is still an experimental thing, and
12844 backward binary compatibility is still not guaranteed.
12846 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12848 * Add support for Subject Information Access extension.
12850 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12852 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12853 additional bytes when new memory had to be allocated, not just
12854 when reusing an existing buffer.
12858 * New command line and configuration option 'utf8' for the req command.
12859 This allows field values to be specified as UTF8 strings.
12863 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12864 runs for the former and machine-readable output for the latter.
12868 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12869 of the e-mail address in the DN (i.e., it will go into a certificate
12870 extension only). The new configuration file option 'email_in_dn = no'
12871 has the same effect.
12873 *Massimiliano Pala madwolf@openca.org*
12875 * Change all functions with names starting with `des_` to be starting
12876 with `DES_` instead. Add wrappers that are compatible with libdes,
12877 but are named `_ossl_old_des_*`. Finally, add macros that map the
12878 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12879 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12880 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12883 Since we provide two compatibility mappings, the user needs to
12884 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12885 compatibility is desired. The default (i.e., when that macro
12886 isn't defined) is OpenSSL 0.9.6c compatibility.
12888 There are also macros that enable and disable the support of old
12889 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12890 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12891 are defined, the default will apply: to support the old des routines.
12893 In either case, one must include openssl/des.h to get the correct
12894 definitions. Do not try to just include openssl/des_old.h, that
12897 NOTE: This is a major break of an old API into a new one. Software
12898 authors are encouraged to switch to the `DES_` style functions. Some
12899 time in the future, des_old.h and the libdes compatibility functions
12900 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12901 default), and then completely removed.
12905 * Test for certificates which contain unsupported critical extensions.
12906 If such a certificate is found during a verify operation it is
12907 rejected by default: this behaviour can be overridden by either
12908 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12909 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12910 X509_supported_extension() has also been added which returns 1 if a
12911 particular extension is supported.
12915 * Modify the behaviour of EVP cipher functions in similar way to digests
12916 to retain compatibility with existing code.
12920 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12921 compatibility with existing code. In particular the 'ctx' parameter does
12922 not have to be to be initialized before the call to EVP_DigestInit() and
12923 it is tidied up after a call to EVP_DigestFinal(). New function
12924 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12925 EVP_MD_CTX_copy() changed to not require the destination to be
12926 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12927 requires the destination to be valid.
12929 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12930 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12934 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12935 so that complete 'Handshake' protocol structures are kept in memory
12936 instead of overwriting 'msg_type' and 'length' with 'body' data.
12940 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12942 *Massimo Santin via Richard Levitte*
12944 * Major restructuring to the underlying ENGINE code. This includes
12945 reduction of linker bloat, separation of pure "ENGINE" manipulation
12946 (initialisation, etc) from functionality dealing with implementations
12947 of specific crypto interfaces. This change also introduces integrated
12948 support for symmetric ciphers and digest implementations - so ENGINEs
12949 can now accelerate these by providing EVP_CIPHER and EVP_MD
12950 implementations of their own. This is detailed in
12951 [crypto/engine/README.md](crypto/engine/README.md)
12952 as it couldn't be adequately described here. However, there are a few
12953 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12954 were changed in the original introduction of ENGINE code have now
12955 reverted back - the hooking from this code to ENGINE is now a good
12956 deal more passive and at run-time, operations deal directly with
12957 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12958 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12959 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12960 they were not being used by the framework as there is no concept of a
12961 BIGNUM_METHOD and they could not be generalised to the new
12962 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12963 ENGINE_cpy() has been removed as it cannot be consistently defined in
12968 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12972 * Change mkdef.pl to sort symbols that get the same entry number,
12973 and make sure the automatically generated functions `ERR_load_*`
12974 become part of libeay.num as well.
12978 * New function SSL_renegotiate_pending(). This returns true once
12979 renegotiation has been requested (either SSL_renegotiate() call
12980 or HelloRequest/ClientHello received from the peer) and becomes
12981 false once a handshake has been completed.
12982 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12983 sends a HelloRequest, but does not ensure that a handshake takes
12984 place. SSL_renegotiate_pending() is useful for checking if the
12985 client has followed the request.)
12989 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12990 By default, clients may request session resumption even during
12991 renegotiation (if session ID contexts permit); with this option,
12992 session resumption is possible only in the first handshake.
12994 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12995 more bits available for options that should not be part of
12996 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13000 * Add some demos for certificate and certificate request creation.
13004 * Make maximum certificate chain size accepted from the peer application
13005 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13006 "Douglas E. Engert" <deengert@anl.gov>.
13010 * Add support for shared libraries for Unixware-7
13011 (Boyd Lynn Gerber <gerberb@zenez.com>).
13015 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13016 be done prior to destruction. Use this to unload error strings from
13017 ENGINEs that load their own error strings. NB: This adds two new API
13018 functions to "get" and "set" this destroy handler in an ENGINE.
13022 * Alter all existing ENGINE implementations (except "openssl" and
13023 "openbsd") to dynamically instantiate their own error strings. This
13024 makes them more flexible to be built both as statically-linked ENGINEs
13025 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13026 Also, add stub code to each that makes building them as self-contained
13027 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13031 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13032 implementations into applications that are completely implemented in
13033 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13034 commands that can be used to configure what shared-library to load and
13035 to control aspects of the way it is handled. Also, made an update to
13036 the [README-Engine.md](README-Engine.md) file
13037 that brings its information up-to-date and
13038 provides some information and instructions on the "dynamic" ENGINE
13039 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13043 * Make it possible to unload ranges of ERR strings with a new
13044 "ERR_unload_strings" function.
13048 * Add a copy() function to EVP_MD.
13052 * Make EVP_MD routines take a context pointer instead of just the
13053 md_data void pointer.
13057 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13058 that the digest can only process a single chunk of data
13059 (typically because it is provided by a piece of
13060 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13061 is only going to provide a single chunk of data, and hence the
13062 framework needn't accumulate the data for oneshot drivers.
13066 * As with "ERR", make it possible to replace the underlying "ex_data"
13067 functions. This change also alters the storage and management of global
13068 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13069 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13070 index counters. The API functions that use this state have been changed
13071 to take a "class_index" rather than pointers to the class's local STACK
13072 and counter, and there is now an API function to dynamically create new
13073 classes. This centralisation allows us to (a) plug a lot of the
13074 thread-safety problems that existed, and (b) makes it possible to clean
13075 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13076 such data would previously have always leaked in application code and
13077 workarounds were in place to make the memory debugging turn a blind eye
13078 to it. Application code that doesn't use this new function will still
13079 leak as before, but their memory debugging output will announce it now
13080 rather than letting it slide.
13082 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13083 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13084 has a return value to indicate success or failure.
13088 * Make it possible to replace the underlying "ERR" functions such that the
13089 global state (2 LHASH tables and 2 locks) is only used by the "default"
13090 implementation. This change also adds two functions to "get" and "set"
13091 the implementation prior to it being automatically set the first time
13092 any other ERR function takes place. Ie. an application can call "get",
13093 pass the return value to a module it has just loaded, and that module
13094 can call its own "set" function using that value. This means the
13095 module's "ERR" operations will use (and modify) the error state in the
13096 application and not in its own statically linked copy of OpenSSL code.
13100 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13101 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13102 the operation, and provides a more encapsulated way for external code
13103 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13104 to use these functions rather than manually incrementing the counts.
13106 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13110 * Add EVP test program.
13114 * Add symmetric cipher support to ENGINE. Expect the API to change!
13118 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13119 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13120 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13121 These allow a CRL to be built without having to access X509_CRL fields
13122 directly. Modify 'ca' application to use new functions.
13126 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13127 bug workarounds. Rollback attack detection is a security feature.
13128 The problem will only arise on OpenSSL servers when TLSv1 is not
13129 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13130 Software authors not wanting to support TLSv1 will have special reasons
13131 for their choice and can explicitly enable this option.
13133 *Bodo Moeller, Lutz Jaenicke*
13135 * Rationalise EVP so it can be extended: don't include a union of
13136 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13137 (similar to those existing for EVP_CIPHER_CTX).
13142 EVP_MD_CTX_init(&md); /* new function call */
13143 EVP_DigestInit(&md, EVP_sha1());
13144 EVP_DigestUpdate(&md, in, len);
13145 EVP_DigestFinal(&md, out, NULL);
13146 EVP_MD_CTX_cleanup(&md); /* new function call */
13150 * Make DES key schedule conform to the usual scheme, as well as
13151 correcting its structure. This means that calls to DES functions
13152 now have to pass a pointer to a des_key_schedule instead of a
13153 plain des_key_schedule (which was actually always a pointer
13156 des_key_schedule ks;
13158 des_set_key_checked(..., &ks);
13159 des_ncbc_encrypt(..., &ks, ...);
13161 (Note that a later change renames 'des_...' into 'DES_...'.)
13165 * Initial reduction of linker bloat: the use of some functions, such as
13166 PEM causes large amounts of unused functions to be linked in due to
13167 poor organisation. For example pem_all.c contains every PEM function
13168 which has a knock on effect of linking in large amounts of (unused)
13169 ASN1 code. Grouping together similar functions and splitting unrelated
13170 functions prevents this.
13174 * Cleanup of EVP macros.
13178 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13179 correct `_ecb suffix`.
13183 * Add initial OCSP responder support to ocsp application. The
13184 revocation information is handled using the text based index
13185 use by the ca application. The responder can either handle
13186 requests generated internally, supplied in files (for example
13187 via a CGI script) or using an internal minimal server.
13191 * Add configuration choices to get zlib compression for TLS.
13195 * Changes to Kerberos SSL for RFC 2712 compliance:
13196 1. Implemented real KerberosWrapper, instead of just using
13197 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13198 2. Implemented optional authenticator field of KerberosWrapper.
13200 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13201 and authenticator structs; see crypto/krb5/.
13203 Generalized Kerberos calls to support multiple Kerberos libraries.
13204 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13205 via Richard Levitte*
13207 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13208 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13209 values for each of the key sizes rather than having just
13210 parameters (and 'speed' generating keys each time).
13214 * Speed up EVP routines.
13217 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13218 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13219 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13220 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13222 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13223 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13224 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13227 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13229 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13233 * Added the OS2-EMX target.
13235 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13237 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13238 New functions to support `NCONF` routines in extension code.
13239 New function `CONF_set_nconf()`
13240 to allow functions which take an `NCONF` to also handle the old `LHASH`
13241 structure: this means that the old `CONF` compatible routines can be
13242 retained (in particular w.rt. extensions) without having to duplicate the
13243 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13247 * Enhance the general user interface with mechanisms for inner control
13248 and with possibilities to have yes/no kind of prompts.
13252 * Change all calls to low-level digest routines in the library and
13253 applications to use EVP. Add missing calls to HMAC_cleanup() and
13254 don't assume HMAC_CTX can be copied using memcpy().
13256 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13258 * Add the possibility to control engines through control names but with
13259 arbitrary arguments instead of just a string.
13260 Change the key loaders to take a UI_METHOD instead of a callback
13261 function pointer. NOTE: this breaks binary compatibility with earlier
13262 versions of OpenSSL [engine].
13263 Adapt the nCipher code for these new conditions and add a card insertion
13268 * Enhance the general user interface with mechanisms to better support
13269 dialog box interfaces, application-defined prompts, the possibility
13270 to use defaults (for example default passwords from somewhere else)
13271 and interrupts/cancellations.
13275 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13276 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13280 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13281 tidy up some unnecessarily weird code in 'sk_new()').
13283 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13285 * Change the key loading routines for ENGINEs to use the same kind
13286 callback (pem_password_cb) as all other routines that need this
13291 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13292 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13293 than this minimum value is recommended.
13297 * New random seeder for OpenVMS, using the system process statistics
13298 that are easily reachable.
13302 * Windows apparently can't transparently handle global
13303 variables defined in DLLs. Initialisations such as:
13305 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13307 won't compile. This is used by the any applications that need to
13308 declare their own ASN1 modules. This was fixed by adding the option
13309 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13310 needed for static libraries under Win32.
13314 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13315 setting of purpose and trust fields. New X509_STORE trust and
13316 purpose functions and tidy up setting in other SSL functions.
13320 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13321 structure. These are inherited by X509_STORE_CTX when it is
13322 initialised. This allows various defaults to be set in the
13323 X509_STORE structure (such as flags for CRL checking and custom
13324 purpose or trust settings) for functions which only use X509_STORE_CTX
13325 internally such as S/MIME.
13327 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13328 trust settings if they are not set in X509_STORE. This allows X509_STORE
13329 purposes and trust (in S/MIME for example) to override any set by default.
13331 Add command line options for CRL checking to smime, s_client and s_server
13336 * Initial CRL based revocation checking. If the CRL checking flag(s)
13337 are set then the CRL is looked up in the X509_STORE structure and
13338 its validity and signature checked, then if the certificate is found
13339 in the CRL the verify fails with a revoked error.
13341 Various new CRL related callbacks added to X509_STORE_CTX structure.
13343 Command line options added to 'verify' application to support this.
13345 This needs some additional work, such as being able to handle multiple
13346 CRLs with different times, extension based lookup (rather than just
13347 by subject name) and ultimately more complete V2 CRL extension
13352 * Add a general user interface API (crypto/ui/). This is designed
13353 to replace things like des_read_password and friends (backward
13354 compatibility functions using this new API are provided).
13355 The purpose is to remove prompting functions from the DES code
13356 section as well as provide for prompting through dialog boxes in
13357 a window system and the like.
13361 * Add "ex_data" support to ENGINE so implementations can add state at a
13362 per-structure level rather than having to store it globally.
13366 * Make it possible for ENGINE structures to be copied when retrieved by
13367 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13368 This causes the "original" ENGINE structure to act like a template,
13369 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13370 operational state can be localised to each ENGINE structure, despite the
13371 fact they all share the same "methods". New ENGINE structures returned in
13372 this case have no functional references and the return value is the single
13373 structural reference. This matches the single structural reference returned
13374 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13379 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13380 needs to match any other type at all we need to manually clear the
13385 * Changes to the "openssl engine" utility to include;
13386 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13387 about an ENGINE's available control commands.
13388 - executing control commands from command line arguments using the
13389 '-pre' and '-post' switches. '-post' is only used if '-t' is
13390 specified and the ENGINE is successfully initialised. The syntax for
13391 the individual commands are colon-separated, for example;
13392 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13396 * New dynamic control command support for ENGINEs. ENGINEs can now
13397 declare their own commands (numbers), names (strings), descriptions,
13398 and input types for run-time discovery by calling applications. A
13399 subset of these commands are implicitly classed as "executable"
13400 depending on their input type, and only these can be invoked through
13401 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13402 can be based on user input, config files, etc). The distinction is
13403 that "executable" commands cannot return anything other than a boolean
13404 result and can only support numeric or string input, whereas some
13405 discoverable commands may only be for direct use through
13406 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13407 pointers, or other custom uses. The "executable" commands are to
13408 support parameterisations of ENGINE behaviour that can be
13409 unambiguously defined by ENGINEs and used consistently across any
13410 OpenSSL-based application. Commands have been added to all the
13411 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13412 control over shared-library paths without source code alterations.
13416 * Changed all ENGINE implementations to dynamically allocate their
13417 ENGINEs rather than declaring them statically. Apart from this being
13418 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13419 this also allows the implementations to compile without using the
13420 internal engine_int.h header.
13424 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13425 'const' value. Any code that should be able to modify a RAND_METHOD
13426 should already have non-const pointers to it (ie. they should only
13427 modify their own ones).
13431 * Made a variety of little tweaks to the ENGINE code.
13432 - "atalla" and "ubsec" string definitions were moved from header files
13433 to C code. "nuron" string definitions were placed in variables
13434 rather than hard-coded - allowing parameterisation of these values
13435 later on via ctrl() commands.
13436 - Removed unused "#if 0"'d code.
13437 - Fixed engine list iteration code so it uses ENGINE_free() to release
13438 structural references.
13439 - Constified the RAND_METHOD element of ENGINE structures.
13440 - Constified various get/set functions as appropriate and added
13441 missing functions (including a catch-all ENGINE_cpy that duplicates
13442 all ENGINE values onto a new ENGINE except reference counts/state).
13443 - Removed NULL parameter checks in get/set functions. Setting a method
13444 or function to NULL is a way of cancelling out a previously set
13445 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13446 and doesn't justify the extra error symbols and code.
13447 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13448 flags from engine_int.h to engine.h.
13449 - Changed prototypes for ENGINE handler functions (init(), finish(),
13450 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13454 * Implement binary inversion algorithm for BN_mod_inverse in addition
13455 to the algorithm using long division. The binary algorithm can be
13456 used only if the modulus is odd. On 32-bit systems, it is faster
13457 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13458 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13459 up to 450 bits. In 64-bit environments, the binary algorithm
13460 appears to be advantageous for much longer moduli; here we use it
13461 for moduli up to 2048 bits.
13465 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13466 could not support the combine flag in choice fields.
13470 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13471 extensions from a certificate request to the certificate.
13475 * Allow multiple 'certopt' and 'nameopt' options to be separated
13476 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13477 file: this allows the display of the certificate about to be
13478 signed to be customised, to allow certain fields to be included
13479 or excluded and extension details. The old system didn't display
13480 multicharacter strings properly, omitted fields not in the policy
13481 and couldn't display additional details such as extensions.
13485 * Function EC_POINTs_mul for multiple scalar multiplication
13486 of an arbitrary number of elliptic curve points
13487 \sum scalars[i]*points[i],
13488 optionally including the generator defined for the EC_GROUP:
13489 scalar*generator + \sum scalars[i]*points[i].
13491 EC_POINT_mul is a simple wrapper function for the typical case
13492 that the point list has just one item (besides the optional
13497 * First EC_METHODs for curves over GF(p):
13499 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13500 operations and provides various method functions that can also
13501 operate with faster implementations of modular arithmetic.
13503 EC_GFp_mont_method() reuses most functions that are part of
13504 EC_GFp_simple_method, but uses Montgomery arithmetic.
13506 *Bodo Moeller; point addition and point doubling
13507 implementation directly derived from source code provided by
13508 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13510 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13511 crypto/ec/ec_lib.c):
13513 Curves are EC_GROUP objects (with an optional group generator)
13514 based on EC_METHODs that are built into the library.
13516 Points are EC_POINT objects based on EC_GROUP objects.
13518 Most of the framework would be able to handle curves over arbitrary
13519 finite fields, but as there are no obvious types for fields other
13520 than GF(p), some functions are limited to that for now.
13524 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13525 that the file contains a complete HTTP response.
13529 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13530 change the def and num file printf format specifier from "%-40sXXX"
13531 to "%-39s XXX". The latter will always guarantee a space after the
13532 field while the former will cause them to run together if the field
13533 is 40 of more characters long.
13537 * Constify the cipher and digest 'method' functions and structures
13538 and modify related functions to take constant EVP_MD and EVP_CIPHER
13543 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13544 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13548 * Modify `EVP_Digest*()` routines so they now return values. Although the
13549 internal software routines can never fail additional hardware versions
13554 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13556 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13557 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13560 ERR_R_NESTED_ASN1_ERROR
13562 ERR_R_MISSING_ASN1_EOS
13563 were 4 .. 9, conflicting with
13564 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13566 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13567 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13569 Add new error code 'ERR_R_INTERNAL_ERROR'.
13573 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13578 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13579 sets the subject name for a new request or supersedes the
13580 subject name in a given request. Formats that can be parsed are
13581 'CN=Some Name, OU=myOU, C=IT'
13583 'CN=Some Name/OU=myOU/C=IT'.
13585 Add options '-batch' and '-verbose' to 'openssl req'.
13587 *Massimiliano Pala <madwolf@hackmasters.net>*
13589 * Introduce the possibility to access global variables through
13590 functions on platform were that's the best way to handle exporting
13591 global variables in shared libraries. To enable this functionality,
13592 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13593 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13594 is normally done by Configure or something similar).
13596 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13597 in the source file (foo.c) like this:
13599 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13600 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13602 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13603 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13605 OPENSSL_DECLARE_GLOBAL(int,foo);
13606 #define foo OPENSSL_GLOBAL_REF(foo)
13607 OPENSSL_DECLARE_GLOBAL(double,bar);
13608 #define bar OPENSSL_GLOBAL_REF(bar)
13610 The #defines are very important, and therefore so is including the
13611 header file everywhere where the defined globals are used.
13613 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13614 of ASN.1 items, but that structure is a bit different.
13616 The largest change is in util/mkdef.pl which has been enhanced with
13617 better and easier to understand logic to choose which symbols should
13618 go into the Windows .def files as well as a number of fixes and code
13619 cleanup (among others, algorithm keywords are now sorted
13620 lexicographically to avoid constant rewrites).
13624 * In BN_div() keep a copy of the sign of 'num' before writing the
13625 result to 'rm' because if rm==num the value will be overwritten
13626 and produce the wrong result if 'num' is negative: this caused
13627 problems with BN_mod() and BN_nnmod().
13631 * Function OCSP_request_verify(). This checks the signature on an
13632 OCSP request and verifies the signer certificate. The signer
13633 certificate is just checked for a generic purpose and OCSP request
13638 * Add OCSP_check_validity() function to check the validity of OCSP
13639 responses. OCSP responses are prepared in real time and may only
13640 be a few seconds old. Simply checking that the current time lies
13641 between thisUpdate and nextUpdate max reject otherwise valid responses
13642 caused by either OCSP responder or client clock inaccuracy. Instead
13643 we allow thisUpdate and nextUpdate to fall within a certain period of
13644 the current time. The age of the response can also optionally be
13645 checked. Two new options -validity_period and -status_age added to
13650 * If signature or public key algorithm is unrecognized print out its
13651 OID rather that just UNKNOWN.
13655 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13656 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13657 ID to be generated from the issuer certificate alone which can then be
13658 passed to OCSP_id_issuer_cmp().
