4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.1.4 and 3.1.5 [xx XXX xxxx]
27 * When function EVP_PKEY_public_check() is called on RSA public keys,
28 a computation is done to confirm that the RSA modulus, n, is composite.
29 For valid RSA keys, n is a product of two or more large primes and this
30 computation completes quickly. However, if n is an overly large prime,
31 then this computation would take a long time.
33 An application that calls EVP_PKEY_public_check() and supplies an RSA key
34 obtained from an untrusted source could be vulnerable to a Denial of Service
37 The function EVP_PKEY_public_check() is not called from other OpenSSL
38 functions however it is called from the OpenSSL pkey command line
39 application. For that reason that application is also vulnerable if used
40 with the "-pubin" and "-check" options on untrusted data.
42 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
43 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
49 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
50 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
55 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
56 for PowerPC CPUs saves the contents of vector registers in different
57 order than they are restored. Thus the contents of some of these vector
58 registers is corrupted when returning to the caller. The vulnerable code is
59 used only on newer PowerPC processors supporting the PowerISA 2.07
62 The consequences of this kind of internal application state corruption can
63 be various - from no consequences, if the calling application does not
64 depend on the contents of non-volatile XMM registers at all, to the worst
65 consequences, where the attacker could get complete control of the
66 application process. However unless the compiler uses the vector registers
67 for storing pointers, the most likely consequence, if any, would be an
68 incorrect result of some application dependent calculations or a crash
69 leading to a denial of service.
75 * Fix excessive time spent in DH check / generation with large Q parameter
78 Applications that use the functions DH_generate_key() to generate an
79 X9.42 DH key may experience long delays. Likewise, applications that use
80 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
81 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
82 Where the key or parameters that are being checked have been obtained from
83 an untrusted source this may lead to a Denial of Service.
89 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
91 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
92 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
93 that alter the key or IV length ([CVE-2023-5363]).
97 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
99 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
101 The POLY1305 MAC (message authentication code) implementation in OpenSSL
102 does not save the contents of non-volatile XMM registers on Windows 64
103 platform when calculating the MAC of data larger than 64 bytes. Before
104 returning to the caller all the XMM registers are set to zero rather than
105 restoring their previous content. The vulnerable code is used only on newer
106 x86_64 processors supporting the AVX512-IFMA instructions.
108 The consequences of this kind of internal application state corruption can
109 be various - from no consequences, if the calling application does not
110 depend on the contents of non-volatile XMM registers at all, to the worst
111 consequences, where the attacker could get complete control of the
112 application process. However given the contents of the registers are just
113 zeroized so the attacker cannot put arbitrary values inside, the most likely
114 consequence, if any, would be an incorrect result of some application
115 dependent calculations or a crash leading to a denial of service.
121 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
123 * Fix excessive time spent checking DH q parameter value.
125 The function DH_check() performs various checks on DH parameters. After
126 fixing CVE-2023-3446 it was discovered that a large q parameter value can
127 also trigger an overly long computation during some of these checks.
128 A correct q value, if present, cannot be larger than the modulus p
129 parameter, thus it is unnecessary to perform these checks if q is larger
132 If DH_check() is called with such q parameter value,
133 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
134 intensive checks are skipped.
140 * Fix DH_check() excessive time with over sized modulus.
142 The function DH_check() performs various checks on DH parameters. One of
143 those checks confirms that the modulus ("p" parameter) is not too large.
144 Trying to use a very large modulus is slow and OpenSSL will not normally use
145 a modulus which is over 10,000 bits in length.
147 However the DH_check() function checks numerous aspects of the key or
148 parameters that have been supplied. Some of those checks use the supplied
149 modulus value even if it has already been found to be too large.
151 A new limit has been added to DH_check of 32,768 bits. Supplying a
152 key/parameters with a modulus over this size will simply cause DH_check() to
159 * Do not ignore empty associated data entries with AES-SIV.
161 The AES-SIV algorithm allows for authentication of multiple associated
162 data entries along with the encryption. To authenticate empty data the
163 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
164 with NULL pointer as the output buffer and 0 as the input buffer length.
165 The AES-SIV implementation in OpenSSL just returns success for such call
166 instead of performing the associated data authentication operation.
167 The empty data thus will not be authenticated. ([CVE-2023-2975])
169 Thanks to Juerg Wullschleger (Google) for discovering the issue.
171 The fix changes the authentication tag value and the ciphertext for
172 applications that use empty associated data entries with AES-SIV.
173 To decrypt data encrypted with previous versions of OpenSSL the application
174 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
179 * When building with the `enable-fips` option and using the resulting
180 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
181 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
182 not operate with truncated digests (FIPS 140-3 IG G.R).
186 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
188 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
189 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
191 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
192 numeric text form. For gigantic sub-identifiers, this would take a very
193 long time, the time complexity being O(n^2) where n is the size of that
194 sub-identifier. ([CVE-2023-2650])
196 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
197 IDENTIFIER to canonical numeric text form if the size of that OBJECT
198 IDENTIFIER is 586 bytes or less, and fail otherwise.
200 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
201 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
202 most 128 sub-identifiers, and that the maximum value that each sub-
203 identifier may have is 2^32-1 (4294967295 decimal).
205 For each byte of every sub-identifier, only the 7 lower bits are part of
206 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
207 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
212 * Multiple algorithm implementation fixes for ARM BE platforms.
216 * Added a -pedantic option to fipsinstall that adjusts the various
217 settings to ensure strict FIPS compliance rather than backwards
222 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
223 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
224 trigger a crash of an application using AES-XTS decryption if the memory
225 just after the buffer being decrypted is not mapped.
226 Thanks to Anton Romanov (Amazon) for discovering the issue.
231 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
232 The previous fix for this timing side channel turned out to cause
233 a severe 2-3x performance regression in the typical use case
234 compared to 3.0.7. The new fix uses existing constant time
235 code paths, and restores the previous performance level while
236 fully eliminating all existing timing side channels.
237 The fix was developed by Bernd Edlinger with testing support
242 * Add FIPS provider configuration option to disallow the use of
243 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
244 The option '-no_drbg_truncated_digests' can optionally be
245 supplied to 'openssl fipsinstall'.
249 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
250 that it does not enable policy checking. Thanks to David Benjamin for
251 discovering this issue.
256 * Fixed an issue where invalid certificate policies in leaf certificates are
257 silently ignored by OpenSSL and other certificate policy checks are skipped
258 for that certificate. A malicious CA could use this to deliberately assert
259 invalid certificate policies in order to circumvent policy checking on the
260 certificate altogether.
265 * Limited the number of nodes created in a policy tree to mitigate
266 against CVE-2023-0464. The default limit is set to 1000 nodes, which
267 should be sufficient for most installations. If required, the limit
268 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
269 time define to a desired maximum number of nodes or zero to allow
275 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
277 * Add FIPS provider configuration option to enforce the
278 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
279 The option '-ems-check' can optionally be supplied to
280 'openssl fipsinstall'.
284 * The FIPS provider includes a few non-approved algorithms for
285 backward compatibility purposes and the "fips=yes" property query
286 must be used for all algorithm fetches to ensure FIPS compliance.
288 The algorithms that are included but not approved are Triple DES ECB,
289 Triple DES CBC and EdDSA.
293 * Added support for KMAC in KBKDF.
297 * RNDR and RNDRRS support in provider functions to provide
298 random number generation for Arm CPUs (aarch64).
302 * s_client and s_server apps now explicitly say when the TLS version
303 does not include the renegotiation mechanism. This avoids confusion
304 between that scenario versus when the TLS version includes secure
305 renegotiation but the peer lacks support for it.
309 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
311 *Tomasz Kantecki, Andrey Matyukov*
313 * The various OBJ_* functions have been made thread safe.
317 * Parallel dual-prime 1536/2048-bit modular exponentiation for
318 AVX512_IFMA capable processors.
320 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
322 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
323 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
324 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
325 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
326 `OPENSSL_NO_DEPRECATED_3_1`.
328 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
329 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
330 definitions for these functions regardless of whether
331 `OPENSSL_NO_DEPRECATED_3_1` is defined.
333 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
334 functions regardless of whether they are using them. It is recommended that
335 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
339 * When generating safe-prime DH parameters set the recommended private key
340 length equivalent to minimum key lengths as in RFC 7919.
344 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
345 maximum size that is smaller or equal to the digest length to comply with
346 FIPS 186-4 section 5. This is implemented by a new option
347 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
348 `rsa_pss_saltlen` parameter, which is now the default. Signature
349 verification is not affected by this change and continues to work as before.
356 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
357 listed here are only a brief description.
358 The migration guide contains more detailed information related to new features,
359 breaking changes, and mappings for the large list of deprecated functions.
361 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
363 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
365 * Fixed NULL dereference during PKCS7 data verification.
367 A NULL pointer can be dereferenced when signatures are being
368 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
369 algorithm used for the signature is known to the OpenSSL library but
370 the implementation of the hash algorithm is not available the digest
371 initialization will fail. There is a missing check for the return
372 value from the initialization function which later leads to invalid
373 usage of the digest API most likely leading to a crash.
376 PKCS7 data is processed by the SMIME library calls and also by the
377 time stamp (TS) library calls. The TLS implementation in OpenSSL does
378 not call these functions however third party applications would be
379 affected if they call these functions to verify signatures on untrusted
384 * Fixed X.400 address type confusion in X.509 GeneralName.
386 There is a type confusion vulnerability relating to X.400 address processing
387 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
388 but the public structure definition for GENERAL_NAME incorrectly specified
389 the type of the x400Address field as ASN1_TYPE. This field is subsequently
390 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
393 When CRL checking is enabled (i.e. the application sets the
394 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
395 pass arbitrary pointers to a memcmp call, enabling them to read memory
396 contents or enact a denial of service.
401 * Fixed NULL dereference validating DSA public key.
403 An invalid pointer dereference on read can be triggered when an
404 application tries to check a malformed DSA public key by the
405 EVP_PKEY_public_check() function. This will most likely lead
406 to an application crash. This function can be called on public
407 keys supplied from untrusted sources which could allow an attacker
408 to cause a denial of service attack.
410 The TLS implementation in OpenSSL does not call this function
411 but applications might call the function if there are additional
412 security requirements imposed by standards such as FIPS 140-3.
415 *Shane Lontis, Tomáš Mráz*
417 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
419 An invalid pointer dereference on read can be triggered when an
420 application tries to load malformed PKCS7 data with the
421 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
423 The result of the dereference is an application crash which could
424 lead to a denial of service attack. The TLS implementation in OpenSSL
425 does not call this function however third party applications might
426 call these functions on untrusted data.
431 * Fixed Use-after-free following BIO_new_NDEF.
433 The public API function BIO_new_NDEF is a helper function used for
434 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
435 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
436 be called directly by end user applications.
438 The function receives a BIO from the caller, prepends a new BIO_f_asn1
439 filter BIO onto the front of it to form a BIO chain, and then returns
440 the new head of the BIO chain to the caller. Under certain conditions,
441 for example if a CMS recipient public key is invalid, the new filter BIO
442 is freed and the function returns a NULL result indicating a failure.
443 However, in this case, the BIO chain is not properly cleaned up and the
444 BIO passed by the caller still retains internal pointers to the previously
445 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
446 then a use-after-free will occur. This will most likely result in a crash.
449 *Viktor Dukhovni, Matt Caswell*
451 * Fixed Double free after calling PEM_read_bio_ex.
453 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
454 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
455 data. If the function succeeds then the "name_out", "header" and "data"
456 arguments are populated with pointers to buffers containing the relevant
457 decoded data. The caller is responsible for freeing those buffers. It is
458 possible to construct a PEM file that results in 0 bytes of payload data.
459 In this case PEM_read_bio_ex() will return a failure code but will populate
460 the header argument with a pointer to a buffer that has already been freed.
461 If the caller also frees this buffer then a double free will occur. This
462 will most likely lead to a crash.
464 The functions PEM_read_bio() and PEM_read() are simple wrappers around
465 PEM_read_bio_ex() and therefore these functions are also directly affected.
467 These functions are also called indirectly by a number of other OpenSSL
468 functions including PEM_X509_INFO_read_bio_ex() and
469 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
470 internal uses of these functions are not vulnerable because the caller does
471 not free the header argument if PEM_read_bio_ex() returns a failure code.
474 *Kurt Roeckx, Matt Caswell*
476 * Fixed Timing Oracle in RSA Decryption.
478 A timing based side channel exists in the OpenSSL RSA Decryption
479 implementation which could be sufficient to recover a plaintext across
480 a network in a Bleichenbacher style attack. To achieve a successful
481 decryption an attacker would have to be able to send a very large number
482 of trial messages for decryption. The vulnerability affects all RSA padding
483 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
486 *Dmitry Belyavsky, Hubert Kario*
488 * Fixed X.509 Name Constraints Read Buffer Overflow.
490 A read buffer overrun can be triggered in X.509 certificate verification,
491 specifically in name constraint checking. The read buffer overrun might
492 result in a crash which could lead to a denial of service attack.
493 In a TLS client, this can be triggered by connecting to a malicious
494 server. In a TLS server, this can be triggered if the server requests
495 client authentication and a malicious client connects.
500 * Fixed X.509 Policy Constraints Double Locking security issue.
502 If an X.509 certificate contains a malformed policy constraint and
503 policy processing is enabled, then a write lock will be taken twice
504 recursively. On some operating systems (most widely: Windows) this
505 results in a denial of service when the affected process hangs. Policy
506 processing being enabled on a publicly facing server is not considered
507 to be a common setup.
512 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
513 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
514 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
515 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
516 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
517 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
518 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
519 for legacy EC and SM2 keys is also changed similarly to honor the
520 equivalent conversion format flag as specified in the underlying
521 `EC_KEY` object being exported to a provider, when this function is
522 called through `EVP_PKEY_export()`.
526 * Fixed a type confusion vulnerability relating to X.400 address processing
527 inside an X.509 GeneralName. X.400 addresses were parsed as an `ASN1_STRING`
528 but subsequently interpreted by `GENERAL_NAME_cmp` as an `ASN1_TYPE`. This
529 vulnerability may allow an attacker who can provide a certificate chain and
530 CRL (neither of which need have a valid signature) to pass arbitrary pointers
531 to a `memcmp` call, creating a possible read primitive, subject to some
532 constraints. Refer to the advisory for more information. Thanks to David
533 Benjamin for discovering this issue. ([CVE-2023-0286])
535 This issue has been fixed by changing the public header file definition of
536 `GENERAL_NAME` so that `x400Address` reflects the implementation. It was not
537 possible for any existing application to successfully use the existing
538 definition; however, if any application references the `x400Address` field
539 (e.g. in dead code), note that the type of this field has changed. There is
544 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
546 * Fixed two buffer overflows in punycode decoding functions.
548 A buffer overrun can be triggered in X.509 certificate verification,
549 specifically in name constraint checking. Note that this occurs after
550 certificate chain signature verification and requires either a CA to
551 have signed the malicious certificate or for the application to continue
552 certificate verification despite failure to construct a path to a trusted
555 In a TLS client, this can be triggered by connecting to a malicious
556 server. In a TLS server, this can be triggered if the server requests
557 client authentication and a malicious client connects.
559 An attacker can craft a malicious email address to overflow
560 an arbitrary number of bytes containing the `.` character (decimal 46)
561 on the stack. This buffer overflow could result in a crash (causing a
565 An attacker can craft a malicious email address to overflow four
566 attacker-controlled bytes on the stack. This buffer overflow could
567 result in a crash (causing a denial of service) or potentially remote code
568 execution depending on stack layout for any given platform/compiler.
573 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
574 parameters in OpenSSL code.
575 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
576 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
577 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
578 Using these invalid names may cause algorithms to use slower methods
579 that ignore the CRT parameters.
583 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
588 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
589 data to be signed before signing the certificate.
593 * Added RIPEMD160 to the default provider.
597 * Ensured that the key share group sent or accepted for the key exchange
598 is allowed for the protocol version.
602 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
604 * OpenSSL supports creating a custom cipher via the legacy
605 EVP_CIPHER_meth_new() function and associated function calls. This function
606 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
607 to use the new provider mechanism in order to implement custom ciphers.
609 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
610 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
611 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
612 and decryption initialisation functions). Instead of using the custom cipher
613 directly it incorrectly tries to fetch an equivalent cipher from the
614 available providers. An equivalent cipher is found based on the NID passed to
615 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
616 given cipher. However it is possible for an application to incorrectly pass
617 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
618 is used in this way the OpenSSL encryption/decryption initialisation function
619 will match the NULL cipher as being equivalent and will fetch this from the
620 available providers. This will succeed if the default provider has been
621 loaded (or if a third party provider has been loaded that offers this
622 cipher). Using the NULL cipher means that the plaintext is emitted as the
625 Applications are only affected by this issue if they call
626 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
627 encryption/decryption initialisation function. Applications that only use
628 SSL/TLS are not impacted by this issue.
633 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
638 * Fixed the linux-mips64 Configure target which was missing the
639 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
644 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
649 * Correctly handle a retransmitted ClientHello in DTLS
653 * Fixed detection of ktls support in cross-compile environment on Linux
657 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
662 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
663 report correct results in some cases
667 * Fix UWP builds by defining VirtualLock
671 * For known safe primes use the minimum key length according to RFC 7919.
672 Longer private key sizes unnecessarily raise the cycles needed to compute the
673 shared secret without any increase of the real security. This fixes a
674 regression from 1.1.1 where these shorter keys were generated for the known
679 * Added the loongarch64 target
683 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
684 only passed to the FIPS provider and not to the default or legacy provider.
688 * Fixed reported performance degradation on aarch64. Restored the
689 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
690 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
691 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
692 The new algorithm is still used for 32 bit targets.
696 * Added a missing header for memcmp that caused compilation failure on some
701 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
703 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
704 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
705 This issue makes the RSA implementation with 2048 bit private keys
706 incorrect on such machines and memory corruption will happen during
707 the computation. As a consequence of the memory corruption an attacker
708 may be able to trigger a remote code execution on the machine performing
711 SSL/TLS servers or other servers using 2048 bit RSA private keys running
712 on machines supporting AVX512IFMA instructions of the X86_64 architecture
713 are affected by this issue.
718 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
719 implementation would not encrypt the entirety of the data under some
720 circumstances. This could reveal sixteen bytes of data that was
721 preexisting in the memory that wasn't written. In the special case of
722 "in place" encryption, sixteen bytes of the plaintext would be revealed.
724 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
725 they are both unaffected.
728 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
730 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
732 * In addition to the c_rehash shell command injection identified in
733 CVE-2022-1292, further bugs where the c_rehash script does not
734 properly sanitise shell metacharacters to prevent command injection have been
737 When the CVE-2022-1292 was fixed it was not discovered that there
738 are other places in the script where the file names of certificates
739 being hashed were possibly passed to a command executed through the shell.
741 This script is distributed by some operating systems in a manner where
742 it is automatically executed. On such operating systems, an attacker
743 could execute arbitrary commands with the privileges of the script.
745 Use of the c_rehash script is considered obsolete and should be replaced
746 by the OpenSSL rehash command line tool.
749 *Daniel Fiala, Tomáš Mráz*
751 * Case insensitive string comparison no longer uses locales. It has instead
752 been directly implemented.
756 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
758 * Case insensitive string comparison is reimplemented via new locale-agnostic
759 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
760 comparison. The previous implementation had problems when the Turkish locale
765 * Fixed a bug in the c_rehash script which was not properly sanitising shell
766 metacharacters to prevent command injection. This script is distributed by
767 some operating systems in a manner where it is automatically executed. On
768 such operating systems, an attacker could execute arbitrary commands with the
769 privileges of the script.
771 Use of the c_rehash script is considered obsolete and should be replaced
772 by the OpenSSL rehash command line tool.
777 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
778 certificate on an OCSP response. The bug caused the function in the case
779 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
780 response (meaning a successful verification) even in the case where the
781 response signing certificate fails to verify.
783 It is anticipated that most users of `OCSP_basic_verify` will not use the
784 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
785 a negative value (indicating a fatal error) in the case of a certificate
786 verification failure. The normal expected return value in this case would be
789 This issue also impacts the command line OpenSSL "ocsp" application. When
790 verifying an ocsp response with the "-no_cert_checks" option the command line
791 application will report that the verification is successful even though it
792 has in fact failed. In this case the incorrect successful response will also
793 be accompanied by error messages showing the failure and contradicting the
794 apparently successful result.
799 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
800 AAD data as the MAC key. This made the MAC key trivially predictable.
802 An attacker could exploit this issue by performing a man-in-the-middle attack
803 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
804 that the modified data would still pass the MAC integrity check.
806 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
807 endpoint will always be rejected by the recipient and the connection will
808 fail at that point. Many application protocols require data to be sent from
809 the client to the server first. Therefore, in such a case, only an OpenSSL
810 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
812 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
813 sent in both directions. In this case both clients and servers could be
814 affected, regardless of the application protocol.
816 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
817 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
818 the handshake when using this ciphersuite.
820 The confidentiality of data is not impacted by this issue, i.e. an attacker
821 cannot decrypt data that has been encrypted using this ciphersuite - they can
824 In order for this attack to work both endpoints must legitimately negotiate
825 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
826 OpenSSL 3.0, and is not available within the default provider or the default
827 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
828 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
829 following must have occurred:
831 1) OpenSSL must have been compiled with the (non-default) compile time option
832 enable-weak-ssl-ciphers
834 2) OpenSSL must have had the legacy provider explicitly loaded (either
835 through application code or via configuration)
837 3) The ciphersuite must have been explicitly added to the ciphersuite list
839 4) The libssl security level must have been set to 0 (default is 1)
841 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
843 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
844 others that both endpoints have in common
849 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
850 occupied by the removed hash table entries.
852 This function is used when decoding certificates or keys. If a long lived
853 process periodically decodes certificates or keys its memory usage will
854 expand without bounds and the process might be terminated by the operating
855 system causing a denial of service. Also traversing the empty hash table
856 entries will take increasingly more time.
858 Typically such long lived processes might be TLS clients or TLS servers
859 configured to accept client certificate authentication.
862 *Hugo Landau, Aliaksei Levin*
864 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
865 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
866 statistics are no longer supported. For compatibility, these statistics are
867 still listed in the output but are now always reported as zero.
871 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
873 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
874 for non-prime moduli.
876 Internally this function is used when parsing certificates that contain
877 elliptic curve public keys in compressed form or explicit elliptic curve
878 parameters with a base point encoded in compressed form.
880 It is possible to trigger the infinite loop by crafting a certificate that
881 has invalid explicit curve parameters.
883 Since certificate parsing happens prior to verification of the certificate
884 signature, any process that parses an externally supplied certificate may thus
885 be subject to a denial of service attack. The infinite loop can also be
886 reached when parsing crafted private keys as they can contain explicit
887 elliptic curve parameters.
889 Thus vulnerable situations include:
891 - TLS clients consuming server certificates
892 - TLS servers consuming client certificates
893 - Hosting providers taking certificates or private keys from customers
894 - Certificate authorities parsing certification requests from subscribers
895 - Anything else which parses ASN.1 elliptic curve parameters
897 Also any other applications that use the BN_mod_sqrt() where the attacker
898 can control the parameter values are vulnerable to this DoS issue.
903 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
904 to the list of ciphersuites providing Perfect Forward Secrecy as
905 required by SECLEVEL >= 3.
907 *Dmitry Belyavskiy, Nicola Tuveri*
909 * Made the AES constant time code for no-asm configurations
910 optional due to the resulting 95% performance degradation.
911 The AES constant time code can be enabled, for no assembly
912 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
916 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
921 * The negative return value handling of the certificate verification callback
922 was reverted. The replacement is to set the verification retry state with
923 the SSL_set_retry_verify() function.
927 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
929 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
930 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
931 verify a certificate supplied by a server. That function may return a
932 negative return value to indicate an internal error (for example out of
933 memory). Such a negative return value is mishandled by OpenSSL and will cause
934 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
935 success and a subsequent call to SSL_get_error() to return the value
936 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
937 returned by OpenSSL if the application has previously called
938 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
939 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
940 totally unexpected and applications may not behave correctly as a result. The
941 exact behaviour will depend on the application but it could result in
942 crashes, infinite loops or other similar incorrect responses.
944 This issue is made more serious in combination with a separate bug in OpenSSL
945 3.0 that will cause X509_verify_cert() to indicate an internal error when
946 processing a certificate chain. This will occur where a certificate does not
947 include the Subject Alternative Name extension but where a Certificate
948 Authority has enforced name constraints. This issue can occur even with valid
954 * Corrected a few file name and file reference bugs in the build,
955 installation and setup scripts, which lead to installation verification
956 failures. Slightly enhanced the installation verification script.
960 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
965 * Fixed PVK encoder to properly query for the passphrase.
969 * Multiple fixes in the OSSL_HTTP API functions.
973 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
974 OSSL_PARAM_INTEGER data type and return error on negative numbers
975 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
976 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
980 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
984 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
988 * Multiple threading fixes.
992 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
996 * Allow fetching an operation from the provider that owns an unexportable key
997 as a fallback if that is still allowed by the property query.
1001 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
1003 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1008 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1009 S390X capability vector to zero. This simplifies testing of different code
1010 paths on S390X architecture.
1014 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1015 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1016 SP 800-38D". The communication will fail at this point.
1020 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1021 confidential in EC_GROUP data.
1025 * The byte order mark (BOM) character is ignored if encountered at the
1026 beginning of a PEM-formatted file.
1030 * Added CMS support for the Russian GOST algorithms.
1034 * Due to move of the implementation of cryptographic operations
1035 to the providers, validation of various operation parameters can
1036 be postponed until the actual operation is executed where previously
1037 it happened immediately when an operation parameter was set.
1039 For example when setting an unsupported curve with
1040 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1041 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1043 *OpenSSL team members and many third party contributors*
1045 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1046 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1047 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1048 instead to retrieve these algorithms from a provider.
1052 * On build targets where the multilib postfix is set in the build
1053 configuration the libdir directory was changing based on whether
1054 the lib directory with the multilib postfix exists on the system
1055 or not. This unpredictable behavior was removed and eventual
1056 multilib postfix is now always added to the default libdir. Use
1057 `--libdir=lib` to override the libdir if adding the postfix is
1062 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1063 no longer interoperable with OpenSSL 1.1.1.
1067 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1068 function codes, this function can only cause problems for calling
1073 * Add a configurable flag to output date formats as ISO 8601. Does not
1074 change the default date format.
1078 * Version of MSVC earlier than 1300 could get link warnings, which could
1079 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1080 Support for this flag has been removed.
1084 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1085 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1086 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1087 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1088 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1092 * The signatures of the functions to get and set options on SSL and
1093 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1094 Some source code changes may be required.
1098 * The public definitions of conf_method_st and conf_st have been
1099 deprecated. They will be made opaque in a future release.
1101 *Rich Salz and Tomáš Mráz*
1103 * Client-initiated renegotiation is disabled by default. To allow it, use
1104 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1105 flag, or the "ClientRenegotiation" config parameter as appropriate.
1109 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1110 or modify relative pathname inclusion.
1114 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1115 validated. Please consult the README-FIPS and
1116 README-PROVIDERS files, as well as the migration guide.
1118 *OpenSSL team members and many third party contributors*
1120 * For the key types DH and DHX the allowed settable parameters are now different.
1124 * The openssl commands that read keys, certificates, and CRLs now
1125 automatically detect the PEM or DER format of the input files.
1127 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1129 * Added enhanced PKCS#12 APIs which accept a library context.
1133 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1137 * Added support for Kernel TLS (KTLS).
1139 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1141 * Support for RFC 5746 secure renegotiation is now required by default for
1142 SSL or TLS connections to succeed.
1146 * The signature of the `copy` functional parameter of the
1147 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1148 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1149 the signature of the `pub_decode` functional parameter of the
1150 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1151 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1155 * The error return values from some control calls (ctrl) have changed.
1159 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1163 * Many functions in the EVP_ namespace that are getters of values from
1164 implementations or contexts were renamed to include get or get0 in their
1165 names. Old names are provided as macro aliases for compatibility and
1170 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1171 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1172 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1177 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1180 * The output from the command line applications may have minor
1185 * The output from numerous "printing" may have minor changes.
1189 * Windows thread synchronization uses read/write primitives (SRWLock) when
1190 supported by the OS, otherwise CriticalSection continues to be used.
1194 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1195 work on read only BIO source/sinks that do not support these functions.
1196 This allows piping or redirection of a file BIO using stdin to be buffered
1197 into memory. This is used internally in OSSL_DECODER_from_bio().
1201 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1202 this function would return one of the values OSSL_STORE_INFO_NAME,
1203 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1204 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1205 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1206 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1207 using this function should be amended to handle the changed return value.
1211 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1212 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1213 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1214 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1215 contains more than one certificate identifier: This means that all
1216 certificates referenced there MUST be part of the validation chain.
1220 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1221 RC5, DESX and DES have been moved to the legacy provider.
1225 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1226 RIPEMD-160 have been moved to the legacy provider.
1230 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1235 * The deprecated functions EVP_PKEY_get0_RSA(),
1236 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1237 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1238 well as the similarly named "get1" functions behave differently in
1243 * A number of functions handling low-level keys or engines were deprecated
1244 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1245 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1246 EVP_PKEY_get0_siphash().
1250 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1251 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1252 will need to load the legacy crypto provider. This includes these PBE
1253 algorithms which use this KDF:
1254 - NID_pbeWithMD2AndDES_CBC
1255 - NID_pbeWithMD5AndDES_CBC
1256 - NID_pbeWithSHA1AndRC2_CBC
1257 - NID_pbeWithMD2AndRC2_CBC
1258 - NID_pbeWithMD5AndRC2_CBC
1259 - NID_pbeWithSHA1AndDES_CBC
1263 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1264 BIO_debug_callback() functions.
1268 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1269 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1273 * The RAND_METHOD APIs have been deprecated.
1277 * The SRP APIs have been deprecated.
1281 * Add a compile time option to prevent the caching of provider fetched
1282 algorithms. This is enabled by including the no-cached-fetch option
1283 at configuration time.
1287 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1288 count of PKCS12_DEFAULT_ITER.
1290 *Tomáš Mráz and Sahana Prasad*
1292 * The openssl speed command does not use low-level API calls anymore.
1296 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1299 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1301 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1305 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1306 providers may supply their own group implementations (using either the "key
1307 exchange" or the "key encapsulation" methods) which will automatically be
1308 detected and used by libssl.
1310 *Matt Caswell, Nicola Tuveri*
1312 * The undocumented function X509_certificate_type() has been deprecated;
1316 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1320 * Removed RSA padding mode for SSLv23 (which was only used for
1321 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1322 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1327 * Deprecated the obsolete X9.31 RSA key generation related functions.
1329 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1330 is not allowed to return a value > 1, this is no more taken as failure.
1332 *Viktor Dukhovni and David von Oheimb*
1334 * Deprecated the obsolete X9.31 RSA key generation related functions
1335 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1336 BN_X931_generate_prime_ex().
1340 * The default key generation method for the regular 2-prime RSA keys was
1341 changed to the FIPS 186-4 B.3.6 method.
1345 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1349 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1353 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1354 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1356 *Rich Salz, Richard Levitte, and David von Oheimb*
1358 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1362 * Deprecated `OCSP_parse_url()`.
1366 * Validation of SM2 keys has been separated from the validation of regular EC
1371 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1372 switches: a validation failure triggers an early exit, returning a failure
1373 exit status to the parent process.
1377 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1378 to ignore unknown ciphers.
1382 * The `-cipher-commands` and `-digest-commands` options
1383 of the command line utility `list` have been deprecated.
1384 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1388 * Added convenience functions for generating asymmetric key pairs:
1389 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1390 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1394 * All of the low-level EC_KEY functions have been deprecated.
1396 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1398 * Deprecated all the libcrypto and libssl error string loading
1403 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1404 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1409 * The `-crypt` option to the `passwd` command line tool has been removed.
1413 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1418 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1422 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1423 EVP_PKEY_get1_tls_encodedpoint().
1427 * The security callback, which can be customised by application code, supports
1428 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1429 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1433 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1434 interface. Their functionality remains unchanged.
1438 * Added new option for 'openssl list', '-providers', which will display the
1439 list of loaded providers, their names, version and status. It optionally
1440 displays their gettable parameters.
1444 * Removed EVP_PKEY_set_alias_type().
1448 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1449 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1453 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1454 parameters are still checked for type safety at compile time via helper
1459 * Remove the RAND_DRBG API
1461 *Paul Dale and Matthias St. Pierre*
1463 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1464 as well as actual hostnames.
1468 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1469 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1470 conversely, silently ignore DTLS protocol version bounds when configuring
1471 TLS-based contexts. The commands can be repeated to set bounds of both
1472 types. The same applies with the corresponding "min_protocol" and
1473 "max_protocol" command-line switches, in case some application uses both TLS
1476 SSL_CTX instances that are created for a fixed protocol version (e.g.
1477 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1478 attempts to apply bounds to these protocol versions would result in an
1479 error. Now only the "version-flexible" SSL_CTX instances are subject to
1480 limits in configuration files in command-line options.
1484 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1489 * Reworked the recorded ERR codes to make better space for system errors.
1490 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1491 given code is a system error (true) or an OpenSSL error (false).
1495 * Reworked the test perl framework to better allow parallel testing.
1497 *Nicola Tuveri and David von Oheimb*
1499 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1500 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1504 * 'Configure' has been changed to figure out the configuration target if
1505 none is given on the command line. Consequently, the 'config' script is
1506 now only a mere wrapper. All documentation is changed to only mention
1509 *Rich Salz and Richard Levitte*
1511 * Added a library context `OSSL_LIB_CTX` that applications as well as
1512 other libraries can use to form a separate context within which
1513 libcrypto operations are performed.
1517 * Added various `_ex` functions to the OpenSSL API that support using
1518 a non-default `OSSL_LIB_CTX`.
1522 * Handshake now fails if Extended Master Secret extension is dropped
1527 * Dropped interactive mode from the `openssl` program.
1531 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1533 *David von Oheimb and Shane Lontis*
1535 * Deprecated `EC_METHOD_get_field_type()`.
1539 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1540 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1541 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1545 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1549 * Add CAdES-BES signature verification support, mostly derived
1550 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1552 *Filipe Raimundo da Silva*
1554 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1558 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1559 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1563 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1567 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1568 EC_KEY_precompute_mult().
1572 * Deprecated EC_POINTs_mul().
1576 * Removed FIPS_mode() and FIPS_mode_set().
1580 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1584 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1585 EC_POINT_get_Jprojective_coordinates_GFp().
1589 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1590 arrays to be more easily constructed via a series of utility functions.
1591 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1592 the various push functions and finally convert to a passable OSSL_PARAM
1593 array using OSSL_PARAM_BLD_to_param().
1597 * The security strength of SHA1 and MD5 based signatures in TLS has been
1602 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1603 contain a provider side internal key.
1607 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1611 * Project text documents not yet having a proper file name extension
1612 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1613 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1617 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1618 have been converted to Markdown with the goal to produce documents
1619 which not only look pretty when viewed online in the browser, but
1620 remain well readable inside a plain text editor.
1622 To achieve this goal, a 'minimalistic' Markdown style has been applied
1623 which avoids formatting elements that interfere too much with the
1624 reading flow in the text file. For example, it
1626 * avoids [ATX headings][] and uses [setext headings][] instead
1627 (which works for `<h1>` and `<h2>` headings only).
1628 * avoids [inline links][] and uses [reference links][] instead.
1629 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1631 [ATX headings]: https://github.github.com/gfm/#atx-headings
1632 [setext headings]: https://github.github.com/gfm/#setext-headings
1633 [inline links]: https://github.github.com/gfm/#inline-link
1634 [reference links]: https://github.github.com/gfm/#reference-link
1635 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1636 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1638 *Matthias St. Pierre*
1640 * The test suite is changed to preserve results of each test recipe.
1641 A new directory test-runs/ with subdirectories named like the
1642 test recipes are created in the build tree for this purpose.
1646 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1647 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1648 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1650 *David von Oheimb, Martin Peylo*
1652 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1653 It supports arbitrary request and response content types, GET redirection,
1654 TLS, connections via HTTP(S) proxies, connections and exchange via
1655 user-defined BIOs (allowing implicit connections), persistent connections,
1656 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1657 The legacy OCSP-focused (and only partly documented) API
1658 is retained for backward compatibility, while most of it is deprecated.
1662 * Added `util/check-format.pl`, a tool for checking adherence to the
1663 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1664 The checks performed are incomplete and yield some false positives.
1665 Still the tool should be useful for detecting most typical glitches.
1669 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1670 If domain name resolution yields multiple IP addresses all of them are tried
1671 after `connect()` failures.
1675 * All of the low-level RSA functions have been deprecated.
1679 * X509 certificates signed using SHA1 are no longer allowed at security
1684 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1685 modified to use PKEY APIs. These commands are now in maintenance mode
1686 and no new features will be added to them.
1690 * The command line utility rsautl has been deprecated.
1694 * The command line utilities genrsa and rsa have been modified to use PKEY
1695 APIs. They now write PKCS#8 keys by default. These commands are now in
1696 maintenance mode and no new features will be added to them.
1700 * All of the low-level DH functions have been deprecated.
1702 *Paul Dale and Matt Caswell*
1704 * All of the low-level DSA functions have been deprecated.
1708 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1709 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1713 * Deprecated low-level ECDH and ECDSA functions.
1717 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1721 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1722 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1723 a new formulation to include all the things it can be used for,
1724 as well as words of caution.
1728 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1732 * All of the low-level HMAC functions have been deprecated.
1734 *Paul Dale and David von Oheimb*
1736 * Over two thousand fixes were made to the documentation, including:
1737 - Common options (such as -rand/-writerand, TLS version control, etc)
1738 were refactored and point to newly-enhanced descriptions in openssl.pod.
1739 - Added style conformance for all options (with help from Richard Levitte),
1740 documented all reported missing options, added a CI build to check
1741 that all options are documented and that no unimplemented options
1743 - Documented some internals, such as all use of environment variables.
1744 - Addressed all internal broken L<> references.
1748 * All of the low-level CMAC functions have been deprecated.
1752 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1753 functions have been deprecated.
1755 *Paul Dale and David von Oheimb*
1757 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1758 set of functions. The documentation mentioned negative values for some
1759 errors, but this was never the case, so the mention of negative values
1762 Code that followed the documentation and thereby check with something
1763 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1767 * All of the low-level cipher functions have been deprecated.
1769 *Matt Caswell and Paul Dale*
1771 * Removed include/openssl/opensslconf.h.in and replaced it with
1772 include/openssl/configuration.h.in, which differs in not including
1773 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1774 was added to include both.
1776 This allows internal hacks where one might need to modify the set
1777 of configured macros, for example this if deprecated symbols are
1778 still supposed to be available internally:
1780 #include <openssl/configuration.h>
1782 #undef OPENSSL_NO_DEPRECATED
1783 #define OPENSSL_SUPPRESS_DEPRECATED
1785 #include <openssl/macros.h>
1787 This should not be used by applications that use the exported
1788 symbols, as that will lead to linking errors.
1792 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1793 used in exponentiation with 512-bit moduli. No EC algorithms are
1794 affected. Analysis suggests that attacks against 2-prime RSA1024,
1795 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1796 difficult to perform and are not believed likely. Attacks against DH512
1797 are considered just feasible. However, for an attack the target would
1798 have to re-use the DH512 private key, which is not recommended anyway.
1799 Also applications directly using the low-level API BN_mod_exp may be
1800 affected if they use BN_FLG_CONSTTIME.
1805 * Most memory-debug features have been deprecated, and the functionality
1806 replaced with no-ops.
1810 * Added documentation for the STACK API.
1814 * Introduced a new method type and API, OSSL_ENCODER, to represent
1815 generic encoders. These do the same sort of job that PEM writers
1816 and d2i functions do, but with support for methods supplied by
1817 providers, and the possibility for providers to support other
1822 * Introduced a new method type and API, OSSL_DECODER, to represent
1823 generic decoders. These do the same sort of job that PEM readers
1824 and i2d functions do, but with support for methods supplied by
1825 providers, and the possibility for providers to support other
1830 * Added a .pragma directive to the syntax of configuration files, to
1831 allow varying behavior in a supported and predictable manner.
1832 Currently added pragma:
1836 This allows dollar signs to be a keyword character unless it's
1837 followed by a opening brace or parenthesis. This is useful for
1838 platforms where dollar signs are commonly used in names, such as
1839 volume names and system directory names on VMS.
1843 * Added functionality to create an EVP_PKEY from user data.
1847 * Change the interpretation of the '--api' configuration option to
1848 mean that this is a desired API compatibility level with no
1849 further meaning. The previous interpretation, that this would
1850 also mean to remove all deprecated symbols up to and including
1851 the given version, no requires that 'no-deprecated' is also used
1852 in the configuration.
1854 When building applications, the desired API compatibility level
1855 can be set with the OPENSSL_API_COMPAT macro like before. For
1856 API compatibility version below 3.0, the old style numerical
1857 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1858 For version 3.0 and on, the value is expected to be the decimal
1859 value calculated from the major and minor version like this:
1861 MAJOR * 10000 + MINOR * 100
1865 -DOPENSSL_API_COMPAT=30000 For 3.0
1866 -DOPENSSL_API_COMPAT=30200 For 3.2
1868 To hide declarations that are deprecated up to and including the
1869 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1870 given when building the application as well.
1874 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1875 access to certificate and CRL stores via URIs and OSSL_STORE
1878 This adds the following functions:
1880 - X509_LOOKUP_store()
1881 - X509_STORE_load_file()
1882 - X509_STORE_load_path()
1883 - X509_STORE_load_store()
1884 - SSL_add_store_cert_subjects_to_stack()
1885 - SSL_CTX_set_default_verify_store()
1886 - SSL_CTX_load_verify_file()
1887 - SSL_CTX_load_verify_dir()
1888 - SSL_CTX_load_verify_store()
1892 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1893 The presence of this system service is determined at run-time.
1897 * Added functionality to create an EVP_PKEY context based on data
1898 for methods from providers. This takes an algorithm name and a
1899 property query string and simply stores them, with the intent
1900 that any operation that uses this context will use those strings
1901 to fetch the needed methods implicitly, thereby making the port
1902 of application written for pre-3.0 OpenSSL easier.
1906 * The undocumented function NCONF_WIN32() has been deprecated; for
1907 conversion details see the HISTORY section of doc/man5/config.pod
1911 * Introduced the new functions EVP_DigestSignInit_ex() and
1912 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1913 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1914 pages for further details.
1918 * Over two thousand fixes were made to the documentation, including:
1919 adding missing command flags, better style conformance, documentation
1922 *Rich Salz, Richard Levitte*
1924 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1925 X25519, X448, Ed25519 and Ed448.
1929 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1934 * Deprecated the public definition of `ERR_STATE` as well as the function
1935 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1940 * Added ERR functionality to give callers access to the stored function
1941 names that have replaced the older function code based functions.
1943 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1944 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1945 ERR_peek_error_all() and ERR_peek_last_error_all().
1947 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1948 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1949 ERR_func_error_string().
1953 * Extended testing to be verbose for failing tests only. The make variables
1954 VERBOSE_FAILURE or VF can be used to enable this:
1956 $ make VF=1 test # Unix
1957 $ mms /macro=(VF=1) test ! OpenVMS
1958 $ nmake VF=1 test # Windows
1962 * Added the `-copy_extensions` option to the `x509` command for use with
1963 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1964 all extensions in the request are copied to the certificate or vice versa.
1966 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1968 * Added the `-copy_extensions` option to the `req` command for use with
1969 `-x509`. When given with the `copy` or `copyall` argument,
1970 all extensions in the certification request are copied to the certificate.
1974 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1975 they generate are by default RFC 5280 compliant in the following sense:
1976 There is a subjectKeyIdentifier extension with a hash value of the public key
1977 and for not self-signed certs there is an authorityKeyIdentifier extension
1978 with a keyIdentifier field or issuer information identifying the signing key.
1979 This is done unless some configuration overrides the new default behavior,
1980 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1984 * Added several checks to `X509_verify_cert()` according to requirements in
1985 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1986 (which may be done by using the CLI option `-x509_strict`):
1987 * The basicConstraints of CA certificates must be marked critical.
1988 * CA certificates must explicitly include the keyUsage extension.
1989 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1990 * The issuer name of any certificate must not be empty.
1991 * The subject name of CA certs, certs with keyUsage crlSign,
1992 and certs without subjectAlternativeName must not be empty.
1993 * If a subjectAlternativeName extension is given it must not be empty.
1994 * The signatureAlgorithm field and the cert signature must be consistent.
1995 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1996 must not be marked critical.
1997 * The authorityKeyIdentifier must be given for X.509v3 certs
1998 unless they are self-signed.
1999 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2003 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2004 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2008 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2009 used even when parsing explicit parameters, when loading a encoded key
2010 or calling `EC_GROUP_new_from_ecpkparameters()`/
2011 `EC_GROUP_new_from_ecparameters()`.
2012 This prevents bypass of security hardening and performance gains,
2013 especially for curves with specialized EC_METHODs.
2014 By default, if a key encoded with explicit parameters is loaded and later
2015 encoded, the output is still encoded with explicit parameters, even if
2016 internally a "named" EC_GROUP is used for computation.
2020 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2021 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2022 NULL. After this change, only the cofactor parameter can be NULL. It also
2023 does some minimal sanity checks on the passed order.
2028 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2029 An attack is simple, if the first CMS_recipientInfo is valid but the
2030 second CMS_recipientInfo is chosen ciphertext. If the second
2031 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2032 encryption key will be replaced by garbage, and the message cannot be
2033 decoded, but if the RSA decryption fails, the correct encryption key is
2034 used and the recipient will not notice the attack.
2035 As a work around for this potential attack the length of the decrypted
2036 key must be equal to the cipher default key length, in case the
2037 certificate is not given and all recipientInfo are tried out.
2038 The old behaviour can be re-enabled in the CMS code by setting the
2039 CMS_DEBUG_DECRYPT flag.
2043 * Early start up entropy quality from the DEVRANDOM seed source has been
2044 improved for older Linux systems. The RAND subsystem will wait for
2045 /dev/random to be producing output before seeding from /dev/urandom.
2046 The seeded state is stored for future library initialisations using
2047 a system global shared memory segment. The shared memory identifier
2048 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2049 the desired value. The default identifier is 114.
2053 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2054 when primes for RSA keys are computed.
2055 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2056 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2057 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2058 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2059 This avoids possible fingerprinting of newly generated RSA modules.
2063 * Correct the extended master secret constant on EBCDIC systems. Without this
2064 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2065 negotiate EMS will fail. Unfortunately this also means that TLS connections
2066 between EBCDIC systems with this fix, and EBCDIC systems without this
2067 fix will fail if they negotiate EMS.
2071 * Changed the library initialisation so that the config file is now loaded
2072 by default. This was already the case for libssl. It now occurs for both
2073 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2074 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2078 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2079 where the former acts as a replacement for `ERR_put_error()`, and the
2080 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2081 `ERR_raise_data()` adds more flexibility by taking a format string and
2082 an arbitrary number of arguments following it, to be processed with
2087 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2088 to check if a named provider is loaded and available. When called, it
2089 will also activate all fallback providers if such are still present.
2093 * Enforce a minimum DH modulus size of 512 bits.
2097 * Changed DH parameters to generate the order q subgroup instead of 2q.
2098 Previously generated DH parameters are still accepted by DH_check
2099 but DH_generate_key works around that by clearing bit 0 of the
2100 private key for those. This avoids leaking bit 0 of the private key.
2104 * Significantly reduce secure memory usage by the randomness pools.
2108 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2113 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2114 algorithms. An implementation of a key exchange algorithm can be obtained
2115 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2116 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2117 the older EVP_PKEY_derive_init() function. See the man pages for the new
2118 functions for further details.
2122 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2126 * Removed the function names from error messages and deprecated the
2131 * Removed NextStep support and the macro OPENSSL_UNISTD
2135 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2136 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2137 Also removed "export var as function" capability; we do not export
2138 variables, only functions.
2142 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2143 an error and 1 indicating success. In previous versions of OpenSSL this
2144 was a void type. If a key was set longer than the maximum possible this
2149 * Support SM2 signing and verification schemes with X509 certificate.
2153 * Use SHA256 as the default digest for TS query in the `ts` app.
2157 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2161 * Default cipher lists/suites are now available via a function, the
2162 #defines are deprecated.
2166 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2167 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2168 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2172 * Join the directories crypto/x509 and crypto/x509v3
2176 * Added command 'openssl kdf' that uses the EVP_KDF API.
2180 * Added command 'openssl mac' that uses the EVP_MAC API.
2184 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2185 as default directories. Also added the command 'openssl info'
2186 for scripting purposes.
2190 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2195 * Add prediction resistance to the DRBG reseeding process.
2199 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2200 mandated by IEEE Std 1619-2018.
2204 * Added newline escaping functionality to a filename when using openssl dgst.
2205 This output format is to replicate the output format found in the `*sum`
2206 checksum programs. This aims to preserve backward compatibility.
2208 *Matt Eaton, Richard Levitte, and Paul Dale*
2210 * Removed the heartbeat message in DTLS feature, as it has very
2211 little usage and doesn't seem to fulfill a valuable purpose.
2212 The configuration option is now deprecated.
2216 * Changed the output of 'openssl {digestname} < file' to display the
2217 digest name in its output.
2221 * Added a new generic trace API which provides support for enabling
2222 instrumentation through trace output.
2224 *Richard Levitte & Matthias St. Pierre*
2226 * Added build tests for C++. These are generated files that only do one
2227 thing, to include one public OpenSSL head file each. This tests that
2228 the public header files can be usefully included in a C++ application.
2230 This test isn't enabled by default. It can be enabled with the option
2231 'enable-buildtest-c++'.
2235 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2239 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2243 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2247 * Added KMAC to EVP_MAC.
2251 * Added property based algorithm implementation selection framework to
2256 * Added SCA hardening for modular field inversion in EC_GROUP through
2257 a new dedicated field_inv() pointer in EC_METHOD.
2258 This also addresses a leakage affecting conversions from projective
2259 to affine coordinates.
2261 *Billy Bob Brumley, Nicola Tuveri*
2263 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2264 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2265 those algorithms that were already supported through the EVP_PKEY API
2266 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2267 and scrypt are now wrappers that call EVP_KDF.
2271 * Build devcrypto engine as a dynamic engine.
2273 *Eneas U de Queiroz*
2275 * Add keyed BLAKE2 to EVP_MAC.
2279 * Fix a bug in the computation of the endpoint-pair shared secret used
2280 by DTLS over SCTP. This breaks interoperability with older versions
2281 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2282 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2283 interoperability with such broken implementations. However, enabling
2284 this switch breaks interoperability with correct implementations.
2286 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2287 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2291 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2295 * Changed the license to the Apache License v2.0.
2299 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2301 - Major releases (indicated by incrementing the MAJOR release number)
2302 may introduce incompatible API/ABI changes.
2303 - Minor releases (indicated by incrementing the MINOR release number)
2304 may introduce new features but retain API/ABI compatibility.
2305 - Patch releases (indicated by incrementing the PATCH number)
2306 are intended for bug fixes and other improvements of existing
2307 features only (like improving performance or adding documentation)
2308 and retain API/ABI compatibility.
2312 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2316 * Remove the 'dist' target and add a tarball building script. The
2317 'dist' target has fallen out of use, and it shouldn't be
2318 necessary to configure just to create a source distribution.
2322 * Recreate the OS390-Unix config target. It no longer relies on a
2323 special script like it did for OpenSSL pre-1.1.0.
2327 * Instead of having the source directories listed in Configure, add
2328 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2333 * Add GMAC to EVP_MAC.
2337 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2341 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2342 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2343 to facilitate the continued use of MACs through raw private keys in
2344 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2348 * Deprecate ECDH_KDF_X9_62().
2352 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2353 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2354 are retained for backwards compatibility.
2358 * AES-XTS mode now enforces that its two keys are different to mitigate
2359 the attacked described in "Efficient Instantiations of Tweakable
2360 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2361 Details of this attack can be obtained from:
2362 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2366 * Rename the object files, i.e. give them other names than in previous
2367 versions. Their names now include the name of the final product, as
2368 well as its type mnemonic (bin, lib, shlib).
2372 * Added new option for 'openssl list', '-objects', which will display the
2373 list of built in objects, i.e. OIDs with names.
2377 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2378 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2383 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2384 improves application performance by removing data copies and providing
2385 applications with zero-copy system calls such as sendfile and splice.
2389 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2393 * `PKCS12_parse` now maintains the order of the parsed certificates
2394 when outputting them via `*ca` (rather than reversing it).
2398 * Deprecated pthread fork support methods.
2402 * Added support for FFDHE key exchange in TLS 1.3.
2406 * Added a new concept for OpenSSL plugability: providers. This
2407 functionality is designed to replace the ENGINE API and ENGINE
2408 implementations, and to be much more dynamic, allowing provider
2409 authors to introduce new algorithms among other things, as long as
2410 there's an API that supports the algorithm type.
2412 With this concept comes a new core API for interaction between
2413 libcrypto and provider implementations. Public libcrypto functions
2414 that want to use providers do so through this core API.
2416 The main documentation for this core API is found in
2417 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2418 refer to other manuals describing the API specific for supported
2419 algorithm types (also called operations).
2426 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2428 * Avoid loading of a dynamic engine twice.
2432 * Prioritise DANE TLSA issuer certs over peer certs
2436 * Fixed random API for MacOS prior to 10.12
2438 These MacOS versions don't support the CommonCrypto APIs
2442 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2444 * Fixed an SM2 Decryption Buffer Overflow.
2446 In order to decrypt SM2 encrypted data an application is expected to
2447 call the API function EVP_PKEY_decrypt(). Typically an application will
2448 call this function twice. The first time, on entry, the "out" parameter
2449 can be NULL and, on exit, the "outlen" parameter is populated with the
2450 buffer size required to hold the decrypted plaintext. The application
2451 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2452 again, but this time passing a non-NULL value for the "out" parameter.
2454 A bug in the implementation of the SM2 decryption code means that the
2455 calculation of the buffer size required to hold the plaintext returned
2456 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2457 size required by the second call. This can lead to a buffer overflow
2458 when EVP_PKEY_decrypt() is called by the application a second time with
2459 a buffer that is too small.
2461 A malicious attacker who is able present SM2 content for decryption to
2462 an application could cause attacker chosen data to overflow the buffer
2463 by up to a maximum of 62 bytes altering the contents of other data held
2464 after the buffer, possibly changing application behaviour or causing
2465 the application to crash. The location of the buffer is application
2466 dependent but is typically heap allocated.
2471 * Fixed various read buffer overruns processing ASN.1 strings
2473 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2474 structure which contains a buffer holding the string data and a field
2475 holding the buffer length. This contrasts with normal C strings which
2476 are represented as a buffer for the string data which is terminated
2477 with a NUL (0) byte.
2479 Although not a strict requirement, ASN.1 strings that are parsed using
2480 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2481 well as any string whose value has been set with the ASN1_STRING_set()
2482 function will additionally NUL terminate the byte array in the
2483 ASN1_STRING structure.
2485 However, it is possible for applications to directly construct valid
2486 ASN1_STRING structures which do not NUL terminate the byte array by
2487 directly setting the "data" and "length" fields in the ASN1_STRING
2488 array. This can also happen by using the ASN1_STRING_set0() function.
2490 Numerous OpenSSL functions that print ASN.1 data have been found to
2491 assume that the ASN1_STRING byte array will be NUL terminated, even
2492 though this is not guaranteed for strings that have been directly
2493 constructed. Where an application requests an ASN.1 structure to be
2494 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2495 been directly constructed by the application without NUL terminating
2496 the "data" field, then a read buffer overrun can occur.
2498 The same thing can also occur during name constraints processing
2499 of certificates (for example if a certificate has been directly
2500 constructed by the application instead of loading it via the OpenSSL
2501 parsing functions, and the certificate contains non NUL terminated
2502 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2503 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2505 If a malicious actor can cause an application to directly construct an
2506 ASN1_STRING and then process it through one of the affected OpenSSL
2507 functions then this issue could be hit. This might result in a crash
2508 (causing a Denial of Service attack). It could also result in the
2509 disclosure of private memory contents (such as private keys, or
2510 sensitive plaintext).
2515 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2517 * Fixed a problem with verifying a certificate chain when using the
2518 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2519 the certificates present in a certificate chain. It is not set by default.
2521 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2522 the chain that have explicitly encoded elliptic curve parameters was added
2523 as an additional strict check.
2525 An error in the implementation of this check meant that the result of a
2526 previous check to confirm that certificates in the chain are valid CA
2527 certificates was overwritten. This effectively bypasses the check
2528 that non-CA certificates must not be able to issue other certificates.
2530 If a "purpose" has been configured then there is a subsequent opportunity
2531 for checks that the certificate is a valid CA. All of the named "purpose"
2532 values implemented in libcrypto perform this check. Therefore, where
2533 a purpose is set the certificate chain will still be rejected even when the
2534 strict flag has been used. A purpose is set by default in libssl client and
2535 server certificate verification routines, but it can be overridden or
2536 removed by an application.
2538 In order to be affected, an application must explicitly set the
2539 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2540 for the certificate verification or, in the case of TLS client or server
2541 applications, override the default purpose.
2546 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2547 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2548 renegotiation ClientHello omits the signature_algorithms extension (where it
2549 was present in the initial ClientHello), but includes a
2550 signature_algorithms_cert extension then a NULL pointer dereference will
2551 result, leading to a crash and a denial of service attack.
2553 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2554 (which is the default configuration). OpenSSL TLS clients are not impacted by
2558 *Peter Kästle and Samuel Sapalski*
2560 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2562 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2563 create a unique hash value based on the issuer and serial number data
2564 contained within an X509 certificate. However, it was failing to correctly
2565 handle any errors that may occur while parsing the issuer field (which might
2566 occur if the issuer field is maliciously constructed). This may subsequently
2567 result in a NULL pointer deref and a crash leading to a potential denial of
2573 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2574 padding mode to correctly check for rollback attacks. This is considered a
2575 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2580 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2581 functions. Previously they could overflow the output length argument in some
2582 cases where the input length is close to the maximum permissible length for
2583 an integer on the platform. In such cases the return value from the function
2584 call would be 1 (indicating success), but the output length value would be
2585 negative. This could cause applications to behave incorrectly or crash.
2590 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2591 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2592 could be exploited in a side channel attack to recover the password. Since
2593 the attack is local host only this is outside of the current OpenSSL
2594 threat model and therefore no CVE is assigned.
2596 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2601 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2603 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2604 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2605 If an attacker can control both items being compared then this could lead
2606 to a possible denial of service attack. OpenSSL itself uses the
2607 GENERAL_NAME_cmp function for two purposes:
2608 1) Comparing CRL distribution point names between an available CRL and a
2609 CRL distribution point embedded in an X509 certificate
2610 2) When verifying that a timestamp response token signer matches the
2611 timestamp authority name (exposed via the API functions
2612 TS_RESP_verify_response and TS_RESP_verify_token)
2617 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2619 * Certificates with explicit curve parameters are now disallowed in
2620 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2624 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2625 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2626 conversely, silently ignore DTLS protocol version bounds when configuring
2627 TLS-based contexts. The commands can be repeated to set bounds of both
2628 types. The same applies with the corresponding "min_protocol" and
2629 "max_protocol" command-line switches, in case some application uses both TLS
2632 SSL_CTX instances that are created for a fixed protocol version (e.g.
2633 TLSv1_server_method()) also silently ignore version bounds. Previously
2634 attempts to apply bounds to these protocol versions would result in an
2635 error. Now only the "version-flexible" SSL_CTX instances are subject to
2636 limits in configuration files in command-line options.
2640 * Handshake now fails if Extended Master Secret extension is dropped
2645 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2647 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2649 * Fixed segmentation fault in SSL_check_chain()
2650 Server or client applications that call the SSL_check_chain() function
2651 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2652 dereference as a result of incorrect handling of the
2653 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2654 or unrecognised signature algorithm is received from the peer. This could
2655 be exploited by a malicious peer in a Denial of Service attack.
2660 * Added AES consttime code for no-asm configurations
2661 an optional constant time support for AES was added
2662 when building openssl for no-asm.
2663 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2664 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2665 At this time this feature is by default disabled.
2666 It will be enabled by default in 3.0.
2670 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2672 * Revert the change of EOF detection while reading in libssl to avoid
2673 regressions in applications depending on the current way of reporting
2674 the EOF. As the existing method is not fully accurate the change to
2675 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2676 branch and will be present in the 3.0 release.
2680 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2681 when primes for RSA keys are computed.
2682 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2683 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2684 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2685 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2686 This avoids possible fingerprinting of newly generated RSA modules.
2690 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2692 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2693 while reading in libssl then we would report an error back to the
2694 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2695 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2696 therefore give a hint as to what went wrong.
2700 * Check that ed25519 and ed448 are allowed by the security level. Previously
2701 signature algorithms not using an MD were not being checked that they were
2702 allowed by the security level.
2706 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2707 was not quite right. The behaviour was not consistent between resumption
2708 and normal handshakes, and also not quite consistent with historical
2709 behaviour. The behaviour in various scenarios has been clarified and
2710 it has been updated to make it match historical behaviour as closely as
2715 * *[VMS only]* The header files that the VMS compilers include automatically,
2716 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2717 that the C++ compiler doesn't understand. This is a shortcoming in the
2718 compiler, but can be worked around with `__cplusplus` guards.
2720 C++ applications that use OpenSSL libraries must be compiled using the
2721 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2722 functions. Otherwise, only functions with symbols of less than 31
2723 characters can be used, as the linker will not be able to successfully
2724 resolve symbols with longer names.
2728 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2729 The presence of this system service is determined at run-time.
2733 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2738 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2740 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2741 number generator (RNG). This was intended to include protection in the
2742 event of a fork() system call in order to ensure that the parent and child
2743 processes did not share the same RNG state. However, this protection was not
2744 being used in the default case.
2746 A partial mitigation for this issue is that the output from a high
2747 precision timer is mixed into the RNG state so the likelihood of a parent
2748 and child process sharing state is significantly reduced.
2750 If an application already calls OPENSSL_init_crypto() explicitly using
2751 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2754 *Matthias St. Pierre*
2756 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2757 used even when parsing explicit parameters, when loading a encoded key
2758 or calling `EC_GROUP_new_from_ecpkparameters()`/
2759 `EC_GROUP_new_from_ecparameters()`.
2760 This prevents bypass of security hardening and performance gains,
2761 especially for curves with specialized EC_METHODs.
2762 By default, if a key encoded with explicit parameters is loaded and later
2763 encoded, the output is still encoded with explicit parameters, even if
2764 internally a "named" EC_GROUP is used for computation.
2768 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2769 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2770 NULL. After this change, only the cofactor parameter can be NULL. It also
2771 does some minimal sanity checks on the passed order.
2776 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2777 An attack is simple, if the first CMS_recipientInfo is valid but the
2778 second CMS_recipientInfo is chosen ciphertext. If the second
2779 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2780 encryption key will be replaced by garbage, and the message cannot be
2781 decoded, but if the RSA decryption fails, the correct encryption key is
2782 used and the recipient will not notice the attack.
2783 As a work around for this potential attack the length of the decrypted
2784 key must be equal to the cipher default key length, in case the
2785 certificate is not given and all recipientInfo are tried out.
2786 The old behaviour can be re-enabled in the CMS code by setting the
2787 CMS_DEBUG_DECRYPT flag.
2792 * Early start up entropy quality from the DEVRANDOM seed source has been
2793 improved for older Linux systems. The RAND subsystem will wait for
2794 /dev/random to be producing output before seeding from /dev/urandom.
2795 The seeded state is stored for future library initialisations using
2796 a system global shared memory segment. The shared memory identifier
2797 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2798 the desired value. The default identifier is 114.
2802 * Correct the extended master secret constant on EBCDIC systems. Without this
2803 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2804 negotiate EMS will fail. Unfortunately this also means that TLS connections
2805 between EBCDIC systems with this fix, and EBCDIC systems without this
2806 fix will fail if they negotiate EMS.
2810 * Use Windows installation paths in the mingw builds
2812 Mingw isn't a POSIX environment per se, which means that Windows
2813 paths should be used for installation.
2818 * Changed DH_check to accept parameters with order q and 2q subgroups.
2819 With order 2q subgroups the bit 0 of the private key is not secret
2820 but DH_generate_key works around that by clearing bit 0 of the
2821 private key for those. This avoids leaking bit 0 of the private key.
2825 * Significantly reduce secure memory usage by the randomness pools.
2829 * Revert the DEVRANDOM_WAIT feature for Linux systems
2831 The DEVRANDOM_WAIT feature added a select() call to wait for the
2832 /dev/random device to become readable before reading from the
2833 /dev/urandom device.
2835 It turned out that this change had negative side effects on
2836 performance which were not acceptable. After some discussion it
2837 was decided to revert this feature and leave it up to the OS
2838 resp. the platform maintainer to ensure a proper initialization
2839 during early boot time.
2841 *Matthias St. Pierre*
2843 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2845 * Add build tests for C++. These are generated files that only do one
2846 thing, to include one public OpenSSL head file each. This tests that
2847 the public header files can be usefully included in a C++ application.
2849 This test isn't enabled by default. It can be enabled with the option
2850 'enable-buildtest-c++'.