13662 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13663 ASN1 modules to export functions returning ASN1_ITEM pointers
13664 instead of the ASN1_ITEM structures themselves. This adds several
13665 new macros which allow the underlying ASN1 function/structure to
13666 be accessed transparently. As a result code should not use ASN1_ITEM
13667 references directly (such as &X509_it) but instead use the relevant
13668 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13669 use of the new ASN1 code on platforms where exporting structures
13670 is problematical (for example in shared libraries) but exporting
13671 functions returning pointers to structures is not.
13675 * Add support for overriding the generation of SSL/TLS session IDs.
13676 These callbacks can be registered either in an SSL_CTX or per SSL.
13677 The purpose of this is to allow applications to control, if they wish,
13678 the arbitrary values chosen for use as session IDs, particularly as it
13679 can be useful for session caching in multiple-server environments. A
13680 command-line switch for testing this (and any client code that wishes
13681 to use such a feature) has been added to "s_server".
13683 *Geoff Thorpe, Lutz Jaenicke*
13685 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13686 of the form `#if defined(...) || defined(...) || ...` and
13687 `#if !defined(...) && !defined(...) && ...`. This also avoids
13688 the growing number of special cases it was previously handling.
13692 * Make all configuration macros available for application by making
13693 sure they are available in opensslconf.h, by giving them names starting
13694 with `OPENSSL_` to avoid conflicts with other packages and by making
13695 sure e_os2.h will cover all platform-specific cases together with
13697 Additionally, it is now possible to define configuration/platform-
13698 specific names (called "system identities"). In the C code, these
13699 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13700 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13701 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13706 * New option -set_serial to 'req' and 'x509' this allows the serial
13707 number to use to be specified on the command line. Previously self
13708 signed certificates were hard coded with serial number 0 and the
13709 CA options of 'x509' had to use a serial number in a file which was
13714 * New options to 'ca' utility to support V2 CRL entry extensions.
13715 Currently CRL reason, invalidity date and hold instruction are
13716 supported. Add new CRL extensions to V3 code and some new objects.
13720 * New function EVP_CIPHER_CTX_set_padding() this is used to
13721 disable standard block padding (aka PKCS#5 padding) in the EVP
13722 API, which was previously mandatory. This means that the data is
13723 not padded in any way and so the total length much be a multiple
13724 of the block size, otherwise an error occurs.
13728 * Initial (incomplete) OCSP SSL support.
13732 * New function OCSP_parse_url(). This splits up a URL into its host,
13733 port and path components: primarily to parse OCSP URLs. New -url
13734 option to ocsp utility.
13738 * New nonce behavior. The return value of OCSP_check_nonce() now
13739 reflects the various checks performed. Applications can decide
13740 whether to tolerate certain situations such as an absent nonce
13741 in a response when one was present in a request: the ocsp application
13742 just prints out a warning. New function OCSP_add1_basic_nonce()
13743 this is to allow responders to include a nonce in a response even if
13744 the request is nonce-less.
13748 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13749 skipped when using openssl x509 multiple times on a single input file,
13750 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13754 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13755 set string type: to handle setting ASN1_TIME structures. Fix ca
13756 utility to correctly initialize revocation date of CRLs.
13760 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13761 the clients preferred ciphersuites and rather use its own preferences.
13762 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13763 Internet Explorer by ensuring unchanged hash method during stepup.
13764 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13768 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13769 to aes and add a new 'exist' option to print out symbols that don't
13774 * Additional options to ocsp utility to allow flags to be set and
13775 additional certificates supplied.
13779 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13780 OCSP client a number of certificate to only verify the response
13785 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13786 handle the new API. Currently only ECB, CBC modes supported. Add new
13789 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13790 Encryption Standard (AES) Ciphersuites for Transport Layer
13791 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13792 not enabled by default and were not part of the "ALL" ciphersuite
13793 alias because they were not yet official; they could be
13794 explicitly requested by specifying the "AESdraft" ciphersuite
13795 group alias. In the final release of OpenSSL 0.9.7, the group
13796 alias is called "AES" and is part of "ALL".)
13798 *Ben Laurie, Steve Henson, Bodo Moeller*
13800 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13801 request to response.
13805 * Functions for OCSP responders. OCSP_request_onereq_count(),
13806 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13807 extract information from a certificate request. OCSP_response_create()
13808 creates a response and optionally adds a basic response structure.
13809 OCSP_basic_add1_status() adds a complete single response to a basic
13810 response and returns the OCSP_SINGLERESP structure just added (to allow
13811 extensions to be included for example). OCSP_basic_add1_cert() adds a
13812 certificate to a basic response and OCSP_basic_sign() signs a basic
13813 response with various flags. New helper functions ASN1_TIME_check()
13814 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13815 (converts ASN1_TIME to GeneralizedTime).
13819 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13820 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13821 structure from a certificate. X509_pubkey_digest() digests the public_key
13822 contents: this is used in various key identifiers.
13826 * Make sk_sort() tolerate a NULL argument.
13828 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13830 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13831 passed by the function are trusted implicitly. If any of them signed the
13832 response then it is assumed to be valid and is not verified.
13836 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13837 to data. This was previously part of the PKCS7 ASN1 code. This
13838 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13839 *Steve Henson, reported by Kenneth R. Robinette
13840 <support@securenetterm.com>*
13842 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13843 routines: without these tracing memory leaks is very painful.
13844 Fix leaks in PKCS12 and PKCS7 routines.
13848 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13849 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13850 effectively meant GeneralizedTime would never be used. Now it
13851 is initialised to -1 but X509_time_adj() now has to check the value
13852 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13853 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13854 *Steve Henson, reported by Kenneth R. Robinette
13855 <support@securenetterm.com>*
13857 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13858 result in a zero length in the ASN1_INTEGER structure which was
13859 not consistent with the structure when d2i_ASN1_INTEGER() was used
13860 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13861 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13862 where it did not print out a minus for negative ASN1_INTEGER.
13866 * Add summary printout to ocsp utility. The various functions which
13867 convert status values to strings have been renamed to:
13868 OCSP_response_status_str(), OCSP_cert_status_str() and
13869 OCSP_crl_reason_str() and are no longer static. New options
13870 to verify nonce values and to disable verification. OCSP response
13871 printout format cleaned up.
13875 * Add additional OCSP certificate checks. These are those specified
13876 in RFC2560. This consists of two separate checks: the CA of the
13877 certificate being checked must either be the OCSP signer certificate
13878 or the issuer of the OCSP signer certificate. In the latter case the
13879 OCSP signer certificate must contain the OCSP signing extended key
13880 usage. This check is performed by attempting to match the OCSP
13881 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13882 in the OCSP_CERTID structures of the response.
13886 * Initial OCSP certificate verification added to OCSP_basic_verify()
13887 and related routines. This uses the standard OpenSSL certificate
13888 verify routines to perform initial checks (just CA validity) and
13889 to obtain the certificate chain. Then additional checks will be
13890 performed on the chain. Currently the root CA is checked to see
13891 if it is explicitly trusted for OCSP signing. This is used to set
13892 a root CA as a global signing root: that is any certificate that
13893 chains to that CA is an acceptable OCSP signing certificate.
13897 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13898 extensions from a separate configuration file.
13899 As when reading extensions from the main configuration file,
13900 the '-extensions ...' option may be used for specifying the
13903 *Massimiliano Pala <madwolf@comune.modena.it>*
13905 * New OCSP utility. Allows OCSP requests to be generated or
13906 read. The request can be sent to a responder and the output
13907 parsed, outputted or printed in text form. Not complete yet:
13908 still needs to check the OCSP response validity.
13912 * New subcommands for 'openssl ca':
13913 `openssl ca -status <serial>` prints the status of the cert with
13914 the given serial number (according to the index file).
13915 `openssl ca -updatedb` updates the expiry status of certificates
13918 *Massimiliano Pala <madwolf@comune.modena.it>*
13920 * New '-newreq-nodes' command option to CA.pl. This is like
13921 '-newreq', but calls 'openssl req' with the '-nodes' option
13922 so that the resulting key is not encrypted.
13924 *Damien Miller <djm@mindrot.org>*
13926 * New configuration for the GNU Hurd.
13928 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13930 * Initial code to implement OCSP basic response verify. This
13931 is currently incomplete. Currently just finds the signer's
13932 certificate and verifies the signature on the response.
13936 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13937 value of OPENSSLDIR. This is available via the new '-d' option
13938 to 'openssl version', and is also included in 'openssl version -a'.
13942 * Allowing defining memory allocation callbacks that will be given
13943 file name and line number information in additional arguments
13944 (a `const char*` and an int). The basic functionality remains, as
13945 well as the original possibility to just replace malloc(),
13946 realloc() and free() by functions that do not know about these
13947 additional arguments. To register and find out the current
13948 settings for extended allocation functions, the following
13949 functions are provided:
13951 CRYPTO_set_mem_ex_functions
13952 CRYPTO_set_locked_mem_ex_functions
13953 CRYPTO_get_mem_ex_functions
13954 CRYPTO_get_locked_mem_ex_functions
13956 These work the same way as CRYPTO_set_mem_functions and friends.
13957 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13958 extended allocation function is enabled.
13959 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13960 a conventional allocation function is enabled.
13962 *Richard Levitte, Bodo Moeller*
13964 * Finish off removing the remaining LHASH function pointer casts.
13965 There should no longer be any prototype-casting required when using
13966 the LHASH abstraction, and any casts that remain are "bugs". See
13967 the callback types and macros at the head of lhash.h for details
13968 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13972 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13973 If /dev/[u]random devices are not available or do not return enough
13974 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13976 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13977 /etc/entropy will be queried once each in this sequence, querying stops
13978 when enough entropy was collected without querying more sockets.
13982 * Change the Unix RAND_poll() variant to be able to poll several
13983 random devices, as specified by DEVRANDOM, until a sufficient amount
13984 of data has been collected. We spend at most 10 ms on each file
13985 (select timeout) and read in non-blocking mode. DEVRANDOM now
13986 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13987 (previously it was just the string "/dev/urandom"), so on typical
13988 platforms the 10 ms delay will never occur.
13989 Also separate out the Unix variant to its own file, rand_unix.c.
13990 For VMS, there's a currently-empty rand_vms.c.
13994 * Move OCSP client related routines to ocsp_cl.c. These
13995 provide utility functions which an application needing
13996 to issue a request to an OCSP responder and analyse the
13997 response will typically need: as opposed to those which an
13998 OCSP responder itself would need which will be added later.
14000 OCSP_request_sign() signs an OCSP request with an API similar
14001 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14002 response. OCSP_response_get1_basic() extracts basic response
14003 from response. OCSP_resp_find_status(): finds and extracts status
14004 information from an OCSP_CERTID structure (which will be created
14005 when the request structure is built). These are built from lower
14006 level functions which work on OCSP_SINGLERESP structures but
14007 won't normally be used unless the application wishes to examine
14008 extensions in the OCSP response for example.
14010 Replace nonce routines with a pair of functions.
14011 OCSP_request_add1_nonce() adds a nonce value and optionally
14012 generates a random value. OCSP_check_nonce() checks the
14013 validity of the nonce in an OCSP response.
14017 * Change function OCSP_request_add() to OCSP_request_add0_id().
14018 This doesn't copy the supplied OCSP_CERTID and avoids the
14019 need to free up the newly created id. Change return type
14020 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14021 This can then be used to add extensions to the request.
14022 Deleted OCSP_request_new(), since most of its functionality
14023 is now in OCSP_REQUEST_new() (and the case insensitive name
14024 clash) apart from the ability to set the request name which
14025 will be added elsewhere.
14029 * Update OCSP API. Remove obsolete extensions argument from
14030 various functions. Extensions are now handled using the new
14031 OCSP extension code. New simple OCSP HTTP function which
14032 can be used to send requests and parse the response.
14036 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14037 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14038 uses the special reorder version of SET OF to sort the attributes
14039 and reorder them to match the encoded order. This resolves a long
14040 standing problem: a verify on a PKCS7 structure just after signing
14041 it used to fail because the attribute order did not match the
14042 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14043 it uses the received order. This is necessary to tolerate some broken
14044 software that does not order SET OF. This is handled by encoding
14045 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14046 to produce the required SET OF.
14050 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14051 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14052 files to get correct declarations of the ASN.1 item variables.
14056 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14057 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14058 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14059 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14060 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14061 ASN1_ITEM and no wrapper functions.
14065 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14066 replace the old function pointer based I/O routines. Change most of
14067 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14071 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14072 lines, recognize more "algorithms" that can be deselected, and make
14073 it complain about algorithm deselection that isn't recognised.
14077 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14078 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14079 to use new functions. Add NO_ASN1_OLD which can be set to remove
14080 some old style ASN1 functions: this can be used to determine if old
14081 code will still work when these eventually go away.
14085 * New extension functions for OCSP structures, these follow the
14086 same conventions as certificates and CRLs.
14090 * New function X509V3_add1_i2d(). This automatically encodes and
14091 adds an extension. Its behaviour can be customised with various
14092 flags to append, replace or delete. Various wrappers added for
14093 certificates and CRLs.
14097 * Fix to avoid calling the underlying ASN1 print routine when
14098 an extension cannot be parsed. Correct a typo in the
14099 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14103 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14104 entries for variables.
14108 * Add functionality to `apps/openssl.c` for detecting locking
14109 problems: As the program is single-threaded, all we have
14110 to do is register a locking callback using an array for
14111 storing which locks are currently held by the program.
14115 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14116 SSL_get_ex_data_X509_STORE_idx(), which is used in
14117 ssl_verify_cert_chain() and thus can be called at any time
14118 during TLS/SSL handshakes so that thread-safety is essential.
14119 Unfortunately, the ex_data design is not at all suited
14120 for multi-threaded use, so it probably should be abolished.
14124 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14126 *Broadcom, tweaked and integrated by Geoff Thorpe*
14128 * Move common extension printing code to new function
14129 X509V3_print_extensions(). Reorganise OCSP print routines and
14130 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14134 * New function X509_signature_print() to remove duplication in some
14139 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14140 set (this was treated exactly the same as SET OF previously). This
14141 is used to reorder the STACK representing the structure to match the
14142 encoding. This will be used to get round a problem where a PKCS7
14143 structure which was signed could not be verified because the STACK
14144 order did not reflect the encoded order.
14148 * Reimplement the OCSP ASN1 module using the new code.
14152 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14153 for its ASN1 operations. The old style function pointers still exist
14154 for now but they will eventually go away.
14158 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14159 completely replaces the old ASN1 functionality with a table driven
14160 encoder and decoder which interprets an ASN1_ITEM structure describing
14161 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14162 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14163 has also been converted to the new form.
14167 * Change BN_mod_exp_recp so that negative moduli are tolerated
14168 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14169 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14170 for negative moduli.
14174 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14175 of not touching the result's sign bit.
14179 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14184 * Changed the LHASH code to use prototypes for callbacks, and created
14185 macros to declare and implement thin (optionally static) functions
14186 that provide type-safety and avoid function pointer casting for the
14187 type-specific callbacks.
14191 * Added Kerberos Cipher Suites to be used with TLS, as written in
14193 *Veers Staats <staatsvr@asc.hpc.mil>,
14194 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14196 * Reformat the FAQ so the different questions and answers can be divided
14197 in sections depending on the subject.
14201 * Have the zlib compression code load ZLIB.DLL dynamically under
14206 * New function BN_mod_sqrt for computing square roots modulo a prime
14207 (using the probabilistic Tonelli-Shanks algorithm unless
14208 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14209 be handled deterministically).
14211 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14213 * Make BN_mod_inverse faster by explicitly handling small quotients
14214 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14215 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14219 * New function BN_kronecker.
14223 * Fix BN_gcd so that it works on negative inputs; the result is
14224 positive unless both parameters are zero.
14225 Previously something reasonably close to an infinite loop was
14226 possible because numbers could be growing instead of shrinking
14227 in the implementation of Euclid's algorithm.
14231 * Fix BN_is_word() and BN_is_one() macros to take into account the
14232 sign of the number in question.
14234 Fix BN_is_word(a,w) to work correctly for w == 0.
14236 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14237 because its test if the absolute value of 'a' equals 'w'.
14238 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14239 it exists mostly for use in the implementations of BN_is_zero(),
14240 BN_is_one(), and BN_is_word().
14244 * New function BN_swap.
14248 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14249 the exponentiation functions are more likely to produce reasonable
14250 results on negative inputs.
14254 * Change BN_mod_mul so that the result is always non-negative.
14255 Previously, it could be negative if one of the factors was negative;
14256 I don't think anyone really wanted that behaviour.
14260 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14261 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14262 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14263 and add new functions:
14272 BN_mod_lshift1_quick
14274 BN_mod_lshift_quick
14276 These functions always generate non-negative results.
14278 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14279 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14281 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14282 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14283 be reduced modulo `m`.
14285 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14288 The following entry accidentally appeared in the CHANGES file
14289 distributed with OpenSSL 0.9.7. The modifications described in
14290 it do *not* apply to OpenSSL 0.9.7.
14292 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14293 was actually never needed) and in BN_mul(). The removal in BN_mul()
14294 required a small change in bn_mul_part_recursive() and the addition
14295 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14296 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14297 bn_sub_words() and bn_add_words() except they take arrays with
14303 * In 'openssl passwd', verify passwords read from the terminal
14304 unless the '-salt' option is used (which usually means that
14305 verification would just waste user's time since the resulting
14306 hash is going to be compared with some given password hash)
14307 or the new '-noverify' option is used.
14309 This is an incompatible change, but it does not affect
14310 non-interactive use of 'openssl passwd' (passwords on the command
14311 line, '-stdin' option, '-in ...' option) and thus should not
14312 cause any problems.
14316 * Remove all references to RSAref, since there's no more need for it.
14320 * Make DSO load along a path given through an environment variable
14321 (SHLIB_PATH) with shl_load().
14325 * Constify the ENGINE code as a result of BIGNUM constification.
14326 Also constify the RSA code and most things related to it. In a
14327 few places, most notable in the depth of the ASN.1 code, ugly
14328 casts back to non-const were required (to be solved at a later
14333 * Make it so the openssl application has all engines loaded by default.
14337 * Constify the BIGNUM routines a little more.
14341 * Add the following functions:
14343 ENGINE_load_cswift()
14345 ENGINE_load_atalla()
14346 ENGINE_load_nuron()
14347 ENGINE_load_builtin_engines()
14349 That way, an application can itself choose if external engines that
14350 are built-in in OpenSSL shall ever be used or not. The benefit is
14351 that applications won't have to be linked with libdl or other dso
14352 libraries unless it's really needed.
14354 Changed 'openssl engine' to load all engines on demand.
14355 Changed the engine header files to avoid the duplication of some
14356 declarations (they differed!).
14360 * 'openssl engine' can now list capabilities.
14364 * Better error reporting in 'openssl engine'.
14368 * Never call load_dh_param(NULL) in s_server.
14372 * Add engine application. It can currently list engines by name and
14373 identity, and test if they are actually available.
14377 * Improve RPM specification file by forcing symbolic linking and making
14378 sure the installed documentation is also owned by root.root.
14380 *Damien Miller <djm@mindrot.org>*
14382 * Give the OpenSSL applications more possibilities to make use of
14383 keys (public as well as private) handled by engines.
14387 * Add OCSP code that comes from CertCo.
14391 * Add VMS support for the Rijndael code.
14395 * Added untested support for Nuron crypto accelerator.
14399 * Add support for external cryptographic devices. This code was
14400 previously distributed separately as the "engine" branch.
14402 *Geoff Thorpe, Richard Levitte*
14404 * Rework the filename-translation in the DSO code. It is now possible to
14405 have far greater control over how a "name" is turned into a filename
14406 depending on the operating environment and any oddities about the
14407 different shared library filenames on each system.
14411 * Support threads on FreeBSD-elf in Configure.
14415 * Fix for SHA1 assembly problem with MASM: it produces
14416 warnings about corrupt line number information when assembling
14417 with debugging information. This is caused by the overlapping
14420 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14423 NCONF_get_number() has no error checking at all. As a replacement,
14424 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14425 promoted strongly. The old NCONF_get_number is kept around for
14426 binary backward compatibility.
14427 Make it possible for methods to load from something other than a BIO,
14428 by providing a function pointer that is given a name instead of a BIO.
14429 For example, this could be used to load configuration data from an
14434 * Fix for non blocking accept BIOs. Added new I/O special reason
14435 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14436 with non blocking I/O was not possible because no retry code was
14437 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14442 * Added the beginnings of Rijndael support.
14446 * Fix for bug in DirectoryString mask setting. Add support for
14447 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14448 to allow certificate printing to more controllable, additional
14449 'certopt' option to 'x509' to allow new printing options to be
14454 * Clean old EAY MD5 hack from e_os.h.
14458 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14460 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14461 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14463 *Joe Orton, Steve Henson*
14465 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14467 * Fix additional bug revealed by the NISCC test suite:
14469 Stop bug triggering large recursion when presented with
14470 certain ASN.1 tags ([CVE-2003-0851])
14474 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14476 * Fix various bugs revealed by running the NISCC test suite:
14478 Stop out of bounds reads in the ASN1 code when presented with
14479 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14481 If verify callback ignores invalid public key errors don't try to check
14482 certificate signature with the NULL public key.
14486 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14487 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14492 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14493 extra data after the compression methods not only for TLS 1.0
14494 but also for SSL 3.0 (as required by the specification).
14496 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14498 * Change X509_certificate_type() to mark the key as exported/exportable
14499 when it's 512 *bits* long, not 512 bytes.
14503 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14505 * Countermeasure against the Klima-Pokorny-Rosa extension of
14506 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14507 a protocol version number mismatch like a decryption error
14508 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14512 * Turn on RSA blinding by default in the default implementation
14513 to avoid a timing attack. Applications that don't want it can call
14514 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14515 They would be ill-advised to do so in most cases.
14517 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14519 * Change RSA blinding code so that it works when the PRNG is not
14520 seeded (in this case, the secret RSA exponent is abused as
14521 an unpredictable seed -- if it is not unpredictable, there
14522 is no point in blinding anyway). Make RSA blinding thread-safe
14523 by remembering the creator's thread ID in rsa->blinding and
14524 having all other threads use local one-time blinding factors
14525 (this requires more computation than sharing rsa->blinding, but
14526 avoids excessive locking; and if an RSA object is not shared
14527 between threads, blinding will still be very fast).
14531 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14533 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14534 via timing by performing a MAC computation even if incorrect
14535 block cipher padding has been found. This is a countermeasure
14536 against active attacks where the attacker has to distinguish
14537 between bad padding and a MAC verification error. ([CVE-2003-0078])
14539 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14540 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14541 Martin Vuagnoux (EPFL, Ilion)*
14543 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14545 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14546 memory from its contents. This is done with a counter that will
14547 place alternating values in each byte. This can be used to solve
14548 two issues: 1) the removal of calls to memset() by highly optimizing
14549 compilers, and 2) cleansing with other values than 0, since those can
14550 be read through on certain media, for example a swap space on disk.
14554 * Bugfix: client side session caching did not work with external caching,
14555 because the session->cipher setting was not restored when reloading
14556 from the external cache. This problem was masked, when
14557 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14558 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14562 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14563 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14565 *Zeev Lieber <zeev-l@yahoo.com>*
14567 * Undo an undocumented change introduced in 0.9.6e which caused
14568 repeated calls to OpenSSL_add_all_ciphers() and
14569 OpenSSL_add_all_digests() to be ignored, even after calling
14574 * Change the default configuration reader to deal with last line not
14575 being properly terminated.
14579 * Change X509_NAME_cmp() so it applies the special rules on handling
14580 DN values that are of type PrintableString, as well as RDNs of type
14581 emailAddress where the value has the type ia5String.
14583 *stefank@valicert.com via Richard Levitte*
14585 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14586 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14587 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14588 the bitwise-OR of the two for use by the majority of applications
14589 wanting this behaviour, and update the docs. The documented
14590 behaviour and actual behaviour were inconsistent and had been
14591 changing anyway, so this is more a bug-fix than a behavioural
14594 *Geoff Thorpe, diagnosed by Nadav Har'El*
14596 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14597 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14601 * Fix initialization code race conditions in
14602 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14603 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14604 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14605 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14606 ssl2_get_cipher_by_char(),
14607 ssl3_get_cipher_by_char().
14609 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14611 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14612 the cached sessions are flushed, as the remove_cb() might use ex_data
14613 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14614 (see [openssl.org #212]).
14616 *Geoff Thorpe, Lutz Jaenicke*
14618 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14619 length, instead of the encoding length to d2i_ASN1_OBJECT.
14623 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14625 * [In 0.9.6g-engine release:]
14626 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14628 *Lynn Gazis <lgazis@rainbow.com>*
14630 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14632 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14633 and get fix the header length calculation.
14634 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14635 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14637 * Use proper error handling instead of 'assertions' in buffer
14638 overflow checks added in 0.9.6e. This prevents DoS (the
14639 assertions could call abort()).
14641 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14643 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14645 * Add various sanity checks to asn1_get_length() to reject
14646 the ASN1 length bytes if they exceed sizeof(long), will appear
14647 negative or the content length exceeds the length of the
14650 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14652 * Fix cipher selection routines: ciphers without encryption had no flags
14653 for the cipher strength set and where therefore not handled correctly
14654 by the selection routines (PR #130).
14658 * Fix EVP_dsa_sha macro.
14663 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14664 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14665 that was added in OpenSSL 0.9.6d.
14667 As the countermeasure turned out to be incompatible with some
14668 broken SSL implementations, the new option is part of SSL_OP_ALL.
14669 SSL_OP_ALL is usually employed when compatibility with weird SSL
14670 implementations is desired (e.g. '-bugs' option to 's_client' and
14671 's_server'), so the new option is automatically set in many
14676 * Changes in security patch:
14678 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14679 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14680 Air Force Materiel Command, USAF, under agreement number
14683 * Add various sanity checks to asn1_get_length() to reject
14684 the ASN1 length bytes if they exceed sizeof(long), will appear
14685 negative or the content length exceeds the length of the
14686 supplied buffer. ([CVE-2002-0659])
14688 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14690 * Assertions for various potential buffer overflows, not known to
14691 happen in practice.
14693 *Ben Laurie (CHATS)*
14695 * Various temporary buffers to hold ASCII versions of integers were
14696 too small for 64 bit platforms. ([CVE-2002-0655])
14697 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14699 * Remote buffer overflow in SSL3 protocol - an attacker could
14700 supply an oversized session ID to a client. ([CVE-2002-0656])
14702 *Ben Laurie (CHATS)*
14704 * Remote buffer overflow in SSL2 protocol - an attacker could
14705 supply an oversized client master key. ([CVE-2002-0656])
14707 *Ben Laurie (CHATS)*
14709 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14711 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14712 encoded as NULL) with id-dsa-with-sha1.
14714 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14716 * Check various `X509_...()` return values in `apps/req.c`.
14718 *Nils Larsch <nla@trustcenter.de>*
14720 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14721 an end-of-file condition would erroneously be flagged, when the CRLF
14722 was just at the end of a processed block. The bug was discovered when
14723 processing data through a buffering memory BIO handing the data to a
14724 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14725 <ptsekov@syntrex.com> and Nedelcho Stanev.
14729 * Implement a countermeasure against a vulnerability recently found
14730 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14731 before application data chunks to avoid the use of known IVs
14732 with data potentially chosen by the attacker.
14736 * Fix length checks in ssl3_get_client_hello().
14740 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14741 to prevent ssl3_read_internal() from incorrectly assuming that
14742 ssl3_read_bytes() found application data while handshake
14743 processing was enabled when in fact s->s3->in_read_app_data was
14744 merely automatically cleared during the initial handshake.
14746 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14748 * Fix object definitions for Private and Enterprise: they were not
14749 recognized in their shortname (=lowercase) representation. Extend
14750 obj_dat.pl to issue an error when using undefined keywords instead
14751 of silently ignoring the problem (Svenning Sorensen
14752 <sss@sss.dnsalias.net>).
14756 * Fix DH_generate_parameters() so that it works for 'non-standard'
14757 generators, i.e. generators other than 2 and 5. (Previously, the
14758 code did not properly initialise the 'add' and 'rem' values to
14759 BN_generate_prime().)
14761 In the new general case, we do not insist that 'generator' is
14762 actually a primitive root: This requirement is rather pointless;
14763 a generator of the order-q subgroup is just as good, if not
14768 * Map new X509 verification errors to alerts. Discovered and submitted by
14769 Tom Wu <tom@arcot.com>.
14773 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14774 returning non-zero before the data has been completely received
14775 when using non-blocking I/O.
14777 *Bodo Moeller; problem pointed out by John Hughes*
14779 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14781 *Ben Laurie, Lutz Jaenicke*
14783 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14784 Yoram Zahavi <YoramZ@gilian.com>).
14788 * Add information about CygWin 1.3 and on, and preserve proper
14789 configuration for the versions before that.
14791 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14793 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14794 check whether we deal with a copy of a session and do not delete from
14795 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14796 <izhar@checkpoint.com>.
14800 * Do not store session data into the internal session cache, if it
14801 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14802 flag is set). Proposed by Aslam <aslam@funk.com>.
14806 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14811 * [In 0.9.6d-engine release:]
14812 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14814 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14816 * Add the configuration target linux-s390x.
14818 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14820 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14821 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14822 variable as an indication that a ClientHello message has been
14823 received. As the flag value will be lost between multiple
14824 invocations of ssl3_accept when using non-blocking I/O, the
14825 function may not be aware that a handshake has actually taken
14826 place, thus preventing a new session from being added to the
14829 To avoid this problem, we now set s->new_session to 2 instead of
14830 using a local variable.
14832 *Lutz Jaenicke, Bodo Moeller*
14834 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14835 if the SSL_R_LENGTH_MISMATCH error is detected.
14837 *Geoff Thorpe, Bodo Moeller*
14839 * New 'shared_ldflag' column in Configure platform table.
14843 * Fix EVP_CIPHER_mode macro.
14845 *"Dan S. Camper" <dan@bti.net>*
14847 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14848 type, we must throw them away by setting rr->length to 0.
14850 *D P Chang <dpc@qualys.com>*
14852 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14854 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14855 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14856 worked incorrectly for those cases where range = `10..._2` and
14857 `3*range` is two bits longer than range.)
14861 * Only add signing time to PKCS7 structures if it is not already
14866 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14867 OBJ_ld_ce should be OBJ_id_ce.
14868 Also some ip-pda OIDs in crypto/objects/objects.txt were
14869 incorrect (cf. RFC 3039).
14871 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14873 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14874 returns early because it has nothing to do.
14876 *Andy Schneider <andy.schneider@bjss.co.uk>*
14878 * [In 0.9.6c-engine release:]
14879 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14881 *Andy Schneider <andy.schneider@bjss.co.uk>*
14883 * [In 0.9.6c-engine release:]
14884 Add support for Cryptographic Appliance's keyserver technology.
14885 (Use engine 'keyclient')
14887 *Cryptographic Appliances and Geoff Thorpe*
14889 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14890 is called via tools/c89.sh because arguments have to be
14891 rearranged (all '-L' options must appear before the first object
14894 *Richard Shapiro <rshapiro@abinitio.com>*
14896 * [In 0.9.6c-engine release:]
14897 Add support for Broadcom crypto accelerator cards, backported
14900 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14902 * [In 0.9.6c-engine release:]
14903 Add support for SureWare crypto accelerator cards from
14904 Baltimore Technologies. (Use engine 'sureware')
14906 *Baltimore Technologies and Mark Cox*
14908 * [In 0.9.6c-engine release:]
14909 Add support for crypto accelerator cards from Accelerated
14910 Encryption Processing, www.aep.ie. (Use engine 'aep')
14912 *AEP Inc. and Mark Cox*
14914 * Add a configuration entry for gcc on UnixWare.
14916 *Gary Benson <gbenson@redhat.com>*
14918 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14919 messages are stored in a single piece (fixed-length part and
14920 variable-length part combined) and fix various bugs found on the way.
14924 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14925 instead. BIO_gethostbyname() does not know what timeouts are
14926 appropriate, so entries would stay in cache even when they have
14928 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14930 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14931 faced with a pathologically small ClientHello fragment that does
14932 not contain client_version: Instead of aborting with an error,
14933 simply choose the highest available protocol version (i.e.,
14934 TLS 1.0 unless it is disabled). In practice, ClientHello
14935 messages are never sent like this, but this change gives us
14936 strictly correct behaviour at least for TLS.
14940 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14941 never resets s->method to s->ctx->method when called from within
14942 one of the SSL handshake functions.
14944 *Bodo Moeller; problem pointed out by Niko Baric*
14946 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14947 (sent using the client's version number) if client_version is
14948 smaller than the protocol version in use. Also change
14949 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14950 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14951 the client will at least see that alert.
14955 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14960 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14961 client receives HelloRequest while in a handshake.
14963 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14965 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14966 should end in 'break', not 'goto end' which circumvents various
14967 cleanups done in state SSL_ST_OK. But session related stuff
14968 must be disabled for SSL_ST_OK in the case that we just sent a
14971 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14972 before just sending a HelloRequest.
14974 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14976 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14977 reveal whether illegal block cipher padding was found or a MAC
14978 verification error occurred. (Neither SSLerr() codes nor alerts
14979 are directly visible to potential attackers, but the information
14980 may leak via logfiles.)
14982 Similar changes are not required for the SSL 2.0 implementation
14983 because the number of padding bytes is sent in clear for SSL 2.0,
14984 and the extra bytes are just ignored. However ssl/s2_pkt.c
14985 failed to verify that the purported number of padding bytes is in
14990 * Add OpenUNIX-8 support including shared libraries
14991 (Boyd Lynn Gerber <gerberb@zenez.com>).
14995 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14996 'wristwatch attack' using huge encoding parameters (cf.
14997 James H. Manger's CRYPTO 2001 paper). Note that the
14998 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14999 encoding parameters and hence was not vulnerable.
15003 * BN_sqr() bug fix.
15005 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15007 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15008 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15009 followed by modular reduction.
15011 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15013 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15014 equivalent based on BN_pseudo_rand() instead of BN_rand().
15018 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15019 This function was broken, as the check for a new client hello message
15020 to handle SGC did not allow these large messages.
15021 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15025 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15029 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15030 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15034 * Rework the configuration and shared library support for Tru64 Unix.
15035 The configuration part makes use of modern compiler features and
15036 still retains old compiler behavior for those that run older versions
15037 of the OS. The shared library support part includes a variant that
15038 uses the RPATH feature, and is available through the special
15039 configuration target "alpha-cc-rpath", which will never be selected
15042 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15044 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15045 with the same message size as in ssl3_get_certificate_request().
15046 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15047 messages might inadvertently be reject as too long.
15049 *Petr Lampa <lampa@fee.vutbr.cz>*
15051 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15055 * Modified SSL library such that the verify_callback that has been set
15056 specifically for an SSL object with SSL_set_verify() is actually being
15057 used. Before the change, a verify_callback set with this function was
15058 ignored and the verify_callback() set in the SSL_CTX at the time of
15059 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15060 to allow the necessary settings.
15064 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15065 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15066 done automatically (in contradiction to the requirements of the C
15067 standard). This made problems when used from OpenSSH.
15071 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15072 dh->length and always used
15074 BN_rand_range(priv_key, dh->p).
15076 BN_rand_range() is not necessary for Diffie-Hellman, and this
15077 specific range makes Diffie-Hellman unnecessarily inefficient if
15078 dh->length (recommended exponent length) is much smaller than the
15079 length of dh->p. We could use BN_rand_range() if the order of
15080 the subgroup was stored in the DH structure, but we only have
15085 BN_rand(priv_key, l, ...)
15087 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15094 RSA_eay_public_encrypt
15095 RSA_eay_private_decrypt
15096 RSA_eay_private_encrypt (signing)
15097 RSA_eay_public_decrypt (signature verification)
15099 (default implementations for RSA_public_encrypt,
15100 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15101 always reject numbers >= n.
15105 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15106 to synchronize access to 'locking_thread'. This is necessary on
15107 systems where access to 'locking_thread' (an 'unsigned long'
15108 variable) is not atomic.
15112 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15113 *before* setting the 'crypto_lock_rand' flag. The previous code had
15114 a race condition if 0 is a valid thread ID.
15116 *Travis Vitek <vitek@roguewave.com>*
15118 * Add support for shared libraries under Irix.
15120 *Albert Chin-A-Young <china@thewrittenword.com>*
15122 * Add configuration option to build on Linux on both big-endian and
15123 little-endian MIPS.
15125 *Ralf Baechle <ralf@uni-koblenz.de>*
15127 * Add the possibility to create shared libraries on HP-UX.
15131 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15133 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15134 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15135 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15136 PRNG state recovery was possible based on the output of
15137 one PRNG request appropriately sized to gain knowledge on
15138 'md' followed by enough consecutive 1-byte PRNG requests
15139 to traverse all of 'state'.
15141 1. When updating 'md_local' (the current thread's copy of 'md')
15142 during PRNG output generation, hash all of the previous
15143 'md_local' value, not just the half used for PRNG output.
15145 2. Make the number of bytes from 'state' included into the hash
15146 independent from the number of PRNG bytes requested.
15148 The first measure alone would be sufficient to avoid
15149 Markku-Juhani's attack. (Actually it had never occurred
15150 to me that the half of 'md_local' used for chaining was the
15151 half from which PRNG output bytes were taken -- I had always
15152 assumed that the secret half would be used.) The second
15153 measure makes sure that additional data from 'state' is never
15154 mixed into 'md_local' in small portions; this heuristically
15155 further strengthens the PRNG.
15159 * Fix crypto/bn/asm/mips3.s.
15163 * When only the key is given to "enc", the IV is undefined. Print out
15164 an error message in this case.
15168 * Handle special case when X509_NAME is empty in X509 printing routines.
15172 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15173 positive and less than q.
15177 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15178 used: it isn't thread safe and the add_lock_callback should handle
15181 *Paul Rose <Paul.Rose@bridge.com>*
15183 * Verify that incoming data obeys the block size in
15184 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15190 *Ulf Möller, Bodo Möller*
15192 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15193 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15194 when fixing the server behaviour for backwards-compatible 'client
15195 hello' messages. (Note that the attack is impractical against
15196 SSL 3.0 and TLS 1.0 anyway because length and version checking
15197 means that the probability of guessing a valid ciphertext is
15198 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15201 Before 0.9.5, the countermeasure (hide the error by generating a
15202 random 'decryption result') did not work properly because
15203 ERR_clear_error() was missing, meaning that SSL_get_error() would
15204 detect the supposedly ignored error.
15206 Both problems are now fixed.
15210 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15211 (previously it was 1024).
15215 * Fix for compatibility mode trust settings: ignore trust settings
15216 unless some valid trust or reject settings are present.
15220 * Fix for blowfish EVP: its a variable length cipher.
15224 * Fix various bugs related to DSA S/MIME verification. Handle missing
15225 parameters in DSA public key structures and return an error in the
15226 DSA routines if parameters are absent.
15230 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15231 in the current directory if neither $RANDFILE nor $HOME was set.
15232 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15233 caused some confusion to Windows users who haven't defined $HOME.
15234 Thus RAND_file_name() is changed again: e_os.h can define a
15235 DEFAULT_HOME, which will be used if $HOME is not set.
15236 For Windows, we use "C:"; on other platforms, we still require
15237 environment variables.
15239 * Move 'if (!initialized) RAND_poll()' into regions protected by
15240 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15241 having multiple threads call RAND_poll() concurrently.
15245 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15246 combination of a flag and a thread ID variable.
15247 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15248 flag), *other* threads can enter ssleay_add_bytes without obeying
15249 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15250 that they do not hold after the first thread unsets add_do_not_lock).
15254 * Change bctest again: '-x' expressions are not available in all
15255 versions of 'test'.
15259 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15261 * Fix a couple of memory leaks in PKCS7_dataDecode()
15263 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15265 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15266 the default extension for executables, if any. Also, make the perl
15267 scripts that use symlink() to test if it really exists and use "cp"
15268 if it doesn't. All this made OpenSSL compilable and installable in
15273 * Fix for asn1_GetSequence() for indefinite length constructed data.
15274 If SEQUENCE is length is indefinite just set c->slen to the total
15275 amount of data available.
15277 *Steve Henson, reported by shige@FreeBSD.org*
15279 *This change does not apply to 0.9.7.*
15281 * Change bctest to avoid here-documents inside command substitution
15282 (workaround for FreeBSD /bin/sh bug).
15283 For compatibility with Ultrix, avoid shell functions (introduced
15284 in the bctest version that searches along $PATH).
15288 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15289 with des_encrypt() defined on some operating systems, like Solaris
15294 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15295 On the Importance of Eliminating Errors in Cryptographic
15296 Computations, J. Cryptology 14 (2001) 2, 101-119,
15297 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15301 * MIPS assembler BIGNUM division bug fix.
15305 * Disabled incorrect Alpha assembler code.
15309 * Fix PKCS#7 decode routines so they correctly update the length
15310 after reading an EOC for the EXPLICIT tag.
15314 *This change does not apply to 0.9.7.*
15316 * Fix bug in PKCS#12 key generation routines. This was triggered
15317 if a 3DES key was generated with a 0 initial byte. Include
15318 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15319 (but broken) behaviour.
15323 * Enhance bctest to search for a working bc along $PATH and print
15326 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15328 * Fix memory leaks in err.c: free err_data string if necessary;
15329 don't write to the wrong index in ERR_set_error_data.
15333 * Implement ssl23_peek (analogous to ssl23_read), which previously
15338 * Replace rdtsc with `_emit` statements for VC++ version 5.
15340 *Jeremy Cooper <jeremy@baymoo.org>*
15342 * Make it possible to reuse SSLv2 sessions.
15346 * In copy_email() check for >= 0 as a return value for
15347 X509_NAME_get_index_by_NID() since 0 is a valid index.
15349 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15351 * Avoid coredump with unsupported or invalid public keys by checking if
15352 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15353 PKCS7_verify() fails with non detached data.
15357 * Don't use getenv in library functions when run as setuid/setgid.