2854 * Enable SHA3 pre-hashing for ECDSA and DSA.
2858 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2859 This changes the size when using the `genpkey` command when no size is given.
2860 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2861 generation commands to use 2048 bits by default.
2865 * Reorganize the manual pages to consistently have RETURN VALUES,
2866 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2867 util/fix-doc-nits accordingly.
2869 *Paul Yang, Joshua Lock*
2871 * Add the missing accessor EVP_PKEY_get0_engine()
2875 * Have commands like `s_client` and `s_server` output the signature scheme
2876 along with other cipher suite parameters when debugging.
2880 * Make OPENSSL_config() error agnostic again.
2884 * Do the error handling in RSA decryption constant time.
2888 * Prevent over long nonces in ChaCha20-Poly1305.
2890 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2891 for every encryption operation. RFC 7539 specifies that the nonce value
2892 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2893 and front pads the nonce with 0 bytes if it is less than 12
2894 bytes. However it also incorrectly allows a nonce to be set of up to 16
2895 bytes. In this case only the last 12 bytes are significant and any
2896 additional leading bytes are ignored.
2898 It is a requirement of using this cipher that nonce values are
2899 unique. Messages encrypted using a reused nonce value are susceptible to
2900 serious confidentiality and integrity attacks. If an application changes
2901 the default nonce length to be longer than 12 bytes and then makes a
2902 change to the leading bytes of the nonce expecting the new value to be a
2903 new unique nonce then such an application could inadvertently encrypt
2904 messages with a reused nonce.
2906 Additionally the ignored bytes in a long nonce are not covered by the
2907 integrity guarantee of this cipher. Any application that relies on the
2908 integrity of these ignored leading bytes of a long nonce may be further
2909 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2910 is safe because no such use sets such a long nonce value. However user
2911 applications that use this cipher directly and set a non-default nonce
2912 length to be longer than 12 bytes may be vulnerable.
2914 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2920 * Add DEVRANDOM_WAIT feature for Linux systems
2922 On older Linux systems where the getrandom() system call is not available,
2923 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2924 Contrary to getrandom(), the /dev/urandom device will not block during
2925 early boot when the kernel CSPRNG has not been seeded yet.
2927 To mitigate this known weakness, use select() to wait for /dev/random to
2928 become readable before reading from /dev/urandom.
2930 * Ensure that SM2 only uses SM3 as digest algorithm
2934 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2936 * Change the info callback signals for the start and end of a post-handshake
2937 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2938 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2939 confused by this and assume that a TLSv1.2 renegotiation has started. This
2940 can break KeyUpdate handling. Instead we no longer signal the start and end
2941 of a post handshake message exchange (although the messages themselves are
2942 still signalled). This could break some applications that were expecting
2943 the old signals. However without this KeyUpdate is not usable for many
2948 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2950 * Timing vulnerability in DSA signature generation
2952 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2953 timing side channel attack. An attacker could use variations in the signing
2954 algorithm to recover the private key.
2956 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2961 * Timing vulnerability in ECDSA signature generation
2963 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2964 timing side channel attack. An attacker could use variations in the signing
2965 algorithm to recover the private key.
2967 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2972 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2973 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2974 of two gigabytes and the error handling improved.
2976 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2977 categorized as a normal bug, not a security issue, because the DRBG reseeds
2978 automatically and is fully functional even without additional randomness
2979 provided by the application.
2981 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2983 * Add a new ClientHello callback. Provides a callback interface that gives
2984 the application the ability to adjust the nascent SSL object at the
2985 earliest stage of ClientHello processing, immediately after extensions have
2986 been collected but before they have been processed. In particular, this
2987 callback can adjust the supported TLS versions in response to the contents
2992 * Add SM2 base algorithm support.
2996 * s390x assembly pack: add (improved) hardware-support for the following
2997 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2998 aes-cfb/cfb8, aes-ecb.
3002 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3003 parameter is no longer accepted, as it leads to a corrupt table. NULL
3004 pem_str is reserved for alias entries only.
3008 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3009 step for prime curves. The new implementation is based on formulae from
3010 differential addition-and-doubling in homogeneous projective coordinates
3011 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3012 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3013 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3014 to work in projective coordinates.
3016 *Billy Bob Brumley, Nicola Tuveri*
3018 * Change generating and checking of primes so that the error rate of not
3019 being prime depends on the intended use based on the size of the input.
3020 For larger primes this will result in more rounds of Miller-Rabin.
3021 The maximal error rate for primes with more than 1080 bits is lowered
3024 *Kurt Roeckx, Annie Yousar*
3026 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3030 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3031 moving between systems, and to avoid confusion when a Windows build is
3032 done with mingw vs with MSVC. For POSIX installs, there's still a
3033 symlink or copy named 'tsget' to avoid that confusion as well.
3037 * Revert blinding in ECDSA sign and instead make problematic addition
3038 length-invariant. Switch even to fixed-length Montgomery multiplication.
3042 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3043 step for binary curves. The new implementation is based on formulae from
3044 differential addition-and-doubling in mixed Lopez-Dahab projective
3045 coordinates, modified to independently blind the operands.
3047 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3049 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3050 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3051 EC_METHODs to implement their own specialized "ladder step", to take
3052 advantage of more favorable coordinate systems or more efficient
3053 differential addition-and-doubling algorithms.
3055 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3057 * Modified the random device based seed sources to keep the relevant
3058 file descriptors open rather than reopening them on each access.
3059 This allows such sources to operate in a chroot() jail without
3060 the associated device nodes being available. This behaviour can be
3061 controlled using RAND_keep_random_devices_open().
3065 * Numerous side-channel attack mitigations have been applied. This may have
3066 performance impacts for some algorithms for the benefit of improved
3067 security. Specific changes are noted in this change log by their respective
3072 * AIX shared library support overhaul. Switch to AIX "natural" way of
3073 handling shared libraries, which means collecting shared objects of
3074 different versions and bitnesses in one common archive. This allows to
3075 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3076 doesn't affect the way 3rd party applications are linked, only how
3077 multi-version installation is managed.
3081 * Make ec_group_do_inverse_ord() more robust and available to other
3082 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3083 mitigations are applied to the fallback BN_mod_inverse().
3084 When using this function rather than BN_mod_inverse() directly, new
3085 EC cryptosystem implementations are then safer-by-default.
3089 * Add coordinate blinding for EC_POINT and implement projective
3090 coordinate blinding for generic prime curves as a countermeasure to
3091 chosen point SCA attacks.
3093 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3095 * Add blinding to ECDSA and DSA signatures to protect against side channel
3096 attacks discovered by Keegan Ryan (NCC Group).
3100 * Enforce checking in the `pkeyutl` command to ensure that the input
3101 length does not exceed the maximum supported digest length when performing
3102 a sign, verify or verifyrecover operation.
3106 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3107 I/O in combination with something like select() or poll() will hang. This
3108 can be turned off again using SSL_CTX_clear_mode().
3109 Many applications do not properly handle non-application data records, and
3110 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3111 around the problems in those applications, but can also break some.
3112 It's recommended to read the manpages about SSL_read(), SSL_write(),
3113 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3114 SSL_CTX_set_read_ahead() again.
3118 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3119 now allow empty (zero character) pass phrases.
3123 * Apply blinding to binary field modular inversion and remove patent
3124 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3128 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3129 binary and prime elliptic curves.
3133 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3134 constant time fixed point multiplication.
3138 * Revise elliptic curve scalar multiplication with timing attack
3139 defenses: ec_wNAF_mul redirects to a constant time implementation
3140 when computing fixed point and variable point multiplication (which
3141 in OpenSSL are mostly used with secret scalars in keygen, sign,
3142 ECDH derive operations).
3143 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3146 * Updated CONTRIBUTING
3150 * Updated DRBG / RAND to request nonce and additional low entropy
3151 randomness from the system.
3153 *Matthias St. Pierre*
3155 * Updated 'openssl rehash' to use OpenSSL consistent default.
3159 * Moved the load of the ssl_conf module to libcrypto, which helps
3160 loading engines that libssl uses before libssl is initialised.
3164 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3168 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3170 *Ingo Schwarze, Rich Salz*
3172 * Added output of accepting IP address and port for 'openssl s_server'
3176 * Added a new API for TLSv1.3 ciphersuites:
3177 SSL_CTX_set_ciphersuites()
3178 SSL_set_ciphersuites()
3182 * Memory allocation failures consistently add an error to the error
3187 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3188 in libcrypto when run as setuid/setgid.
3192 * Load any config file by default when libssl is used.
3196 * Added new public header file <openssl/rand_drbg.h> and documentation
3197 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3199 *Matthias St. Pierre*
3201 * QNX support removed (cannot find contributors to get their approval
3202 for the license change).
3206 * TLSv1.3 replay protection for early data has been implemented. See the
3207 SSL_read_early_data() man page for further details.
3211 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3212 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3213 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3214 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3215 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3216 configuration has been separated out. See the ciphers man page or the
3217 SSL_CTX_set_ciphersuites() man page for more information.
3221 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3222 in responder mode now supports the new "-multi" option, which
3223 spawns the specified number of child processes to handle OCSP
3224 requests. The "-timeout" option now also limits the OCSP
3225 responder's patience to wait to receive the full client request
3226 on a newly accepted connection. Child processes are respawned
3227 as needed, and the CA index file is automatically reloaded
3228 when changed. This makes it possible to run the "ocsp" responder
3229 as a long-running service, making the OpenSSL CA somewhat more
3230 feature-complete. In this mode, most diagnostic messages logged
3231 after entering the event loop are logged via syslog(3) rather than
3236 * Added support for X448 and Ed448. Heavily based on original work by
3241 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3242 objects loaded. This adds the functions OSSL_STORE_expect() and
3243 OSSL_STORE_find() as well as needed tools to construct searches and
3244 get the search data out of them.
3248 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3249 version of OpenSSL should review their configuration settings to ensure
3250 that they are still appropriate for TLSv1.3. For further information see:
3251 <https://wiki.openssl.org/index.php/TLS1.3>
3255 * Grand redesign of the OpenSSL random generator
3257 The default RAND method now utilizes an AES-CTR DRBG according to
3258 NIST standard SP 800-90Ar1. The new random generator is essentially
3259 a port of the default random generator from the OpenSSL FIPS 2.0
3260 object module. It is a hybrid deterministic random bit generator
3261 using an AES-CTR bit stream and which seeds and reseeds itself
3262 automatically using trusted system entropy sources.
3264 Some of its new features are:
3265 - Support for multiple DRBG instances with seed chaining.
3266 - The default RAND method makes use of a DRBG.
3267 - There is a public and private DRBG instance.
3268 - The DRBG instances are fork-safe.
3269 - Keep all global DRBG instances on the secure heap if it is enabled.
3270 - The public and private DRBG instance are per thread for lock free
3273 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3275 * Changed Configure so it only says what it does and doesn't dump
3276 so much data. Instead, ./configdata.pm should be used as a script
3277 to display all sorts of configuration data.
3281 * Added processing of "make variables" to Configure.
3285 * Added SHA512/224 and SHA512/256 algorithm support.
3289 * The last traces of Netware support, first removed in 1.1.0, have
3294 * Get rid of Makefile.shared, and in the process, make the processing
3295 of certain files (rc.obj, or the .def/.map/.opt files produced from
3296 the ordinal files) more visible and hopefully easier to trace and
3297 debug (or make silent).
3301 * Make it possible to have environment variable assignments as
3302 arguments to config / Configure.
3306 * Add multi-prime RSA (RFC 8017) support.
3310 * Add SM3 implemented according to GB/T 32905-2016
3311 *Jack Lloyd <jack.lloyd@ribose.com>,*
3312 *Ronald Tse <ronald.tse@ribose.com>,*
3313 *Erick Borsboom <erick.borsboom@ribose.com>*
3315 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3316 as documented in RFC6066.
3317 Based on a patch from Tomasz Moń
3319 *Filipe Raimundo da Silva*
3321 * Add SM4 implemented according to GB/T 32907-2016.
3322 *Jack Lloyd <jack.lloyd@ribose.com>,*
3323 *Ronald Tse <ronald.tse@ribose.com>,*
3324 *Erick Borsboom <erick.borsboom@ribose.com>*
3326 * Reimplement -newreq-nodes and ERR_error_string_n; the
3327 original author does not agree with the license change.
3331 * Add ARIA AEAD TLS support.
3335 * Some macro definitions to support VS6 have been removed. Visual
3336 Studio 6 has not worked since 1.1.0
3340 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3341 without clearing the errors.
3345 * Add "atfork" functions. If building on a system that without
3346 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3347 requirements. The RAND facility now uses/requires this.
3355 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3356 not possible to disable entirely. However, it's still possible to
3357 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3360 To disable, configure with 'no-ui-console'. 'no-ui' is still
3361 possible to use as an alias. Check at compile time with the
3362 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3363 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3367 * Add a STORE module, which implements a uniform and URI based reader of
3368 stores that can contain keys, certificates, CRLs and numerous other
3369 objects. The main API is loosely based on a few stdio functions,
3370 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3371 OSSL_STORE_error and OSSL_STORE_close.
3372 The implementation uses backends called "loaders" to implement arbitrary
3373 URI schemes. There is one built in "loader" for the 'file' scheme.
3377 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3378 then adjusted to work on FreeBSD 8.4 as well.
3379 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3380 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3384 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3385 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3386 error code calls like this:
3388 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3390 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3391 that can be encoded in C. For the foreseeable future, this will only
3394 *Richard Levitte and Tim Hudson*
3396 * Removed BSD cryptodev engine.
3400 * Add a build target 'build_all_generated', to build all generated files
3401 and only that. This can be used to prepare everything that requires
3402 things like perl for a system that lacks perl and then move everything
3403 to that system and do the rest of the build there.
3407 * In the UI interface, make it possible to duplicate the user data. This
3408 can be used by engines that need to retain the data for a longer time
3409 than just the call where this user data is passed.
3413 * Ignore the '-named_curve auto' value for compatibility of applications
3416 *Tomáš Mráz <tmraz@fedoraproject.org>*
3418 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3419 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3420 alerts across multiple records (some of which could be empty). In practice
3421 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3422 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3423 support this at all. Supporting it adds significant complexity to the
3424 record layer, and its removal is unlikely to cause interoperability
3429 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3430 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3431 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3436 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3437 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3439 *Richard Levitte, Andy Polyakov*
3441 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3446 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3447 platform rather than 'mingw'.
3451 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3452 success if they are asked to add an object which already exists
3453 in the store. This change cascades to other functions which load
3454 certificates and CRLs.
3458 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3459 facilitate stack unwinding even from assembly subroutines.
3463 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3464 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3468 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3469 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3470 which is the minimum version we support.
3474 * Certificate time validation (X509_cmp_time) enforces stricter
3475 compliance with RFC 5280. Fractional seconds and timezone offsets
3476 are no longer allowed.
3480 * Add support for ARIA
3484 * s_client will now send the Server Name Indication (SNI) extension by
3485 default unless the new "-noservername" option is used. The server name is
3486 based on the host provided to the "-connect" option unless overridden by
3487 using "-servername".
3491 * Add support for SipHash
3495 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3496 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3497 prevent issues where no progress is being made and the peer continually
3498 sends unrecognised record types, using up resources processing them.
3502 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3503 using the algorithm defined in
3504 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3508 * Heartbeat support has been removed; the ABI is changed for now.
3510 *Richard Levitte, Rich Salz*
3512 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3516 * The RSA "null" method, which was partially supported to avoid patent
3517 issues, has been replaced to always returns NULL.
3524 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3526 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3527 used even when parsing explicit parameters, when loading a encoded key
3528 or calling `EC_GROUP_new_from_ecpkparameters()`/
3529 `EC_GROUP_new_from_ecparameters()`.
3530 This prevents bypass of security hardening and performance gains,
3531 especially for curves with specialized EC_METHODs.
3532 By default, if a key encoded with explicit parameters is loaded and later
3533 encoded, the output is still encoded with explicit parameters, even if
3534 internally a "named" EC_GROUP is used for computation.
3538 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3539 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3540 NULL. After this change, only the cofactor parameter can be NULL. It also
3541 does some minimal sanity checks on the passed order.
3546 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3547 An attack is simple, if the first CMS_recipientInfo is valid but the
3548 second CMS_recipientInfo is chosen ciphertext. If the second
3549 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3550 encryption key will be replaced by garbage, and the message cannot be
3551 decoded, but if the RSA decryption fails, the correct encryption key is
3552 used and the recipient will not notice the attack.
3553 As a work around for this potential attack the length of the decrypted
3554 key must be equal to the cipher default key length, in case the
3555 certificate is not given and all recipientInfo are tried out.
3556 The old behaviour can be re-enabled in the CMS code by setting the
3557 CMS_DEBUG_DECRYPT flag.
3562 * Use Windows installation paths in the mingw builds
3564 Mingw isn't a POSIX environment per se, which means that Windows
3565 paths should be used for installation.
3570 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3572 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3573 This changes the size when using the `genpkey` command when no size is given.
3574 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3575 generation commands to use 2048 bits by default.
3579 * Prevent over long nonces in ChaCha20-Poly1305.
3581 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3582 for every encryption operation. RFC 7539 specifies that the nonce value
3583 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3584 and front pads the nonce with 0 bytes if it is less than 12
3585 bytes. However it also incorrectly allows a nonce to be set of up to 16
3586 bytes. In this case only the last 12 bytes are significant and any
3587 additional leading bytes are ignored.
3589 It is a requirement of using this cipher that nonce values are
3590 unique. Messages encrypted using a reused nonce value are susceptible to
3591 serious confidentiality and integrity attacks. If an application changes
3592 the default nonce length to be longer than 12 bytes and then makes a
3593 change to the leading bytes of the nonce expecting the new value to be a
3594 new unique nonce then such an application could inadvertently encrypt
3595 messages with a reused nonce.
3597 Additionally the ignored bytes in a long nonce are not covered by the
3598 integrity guarantee of this cipher. Any application that relies on the
3599 integrity of these ignored leading bytes of a long nonce may be further
3600 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3601 is safe because no such use sets such a long nonce value. However user
3602 applications that use this cipher directly and set a non-default nonce
3603 length to be longer than 12 bytes may be vulnerable.
3605 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3611 * Added SCA hardening for modular field inversion in EC_GROUP through
3612 a new dedicated field_inv() pointer in EC_METHOD.
3613 This also addresses a leakage affecting conversions from projective
3614 to affine coordinates.
3616 *Billy Bob Brumley, Nicola Tuveri*
3618 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3619 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3623 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3627 * Remove the 'dist' target and add a tarball building script. The
3628 'dist' target has fallen out of use, and it shouldn't be
3629 necessary to configure just to create a source distribution.
3633 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3635 * Timing vulnerability in DSA signature generation
3637 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3638 timing side channel attack. An attacker could use variations in the signing
3639 algorithm to recover the private key.
3641 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3646 * Timing vulnerability in ECDSA signature generation
3648 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3649 timing side channel attack. An attacker could use variations in the signing
3650 algorithm to recover the private key.
3652 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3657 * Add coordinate blinding for EC_POINT and implement projective
3658 coordinate blinding for generic prime curves as a countermeasure to
3659 chosen point SCA attacks.
3661 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3663 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3665 * Client DoS due to large DH parameter
3667 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3668 malicious server can send a very large prime value to the client. This will
3669 cause the client to spend an unreasonably long period of time generating a
3670 key for this prime resulting in a hang until the client has finished. This
3671 could be exploited in a Denial Of Service attack.
3673 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3678 * Cache timing vulnerability in RSA Key Generation
3680 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3681 a cache timing side channel attack. An attacker with sufficient access to
3682 mount cache timing attacks during the RSA key generation process could
3683 recover the private key.
3685 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3686 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3691 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3692 parameter is no longer accepted, as it leads to a corrupt table. NULL
3693 pem_str is reserved for alias entries only.
3697 * Revert blinding in ECDSA sign and instead make problematic addition
3698 length-invariant. Switch even to fixed-length Montgomery multiplication.
3702 * Change generating and checking of primes so that the error rate of not
3703 being prime depends on the intended use based on the size of the input.
3704 For larger primes this will result in more rounds of Miller-Rabin.
3705 The maximal error rate for primes with more than 1080 bits is lowered
3708 *Kurt Roeckx, Annie Yousar*
3710 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3714 * Add blinding to ECDSA and DSA signatures to protect against side channel
3715 attacks discovered by Keegan Ryan (NCC Group).
3719 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3720 now allow empty (zero character) pass phrases.
3724 * Certificate time validation (X509_cmp_time) enforces stricter
3725 compliance with RFC 5280. Fractional seconds and timezone offsets
3726 are no longer allowed.
3730 * Fixed a text canonicalisation bug in CMS
3732 Where a CMS detached signature is used with text content the text goes
3733 through a canonicalisation process first prior to signing or verifying a
3734 signature. This process strips trailing space at the end of lines, converts
3735 line terminators to CRLF and removes additional trailing line terminators
3736 at the end of a file. A bug in the canonicalisation process meant that
3737 some characters, such as form-feed, were incorrectly treated as whitespace
3738 and removed. This is contrary to the specification (RFC5485). This fix
3739 could mean that detached text data signed with an earlier version of
3740 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3741 signed with a fixed OpenSSL may fail to verify with an earlier version of
3742 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3743 and use the "-binary" flag (for the "cms" command line application) or set
3744 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3748 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3750 * Constructed ASN.1 types with a recursive definition could exceed the stack
3752 Constructed ASN.1 types with a recursive definition (such as can be found
3753 in PKCS7) could eventually exceed the stack given malicious input with
3754 excessive recursion. This could result in a Denial Of Service attack. There
3755 are no such structures used within SSL/TLS that come from untrusted sources
3756 so this is considered safe.
3758 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3764 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3766 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3767 effectively reduced to only comparing the least significant bit of each
3768 byte. This allows an attacker to forge messages that would be considered as
3769 authenticated in an amount of tries lower than that guaranteed by the
3770 security claims of the scheme. The module can only be compiled by the
3771 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3773 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3779 * Add a build target 'build_all_generated', to build all generated files
3780 and only that. This can be used to prepare everything that requires
3781 things like perl for a system that lacks perl and then move everything
3782 to that system and do the rest of the build there.
3786 * Backport SSL_OP_NO_RENGOTIATION
3788 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3789 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3790 changes this is no longer possible in 1.1.0. Therefore, the new
3791 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3792 1.1.0 to provide equivalent functionality.
3794 Note that if an application built against 1.1.0h headers (or above) is run
3795 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3796 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3800 * Removed the OS390-Unix config target. It relied on a script that doesn't
3805 * rsaz_1024_mul_avx2 overflow bug on x86_64
3807 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3808 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3809 Analysis suggests that attacks against RSA and DSA as a result of this
3810 defect would be very difficult to perform and are not believed likely.
3811 Attacks against DH1024 are considered just feasible, because most of the
3812 work necessary to deduce information about a private key may be performed
3813 offline. The amount of resources required for such an attack would be
3814 significant. However, for an attack on TLS to be meaningful, the server
3815 would have to share the DH1024 private key among multiple clients, which is
3816 no longer an option since CVE-2016-0701.
3818 This only affects processors that support the AVX2 but not ADX extensions
3819 like Intel Haswell (4th generation).
3821 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3822 was originally found via the OSS-Fuzz project.
3827 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3829 * bn_sqrx8x_internal carry bug on x86_64
3831 There is a carry propagating bug in the x86_64 Montgomery squaring
3832 procedure. No EC algorithms are affected. Analysis suggests that attacks
3833 against RSA and DSA as a result of this defect would be very difficult to
3834 perform and are not believed likely. Attacks against DH are considered just
3835 feasible (although very difficult) because most of the work necessary to
3836 deduce information about a private key may be performed offline. The amount
3837 of resources required for such an attack would be very significant and
3838 likely only accessible to a limited number of attackers. An attacker would
3839 additionally need online access to an unpatched system using the target
3840 private key in a scenario with persistent DH parameters and a private
3841 key that is shared between multiple clients.
3843 This only affects processors that support the BMI1, BMI2 and ADX extensions
3844 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3846 This issue was reported to OpenSSL by the OSS-Fuzz project.
3851 * Malformed X.509 IPAddressFamily could cause OOB read
3853 If an X.509 certificate has a malformed IPAddressFamily extension,
3854 OpenSSL could do a one-byte buffer overread. The most likely result
3855 would be an erroneous display of the certificate in text format.
3857 This issue was reported to OpenSSL by the OSS-Fuzz project.
3862 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3864 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3865 platform rather than 'mingw'.
3869 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3870 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3871 which is the minimum version we support.
3875 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3877 * Encrypt-Then-Mac renegotiation crash
3879 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3880 negotiated where it was not in the original handshake (or vice-versa) then
3881 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3882 and servers are affected.
3884 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3889 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3891 * Truncated packet could crash via OOB read
3893 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3894 cipher is being used, then a truncated packet can cause that host to
3895 perform an out-of-bounds read, usually resulting in a crash.
3897 This issue was reported to OpenSSL by Robert Święcki of Google.
3902 * Bad (EC)DHE parameters cause a client crash
3904 If a malicious server supplies bad parameters for a DHE or ECDHE key
3905 exchange then this can result in the client attempting to dereference a
3906 NULL pointer leading to a client crash. This could be exploited in a Denial
3909 This issue was reported to OpenSSL by Guido Vranken.
3914 * BN_mod_exp may produce incorrect results on x86_64
3916 There is a carry propagating bug in the x86_64 Montgomery squaring
3917 procedure. No EC algorithms are affected. Analysis suggests that attacks
3918 against RSA and DSA as a result of this defect would be very difficult to
3919 perform and are not believed likely. Attacks against DH are considered just
3920 feasible (although very difficult) because most of the work necessary to
3921 deduce information about a private key may be performed offline. The amount
3922 of resources required for such an attack would be very significant and
3923 likely only accessible to a limited number of attackers. An attacker would
3924 additionally need online access to an unpatched system using the target
3925 private key in a scenario with persistent DH parameters and a private
3926 key that is shared between multiple clients. For example this can occur by
3927 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3928 similar to CVE-2015-3193 but must be treated as a separate problem.
3930 This issue was reported to OpenSSL by the OSS-Fuzz project.
3935 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3937 * ChaCha20/Poly1305 heap-buffer-overflow
3939 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3940 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3941 crash. This issue is not considered to be exploitable beyond a DoS.
3943 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3948 * CMS Null dereference
3950 Applications parsing invalid CMS structures can crash with a NULL pointer
3951 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3952 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3953 structure callback if an attempt is made to free certain invalid encodings.
3954 Only CHOICE structures using a callback which do not handle NULL value are
3957 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3962 * Montgomery multiplication may produce incorrect results
3964 There is a carry propagating bug in the Broadwell-specific Montgomery
3965 multiplication procedure that handles input lengths divisible by, but
3966 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3967 and DH private keys are impossible. This is because the subroutine in
3968 question is not used in operations with the private key itself and an input
3969 of the attacker's direct choice. Otherwise the bug can manifest itself as
3970 transient authentication and key negotiation failures or reproducible
3971 erroneous outcome of public-key operations with specially crafted input.
3972 Among EC algorithms only Brainpool P-512 curves are affected and one
3973 presumably can attack ECDH key negotiation. Impact was not analyzed in
3974 detail, because pre-requisites for attack are considered unlikely. Namely
3975 multiple clients have to choose the curve in question and the server has to
3976 share the private key among them, neither of which is default behaviour.
3977 Even then only clients that chose the curve will be affected.
3979 This issue was publicly reported as transient failures and was not
3980 initially recognized as a security issue. Thanks to Richard Morgan for
3981 providing reproducible case.
3986 * Removed automatic addition of RPATH in shared libraries and executables,
3987 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3991 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3993 * Fix Use After Free for large message sizes
3995 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3996 message larger than approx 16k is received then the underlying buffer to
3997 store the incoming message is reallocated and moved. Unfortunately a
3998 dangling pointer to the old location is left which results in an attempt to
3999 write to the previously freed location. This is likely to result in a
4000 crash, however it could potentially lead to execution of arbitrary code.
4002 This issue only affects OpenSSL 1.1.0a.
4004 This issue was reported to OpenSSL by Robert Święcki.
4009 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4011 * OCSP Status Request extension unbounded memory growth
4013 A malicious client can send an excessively large OCSP Status Request
4014 extension. If that client continually requests renegotiation, sending a
4015 large OCSP Status Request extension each time, then there will be unbounded
4016 memory growth on the server. This will eventually lead to a Denial Of
4017 Service attack through memory exhaustion. Servers with a default
4018 configuration are vulnerable even if they do not support OCSP. Builds using
4019 the "no-ocsp" build time option are not affected.
4021 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4026 * SSL_peek() hang on empty record
4028 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4029 sends an empty record. This could be exploited by a malicious peer in a
4030 Denial Of Service attack.
4032 This issue was reported to OpenSSL by Alex Gaynor.
4037 * Excessive allocation of memory in tls_get_message_header() and
4038 dtls1_preprocess_fragment()
4040 A (D)TLS message includes 3 bytes for its length in the header for the
4041 message. This would allow for messages up to 16Mb in length. Messages of
4042 this length are excessive and OpenSSL includes a check to ensure that a
4043 peer is sending reasonably sized messages in order to avoid too much memory
4044 being consumed to service a connection. A flaw in the logic of version
4045 1.1.0 means that memory for the message is allocated too early, prior to
4046 the excessive message length check. Due to way memory is allocated in
4047 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4048 to service a connection. This could lead to a Denial of Service through
4049 memory exhaustion. However, the excessive message length check still takes
4050 place, and this would cause the connection to immediately fail. Assuming
4051 that the application calls SSL_free() on the failed connection in a timely
4052 manner then the 21Mb of allocated memory will then be immediately freed
4053 again. Therefore, the excessive memory allocation will be transitory in
4054 nature. This then means that there is only a security impact if:
4056 1) The application does not call SSL_free() in a timely manner in the event
4057 that the connection fails
4059 2) The application is working in a constrained environment where there is
4060 very little free memory
4062 3) The attacker initiates multiple connection attempts such that there are
4063 multiple connections in a state where memory has been allocated for the
4064 connection; SSL_free() has not yet been called; and there is insufficient
4065 memory to service the multiple requests.
4067 Except in the instance of (1) above any Denial Of Service is likely to be
4068 transitory because as soon as the connection fails the memory is
4069 subsequently freed again in the SSL_free() call. However there is an
4070 increased risk during this period of application crashes due to the lack of
4071 memory - which would then mean a more serious Denial of Service.
4073 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4074 (CVE-2016-6307 and CVE-2016-6308)
4078 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4079 had to be removed. Primary reason is that vendor assembler can't
4080 assemble our modules with -KPIC flag. As result it, assembly
4081 support, was not even available as option. But its lack means
4082 lack of side-channel resistant code, which is incompatible with
4083 security by todays standards. Fortunately gcc is readily available
4084 prepackaged option, which we firmly point at...
4088 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4090 * Windows command-line tool supports UTF-8 opt-in option for arguments
4091 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4092 (to any value) allows Windows user to access PKCS#12 file generated
4093 with Windows CryptoAPI and protected with non-ASCII password, as well
4094 as files generated under UTF-8 locale on Linux also protected with
4099 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4100 have been disabled by default and removed from DEFAULT, just like RC4.
4101 See the RC4 item below to re-enable both.
4105 * The method for finding the storage location for the Windows RAND seed file
4106 has changed. First we check %RANDFILE%. If that is not set then we check
4107 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4108 all else fails we fall back to C:\.
4112 * The EVP_EncryptUpdate() function has had its return type changed from void
4113 to int. A return of 0 indicates and error while a return of 1 indicates
4118 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4119 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4120 off the constant time implementation for RSA, DSA and DH have been made
4121 no-ops and deprecated.
4125 * Windows RAND implementation was simplified to only get entropy by
4126 calling CryptGenRandom(). Various other RAND-related tickets
4129 *Joseph Wylie Yandle, Rich Salz*
4131 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4132 and `OPENSSL_LH_`, respectively. The old names are available
4133 with API compatibility. They new names are now completely documented.
4137 * Unify TYPE_up_ref(obj) methods signature.
4138 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4139 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4140 int (instead of void) like all others TYPE_up_ref() methods.
4141 So now these methods also check the return value of CRYPTO_atomic_add(),
4142 and the validity of object reference counter.
4144 *fdasilvayy@gmail.com*
4146 * With Windows Visual Studio builds, the .pdb files are installed
4147 alongside the installed libraries and executables. For a static
4148 library installation, ossl_static.pdb is the associate compiler
4149 generated .pdb file to be used when linking programs.
4153 * Remove openssl.spec. Packaging files belong with the packagers.
4157 * Automatic Darwin/OSX configuration has had a refresh, it will now
4158 recognise x86_64 architectures automatically. You can still decide
4159 to build for a different bitness with the environment variable
4160 KERNEL_BITS (can be 32 or 64), for example:
4162 KERNEL_BITS=32 ./config
4166 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4167 256 bit AES and HMAC with SHA256.