15358 New function OPENSSL_issetugid().
15362 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15363 due to incorrect handling of multi-threading:
15365 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15367 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15369 3. Count how many times MemCheck_off() has been called so that
15370 nested use can be treated correctly. This also avoids
15371 inband-signalling in the previous code (which relied on the
15372 assumption that thread ID 0 is impossible).
15376 * Add "-rand" option also to s_client and s_server.
15380 * Fix CPU detection on Irix 6.x.
15381 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15382 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15384 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15389 *This change does not apply to 0.9.7.*
15391 * Use the cached encoding of an X509_NAME structure rather than
15392 copying it. This is apparently the reason for the libsafe "errors"
15393 but the code is actually correct.
15397 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15398 Bleichenbacher's DSA attack.
15399 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15400 to be set and top=0 forces the highest bit to be set; top=-1 is new
15401 and leaves the highest bit random.
15403 *Ulf Moeller, Bodo Moeller*
15405 * In the `NCONF_...`-based implementations for `CONF_...` queries
15406 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15407 a temporary CONF structure with the data component set to NULL
15408 (which gives segmentation faults in lh_retrieve).
15409 Instead, use NULL for the CONF pointer in CONF_get_string and
15410 CONF_get_number (which may use environment variables) and directly
15411 return NULL from CONF_get_section.
15415 * Fix potential buffer overrun for EBCDIC.
15419 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15420 keyUsage if basicConstraints absent for a CA.
15424 * Make SMIME_write_PKCS7() write mail header values with a format that
15425 is more generally accepted (no spaces before the semicolon), since
15426 some programs can't parse those values properly otherwise. Also make
15427 sure BIO's that break lines after each write do not create invalid
15432 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15433 macros previously used would not encode an empty SEQUENCE OF
15434 and break the signature.
15438 *This change does not apply to 0.9.7.*
15440 * Zero the premaster secret after deriving the master secret in
15445 * Add some EVP_add_digest_alias registrations (as found in
15446 OpenSSL_add_all_digests()) to SSL_library_init()
15447 aka OpenSSL_add_ssl_algorithms(). This provides improved
15448 compatibility with peers using X.509 certificates
15449 with unconventional AlgorithmIdentifier OIDs.
15453 * Fix for Irix with NO_ASM.
15455 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15457 * ./config script fixes.
15459 *Ulf Moeller, Richard Levitte*
15461 * Fix 'openssl passwd -1'.
15465 * Change PKCS12_key_gen_asc() so it can cope with non null
15466 terminated strings whose length is passed in the passlen
15467 parameter, for example from PEM callbacks. This was done
15468 by adding an extra length parameter to asc2uni().
15470 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15472 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15473 call failed, free the DSA structure.
15477 * Fix to uni2asc() to cope with zero length Unicode strings.
15478 These are present in some PKCS#12 files.
15482 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15483 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15484 when writing a 32767 byte record.
15486 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15488 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15489 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15491 (RSA objects have a reference count access to which is protected
15492 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15493 so they are meant to be shared between threads.)
15494 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15495 "Reddie, Steven" <Steven.Reddie@ca.com>*
15497 * Fix a deadlock in CRYPTO_mem_leaks().
15501 * Use better test patterns in bntest.
15505 * rand_win.c fix for Borland C.
15509 * BN_rshift bugfix for n == 0.
15513 * Add a 'bctest' script that checks for some known 'bc' bugs
15514 so that 'make test' does not abort just because 'bc' is broken.
15518 * Store verify_result within SSL_SESSION also for client side to
15519 avoid potential security hole. (Re-used sessions on the client side
15520 always resulted in verify_result==X509_V_OK, not using the original
15521 result of the server certificate verification.)
15525 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15526 SSL3_RT_APPLICATION_DATA, return 0.
15527 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15532 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15533 releases, have been re-implemented by renaming the previous
15534 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15535 and ssl3_read_internal, respectively, and adding 'peek' parameters
15536 to them. The new ssl[23]_{read,peek} functions are calls to
15537 ssl[23]_read_internal with the 'peek' flag set appropriately.
15538 A 'peek' parameter has also been added to ssl3_read_bytes, which
15539 does the actual work for ssl3_read_internal.
15543 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15544 the method-specific "init()" handler. Also clean up ex_data after
15545 calling the method-specific "finish()" handler. Previously, this was
15546 happening the other way round.
15550 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15551 The previous value, 12, was not always sufficient for BN_mod_exp().
15555 * Make sure that shared libraries get the internal name engine with
15556 the full version number and not just 0. This should mark the
15557 shared libraries as not backward compatible. Of course, this should
15558 be changed again when we can guarantee backward binary compatibility.
15562 * Fix typo in get_cert_by_subject() in by_dir.c
15564 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15566 * Rework the system to generate shared libraries:
15568 - Make note of the expected extension for the shared libraries and
15569 if there is a need for symbolic links from for example libcrypto.so.0
15570 to libcrypto.so.0.9.7. There is extended info in Configure for
15573 - Make as few rebuilds of the shared libraries as possible.
15575 - Still avoid linking the OpenSSL programs with the shared libraries.
15577 - When installing, install the shared libraries separately from the
15582 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15584 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15585 and not in SSL_clear because the latter is also used by the
15586 accept/connect functions; previously, the settings made by
15587 SSL_set_read_ahead would be lost during the handshake.
15589 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15591 * Correct util/mkdef.pl to be selective about disabled algorithms.
15592 Previously, it would create entries for disabled algorithms no
15597 * Added several new manual pages for SSL_* function.
15601 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15603 * In ssl23_get_client_hello, generate an error message when faced
15604 with an initial SSL 3.0/TLS record that is too small to contain the
15605 first two bytes of the ClientHello message, i.e. client_version.
15606 (Note that this is a pathologic case that probably has never happened
15607 in real life.) The previous approach was to use the version number
15608 from the record header as a substitute; but our protocol choice
15609 should not depend on that one because it is not authenticated
15610 by the Finished messages.
15614 * More robust randomness gathering functions for Windows.
15616 *Jeffrey Altman <jaltman@columbia.edu>*
15618 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15619 not set then we don't setup the error code for issuer check errors
15620 to avoid possibly overwriting other errors which the callback does
15621 handle. If an application does set the flag then we assume it knows
15622 what it is doing and can handle the new informational codes
15627 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15628 a general "ANY" type, as such it should be able to decode anything
15629 including tagged types. However it didn't check the class so it would
15630 wrongly interpret tagged types in the same way as their universal
15631 counterpart and unknown types were just rejected. Changed so that the
15632 tagged and unknown types are handled in the same way as a SEQUENCE:
15633 that is the encoding is stored intact. There is also a new type
15634 "V_ASN1_OTHER" which is used when the class is not universal, in this
15635 case we have no idea what the actual type is so we just lump them all
15640 * On VMS, stdout may very well lead to a file that is written to
15641 in a record-oriented fashion. That means that every write() will
15642 write a separate record, which will be read separately by the
15643 programs trying to read from it. This can be very confusing.
15645 The solution is to put a BIO filter in the way that will buffer
15646 text until a linefeed is reached, and then write everything a
15647 line at a time, so every record written will be an actual line,
15648 not chunks of lines and not (usually doesn't happen, but I've
15649 seen it once) several lines in one record. BIO_f_linebuffer() is
15652 Currently, it's a VMS-only method, because that's where it has
15653 been tested well enough.
15657 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15658 it can return incorrect results.
15659 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15660 but it was in 0.9.6-beta[12].)
15664 * Disable the check for content being present when verifying detached
15665 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15666 include zero length content when signing messages.
15670 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15671 BIO_ctrl (for BIO pairs).
15675 * Add DSO method for VMS.
15679 * Bug fix: Montgomery multiplication could produce results with the
15684 * Add RPM specification openssl.spec and modify it to build three
15685 packages. The default package contains applications, application
15686 documentation and run-time libraries. The devel package contains
15687 include files, static libraries and function documentation. The
15688 doc package contains the contents of the doc directory. The original
15689 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15693 * Add a large number of documentation files for many SSL routines.
15695 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15697 * Add a configuration entry for Sony News 4.
15699 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15701 * Don't set the two most significant bits to one when generating a
15702 random number < q in the DSA library.
15706 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15707 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15708 the underlying transport is blocking) if a handshake took place.
15709 (The default behaviour is needed by applications such as s_client
15710 and s_server that use select() to determine when to use SSL_read;
15711 but for applications that know in advance when to expect data, it
15712 just makes things more complicated.)
15716 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15721 * Add a few more EBCDIC conditionals that make `req` and `x509`
15722 work better on such systems.
15724 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15726 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15727 Update PKCS12_parse() so it copies the friendlyName and the
15728 keyid to the certificates aux info.
15732 * Fix bug in PKCS7_verify() which caused an infinite loop
15733 if there was more than one signature.
15735 *Sven Uszpelkat <su@celocom.de>*
15737 * Major change in util/mkdef.pl to include extra information
15738 about each symbol, as well as presenting variables as well
15739 as functions. This change means that there's n more need
15740 to rebuild the .num files when some algorithms are excluded.
15744 * Allow the verify time to be set by an application,
15745 rather than always using the current time.
15749 * Phase 2 verify code reorganisation. The certificate
15750 verify code now looks up an issuer certificate by a
15751 number of criteria: subject name, authority key id
15752 and key usage. It also verifies self signed certificates
15753 by the same criteria. The main comparison function is
15754 X509_check_issued() which performs these checks.
15756 Lot of changes were necessary in order to support this
15757 without completely rewriting the lookup code.
15759 Authority and subject key identifier are now cached.
15761 The LHASH 'certs' is X509_STORE has now been replaced
15762 by a STACK_OF(X509_OBJECT). This is mainly because an
15763 LHASH can't store or retrieve multiple objects with
15764 the same hash value.
15766 As a result various functions (which were all internal
15767 use only) have changed to handle the new X509_STORE
15768 structure. This will break anything that messed round
15769 with X509_STORE internally.
15771 The functions X509_STORE_add_cert() now checks for an
15772 exact match, rather than just subject name.
15774 The X509_STORE API doesn't directly support the retrieval
15775 of multiple certificates matching a given criteria, however
15776 this can be worked round by performing a lookup first
15777 (which will fill the cache with candidate certificates)
15778 and then examining the cache for matches. This is probably
15779 the best we can do without throwing out X509_LOOKUP
15780 entirely (maybe later...).
15782 The X509_VERIFY_CTX structure has been enhanced considerably.
15784 All certificate lookup operations now go via a get_issuer()
15785 callback. Although this currently uses an X509_STORE it
15786 can be replaced by custom lookups. This is a simple way
15787 to bypass the X509_STORE hackery necessary to make this
15788 work and makes it possible to use more efficient techniques
15789 in future. A very simple version which uses a simple
15790 STACK for its trusted certificate store is also provided
15791 using X509_STORE_CTX_trusted_stack().
15793 The verify_cb() and verify() callbacks now have equivalents
15794 in the X509_STORE_CTX structure.
15796 X509_STORE_CTX also has a 'flags' field which can be used
15797 to customise the verify behaviour.
15801 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15802 excludes S/MIME capabilities.
15806 * When a certificate request is read in keep a copy of the
15807 original encoding of the signed data and use it when outputting
15808 again. Signatures then use the original encoding rather than
15809 a decoded, encoded version which may cause problems if the
15810 request is improperly encoded.
15814 * For consistency with other BIO_puts implementations, call
15815 buffer_write(b, ...) directly in buffer_puts instead of calling
15818 In BIO_puts, increment b->num_write as in BIO_write.
15820 *Peter.Sylvester@EdelWeb.fr*
15822 * Fix BN_mul_word for the case where the word is 0. (We have to use
15823 BN_zero, we may not return a BIGNUM with an array consisting of
15824 words set to zero.)
15828 * Avoid calling abort() from within the library when problems are
15829 detected, except if preprocessor symbols have been defined
15830 (such as REF_CHECK, BN_DEBUG etc.).
15834 * New openssl application 'rsautl'. This utility can be
15835 used for low-level RSA operations. DER public key
15836 BIO/fp routines also added.
15840 * New Configure entry and patches for compiling on QNX 4.
15842 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15844 * A demo state-machine implementation was sponsored by
15845 Nuron (<http://www.nuron.com/>) and is now available in
15846 demos/state_machine.
15850 * New options added to the 'dgst' utility for signature
15851 generation and verification.
15855 * Unrecognized PKCS#7 content types are now handled via a
15856 catch all ASN1_TYPE structure. This allows unsupported
15857 types to be stored as a "blob" and an application can
15858 encode and decode it manually.
15862 * Fix various signed/unsigned issues to make a_strex.c
15863 compile under VC++.
15865 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15867 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15868 length if passed a buffer. ASN1_INTEGER_to_BN failed
15869 if passed a NULL BN and its argument was negative.
15871 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15873 * Modification to PKCS#7 encoding routines to output definite
15874 length encoding. Since currently the whole structures are in
15875 memory there's not real point in using indefinite length
15876 constructed encoding. However if OpenSSL is compiled with
15877 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15881 * Added BIO_vprintf() and BIO_vsnprintf().
15885 * Added more prefixes to parse for in the strings written
15886 through a logging bio, to cover all the levels that are available
15887 through syslog. The prefixes are now:
15889 PANIC, EMERG, EMR => LOG_EMERG
15890 ALERT, ALR => LOG_ALERT
15891 CRIT, CRI => LOG_CRIT
15892 ERROR, ERR => LOG_ERR
15893 WARNING, WARN, WAR => LOG_WARNING
15894 NOTICE, NOTE, NOT => LOG_NOTICE
15895 INFO, INF => LOG_INFO
15896 DEBUG, DBG => LOG_DEBUG
15898 and as before, if none of those prefixes are present at the
15899 beginning of the string, LOG_ERR is chosen.
15901 On Win32, the `LOG_*` levels are mapped according to this:
15903 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15904 LOG_WARNING => EVENTLOG_WARNING_TYPE
15905 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15909 * Made it possible to reconfigure with just the configuration
15910 argument "reconf" or "reconfigure". The command line arguments
15911 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15912 and are retrieved from there when reconfiguring.
15918 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15920 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15924 * The obj_dat.pl script was messing up the sorting of object
15925 names. The reason was that it compared the quoted version
15926 of strings as a result "OCSP" > "OCSP Signing" because
15927 " > SPACE. Changed script to store unquoted versions of
15928 names and add quotes on output. It was also omitting some
15929 names from the lookup table if they were given a default
15930 value (that is if SN is missing it is given the same
15931 value as LN and vice versa), these are now added on the
15932 grounds that if an object has a name we should be able to
15933 look it up. Finally added warning output when duplicate
15934 short or long names are found.
15938 * Changes needed for Tandem NSK.
15940 *Scott Uroff <scott@xypro.com>*
15942 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15943 RSA_padding_check_SSLv23(), special padding was never detected
15944 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15945 version rollback attacks was not effective.
15947 In s23_clnt.c, don't use special rollback-attack detection padding
15948 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15949 client; similarly, in s23_srvr.c, don't do the rollback check if
15950 SSL 2.0 is the only protocol enabled in the server.
15954 * Make it possible to get hexdumps of unprintable data with 'openssl
15955 asn1parse'. By implication, the functions ASN1_parse_dump() and
15956 BIO_dump_indent() are added.
15960 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15961 these print out strings and name structures based on various
15962 flags including RFC2253 support and proper handling of
15963 multibyte characters. Added options to the 'x509' utility
15964 to allow the various flags to be set.
15968 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15969 Also change the functions X509_cmp_current_time() and
15970 X509_gmtime_adj() work with an ASN1_TIME structure,
15971 this will enable certificates using GeneralizedTime in validity
15972 dates to be checked.
15976 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15977 negative public key encodings) on by default,
15978 NO_NEG_PUBKEY_BUG can be set to disable it.
15982 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15983 content octets. An i2c_ASN1_OBJECT is unnecessary because
15984 the encoding can be trivially obtained from the structure.
15988 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15989 not read locks (`CRYPTO_r_[un]lock`).
15993 * A first attempt at creating official support for shared
15994 libraries through configuration. I've kept it so the
15995 default is static libraries only, and the OpenSSL programs
15996 are always statically linked for now, but there are
15997 preparations for dynamic linking in place.
15998 This has been tested on Linux and Tru64.
16002 * Randomness polling function for Win9x, as described in:
16003 Peter Gutmann, Software Generation of Practically Strong
16008 * Fix so PRNG is seeded in req if using an already existing
16013 * New options to smime application. -inform and -outform
16014 allow alternative formats for the S/MIME message including
16015 PEM and DER. The -content option allows the content to be
16016 specified separately. This should allow things like Netscape
16017 form signing output easier to verify.
16021 * Fix the ASN1 encoding of tags using the 'long form'.
16025 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16026 STRING types. These convert content octets to and from the
16027 underlying type. The actual tag and length octets are
16028 already assumed to have been read in and checked. These
16029 are needed because all other string types have virtually
16030 identical handling apart from the tag. By having versions
16031 of the ASN1 functions that just operate on content octets
16032 IMPLICIT tagging can be handled properly. It also allows
16033 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16034 and ASN1_INTEGER are identical apart from the tag.
16038 * Change the handling of OID objects as follows:
16040 - New object identifiers are inserted in objects.txt, following
16041 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16042 - objects.pl is used to process obj_mac.num and create a new
16044 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16047 This is currently kind of a hack, and the perl code in objects.pl
16048 isn't very elegant, but it works as I intended. The simplest way
16049 to check that it worked correctly is to look in obj_dat.h and
16050 check the array nid_objs and make sure the objects haven't moved
16051 around (this is important!). Additions are OK, as well as
16052 consistent name changes.
16056 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16060 * Addition of the command line parameter '-rand file' to 'openssl req'.
16061 The given file adds to whatever has already been seeded into the
16062 random pool through the RANDFILE configuration file option or
16063 environment variable, or the default random state file.
16067 * mkstack.pl now sorts each macro group into lexical order.
16068 Previously the output order depended on the order the files
16069 appeared in the directory, resulting in needless rewriting
16074 * Patches to make OpenSSL compile under Win32 again. Mostly
16075 work arounds for the VC++ problem that it treats func() as
16076 func(void). Also stripped out the parts of mkdef.pl that
16077 added extra typesafe functions: these no longer exist.
16081 * Reorganisation of the stack code. The macros are now all
16082 collected in safestack.h . Each macro is defined in terms of
16083 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16084 DEBUG_SAFESTACK is now handled in terms of function casts,
16085 this has the advantage of retaining type safety without the
16086 use of additional functions. If DEBUG_SAFESTACK is not defined
16087 then the non typesafe macros are used instead. Also modified the
16088 mkstack.pl script to handle the new form. Needs testing to see
16089 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16090 the default if no major problems. Similar behaviour for ASN1_SET_OF
16091 and PKCS12_STACK_OF.
16095 * When some versions of IIS use the 'NET' form of private key the
16096 key derivation algorithm is different. Normally MD5(password) is
16097 used as a 128 bit RC4 key. In the modified case
16098 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16099 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16100 as the old Netscape_RSA functions except they have an additional
16101 'sgckey' parameter which uses the modified algorithm. Also added
16102 an -sgckey command line option to the rsa utility. Thanks to
16103 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16104 algorithm to openssl-dev.
16108 * The evp_local.h macros were using 'c.##kname' which resulted in
16109 invalid expansion on some systems (SCO 5.0.5 for example).
16110 Corrected to 'c.kname'.
16112 *Phillip Porch <root@theporch.com>*
16114 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16115 a STACK of email addresses from a certificate or request, these look
16116 in the subject name and the subject alternative name extensions and
16117 omit any duplicate addresses.
16121 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16122 This makes DSA verification about 2 % faster.
16126 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16127 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16128 plus overhead for 1024 bit moduli).
16129 This makes exponentiations about 0.5 % faster for 1024 bit
16130 exponents (as measured by "openssl speed rsa2048").
16134 * Rename memory handling macros to avoid conflicts with other
16136 Malloc => OPENSSL_malloc
16137 Malloc_locked => OPENSSL_malloc_locked
16138 Realloc => OPENSSL_realloc
16139 Free => OPENSSL_free
16143 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16144 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16148 * CygWin32 support.
16150 *John Jarvie <jjarvie@newsguy.com>*
16152 * The type-safe stack code has been rejigged. It is now only compiled
16153 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16154 by default all type-specific stack functions are "#define"d back to
16155 standard stack functions. This results in more streamlined output
16156 but retains the type-safety checking possibilities of the original
16161 * The STACK code has been cleaned up, and certain type declarations
16162 that didn't make a lot of sense have been brought in line. This has
16163 also involved a cleanup of sorts in safestack.h to more correctly
16164 map type-safe stack functions onto their plain stack counterparts.
16165 This work has also resulted in a variety of "const"ifications of
16166 lots of the code, especially `_cmp` operations which should normally
16167 be prototyped with "const" parameters anyway.
16171 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16172 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16173 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16174 where all of 'md' is used each time the PRNG is used, but 'state'
16175 is used only indexed by a cyclic counter. As entropy may not be
16176 well distributed from the beginning, 'md' is important as a
16177 chaining variable. However, the output function chains only half
16178 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16179 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16180 in all of 'state' being rewritten, with the new values depending
16181 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16185 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16186 the handshake is continued after ssl_verify_cert_chain();
16187 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16188 can lead to 'unexplainable' connection aborts later.
16190 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16192 * Major EVP API cipher revision.
16193 Add hooks for extra EVP features. This allows various cipher
16194 parameters to be set in the EVP interface. Support added for variable
16195 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16196 setting of RC2 and RC5 parameters.
16198 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16201 Remove lots of duplicated code from the EVP library. For example *every*
16202 cipher init() function handles the 'iv' in the same way according to the
16203 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16204 for CFB and OFB modes they zero ctx->num.
16206 New functionality allows removal of S/MIME code RC2 hack.
16208 Most of the routines have the same form and so can be declared in terms
16211 By shifting this to the top level EVP_CipherInit() it can be removed from
16212 all individual ciphers. If the cipher wants to handle IVs or keys
16213 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16216 Change lots of functions like EVP_EncryptUpdate() to now return a
16217 value: although software versions of the algorithms cannot fail
16218 any installed hardware versions can.
16222 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16223 this option is set, tolerate broken clients that send the negotiated
16224 protocol version number instead of the requested protocol version
16229 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16230 i.e. non-zero for export ciphersuites, zero otherwise.
16231 Previous versions had this flag inverted, inconsistent with
16232 rsa_tmp_cb (..._TMP_RSA_CB).
16234 *Bodo Moeller; problem reported by Amit Chopra*
16236 * Add missing DSA library text string. Work around for some IIS
16237 key files with invalid SEQUENCE encoding.
16241 * Add a document (doc/standards.txt) that list all kinds of standards
16242 and so on that are implemented in OpenSSL.
16246 * Enhance c_rehash script. Old version would mishandle certificates
16247 with the same subject name hash and wouldn't handle CRLs at all.
16248 Added -fingerprint option to crl utility, to support new c_rehash
16253 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16257 * Fix for SSL server purpose checking. Server checking was
16258 rejecting certificates which had extended key usage present
16259 but no ssl client purpose.
16261 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16263 * Make PKCS#12 code work with no password. The PKCS#12 spec
16264 is a little unclear about how a blank password is handled.
16265 Since the password in encoded as a BMPString with terminating
16266 double NULL a zero length password would end up as just the
16267 double NULL. However no password at all is different and is
16268 handled differently in the PKCS#12 key generation code. NS
16269 treats a blank password as zero length. MSIE treats it as no
16270 password on export: but it will try both on import. We now do
16271 the same: PKCS12_parse() tries zero length and no password if
16272 the password is set to "" or NULL (NULL is now a valid password:
16273 it wasn't before) as does the pkcs12 application.
16277 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16278 perror when PEM_read_bio_X509_REQ fails, the error message must
16279 be obtained from the error queue.
16283 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16284 it in ERR_remove_state if appropriate, and change ERR_get_state
16285 accordingly to avoid race conditions (this is necessary because
16286 thread_hash is no longer constant once set).
16290 * Bugfix for linux-elf makefile.one.
16294 * RSA_get_default_method() will now cause a default
16295 RSA_METHOD to be chosen if one doesn't exist already.
16296 Previously this was only set during a call to RSA_new()
16297 or RSA_new_method(NULL) meaning it was possible for
16298 RSA_get_default_method() to return NULL.
16302 * Added native name translation to the existing DSO code
16303 that will convert (if the flag to do so is set) filenames
16304 that are sufficiently small and have no path information
16305 into a canonical native form. Eg. "blah" converted to
16306 "libblah.so" or "blah.dll" etc.
16310 * New function ERR_error_string_n(e, buf, len) which is like
16311 ERR_error_string(e, buf), but writes at most 'len' bytes
16312 including the 0 terminator. For ERR_error_string_n, 'buf'
16315 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16317 * CONF library reworked to become more general. A new CONF
16318 configuration file reader "class" is implemented as well as a
16319 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16320 old `CONF_*` functions are still there, but are reimplemented to
16321 work in terms of the new functions. Also, a set of functions
16322 to handle the internal storage of the configuration data is
16323 provided to make it easier to write new configuration file
16324 reader "classes" (I can definitely see something reading a
16325 configuration file in XML format, for example), called `_CONF_*`,
16326 or "the configuration storage API"...
16328 The new configuration file reading functions are:
16330 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16331 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16333 NCONF_default, NCONF_WIN32
16335 NCONF_dump_fp, NCONF_dump_bio
16337 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16338 NCONF_new creates a new CONF object. This works in the same way
16339 as other interfaces in OpenSSL, like the BIO interface.
16340 `NCONF_dump_*` dump the internal storage of the configuration file,
16341 which is useful for debugging. All other functions take the same
16342 arguments as the old `CONF_*` functions with the exception of the
16343 first that must be a `CONF *` instead of a `LHASH *`.
16345 To make it easier to use the new classes with the old `CONF_*` functions,
16346 the function CONF_set_default_method is provided.
16350 * Add '-tls1' option to 'openssl ciphers', which was already
16351 mentioned in the documentation but had not been implemented.
16352 (This option is not yet really useful because even the additional
16353 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16357 * Initial DSO code added into libcrypto for letting OpenSSL (and
16358 OpenSSL-based applications) load shared libraries and bind to
16359 them in a portable way.
16361 *Geoff Thorpe, with contributions from Richard Levitte*
16363 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16365 * Make sure _lrotl and _lrotr are only used with MSVC.
16367 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16368 (the default implementation of RAND_status).
16370 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16371 to '-clrext' (= clear extensions), as intended and documented.
16372 *Bodo Moeller; inconsistency pointed out by Michael Attili
16373 <attili@amaxo.com>*
16375 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16376 was larger than the MD block size.
16378 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16380 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16381 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16382 using the passed key: if the passed key was a private key the result
16383 of X509_print(), for example, would be to print out all the private key
16388 * des_quad_cksum() byte order bug fix.
16389 *Ulf Möller, using the problem description in krb4-0.9.7, where
16390 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16392 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16395 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16397 * For easily testing in shell scripts whether some command
16398 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16399 returns with exit code 0 iff no command of the given name is available.
16400 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16401 the output goes to stdout and nothing is printed to stderr.
16402 Additional arguments are always ignored.
16404 Since for each cipher there is a command of the same name,
16405 the 'no-cipher' compilation switches can be tested this way.
16407 ('openssl no-XXX' is not able to detect pseudo-commands such
16408 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16412 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16416 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16417 is set; it will be thrown away anyway because each handshake creates
16419 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16420 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16421 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16422 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16426 * New s_client option -ign_eof: EOF at stdin is ignored, and
16427 'Q' and 'R' lose their special meanings (quit/renegotiate).
16428 This is part of what -quiet does; unlike -quiet, -ign_eof
16429 does not suppress any output.
16433 * Add compatibility options to the purpose and trust code. The
16434 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16435 accepts a certificate or CA, this was the previous behaviour,
16436 with all the associated security issues.
16438 X509_TRUST_COMPAT is the old trust behaviour: only and
16439 automatically trust self signed roots in certificate store. A
16440 new trust setting X509_TRUST_DEFAULT is used to specify that
16441 a purpose has no associated trust setting and it should instead
16442 use the value in the default purpose.
16446 * Fix the PKCS#8 DSA private key code so it decodes keys again
16447 and fix a memory leak.
16451 * In util/mkerr.pl (which implements 'make errors'), preserve
16452 reason strings from the previous version of the .c file, as
16453 the default to have only downcase letters (and digits) in
16454 automatically generated reasons codes is not always appropriate.
16458 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16459 using strerror. Previously, ERR_reason_error_string() returned
16460 library names as reason strings for SYSerr; but SYSerr is a special
16461 case where small numbers are errno values, not library numbers.
16465 * Add '-dsaparam' option to 'openssl dhparam' application. This
16466 converts DSA parameters into DH parameters. (When creating parameters,
16467 DSA_generate_parameters is used.)
16471 * Include 'length' (recommended exponent length) in C code generated
16472 by 'openssl dhparam -C'.
16476 * The second argument to set_label in perlasm was already being used
16477 so couldn't be used as a "file scope" flag. Moved to third argument
16482 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16483 instead of RAND_bytes for encryption IVs and salts.
16487 * Include RAND_status() into RAND_METHOD instead of implementing
16488 it only for md_rand.c Otherwise replacing the PRNG by calling
16489 RAND_set_rand_method would be impossible.
16493 * Don't let DSA_generate_key() enter an infinite loop if the random
16494 number generation fails.
16498 * New 'rand' application for creating pseudo-random output.
16502 * Added configuration support for Linux/IA64
16504 *Rolf Haberrecker <rolf@suse.de>*
16506 * Assembler module support for Mingw32.
16510 * Shared library support for HPUX (in shlib/).
16512 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16514 * Shared library support for Solaris gcc.
16516 *Lutz Behnke <behnke@trustcenter.de>*
16518 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16520 * PKCS7_encrypt() was adding text MIME headers twice because they
16521 were added manually and by SMIME_crlf_copy().
16525 * In bntest.c don't call BN_rand with zero bits argument.
16527 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16529 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16530 case was implemented. This caused BN_div_recp() to fail occasionally.
16534 * Add an optional second argument to the set_label() in the perl
16535 assembly language builder. If this argument exists and is set
16536 to 1 it signals that the assembler should use a symbol whose
16537 scope is the entire file, not just the current function. This
16538 is needed with MASM which uses the format label:: for this scope.
16540 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16542 * Change the ASN1 types so they are typedefs by default. Before
16543 almost all types were #define'd to ASN1_STRING which was causing
16544 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16549 * Change names of new functions to the new get1/get0 naming
16550 convention: After 'get1', the caller owns a reference count
16551 and has to call `..._free`; 'get0' returns a pointer to some
16552 data structure without incrementing reference counters.
16553 (Some of the existing 'get' functions increment a reference
16554 counter, some don't.)
16555 Similarly, 'set1' and 'add1' functions increase reference
16556 counters or duplicate objects.
16560 * Allow for the possibility of temp RSA key generation failure:
16561 the code used to assume it always worked and crashed on failure.
16565 * Fix potential buffer overrun problem in BIO_printf().
16566 *Ulf Möller, using public domain code by Patrick Powell; problem
16567 pointed out by David Sacerdote <das33@cornell.edu>*
16569 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16570 RAND_egd() and RAND_status(). In the command line application,
16571 the EGD socket can be specified like a seed file using RANDFILE
16576 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16577 Some CAs (e.g. Verisign) distribute certificates in this form.
16581 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16582 list to exclude them. This means that no special compilation option
16583 is needed to use anonymous DH: it just needs to be included in the
16588 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16589 EVP_MD_type. The old functionality is available in a new macro called
16590 EVP_MD_md(). Change code that uses it and update docs.
16594 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16595 where the `void *` argument is replaced by a function pointer argument.
16596 Previously `void *` was abused to point to functions, which works on
16597 many platforms, but is not correct. As these functions are usually
16598 called by macros defined in OpenSSL header files, most source code
16599 should work without changes.
16603 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16604 sections with information on -D... compiler switches used for
16605 compiling the library so that applications can see them. To enable
16606 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16607 must be defined. E.g.,
16608 #define OPENSSL_ALGORITHM_DEFINES
16609 #include <openssl/opensslconf.h>
16610 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16612 *Richard Levitte, Ulf and Bodo Möller*
16614 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16619 * Change the 'other' type in certificate aux info to a STACK_OF
16620 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16621 the required ASN1 format: arbitrary types determined by an OID.
16625 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16626 argument to 'req'. This is not because the function is newer or
16627 better than others it just uses the work 'NEW' in the certificate
16628 request header lines. Some software needs this.
16632 * Reorganise password command line arguments: now passwords can be
16633 obtained from various sources. Delete the PEM_cb function and make
16634 it the default behaviour: i.e. if the callback is NULL and the
16635 usrdata argument is not NULL interpret it as a null terminated pass
16636 phrase. If usrdata and the callback are NULL then the pass phrase
16637 is prompted for as usual.
16641 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16642 the support is automatically enabled. The resulting binaries will
16643 autodetect the card and use it if present.
16645 *Ben Laurie and Compaq Inc.*
16647 * Work around for Netscape hang bug. This sends certificate request
16648 and server done in one record. Since this is perfectly legal in the
16649 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16650 the bugs/SSLv3 entry for more info.
16654 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16658 * Add -rand argument to smime and pkcs12 applications and read/write
16663 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16667 * Add command line password options to the remaining applications.
16671 * Bug fix for BN_div_recp() for numerators with an even number of
16676 * More tests in bntest.c, and changed test_bn output.
16680 * ./config recognizes MacOS X now.
16684 * Bug fix for BN_div() when the first words of num and divisor are
16685 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16689 * Add support for various broken PKCS#8 formats, and command line
16690 options to produce them.
16694 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16695 get temporary BIGNUMs from a BN_CTX.
16699 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16704 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16705 include a #define from the old name to the new. The original intent
16706 was that statically linked binaries could for example just call
16707 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16708 link with digests. This never worked because SSLeay_add_all_digests()
16709 and SSLeay_add_all_ciphers() were in the same source file so calling
16710 one would link with the other. They are now in separate source files.
16714 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16718 * Use a less unusual form of the Miller-Rabin primality test (it used
16719 a binary algorithm for exponentiation integrated into the Miller-Rabin
16720 loop, our standard modexp algorithms are faster).
16724 * Support for the EBCDIC character set completed.
16726 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16728 * Source code cleanups: use const where appropriate, eliminate casts,
16729 use `void *` instead of `char *` in lhash.
16733 * Bugfix: ssl3_send_server_key_exchange was not restartable
16734 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16735 this the server could overwrite ephemeral keys that the client
16740 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16741 using 50 iterations of the Rabin-Miller test.
16743 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16744 iterations of the Rabin-Miller test as required by the appendix
16745 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16746 As BN_is_prime_fasttest includes trial division, DSA parameter
16747 generation becomes much faster.
16749 This implies a change for the callback functions in DSA_is_prime
16750 and DSA_generate_parameters: The callback function is called once
16751 for each positive witness in the Rabin-Miller test, not just
16752 occasionally in the inner loop; and the parameters to the
16753 callback function now provide an iteration count for the outer
16754 loop rather than for the current invocation of the inner loop.
16755 DSA_generate_parameters additionally can call the callback
16756 function with an 'iteration count' of -1, meaning that a
16757 candidate has passed the trial division test (when q is generated
16758 from an application-provided seed, trial division is skipped).
16762 * New function BN_is_prime_fasttest that optionally does trial
16763 division before starting the Rabin-Miller test and has
16764 an additional BN_CTX * argument (whereas BN_is_prime always
16765 has to allocate at least one BN_CTX).
16766 'callback(1, -1, cb_arg)' is called when a number has passed the
16767 trial division stage.
16771 * Fix for bug in CRL encoding. The validity dates weren't being handled
16776 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16780 * New function BN_pseudo_rand().
16784 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16785 bignum version of BN_from_montgomery() with the working code from
16786 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16791 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16792 made it impossible to use the same SSL_SESSION data structure in
16793 SSL2 clients in multiple threads.
16797 * The return value of RAND_load_file() no longer counts bytes obtained
16798 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16799 to seed the PRNG (previously an explicit byte count was required).
16801 *Ulf Möller, Bodo Möller*
16803 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16804 used `char *` instead of `void *` and had casts all over the place.
16808 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16812 * Retain source code compatibility for BN_prime_checks macro:
16813 BN_is_prime(..., BN_prime_checks, ...) now uses
16814 BN_prime_checks_for_size to determine the appropriate number of
16815 Rabin-Miller iterations.
16819 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16820 DH_CHECK_P_NOT_SAFE_PRIME.
16821 (Check if this is true? OpenPGP calls them "strong".)
16825 * Merge the functionality of "dh" and "gendh" programs into a new program
16826 "dhparam". The old programs are retained for now but will handle DH keys
16827 (instead of parameters) in future.
16831 * Make the ciphers, s_server and s_client programs check the return values
16832 when a new cipher list is set.
16836 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16837 ciphers. Before when the 56bit ciphers were enabled the sorting was
16840 The syntax for the cipher sorting has been extended to support sorting by
16841 cipher-strength (using the strength_bits hard coded in the tables).
16842 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16844 Fix a bug in the cipher-command parser: when supplying a cipher command
16845 string with an "undefined" symbol (neither command nor alphanumeric
16846 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16847 an error is flagged.
16849 Due to the strength-sorting extension, the code of the
16850 ssl_create_cipher_list() function was completely rearranged. I hope that
16851 the readability was also increased :-)
16853 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16855 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16856 for the first serial number and places 2 in the serial number file. This
16857 avoids problems when the root CA is created with serial number zero and
16858 the first user certificate has the same issuer name and serial number
16863 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16864 the new code. Add documentation for this stuff.
16868 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16869 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16870 structures and behave in an analogous way to the X509v3 functions:
16871 they shouldn't be called directly but wrapper functions should be used
16874 So we also now have some wrapper functions that call the X509at functions
16875 when passed certificate requests. (TO DO: similar things can be done with
16876 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16877 things. Some of these need some d2i or i2d and print functionality
16878 because they handle more complex structures.)
16882 * Add missing #ifndefs that caused missing symbols when building libssl
16883 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16884 NO_RSA in `ssl/s2*.c`.
16886 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16888 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16889 has a return value which indicates the quality of the random data
16890 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16891 error queue. New function RAND_pseudo_bytes() generates output that is
16892 guaranteed to be unique but not unpredictable. RAND_add is like
16893 RAND_seed, but takes an extra argument for an entropy estimate
16894 (RAND_seed always assumes full entropy).
16898 * Do more iterations of Rabin-Miller probable prime test (specifically,
16899 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16900 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16901 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16902 false-positive rate of at most 2^-80 for random input.
16906 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16910 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16911 in the 0.9.5 release), this returns the chain
16912 from an X509_CTX structure with a dup of the stack and all
16913 the X509 reference counts upped: so the stack will exist
16914 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16917 Also make SSL_SESSION_print() print out the verify return
16922 * Add manpage for the pkcs12 command. Also change the default
16923 behaviour so MAC iteration counts are used unless the new
16924 -nomaciter option is used. This improves file security and
16925 only older versions of MSIE (4.0 for example) need it.
16929 * Honor the no-xxx Configure options when creating .DEF files.
16933 * Add PKCS#10 attributes to field table: challengePassword,
16934 unstructuredName and unstructuredAddress. These are taken from
16935 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16936 international characters are used.
16938 More changes to X509_ATTRIBUTE code: allow the setting of types
16939 based on strings. Remove the 'loc' parameter when adding
16940 attributes because these will be a SET OF encoding which is sorted
16945 * Initial changes to the 'req' utility to allow request generation
16946 automation. This will allow an application to just generate a template
16947 file containing all the field values and have req construct the
16950 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16951 used all over the place including certificate requests and PKCS#7
16952 structures. They are currently handled manually where necessary with
16953 some primitive wrappers for PKCS#7. The new functions behave in a
16954 manner analogous to the X509 extension functions: they allow
16955 attributes to be looked up by NID and added.
16957 Later something similar to the X509V3 code would be desirable to
16958 automatically handle the encoding, decoding and printing of the
16959 more complex types. The string types like challengePassword can
16960 be handled by the string table functions.
16962 Also modified the multi byte string table handling. Now there is
16963 a 'global mask' which masks out certain types. The table itself
16964 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16965 is useful when for example there is only one permissible type
16966 (as in countryName) and using the mask might result in no valid
16971 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16972 SSL_get_peer_finished to allow applications to obtain the latest
16973 Finished messages sent to the peer or expected from the peer,
16974 respectively. (SSL_get_peer_finished is usually the Finished message
16975 actually received from the peer, otherwise the protocol will be aborted.)
16977 As the Finished message are message digests of the complete handshake
16978 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16979 be used for external authentication procedures when the authentication
16980 provided by SSL/TLS is not desired or is not enough.
16984 * Enhanced support for Alpha Linux is added. Now ./config checks if
16985 the host supports BWX extension and if Compaq C is present on the
16986 $PATH. Just exploiting of the BWX extension results in 20-30%
16987 performance kick for some algorithms, e.g. DES and RC4 to mention
16988 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16993 * Add support for MS "fast SGC". This is arguably a violation of the
16994 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16995 weak crypto and after checking the certificate is SGC a second one
16996 with strong crypto. MS SGC stops the first handshake after receiving
16997 the server certificate message and sends a second client hello. Since
16998 a server will typically do all the time consuming operations before
16999 expecting any further messages from the client (server key exchange
17000 is the most expensive) there is little difference between the two.
17002 To get OpenSSL to support MS SGC we have to permit a second client
17003 hello message after we have sent server done. In addition we have to
17004 reset the MAC if we do get this second client hello.
17008 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17009 if a DER encoded private key is RSA or DSA traditional format. Changed
17010 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17011 format DER encoded private key. Newer code should use PKCS#8 format which
17012 has the key type encoded in the ASN1 structure. Added DER private key
17013 support to pkcs8 application.
17017 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17018 ciphersuites has been selected (as required by the SSL 3/TLS 1
17019 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17020 is set, we interpret this as a request to violate the specification
17021 (the worst that can happen is a handshake failure, and 'correct'
17022 behaviour would result in a handshake failure anyway).
17026 * In SSL_CTX_add_session, take into account that there might be multiple
17027 SSL_SESSION structures with the same session ID (e.g. when two threads
17028 concurrently obtain them from an external cache).
17029 The internal cache can handle only one SSL_SESSION with a given ID,
17030 so if there's a conflict, we now throw out the old one to achieve
17035 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17036 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17037 some routines that use cipher OIDs: some ciphers do not have OIDs
17038 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17043 * Simplify the trust setting structure and code. Now we just have
17044 two sequences of OIDs for trusted and rejected settings. These will
17045 typically have values the same as the extended key usage extension
17046 and any application specific purposes.