4171 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4175 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4179 * To enable users to have their own config files and build file templates,
4180 Configure looks in the directory indicated by the environment variable
4181 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4182 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4183 name and is used as is.
4187 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4188 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4189 X509_CERT_FILE_CTX was removed.
4193 * "shared" builds are now the default. To create only static libraries use
4194 the "no-shared" Configure option.
4198 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4199 All of these option have not worked for some while and are fundamental
4204 * Make various cleanup routines no-ops and mark them as deprecated. Most
4205 global cleanup functions are no longer required because they are handled
4206 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4207 Explicitly de-initing can cause problems (e.g. where a library that uses
4208 OpenSSL de-inits, but an application is still using it). The affected
4209 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4210 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4211 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4212 COMP_zlib_cleanup().
4216 * --strict-warnings no longer enables runtime debugging options
4217 such as REF_DEBUG. Instead, debug options are automatically
4218 enabled with '--debug' builds.
4220 *Andy Polyakov, Emilia Käsper*
4222 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4223 have been moved out of the public header files. New functions for managing
4224 these have been added.
4228 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4229 objects have been moved out of the public header files. New
4230 functions for managing these have been added.
4234 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4235 have been moved out of the public header files. New functions for managing
4236 these have been added.
4240 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4241 moved out of the public header files. New functions for managing these
4246 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4250 * Removed the mk1mf build scripts.
4254 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4255 it is always safe to #include a header now.
4259 * Removed the aged BC-32 config and all its supporting scripts
4263 * Removed support for Ultrix, Netware, and OS/2.
4267 * Add support for HKDF.
4269 *Alessandro Ghedini*
4271 * Add support for blake2b and blake2s
4275 * Added support for "pipelining". Ciphers that have the
4276 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4277 encryptions/decryptions simultaneously. There are currently no built-in
4278 ciphers with this property but the expectation is that engines will be able
4279 to offer it to significantly improve throughput. Support has been extended
4280 into libssl so that multiple records for a single connection can be
4281 processed in one go (for >=TLS 1.1).
4285 * Added the AFALG engine. This is an async capable engine which is able to
4286 offload work to the Linux kernel. In this initial version it only supports
4287 AES128-CBC. The kernel must be version 4.1.0 or greater.
4291 * OpenSSL now uses a new threading API. It is no longer necessary to
4292 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4293 are two supported threading models: pthreads and windows threads. It is
4294 also possible to configure OpenSSL at compile time for "no-threads". The
4295 old threading API should no longer be used. The functions have been
4296 replaced with "no-op" compatibility macros.
4298 *Alessandro Ghedini, Matt Caswell*
4300 * Modify behavior of ALPN to invoke callback after SNI/servername
4301 callback, such that updates to the SSL_CTX affect ALPN.
4305 * Add SSL_CIPHER queries for authentication and key-exchange.
4309 * Changes to the DEFAULT cipherlist:
4310 - Prefer (EC)DHE handshakes over plain RSA.
4311 - Prefer AEAD ciphers over legacy ciphers.
4312 - Prefer ECDSA over RSA when both certificates are available.
4313 - Prefer TLSv1.2 ciphers/PRF.
4314 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4319 * Change the ECC default curve list to be this, in order: x25519,
4320 secp256r1, secp521r1, secp384r1.
4324 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4325 disabled by default. They can be re-enabled using the
4326 enable-weak-ssl-ciphers option to Configure.
4330 * If the server has ALPN configured, but supports no protocols that the
4331 client advertises, send a fatal "no_application_protocol" alert.
4332 This behaviour is SHALL in RFC 7301, though it isn't universally
4333 implemented by other servers.
4337 * Add X25519 support.
4338 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4339 for public and private key encoding using the format documented in
4340 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4341 key generation and key derivation.
4343 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4348 * Deprecate SRP_VBASE_get_by_user.
4349 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4350 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4351 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4352 seed, even if the seed is configured.
4354 Users should use SRP_VBASE_get1_by_user instead. Note that in
4355 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4356 also that even though configuring the SRP seed attempts to hide
4357 invalid usernames by continuing the handshake with fake
4358 credentials, this behaviour is not constant time and no strong
4359 guarantees are made that the handshake is indistinguishable from
4360 that of a valid user.
4364 * Configuration change; it's now possible to build dynamic engines
4365 without having to build shared libraries and vice versa. This
4366 only applies to the engines in `engines/`, those in `crypto/engine/`
4367 will always be built into libcrypto (i.e. "static").
4369 Building dynamic engines is enabled by default; to disable, use
4370 the configuration option "disable-dynamic-engine".
4372 The only requirements for building dynamic engines are the
4373 presence of the DSO module and building with position independent
4374 code, so they will also automatically be disabled if configuring
4375 with "disable-dso" or "disable-pic".
4377 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4378 are also taken away from openssl/opensslconf.h, as they are
4383 * Configuration change; if there is a known flag to compile
4384 position independent code, it will always be applied on the
4385 libcrypto and libssl object files, and never on the application
4386 object files. This means other libraries that use routines from
4387 libcrypto / libssl can be made into shared libraries regardless
4388 of how OpenSSL was configured.
4390 If this isn't desirable, the configuration options "disable-pic"
4391 or "no-pic" can be used to disable the use of PIC. This will
4392 also disable building shared libraries and dynamic engines.
4396 * Removed JPAKE code. It was experimental and has no wide use.
4400 * The INSTALL_PREFIX Makefile variable has been renamed to
4401 DESTDIR. That makes for less confusion on what this variable
4402 is for. Also, the configuration option --install_prefix is
4407 * Heartbeat for TLS has been removed and is disabled by default
4408 for DTLS; configure with enable-heartbeats. Code that uses the
4409 old #define's might need to be updated.
4411 *Emilia Käsper, Rich Salz*
4413 * Rename REF_CHECK to REF_DEBUG.
4417 * New "unified" build system
4419 The "unified" build system is aimed to be a common system for all
4420 platforms we support. With it comes new support for VMS.
4422 This system builds supports building in a different directory tree
4423 than the source tree. It produces one Makefile (for unix family
4424 or lookalikes), or one descrip.mms (for VMS).
4426 The source of information to make the Makefile / descrip.mms is
4427 small files called 'build.info', holding the necessary
4428 information for each directory with source to compile, and a
4429 template in Configurations, like unix-Makefile.tmpl or
4432 With this change, the library names were also renamed on Windows
4433 and on VMS. They now have names that are closer to the standard
4434 on Unix, and include the major version number, and in certain
4435 cases, the architecture they are built for. See "Notes on shared
4436 libraries" in INSTALL.
4438 We rely heavily on the perl module Text::Template.
4442 * Added support for auto-initialisation and de-initialisation of the library.
4443 OpenSSL no longer requires explicit init or deinit routines to be called,
4444 except in certain circumstances. See the OPENSSL_init_crypto() and
4445 OPENSSL_init_ssl() man pages for further information.
4449 * The arguments to the DTLSv1_listen function have changed. Specifically the
4450 "peer" argument is now expected to be a BIO_ADDR object.
4452 * Rewrite of BIO networking library. The BIO library lacked consistent
4453 support of IPv6, and adding it required some more extensive
4454 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4455 which hold all types of addresses and chains of address information.
4456 It also introduces a new API, with functions like BIO_socket,
4457 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4458 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4459 have been adapted accordingly.
4463 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4468 * CRIME protection: disable compression by default, even if OpenSSL is
4469 compiled with zlib enabled. Applications can still enable compression
4470 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4471 using the SSL_CONF library to configure compression.
4475 * The signature of the session callback configured with
4476 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4477 was explicitly marked as `const unsigned char*` instead of
4482 * Always DPURIFY. Remove the use of uninitialized memory in the
4483 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4487 * Removed many obsolete configuration items, including
4488 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4489 MD2_CHAR, MD2_INT, MD2_LONG
4491 IDEA_SHORT, IDEA_LONG
4492 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4494 *Rich Salz, with advice from Andy Polyakov*
4496 * Many BN internals have been moved to an internal header file.
4498 *Rich Salz with help from Andy Polyakov*
4500 * Configuration and writing out the results from it has changed.
4501 Files such as Makefile include/openssl/opensslconf.h and are now
4502 produced through general templates, such as Makefile.in and
4503 crypto/opensslconf.h.in and some help from the perl module
4506 Also, the center of configuration information is no longer
4507 Makefile. Instead, Configure produces a perl module in
4508 configdata.pm which holds most of the config data (in the hash
4509 table %config), the target data that comes from the target
4510 configuration in one of the `Configurations/*.conf` files (in
4515 * To clarify their intended purposes, the Configure options
4516 --prefix and --openssldir change their semantics, and become more
4517 straightforward and less interdependent.
4519 --prefix shall be used exclusively to give the location INSTALLTOP
4520 where programs, scripts, libraries, include files and manuals are
4521 going to be installed. The default is now /usr/local.
4523 --openssldir shall be used exclusively to give the default
4524 location OPENSSLDIR where certificates, private keys, CRLs are
4525 managed. This is also where the default openssl.cnf gets
4527 If the directory given with this option is a relative path, the
4528 values of both the --prefix value and the --openssldir value will
4529 be combined to become OPENSSLDIR.
4530 The default for --openssldir is INSTALLTOP/ssl.
4532 Anyone who uses --openssldir to specify where OpenSSL is to be
4533 installed MUST change to use --prefix instead.
4537 * The GOST engine was out of date and therefore it has been removed. An up
4538 to date GOST engine is now being maintained in an external repository.
4539 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4540 support for GOST ciphersuites (these are only activated if a GOST engine
4545 * EGD is no longer supported by default; use enable-egd when
4548 *Ben Kaduk and Rich Salz*
4550 * The distribution now has Makefile.in files, which are used to
4551 create Makefile's when Configure is run. *Configure must be run
4552 before trying to build now.*
4556 * The return value for SSL_CIPHER_description() for error conditions
4561 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4563 Obtaining and performing DNSSEC validation of TLSA records is
4564 the application's responsibility. The application provides
4565 the TLSA records of its choice to OpenSSL, and these are then
4566 used to authenticate the peer.
4568 The TLSA records need not even come from DNS. They can, for
4569 example, be used to implement local end-entity certificate or
4570 trust-anchor "pinning", where the "pin" data takes the form
4571 of TLSA records, which can augment or replace verification
4572 based on the usual WebPKI public certification authorities.
4576 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4577 continues to support deprecated interfaces in default builds.
4578 However, applications are strongly advised to compile their
4579 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4580 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4581 or the 1.1.0 releases.
4583 In environments in which all applications have been ported to
4584 not use any deprecated interfaces OpenSSL's Configure script
4585 should be used with the --api=1.1.0 option to entirely remove
4586 support for the deprecated features from the library and
4587 unconditionally disable them in the installed headers.
4588 Essentially the same effect can be achieved with the "no-deprecated"
4589 argument to Configure, except that this will always restrict
4590 the build to just the latest API, rather than a fixed API
4593 As applications are ported to future revisions of the API,
4594 they should update their compile-time OPENSSL_API_COMPAT define
4595 accordingly, but in most cases should be able to continue to
4596 compile with later releases.
4598 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4599 0x10000000L and 0x00908000L, respectively. However those
4600 versions did not support the OPENSSL_API_COMPAT feature, and
4601 so applications are not typically tested for explicit support
4602 of just the undeprecated features of either release.
4606 * Add support for setting the minimum and maximum supported protocol.
4607 It can bet set via the SSL_set_min_proto_version() and
4608 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4609 MaxProtocol. It's recommended to use the new APIs to disable
4610 protocols instead of disabling individual protocols using
4611 SSL_set_options() or SSL_CONF's Protocol. This change also
4612 removes support for disabling TLS 1.2 in the OpenSSL TLS
4613 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4617 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4621 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4622 and integrates ECDSA and ECDH functionality into EC. Implementations can
4623 now redirect key generation and no longer need to convert to or from
4626 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4627 include the ec.h header file instead.
4631 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4632 ciphers who are no longer supported and drops support the ephemeral RSA key
4633 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4637 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4638 opaque. For HMAC_CTX, the following constructors and destructors
4641 HMAC_CTX *HMAC_CTX_new(void);
4642 void HMAC_CTX_free(HMAC_CTX *ctx);
4644 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4645 destroy such methods has been added. See EVP_MD_meth_new(3) and
4646 EVP_CIPHER_meth_new(3) for documentation.
4649 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4650 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4651 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4652 an already created structure.
4653 2) For consistency with the majority of our object creators and
4654 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4655 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4656 for deprecated builds.
4660 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4661 cryptographic operations to be performed asynchronously as long as an
4662 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4663 further details. Libssl has also had this capability integrated with the
4664 introduction of the new mode SSL_MODE_ASYNC and associated error
4665 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4666 pages. This work was developed in partnership with Intel Corp.
4670 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4671 always enabled now. If you want to disable the support you should
4672 exclude it using the list of supported ciphers. This also means that the
4673 "-no_ecdhe" option has been removed from s_server.
4677 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4678 SSL_{CTX_}set1_curves() which can set a list.
4682 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4683 curve you want to support using SSL_{CTX_}set1_curves().
4687 * State machine rewrite. The state machine code has been significantly
4688 refactored in order to remove much duplication of code and solve issues
4689 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4690 further details). This change does have some associated API changes.
4691 Notably the SSL_state() function has been removed and replaced by
4692 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4693 SSL_set_state() has been removed altogether. The previous handshake states
4694 defined in ssl.h and ssl3.h have also been removed.
4698 * All instances of the string "ssleay" in the public API were replaced
4699 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4700 Some error codes related to internal RSA_eay API's were renamed.
4704 * The demo files in crypto/threads were moved to demo/threads.
4708 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4711 *Matt Caswell, Rich Salz*
4713 * New ASN.1 embed macro.
4715 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4716 structure is not allocated: it is part of the parent. That is instead of
4724 This reduces memory fragmentation and make it impossible to accidentally
4725 set a mandatory field to NULL.
4727 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4728 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4729 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4734 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4738 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4739 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4740 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4741 DES and RC4 ciphersuites.
4745 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4746 This changes the decoding behaviour for some invalid messages,
4747 though the change is mostly in the more lenient direction, and
4748 legacy behaviour is preserved as much as possible.
4752 * Fix no-stdio build.
4753 *David Woodhouse <David.Woodhouse@intel.com> and also*
4754 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4756 * New testing framework
4757 The testing framework has been largely rewritten and is now using
4758 perl and the perl modules Test::Harness and an extended variant of
4759 Test::More called OpenSSL::Test to do its work. All test scripts in
4760 test/ have been rewritten into test recipes, and all direct calls to
4761 executables in test/Makefile have become individual recipes using the
4762 simplified testing OpenSSL::Test::Simple.
4764 For documentation on our testing modules, do:
4766 perldoc test/testlib/OpenSSL/Test/Simple.pm
4767 perldoc test/testlib/OpenSSL/Test.pm
4771 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4772 are used; the latter aborts on memory leaks (usually checked on exit).
4773 Some undocumented "set malloc, etc., hooks" functions were removed
4774 and others were changed. All are now documented.
4778 * In DSA_generate_parameters_ex, if the provided seed is too short,
4781 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4783 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4784 from RFC4279, RFC4785, RFC5487, RFC5489.
4786 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4787 original RSA_PSK patch.
4791 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4792 era flag was never set throughout the codebase (only read). Also removed
4793 SSL3_FLAGS_POP_BUFFER which was only used if
4794 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4798 * Changed the default name options in the "ca", "crl", "req" and "x509"
4799 to be "oneline" instead of "compat".
4803 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4804 not aware of clients that still exhibit this bug, and the workaround
4805 hasn't been working properly for a while.
4809 * The return type of BIO_number_read() and BIO_number_written() as well as
4810 the corresponding num_read and num_write members in the BIO structure has
4811 changed from unsigned long to uint64_t. On platforms where an unsigned
4812 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4817 * Given the pervasive nature of TLS extensions it is inadvisable to run
4818 OpenSSL without support for them. It also means that maintaining
4819 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4820 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4824 * Removed support for the two export grade static DH ciphersuites
4825 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4826 were newly added (along with a number of other static DH ciphersuites) to
4827 1.0.2. However the two export ones have *never* worked since they were
4828 introduced. It seems strange in any case to be adding new export
4829 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4833 * Version negotiation has been rewritten. In particular SSLv23_method(),
4834 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4835 and turned into macros which simply call the new preferred function names
4836 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4837 should use the new names instead. Also as part of this change the ssl23.h
4838 header file has been removed.
4842 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4843 code and the associated standard is no longer considered fit-for-purpose.
4847 * RT2547 was closed. When generating a private key, try to make the
4848 output file readable only by the owner. This behavior change might
4849 be noticeable when interacting with other software.
4851 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4856 * Added HTTP GET support to the ocsp command.
4860 * Changed default digest for the dgst and enc commands from MD5 to
4865 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4869 * Added support for TLS extended master secret from
4870 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4871 initial patch which was a great help during development.
4875 * All libssl internal structures have been removed from the public header
4876 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4877 now redundant). Users should not attempt to access internal structures
4878 directly. Instead they should use the provided API functions.
4882 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4883 Access to deprecated functions can be re-enabled by running config with
4884 "enable-deprecated". In addition applications wishing to use deprecated
4885 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4886 will, by default, disable some transitive includes that previously existed
4887 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4891 * Added support for OCB mode. OpenSSL has been granted a patent license
4892 compatible with the OpenSSL license for use of OCB. Details are available
4893 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4894 for OCB can be removed by calling config with no-ocb.
4898 * SSLv2 support has been removed. It still supports receiving an SSLv2
4899 compatible client hello.
4903 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4904 done while fixing the error code for the key-too-small case.
4906 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4908 * CA.sh has been removed; use CA.pl instead.
4912 * Removed old DES API.
4916 * Remove various unsupported platforms:
4922 Sinix/ReliantUNIX RM400
4927 16-bit platforms such as WIN16
4931 * Clean up OPENSSL_NO_xxx #define's
4932 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4933 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4934 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4935 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4936 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4937 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4938 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4939 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4940 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4941 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4945 * Cleaned up dead code
4946 Remove all but one '#ifdef undef' which is to be looked at.
4950 * Clean up calling of xxx_free routines.
4951 Just like free(), fix most of the xxx_free routines to accept
4952 NULL. Remove the non-null checks from callers. Save much code.
4956 * Add secure heap for storage of private keys (when possible).
4957 Add BIO_s_secmem(), CBIGNUM, etc.
4958 Contributed by Akamai Technologies under our Corporate CLA.
4962 * Experimental support for a new, fast, unbiased prime candidate generator,
4963 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4965 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4967 * New output format NSS in the sess_id command line tool. This allows
4968 exporting the session id and the master key in NSS keylog format.
4970 *Martin Kaiser <martin@kaiser.cx>*
4972 * Harmonize version and its documentation. -f flag is used to display
4975 *mancha <mancha1@zoho.com>*
4977 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4978 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4980 *mancha <mancha1@zoho.com>*
4982 * Fix some double frees. These are not thought to be exploitable.
4984 *mancha <mancha1@zoho.com>*
4986 * A missing bounds check in the handling of the TLS heartbeat extension
4987 can be used to reveal up to 64k of memory to a connected client or
4990 Thanks for Neel Mehta of Google Security for discovering this bug and to
4991 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4992 preparing the fix ([CVE-2014-0160])
4994 *Adam Langley, Bodo Moeller*
4996 * Fix for the attack described in the paper "Recovering OpenSSL
4997 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4998 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4999 <http://eprint.iacr.org/2014/140>
5001 Thanks to Yuval Yarom and Naomi Benger for discovering this
5002 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5004 *Yuval Yarom and Naomi Benger*
5006 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5007 this fixes a limitation in previous versions of OpenSSL.
5011 * Experimental encrypt-then-mac support.
5013 Experimental support for encrypt then mac from
5014 draft-gutmann-tls-encrypt-then-mac-02.txt
5016 To enable it set the appropriate extension number (0x42 for the test
5017 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5019 For non-compliant peers (i.e. just about everything) this should have no
5022 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5026 * Add EVP support for key wrapping algorithms, to avoid problems with
5027 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5028 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5029 algorithms and include tests cases.
5033 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5038 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5039 MGF1 digest and OAEP label.
5043 * Make openssl verify return errors.
5045 *Chris Palmer <palmer@google.com> and Ben Laurie*
5047 * New function ASN1_TIME_diff to calculate the difference between two
5048 ASN1_TIME structures or one structure and the current time.
5052 * Update fips_test_suite to support multiple command line options. New
5053 test to induce all self test errors in sequence and check expected
5058 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5059 sign or verify all in one operation.
5063 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5064 test programs and fips_test_suite. Includes functionality to parse
5065 the minimal script output of fipsalgest.pl directly.
5069 * Add authorisation parameter to FIPS_module_mode_set().
5073 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5077 * Use separate DRBG fields for internal and external flags. New function
5078 FIPS_drbg_health_check() to perform on demand health checking. Add
5079 generation tests to fips_test_suite with reduced health check interval to
5080 demonstrate periodic health checking. Add "nodh" option to
5081 fips_test_suite to skip very slow DH test.
5085 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5090 * More extensive health check for DRBG checking many more failure modes.
5091 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5092 combination: call this in fips_test_suite.
5096 * Add support for canonical generation of DSA parameter 'g'. See
5099 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5100 POST to handle HMAC cases.
5104 * Add functions FIPS_module_version() and FIPS_module_version_text()
5105 to return numerical and string versions of the FIPS module number.
5109 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5110 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5111 outside the validated module in the FIPS capable OpenSSL.
5115 * Minor change to DRBG entropy callback semantics. In some cases
5116 there is no multiple of the block length between min_len and
5117 max_len. Allow the callback to return more than max_len bytes
5118 of entropy but discard any extra: it is the callback's responsibility
5119 to ensure that the extra data discarded does not impact the
5120 requested amount of entropy.
5124 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5125 information in FIPS186-3, SP800-57 and SP800-131A.
5129 * CCM support via EVP. Interface is very similar to GCM case except we
5130 must supply all data in one chunk (i.e. no update, final) and the
5131 message length must be supplied if AAD is used. Add algorithm test
5136 * Initial version of POST overhaul. Add POST callback to allow the status
5137 of POST to be monitored and/or failures induced. Modify fips_test_suite
5138 to use callback. Always run all selftests even if one fails.
5142 * XTS support including algorithm test driver in the fips_gcmtest program.
5143 Note: this does increase the maximum key length from 32 to 64 bytes but
5144 there should be no binary compatibility issues as existing applications
5145 will never use XTS mode.
5149 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5150 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5151 performs algorithm blocking for unapproved PRNG types. Also do not
5152 set PRNG type in FIPS_mode_set(): leave this to the application.
5153 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5154 the standard OpenSSL PRNG: set additional data to a date time vector.
5158 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5159 This shouldn't present any incompatibility problems because applications
5160 shouldn't be using these directly and any that are will need to rethink
5161 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5165 * Extensive self tests and health checking required by SP800-90 DRBG.
5166 Remove strength parameter from FIPS_drbg_instantiate and always
5167 instantiate at maximum supported strength.
5171 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5175 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5179 * New function DH_compute_key_padded() to compute a DH key and pad with
5180 leading zeroes if needed: this complies with SP800-56A et al.
5184 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5185 anything, incomplete, subject to change and largely untested at present.
5189 * Modify fipscanisteronly build option to only build the necessary object
5190 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5194 * Add experimental option FIPSSYMS to give all symbols in
5195 fipscanister.o and FIPS or fips prefix. This will avoid
5196 conflicts with future versions of OpenSSL. Add perl script
5197 util/fipsas.pl to preprocess assembly language source files
5198 and rename any affected symbols.
5202 * Add selftest checks and algorithm block of non-fips algorithms in
5203 FIPS mode. Remove DES2 from selftests.
5207 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5208 return internal method without any ENGINE dependencies. Add new
5209 tiny fips sign and verify functions.
5213 * New build option no-ec2m to disable characteristic 2 code.
5217 * New build option "fipscanisteronly". This only builds fipscanister.o
5218 and (currently) associated fips utilities. Uses the file Makefile.fips
5219 instead of Makefile.org as the prototype.
5223 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5224 Update fips_gcmtest to use IV generator.
5228 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5229 setting output buffer to NULL. The `*Final` function must be
5230 called although it will not retrieve any additional data. The tag
5231 can be set or retrieved with a ctrl. The IV length is by default 12
5232 bytes (96 bits) but can be set to an alternative value. If the IV
5233 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5238 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5239 underlying do_cipher function handles all cipher semantics itself
5240 including padding and finalisation. This is useful if (for example)
5241 an ENGINE cipher handles block padding itself. The behaviour of
5242 do_cipher is subtly changed if this flag is set: the return value
5243 is the number of characters written to the output buffer (zero is
5244 no longer an error code) or a negative error code. Also if the
5245 input buffer is NULL and length 0 finalisation should be performed.
5249 * If a candidate issuer certificate is already part of the constructed
5250 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5254 * Improve forward-security support: add functions
5256 void SSL_CTX_set_not_resumable_session_callback(
5257 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5258 void SSL_set_not_resumable_session_callback(
5259 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5261 for use by SSL/TLS servers; the callback function will be called whenever a
5262 new session is created, and gets to decide whether the session may be
5263 cached to make it resumable (return 0) or not (return 1). (As by the
5264 SSL/TLS protocol specifications, the session_id sent by the server will be
5265 empty to indicate that the session is not resumable; also, the server will
5266 not generate RFC 4507 (RFC 5077) session tickets.)
5268 A simple reasonable callback implementation is to return is_forward_secure.
5269 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5270 by the SSL/TLS server library, indicating whether it can provide forward
5273 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5275 * New -verify_name option in command line utilities to set verification
5280 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5281 Add CMAC pkey methods.
5285 * Experimental renegotiation in s_server -www mode. If the client
5286 browses /reneg connection is renegotiated. If /renegcert it is
5287 renegotiated requesting a certificate.
5291 * Add an "external" session cache for debugging purposes to s_server. This
5292 should help trace issues which normally are only apparent in deployed
5293 multi-process servers.
5297 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5298 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5299 BIO_set_cipher() and some obscure PEM functions were changed so they
5300 can now return an error. The RAND changes required a change to the
5301 RAND_METHOD structure.
5305 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5306 a gcc attribute to warn if the result of a function is ignored. This
5307 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5308 whose return value is often ignored.
5312 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5313 These allow SCTs (signed certificate timestamps) to be requested and
5314 validated when establishing a connection.
5316 *Rob Percival <robpercival@google.com>*
5321 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5323 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5324 used even when parsing explicit parameters, when loading a encoded key
5325 or calling `EC_GROUP_new_from_ecpkparameters()`/
5326 `EC_GROUP_new_from_ecparameters()`.
5327 This prevents bypass of security hardening and performance gains,
5328 especially for curves with specialized EC_METHODs.
5329 By default, if a key encoded with explicit parameters is loaded and later
5330 encoded, the output is still encoded with explicit parameters, even if
5331 internally a "named" EC_GROUP is used for computation.
5335 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5336 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5337 NULL. After this change, only the cofactor parameter can be NULL. It also
5338 does some minimal sanity checks on the passed order.
5343 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5344 An attack is simple, if the first CMS_recipientInfo is valid but the
5345 second CMS_recipientInfo is chosen ciphertext. If the second
5346 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5347 encryption key will be replaced by garbage, and the message cannot be
5348 decoded, but if the RSA decryption fails, the correct encryption key is
5349 used and the recipient will not notice the attack.
5350 As a work around for this potential attack the length of the decrypted
5351 key must be equal to the cipher default key length, in case the
5352 certificate is not given and all recipientInfo are tried out.
5353 The old behaviour can be re-enabled in the CMS code by setting the
5354 CMS_DEBUG_DECRYPT flag.
5359 * Document issue with installation paths in diverse Windows builds
5361 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5362 binaries and run-time config file.
5367 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5369 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5370 This changes the size when using the `genpkey` command when no size is given.
5371 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5372 generation commands to use 2048 bits by default.
5376 * Add FIPS support for Android Arm 64-bit
5378 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5379 Module in Version 2.0.10. For some reason, the corresponding target
5380 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5381 built with FIPS support on Android Arm 64-bit. This omission has been
5384 *Matthias St. Pierre*
5386 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5388 * 0-byte record padding oracle
5390 If an application encounters a fatal protocol error and then calls
5391 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5392 then OpenSSL can respond differently to the calling application if a 0 byte
5393 record is received with invalid padding compared to if a 0 byte record is
5394 received with an invalid MAC. If the application then behaves differently
5395 based on that in a way that is detectable to the remote peer, then this
5396 amounts to a padding oracle that could be used to decrypt data.
5398 In order for this to be exploitable "non-stitched" ciphersuites must be in
5399 use. Stitched ciphersuites are optimised implementations of certain
5400 commonly used ciphersuites. Also the application must call SSL_shutdown()
5401 twice even if a protocol error has occurred (applications should not do
5402 this but some do anyway).
5404 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5405 Aviram, with additional investigation by Steven Collison and Andrew
5406 Hourselt. It was reported to OpenSSL on 10th December 2018.
5411 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5415 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5417 * Microarchitecture timing vulnerability in ECC scalar multiplication
5419 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5420 shown to be vulnerable to a microarchitecture timing side channel attack.
5421 An attacker with sufficient access to mount local timing attacks during
5422 ECDSA signature generation could recover the private key.
5424 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5425 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5431 * Timing vulnerability in DSA signature generation
5433 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5434 timing side channel attack. An attacker could use variations in the signing
5435 algorithm to recover the private key.
5437 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5442 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5443 Module, accidentally introduced while backporting security fixes from the
5444 development branch and hindering the use of ECC in FIPS mode.
5448 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5450 * Client DoS due to large DH parameter
5452 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5453 malicious server can send a very large prime value to the client. This will
5454 cause the client to spend an unreasonably long period of time generating a
5455 key for this prime resulting in a hang until the client has finished. This
5456 could be exploited in a Denial Of Service attack.
5458 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5463 * Cache timing vulnerability in RSA Key Generation
5465 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5466 a cache timing side channel attack. An attacker with sufficient access to
5467 mount cache timing attacks during the RSA key generation process could
5468 recover the private key.
5470 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5471 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5476 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5477 parameter is no longer accepted, as it leads to a corrupt table. NULL
5478 pem_str is reserved for alias entries only.
5482 * Revert blinding in ECDSA sign and instead make problematic addition
5483 length-invariant. Switch even to fixed-length Montgomery multiplication.
5487 * Change generating and checking of primes so that the error rate of not
5488 being prime depends on the intended use based on the size of the input.
5489 For larger primes this will result in more rounds of Miller-Rabin.
5490 The maximal error rate for primes with more than 1080 bits is lowered
5493 *Kurt Roeckx, Annie Yousar*
5495 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5499 * Add blinding to ECDSA and DSA signatures to protect against side channel
5500 attacks discovered by Keegan Ryan (NCC Group).
5504 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5505 now allow empty (zero character) pass phrases.
5509 * Certificate time validation (X509_cmp_time) enforces stricter
5510 compliance with RFC 5280. Fractional seconds and timezone offsets
5511 are no longer allowed.
5515 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5517 * Constructed ASN.1 types with a recursive definition could exceed the stack
5519 Constructed ASN.1 types with a recursive definition (such as can be found
5520 in PKCS7) could eventually exceed the stack given malicious input with
5521 excessive recursion. This could result in a Denial Of Service attack. There
5522 are no such structures used within SSL/TLS that come from untrusted sources
5523 so this is considered safe.
5525 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5531 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5533 * Read/write after SSL object in error state
5535 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5536 mechanism. The intent was that if a fatal error occurred during a handshake
5537 then OpenSSL would move into the error state and would immediately fail if
5538 you attempted to continue the handshake. This works as designed for the
5539 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5540 SSL_connect()), however due to a bug it does not work correctly if
5541 SSL_read() or SSL_write() is called directly. In that scenario, if the
5542 handshake fails then a fatal error will be returned in the initial function
5543 call. If SSL_read()/SSL_write() is subsequently called by the application
5544 for the same SSL object then it will succeed and the data is passed without
5545 being decrypted/encrypted directly from the SSL/TLS record layer.
5547 In order to exploit this issue an application bug would have to be present
5548 that resulted in a call to SSL_read()/SSL_write() being issued after having
5549 already received a fatal error.
5551 This issue was reported to OpenSSL by David Benjamin (Google).
5556 * rsaz_1024_mul_avx2 overflow bug on x86_64
5558 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5559 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5560 Analysis suggests that attacks against RSA and DSA as a result of this
5561 defect would be very difficult to perform and are not believed likely.
5562 Attacks against DH1024 are considered just feasible, because most of the
5563 work necessary to deduce information about a private key may be performed
5564 offline. The amount of resources required for such an attack would be
5565 significant. However, for an attack on TLS to be meaningful, the server
5566 would have to share the DH1024 private key among multiple clients, which is
5567 no longer an option since CVE-2016-0701.
5569 This only affects processors that support the AVX2 but not ADX extensions
5570 like Intel Haswell (4th generation).
5572 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5573 was originally found via the OSS-Fuzz project.
5578 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5580 * bn_sqrx8x_internal carry bug on x86_64
5582 There is a carry propagating bug in the x86_64 Montgomery squaring
5583 procedure. No EC algorithms are affected. Analysis suggests that attacks
5584 against RSA and DSA as a result of this defect would be very difficult to
5585 perform and are not believed likely. Attacks against DH are considered just
5586 feasible (although very difficult) because most of the work necessary to
5587 deduce information about a private key may be performed offline. The amount
5588 of resources required for such an attack would be very significant and
5589 likely only accessible to a limited number of attackers. An attacker would
5590 additionally need online access to an unpatched system using the target
5591 private key in a scenario with persistent DH parameters and a private
5592 key that is shared between multiple clients.
5594 This only affects processors that support the BMI1, BMI2 and ADX extensions
5595 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5597 This issue was reported to OpenSSL by the OSS-Fuzz project.
5602 * Malformed X.509 IPAddressFamily could cause OOB read
5604 If an X.509 certificate has a malformed IPAddressFamily extension,
5605 OpenSSL could do a one-byte buffer overread. The most likely result
5606 would be an erroneous display of the certificate in text format.
5608 This issue was reported to OpenSSL by the OSS-Fuzz project.
5612 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5614 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5615 platform rather than 'mingw'.
5619 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5621 * Truncated packet could crash via OOB read
5623 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5624 cipher is being used, then a truncated packet can cause that host to
5625 perform an out-of-bounds read, usually resulting in a crash.
5627 This issue was reported to OpenSSL by Robert Święcki of Google.
5632 * BN_mod_exp may produce incorrect results on x86_64
5634 There is a carry propagating bug in the x86_64 Montgomery squaring
5635 procedure. No EC algorithms are affected. Analysis suggests that attacks
5636 against RSA and DSA as a result of this defect would be very difficult to
5637 perform and are not believed likely. Attacks against DH are considered just
5638 feasible (although very difficult) because most of the work necessary to
5639 deduce information about a private key may be performed offline. The amount
5640 of resources required for such an attack would be very significant and
5641 likely only accessible to a limited number of attackers. An attacker would
5642 additionally need online access to an unpatched system using the target
5643 private key in a scenario with persistent DH parameters and a private
5644 key that is shared between multiple clients. For example this can occur by
5645 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5646 similar to CVE-2015-3193 but must be treated as a separate problem.
5648 This issue was reported to OpenSSL by the OSS-Fuzz project.
5653 * Montgomery multiplication may produce incorrect results
5655 There is a carry propagating bug in the Broadwell-specific Montgomery
5656 multiplication procedure that handles input lengths divisible by, but
5657 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5658 and DH private keys are impossible. This is because the subroutine in
5659 question is not used in operations with the private key itself and an input
5660 of the attacker's direct choice. Otherwise the bug can manifest itself as
5661 transient authentication and key negotiation failures or reproducible
5662 erroneous outcome of public-key operations with specially crafted input.
5663 Among EC algorithms only Brainpool P-512 curves are affected and one
5664 presumably can attack ECDH key negotiation. Impact was not analyzed in
5665 detail, because pre-requisites for attack are considered unlikely. Namely
5666 multiple clients have to choose the curve in question and the server has to
5667 share the private key among them, neither of which is default behaviour.
5668 Even then only clients that chose the curve will be affected.
5670 This issue was publicly reported as transient failures and was not
5671 initially recognized as a security issue. Thanks to Richard Morgan for
5672 providing reproducible case.
5677 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5678 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5679 prevent issues where no progress is being made and the peer continually
5680 sends unrecognised record types, using up resources processing them.
5684 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5686 * Missing CRL sanity check
5688 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5689 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5690 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5692 This issue only affects the OpenSSL 1.0.2i
5697 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5699 * OCSP Status Request extension unbounded memory growth
5701 A malicious client can send an excessively large OCSP Status Request
5702 extension. If that client continually requests renegotiation, sending a
5703 large OCSP Status Request extension each time, then there will be unbounded
5704 memory growth on the server. This will eventually lead to a Denial Of
5705 Service attack through memory exhaustion. Servers with a default
5706 configuration are vulnerable even if they do not support OCSP. Builds using
5707 the "no-ocsp" build time option are not affected.
5709 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5714 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5717 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5723 * OOB write in MDC2_Update()
5725 An overflow can occur in MDC2_Update() either if called directly or
5726 through the EVP_DigestUpdate() function using MDC2. If an attacker
5727 is able to supply very large amounts of input data after a previous
5728 call to EVP_EncryptUpdate() with a partial block then a length check
5729 can overflow resulting in a heap corruption.
5731 The amount of data needed is comparable to SIZE_MAX which is impractical
5734 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5739 * Malformed SHA512 ticket DoS
5741 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5742 DoS attack where a malformed ticket will result in an OOB read which will
5745 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5746 a custom server callback and ticket lookup mechanism.
5748 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5753 * OOB write in BN_bn2dec()
5755 The function BN_bn2dec() does not check the return value of BN_div_word().
5756 This can cause an OOB write if an application uses this function with an
5757 overly large BIGNUM. This could be a problem if an overly large certificate
5758 or CRL is printed out from an untrusted source. TLS is not affected because
5759 record limits will reject an oversized certificate before it is parsed.
5761 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5766 * OOB read in TS_OBJ_print_bio()
5768 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5769 the total length the OID text representation would use and not the amount
5770 of data written. This will result in OOB reads when large OIDs are
5773 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5778 * Pointer arithmetic undefined behaviour
5780 Avoid some undefined pointer arithmetic
5782 A common idiom in the codebase is to check limits in the following manner:
5785 Where "p" points to some malloc'd data of SIZE bytes and
5788 "len" here could be from some externally supplied data (e.g. from a TLS
5791 The rules of C pointer arithmetic are such that "p + len" is only well
5792 defined where len <= SIZE. Therefore the above idiom is actually
5793 undefined behaviour.
5795 For example this could cause problems if some malloc implementation
5796 provides an address for "p" such that "p + len" actually overflows for
5797 values of len that are too big and therefore p + len < limit.
5799 This issue was reported to OpenSSL by Guido Vranken
5804 * Constant time flag not preserved in DSA signing
5806 Operations in the DSA signing algorithm should run in constant time in
5807 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5808 implementation means that a non-constant time codepath is followed for
5809 certain operations. This has been demonstrated through a cache-timing
5810 attack to be sufficient for an attacker to recover the private DSA key.
5812 This issue was reported by César Pereida (Aalto University), Billy Brumley
5813 (Tampere University of Technology), and Yuval Yarom (The University of
5814 Adelaide and NICTA).
5819 * DTLS buffered message DoS
5821 In a DTLS connection where handshake messages are delivered out-of-order
5822 those messages that OpenSSL is not yet ready to process will be buffered
5823 for later use. Under certain circumstances, a flaw in the logic means that
5824 those messages do not get removed from the buffer even though the handshake
5825 has been completed. An attacker could force up to approx. 15 messages to
5826 remain in the buffer when they are no longer required. These messages will
5827 be cleared when the DTLS connection is closed. The default maximum size for
5828 a message is 100k. Therefore, the attacker could force an additional 1500k
5829 to be consumed per connection. By opening many simultaneous connections an
5830 attacker could cause a DoS attack through memory exhaustion.
5832 This issue was reported to OpenSSL by Quan Luo.
5837 * DTLS replay protection DoS
5839 A flaw in the DTLS replay attack protection mechanism means that records
5840 that arrive for future epochs update the replay protection "window" before
5841 the MAC for the record has been validated. This could be exploited by an
5842 attacker by sending a record for the next epoch (which does not have to
5843 decrypt or have a valid MAC), with a very large sequence number. This means
5844 that all subsequent legitimate packets are dropped causing a denial of
5845 service for a specific DTLS connection.
5847 This issue was reported to OpenSSL by the OCAP audit team.
5852 * Certificate message OOB reads
5854 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5855 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5856 theoretical DoS risk but this has not been observed in practice on common
5859 The messages affected are client certificate, client certificate request
5860 and server certificate. As a result the attack can only be performed
5861 against a client or a server which enables client authentication.
5863 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5868 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5870 * Prevent padding oracle in AES-NI CBC MAC check
5872 A MITM attacker can use a padding oracle attack to decrypt traffic
5873 when the connection uses an AES CBC cipher and the server support
5876 This issue was introduced as part of the fix for Lucky 13 padding
5877 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5878 constant time by making sure that always the same bytes are read and
5879 compared against either the MAC or padding bytes. But it no longer
5880 checked that there was enough data to have both the MAC and padding
5883 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5887 * Fix EVP_EncodeUpdate overflow
5889 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5890 Base64 encoding of binary data. If an attacker is able to supply very large
5891 amounts of input data then a length check can overflow resulting in a heap
5894 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5895 the `PEM_write_bio*` family of functions. These are mainly used within the
5896 OpenSSL command line applications, so any application which processes data
5897 from an untrusted source and outputs it as a PEM file should be considered
5898 vulnerable to this issue. User applications that call these APIs directly
5899 with large amounts of untrusted data may also be vulnerable.
5901 This issue was reported by Guido Vranken.
5906 * Fix EVP_EncryptUpdate overflow
5908 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5909 is able to supply very large amounts of input data after a previous call to
5910 EVP_EncryptUpdate() with a partial block then a length check can overflow
5911 resulting in a heap corruption. Following an analysis of all OpenSSL
5912 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5913 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5914 the first called function after an EVP_EncryptInit(), and therefore that
5915 specific call must be safe. The second form is where the length passed to
5916 EVP_EncryptUpdate() can be seen from the code to be some small value and
5917 therefore there is no possibility of an overflow. Since all instances are
5918 one of these two forms, it is believed that there can be no overflows in
5919 internal code due to this problem. It should be noted that
5920 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5921 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5922 of these calls have also been analysed too and it is believed there are no
5923 instances in internal usage where an overflow could occur.
5925 This issue was reported by Guido Vranken.
5930 * Prevent ASN.1 BIO excessive memory allocation
5932 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5933 a short invalid encoding can cause allocation of large amounts of memory
5934 potentially consuming excessive resources or exhausting memory.
5936 Any application parsing untrusted data through d2i BIO functions is
5937 affected. The memory based functions such as d2i_X509() are *not* affected.
5938 Since the memory based functions are used by the TLS library, TLS
5939 applications are not affected.
5941 This issue was reported by Brian Carpenter.
5948 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5949 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5950 in arbitrary stack data being returned in the buffer.
5952 This issue was reported by Guido Vranken.
5957 * Modify behavior of ALPN to invoke callback after SNI/servername
5958 callback, such that updates to the SSL_CTX affect ALPN.
5962 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5967 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5968 methods are enabled and ssl2 is disabled the methods return NULL.
5972 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5974 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5975 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5976 provide any "EXPORT" or "LOW" strength ciphers.
5980 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5981 is by default disabled at build-time. Builds that are not configured with
5982 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5983 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5984 will need to explicitly call either of:
5986 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5988 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5990 as appropriate. Even if either of those is used, or the application
5991 explicitly uses the version-specific SSLv2_method() or its client and
5992 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5993 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5994 ciphers, and SSLv2 56-bit DES are no longer available.
5999 * Fix a double-free in DSA code
6001 A double free bug was discovered when OpenSSL parses malformed DSA private
6002 keys and could lead to a DoS attack or memory corruption for applications
6003 that receive DSA private keys from untrusted sources. This scenario is
6006 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6012 * Disable SRP fake user seed to address a server memory leak.
6014 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6016 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6017 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6018 was changed to ignore the "fake user" SRP seed, even if the seed
6021 Users should use SRP_VBASE_get1_by_user instead. Note that in
6022 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6023 also that even though configuring the SRP seed attempts to hide
6024 invalid usernames by continuing the handshake with fake
6025 credentials, this behaviour is not constant time and no strong
6026 guarantees are made that the handshake is indistinguishable from
6027 that of a valid user.
6032 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6034 In the BN_hex2bn function the number of hex digits is calculated using an
6035 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6036 large values of `i` this can result in `bn_expand` not allocating any
6037 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6038 field as NULL leading to a subsequent NULL ptr deref. For very large values
6039 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6040 In this case memory is allocated to the internal BIGNUM data field, but it
6041 is insufficiently sized leading to heap corruption. A similar issue exists
6042 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6043 is ever called by user applications with very large untrusted hex/dec data.
6044 This is anticipated to be a rare occurrence.
6046 All OpenSSL internal usage of these functions use data that is not expected
6047 to be untrusted, e.g. config file data or application command line
6048 arguments. If user developed applications generate config file data based
6049 on untrusted data then it is possible that this could also lead to security
6050 consequences. This is also anticipated to be rare.
6052 This issue was reported to OpenSSL by Guido Vranken.
6057 * Fix memory issues in `BIO_*printf` functions
6059 The internal `fmtstr` function used in processing a "%s" format string in
6060 the `BIO_*printf` functions could overflow while calculating the length of a
6061 string and cause an OOB read when printing very long strings.
6063 Additionally the internal `doapr_outch` function can attempt to write to an
6064 OOB memory location (at an offset from the NULL pointer) in the event of a
6065 memory allocation failure. In 1.0.2 and below this could be caused where
6066 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6067 could be in processing a very long "%s" format string. Memory leaks can
6070 The first issue may mask the second issue dependent on compiler behaviour.
6071 These problems could enable attacks where large amounts of untrusted data
6072 is passed to the `BIO_*printf` functions. If applications use these functions
6073 in this way then they could be vulnerable. OpenSSL itself uses these
6074 functions when printing out human-readable dumps of ASN.1 data. Therefore
6075 applications that print this data could be vulnerable if the data is from
6076 untrusted sources. OpenSSL command line applications could also be
6077 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6078 as command line arguments.
6080 Libssl is not considered directly vulnerable. Additionally certificates etc
6081 received via remote connections via libssl are also unlikely to be able to
6082 trigger these issues because of message size limits enforced within libssl.
6084 This issue was reported to OpenSSL Guido Vranken.
6089 * Side channel attack on modular exponentiation
6091 A side-channel attack was found which makes use of cache-bank conflicts on
6092 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6093 of RSA keys. The ability to exploit this issue is limited as it relies on
6094 an attacker who has control of code in a thread running on the same
6095 hyper-threaded core as the victim thread which is performing decryptions.
6097 This issue was reported to OpenSSL by Yuval Yarom, The University of
6098 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6099 Nadia Heninger, University of Pennsylvania with more information at
6100 <http://cachebleed.info>.
6105 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6106 if no keysize is specified with default_bits. This fixes an
6107 omission in an earlier change that changed all RSA/DSA key generation
6108 commands to use 2048 bits by default.
6112 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6114 * DH small subgroups
6116 Historically OpenSSL only ever generated DH parameters based on "safe"
6117 primes. More recently (in version 1.0.2) support was provided for
6118 generating X9.42 style parameter files such as those required for RFC 5114
6119 support. The primes used in such files may not be "safe". Where an
6120 application is using DH configured with parameters based on primes that are
6121 not "safe" then an attacker could use this fact to find a peer's private
6122 DH exponent. This attack requires that the attacker complete multiple
6123 handshakes in which the peer uses the same private DH exponent. For example
6124 this could be used to discover a TLS server's private DH exponent if it's
6125 reusing the private DH exponent or it's using a static DH ciphersuite.
6127 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6128 TLS. It is not on by default. If the option is not set then the server
6129 reuses the same private DH exponent for the life of the server process and
6130 would be vulnerable to this attack. It is believed that many popular
6131 applications do set this option and would therefore not be at risk.
6133 The fix for this issue adds an additional check where a "q" parameter is
6134 available (as is the case in X9.42 based parameters). This detects the
6135 only known attack, and is the only possible defense for static DH
6136 ciphersuites. This could have some performance impact.
6138 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6139 default and cannot be disabled. This could have some performance impact.
6141 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6146 * SSLv2 doesn't block disabled ciphers
6148 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6149 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6150 been disabled, provided that the SSLv2 protocol was not also disabled via
6153 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6154 and Sebastian Schinzel.
6159 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6161 * BN_mod_exp may produce incorrect results on x86_64
6163 There is a carry propagating bug in the x86_64 Montgomery squaring
6164 procedure. No EC algorithms are affected. Analysis suggests that attacks
6165 against RSA and DSA as a result of this defect would be very difficult to
6166 perform and are not believed likely. Attacks against DH are considered just
6167 feasible (although very difficult) because most of the work necessary to
6168 deduce information about a private key may be performed offline. The amount
6169 of resources required for such an attack would be very significant and
6170 likely only accessible to a limited number of attackers. An attacker would
6171 additionally need online access to an unpatched system using the target
6172 private key in a scenario with persistent DH parameters and a private
6173 key that is shared between multiple clients. For example this can occur by
6174 default in OpenSSL DHE based SSL/TLS ciphersuites.
6176 This issue was reported to OpenSSL by Hanno Böck.
6181 * Certificate verify crash with missing PSS parameter
6183 The signature verification routines will crash with a NULL pointer
6184 dereference if presented with an ASN.1 signature using the RSA PSS
6185 algorithm and absent mask generation function parameter. Since these
6186 routines are used to verify certificate signature algorithms this can be
6187 used to crash any certificate verification operation and exploited in a
6188 DoS attack. Any application which performs certificate verification is
6189 vulnerable including OpenSSL clients and servers which enable client
6192 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6197 * X509_ATTRIBUTE memory leak
6199 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6200 memory. This structure is used by the PKCS#7 and CMS routines so any
6201 application which reads PKCS#7 or CMS data from untrusted sources is
6202 affected. SSL/TLS is not affected.
6204 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6210 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6211 This changes the decoding behaviour for some invalid messages,
6212 though the change is mostly in the more lenient direction, and
6213 legacy behaviour is preserved as much as possible.
6217 * In DSA_generate_parameters_ex, if the provided seed is too short,
6220 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6222 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6224 * Alternate chains certificate forgery
6226 During certificate verification, OpenSSL will attempt to find an
6227 alternative certificate chain if the first attempt to build such a chain
6228 fails. An error in the implementation of this logic can mean that an
6229 attacker could cause certain checks on untrusted certificates to be
6230 bypassed, such as the CA flag, enabling them to use a valid leaf
6231 certificate to act as a CA and "issue" an invalid certificate.
6233 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6238 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6240 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6241 incompatibility in the handling of HMAC. The previous ABI has now been
6246 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6248 * Malformed ECParameters causes infinite loop
6250 When processing an ECParameters structure OpenSSL enters an infinite loop
6251 if the curve specified is over a specially malformed binary polynomial
6254 This can be used to perform denial of service against any
6255 system which processes public keys, certificate requests or
6256 certificates. This includes TLS clients and TLS servers with
6257 client authentication enabled.
6259 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6264 * Exploitable out-of-bounds read in X509_cmp_time
6266 X509_cmp_time does not properly check the length of the ASN1_TIME
6267 string and can read a few bytes out of bounds. In addition,
6268 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6271 An attacker can use this to craft malformed certificates and CRLs of
6272 various sizes and potentially cause a segmentation fault, resulting in
6273 a DoS on applications that verify certificates or CRLs. TLS clients
6274 that verify CRLs are affected. TLS clients and servers with client
6275 authentication enabled may be affected if they use custom verification
6278 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6279 independently by Hanno Böck.
6284 * PKCS7 crash with missing EnvelopedContent
6286 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6287 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6288 with missing content and trigger a NULL pointer dereference on parsing.
6290 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6291 structures from untrusted sources are affected. OpenSSL clients and
6292 servers are not affected.
6294 This issue was reported to OpenSSL by Michal Zalewski (Google).
6299 * CMS verify infinite loop with unknown hash function
6301 When verifying a signedData message the CMS code can enter an infinite loop
6302 if presented with an unknown hash function OID. This can be used to perform
6303 denial of service against any system which verifies signedData messages using
6305 This issue was reported to OpenSSL by Johannes Bauer.
6310 * Race condition handling NewSessionTicket
6312 If a NewSessionTicket is received by a multi-threaded client when attempting to
6313 reuse a previous ticket then a race condition can occur potentially leading to
6314 a double free of the ticket data.
6319 * Only support 256-bit or stronger elliptic curves with the
6320 'ecdh_auto' setting (server) or by default (client). Of supported
6321 curves, prefer P-256 (both).
6325 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6327 * ClientHello sigalgs DoS fix
6329 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6330 invalid signature algorithms extension a NULL pointer dereference will
6331 occur. This can be exploited in a DoS attack against the server.
6333 This issue was was reported to OpenSSL by David Ramos of Stanford
6337 *Stephen Henson and Matt Caswell*
6339 * Multiblock corrupted pointer fix
6341 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6342 feature only applies on 64 bit x86 architecture platforms that support AES
6343 NI instructions. A defect in the implementation of "multiblock" can cause
6344 OpenSSL's internal write buffer to become incorrectly set to NULL when
6345 using non-blocking IO. Typically, when the user application is using a
6346 socket BIO for writing, this will only result in a failed connection.
6347 However if some other BIO is used then it is likely that a segmentation
6348 fault will be triggered, thus enabling a potential DoS attack.
6350 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6355 * Segmentation fault in DTLSv1_listen fix
6357 The DTLSv1_listen function is intended to be stateless and processes the
6358 initial ClientHello from many peers. It is common for user code to loop
6359 over the call to DTLSv1_listen until a valid ClientHello is received with
6360 an associated cookie. A defect in the implementation of DTLSv1_listen means
6361 that state is preserved in the SSL object from one invocation to the next
6362 that can lead to a segmentation fault. Errors processing the initial
6363 ClientHello can trigger this scenario. An example of such an error could be
6364 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6367 This issue was reported to OpenSSL by Per Allansson.
6372 * Segmentation fault in ASN1_TYPE_cmp fix
6374 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6375 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6376 certificate signature algorithm consistency this can be used to crash any
6377 certificate verification operation and exploited in a DoS attack. Any
6378 application which performs certificate verification is vulnerable including
6379 OpenSSL clients and servers which enable client authentication.
6384 * Segmentation fault for invalid PSS parameters fix
6386 The signature verification routines will crash with a NULL pointer
6387 dereference if presented with an ASN.1 signature using the RSA PSS
6388 algorithm and invalid parameters. Since these routines are used to verify
6389 certificate signature algorithms this can be used to crash any
6390 certificate verification operation and exploited in a DoS attack. Any
6391 application which performs certificate verification is vulnerable including
6392 OpenSSL clients and servers which enable client authentication.
6394 This issue was was reported to OpenSSL by Brian Carpenter.
6399 * ASN.1 structure reuse memory corruption fix
6401 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6402 memory corruption via an invalid write. Such reuse is and has been
6403 strongly discouraged and is believed to be rare.
6405 Applications that parse structures containing CHOICE or ANY DEFINED BY
6406 components may be affected. Certificate parsing (d2i_X509 and related
6407 functions) are however not affected. OpenSSL clients and servers are
6413 * PKCS7 NULL pointer dereferences fix
6415 The PKCS#7 parsing code does not handle missing outer ContentInfo
6416 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6417 missing content and trigger a NULL pointer dereference on parsing.
6419 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6420 otherwise parse PKCS#7 structures from untrusted sources are
6421 affected. OpenSSL clients and servers are not affected.
6423 This issue was reported to OpenSSL by Michal Zalewski (Google).
6428 * DoS via reachable assert in SSLv2 servers fix
6430 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6431 servers that both support SSLv2 and enable export cipher suites by sending
6432 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6434 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6435 (OpenSSL development team).
6440 * Empty CKE with client auth and DHE fix
6442 If client auth is used then a server can seg fault in the event of a DHE
6443 ciphersuite being selected and a zero length ClientKeyExchange message
6444 being sent by the client. This could be exploited in a DoS attack.
6449 * Handshake with unseeded PRNG fix
6451 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6452 with an unseeded PRNG. The conditions are:
6453 - The client is on a platform where the PRNG has not been seeded
6454 automatically, and the user has not seeded manually
6455 - A protocol specific client method version has been used (i.e. not
6456 SSL_client_methodv23)
6457 - A ciphersuite is used that does not require additional random data from
6458 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6460 If the handshake succeeds then the client random that has been used will
6461 have been generated from a PRNG with insufficient entropy and therefore the
6462 output may be predictable.
6464 For example using the following command with an unseeded openssl will
6465 succeed on an unpatched platform:
6467 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6472 * Use After Free following d2i_ECPrivatekey error fix
6474 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6475 could cause a use after free condition. This, in turn, could cause a double
6476 free in several private key parsing functions (such as d2i_PrivateKey
6477 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6478 for applications that receive EC private keys from untrusted
6479 sources. This scenario is considered rare.
6481 This issue was discovered by the BoringSSL project and fixed in their
6487 * X509_to_X509_REQ NULL pointer deref fix
6489 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6490 the certificate key is invalid. This function is rarely used in practice.
6492 This issue was discovered by Brian Carpenter.
6497 * Removed the export ciphers from the DEFAULT ciphers
6501 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6503 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6504 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6505 So far those who have to target multiple platforms would compromise
6506 and argue that binary targeting say ARMv5 would still execute on
6507 ARMv8. "Universal" build resolves this compromise by providing
6508 near-optimal performance even on newer platforms.
6512 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6513 (other platforms pending).
6515 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6517 * Add support for the SignedCertificateTimestampList certificate and
6518 OCSP response extensions from RFC6962.
6522 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6523 for corner cases. (Certain input points at infinity could lead to
6524 bogus results, with non-infinity inputs mapped to infinity too.)
6528 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6529 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6530 common cases are optimized and there still is room for further
6531 improvements. Vector Permutation AES for Altivec is also added.
6535 * Add support for little-endian ppc64 Linux target.
6537 *Marcelo Cerri (IBM)*
6539 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6540 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6541 are optimized and there still is room for further improvements.
6542 Both 32- and 64-bit modes are supported.
6544 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6546 * Improved ARMv7 NEON support.
6550 * Support for SPARC Architecture 2011 crypto extensions, first
6551 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6552 SHA256/512, MD5, GHASH and modular exponentiation.
6554 *Andy Polyakov, David Miller*
6556 * Accelerated modular exponentiation for Intel processors, a.k.a.
6559 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6561 * Support for new and upcoming Intel processors, including AVX2,
6562 BMI and SHA ISA extensions. This includes additional "stitched"
6563 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6566 This work was sponsored by Intel Corp.
6570 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6571 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6572 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6576 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6577 this fixes a limitation in previous versions of OpenSSL.
6581 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6582 MGF1 digest and OAEP label.
6586 * Add EVP support for key wrapping algorithms, to avoid problems with
6587 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6588 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6589 algorithms and include tests cases.
6593 * Add functions to allocate and set the fields of an ECDSA_METHOD
6596 *Douglas E. Engert, Steve Henson*
6598 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6599 difference in days and seconds between two tm or ASN1_TIME structures.
6603 * Add -rev test option to s_server to just reverse order of characters
6604 received by client and send back to server. Also prints an abbreviated
6605 summary of the connection parameters.
6609 * New option -brief for s_client and s_server to print out a brief summary
6610 of connection parameters.
6614 * Add callbacks for arbitrary TLS extensions.
6616 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6618 * New option -crl_download in several openssl utilities to download CRLs
6619 from CRLDP extension in certificates.
6623 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6627 * New function X509_CRL_diff to generate a delta CRL from the difference
6628 of two full CRLs. Add support to "crl" utility.
6632 * New functions to set lookup_crls function and to retrieve
6633 X509_STORE from X509_STORE_CTX.
6637 * Print out deprecated issuer and subject unique ID fields in
6642 * Extend OCSP I/O functions so they can be used for simple general purpose
6643 HTTP as well as OCSP. New wrapper function which can be used to download
6644 CRLs using the OCSP API.
6648 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6652 * `SSL_CONF*` functions. These provide a common framework for application
6653 configuration using configuration files or command lines.
6657 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6658 message callback and prints the results. Needs compile time option
6659 "enable-ssl-trace". New options to s_client and s_server to enable
6664 * New ctrl and macro to retrieve supported points extensions.
6665 Print out extension in s_server and s_client.
6669 * New functions to retrieve certificate signature and signature
6674 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6679 * New Suite B modes for TLS code. These use and enforce the requirements
6680 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6681 only use Suite B curves. The Suite B modes can be set by using the
6682 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6686 * New chain verification flags for Suite B levels of security. Check
6687 algorithms are acceptable when flags are set in X509_verify_cert.
6691 * Make tls1_check_chain return a set of flags indicating checks passed
6692 by a certificate chain. Add additional tests to handle client
6693 certificates: checks for matching certificate type and issuer name
6698 * If an attempt is made to use a signature algorithm not in the peer
6699 preference list abort the handshake. If client has no suitable
6700 signature algorithms in response to a certificate request do not
6701 use the certificate.
6705 * If server EC tmp key is not in client preference list abort handshake.
6709 * Add support for certificate stores in CERT structure. This makes it
6710 possible to have different stores per SSL structure or one store in
6711 the parent SSL_CTX. Include distinct stores for certificate chain
6712 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6713 to build and store a certificate chain in CERT structure: returning
6714 an error if the chain cannot be built: this will allow applications
6715 to test if a chain is correctly configured.
6717 Note: if the CERT based stores are not set then the parent SSL_CTX
6718 store is used to retain compatibility with existing behaviour.
6722 * New function ssl_set_client_disabled to set a ciphersuite disabled
6723 mask based on the current session, check mask when sending client
6724 hello and checking the requested ciphersuite.
6728 * New ctrls to retrieve and set certificate types in a certificate
6729 request message. Print out received values in s_client. If certificate
6730 types is not set with custom values set sensible values based on
6731 supported signature algorithms.
6735 * Support for distinct client and server supported signature algorithms.
6739 * Add certificate callback. If set this is called whenever a certificate
6740 is required by client or server. An application can decide which
6741 certificate chain to present based on arbitrary criteria: for example
6742 supported signature algorithms. Add very simple example to s_server.
6743 This fixes many of the problems and restrictions of the existing client
6744 certificate callback: for example you can now clear an existing
6745 certificate and specify the whole chain.
6749 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6750 the certificate can be used for (if anything). Set valid_flags field
6751 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6752 to have similar checks in it.
6754 Add new "cert_flags" field to CERT structure and include a "strict mode".
6755 This enforces some TLS certificate requirements (such as only permitting
6756 certificate signature algorithms contained in the supported algorithms
6757 extension) which some implementations ignore: this option should be used
6758 with caution as it could cause interoperability issues.
6762 * Update and tidy signature algorithm extension processing. Work out
6763 shared signature algorithms based on preferences and peer algorithms
6764 and print them out in s_client and s_server. Abort handshake if no
6765 shared signature algorithms.
6769 * Add new functions to allow customised supported signature algorithms
6770 for SSL and SSL_CTX structures. Add options to s_client and s_server
6775 * New function SSL_certs_clear() to delete all references to certificates
6776 from an SSL structure. Before this once a certificate had been added
6777 it couldn't be removed.
6781 * Integrate hostname, email address and IP address checking with certificate
6782 verification. New verify options supporting checking in openssl utility.
6786 * Fixes and wildcard matching support to hostname and email checking
6787 functions. Add manual page.
6789 *Florian Weimer (Red Hat Product Security Team)*
6791 * New functions to check a hostname email or IP address against a
6792 certificate. Add options x509 utility to print results of checks against
6797 * Fix OCSP checking.
6799 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6801 * Initial experimental support for explicitly trusted non-root CAs.
6802 OpenSSL still tries to build a complete chain to a root but if an
6803 intermediate CA has a trust setting included that is used. The first
6804 setting is used: whether to trust (e.g., -addtrust option to the x509
6809 * Add -trusted_first option which attempts to find certificates in the
6810 trusted store even if an untrusted chain is also supplied.
6814 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6815 platform support for Linux and Android.
6819 * Support for linux-x32, ILP32 environment in x86_64 framework.
6823 * Experimental multi-implementation support for FIPS capable OpenSSL.
6824 When in FIPS mode the approved implementations are used as normal,
6825 when not in FIPS mode the internal unapproved versions are used instead.
6826 This means that the FIPS capable OpenSSL isn't forced to use the
6827 (often lower performance) FIPS implementations outside FIPS mode.
6831 * Transparently support X9.42 DH parameters when calling
6832 PEM_read_bio_DHparameters. This means existing applications can handle
6833 the new parameter format automatically.
6837 * Initial experimental support for X9.42 DH parameter format: mainly
6838 to support use of 'q' parameter for RFC5114 parameters.