17048 The trust checking code now has a default behaviour: it will just
17049 check for an object with the same NID as the passed id. Functions can
17050 be provided to override either the default behaviour or the behaviour
17051 for a given id. SSL client, server and email already have functions
17052 in place for compatibility: they check the NID and also return "trusted"
17053 if the certificate is self signed.
17057 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17058 traditional format into an EVP_PKEY structure.
17062 * Add a password callback function PEM_cb() which either prompts for
17063 a password if usr_data is NULL or otherwise assumes it is a null
17064 terminated password. Allow passwords to be passed on command line
17065 environment or config files in a few more utilities.
17069 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17070 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17071 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17072 Update documentation.
17076 * Support for ASN1 "NULL" type. This could be handled before by using
17077 ASN1_TYPE but there wasn't any function that would try to read a NULL
17078 and produce an error if it couldn't. For compatibility we also have
17079 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17080 don't allocate anything because they don't need to.
17084 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17087 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17089 * Rebuild of the memory allocation routines used by OpenSSL code and
17090 possibly others as well. The purpose is to make an interface that
17091 provide hooks so anyone can build a separate set of allocation and
17092 deallocation routines to be used by OpenSSL, for example memory
17093 pool implementations, or something else, which was previously hard
17094 since Malloc(), Realloc() and Free() were defined as macros having
17095 the values malloc, realloc and free, respectively (except for Win32
17096 compilations). The same is provided for memory debugging code.
17097 OpenSSL already comes with functionality to find memory leaks, but
17098 this gives people a chance to debug other memory problems.
17100 With these changes, a new set of functions and macros have appeared:
17102 CRYPTO_set_mem_debug_functions() [F]
17103 CRYPTO_get_mem_debug_functions() [F]
17104 CRYPTO_dbg_set_options() [F]
17105 CRYPTO_dbg_get_options() [F]
17106 CRYPTO_malloc_debug_init() [M]
17108 The memory debug functions are NULL by default, unless the library
17109 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17110 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17111 gives the standard debugging functions that come with OpenSSL) or
17112 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17113 provided by the library user) must be used. When the standard
17114 debugging functions are used, CRYPTO_dbg_set_options can be used to
17115 request additional information:
17116 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17117 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17119 Also, things like CRYPTO_set_mem_functions will always give the
17120 expected result (the new set of functions is used for allocation
17121 and deallocation) at all times, regardless of platform and compiler
17124 To finish it up, some functions that were never use in any other
17125 way than through macros have a new API and new semantic:
17127 CRYPTO_dbg_malloc()
17128 CRYPTO_dbg_realloc()
17131 All macros of value have retained their old syntax.
17133 *Richard Levitte and Bodo Moeller*
17135 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17136 ordering of SMIMECapabilities wasn't in "strength order" and there
17137 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17142 * Some ASN1 types with illegal zero length encoding (INTEGER,
17143 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17145 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17147 * Merge in my S/MIME library for OpenSSL. This provides a simple
17148 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17149 functionality to handle multipart/signed properly) and a utility
17150 called 'smime' to call all this stuff. This is based on code I
17151 originally wrote for Celo who have kindly allowed it to be
17152 included in OpenSSL.
17156 * Add variants des_set_key_checked and des_set_key_unchecked of
17157 des_set_key (aka des_key_sched). Global variable des_check_key
17158 decides which of these is called by des_set_key; this way
17159 des_check_key behaves as it always did, but applications and
17160 the library itself, which was buggy for des_check_key == 1,
17161 have a cleaner way to pick the version they need.
17165 * New function PKCS12_newpass() which changes the password of a
17170 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17171 dynamic mix. In both cases the ids can be used as an index into the
17172 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17173 functions so they accept a list of the field values and the
17174 application doesn't need to directly manipulate the X509_TRUST
17179 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17184 * Modify the way the V3 extension code looks up extensions. This now
17185 works in a similar way to the object code: we have some "standard"
17186 extensions in a static table which is searched with OBJ_bsearch()
17187 and the application can add dynamic ones if needed. The file
17188 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17189 updated whenever a new extension is added to the core code and kept
17190 in ext_nid order. There is a simple program 'tabtest.c' which checks
17191 this. New extensions are not added too often so this file can readily
17192 be maintained manually.
17194 There are two big advantages in doing things this way. The extensions
17195 can be looked up immediately and no longer need to be "added" using
17196 X509V3_add_standard_extensions(): this function now does nothing.
17197 Side note: I get *lots* of email saying the extension code doesn't
17198 work because people forget to call this function.
17199 Also no dynamic allocation is done unless new extensions are added:
17200 so if we don't add custom extensions there is no need to call
17201 X509V3_EXT_cleanup().
17205 * Modify enc utility's salting as follows: make salting the default. Add a
17206 magic header, so unsalted files fail gracefully instead of just decrypting
17207 to garbage. This is because not salting is a big security hole, so people
17208 should be discouraged from doing it.
17212 * Fixes and enhancements to the 'x509' utility. It allowed a message
17213 digest to be passed on the command line but it only used this
17214 parameter when signing a certificate. Modified so all relevant
17215 operations are affected by the digest parameter including the
17216 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17217 DSA key was used because it didn't fix the digest.
17221 * Initial certificate chain verify code. Currently tests the untrusted
17222 certificates for consistency with the verify purpose (which is set
17223 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17225 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17226 this is because it will reject chains with invalid extensions whereas
17227 every previous version of OpenSSL and SSLeay made no checks at all.
17229 Trust code: checks the root CA for the relevant trust settings. Trust
17230 settings have an initial value consistent with the verify purpose: e.g.
17231 if the verify purpose is for SSL client use it expects the CA to be
17232 trusted for SSL client use. However the default value can be changed to
17233 permit custom trust settings: one example of this would be to only trust
17234 certificates from a specific "secure" set of CAs.
17236 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17237 which should be used for version portability: especially since the
17238 verify structure is likely to change more often now.
17240 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17241 to set them. If not set then assume SSL clients will verify SSL servers
17244 Two new options to the verify program: -untrusted allows a set of
17245 untrusted certificates to be passed in and -purpose which sets the
17246 intended purpose of the certificate. If a purpose is set then the
17247 new chain verify code is used to check extension consistency.
17251 * Support for the authority information access extension.
17255 * Modify RSA and DSA PEM read routines to transparently handle
17256 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17257 public keys in a format compatible with certificate
17258 SubjectPublicKeyInfo structures. Unfortunately there were already
17259 functions called *_PublicKey_* which used various odd formats so
17260 these are retained for compatibility: however the DSA variants were
17261 never in a public release so they have been deleted. Changed dsa/rsa
17262 utilities to handle the new format: note no releases ever handled public
17263 keys so we should be OK.
17265 The primary motivation for this change is to avoid the same fiasco
17266 that dogs private keys: there are several incompatible private key
17267 formats some of which are standard and some OpenSSL specific and
17268 require various evil hacks to allow partial transparent handling and
17269 even then it doesn't work with DER formats. Given the option anything
17270 other than PKCS#8 should be dumped: but the other formats have to
17271 stay in the name of compatibility.
17273 With public keys and the benefit of hindsight one standard format
17274 is used which works with EVP_PKEY, RSA or DSA structures: though
17275 it clearly returns an error if you try to read the wrong kind of key.
17277 Added a -pubkey option to the 'x509' utility to output the public key.
17278 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17279 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17280 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17281 that do the same as the `EVP_PKEY_assign_*()` except they up the
17282 reference count of the added key (they don't "swallow" the
17287 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17288 CRLs would fail if the file contained no certificates or no CRLs:
17289 added a new function to read in both types and return the number
17290 read: this means that if none are read it will be an error. The
17291 DER versions of the certificate and CRL reader would always fail
17292 because it isn't possible to mix certificates and CRLs in DER format
17293 without choking one or the other routine. Changed this to just read
17294 a certificate: this is the best we can do. Also modified the code
17295 in `apps/verify.c` to take notice of return codes: it was previously
17296 attempting to read in certificates from NULL pointers and ignoring
17297 any errors: this is one reason why the cert and CRL reader seemed
17298 to work. It doesn't check return codes from the default certificate
17299 routines: these may well fail if the certificates aren't installed.
17303 * Code to support otherName option in GeneralName.
17307 * First update to verify code. Change the verify utility
17308 so it warns if it is passed a self signed certificate:
17309 for consistency with the normal behaviour. X509_verify
17310 has been modified to it will now verify a self signed
17311 certificate if *exactly* the same certificate appears
17312 in the store: it was previously impossible to trust a
17313 single self signed certificate. This means that:
17314 openssl verify ss.pem
17315 now gives a warning about a self signed certificate but
17316 openssl verify -CAfile ss.pem ss.pem
17321 * For servers, store verify_result in SSL_SESSION data structure
17322 (and add it to external session representation).
17323 This is needed when client certificate verifications fails,
17324 but an application-provided verification callback (set by
17325 SSL_CTX_set_cert_verify_callback) allows accepting the session
17326 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17327 but returns 1): When the session is reused, we have to set
17328 ssl->verify_result to the appropriate error code to avoid
17331 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17333 * Fix a bug in the new PKCS#7 code: it didn't consider the
17334 case in PKCS7_dataInit() where the signed PKCS7 structure
17335 didn't contain any existing data because it was being created.
17337 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17339 * Add a salt to the key derivation routines in enc.c. This
17340 forms the first 8 bytes of the encrypted file. Also add a
17341 -S option to allow a salt to be input on the command line.
17345 * New function X509_cmp(). Oddly enough there wasn't a function
17346 to compare two certificates. We do this by working out the SHA1
17347 hash and comparing that. X509_cmp() will be needed by the trust
17352 * SSL_get1_session() is like SSL_get_session(), but increments
17353 the reference count in the SSL_SESSION returned.
17355 *Geoff Thorpe <geoff@eu.c2.net>*
17357 * Fix for 'req': it was adding a null to request attributes.
17358 Also change the X509_LOOKUP and X509_INFO code to handle
17359 certificate auxiliary information.
17363 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17368 * Add the possibility to add extra information to the memory leak
17369 detecting output, to form tracebacks, showing from where each
17370 allocation was originated: CRYPTO_push_info("constant string") adds
17371 the string plus current file name and line number to a per-thread
17372 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17373 is like calling CYRPTO_pop_info() until the stack is empty.
17374 Also updated memory leak detection code to be multi-thread-safe.
17378 * Add options -text and -noout to pkcs7 utility and delete the
17379 encryption options which never did anything. Update docs.
17383 * Add options to some of the utilities to allow the pass phrase
17384 to be included on either the command line (not recommended on
17385 OSes like Unix) or read from the environment. Update the
17386 manpages and fix a few bugs.
17390 * Add a few manpages for some of the openssl commands.
17394 * Fix the -revoke option in ca. It was freeing up memory twice,
17395 leaking and not finding already revoked certificates.
17399 * Extensive changes to support certificate auxiliary information.
17400 This involves the use of X509_CERT_AUX structure and X509_AUX
17401 functions. An X509_AUX function such as PEM_read_X509_AUX()
17402 can still read in a certificate file in the usual way but it
17403 will also read in any additional "auxiliary information". By
17404 doing things this way a fair degree of compatibility can be
17405 retained: existing certificates can have this information added
17406 using the new 'x509' options.
17408 Current auxiliary information includes an "alias" and some trust
17409 settings. The trust settings will ultimately be used in enhanced
17410 certificate chain verification routines: currently a certificate
17411 can only be trusted if it is self signed and then it is trusted
17416 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17417 The problem was that one of the replacement routines had not been working
17418 since SSLeay releases. For now the offending routine has been replaced
17419 with non-optimised assembler. Even so, this now gives around 95%
17420 performance improvement for 1024 bit RSA signs.
17424 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17425 handling. Most clients have the effective key size in bits equal to
17426 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17427 A few however don't do this and instead use the size of the decrypted key
17428 to determine the RC2 key length and the AlgorithmIdentifier to determine
17429 the effective key length. In this case the effective key length can still
17430 be 40 bits but the key length can be 168 bits for example. This is fixed
17431 by manually forcing an RC2 key into the EVP_PKEY structure because the
17432 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17433 the key length and effective key length are equal.
17437 * Add a bunch of functions that should simplify the creation of
17438 X509_NAME structures. Now you should be able to do:
17439 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17440 and have it automatically work out the correct field type and fill in
17441 the structures. The more adventurous can try:
17442 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17443 and it will (hopefully) work out the correct multibyte encoding.
17447 * Change the 'req' utility to use the new field handling and multibyte
17448 copy routines. Before the DN field creation was handled in an ad hoc
17449 way in req, ca, and x509 which was rather broken and didn't support
17450 BMPStrings or UTF8Strings. Since some software doesn't implement
17451 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17452 using the dirstring_type option. See the new comment in the default
17453 openssl.cnf for more info.
17457 * Make crypto/rand/md_rand.c more robust:
17458 - Assure unique random numbers after fork().
17459 - Make sure that concurrent threads access the global counter and
17460 md serializably so that we never lose entropy in them
17461 or use exactly the same state in multiple threads.
17462 Access to the large state is not always serializable because
17463 the additional locking could be a performance killer, and
17464 md should be large enough anyway.
17468 * New file `apps/app_rand.c` with commonly needed functionality
17469 for handling the random seed file.
17471 Use the random seed file in some applications that previously did not:
17473 dsaparam -genkey (which also ignored its '-rand' option),
17476 x509 (when signing).
17477 Except on systems with /dev/urandom, it is crucial to have a random
17478 seed file at least for key creation, DSA signing, and for DH exchanges;
17479 for RSA signatures we could do without one.
17481 gendh and gendsa (unlike genrsa) used to read only the first byte
17482 of each file listed in the '-rand' option. The function as previously
17483 found in genrsa is now in app_rand.c and is used by all programs
17484 that support '-rand'.
17488 * In RAND_write_file, use mode 0600 for creating files;
17489 don't just chmod when it may be too late.
17493 * Report an error from X509_STORE_load_locations
17494 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17498 * New function ASN1_mbstring_copy() this copies a string in either
17499 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17500 into an ASN1_STRING type. A mask of permissible types is passed
17501 and it chooses the "minimal" type to use or an error if not type
17506 * Add function equivalents to the various macros in asn1.h. The old
17507 macros are retained with an `M_` prefix. Code inside the library can
17508 use the `M_` macros. External code (including the openssl utility)
17509 should *NOT* in order to be "shared library friendly".
17513 * Add various functions that can check a certificate's extensions
17514 to see if it usable for various purposes such as SSL client,
17515 server or S/MIME and CAs of these types. This is currently
17516 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17517 verification. Also added a -purpose flag to x509 utility to
17518 print out all the purposes.
17522 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17527 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17528 for, obtain and decode and extension and obtain its critical flag.
17529 This allows all the necessary extension code to be handled in a
17530 single function call.
17534 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17535 platforms. See crypto/rc4/rc4_enc.c for further details.
17539 * New -noout option to asn1parse. This causes no output to be produced
17540 its main use is when combined with -strparse and -out to extract data
17541 from a file (which may not be in ASN.1 format).
17545 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17546 when producing the local key id.
17548 *Richard Levitte <levitte@stacken.kth.se>*
17550 * New option -dhparam in s_server. This allows a DH parameter file to be
17551 stated explicitly. If it is not stated then it tries the first server
17552 certificate file. The previous behaviour hard coded the filename
17557 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17558 a public key to be input or output. For example:
17559 openssl rsa -in key.pem -pubout -out pubkey.pem
17560 Also added necessary DSA public key functions to handle this.
17564 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17565 in the message. This was handled by allowing
17566 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17568 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17570 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17571 to the end of the strings whereas this didn't. This would cause problems
17572 if strings read with d2i_ASN1_bytes() were later modified.
17574 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17576 * Fix for base64 decode bug. When a base64 bio reads only one line of
17577 data and it contains EOF it will end up returning an error. This is
17578 caused by input 46 bytes long. The cause is due to the way base64
17579 BIOs find the start of base64 encoded data. They do this by trying a
17580 trial decode on each line until they find one that works. When they
17581 do a flag is set and it starts again knowing it can pass all the
17582 data directly through the decoder. Unfortunately it doesn't reset
17583 the context it uses. This means that if EOF is reached an attempt
17584 is made to pass two EOFs through the context and this causes the
17585 resulting error. This can also cause other problems as well. As is
17586 usual with these problems it takes *ages* to find and the fix is
17587 trivial: move one line.
17589 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17591 * Ugly workaround to get s_client and s_server working under Windows. The
17592 old code wouldn't work because it needed to select() on sockets and the
17593 tty (for keypresses and to see if data could be written). Win32 only
17594 supports select() on sockets so we select() with a 1s timeout on the
17595 sockets and then see if any characters are waiting to be read, if none
17596 are present then we retry, we also assume we can always write data to
17597 the tty. This isn't nice because the code then blocks until we've
17598 received a complete line of data and it is effectively polling the
17599 keyboard at 1s intervals: however it's quite a bit better than not
17600 working at all :-) A dedicated Windows application might handle this
17601 with an event loop for example.
17605 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17606 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17607 will be called when RSA_sign() and RSA_verify() are used. This is useful
17608 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17609 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17610 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17611 This necessitated the support of an extra signature type NID_md5_sha1
17612 for SSL signatures and modifications to the SSL library to use it instead
17613 of calling RSA_public_decrypt() and RSA_private_encrypt().
17617 * Add new -verify -CAfile and -CApath options to the crl program, these
17618 will lookup a CRL issuers certificate and verify the signature in a
17619 similar way to the verify program. Tidy up the crl program so it
17620 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17621 less strict. It will now permit CRL extensions even if it is not
17622 a V2 CRL: this will allow it to tolerate some broken CRLs.
17626 * Initialize all non-automatic variables each time one of the openssl
17627 sub-programs is started (this is necessary as they may be started
17628 multiple times from the "OpenSSL>" prompt).
17630 *Lennart Bang, Bodo Moeller*
17632 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17633 removing all other RSA functionality (this is what NO_RSA does). This
17634 is so (for example) those in the US can disable those operations covered
17635 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17640 * Non-copying interface to BIO pairs.
17641 (still largely untested)
17645 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17646 ASCII string. This was handled independently in various places before.
17650 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17651 UTF8 strings a character at a time.
17655 * Use client_version from client hello to select the protocol
17656 (s23_srvr.c) and for RSA client key exchange verification
17657 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17661 * Add various utility functions to handle SPKACs, these were previously
17662 handled by poking round in the structure internals. Added new function
17663 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17664 print, verify and generate SPKACs. Based on an original idea from
17665 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17669 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17673 * Allow the config file extension section to be overwritten on the
17674 command line. Based on an original idea from Massimiliano Pala
17675 <madwolf@comune.modena.it>. The new option is called -extensions
17676 and can be applied to ca, req and x509. Also -reqexts to override
17677 the request extensions in req and -crlexts to override the crl extensions
17682 * Add new feature to the SPKAC handling in ca. Now you can include
17683 the same field multiple times by preceding it by "XXXX." for example:
17686 this is the same syntax as used in the req config file.
17690 * Allow certificate extensions to be added to certificate requests. These
17691 are specified in a 'req_extensions' option of the req section of the
17692 config file. They can be printed out with the -text option to req but
17693 are otherwise ignored at present.
17697 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17698 data read consists of only the final block it would not decrypted because
17699 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17700 A misplaced 'break' also meant the decrypted final block might not be
17701 copied until the next read.
17705 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17706 a few extra parameters to the DH structure: these will be useful if
17707 for example we want the value of 'q' or implement X9.42 DH.
17711 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17712 provides hooks that allow the default DSA functions or functions on a
17713 "per key" basis to be replaced. This allows hardware acceleration and
17714 hardware key storage to be handled without major modification to the
17715 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17716 associated functions.
17720 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17721 as "read only": it can't be written to and the buffer it points to will
17722 not be freed. Reading from a read only BIO is much more efficient than
17723 a normal memory BIO. This was added because there are several times when
17724 an area of memory needs to be read from a BIO. The previous method was
17725 to create a memory BIO and write the data to it, this results in two
17726 copies of the data and an O(n^2) reading algorithm. There is a new
17727 function BIO_new_mem_buf() which creates a read only memory BIO from
17728 an area of memory. Also modified the PKCS#7 routines to use read only
17733 * Bugfix: ssl23_get_client_hello did not work properly when called in
17734 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17735 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17736 but a retry condition occurred while trying to read the rest.
17740 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17741 NID_pkcs7_encrypted by default: this was wrong since this should almost
17742 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17743 the encrypted data type: this is a more sensible place to put it and it
17744 allows the PKCS#12 code to be tidied up that duplicated this
17749 * Changed obj_dat.pl script so it takes its input and output files on
17750 the command line. This should avoid shell escape redirection problems
17755 * Initial support for certificate extension requests, these are included
17756 in things like Xenroll certificate requests. Included functions to allow
17757 extensions to be obtained and added.
17761 * -crlf option to s_client and s_server for sending newlines as
17762 CRLF (as required by many protocols).
17766 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17768 * Install libRSAglue.a when OpenSSL is built with RSAref.
17770 *Ralf S. Engelschall*
17772 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17774 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17776 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17781 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17782 DH parameters/keys (q is lost during that conversion, but the resulting
17783 DH parameters contain its length).
17785 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17786 much faster than DH_generate_parameters (which creates parameters
17787 where `p = 2*q + 1`), and also the smaller q makes DH computations
17788 much more efficient (160-bit exponentiation instead of 1024-bit
17789 exponentiation); so this provides a convenient way to support DHE
17790 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17791 utter importance to use
17792 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17794 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17795 when such DH parameters are used, because otherwise small subgroup
17796 attacks may become possible!
17800 * Avoid memory leak in i2d_DHparams.
17804 * Allow the -k option to be used more than once in the enc program:
17805 this allows the same encrypted message to be read by multiple recipients.
17809 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17810 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17811 it will always use the numerical form of the OID, even if it has a short
17816 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17817 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17818 otherwise bn_mod_exp was called. In the case of hardware keys for example
17819 no private key components need be present and it might store extra data
17820 in the RSA structure, which cannot be accessed from bn_mod_exp.
17821 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17822 private key operations.
17826 * Added support for SPARC Linux.
17830 * pem_password_cb function type incompatibly changed from
17831 typedef int pem_password_cb(char *buf, int size, int rwflag);
17833 ....(char *buf, int size, int rwflag, void *userdata);
17834 so that applications can pass data to their callbacks:
17835 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17836 additional void * argument, which is just handed through whenever
17837 the password callback is called.
17839 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17841 New function SSL_CTX_set_default_passwd_cb_userdata.
17843 Compatibility note: As many C implementations push function arguments
17844 onto the stack in reverse order, the new library version is likely to
17845 interoperate with programs that have been compiled with the old
17846 pem_password_cb definition (PEM_whatever takes some data that
17847 happens to be on the stack as its last argument, and the callback
17848 just ignores this garbage); but there is no guarantee whatsoever that
17851 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17852 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17853 problems not only on Windows, but also on some Unix platforms.
17854 To avoid problematic command lines, these definitions are now in an
17855 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17856 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17860 * MIPS III/IV assembler module is reimplemented.
17864 * More DES library cleanups: remove references to srand/rand and
17865 delete an unused file.
17869 * Add support for the free Netwide assembler (NASM) under Win32,
17870 since not many people have MASM (ml) and it can be hard to obtain.
17871 This is currently experimental but it seems to work OK and pass all
17872 the tests. Check out INSTALL.W32 for info.
17876 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17877 without temporary keys kept an extra copy of the server key,
17878 and connections with temporary keys did not free everything in case
17883 * New function RSA_check_key and new openssl rsa option -check
17884 for verifying the consistency of RSA keys.
17886 *Ulf Moeller, Bodo Moeller*
17888 * Various changes to make Win32 compile work:
17889 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17890 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17891 comparison" warnings.
17892 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17896 * Add a debugging option to PKCS#5 v2 key generation function: when
17897 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17898 derived keys are printed to stderr.
17902 * Copy the flags in ASN1_STRING_dup().
17904 *Roman E. Pavlov <pre@mo.msk.ru>*
17906 * The x509 application mishandled signing requests containing DSA
17907 keys when the signing key was also DSA and the parameters didn't match.
17909 It was supposed to omit the parameters when they matched the signing key:
17910 the verifying software was then supposed to automatically use the CA's
17911 parameters if they were absent from the end user certificate.
17913 Omitting parameters is no longer recommended. The test was also
17914 the wrong way round! This was probably due to unusual behaviour in
17915 EVP_cmp_parameters() which returns 1 if the parameters match.
17916 This meant that parameters were omitted when they *didn't* match and
17917 the certificate was useless. Certificates signed with 'ca' didn't have
17920 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17922 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17923 The interface is as follows:
17924 Applications can use
17925 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17926 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17927 "off" is now the default.
17928 The library internally uses
17929 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17930 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17931 to disable memory-checking temporarily.
17933 Some inconsistent states that previously were possible (and were
17934 even the default) are now avoided.
17936 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17937 with each memory chunk allocated; this is occasionally more helpful
17938 than just having a counter.
17940 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17942 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17947 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17948 which largely parallels "options", but is for changing API behaviour,
17949 whereas "options" are about protocol behaviour.
17950 Initial "mode" flags are:
17952 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17953 a single record has been written.
17954 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17955 retries use the same buffer location.
17956 (But all of the contents must be
17961 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17964 * Fix problems with no-hmac etc.
17966 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17968 * New functions RSA_get_default_method(), RSA_set_method() and
17969 RSA_get_method(). These allows replacement of RSA_METHODs without having
17970 to mess around with the internals of an RSA structure.
17974 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17975 Also really enable memory leak checks in openssl.c and in some
17978 *Chad C. Mulligan, Bodo Moeller*
17980 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17981 up the length of negative integers. This has now been simplified to just
17982 store the length when it is first determined and use it later, rather
17983 than trying to keep track of where data is copied and updating it to
17985 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17987 * Add a new function PKCS7_signatureVerify. This allows the verification
17988 of a PKCS#7 signature but with the signing certificate passed to the
17989 function itself. This contrasts with PKCS7_dataVerify which assumes the
17990 certificate is present in the PKCS#7 structure. This isn't always the
17991 case: certificates can be omitted from a PKCS#7 structure and be
17992 distributed by "out of band" means (such as a certificate database).
17996 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17997 function prototypes in pem.h, also change util/mkdef.pl to add the
17998 necessary function names.
18002 * mk1mf.pl (used by Windows builds) did not properly read the
18003 options set by Configure in the top level Makefile, and Configure
18004 was not even able to write more than one option correctly.
18005 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18009 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18010 file to be loaded from a BIO or FILE pointer. The BIO version will
18011 for example allow memory BIOs to contain config info.
18015 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18016 Whoever hopes to achieve shared-library compatibility across versions
18017 must use this, not the compile-time macro.
18018 (Exercise 0.9.4: Which is the minimum library version required by
18020 Note: All this applies only to multi-threaded programs, others don't
18025 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18026 through a BIO pair triggered the default case, i.e.
18027 SSLerr(...,SSL_R_UNKNOWN_STATE).
18031 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18032 can use the SSL library even if none of the specific BIOs is
18037 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18038 for the encoded length.
18040 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18042 * Add initial documentation of the X509V3 functions.
18046 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18047 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18048 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18049 secure PKCS#8 private key format with a high iteration count.
18053 * Fix determination of Perl interpreter: A perl or perl5
18054 *directory* in $PATH was also accepted as the interpreter.
18056 *Ralf S. Engelschall*
18058 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18059 wrong with it but it was very old and did things like calling
18060 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18061 unusual formatting.
18065 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18066 to use the new extension code.
18070 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18071 with macros. This should make it easier to change their form, add extra
18072 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18077 * Add to configuration table a new entry that can specify an alternative
18078 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18079 according to Mark Crispin <MRC@Panda.COM>.
18083 * DES CBC did not update the IV. Weird.
18087 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18088 Changing the behaviour of the former might break existing programs --
18089 where IV updating is needed, des_ncbc_encrypt can be used.
18092 * When bntest is run from "make test" it drives bc to check its
18093 calculations, as well as internally checking them. If an internal check
18094 fails, it needs to cause bc to give a non-zero result or make test carries
18095 on without noticing the failure. Fixed.
18099 * DES library cleanups.
18103 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18104 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18105 ciphers. NOTE: although the key derivation function has been verified
18106 against some published test vectors it has not been extensively tested
18107 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18112 * Instead of "mkdir -p", which is not fully portable, use new
18113 Perl script "util/mkdir-p.pl".
18117 * Rewrite the way password based encryption (PBE) is handled. It used to
18118 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18119 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18120 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18121 the 'parameter' field of the AlgorithmIdentifier is passed to the
18122 underlying key generation function so it must do its own ASN1 parsing.
18123 This has also changed the EVP_PBE_CipherInit() function which now has a
18124 'parameter' argument instead of literal salt and iteration count values
18125 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18129 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18130 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18131 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18132 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18133 value was just used as a "magic string" and not used directly its
18134 value doesn't matter.
18138 * Introduce some semblance of const correctness to BN. Shame C doesn't
18143 * "linux-sparc64" configuration (ultrapenguin).
18145 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18146 "linux-sparc" configuration.
18148 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18150 * config now generates no-xxx options for missing ciphers.
18154 * Support the EBCDIC character set (work in progress).
18155 File ebcdic.c not yet included because it has a different license.
18157 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18159 * Support BS2000/OSD-POSIX.
18161 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18163 * Make callbacks for key generation use `void *` instead of `char *`.
18167 * Make S/MIME samples compile (not yet tested).
18171 * Additional typesafe stacks.
18175 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18179 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18181 * New configuration variant "sco5-gcc".
18183 * Updated some demos.
18185 *Sean O Riordain, Wade Scholine*
18187 * Add missing BIO_free at exit of pkcs12 application.
18191 * Fix memory leak in conf.c.
18195 * Updates for Win32 to assembler version of MD5.
18199 * Set #! path to perl in `apps/der_chop` to where we found it
18200 instead of using a fixed path.
18204 * SHA library changes for irix64-mips4-cc.
18208 * Improvements for VMS support.
18212 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18214 * Bignum library bug fix. IRIX 6 passes "make test" now!
18215 This also avoids the problems with SC4.2 and unpatched SC5.
18217 *Andy Polyakov <appro@fy.chalmers.se>*
18219 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18220 These are required because of the typesafe stack would otherwise break
18221 existing code. If old code used a structure member which used to be STACK
18222 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18223 sk_num or sk_value it would produce an error because the num, data members
18224 are not present in STACK_OF. Now it just produces a warning. sk_set
18225 replaces the old method of assigning a value to sk_value
18226 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18227 that does this will no longer work (and should use sk_set instead) but
18228 this could be regarded as a "questionable" behaviour anyway.
18232 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18233 correctly handle encrypted S/MIME data.
18237 * Change type of various DES function arguments from des_cblock
18238 (which means, in function argument declarations, pointer to char)
18239 to des_cblock * (meaning pointer to array with 8 char elements),
18240 which allows the compiler to do more typechecking; it was like
18241 that back in SSLeay, but with lots of ugly casts.
18243 Introduce new type const_des_cblock.
18247 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18248 problems: find RecipientInfo structure that matches recipient certificate
18249 and initialise the ASN1 structures properly based on passed cipher.
18253 * Belatedly make the BN tests actually check the results.
18257 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18258 to and from BNs: it was completely broken. New compilation option
18259 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18260 key elements as negative integers.
18264 * Reorganize and speed up MD5.
18266 *Andy Polyakov <appro@fy.chalmers.se>*
18270 *Richard Levitte <richard@levitte.org>*
18272 * New option -out to asn1parse to allow the parsed structure to be
18273 output to a file. This is most useful when combined with the -strparse
18274 option to examine the output of things like OCTET STRINGS.
18278 * Make SSL library a little more fool-proof by not requiring any longer
18279 that `SSL_set_{accept,connect}_state` be called before
18280 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18281 in many applications because usually everything *appeared* to work as
18282 intended anyway -- now it really works as intended).
18286 * Move openssl.cnf out of lib/.
18290 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18291 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18292 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18294 *Ralf S. Engelschall*
18296 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18297 handle PKCS#7 enveloped data properly.
18299 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18301 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18302 copying pointers. The cert_st handling is changed by this in
18303 various ways (and thus what used to be known as ctx->default_cert
18304 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18305 any longer when s->cert does not give us what we need).
18306 ssl_cert_instantiate becomes obsolete by this change.
18307 As soon as we've got the new code right (possibly it already is?),
18308 we have solved a couple of bugs of the earlier code where s->cert
18309 was used as if it could not have been shared with other SSL structures.
18311 Note that using the SSL API in certain dirty ways now will result
18312 in different behaviour than observed with earlier library versions:
18313 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18314 does not influence s as it used to.
18316 In order to clean up things more thoroughly, inside SSL_SESSION
18317 we don't use CERT any longer, but a new structure SESS_CERT
18318 that holds per-session data (if available); currently, this is
18319 the peer's certificate chain and, for clients, the server's certificate
18320 and temporary key. CERT holds only those values that can have
18321 meaningful defaults in an SSL_CTX.
18325 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18326 from the internal representation. Various PKCS#7 fixes: remove some
18327 evil casts and set the enc_dig_alg field properly based on the signing
18332 * Allow PKCS#12 password to be set from the command line or the
18333 environment. Let 'ca' get its config file name from the environment
18334 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18339 * Allow certificate policies extension to use an IA5STRING for the
18340 organization field. This is contrary to the PKIX definition but
18341 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18346 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18347 without disallowing inline assembler and the like for non-pedantic builds.
18351 * Support Borland C++ builder.
18353 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18359 * SHA-1 cleanups and performance enhancements.
18361 *Andy Polyakov <appro@fy.chalmers.se>*
18363 * Sparc v8plus assembler for the bignum library.
18365 *Andy Polyakov <appro@fy.chalmers.se>*
18367 * Accept any -xxx and +xxx compiler options in Configure.
18371 * Update HPUX configuration.
18375 * Add missing `sk_<type>_unshift()` function to safestack.h
18377 *Ralf S. Engelschall*
18379 * New function SSL_CTX_use_certificate_chain_file that sets the
18380 "extra_cert"s in addition to the certificate. (This makes sense
18381 only for "PEM" format files, as chains as a whole are not
18386 * Support verify_depth from the SSL API.
18387 x509_vfy.c had what can be considered an off-by-one-error:
18388 Its depth (which was not part of the external interface)
18389 was actually counting the number of certificates in a chain;
18390 now it really counts the depth.
18394 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18395 instead of X509err, which often resulted in confusing error
18396 messages since the error codes are not globally unique
18397 (e.g. an alleged error in ssl3_accept when a certificate
18398 didn't match the private key).
18400 * New function SSL_CTX_set_session_id_context that allows to set a default
18401 value (so that you don't need SSL_set_session_id_context for each
18402 connection using the SSL_CTX).
18406 * OAEP decoding bug fix.
18410 * Support INSTALL_PREFIX for package builders, as proposed by
18415 * New Configure options "threads" and "no-threads". For systems
18416 where the proper compiler options are known (currently Solaris
18417 and Linux), "threads" is the default.
18421 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18425 * Install various scripts to $(OPENSSLDIR)/misc, not to
18426 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18427 such as /usr/local/bin.
18431 * "make linux-shared" to build shared libraries.
18433 *Niels Poppe <niels@netbox.org>*
18435 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18439 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18440 extension adding in x509 utility.
18444 * Remove NOPROTO sections and error code comments.
18448 * Partial rewrite of the DEF file generator to now parse the ANSI
18453 * New Configure options --prefix=DIR and --openssldir=DIR.
18457 * Complete rewrite of the error code script(s). It is all now handled
18458 by one script at the top level which handles error code gathering,
18459 header rewriting and C source file generation. It should be much better
18460 than the old method: it now uses a modified version of Ulf's parser to
18461 read the ANSI prototypes in all header files (thus the old K&R definitions
18462 aren't needed for error creation any more) and do a better job of
18463 translating function codes into names. The old 'ASN1 error code embedded
18464 in a comment' is no longer necessary and it doesn't use .err files which
18465 have now been deleted. Also the error code call doesn't have to appear all
18466 on one line (which resulted in some large lines...).
18470 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18474 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18475 0 (which usually indicates a closed connection), but continue reading.
18479 * Fix some race conditions.
18483 * Add support for CRL distribution points extension. Add Certificate
18484 Policies and CRL distribution points documentation.
18488 * Move the autogenerated header file parts to crypto/opensslconf.h.
18492 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18493 8 of keying material. Merlin has also confirmed interop with this fix
18494 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18496 *Merlin Hughes <merlin@baltimore.ie>*
18498 * Fix lots of warnings.
18500 *Richard Levitte <levitte@stacken.kth.se>*
18502 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18503 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18505 *Richard Levitte <levitte@stacken.kth.se>*
18507 * Fix problems with sizeof(long) == 8.
18509 *Andy Polyakov <appro@fy.chalmers.se>*
18511 * Change functions to ANSI C.
18515 * Fix typos in error codes.
18517 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18519 * Remove defunct assembler files from Configure.
18523 * SPARC v8 assembler BIGNUM implementation.
18525 *Andy Polyakov <appro@fy.chalmers.se>*
18527 * Support for Certificate Policies extension: both print and set.
18528 Various additions to support the r2i method this uses.
18532 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18533 return a const string when you are expecting an allocated buffer.
18537 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18538 types DirectoryString and DisplayText.
18542 * Add code to allow r2i extensions to access the configuration database,
18543 add an LHASH database driver and add several ctx helper functions.
18547 * Fix an evil bug in bn_expand2() which caused various BN functions to
18548 fail when they extended the size of a BIGNUM.
18552 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18553 support typesafe stack.
18557 * Fix typo in SSL_[gs]et_options().
18559 *Nils Frostberg <nils@medcom.se>*
18561 * Delete various functions and files that belonged to the (now obsolete)
18562 old X509V3 handling code.
18566 * New Configure option "rsaref".
18570 * Don't auto-generate pem.h.
18574 * Introduce type-safe ASN.1 SETs.
18578 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18580 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18582 * Introduce type-safe STACKs. This will almost certainly break lots of code
18583 that links with OpenSSL (well at least cause lots of warnings), but fear
18584 not: the conversion is trivial, and it eliminates loads of evil casts. A
18585 few STACKed things have been converted already. Feel free to convert more.
18586 In the fullness of time, I'll do away with the STACK type altogether.
18590 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18591 specified in `<certfile>` by updating the entry in the index.txt file.
18592 This way one no longer has to edit the index.txt file manually for
18593 revoking a certificate. The -revoke option does the gory details now.
18595 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18597 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18598 `-text` option at all and this way the `-noout -text` combination was
18599 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18601 *Ralf S. Engelschall*
18603 * Make sure a corresponding plain text error message exists for the
18604 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18605 verify callback function determined that a certificate was revoked.
18607 *Ralf S. Engelschall*
18609 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18610 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18611 all available ciphers including rc5, which was forgotten until now.
18612 In order to let the testing shell script know which algorithms
18613 are available, a new (up to now undocumented) command
18614 `openssl list-cipher-commands` is used.
18618 * Bugfix: s_client occasionally would sleep in select() when
18619 it should have checked SSL_pending() first.
18623 * New functions DSA_do_sign and DSA_do_verify to provide access to
18624 the raw DSA values prior to ASN.1 encoding.
18628 * Tweaks to Configure
18630 *Niels Poppe <niels@netbox.org>*
18632 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18637 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18641 * New config option to avoid instructions that are illegal on the 80386.
18642 The default code is faster, but requires at least a 486.
18646 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18647 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18648 same as SSL2_VERSION anyway.
18652 * New "-showcerts" option for s_client.
18656 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18657 application. Various cleanups and fixes.
18661 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18662 modify error routines to work internally. Add error codes and PBE init
18663 to library startup routines.
18667 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18668 packing functions to asn1 and evp. Changed function names and error
18669 codes along the way.
18673 * PKCS12 integration: and so it begins... First of several patches to
18674 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18675 objects to objects.h
18679 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18680 and display support for Thawte strong extranet extension.
18684 * Add LinuxPPC support.
18686 *Jeff Dubrule <igor@pobox.org>*
18688 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18689 bn_div_words in alpha.s.
18691 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18693 * Make sure the RSA OAEP test is skipped under -DRSAref because
18694 OAEP isn't supported when OpenSSL is built with RSAref.
18696 *Ulf Moeller <ulf@fitug.de>*
18698 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18699 so they no longer are missing under -DNOPROTO.
18701 *Soren S. Jorvang <soren@t.dk>*
18703 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18705 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18706 doesn't work when the session is reused. Coming soon!
18710 * Fix a security hole, that allows sessions to be reused in the wrong
18711 context thus bypassing client cert protection! All software that uses
18712 client certs and session caches in multiple contexts NEEDS PATCHING to
18713 allow session reuse! A fuller solution is in the works.
18715 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18717 * Some more source tree cleanups (removed obsolete files
18718 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18719 permission on "config" script to be executable) and a fix for the INSTALL
18722 *Ulf Moeller <ulf@fitug.de>*
18724 * Remove some legacy and erroneous uses of malloc, free instead of
18727 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18729 * Make rsa_oaep_test return non-zero on error.
18731 *Ulf Moeller <ulf@fitug.de>*
18733 * Add support for native Solaris shared libraries. Configure
18734 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18735 if someone would make that last step automatic.
18737 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18739 * ctx_size was not built with the right compiler during "make links". Fixed.
18743 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18744 except NULL ciphers". This means the default cipher list will no longer
18745 enable NULL ciphers. They need to be specifically enabled e.g. with
18746 the string "DEFAULT:eNULL".
18750 * Fix to RSA private encryption routines: if p < q then it would
18751 occasionally produce an invalid result. This will only happen with
18752 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18756 * Be less restrictive and allow also `perl util/perlpath.pl
18757 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18758 because this way one can also use an interpreter named `perl5` (which is
18759 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18760 installed as `perl`).
18762 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18764 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18766 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18768 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18769 advapi32.lib to Win32 build and change the pem test comparison
18770 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18771 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18772 and crypto/des/ede_cbcm_enc.c.
18776 * DES quad checksum was broken on big-endian architectures. Fixed.
18780 * Comment out two functions in bio.h that aren't implemented. Fix up the
18781 Win32 test batch file so it (might) work again. The Win32 test batch file
18782 is horrible: I feel ill....
18786 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18787 in e_os.h. Audit of header files to check ANSI and non ANSI
18788 sections: 10 functions were absent from non ANSI section and not exported
18789 from Windows DLLs. Fixed up libeay.num for new functions.