6842 * Add DH parameters from RFC5114 including test data to dhtest.
6846 * Support for automatic EC temporary key parameter selection. If enabled
6847 the most preferred EC parameters are automatically used instead of
6848 hardcoded fixed parameters. Now a server just has to call:
6849 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6850 support ECDH and use the most appropriate parameters.
6854 * Enhance and tidy EC curve and point format TLS extension code. Use
6855 static structures instead of allocation if default values are used.
6856 New ctrls to set curves we wish to support and to retrieve shared curves.
6857 Print out shared curves in s_server. New options to s_server and s_client
6858 to set list of supported curves.
6862 * New ctrls to retrieve supported signature algorithms and
6863 supported curve values as an array of NIDs. Extend openssl utility
6864 to print out received values.
6868 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6869 between NIDs and the more common NIST names such as "P-256". Enhance
6870 ecparam utility and ECC method to recognise the NIST names for curves.
6874 * Enhance SSL/TLS certificate chain handling to support different
6875 chains for each certificate instead of one chain in the parent SSL_CTX.
6879 * Support for fixed DH ciphersuite client authentication: where both
6880 server and client use DH certificates with common parameters.
6884 * Support for fixed DH ciphersuites: those requiring DH server
6889 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6891 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6892 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6893 X509_CINF_get_signature were reverted post internal team review.
6898 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6900 * OCSP Status Request extension unbounded memory growth
6902 A malicious client can send an excessively large OCSP Status Request
6903 extension. If that client continually requests renegotiation, sending a
6904 large OCSP Status Request extension each time, then there will be unbounded
6905 memory growth on the server. This will eventually lead to a Denial Of
6906 Service attack through memory exhaustion. Servers with a default
6907 configuration are vulnerable even if they do not support OCSP. Builds using
6908 the "no-ocsp" build time option are not affected.
6910 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6915 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6918 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6924 * OOB write in MDC2_Update()
6926 An overflow can occur in MDC2_Update() either if called directly or
6927 through the EVP_DigestUpdate() function using MDC2. If an attacker
6928 is able to supply very large amounts of input data after a previous
6929 call to EVP_EncryptUpdate() with a partial block then a length check
6930 can overflow resulting in a heap corruption.
6932 The amount of data needed is comparable to SIZE_MAX which is impractical
6935 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6940 * Malformed SHA512 ticket DoS
6942 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6943 DoS attack where a malformed ticket will result in an OOB read which will
6946 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6947 a custom server callback and ticket lookup mechanism.
6949 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6954 * OOB write in BN_bn2dec()
6956 The function BN_bn2dec() does not check the return value of BN_div_word().
6957 This can cause an OOB write if an application uses this function with an
6958 overly large BIGNUM. This could be a problem if an overly large certificate
6959 or CRL is printed out from an untrusted source. TLS is not affected because
6960 record limits will reject an oversized certificate before it is parsed.
6962 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6967 * OOB read in TS_OBJ_print_bio()
6969 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6970 the total length the OID text representation would use and not the amount
6971 of data written. This will result in OOB reads when large OIDs are
6974 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6979 * Pointer arithmetic undefined behaviour
6981 Avoid some undefined pointer arithmetic
6983 A common idiom in the codebase is to check limits in the following manner:
6986 Where "p" points to some malloc'd data of SIZE bytes and
6989 "len" here could be from some externally supplied data (e.g. from a TLS
6992 The rules of C pointer arithmetic are such that "p + len" is only well
6993 defined where len <= SIZE. Therefore, the above idiom is actually
6994 undefined behaviour.
6996 For example this could cause problems if some malloc implementation
6997 provides an address for "p" such that "p + len" actually overflows for
6998 values of len that are too big and therefore p + len < limit.
7000 This issue was reported to OpenSSL by Guido Vranken
7005 * Constant time flag not preserved in DSA signing
7007 Operations in the DSA signing algorithm should run in constant time in
7008 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7009 implementation means that a non-constant time codepath is followed for
7010 certain operations. This has been demonstrated through a cache-timing
7011 attack to be sufficient for an attacker to recover the private DSA key.
7013 This issue was reported by César Pereida (Aalto University), Billy Brumley
7014 (Tampere University of Technology), and Yuval Yarom (The University of
7015 Adelaide and NICTA).
7020 * DTLS buffered message DoS
7022 In a DTLS connection where handshake messages are delivered out-of-order
7023 those messages that OpenSSL is not yet ready to process will be buffered
7024 for later use. Under certain circumstances, a flaw in the logic means that
7025 those messages do not get removed from the buffer even though the handshake
7026 has been completed. An attacker could force up to approx. 15 messages to
7027 remain in the buffer when they are no longer required. These messages will
7028 be cleared when the DTLS connection is closed. The default maximum size for
7029 a message is 100k. Therefore, the attacker could force an additional 1500k
7030 to be consumed per connection. By opening many simultaneous connections an
7031 attacker could cause a DoS attack through memory exhaustion.
7033 This issue was reported to OpenSSL by Quan Luo.
7038 * DTLS replay protection DoS
7040 A flaw in the DTLS replay attack protection mechanism means that records
7041 that arrive for future epochs update the replay protection "window" before
7042 the MAC for the record has been validated. This could be exploited by an
7043 attacker by sending a record for the next epoch (which does not have to
7044 decrypt or have a valid MAC), with a very large sequence number. This means
7045 that all subsequent legitimate packets are dropped causing a denial of
7046 service for a specific DTLS connection.
7048 This issue was reported to OpenSSL by the OCAP audit team.
7053 * Certificate message OOB reads
7055 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7056 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7057 theoretical DoS risk but this has not been observed in practice on common
7060 The messages affected are client certificate, client certificate request
7061 and server certificate. As a result the attack can only be performed
7062 against a client or a server which enables client authentication.
7064 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7069 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7071 * Prevent padding oracle in AES-NI CBC MAC check
7073 A MITM attacker can use a padding oracle attack to decrypt traffic
7074 when the connection uses an AES CBC cipher and the server support
7077 This issue was introduced as part of the fix for Lucky 13 padding
7078 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7079 constant time by making sure that always the same bytes are read and
7080 compared against either the MAC or padding bytes. But it no longer
7081 checked that there was enough data to have both the MAC and padding
7084 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7089 * Fix EVP_EncodeUpdate overflow
7091 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7092 Base64 encoding of binary data. If an attacker is able to supply very large
7093 amounts of input data then a length check can overflow resulting in a heap
7096 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7097 the `PEM_write_bio*` family of functions. These are mainly used within the
7098 OpenSSL command line applications, so any application which processes data
7099 from an untrusted source and outputs it as a PEM file should be considered
7100 vulnerable to this issue. User applications that call these APIs directly
7101 with large amounts of untrusted data may also be vulnerable.
7103 This issue was reported by Guido Vranken.
7108 * Fix EVP_EncryptUpdate overflow
7110 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7111 is able to supply very large amounts of input data after a previous call to
7112 EVP_EncryptUpdate() with a partial block then a length check can overflow
7113 resulting in a heap corruption. Following an analysis of all OpenSSL
7114 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7115 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7116 the first called function after an EVP_EncryptInit(), and therefore that
7117 specific call must be safe. The second form is where the length passed to
7118 EVP_EncryptUpdate() can be seen from the code to be some small value and
7119 therefore there is no possibility of an overflow. Since all instances are
7120 one of these two forms, it is believed that there can be no overflows in
7121 internal code due to this problem. It should be noted that
7122 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7123 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7124 of these calls have also been analysed too and it is believed there are no
7125 instances in internal usage where an overflow could occur.
7127 This issue was reported by Guido Vranken.
7132 * Prevent ASN.1 BIO excessive memory allocation
7134 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7135 a short invalid encoding can casuse allocation of large amounts of memory
7136 potentially consuming excessive resources or exhausting memory.
7138 Any application parsing untrusted data through d2i BIO functions is
7139 affected. The memory based functions such as d2i_X509() are *not* affected.
7140 Since the memory based functions are used by the TLS library, TLS
7141 applications are not affected.
7143 This issue was reported by Brian Carpenter.
7150 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7151 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7152 in arbitrary stack data being returned in the buffer.
7154 This issue was reported by Guido Vranken.
7159 * Modify behavior of ALPN to invoke callback after SNI/servername
7160 callback, such that updates to the SSL_CTX affect ALPN.
7164 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7169 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7170 methods are enabled and ssl2 is disabled the methods return NULL.
7174 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7176 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7177 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7178 provide any "EXPORT" or "LOW" strength ciphers.
7182 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7183 is by default disabled at build-time. Builds that are not configured with
7184 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7185 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7186 will need to explicitly call either of:
7188 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7190 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7192 as appropriate. Even if either of those is used, or the application
7193 explicitly uses the version-specific SSLv2_method() or its client and
7194 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7195 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7196 ciphers, and SSLv2 56-bit DES are no longer available.
7201 * Fix a double-free in DSA code
7203 A double free bug was discovered when OpenSSL parses malformed DSA private
7204 keys and could lead to a DoS attack or memory corruption for applications
7205 that receive DSA private keys from untrusted sources. This scenario is
7208 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7214 * Disable SRP fake user seed to address a server memory leak.
7216 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7218 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7219 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7220 was changed to ignore the "fake user" SRP seed, even if the seed
7223 Users should use SRP_VBASE_get1_by_user instead. Note that in
7224 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7225 also that even though configuring the SRP seed attempts to hide
7226 invalid usernames by continuing the handshake with fake
7227 credentials, this behaviour is not constant time and no strong
7228 guarantees are made that the handshake is indistinguishable from
7229 that of a valid user.
7234 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7236 In the BN_hex2bn function the number of hex digits is calculated using an
7237 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7238 large values of `i` this can result in `bn_expand` not allocating any
7239 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7240 field as NULL leading to a subsequent NULL ptr deref. For very large values
7241 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7242 In this case memory is allocated to the internal BIGNUM data field, but it
7243 is insufficiently sized leading to heap corruption. A similar issue exists
7244 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7245 is ever called by user applications with very large untrusted hex/dec data.
7246 This is anticipated to be a rare occurrence.
7248 All OpenSSL internal usage of these functions use data that is not expected
7249 to be untrusted, e.g. config file data or application command line
7250 arguments. If user developed applications generate config file data based
7251 on untrusted data then it is possible that this could also lead to security
7252 consequences. This is also anticipated to be rare.
7254 This issue was reported to OpenSSL by Guido Vranken.
7259 * Fix memory issues in `BIO_*printf` functions
7261 The internal `fmtstr` function used in processing a "%s" format string in
7262 the `BIO_*printf` functions could overflow while calculating the length of a
7263 string and cause an OOB read when printing very long strings.
7265 Additionally the internal `doapr_outch` function can attempt to write to an
7266 OOB memory location (at an offset from the NULL pointer) in the event of a
7267 memory allocation failure. In 1.0.2 and below this could be caused where
7268 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7269 could be in processing a very long "%s" format string. Memory leaks can
7272 The first issue may mask the second issue dependent on compiler behaviour.
7273 These problems could enable attacks where large amounts of untrusted data
7274 is passed to the `BIO_*printf` functions. If applications use these functions
7275 in this way then they could be vulnerable. OpenSSL itself uses these
7276 functions when printing out human-readable dumps of ASN.1 data. Therefore
7277 applications that print this data could be vulnerable if the data is from
7278 untrusted sources. OpenSSL command line applications could also be
7279 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7280 as command line arguments.
7282 Libssl is not considered directly vulnerable. Additionally certificates etc
7283 received via remote connections via libssl are also unlikely to be able to
7284 trigger these issues because of message size limits enforced within libssl.
7286 This issue was reported to OpenSSL Guido Vranken.
7291 * Side channel attack on modular exponentiation
7293 A side-channel attack was found which makes use of cache-bank conflicts on
7294 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7295 of RSA keys. The ability to exploit this issue is limited as it relies on
7296 an attacker who has control of code in a thread running on the same
7297 hyper-threaded core as the victim thread which is performing decryptions.
7299 This issue was reported to OpenSSL by Yuval Yarom, The University of
7300 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7301 Nadia Heninger, University of Pennsylvania with more information at
7302 <http://cachebleed.info>.
7307 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7308 if no keysize is specified with default_bits. This fixes an
7309 omission in an earlier change that changed all RSA/DSA key generation
7310 commands to use 2048 bits by default.
7314 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7316 * Protection for DH small subgroup attacks
7318 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7319 switched on by default and cannot be disabled. This could have some
7324 * SSLv2 doesn't block disabled ciphers
7326 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7327 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7328 been disabled, provided that the SSLv2 protocol was not also disabled via
7331 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7332 and Sebastian Schinzel.
7337 * Reject DH handshakes with parameters shorter than 1024 bits.
7341 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7343 * Certificate verify crash with missing PSS parameter
7345 The signature verification routines will crash with a NULL pointer
7346 dereference if presented with an ASN.1 signature using the RSA PSS
7347 algorithm and absent mask generation function parameter. Since these
7348 routines are used to verify certificate signature algorithms this can be
7349 used to crash any certificate verification operation and exploited in a
7350 DoS attack. Any application which performs certificate verification is
7351 vulnerable including OpenSSL clients and servers which enable client
7354 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7359 * X509_ATTRIBUTE memory leak
7361 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7362 memory. This structure is used by the PKCS#7 and CMS routines so any
7363 application which reads PKCS#7 or CMS data from untrusted sources is
7364 affected. SSL/TLS is not affected.
7366 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7372 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7373 This changes the decoding behaviour for some invalid messages,
7374 though the change is mostly in the more lenient direction, and
7375 legacy behaviour is preserved as much as possible.
7379 * In DSA_generate_parameters_ex, if the provided seed is too short,
7380 use a random seed, as already documented.
7382 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7384 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7386 * Alternate chains certificate forgery
7388 During certificate verfification, OpenSSL will attempt to find an
7389 alternative certificate chain if the first attempt to build such a chain
7390 fails. An error in the implementation of this logic can mean that an
7391 attacker could cause certain checks on untrusted certificates to be
7392 bypassed, such as the CA flag, enabling them to use a valid leaf
7393 certificate to act as a CA and "issue" an invalid certificate.
7395 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7401 * Race condition handling PSK identify hint
7403 If PSK identity hints are received by a multi-threaded client then
7404 the values are wrongly updated in the parent SSL_CTX structure. This can
7405 result in a race condition potentially leading to a double free of the
7411 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7413 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7414 incompatibility in the handling of HMAC. The previous ABI has now been
7417 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7419 * Malformed ECParameters causes infinite loop
7421 When processing an ECParameters structure OpenSSL enters an infinite loop
7422 if the curve specified is over a specially malformed binary polynomial
7425 This can be used to perform denial of service against any
7426 system which processes public keys, certificate requests or
7427 certificates. This includes TLS clients and TLS servers with
7428 client authentication enabled.
7430 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7435 * Exploitable out-of-bounds read in X509_cmp_time
7437 X509_cmp_time does not properly check the length of the ASN1_TIME
7438 string and can read a few bytes out of bounds. In addition,
7439 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7442 An attacker can use this to craft malformed certificates and CRLs of
7443 various sizes and potentially cause a segmentation fault, resulting in
7444 a DoS on applications that verify certificates or CRLs. TLS clients
7445 that verify CRLs are affected. TLS clients and servers with client
7446 authentication enabled may be affected if they use custom verification
7449 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7450 independently by Hanno Böck.
7455 * PKCS7 crash with missing EnvelopedContent
7457 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7458 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7459 with missing content and trigger a NULL pointer dereference on parsing.
7461 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7462 structures from untrusted sources are affected. OpenSSL clients and
7463 servers are not affected.
7465 This issue was reported to OpenSSL by Michal Zalewski (Google).
7470 * CMS verify infinite loop with unknown hash function
7472 When verifying a signedData message the CMS code can enter an infinite loop
7473 if presented with an unknown hash function OID. This can be used to perform
7474 denial of service against any system which verifies signedData messages using
7476 This issue was reported to OpenSSL by Johannes Bauer.
7481 * Race condition handling NewSessionTicket
7483 If a NewSessionTicket is received by a multi-threaded client when attempting to
7484 reuse a previous ticket then a race condition can occur potentially leading to
7485 a double free of the ticket data.
7490 * Reject DH handshakes with parameters shorter than 768 bits.
7492 *Kurt Roeckx and Emilia Kasper*
7494 * dhparam: generate 2048-bit parameters by default.
7496 *Kurt Roeckx and Emilia Kasper*
7498 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7500 * Segmentation fault in ASN1_TYPE_cmp fix
7502 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7503 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7504 certificate signature algorithm consistency this can be used to crash any
7505 certificate verification operation and exploited in a DoS attack. Any
7506 application which performs certificate verification is vulnerable including
7507 OpenSSL clients and servers which enable client authentication.
7512 * ASN.1 structure reuse memory corruption fix
7514 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7515 memory corruption via an invalid write. Such reuse is and has been
7516 strongly discouraged and is believed to be rare.
7518 Applications that parse structures containing CHOICE or ANY DEFINED BY
7519 components may be affected. Certificate parsing (d2i_X509 and related
7520 functions) are however not affected. OpenSSL clients and servers are
7526 * PKCS7 NULL pointer dereferences fix
7528 The PKCS#7 parsing code does not handle missing outer ContentInfo
7529 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7530 missing content and trigger a NULL pointer dereference on parsing.
7532 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7533 otherwise parse PKCS#7 structures from untrusted sources are
7534 affected. OpenSSL clients and servers are not affected.
7536 This issue was reported to OpenSSL by Michal Zalewski (Google).
7541 * DoS via reachable assert in SSLv2 servers fix
7543 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7544 servers that both support SSLv2 and enable export cipher suites by sending
7545 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7547 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7548 (OpenSSL development team).
7553 * Use After Free following d2i_ECPrivatekey error fix
7555 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7556 could cause a use after free condition. This, in turn, could cause a double
7557 free in several private key parsing functions (such as d2i_PrivateKey
7558 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7559 for applications that receive EC private keys from untrusted
7560 sources. This scenario is considered rare.
7562 This issue was discovered by the BoringSSL project and fixed in their
7568 * X509_to_X509_REQ NULL pointer deref fix
7570 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7571 the certificate key is invalid. This function is rarely used in practice.
7573 This issue was discovered by Brian Carpenter.
7578 * Removed the export ciphers from the DEFAULT ciphers
7582 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7584 * Build fixes for the Windows and OpenVMS platforms
7586 *Matt Caswell and Richard Levitte*
7588 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7590 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7591 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7592 dereference. This could lead to a Denial Of Service attack. Thanks to
7593 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7598 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7599 dtls1_buffer_record function under certain conditions. In particular this
7600 could occur if an attacker sent repeated DTLS records with the same
7601 sequence number but for the next epoch. The memory leak could be exploited
7602 by an attacker in a Denial of Service attack through memory exhaustion.
7603 Thanks to Chris Mueller for reporting this issue.
7608 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7609 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7610 method would be set to NULL which could later result in a NULL pointer
7611 dereference. Thanks to Frank Schmirler for reporting this issue.
7616 * Abort handshake if server key exchange message is omitted for ephemeral
7619 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7620 reporting this issue.
7625 * Remove non-export ephemeral RSA code on client and server. This code
7626 violated the TLS standard by allowing the use of temporary RSA keys in
7627 non-export ciphersuites and could be used by a server to effectively
7628 downgrade the RSA key length used to a value smaller than the server
7629 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7630 INRIA or reporting this issue.
7635 * Fixed issue where DH client certificates are accepted without verification.
7636 An OpenSSL server will accept a DH certificate for client authentication
7637 without the certificate verify message. This effectively allows a client to
7638 authenticate without the use of a private key. This only affects servers
7639 which trust a client certificate authority which issues certificates
7640 containing DH keys: these are extremely rare and hardly ever encountered.
7641 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7647 * Ensure that the session ID context of an SSL is updated when its
7648 SSL_CTX is updated via SSL_set_SSL_CTX.
7650 The session ID context is typically set from the parent SSL_CTX,
7651 and can vary with the CTX.
7655 * Fix various certificate fingerprint issues.
7657 By using non-DER or invalid encodings outside the signed portion of a
7658 certificate the fingerprint can be changed without breaking the signature.
7659 Although no details of the signed portion of the certificate can be changed
7660 this can cause problems with some applications: e.g. those using the
7661 certificate fingerprint for blacklists.
7663 1. Reject signatures with non zero unused bits.
7665 If the BIT STRING containing the signature has non zero unused bits reject
7666 the signature. All current signature algorithms require zero unused bits.
7668 2. Check certificate algorithm consistency.
7670 Check the AlgorithmIdentifier inside TBS matches the one in the
7671 certificate signature. NB: this will result in signature failure
7672 errors for some broken certificates.
7674 Thanks to Konrad Kraszewski from Google for reporting this issue.
7676 3. Check DSA/ECDSA signatures use DER.
7678 Re-encode DSA/ECDSA signatures and compare with the original received
7679 signature. Return an error if there is a mismatch.
7681 This will reject various cases including garbage after signature
7682 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7683 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7684 (negative or with leading zeroes).
7686 Further analysis was conducted and fixes were developed by Stephen Henson
7687 of the OpenSSL core team.
7693 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7694 results on some platforms, including x86_64. This bug occurs at random
7695 with a very low probability, and is not known to be exploitable in any
7696 way, though its exact impact is difficult to determine. Thanks to Pieter
7697 Wuille (Blockstream) who reported this issue and also suggested an initial
7698 fix. Further analysis was conducted by the OpenSSL development team and
7699 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7700 the OpenSSL core team.
7705 * Do not resume sessions on the server if the negotiated protocol
7706 version does not match the session's version. Resuming with a different
7707 version, while not strictly forbidden by the RFC, is of questionable
7708 sanity and breaks all known clients.
7710 *David Benjamin, Emilia Käsper*
7712 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7713 early CCS messages during renegotiation. (Note that because
7714 renegotiation is encrypted, this early CCS was not exploitable.)
7718 * Tighten client-side session ticket handling during renegotiation:
7719 ensure that the client only accepts a session ticket if the server sends
7720 the extension anew in the ServerHello. Previously, a TLS client would
7721 reuse the old extension state and thus accept a session ticket if one was
7722 announced in the initial ServerHello.
7724 Similarly, ensure that the client requires a session ticket if one
7725 was advertised in the ServerHello. Previously, a TLS client would
7726 ignore a missing NewSessionTicket message.
7730 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7734 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7735 sends a carefully crafted handshake message, to cause OpenSSL to fail
7736 to free up to 64k of memory causing a memory leak. This could be
7737 exploited in a Denial Of Service attack. This issue affects OpenSSL
7738 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7739 whether SRTP is used or configured. Implementations of OpenSSL that
7740 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7742 The fix was developed by the OpenSSL team.
7747 * Session Ticket Memory Leak.
7749 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7750 integrity of that ticket is first verified. In the event of a session
7751 ticket integrity check failing, OpenSSL will fail to free memory
7752 causing a memory leak. By sending a large number of invalid session
7753 tickets an attacker could exploit this issue in a Denial Of Service
7759 * Build option no-ssl3 is incomplete.
7761 When OpenSSL is configured with "no-ssl3" as a build option, servers
7762 could accept and complete an SSL 3.0 handshake, and clients could be
7763 configured to send them.
7766 *Akamai and the OpenSSL team*
7768 * Add support for TLS_FALLBACK_SCSV.
7769 Client applications doing fallback retries should call
7770 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7773 *Adam Langley, Bodo Moeller*
7775 * Add additional DigestInfo checks.
7777 Re-encode DigestInto in DER and check against the original when
7778 verifying RSA signature: this will reject any improperly encoded
7779 DigestInfo structures.
7781 Note: this is a precautionary measure and no attacks are currently known.
7785 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7787 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7788 SRP code can be overrun an internal buffer. Add sanity check that
7789 g, A, B < N to SRP code.
7791 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7792 Group for discovering this issue.
7797 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7798 TLS 1.0 instead of higher protocol versions when the ClientHello message
7799 is badly fragmented. This allows a man-in-the-middle attacker to force a
7800 downgrade to TLS 1.0 even if both the server and the client support a
7801 higher protocol version, by modifying the client's TLS records.
7803 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7804 researching this issue.
7809 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7810 to a denial of service attack. A malicious server can crash the client
7811 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7812 ciphersuite and sending carefully crafted handshake messages.
7814 Thanks to Felix Gröbert (Google) for discovering and researching this
7820 * By sending carefully crafted DTLS packets an attacker could cause openssl
7821 to leak memory. This can be exploited through a Denial of Service attack.
7822 Thanks to Adam Langley for discovering and researching this issue.
7827 * An attacker can force openssl to consume large amounts of memory whilst
7828 processing DTLS handshake messages. This can be exploited through a
7829 Denial of Service attack.
7830 Thanks to Adam Langley for discovering and researching this issue.
7835 * An attacker can force an error condition which causes openssl to crash
7836 whilst processing DTLS packets due to memory being freed twice. This
7837 can be exploited through a Denial of Service attack.
7838 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7844 * If a multithreaded client connects to a malicious server using a resumed
7845 session and the server sends an ec point format extension it could write
7846 up to 255 bytes to freed memory.
7848 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7854 * A malicious server can crash an OpenSSL client with a null pointer
7855 dereference (read) by specifying an SRP ciphersuite even though it was not
7856 properly negotiated with the client. This can be exploited through a
7857 Denial of Service attack.
7859 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7860 discovering and researching this issue.
7865 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7866 X509_name_oneline, X509_name_print_ex et al. to leak some information
7867 from the stack. Applications may be affected if they echo pretty printing
7868 output to the attacker.
7870 Thanks to Ivan Fratric (Google) for discovering this issue.
7873 *Emilia Käsper, and Steve Henson*
7875 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7876 for corner cases. (Certain input points at infinity could lead to
7877 bogus results, with non-infinity inputs mapped to infinity too.)
7881 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7883 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7884 handshake can force the use of weak keying material in OpenSSL
7885 SSL/TLS clients and servers.
7887 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7888 researching this issue. ([CVE-2014-0224])
7890 *KIKUCHI Masashi, Steve Henson*
7892 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7893 OpenSSL DTLS client the code can be made to recurse eventually crashing
7896 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7899 *Imre Rad, Steve Henson*
7901 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7902 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7903 client or server. This is potentially exploitable to run arbitrary
7904 code on a vulnerable client or server.
7906 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7908 *Jüri Aedla, Steve Henson*
7910 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7911 are subject to a denial of service attack.
7913 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7914 this issue. ([CVE-2014-3470])
7916 *Felix Gröbert, Ivan Fratric, Steve Henson*
7918 * Harmonize version and its documentation. -f flag is used to display
7921 *mancha <mancha1@zoho.com>*
7923 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7924 in i2d_ECPrivateKey.
7926 *mancha <mancha1@zoho.com>*
7928 * Fix some double frees. These are not thought to be exploitable.
7930 *mancha <mancha1@zoho.com>*
7932 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7934 * A missing bounds check in the handling of the TLS heartbeat extension
7935 can be used to reveal up to 64k of memory to a connected client or
7938 Thanks for Neel Mehta of Google Security for discovering this bug and to
7939 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7940 preparing the fix ([CVE-2014-0160])
7942 *Adam Langley, Bodo Moeller*
7944 * Fix for the attack described in the paper "Recovering OpenSSL
7945 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7946 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7947 <http://eprint.iacr.org/2014/140>
7949 Thanks to Yuval Yarom and Naomi Benger for discovering this
7950 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7952 *Yuval Yarom and Naomi Benger*
7954 * TLS pad extension: draft-agl-tls-padding-03
7956 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7957 TLS client Hello record length value would otherwise be > 255 and
7958 less that 512 pad with a dummy extension containing zeroes so it
7959 is at least 512 bytes long.
7961 *Adam Langley, Steve Henson*
7963 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7965 * Fix for TLS record tampering bug. A carefully crafted invalid
7966 handshake could crash OpenSSL with a NULL pointer exception.
7967 Thanks to Anton Johansson for reporting this issues.
7970 * Keep original DTLS digest and encryption contexts in retransmission
7971 structures so we can use the previous session parameters if they need
7972 to be resent. ([CVE-2013-6450])
7976 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7977 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7978 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7979 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7980 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7981 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7983 *Rob Stradling, Adam Langley*
7985 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7987 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7988 supporting platforms or when small records were transferred.
7990 *Andy Polyakov, Steve Henson*
7992 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7994 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7996 This addresses the flaw in CBC record processing discovered by
7997 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7998 at: <http://www.isg.rhul.ac.uk/tls/>
8000 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8001 Security Group at Royal Holloway, University of London
8002 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8003 Emilia Käsper for the initial patch.
8006 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8008 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8009 ciphersuites which can be exploited in a denial of service attack.
8010 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8011 and detecting this bug and to Wolfgang Ettlinger
8012 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8017 * Return an error when checking OCSP signatures when key is NULL.
8018 This fixes a DoS attack. ([CVE-2013-0166])
8022 * Make openssl verify return errors.
8024 *Chris Palmer <palmer@google.com> and Ben Laurie*
8026 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8027 the right response is stapled. Also change SSL_get_certificate()
8028 so it returns the certificate actually sent.
8029 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8031 *Rob Stradling <rob.stradling@comodo.com>*
8033 * Fix possible deadlock when decoding public keys.
8037 * Don't use TLS 1.0 record version number in initial client hello
8042 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8044 * Sanity check record length before skipping explicit IV in TLS
8045 1.2, 1.1 and DTLS to fix DoS attack.
8047 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8048 fuzzing as a service testing platform.
8053 * Initialise tkeylen properly when encrypting CMS messages.
8054 Thanks to Solar Designer of Openwall for reporting this issue.
8058 * In FIPS mode don't try to use composite ciphers as they are not
8063 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8065 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8066 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8067 mean any application compiled against OpenSSL 1.0.0 headers setting
8068 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8069 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8070 0x10000000L Any application which was previously compiled against
8071 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8072 will need to be recompiled as a result. Letting be results in
8073 inability to disable specifically TLS 1.1 and in client context,
8074 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8078 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8079 disable just protocol X, but all protocols above X *if* there are
8080 protocols *below* X still enabled. In more practical terms it means
8081 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8082 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8083 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8088 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8090 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8091 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8092 in CRYPTO_realloc_clean.
8094 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8095 issue and to Adam Langley <agl@chromium.org> for fixing it.
8098 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8100 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8104 * Workarounds for some broken servers that "hang" if a client hello
8105 record length exceeds 255 bytes.
8107 1. Do not use record version number > TLS 1.0 in initial client
8108 hello: some (but not all) hanging servers will now work.
8109 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8110 the number of ciphers sent in the client hello. This should be
8111 set to an even number, such as 50, for example by passing:
8112 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8113 Most broken servers should now work.
8114 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8115 TLS 1.2 client support entirely.
8119 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8123 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8125 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8126 STRING form instead of a DigestInfo.
8130 * The format used for MDC2 RSA signatures is inconsistent between EVP
8131 and the RSA_sign/RSA_verify functions. This was made more apparent when
8132 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8133 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8134 the correct format in RSA_verify so both forms transparently work.
8138 * Some servers which support TLS 1.0 can choke if we initially indicate
8139 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8140 encrypted premaster secret. As a workaround use the maximum permitted
8141 client version in client hello, this should keep such servers happy
8142 and still work with previous versions of OpenSSL.
8146 * Add support for TLS/DTLS heartbeats.
8148 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8150 * Add support for SCTP.
8152 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8154 * Improved PRNG seeding for VOS.
8156 *Paul Green <Paul.Green@stratus.com>*
8158 * Extensive assembler packs updates, most notably:
8160 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8161 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8162 - x86_64: bit-sliced AES implementation;
8163 - ARM: NEON support, contemporary platforms optimizations;
8164 - s390x: z196 support;
8165 - `*`: GHASH and GF(2^m) multiplication implementations;
8169 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8170 (removal of unnecessary code)
8172 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8174 * Add TLS key material exporter from RFC 5705.
8178 * Add DTLS-SRTP negotiation from RFC 5764.
8182 * Add Next Protocol Negotiation,
8183 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8184 disabled with a no-npn flag to config or Configure. Code donated
8187 *Adam Langley <agl@google.com> and Ben Laurie*
8189 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8190 NIST-P256, NIST-P521, with constant-time single point multiplication on
8191 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8192 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8193 Code made available under Apache License version 2.0.
8195 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8196 line to include this in your build of OpenSSL, and run "make depend" (or
8197 "make update"). This enables the following EC_METHODs:
8199 EC_GFp_nistp224_method()
8200 EC_GFp_nistp256_method()
8201 EC_GFp_nistp521_method()
8203 EC_GROUP_new_by_curve_name() will automatically use these (while
8204 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8207 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8209 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8210 all platforms. Move ssize_t definition from e_os.h to the public
8211 header file e_os2.h as it now appears in public header file cms.h
8215 * New -sigopt option to the ca, req and x509 utilities. Additional
8216 signature parameters can be passed using this option and in
8221 * Add RSA PSS signing function. This will generate and set the
8222 appropriate AlgorithmIdentifiers for PSS based on those in the
8223 corresponding EVP_MD_CTX structure. No application support yet.
8227 * Support for companion algorithm specific ASN1 signing routines.
8228 New function ASN1_item_sign_ctx() signs a pre-initialised
8229 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8230 the appropriate parameters.
8234 * Add new algorithm specific ASN1 verification initialisation function
8235 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8236 handling will be the same no matter what EVP_PKEY_METHOD is used.
8237 Add a PSS handler to support verification of PSS signatures: checked
8238 against a number of sample certificates.
8242 * Add signature printing for PSS. Add PSS OIDs.
8244 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8246 * Add algorithm specific signature printing. An individual ASN1 method
8247 can now print out signatures instead of the standard hex dump.
8249 More complex signatures (e.g. PSS) can print out more meaningful
8250 information. Include DSA version that prints out the signature
8255 * Password based recipient info support for CMS library: implementing
8260 * Split password based encryption into PBES2 and PBKDF2 functions. This
8261 neatly separates the code into cipher and PBE sections and is required
8262 for some algorithms that split PBES2 into separate pieces (such as
8263 password based CMS).
8267 * Session-handling fixes:
8268 - Fix handling of connections that are resuming with a session ID,
8269 but also support Session Tickets.
8270 - Fix a bug that suppressed issuing of a new ticket if the client
8271 presented a ticket with an expired session.
8272 - Try to set the ticket lifetime hint to something reasonable.
8273 - Make tickets shorter by excluding irrelevant information.
8274 - On the client side, don't ignore renewed tickets.
8276 *Adam Langley, Bodo Moeller (Google)*
8278 * Fix PSK session representation.
8282 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8284 This work was sponsored by Intel.
8288 * Add GCM support to TLS library. Some custom code is needed to split
8289 the IV between the fixed (from PRF) and explicit (from TLS record)
8290 portions. This adds all GCM ciphersuites supported by RFC5288 and
8291 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8292 add a special AESGCM string for GCM only.
8296 * Expand range of ctrls for AES GCM. Permit setting invocation
8297 field on decrypt and retrieval of invocation field only on encrypt.
8301 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8302 As required by RFC5289 these ciphersuites cannot be used if for
8303 versions of TLS earlier than 1.2.
8307 * For FIPS capable OpenSSL interpret a NULL default public key method
8308 as unset and return the appropriate default but do *not* set the default.
8309 This means we can return the appropriate method in applications that
8310 switch between FIPS and non-FIPS modes.
8314 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8315 ENGINE is used then we cannot handle that in the FIPS module so we
8316 keep original code iff non-FIPS operations are allowed.
8320 * Add -attime option to openssl utilities.
8322 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8324 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8328 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8329 FIPS EC methods unconditionally for now.
8333 * New build option no-ec2m to disable characteristic 2 code.
8337 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8338 all cases can be covered as some introduce binary incompatibilities.
8342 * Redirect RSA operations to FIPS module including keygen,
8343 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8347 * Add similar low-level API blocking to ciphers.
8351 * low-level digest APIs are not approved in FIPS mode: any attempt
8352 to use these will cause a fatal error. Applications that *really* want
8353 to use them can use the `private_*` version instead.
8357 * Redirect cipher operations to FIPS module for FIPS builds.
8361 * Redirect digest operations to FIPS module for FIPS builds.
8365 * Update build system to add "fips" flag which will link in fipscanister.o
8366 for static and shared library builds embedding a signature if needed.
8370 * Output TLS supported curves in preference order instead of numerical
8371 order. This is currently hardcoded for the highest order curves first.
8372 This should be configurable so applications can judge speed vs strength.
8376 * Add TLS v1.2 server support for client authentication.
8380 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8385 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8386 FIPS modules versions.
8390 * Add TLS v1.2 client side support for client authentication. Keep cache
8391 of handshake records longer as we don't know the hash algorithm to use
8392 until after the certificate request message is received.
8396 * Initial TLS v1.2 client support. Add a default signature algorithms
8397 extension including all the algorithms we support. Parse new signature
8398 format in client key exchange. Relax some ECC signing restrictions for
8399 TLS v1.2 as indicated in RFC5246.
8403 * Add server support for TLS v1.2 signature algorithms extension. Switch
8404 to new signature format when needed using client digest preference.
8405 All server ciphersuites should now work correctly in TLS v1.2. No client
8406 support yet and no support for client certificates.
8410 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8411 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8412 ciphersuites. At present only RSA key exchange ciphersuites work with
8413 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8414 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8415 and version checking.
8419 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8420 with this defined it will not be affected by any changes to ssl internal
8421 structures. Add several utility functions to allow openssl application
8422 to work with OPENSSL_NO_SSL_INTERN defined.
8426 * A long standing patch to add support for SRP from EdelWeb (Peter
8427 Sylvester and Christophe Renou) was integrated.
8428 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8429 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8432 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8436 * Permit abbreviated handshakes when renegotiating using the function
8437 SSL_renegotiate_abbreviated().
8439 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8441 * Add call to ENGINE_register_all_complete() to
8442 ENGINE_load_builtin_engines(), so some implementations get used
8443 automatically instead of needing explicit application support.
8447 * Add support for TLS key exporter as described in RFC5705.
8449 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8451 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8452 a few changes are required:
8454 Add SSL_OP_NO_TLSv1_1 flag.
8455 Add TLSv1_1 methods.
8456 Update version checking logic to handle version 1.1.
8457 Add explicit IV handling (ported from DTLS code).
8458 Add command line options to s_client/s_server.
8465 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8467 * X509_ATTRIBUTE memory leak
8469 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8470 memory. This structure is used by the PKCS#7 and CMS routines so any
8471 application which reads PKCS#7 or CMS data from untrusted sources is
8472 affected. SSL/TLS is not affected.
8474 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8480 * Race condition handling PSK identify hint
8482 If PSK identity hints are received by a multi-threaded client then
8483 the values are wrongly updated in the parent SSL_CTX structure. This can
8484 result in a race condition potentially leading to a double free of the
8490 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8492 * Malformed ECParameters causes infinite loop
8494 When processing an ECParameters structure OpenSSL enters an infinite loop
8495 if the curve specified is over a specially malformed binary polynomial
8498 This can be used to perform denial of service against any
8499 system which processes public keys, certificate requests or
8500 certificates. This includes TLS clients and TLS servers with
8501 client authentication enabled.
8503 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8508 * Exploitable out-of-bounds read in X509_cmp_time
8510 X509_cmp_time does not properly check the length of the ASN1_TIME
8511 string and can read a few bytes out of bounds. In addition,
8512 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8515 An attacker can use this to craft malformed certificates and CRLs of
8516 various sizes and potentially cause a segmentation fault, resulting in
8517 a DoS on applications that verify certificates or CRLs. TLS clients
8518 that verify CRLs are affected. TLS clients and servers with client
8519 authentication enabled may be affected if they use custom verification
8522 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8523 independently by Hanno Böck.
8528 * PKCS7 crash with missing EnvelopedContent
8530 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8531 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8532 with missing content and trigger a NULL pointer dereference on parsing.
8534 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8535 structures from untrusted sources are affected. OpenSSL clients and
8536 servers are not affected.
8538 This issue was reported to OpenSSL by Michal Zalewski (Google).
8543 * CMS verify infinite loop with unknown hash function
8545 When verifying a signedData message the CMS code can enter an infinite loop
8546 if presented with an unknown hash function OID. This can be used to perform
8547 denial of service against any system which verifies signedData messages using
8549 This issue was reported to OpenSSL by Johannes Bauer.
8554 * Race condition handling NewSessionTicket
8556 If a NewSessionTicket is received by a multi-threaded client when attempting to
8557 reuse a previous ticket then a race condition can occur potentially leading to
8558 a double free of the ticket data.
8563 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8565 * Segmentation fault in ASN1_TYPE_cmp fix
8567 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8568 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8569 certificate signature algorithm consistency this can be used to crash any
8570 certificate verification operation and exploited in a DoS attack. Any
8571 application which performs certificate verification is vulnerable including
8572 OpenSSL clients and servers which enable client authentication.
8577 * ASN.1 structure reuse memory corruption fix
8579 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8580 memory corruption via an invalid write. Such reuse is and has been
8581 strongly discouraged and is believed to be rare.
8583 Applications that parse structures containing CHOICE or ANY DEFINED BY
8584 components may be affected. Certificate parsing (d2i_X509 and related
8585 functions) are however not affected. OpenSSL clients and servers are
8591 * PKCS7 NULL pointer dereferences fix
8593 The PKCS#7 parsing code does not handle missing outer ContentInfo
8594 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8595 missing content and trigger a NULL pointer dereference on parsing.
8597 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8598 otherwise parse PKCS#7 structures from untrusted sources are
8599 affected. OpenSSL clients and servers are not affected.
8601 This issue was reported to OpenSSL by Michal Zalewski (Google).
8606 * DoS via reachable assert in SSLv2 servers fix
8608 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8609 servers that both support SSLv2 and enable export cipher suites by sending
8610 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8612 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8613 (OpenSSL development team).
8618 * Use After Free following d2i_ECPrivatekey error fix
8620 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8621 could cause a use after free condition. This, in turn, could cause a double
8622 free in several private key parsing functions (such as d2i_PrivateKey
8623 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8624 for applications that receive EC private keys from untrusted
8625 sources. This scenario is considered rare.
8627 This issue was discovered by the BoringSSL project and fixed in their
8633 * X509_to_X509_REQ NULL pointer deref fix
8635 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8636 the certificate key is invalid. This function is rarely used in practice.
8638 This issue was discovered by Brian Carpenter.
8643 * Removed the export ciphers from the DEFAULT ciphers
8647 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8649 * Build fixes for the Windows and OpenVMS platforms
8651 *Matt Caswell and Richard Levitte*
8653 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8655 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8656 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8657 dereference. This could lead to a Denial Of Service attack. Thanks to
8658 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8663 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8664 dtls1_buffer_record function under certain conditions. In particular this
8665 could occur if an attacker sent repeated DTLS records with the same
8666 sequence number but for the next epoch. The memory leak could be exploited
8667 by an attacker in a Denial of Service attack through memory exhaustion.
8668 Thanks to Chris Mueller for reporting this issue.
8673 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8674 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8675 method would be set to NULL which could later result in a NULL pointer
8676 dereference. Thanks to Frank Schmirler for reporting this issue.
8681 * Abort handshake if server key exchange message is omitted for ephemeral
8684 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8685 reporting this issue.
8690 * Remove non-export ephemeral RSA code on client and server. This code
8691 violated the TLS standard by allowing the use of temporary RSA keys in
8692 non-export ciphersuites and could be used by a server to effectively
8693 downgrade the RSA key length used to a value smaller than the server
8694 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8695 INRIA or reporting this issue.
8700 * Fixed issue where DH client certificates are accepted without verification.
8701 An OpenSSL server will accept a DH certificate for client authentication
8702 without the certificate verify message. This effectively allows a client to
8703 authenticate without the use of a private key. This only affects servers
8704 which trust a client certificate authority which issues certificates
8705 containing DH keys: these are extremely rare and hardly ever encountered.
8706 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8712 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8713 results on some platforms, including x86_64. This bug occurs at random
8714 with a very low probability, and is not known to be exploitable in any
8715 way, though its exact impact is difficult to determine. Thanks to Pieter
8716 Wuille (Blockstream) who reported this issue and also suggested an initial
8717 fix. Further analysis was conducted by the OpenSSL development team and
8718 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8719 the OpenSSL core team.
8724 * Fix various certificate fingerprint issues.
8726 By using non-DER or invalid encodings outside the signed portion of a
8727 certificate the fingerprint can be changed without breaking the signature.
8728 Although no details of the signed portion of the certificate can be changed
8729 this can cause problems with some applications: e.g. those using the
8730 certificate fingerprint for blacklists.
8732 1. Reject signatures with non zero unused bits.
8734 If the BIT STRING containing the signature has non zero unused bits reject
8735 the signature. All current signature algorithms require zero unused bits.
8737 2. Check certificate algorithm consistency.
8739 Check the AlgorithmIdentifier inside TBS matches the one in the
8740 certificate signature. NB: this will result in signature failure
8741 errors for some broken certificates.
8743 Thanks to Konrad Kraszewski from Google for reporting this issue.
8745 3. Check DSA/ECDSA signatures use DER.
8747 Reencode DSA/ECDSA signatures and compare with the original received
8748 signature. Return an error if there is a mismatch.
8750 This will reject various cases including garbage after signature
8751 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8752 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8753 (negative or with leading zeroes).
8755 Further analysis was conducted and fixes were developed by Stephen Henson
8756 of the OpenSSL core team.
8762 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8764 * Session Ticket Memory Leak.
8766 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8767 integrity of that ticket is first verified. In the event of a session
8768 ticket integrity check failing, OpenSSL will fail to free memory
8769 causing a memory leak. By sending a large number of invalid session
8770 tickets an attacker could exploit this issue in a Denial Of Service
8776 * Build option no-ssl3 is incomplete.
8778 When OpenSSL is configured with "no-ssl3" as a build option, servers
8779 could accept and complete an SSL 3.0 handshake, and clients could be
8780 configured to send them.
8783 *Akamai and the OpenSSL team*
8785 * Add support for TLS_FALLBACK_SCSV.
8786 Client applications doing fallback retries should call
8787 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8790 *Adam Langley, Bodo Moeller*
8792 * Add additional DigestInfo checks.
8794 Reencode DigestInto in DER and check against the original when
8795 verifying RSA signature: this will reject any improperly encoded
8796 DigestInfo structures.
8798 Note: this is a precautionary measure and no attacks are currently known.
8802 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8804 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8805 to a denial of service attack. A malicious server can crash the client
8806 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8807 ciphersuite and sending carefully crafted handshake messages.
8809 Thanks to Felix Gröbert (Google) for discovering and researching this
8815 * By sending carefully crafted DTLS packets an attacker could cause openssl
8816 to leak memory. This can be exploited through a Denial of Service attack.
8817 Thanks to Adam Langley for discovering and researching this issue.
8822 * An attacker can force openssl to consume large amounts of memory whilst
8823 processing DTLS handshake messages. This can be exploited through a
8824 Denial of Service attack.
8825 Thanks to Adam Langley for discovering and researching this issue.
8830 * An attacker can force an error condition which causes openssl to crash
8831 whilst processing DTLS packets due to memory being freed twice. This
8832 can be exploited through a Denial of Service attack.
8833 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8839 * If a multithreaded client connects to a malicious server using a resumed
8840 session and the server sends an ec point format extension it could write
8841 up to 255 bytes to freed memory.
8843 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8849 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8850 X509_name_oneline, X509_name_print_ex et al. to leak some information
8851 from the stack. Applications may be affected if they echo pretty printing
8852 output to the attacker.
8854 Thanks to Ivan Fratric (Google) for discovering this issue.
8857 *Emilia Käsper, and Steve Henson*
8859 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8860 for corner cases. (Certain input points at infinity could lead to
8861 bogus results, with non-infinity inputs mapped to infinity too.)
8865 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8867 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8868 handshake can force the use of weak keying material in OpenSSL
8869 SSL/TLS clients and servers.
8871 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8872 researching this issue. ([CVE-2014-0224])
8874 *KIKUCHI Masashi, Steve Henson*
8876 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8877 OpenSSL DTLS client the code can be made to recurse eventually crashing
8880 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8883 *Imre Rad, Steve Henson*
8885 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8886 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8887 client or server. This is potentially exploitable to run arbitrary
8888 code on a vulnerable client or server.
8890 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8892 *Jüri Aedla, Steve Henson*
8894 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8895 are subject to a denial of service attack.
8897 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8898 this issue. ([CVE-2014-3470])
8900 *Felix Gröbert, Ivan Fratric, Steve Henson*
8902 * Harmonize version and its documentation. -f flag is used to display
8905 *mancha <mancha1@zoho.com>*
8907 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8908 in i2d_ECPrivateKey.
8910 *mancha <mancha1@zoho.com>*
8912 * Fix some double frees. These are not thought to be exploitable.
8914 *mancha <mancha1@zoho.com>*
8916 * Fix for the attack described in the paper "Recovering OpenSSL
8917 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8918 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8919 <http://eprint.iacr.org/2014/140>
8921 Thanks to Yuval Yarom and Naomi Benger for discovering this
8922 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8924 *Yuval Yarom and Naomi Benger*
8926 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8928 * Keep original DTLS digest and encryption contexts in retransmission
8929 structures so we can use the previous session parameters if they need
8930 to be resent. ([CVE-2013-6450])
8934 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8935 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8936 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8937 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8938 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8939 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8941 *Rob Stradling, Adam Langley*
8943 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8945 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8947 This addresses the flaw in CBC record processing discovered by
8948 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8949 at: <http://www.isg.rhul.ac.uk/tls/>
8951 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8952 Security Group at Royal Holloway, University of London
8953 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8954 Emilia Käsper for the initial patch.
8957 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8959 * Return an error when checking OCSP signatures when key is NULL.
8960 This fixes a DoS attack. ([CVE-2013-0166])
8964 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8965 the right response is stapled. Also change SSL_get_certificate()
8966 so it returns the certificate actually sent.
8967 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8968 (This is a backport)
8970 *Rob Stradling <rob.stradling@comodo.com>*
8972 * Fix possible deadlock when decoding public keys.
8976 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8978 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8981 * Sanity check record length before skipping explicit IV in DTLS
8984 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8985 fuzzing as a service testing platform.
8990 * Initialise tkeylen properly when encrypting CMS messages.
8991 Thanks to Solar Designer of Openwall for reporting this issue.
8995 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8997 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8998 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8999 in CRYPTO_realloc_clean.
9001 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9002 issue and to Adam Langley <agl@chromium.org> for fixing it.
9005 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9007 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9009 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9010 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9011 content decryption and always return the same error. Note: this attack
9012 needs on average 2^20 messages so it only affects automated senders. The
9013 old behaviour can be re-enabled in the CMS code by setting the
9014 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9015 an MMA defence is not necessary.
9016 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9017 this issue. ([CVE-2012-0884])
9021 * Fix CVE-2011-4619: make sure we really are receiving a
9022 client hello before rejecting multiple SGC restarts. Thanks to
9023 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9027 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9029 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9030 Thanks to Antonio Martin, Enterprise Secure Access Research and
9031 Development, Cisco Systems, Inc. for discovering this bug and
9032 preparing a fix. ([CVE-2012-0050])
9036 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9038 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9039 of the Vaudenay padding oracle attack on CBC mode encryption
9040 which enables an efficient plaintext recovery attack against
9041 the OpenSSL implementation of DTLS. Their attack exploits timing
9042 differences arising during decryption processing. A research
9043 paper describing this attack can be found at:
9044 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9045 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9046 Security Group at Royal Holloway, University of London
9047 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9048 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9049 for preparing the fix. ([CVE-2011-4108])
9051 *Robin Seggelmann, Michael Tuexen*
9053 * Clear bytes used for block padding of SSL 3.0 records.
9056 *Adam Langley (Google)*
9058 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9059 Kadianakis <desnacked@gmail.com> for discovering this issue and
9060 Adam Langley for preparing the fix. ([CVE-2011-4619])
9062 *Adam Langley (Google)*
9064 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9066 *Andrey Kulikov <amdeich@gmail.com>*
9068 * Prevent malformed RFC3779 data triggering an assertion failure.
9069 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9070 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9072 *Rob Austein <sra@hactrn.net>*
9074 * Improved PRNG seeding for VOS.
9076 *Paul Green <Paul.Green@stratus.com>*
9078 * Fix ssl_ciph.c set-up race.
9080 *Adam Langley (Google)*
9082 * Fix spurious failures in ecdsatest.c.
9084 *Emilia Käsper (Google)*
9086 * Fix the BIO_f_buffer() implementation (which was mixing different
9087 interpretations of the `..._len` fields).
9089 *Adam Langley (Google)*
9091 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9092 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9093 threads won't reuse the same blinding coefficients.
9095 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9096 lock to call BN_BLINDING_invert_ex, and avoids one use of
9097 BN_BLINDING_update for each BN_BLINDING structure (previously,
9098 the last update always remained unused).
9100 *Emilia Käsper (Google)*
9102 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9104 *Bob Buckholz (Google)*
9106 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9108 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9109 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9111 *Kaspar Brand <ossl@velox.ch>*
9113 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9114 for multi-threaded use of ECDH. ([CVE-2011-3210])
9116 *Adam Langley (Google)*
9118 * Fix x509_name_ex_d2i memory leak on bad inputs.
9122 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9123 signature public key algorithm by using OID xref utilities instead.
9124 Before this you could only use some ECC ciphersuites with SHA1 only.
9128 * Add protection against ECDSA timing attacks as mentioned in the paper
9129 by Billy Bob Brumley and Nicola Tuveri, see:
9130 <http://eprint.iacr.org/2011/232.pdf>
9132 *Billy Bob Brumley and Nicola Tuveri*
9134 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9136 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9138 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9140 * Fix bug in string printing code: if *any* escaping is enabled we must
9141 escape the escape character (backslash) or the resulting string is
9146 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9148 * Disable code workaround for ancient and obsolete Netscape browsers
9149 and servers: an attacker can use it in a ciphersuite downgrade attack.
9150 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9154 * Fixed J-PAKE implementation error, originally discovered by
9155 Sebastien Martini, further info and confirmation from Stefan
9156 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9160 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9162 * Fix extension code to avoid race conditions which can result in a buffer
9163 overrun vulnerability: resumed sessions must not be modified as they can
9164 be shared by multiple threads. CVE-2010-3864
9168 * Fix WIN32 build system to correctly link an ENGINE directory into
9173 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9175 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9178 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9180 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9182 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9183 context. The operation can be customised via the ctrl mechanism in
9184 case ENGINEs want to include additional functionality.
9188 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9192 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9193 output hashes compatible with older versions of OpenSSL.
9195 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9197 * Fix compression algorithm handling: if resuming a session use the
9198 compression algorithm of the resumed session instead of determining
9199 it from client hello again. Don't allow server to change algorithm.
9203 * Add load_crls() function to commands tidying load_certs() too. Add option
9204 to verify utility to allow additional CRLs to be included.
9208 * Update OCSP request code to permit adding custom headers to the request:
9209 some responders need this.
9213 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9216 *Julia Lawall <julia@diku.dk>*
9218 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9219 needlessly dereferenced structures, used obsolete functions and
9220 didn't handle all updated verify codes correctly.
9224 * Disable MD2 in the default configuration.
9228 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9229 indicate the initial BIO being pushed or popped. This makes it possible
9230 to determine whether the BIO is the one explicitly called or as a result
9231 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9232 it handles reference counts correctly and doesn't zero out the I/O bio
9233 when it is not being explicitly popped. WARNING: applications which
9234 included workarounds for the old buggy behaviour will need to be modified
9235 or they could free up already freed BIOs.
9239 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9240 renaming to all platforms (within the 0.9.8 branch, this was
9241 done conditionally on Netware platforms to avoid a name clash).
9243 *Guenter <lists@gknw.net>*
9245 * Add ECDHE and PSK support to DTLS.
9247 *Michael Tuexen <tuexen@fh-muenster.de>*
9249 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9254 * Add "missing" function EVP_MD_flags() (without this the only way to
9255 retrieve a digest flags is by accessing the structure directly. Update
9256 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9257 or cipher is registered as in the "from" argument. Print out all
9258 registered digests in the dgst usage message instead of manually
9259 attempting to work them out.
9263 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9264 this allows the use of compression and extensions. Change default cipher
9265 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9266 by default unless an application cipher string requests it.
9270 * Alter match criteria in PKCS12_parse(). It used to try to use local
9271 key ids to find matching certificates and keys but some PKCS#12 files
9272 don't follow the (somewhat unwritten) rules and this strategy fails.
9273 Now just gather all certificates together and the first private key
9274 then look for the first certificate that matches the key.
9278 * Support use of registered digest and cipher names for dgst and cipher
9279 commands instead of having to add each one as a special case. So now
9286 openssl dgst -sha256 foo
9288 and this works for ENGINE based algorithms too.
9292 * Update Gost ENGINE to support parameter files.
9294 *Victor B. Wagner <vitus@cryptocom.ru>*
9296 * Support GeneralizedTime in ca utility.
9298 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9300 * Enhance the hash format used for certificate directory links. The new
9301 form uses the canonical encoding (meaning equivalent names will work
9302 even if they aren't identical) and uses SHA1 instead of MD5. This form
9303 is incompatible with the older format and as a result c_rehash should
9304 be used to rebuild symbolic links.
9308 * Make PKCS#8 the default write format for private keys, replacing the
9309 traditional format. This form is standardised, more secure and doesn't
9310 include an implicit MD5 dependency.
9314 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9315 committed to OpenSSL should pass this lot as a minimum.
9319 * Add session ticket override functionality for use by EAP-FAST.
9321 *Jouni Malinen <j@w1.fi>*
9323 * Modify HMAC functions to return a value. Since these can be implemented
9324 in an ENGINE errors can occur.
9328 * Type-checked OBJ_bsearch_ex.
9332 * Type-checked OBJ_bsearch. Also some constification necessitated
9333 by type-checking. Still to come: TXT_DB, bsearch(?),
9334 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9339 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9340 seconds to a tm structure directly, instead of going through OS
9341 specific date routines. This avoids any issues with OS routines such
9342 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9343 and X509_time_adj_ex() to cover the extended range. The existing
9344 X509_time_adj() is still usable and will no longer have any date issues.
9348 * Delta CRL support. New use deltas option which will attempt to locate
9349 and search any appropriate delta CRLs available.
9351 This work was sponsored by Google.
9355 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9356 code and add additional score elements. Validate alternate CRL paths
9357 as part of the CRL checking and indicate a new error "CRL path validation
9358 error" in this case. Applications wanting additional details can use
9359 the verify callback and check the new "parent" field. If this is not
9360 NULL CRL path validation is taking place. Existing applications won't
9361 see this because it requires extended CRL support which is off by
9364 This work was sponsored by Google.
9368 * Support for freshest CRL extension.
9370 This work was sponsored by Google.
9374 * Initial indirect CRL support. Currently only supported in the CRLs
9375 passed directly and not via lookup. Process certificate issuer
9376 CRL entry extension and lookup CRL entries by bother issuer name
9377 and serial number. Check and process CRL issuer entry in IDP extension.
9379 This work was sponsored by Google.
9383 * Add support for distinct certificate and CRL paths. The CRL issuer
9384 certificate is validated separately in this case. Only enabled if
9385 an extended CRL support flag is set: this flag will enable additional
9386 CRL functionality in future.
9388 This work was sponsored by Google.
9392 * Add support for policy mappings extension.
9394 This work was sponsored by Google.
9398 * Fixes to pathlength constraint, self issued certificate handling,
9399 policy processing to align with RFC3280 and PKITS tests.
9401 This work was sponsored by Google.
9405 * Support for name constraints certificate extension. DN, email, DNS
9406 and URI types are currently supported.
9408 This work was sponsored by Google.
9412 * To cater for systems that provide a pointer-based thread ID rather
9413 than numeric, deprecate the current numeric thread ID mechanism and
9414 replace it with a structure and associated callback type. This
9415 mechanism allows a numeric "hash" to be extracted from a thread ID in
9416 either case, and on platforms where pointers are larger than 'long',
9417 mixing is done to help ensure the numeric 'hash' is usable even if it
9418 can't be guaranteed unique. The default mechanism is to use "&errno"
9419 as a pointer-based thread ID to distinguish between threads.
9421 Applications that want to provide their own thread IDs should now use
9422 CRYPTO_THREADID_set_callback() to register a callback that will call
9423 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9425 Note that ERR_remove_state() is now deprecated, because it is tied
9426 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9427 to free the current thread's error state should be replaced by
9428 ERR_remove_thread_state(NULL).
9430 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9431 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9432 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9433 application was previously providing a numeric thread callback that
9434 was inappropriate for distinguishing threads, then uniqueness might
9435 have been obtained with &errno that happened immediately in the
9436 intermediate development versions of OpenSSL; this is no longer the
9437 case, the numeric thread callback will now override the automatic use
9440 *Geoff Thorpe, with help from Bodo Moeller*
9442 * Initial support for different CRL issuing certificates. This covers a
9443 simple case where the self issued certificates in the chain exist and
9444 the real CRL issuer is higher in the existing chain.
9446 This work was sponsored by Google.
9450 * Removed effectively defunct crypto/store from the build.
9454 * Revamp of STACK to provide stronger type-checking. Still to come:
9455 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9456 ASN1_STRING, CONF_VALUE.
9460 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9461 RAM on SSL connections. This option can save about 34k per idle SSL.
9465 * Revamp of LHASH to provide stronger type-checking. Still to come:
9466 STACK, TXT_DB, bsearch, qsort.
9470 * Initial support for Cryptographic Message Syntax (aka CMS) based
9471 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9472 support for data, signedData, compressedData, digestedData and
9473 encryptedData, envelopedData types included. Scripts to check against
9474 RFC4134 examples draft and interop and consistency checks of many
9475 content types and variants.
9479 * Add options to enc utility to support use of zlib compression BIO.
9483 * Extend mk1mf to support importing of options and assembly language
9484 files from Configure script, currently only included in VC-WIN32.
9485 The assembly language rules can now optionally generate the source
9486 files from the associated perl scripts.
9490 * Implement remaining functionality needed to support GOST ciphersuites.
9491 Interop testing has been performed using CryptoPro implementations.
9493 *Victor B. Wagner <vitus@cryptocom.ru>*
9495 * s390x assembler pack.
9499 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9504 * Implement Opaque PRF Input TLS extension as specified in
9505 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9506 official specification yet and no extension type assignment by
9507 IANA exists, this extension (for now) will have to be explicitly
9508 enabled when building OpenSSL by providing the extension number
9509 to use. For example, specify an option
9511 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9513 to the "config" or "Configure" script to enable the extension,
9514 assuming extension number 0x9527 (which is a completely arbitrary
9515 and unofficial assignment based on the MD5 hash of the Internet
9516 Draft). Note that by doing so, you potentially lose
9517 interoperability with other TLS implementations since these might
9518 be using the same extension number for other purposes.
9520 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9521 opaque PRF input value to use in the handshake. This will create
9522 an internal copy of the length-'len' string at 'src', and will
9523 return non-zero for success.
9525 To get more control and flexibility, provide a callback function
9528 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9529 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9533 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9536 Callback function 'cb' will be called in handshakes, and is
9537 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9538 Argument 'arg' is for application purposes (the value as given to
9539 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9540 be provided to the callback function). The callback function
9541 has to return non-zero to report success: usually 1 to use opaque
9542 PRF input just if possible, or 2 to enforce use of the opaque PRF
9543 input. In the latter case, the library will abort the handshake
9544 if opaque PRF input is not successfully negotiated.
9546 Arguments 'peerinput' and 'len' given to the callback function
9547 will always be NULL and 0 in the case of a client. A server will
9548 see the client's opaque PRF input through these variables if
9549 available (NULL and 0 otherwise). Note that if the server
9550 provides an opaque PRF input, the length must be the same as the
9551 length of the client's opaque PRF input.
9553 Note that the callback function will only be called when creating
9554 a new session (session resumption can resume whatever was
9555 previously negotiated), and will not be called in SSL 2.0
9556 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9557 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9558 for applications that need to enforce opaque PRF input.
9562 * Update ssl code to support digests other than SHA1+MD5 for handshake
9565 *Victor B. Wagner <vitus@cryptocom.ru>*
9567 * Add RFC4507 support to OpenSSL. This includes the corrections in
9568 RFC4507bis. The encrypted ticket format is an encrypted encoded
9569 SSL_SESSION structure, that way new session features are automatically
9572 If a client application caches session in an SSL_SESSION structure
9573 support is transparent because tickets are now stored in the encoded
9576 The SSL_CTX structure automatically generates keys for ticket
9577 protection in servers so again support should be possible
9578 with no application modification.
9580 If a client or server wishes to disable RFC4507 support then the option
9581 SSL_OP_NO_TICKET can be set.
9583 Add a TLS extension debugging callback to allow the contents of any client
9584 or server extensions to be examined.
9586 This work was sponsored by Google.
9590 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9591 OpenSSL should now compile cleanly on gcc 4.2
9593 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9595 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9596 support including streaming MAC support: this is required for GOST
9597 ciphersuite support.
9599 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9601 * Add option -stream to use PKCS#7 streaming in smime utility. New
9602 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9603 to output in BER and PEM format.
9607 * Experimental support for use of HMAC via EVP_PKEY interface. This
9608 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9609 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9610 ENGINE support for HMAC keys which are unextractable. New -mac and
9611 -macopt options to dgst utility.