18793 * Make `openssl version` output lines consistent.
18795 *Ralf S. Engelschall*
18797 * Fix Win32 symbol export lists for BIO functions: Added
18798 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18799 to ms/libeay{16,32}.def.
18801 *Ralf S. Engelschall*
18803 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18804 fine under Unix and passes some trivial tests I've now added. But the
18805 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18806 added to make sure no one expects that this stuff really works in the
18807 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18808 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18811 *Ralf S. Engelschall*
18813 * Fix the generation of two part addresses in perl.
18815 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18817 * Add config entry for Linux on MIPS.
18819 *John Tobey <jtobey@channel1.com>*
18821 * Make links whenever Configure is run, unless we are on Windoze.
18825 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18826 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18831 * Add a useful kludge to allow package maintainers to specify compiler and
18832 other platforms details on the command line without having to patch the
18833 Configure script every time: One now can use
18834 `perl Configure <id>:<details>`,
18835 i.e. platform ids are allowed to have details appended
18836 to them (separated by colons). This is treated as there would be a static
18837 pre-configured entry in Configure's %table under key `<id>` with value
18838 `<details>` and `perl Configure <id>` is called. So, when you want to
18839 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18840 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18841 now, which overrides the FreeBSD-elf entry on-the-fly.
18843 *Ralf S. Engelschall*
18845 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18849 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18850 on the `perl Configure ...` command line. This way one can compile
18851 OpenSSL libraries with Position Independent Code (PIC) which is needed
18852 for linking it into DSOs.
18854 *Ralf S. Engelschall*
18856 * Remarkably, export ciphers were totally broken and no-one had noticed!
18861 * Cleaned up the LICENSE document: The official contact for any license
18862 questions now is the OpenSSL core team under openssl-core@openssl.org.
18863 And add a paragraph about the dual-license situation to make sure people
18864 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18865 to the OpenSSL toolkit.
18867 *Ralf S. Engelschall*
18869 * General source tree makefile cleanups: Made `making xxx in yyy...`
18870 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18871 Additionally cleaned up the `make links` target: Remove unnecessary
18872 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18873 to speed processing and no longer clutter the display with confusing
18874 stuff. Instead only the actually done links are displayed.
18876 *Ralf S. Engelschall*
18878 * Permit null encryption ciphersuites, used for authentication only. It used
18879 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18880 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18885 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18886 signed attributes when verifying signatures (this would break them),
18887 the detached data encoding was wrong and public keys obtained using
18888 X509_get_pubkey() weren't freed.
18892 * Add text documentation for the BUFFER functions. Also added a work around
18893 to a Win95 console bug. This was triggered by the password read stuff: the
18894 last character typed gets carried over to the next fread(). If you were
18895 generating a new cert request using 'req' for example then the last
18896 character of the passphrase would be CR which would then enter the first
18901 * Added the new 'Includes OpenSSL Cryptography Software' button as
18902 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18903 button and can be used by applications based on OpenSSL to show the
18904 relationship to the OpenSSL project.
18906 *Ralf S. Engelschall*
18908 * Remove confusing variables in function signatures in files
18909 ssl/ssl_lib.c and ssl/ssl.h.
18911 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18913 * Don't install bss_file.c under PREFIX/include/
18915 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18917 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18918 functions that return function pointers and has support for NT specific
18919 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18920 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18921 unsigned to signed types: this was killing the Win32 compile.
18925 * Add new certificate file to stack functions,
18926 SSL_add_dir_cert_subjects_to_stack() and
18927 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18928 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18929 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18930 This means that Apache-SSL and similar packages don't have to mess around
18931 to add as many CAs as they want to the preferred list.
18935 * Experiment with doxygen documentation. Currently only partially applied to
18937 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18938 openssl.doxy as the configuration file.
18942 * Get rid of remaining C++-style comments which strict C compilers hate.
18944 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18946 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18947 compiled in by default: it has problems with large keys.
18951 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18952 DH private keys and/or callback functions which directly correspond to
18953 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18954 is needed for applications which have to configure certificates on a
18955 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18957 For the RSA certificate situation is makes no difference, but
18958 for the DSA certificate situation this fixes the "no shared cipher"
18959 problem where the OpenSSL cipher selection procedure failed because the
18960 temporary keys were not overtaken from the context and the API provided
18961 no way to reconfigure them.
18962 The new functions now let applications reconfigure the stuff and they
18963 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18964 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18965 non-public-API function ssl_cert_instantiate() is used as a helper
18966 function and also to reduce code redundancy inside ssl_rsa.c.
18968 *Ralf S. Engelschall*
18970 * Move s_server -dcert and -dkey options out of the undocumented feature
18971 area because they are useful for the DSA situation and should be
18972 recognized by the users.
18974 *Ralf S. Engelschall*
18976 * Fix the cipher decision scheme for export ciphers: the export bits are
18977 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18978 SSL_EXP_MASK. So, the original variable has to be used instead of the
18979 already masked variable.
18981 *Richard Levitte <levitte@stacken.kth.se>*
18983 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18985 *Richard Levitte <levitte@stacken.kth.se>*
18987 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18988 from `int` to `unsigned int` because it is a length and initialized by
18989 EVP_DigestFinal() which expects an `unsigned int *`.
18991 *Richard Levitte <levitte@stacken.kth.se>*
18993 * Don't hard-code path to Perl interpreter on shebang line of Configure
18994 script. Instead use the usual Shell->Perl transition trick.
18996 *Ralf S. Engelschall*
18998 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18999 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19000 -noout -modulus` as it's already the case for `openssl rsa -noout
19001 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19002 currently the public key is printed (a decision which was already done by
19003 `openssl dsa -modulus` in the past) which serves a similar purpose.
19004 Additionally the NO_RSA no longer completely removes the whole -modulus
19005 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19008 *Ralf S. Engelschall*
19010 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19011 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19013 *Arne Ansper <arne@ats.cyber.ee>*
19015 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19016 to be added. Now both 'req' and 'ca' can use new objects defined in the
19021 * Add cool BIO that does syslog (or event log on NT).
19023 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19025 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19026 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19027 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19028 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19032 * Add preliminary config info for new extension code.
19036 * Make RSA_NO_PADDING really use no padding.
19038 *Ulf Moeller <ulf@fitug.de>*
19040 * Generate errors when private/public key check is done.
19044 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19045 for some CRL extensions and new objects added.
19049 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19050 key usage extension and fuller support for authority key id.
19054 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19055 padding method for RSA, which is recommended for new applications in PKCS
19056 #1 v2.0 (RFC 2437, October 1998).
19057 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19058 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19059 against Bleichbacher's attack on RSA.
19060 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19063 * Updates to the new SSL compression code
19065 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19067 * Fix so that the version number in the master secret, when passed
19068 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19069 (because the server will not accept higher), that the version number
19070 is 0x03,0x01, not 0x03,0x00
19072 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19074 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19075 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19076 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19080 * Support for RAW extensions where an arbitrary extension can be
19081 created by including its DER encoding. See `apps/openssl.cnf` for
19086 * Make sure latest Perl versions don't interpret some generated C array
19087 code as Perl array code in the crypto/err/err_genc.pl script.
19089 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19091 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19092 not many people have the assembler. Various Win32 compilation fixes and
19093 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19094 build instructions.
19098 * Modify configure script 'Configure' to automatically create crypto/date.h
19099 file under Win32 and also build pem.h from pem.org. New script
19100 util/mkfiles.pl to create the MINFO file on environments that can't do a
19101 'make files': perl util/mkfiles.pl >MINFO should work.
19105 * Major rework of DES function declarations, in the pursuit of correctness
19106 and purity. As a result, many evil casts evaporated, and some weirdness,
19107 too. You may find this causes warnings in your code. Zapping your evil
19108 casts will probably fix them. Mostly.
19112 * Fix for a typo in asn1.h. Bug fix to object creation script
19113 obj_dat.pl. It considered a zero in an object definition to mean
19114 "end of object": none of the objects in objects.h have any zeros
19115 so it wasn't spotted.
19117 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19119 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19120 Masking (CBCM). In the absence of test vectors, the best I have been able
19121 to do is check that the decrypt undoes the encrypt, so far. Send me test
19122 vectors if you have them.
19126 * Correct calculation of key length for export ciphers (too much space was
19127 allocated for null ciphers). This has not been tested!
19131 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19132 message is now correct (it understands "crypto" and "ssl" on its
19133 command line). There is also now an "update" option. This will update
19134 the util/ssleay.num and util/libeay.num files with any new functions.
19136 perl util/mkdef.pl crypto ssl update
19137 it will update them.
19141 * Overhauled the Perl interface:
19142 - ported BN stuff to OpenSSL's different BN library
19143 - made the perl/ source tree CVS-aware
19144 - renamed the package from SSLeay to OpenSSL (the files still contain
19145 their history because I've copied them in the repository)
19146 - removed obsolete files (the test scripts will be replaced
19147 by better Test::Harness variants in the future)
19149 *Ralf S. Engelschall*
19151 * First cut for a very conservative source tree cleanup:
19152 1. merge various obsolete readme texts into doc/ssleay.txt
19153 where we collect the old documents and readme texts.
19154 2. remove the first part of files where I'm already sure that we no
19155 longer need them because of three reasons: either they are just temporary
19156 files which were left by Eric or they are preserved original files where
19157 I've verified that the diff is also available in the CVS via "cvs diff
19158 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19159 the crypto/md/ stuff).
19161 *Ralf S. Engelschall*
19163 * More extension code. Incomplete support for subject and issuer alt
19164 name, issuer and authority key id. Change the i2v function parameters
19165 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19166 what that's for :-) Fix to ASN1 macro which messed up
19167 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19171 * Preliminary support for ENUMERATED type. This is largely copied from the
19176 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19178 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19180 * Make sure `make rehash` target really finds the `openssl` program.
19182 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19184 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19185 like to hear about it if this slows down other processors.
19189 * Add CygWin32 platform information to Configure script.
19191 *Alan Batie <batie@aahz.jf.intel.com>*
19193 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19195 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19197 * New program nseq to manipulate netscape certificate sequences
19201 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19206 * Fixes to BN code. Previously the default was to define BN_RECURSION
19207 but the BN code had some problems that would cause failures when
19208 doing certificate verification and some other functions.
19210 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19212 * Add ASN1 and PEM code to support netscape certificate sequences.
19216 * Add ASN1 and PEM code to support netscape certificate sequences.
19220 * Add several PKIX and private extended key usage OIDs.
19224 * Modify the 'ca' program to handle the new extension code. Modify
19225 openssl.cnf for new extension format, add comments.
19229 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19230 and add a sample to openssl.cnf so req -x509 now adds appropriate
19235 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19236 error code, add initial support to X509_print() and x509 application.
19240 * Takes a deep breath and start adding X509 V3 extension support code. Add
19241 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19242 stuff is currently isolated and isn't even compiled yet.
19246 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19247 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19248 Removed the versions check from X509 routines when loading extensions:
19249 this allows certain broken certificates that don't set the version
19250 properly to be processed.
19254 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19255 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19256 can still be regenerated with "make depend".
19260 * Spelling mistake in C version of CAST-128.
19262 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19264 * Changes to the error generation code. The perl script err-code.pl
19265 now reads in the old error codes and retains the old numbers, only
19266 adding new ones if necessary. It also only changes the .err files if new
19267 codes are added. The makefiles have been modified to only insert errors
19268 when needed (to avoid needlessly modifying header files). This is done
19269 by only inserting errors if the .err file is newer than the auto generated
19270 C file. To rebuild all the error codes from scratch (the old behaviour)
19271 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19272 or delete all the .err files.
19276 * CAST-128 was incorrectly implemented for short keys. The C version has
19277 been fixed, but is untested. The assembler versions are also fixed, but
19278 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19279 to regenerate it if needed.
19280 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19281 Hagino <itojun@kame.net>*
19283 * File was opened incorrectly in randfile.c.
19285 *Ulf Möller <ulf@fitug.de>*
19287 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19288 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19289 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19290 al: it's just almost always a UTCTime. Note this patch adds new error
19291 codes so do a "make errors" if there are problems.
19295 * Correct Linux 1 recognition in config.
19297 *Ulf Möller <ulf@fitug.de>*
19299 * Remove pointless MD5 hash when using DSA keys in ca.
19301 *Anonymous <nobody@replay.com>*
19303 * Generate an error if given an empty string as a cert directory. Also
19304 generate an error if handed NULL (previously returned 0 to indicate an
19305 error, but didn't set one).
19307 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19309 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19313 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19314 parameters. This was causing a warning which killed off the Win32 compile.
19318 * Remove C++ style comments from crypto/bn/bn_local.h.
19320 *Neil Costigan <neil.costigan@celocom.com>*
19322 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19323 based on a text string, looking up short and long names and finally
19324 "dot" format. The "dot" format stuff didn't work. Added new function
19325 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19326 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19327 OID is not part of the table.
19331 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19332 X509_LOOKUP_by_alias().
19336 * Sort openssl functions by name.
19340 * Get the `gendsa` command working and add it to the `list` command. Remove
19341 encryption from sample DSA keys (in case anyone is interested the password
19346 * Make *all* `*_free` functions accept a NULL pointer.
19348 *Frans Heymans <fheymans@isaserver.be>*
19350 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19353 *Anonymous <nobody@replay.com>*
19355 * s_server should send the CAfile as acceptable CAs, not its own cert.
19357 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19359 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19361 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19363 * Temp key "for export" tests were wrong in s3_srvr.c.
19365 *Anonymous <nobody@replay.com>*
19367 * Add prototype for temp key callback functions
19368 SSL_CTX_set_tmp_{rsa,dh}_callback().
19372 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19373 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19377 * X509_name_add_entry() freed the wrong thing after an error.
19379 *Arne Ansper <arne@ats.cyber.ee>*
19381 * rsa_eay.c would attempt to free a NULL context.
19383 *Arne Ansper <arne@ats.cyber.ee>*
19385 * BIO_s_socket() had a broken should_retry() on Windoze.
19387 *Arne Ansper <arne@ats.cyber.ee>*
19389 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19391 *Arne Ansper <arne@ats.cyber.ee>*
19393 * Make sure the already existing X509_STORE->depth variable is initialized
19394 in X509_STORE_new(), but document the fact that this variable is still
19395 unused in the certificate verification process.
19397 *Ralf S. Engelschall*
19399 * Fix the various library and `apps/` files to free up pkeys obtained from
19400 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19404 * Fix reference counting in X509_PUBKEY_get(). This makes
19405 demos/maurice/example2.c work, amongst others, probably.
19407 *Steve Henson and Ben Laurie*
19409 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19410 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19411 are no longer created. This way we have a single and consistent command
19412 line interface `openssl <command>`, similar to `cvs <command>`.
19414 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19416 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19417 BIT STRING wrapper always have zero unused bits.
19421 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19425 * Make the top-level INSTALL documentation easier to understand.
19429 * Makefiles updated to exit if an error occurs in a sub-directory
19430 make (including if user presses ^C) [Paul Sutton]
19432 * Make Montgomery context stuff explicit in RSA data structure.
19436 * Fix build order of pem and err to allow for generated pem.h.
19440 * Fix renumbering bug in X509_NAME_delete_entry().
19444 * Enhanced the err-ins.pl script so it makes the error library number
19445 global and can add a library name. This is needed for external ASN1 and
19446 other error libraries.
19450 * Fixed sk_insert which never worked properly.
19454 * Fix ASN1 macros so they can handle indefinite length constructed
19455 EXPLICIT tags. Some non standard certificates use these: they can now
19460 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19461 into a single doc/ssleay.txt bundle. This way the information is still
19462 preserved but no longer messes up this directory. Now it's new room for
19463 the new set of documentation files.
19465 *Ralf S. Engelschall*
19467 * SETs were incorrectly DER encoded. This was a major pain, because they
19468 shared code with SEQUENCEs, which aren't coded the same. This means that
19469 almost everything to do with SETs or SEQUENCEs has either changed name or
19470 number of arguments.
19472 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19474 * Fix test data to work with the above.
19478 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19479 was already fixed by Eric for 0.9.1 it seems.
19481 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19483 * Autodetect FreeBSD3.
19487 * Fix various bugs in Configure. This affects the following platforms:
19491 unixware-2.0-pentium
19496 * Eliminate generated files from CVS. Reorder tests to regenerate files
19497 before they are needed.
19501 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19505 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19507 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19508 changed SSLeay to OpenSSL in version strings.
19510 *Ralf S. Engelschall*
19512 * Some fixups to the top-level documents.
19516 * Fixed the nasty bug where rsaref.h was not found under compile-time
19517 because the symlink to include/ was missing.
19519 *Ralf S. Engelschall*
19521 * Incorporated the popular no-RSA/DSA-only patches
19522 which allow to compile an RSA-free SSLeay.
19524 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19526 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19527 when "ssleay" is still not found.
19529 *Ralf S. Engelschall*
19531 * Added more platforms to Configure: Cray T3E, HPUX 11,
19533 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19535 * Updated the README file.
19537 *Ralf S. Engelschall*
19539 * Added various .cvsignore files in the CVS repository subdirs
19540 to make a "cvs update" really silent.
19542 *Ralf S. Engelschall*
19544 * Recompiled the error-definition header files and added
19545 missing symbols to the Win32 linker tables.
19547 *Ralf S. Engelschall*
19549 * Cleaned up the top-level documents;
19550 o new files: CHANGES and LICENSE
19551 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19552 o merged COPYRIGHT into LICENSE
19553 o removed obsolete TODO file
19554 o renamed MICROSOFT to INSTALL.W32
19556 *Ralf S. Engelschall*
19558 * Removed dummy files from the 0.9.1b source tree:
19559 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19560 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19561 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19562 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19563 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19565 *Ralf S. Engelschall*
19567 * Added various platform portability fixes.
19571 * The Genesis of the OpenSSL rpject:
19572 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19573 Young and Tim J. Hudson created while they were working for C2Net until
19576 *The OpenSSL Project*
19578 ### Changes between 0.9.0b and 0.9.1b [not released]
19580 * Updated a few CA certificates under certs/
19584 * Changed some BIGNUM api stuff.
19588 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19589 DGUX x86, Linux Alpha, etc.
19593 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19594 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19599 * Add -strparse option to asn1pars program which parses nested
19602 *Dr Stephen Henson <shenson@bigfoot.com>*
19604 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19608 * DSA fix for "ca" program.
19612 * Added "-genkey" option to "dsaparam" program.
19616 * Added RIPE MD160 (rmd160) message digest.
19620 * Added -a (all) option to "ssleay version" command.
19624 * Added PLATFORM define which is the id given to Configure.
19628 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19632 * Extended the ASN.1 parser routines.
19636 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19640 * Added a BN_CTX to the BN library.
19644 * Fixed the weak key values in DES library
19648 * Changed API in EVP library for cipher aliases.
19652 * Added support for RC2/64bit cipher.
19656 * Converted the lhash library to the crypto/mem.c functions.
19660 * Added more recognized ASN.1 object ids.
19664 * Added more RSA padding checks for SSL/TLS.
19668 * Added BIO proxy/filter functionality.
19672 * Added extra_certs to SSL_CTX which can be used
19673 send extra CA certificates to the client in the CA cert chain sending
19674 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19678 * Now Fortezza is denied in the authentication phase because
19679 this is key exchange mechanism is not supported by SSLeay at all.
19683 * Additional PKCS1 checks.
19687 * Support the string "TLSv1" for all TLS v1 ciphers.
19691 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19692 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19696 * Fixed a few memory leaks.
19700 * Fixed various code and comment typos.
19704 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19705 bytes sent in the client random.
19707 *Edward Bishop <ebishop@spyglass.com>*
19711 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
19712 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
19713 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19714 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19715 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19716 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19717 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19718 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19719 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19720 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19721 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19722 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19723 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19724 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19725 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19726 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19727 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19728 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19729 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19730 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19731 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19732 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19733 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19734 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19735 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19736 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19737 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19738 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19739 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19740 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19741 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19742 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19743 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19744 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19745 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19746 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19747 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19748 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19749 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19750 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19751 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19752 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19753 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19754 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19755 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19756 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19757 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19758 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19759 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19760 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19761 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19762 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19763 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19764 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19765 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19766 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19767 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19768 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19769 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19770 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19771 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19772 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19773 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19774 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19775 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19776 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19777 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19778 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19779 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19780 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19781 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19782 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19783 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19784 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19785 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19786 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19787 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19788 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19789 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19790 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19791 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19792 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19793 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19794 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19795 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19796 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19797 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19798 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19799 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19800 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19801 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19802 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19803 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19804 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19805 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19806 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19807 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19808 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19809 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19810 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19811 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19812 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19813 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19814 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19815 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19816 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19817 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19818 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19819 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19820 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19821 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19822 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19823 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19824 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19825 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19826 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19827 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19828 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19829 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19830 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19831 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19832 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19833 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19834 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19835 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19836 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19837 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19838 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19839 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19840 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19841 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19842 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19843 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19844 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19845 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19846 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19847 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19848 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19849 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19850 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19851 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19852 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19853 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19854 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19855 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19856 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19857 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19858 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19859 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19860 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19861 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19862 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19863 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19864 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19865 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19866 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19867 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19868 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19869 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19870 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19871 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19872 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19873 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19874 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19875 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19876 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19877 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19878 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19879 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19880 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19881 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19882 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19883 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19884 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19885 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19886 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19887 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19888 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19889 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19890 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19891 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19892 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655