9615 * New option -sigopt to dgst utility. Update dgst to use
9616 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9617 alternative signing parameters such as X9.31 or PSS in the dgst
9622 * Change ssl_cipher_apply_rule(), the internal function that does
9623 the work each time a ciphersuite string requests enabling
9624 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9625 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9626 the order of disabled ciphersuites such that those ciphersuites
9627 that most recently went from enabled to disabled not only stay
9628 in order with respect to each other, but also have higher priority
9629 than other disabled ciphersuites the next time ciphersuites are
9632 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9633 the same ciphersuites as with "HIGH" alone, but in a specific
9634 order where the PSK ciphersuites come first (since they are the
9635 most recently disabled ciphersuites when "HIGH" is parsed).
9637 Also, change ssl_create_cipher_list() (using this new
9638 functionality) such that between otherwise identical
9639 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9644 * Change ssl_create_cipher_list() so that it automatically
9645 arranges the ciphersuites in reasonable order before starting
9646 to process the rule string. Thus, the definition for "DEFAULT"
9647 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9648 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9649 This makes it much easier to arrive at a reasonable default order
9650 in applications for which anonymous ciphers are OK (meaning
9651 that you can't actually use DEFAULT).
9653 *Bodo Moeller; suggested by Victor Duchovni*
9655 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9656 processing) into multiple integers instead of setting
9657 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9658 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9659 (These masks as well as the individual bit definitions are hidden
9660 away into the non-exported interface ssl/ssl_locl.h, so this
9661 change to the definition of the SSL_CIPHER structure shouldn't
9662 affect applications.) This give us more bits for each of these
9663 categories, so there is no longer a need to coagulate AES128 and
9664 AES256 into a single algorithm bit, and to coagulate Camellia128
9665 and Camellia256 into a single algorithm bit, which has led to all
9668 Thus, among other things, the kludge introduced in 0.9.7m and
9669 0.9.8e for masking out AES256 independently of AES128 or masking
9670 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9672 With the change, we also introduce new ciphersuite aliases that
9673 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9678 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9679 Use the leftmost N bytes of the signature input if the input is
9680 larger than the prime q (with N being the size in bytes of q).
9684 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9685 it yet and it is largely untested.
9689 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9693 * Initial incomplete changes to avoid need for function casts in OpenSSL
9694 some compilers (gcc 4.2 and later) reject their use. Safestack is
9695 reimplemented. Update ASN1 to avoid use of legacy functions.
9699 * Win32/64 targets are linked with Winsock2.
9703 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9704 to external functions. This can be used to increase CRL handling
9705 efficiency especially when CRLs are very large by (for example) storing
9706 the CRL revoked certificates in a database.
9710 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9711 new CRLs added to a directory can be used. New command line option
9712 -verify_return_error to s_client and s_server. This causes real errors
9713 to be returned by the verify callback instead of carrying on no matter
9714 what. This reflects the way a "real world" verify callback would behave.
9718 * GOST engine, supporting several GOST algorithms and public key formats.
9719 Kindly donated by Cryptocom.
9723 * Partial support for Issuing Distribution Point CRL extension. CRLs
9724 partitioned by DP are handled but no indirect CRL or reason partitioning
9725 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9726 selected via a scoring technique which handles IDP and AKID in CRLs.
9730 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9731 will ultimately be used for all verify operations: this will remove the
9732 X509_STORE dependency on certificate verification and allow alternative
9733 lookup methods. X509_STORE based implementations of these two callbacks.
9737 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9738 Modify get_crl() to find a valid (unexpired) CRL if possible.
9742 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9743 this would be called X509_CRL_cmp() but that name is already used by
9744 a function that just compares CRL issuer names. Cache several CRL
9745 extensions in X509_CRL structure and cache CRLDP in X509.
9749 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9750 this maps equivalent X509_NAME structures into a consistent structure.
9751 Name comparison can then be performed rapidly using memcmp().
9755 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9760 * Allow digests to supply their own micalg string for S/MIME type using
9761 the ctrl EVP_MD_CTRL_MICALG.
9765 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9766 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9767 ctrl. It can then customise the structure before and/or after signing
9772 * New function OBJ_add_sigid() to allow application defined signature OIDs
9773 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9774 to free up any added signature OIDs.
9778 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9779 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9780 digest and cipher tables. New options added to openssl utility:
9781 list-message-digest-algorithms and list-cipher-algorithms.
9785 * Change the array representation of binary polynomials: the list
9786 of degrees of non-zero coefficients is now terminated with -1.
9787 Previously it was terminated with 0, which was also part of the
9788 value; thus, the array representation was not applicable to
9789 polynomials where t^0 has coefficient zero. This change makes
9790 the array representation useful in a more general context.
9794 * Various modifications and fixes to SSL/TLS cipher string
9795 handling. For ECC, the code now distinguishes between fixed ECDH
9796 with RSA certificates on the one hand and with ECDSA certificates
9797 on the other hand, since these are separate ciphersuites. The
9798 unused code for Fortezza ciphersuites has been removed.
9800 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9801 (not "ECDHE"). For consistency with the code for DH
9802 certificates, use of ECDH certificates is now considered ECDH
9803 authentication, not RSA or ECDSA authentication (the latter is
9804 merely the CA's signing algorithm and not actively used in the
9807 The temporary ciphersuite alias "ECCdraft" is no longer
9808 available, and ECC ciphersuites are no longer excluded from "ALL"
9809 and "DEFAULT". The following aliases now exist for RFC 4492
9810 ciphersuites, most of these by analogy with the DH case:
9812 kECDHr - ECDH cert, signed with RSA
9813 kECDHe - ECDH cert, signed with ECDSA
9814 kECDH - ECDH cert (signed with either RSA or ECDSA)
9815 kEECDH - ephemeral ECDH
9816 ECDH - ECDH cert or ephemeral ECDH
9822 AECDH - anonymous ECDH
9823 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9827 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9828 Use correct micalg parameters depending on digest(s) in signed message.
9832 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9833 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9837 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9838 an engine to register a method. Add ENGINE lookups for methods and
9839 functional reference processing.
9843 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9844 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9849 * New -resign option to smime utility. This adds one or more signers
9850 to an existing PKCS#7 signedData structure. Also -md option to use an
9851 alternative message digest algorithm for signing.
9855 * Tidy up PKCS#7 routines and add new functions to make it easier to
9856 create PKCS7 structures containing multiple signers. Update smime
9857 application to support multiple signers.
9861 * New -macalg option to pkcs12 utility to allow setting of an alternative
9866 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9867 Reorganize PBE internals to lookup from a static table using NIDs,
9868 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9869 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9870 PRF which will be automatically used with PBES2.
9874 * Replace the algorithm specific calls to generate keys in "req" with the
9879 * Update PKCS#7 enveloped data routines to use new API. This is now
9880 supported by any public key method supporting the encrypt operation. A
9881 ctrl is added to allow the public key algorithm to examine or modify
9882 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9887 * Add a ctrl to asn1 method to allow a public key algorithm to express
9888 a default digest type to use. In most cases this will be SHA1 but some
9889 algorithms (such as GOST) need to specify an alternative digest. The
9890 return value indicates how strong the preference is 1 means optional and
9891 2 is mandatory (that is it is the only supported type). Modify
9892 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9893 use the default md. Update openssl utilities to use the default digest
9894 type for signing if it is not explicitly indicated.
9898 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9899 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9900 signing method from the key type. This effectively removes the link
9901 between digests and public key types.
9905 * Add an OID cross reference table and utility functions. Its purpose is to
9906 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9907 rsaEncryption. This will allow some of the algorithm specific hackery
9908 needed to use the correct OID to be removed.
9912 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9913 structures for PKCS7_sign(). They are now set up by the relevant public
9918 * Add provisional EC pkey method with support for ECDSA and ECDH.
9922 * Add support for key derivation (agreement) in the API, DH method and
9927 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9928 public and private key formats. As a side effect these add additional
9929 command line functionality not previously available: DSA signatures can be
9930 generated and verified using pkeyutl and DH key support and generation in
9937 *Oliver Tappe <zooey@hirschkaefer.de>*
9939 * New make target "install_html_docs" installs HTML renditions of the
9942 *Oliver Tappe <zooey@hirschkaefer.de>*
9944 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9945 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9946 support key and parameter generation and add initial key generation
9947 functionality for RSA.
9951 * Add functions for main EVP_PKEY_method operations. The undocumented
9952 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9953 `EVP_PKEY_{encrypt,decrypt}_old`.
9957 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9958 key API, doesn't do much yet.
9962 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9963 public key algorithms. New option to openssl utility:
9964 "list-public-key-algorithms" to print out info.
9968 * Implement the Supported Elliptic Curves Extension for
9969 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9973 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9974 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9978 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9979 utilities such as rsa, dsa, dsaparam etc except they process any key
9984 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9985 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9986 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9991 * Initial support for pluggable public key ASN1.
9992 De-spaghettify the public key ASN1 handling. Move public and private
9993 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9994 algorithm specific handling to a single module within the relevant
9995 algorithm directory. Add functions to allow (near) opaque processing
9996 of public and private key structures.
10000 * Implement the Supported Point Formats Extension for
10001 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10005 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10006 for the psk identity [hint] and the psk callback functions to the
10007 SSL_SESSION, SSL and SSL_CTX structure.
10010 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10014 SSL_CTX_use_psk_identity_hint
10015 SSL_get_psk_identity_hint
10016 SSL_get_psk_identity
10017 SSL_use_psk_identity_hint
10019 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10021 * Add RFC 3161 compliant time stamp request creation, response generation
10022 and response verification functionality.
10024 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10026 * Add initial support for TLS extensions, specifically for the server_name
10027 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10028 have new members for a hostname. The SSL data structure has an
10029 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10030 stored in that context to allow for session resumption, even after the
10031 SSL has been switched to a new SSL_CTX in reaction to a client's
10032 server_name extension.
10034 New functions (subject to change):
10036 SSL_get_servername()
10037 SSL_get_servername_type()
10040 New CTRL codes and macros (subject to change):
10042 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10043 - SSL_CTX_set_tlsext_servername_callback()
10044 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10045 - SSL_CTX_set_tlsext_servername_arg()
10046 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10048 openssl s_client has a new '-servername ...' option.
10050 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10051 '-key2 ...', '-servername_fatal' (subject to change). This allows
10052 testing the HostName extension for a specific single hostname ('-cert'
10053 and '-key' remain fallbacks for handshakes without HostName
10054 negotiation). If the unrecognized_name alert has to be sent, this by
10055 default is a warning; it becomes fatal with the '-servername_fatal'
10058 *Peter Sylvester, Remy Allais, Christophe Renou*
10060 * Whirlpool hash implementation is added.
10064 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10065 bn(64,32). Because of instruction set limitations it doesn't have
10066 any negative impact on performance. This was done mostly in order
10067 to make it possible to share assembler modules, such as bn_mul_mont
10068 implementations, between 32- and 64-bit builds without hassle.
10072 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10073 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10078 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10079 dedicated Montgomery multiplication procedure, is introduced.
10080 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10081 "64-bit" performance on certain 32-bit targets.
10085 * New option SSL_OP_NO_COMP to disable use of compression selectively
10086 in SSL structures. New SSL ctrl to set maximum send fragment size.
10087 Save memory by setting the I/O buffer sizes dynamically instead of
10088 using the maximum available value.
10092 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10093 in addition to the text details.
10097 * Very, very preliminary EXPERIMENTAL support for printing of general
10098 ASN1 structures. This currently produces rather ugly output and doesn't
10099 handle several customised structures at all.
10103 * Integrated support for PVK file format and some related formats such
10104 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10105 these in the 'rsa' and 'dsa' utilities.
10109 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10113 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10114 place for the (very old) "NETSCAPE" format certificates which are now
10115 handled using new ASN1 code equivalents.
10119 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10120 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10121 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10125 * Modify CRL distribution points extension code to print out previously
10126 unsupported fields. Enhance extension setting code to allow setting of
10131 * Add print and set support for Issuing Distribution Point CRL extension.
10135 * Change 'Configure' script to enable Camellia by default.
10142 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10144 * When rejecting SSL/TLS records due to an incorrect version number, never
10145 update s->server with a new major version number. As of
10146 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10147 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10148 the previous behavior could result in a read attempt at NULL when
10149 receiving specific incorrect SSL/TLS records once record payload
10150 protection is active. ([CVE-2010-0740])
10152 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10154 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10155 could be crashed if the relevant tables were not present (e.g. chrooted).
10157 *Tomas Hoger <thoger@redhat.com>*
10159 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10161 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10163 *Martin Olsson, Neel Mehta*
10165 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10166 accommodate for stack sorting, always a write lock!).
10170 * On some versions of WIN32 Heap32Next is very slow. This can cause
10171 excessive delays in the RAND_poll(): over a minute. As a workaround
10172 include a time check in the inner Heap32Next loop too.
10176 * The code that handled flushing of data in SSL/TLS originally used the
10177 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10178 the problem outlined in PR#1949. The fix suggested there however can
10179 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10180 of Apache). So instead simplify the code to flush unconditionally.
10181 This should be fine since flushing with no data to flush is a no op.
10185 * Handle TLS versions 2.0 and later properly and correctly use the
10186 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10187 off ancient servers have a habit of sticking around for a while...
10191 * Modify compression code so it frees up structures without using the
10192 ex_data callbacks. This works around a problem where some applications
10193 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10194 restarting) then use compression (e.g. SSL with compression) later.
10195 This results in significant per-connection memory leaks and
10196 has caused some security issues including CVE-2008-1678 and
10201 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10202 change when encrypting or decrypting.
10206 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10207 connect and renegotiate with servers which do not support RI.
10208 Until RI is more widely deployed this option is enabled by default.
10212 * Add "missing" ssl ctrls to clear options and mode.
10216 * If client attempts to renegotiate and doesn't support RI respond with
10217 a no_renegotiation alert as required by RFC5746. Some renegotiating
10218 TLS clients will continue a connection gracefully when they receive
10219 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10220 waiting for a server hello which it will never receive. Now we treat a
10221 received no_renegotiation alert as a fatal error. This is because
10222 applications requesting a renegotiation might well expect it to succeed
10223 and would have no code in place to handle the server denying it so the
10224 only safe thing to do is to terminate the connection.
10228 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10229 peer supports secure renegotiation and 0 otherwise. Print out peer
10230 renegotiation support in s_client/s_server.
10234 * Replace the highly broken and deprecated SPKAC certification method with
10235 the updated NID creation version. This should correctly handle UTF8.
10239 * Implement RFC5746. Re-enable renegotiation but require the extension
10240 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10241 turns out to be a bad idea. It has been replaced by
10242 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10243 SSL_CTX_set_options(). This is really not recommended unless you
10244 know what you are doing.
10246 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10248 * Fixes to stateless session resumption handling. Use initial_ctx when
10249 issuing and attempting to decrypt tickets in case it has changed during
10250 servername handling. Use a non-zero length session ID when attempting
10251 stateless session resumption: this makes it possible to determine if
10252 a resumption has occurred immediately after receiving server hello
10253 (several places in OpenSSL subtly assume this) instead of later in
10258 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10259 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10260 fixes for a few places where the return code is not checked
10263 *Julia Lawall <julia@diku.dk>*
10265 * Add --strict-warnings option to Configure script to include devteam
10266 warnings in other configurations.
10270 * Add support for --libdir option and LIBDIR variable in makefiles. This
10271 makes it possible to install openssl libraries in locations which
10272 have names other than "lib", for example "/usr/lib64" which some
10275 *Steve Henson, based on patch from Jeremy Utley*
10277 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10278 X690 8.9.12 and can produce some misleading textual output of OIDs.
10280 *Steve Henson, reported by Dan Kaminsky*
10282 * Delete MD2 from algorithm tables. This follows the recommendation in
10283 several standards that it is not used in new applications due to
10284 several cryptographic weaknesses. For binary compatibility reasons
10285 the MD2 API is still compiled in by default.
10289 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10294 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10295 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10298 *Guenter <lists@gknw.net>*
10300 * Fix the server certificate chain building code to use X509_verify_cert(),
10301 it used to have an ad-hoc builder which was unable to cope with anything
10302 other than a simple chain.
10304 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10306 * Don't check self signed certificate signatures in X509_verify_cert()
10307 by default (a flag can override this): it just wastes time without
10308 adding any security. As a useful side effect self signed root CAs
10309 with non-FIPS digests are now usable in FIPS mode.
10313 * In dtls1_process_out_of_seq_message() the check if the current message
10314 is already buffered was missing. For every new message was memory
10315 allocated, allowing an attacker to perform an denial of service attack
10316 with sending out of seq handshake messages until there is no memory
10317 left. Additionally every future message was buffered, even if the
10318 sequence number made no sense and would be part of another handshake.
10319 So only messages with sequence numbers less than 10 in advance will be
10320 buffered. ([CVE-2009-1378])
10322 *Robin Seggelmann, discovered by Daniel Mentz*
10324 * Records are buffered if they arrive with a future epoch to be
10325 processed after finishing the corresponding handshake. There is
10326 currently no limitation to this buffer allowing an attacker to perform
10327 a DOS attack with sending records with future epochs until there is no
10328 memory left. This patch adds the pqueue_size() function to determine
10329 the size of a buffer and limits the record buffer to 100 entries.
10332 *Robin Seggelmann, discovered by Daniel Mentz*
10334 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10335 parent structure is freed. ([CVE-2009-1379])
10339 * Handle non-blocking I/O properly in SSL_shutdown() call.
10341 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10343 * Add `2.5.4.*` OIDs
10345 *Ilya O. <vrghost@gmail.com>*
10347 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10349 * Disable renegotiation completely - this fixes a severe security
10350 problem ([CVE-2009-3555]) at the cost of breaking all
10351 renegotiation. Renegotiation can be re-enabled by setting
10352 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10353 run-time. This is really not recommended unless you know what
10358 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10360 * Don't set val to NULL when freeing up structures, it is freed up by
10361 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10362 zeroing past the valid field. ([CVE-2009-0789])
10364 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10366 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10367 checked correctly. This would allow some invalid signed attributes to
10368 appear to verify correctly. ([CVE-2009-0591])
10370 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10372 * Reject UniversalString and BMPString types with invalid lengths. This
10373 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10374 a legal length. ([CVE-2009-0590])
10378 * Set S/MIME signing as the default purpose rather than setting it
10379 unconditionally. This allows applications to override it at the store
10384 * Permit restricted recursion of ASN1 strings. This is needed in practice
10385 to handle some structures.
10389 * Improve efficiency of mem_gets: don't search whole buffer each time
10392 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10394 * New -hex option for openssl rand.
10398 * Print out UTF8String and NumericString when parsing ASN1.
10402 * Support NumericString type for name components.
10406 * Allow CC in the environment to override the automatically chosen
10407 compiler. Note that nothing is done to ensure flags work with the
10412 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10414 * Properly check EVP_VerifyFinal() and similar return values
10417 *Ben Laurie, Bodo Moeller, Google Security Team*
10419 * Enable TLS extensions by default.
10423 * Allow the CHIL engine to be loaded, whether the application is
10424 multithreaded or not. (This does not release the developer from the
10425 obligation to set up the dynamic locking callbacks.)
10427 *Sander Temme <sander@temme.net>*
10429 * Use correct exit code if there is an error in dgst command.
10431 *Steve Henson; problem pointed out by Roland Dirlewanger*
10433 * Tweak Configure so that you need to say "experimental-jpake" to enable
10434 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10438 * Add experimental JPAKE support, including demo authentication in
10439 s_client and s_server.
10443 * Set the comparison function in v3_addr_canonize().
10445 *Rob Austein <sra@hactrn.net>*
10447 * Add support for XMPP STARTTLS in s_client.
10449 *Philip Paeps <philip@freebsd.org>*
10451 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10452 to ensure that even with this option, only ciphersuites in the
10453 server's preference list will be accepted. (Note that the option
10454 applies only when resuming a session, so the earlier behavior was
10455 just about the algorithm choice for symmetric cryptography.)
10459 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10461 * Fix NULL pointer dereference if a DTLS server received
10462 ChangeCipherSpec as first record ([CVE-2009-1386]).
10466 * Fix a state transition in s3_srvr.c and d1_srvr.c
10467 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10469 *Nagendra Modadugu*
10471 * The fix in 0.9.8c that supposedly got rid of unsafe
10472 double-checked locking was incomplete for RSA blinding,
10473 addressing just one layer of what turns out to have been
10474 doubly unsafe triple-checked locking.
10476 So now fix this for real by retiring the MONT_HELPER macro
10477 in crypto/rsa/rsa_eay.c.
10479 *Bodo Moeller; problem pointed out by Marius Schilder*
10481 * Various precautionary measures:
10483 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10485 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10486 (NB: This would require knowledge of the secret session ticket key
10487 to exploit, in which case you'd be SOL either way.)
10489 - Change bn_nist.c so that it will properly handle input BIGNUMs
10490 outside the expected range.
10492 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10495 *Neel Mehta, Bodo Moeller*
10497 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10498 the load fails. Useful for distros.
10500 *Ben Laurie and the FreeBSD team*
10502 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10506 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10510 * Expand ENGINE to support engine supplied SSL client certificate functions.
10512 This work was sponsored by Logica.
10516 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10517 keystores. Support for SSL/TLS client authentication too.
10518 Not compiled unless enable-capieng specified to Configure.
10520 This work was sponsored by Logica.
10524 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10525 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10526 attribute creation routines such as certificate requests and PKCS#12
10531 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10533 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10534 handshake which could lead to a client crash as found using the
10535 Codenomicon TLS test suite ([CVE-2008-1672])
10537 *Steve Henson, Mark Cox*
10539 * Fix double free in TLS server name extensions which could lead to
10540 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10544 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10546 Clear the error queue to ensure that error entries left from
10547 older function calls do not interfere with the correct operation.
10549 *Lutz Jaenicke, Erik de Castro Lopo*
10551 * Remove root CA certificates of commercial CAs:
10553 The OpenSSL project does not recommend any specific CA and does not
10554 have any policy with respect to including or excluding any CA.
10555 Therefore, it does not make any sense to ship an arbitrary selection
10556 of root CA certificates with the OpenSSL software.
10560 * RSA OAEP patches to fix two separate invalid memory reads.
10561 The first one involves inputs when 'lzero' is greater than
10562 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10563 before the beginning of from). The second one involves inputs where
10564 the 'db' section contains nothing but zeroes (there is a one-byte
10565 invalid read after the end of 'db').
10567 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10569 * Partial backport from 0.9.9-dev:
10571 Introduce bn_mul_mont (dedicated Montgomery multiplication
10572 procedure) as a candidate for BIGNUM assembler implementation.
10573 While 0.9.9-dev uses assembler for various architectures, only
10574 x86_64 is available by default here in the 0.9.8 branch, and
10575 32-bit x86 is available through a compile-time setting.
10577 To try the 32-bit x86 assembler implementation, use Configure
10578 option "enable-montasm" (which exists only for this backport).
10580 As "enable-montasm" for 32-bit x86 disclaims code stability
10581 anyway, in this constellation we activate additional code
10582 backported from 0.9.9-dev for further performance improvements,
10583 namely BN_from_montgomery_word. (To enable this otherwise,
10584 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10586 *Andy Polyakov (backport partially by Bodo Moeller)*
10588 * Add TLS session ticket callback. This allows an application to set
10589 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10590 values. This is useful for key rollover for example where several key
10591 sets may exist with different names.
10595 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10596 This was broken until now in 0.9.8 releases, such that the only way
10597 a registered ENGINE could be used (assuming it initialises
10598 successfully on the host) was to explicitly set it as the default
10599 for the relevant algorithms. This is in contradiction with 0.9.7
10600 behaviour and the documentation. With this fix, when an ENGINE is
10601 registered into a given algorithm's table of implementations, the
10602 'uptodate' flag is reset so that auto-discovery will be used next
10603 time a new context for that algorithm attempts to select an
10606 *Ian Lister (tweaked by Geoff Thorpe)*
10608 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10609 implementation in the following ways:
10611 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10614 Lack of BER streaming support means one pass streaming processing is
10615 only supported if data is detached: setting the streaming flag is
10616 ignored for embedded content.
10618 CMS support is disabled by default and must be explicitly enabled
10619 with the enable-cms configuration option.
10623 * Update the GMP engine glue to do direct copies between BIGNUM and
10624 mpz_t when openssl and GMP use the same limb size. Otherwise the
10625 existing "conversion via a text string export" trick is still used.
10627 *Paul Sheer <paulsheer@gmail.com>*
10629 * Zlib compression BIO. This is a filter BIO which compressed and
10630 uncompresses any data passed through it.
10634 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10635 RFC3394 compatible AES key wrapping.
10639 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10640 sets string data without copying. X509_ALGOR_set0() and
10641 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10642 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10643 from an X509_ATTRIBUTE structure optionally checking it occurs only
10644 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10649 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10650 to get the expected BN_FLG_CONSTTIME behavior.
10652 *Bodo Moeller (Google)*
10656 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10657 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10658 - added some more tests to do_tests.pl
10659 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10660 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10661 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10662 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10663 - various changes to netware.pl to enable gcc-cross builds on Win32
10665 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10666 - various changes to fix missing prototype warnings
10667 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10668 - added AES, WHIRLPOOL and CPUID assembler code to build files
10669 - added missing AES assembler make rules to mk1mf.pl
10670 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10672 *Guenter Knauf <eflash@gmx.net>*
10674 * Implement certificate status request TLS extension defined in RFC3546.
10675 A client can set the appropriate parameters and receive the encoded
10676 OCSP response via a callback. A server can query the supplied parameters
10677 and set the encoded OCSP response in the callback. Add simplified examples
10678 to s_client and s_server.
10682 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10684 * Fix various bugs:
10685 + Binary incompatibility of ssl_ctx_st structure
10686 + DTLS interoperation with non-compliant servers
10687 + Don't call get_session_cb() without proposed session
10688 + Fix ia64 assembler code
10690 *Andy Polyakov, Steve Henson*
10692 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10694 * DTLS Handshake overhaul. There were longstanding issues with
10695 OpenSSL DTLS implementation, which were making it impossible for
10696 RFC 4347 compliant client to communicate with OpenSSL server.
10697 Unfortunately just fixing these incompatibilities would "cut off"
10698 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10699 server keeps tolerating non RFC compliant syntax. The opposite is
10700 not true, 0.9.8f client can not communicate with earlier server.
10701 This update even addresses CVE-2007-4995.
10705 * Changes to avoid need for function casts in OpenSSL: some compilers
10706 (gcc 4.2 and later) reject their use.
10707 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10710 * Add RFC4507 support to OpenSSL. This includes the corrections in
10711 RFC4507bis. The encrypted ticket format is an encrypted encoded
10712 SSL_SESSION structure, that way new session features are automatically
10715 If a client application caches session in an SSL_SESSION structure
10716 support is transparent because tickets are now stored in the encoded
10719 The SSL_CTX structure automatically generates keys for ticket
10720 protection in servers so again support should be possible
10721 with no application modification.
10723 If a client or server wishes to disable RFC4507 support then the option
10724 SSL_OP_NO_TICKET can be set.
10726 Add a TLS extension debugging callback to allow the contents of any client
10727 or server extensions to be examined.
10729 This work was sponsored by Google.
10733 * Add initial support for TLS extensions, specifically for the server_name
10734 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10735 have new members for a hostname. The SSL data structure has an
10736 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10737 stored in that context to allow for session resumption, even after the
10738 SSL has been switched to a new SSL_CTX in reaction to a client's
10739 server_name extension.
10741 New functions (subject to change):
10743 SSL_get_servername()
10744 SSL_get_servername_type()
10747 New CTRL codes and macros (subject to change):
10749 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10750 - SSL_CTX_set_tlsext_servername_callback()
10751 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10752 - SSL_CTX_set_tlsext_servername_arg()
10753 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10755 openssl s_client has a new '-servername ...' option.
10757 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10758 '-key2 ...', '-servername_fatal' (subject to change). This allows
10759 testing the HostName extension for a specific single hostname ('-cert'
10760 and '-key' remain fallbacks for handshakes without HostName
10761 negotiation). If the unrecognized_name alert has to be sent, this by
10762 default is a warning; it becomes fatal with the '-servername_fatal'
10765 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10767 * Add AES and SSE2 assembly language support to VC++ build.
10771 * Mitigate attack on final subtraction in Montgomery reduction.
10775 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10776 (which previously caused an internal error).
10780 * Squeeze another 10% out of IGE mode when in != out.
10784 * AES IGE mode speedup.
10786 *Dean Gaudet (Google)*
10788 * Add the Korean symmetric 128-bit cipher SEED (see
10789 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10790 add SEED ciphersuites from RFC 4162:
10792 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10793 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10794 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10795 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10797 To minimize changes between patchlevels in the OpenSSL 0.9.8
10798 series, SEED remains excluded from compilation unless OpenSSL
10799 is configured with 'enable-seed'.
10801 *KISA, Bodo Moeller*
10803 * Mitigate branch prediction attacks, which can be practical if a
10804 single processor is shared, allowing a spy process to extract
10805 information. For detailed background information, see
10806 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10807 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10808 and Necessary Software Countermeasures"). The core of the change
10809 are new versions BN_div_no_branch() and
10810 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10811 respectively, which are slower, but avoid the security-relevant
10812 conditional branches. These are automatically called by BN_div()
10813 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10814 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10815 remove a conditional branch.
10817 BN_FLG_CONSTTIME is the new name for the previous
10818 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10819 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10820 in the exponent causes BN_mod_exp_mont() to use the alternative
10821 implementation in BN_mod_exp_mont_consttime().) The old name
10822 remains as a deprecated alias.
10824 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10825 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10826 constant-time implementations for more than just exponentiation.
10827 Here too the old name is kept as a deprecated alias.
10829 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10830 the BN_BLINDING structure gets an independent copy of the
10831 modulus. This means that the previous `BIGNUM *m` argument to
10832 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10833 essentially becomes `const BIGNUM *m`, although we can't actually
10834 change this in the header file before 0.9.9. It allows
10835 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10836 enable BN_FLG_CONSTTIME.
10838 *Matthew D Wood (Intel Corp)*
10840 * In the SSL/TLS server implementation, be strict about session ID
10841 context matching (which matters if an application uses a single
10842 external cache for different purposes). Previously,
10843 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10844 set. This did ensure strict client verification, but meant that,
10845 with applications using a single external cache for quite
10846 different requirements, clients could circumvent ciphersuite
10847 restrictions for a given session ID context by starting a session
10848 in a different context.
10852 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10853 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10854 authentication-only ciphersuites.
10858 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10859 not complete and could lead to a possible single byte overflow
10860 ([CVE-2007-5135]) [Ben Laurie]
10862 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10864 * Since AES128 and AES256 (and similarly Camellia128 and
10865 Camellia256) share a single mask bit in the logic of
10866 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10867 kludge to work properly if AES128 is available and AES256 isn't
10868 (or if Camellia128 is available and Camellia256 isn't).
10872 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10873 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10874 When a point or a seed is encoded in a BIT STRING, we need to
10875 prevent the removal of trailing zero bits to get the proper DER
10876 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10877 of a NamedBitList, for which trailing 0 bits need to be removed.)
10881 * Have SSL/TLS server implementation tolerate "mismatched" record
10882 protocol version while receiving ClientHello even if the
10883 ClientHello is fragmented. (The server can't insist on the
10884 particular protocol version it has chosen before the ServerHello
10885 message has informed the client about his choice.)
10889 * Add RFC 3779 support.
10891 *Rob Austein for ARIN, Ben Laurie*
10893 * Load error codes if they are not already present instead of using a
10894 static variable. This allows them to be cleanly unloaded and reloaded.
10895 Improve header file function name parsing.
10899 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10900 or CAPABILITY handshake as required by RFCs.
10902 *Goetz Babin-Ebell*
10904 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10906 * Introduce limits to prevent malicious keys being able to
10907 cause a denial of service. ([CVE-2006-2940])
10909 *Steve Henson, Bodo Moeller*
10911 * Fix ASN.1 parsing of certain invalid structures that can result
10912 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10914 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10915 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10917 * Fix SSL client code which could crash if connecting to a
10918 malicious SSLv2 server. ([CVE-2006-4343])
10920 *Tavis Ormandy and Will Drewry, Google Security Team*
10922 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10923 match only those. Before that, "AES256-SHA" would be interpreted
10924 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10925 the same strength classification in 0.9.7h) as we currently only
10926 have a single AES bit in the ciphersuite description bitmap.
10927 That change, however, also applied to ciphersuite strings such as
10928 "RC4-MD5" that intentionally matched multiple ciphersuites --
10929 namely, SSL 2.0 ciphersuites in addition to the more common ones
10930 from SSL 3.0/TLS 1.0.
10932 So we change the selection algorithm again: Naming an explicit
10933 ciphersuite selects this one ciphersuite, and any other similar
10934 ciphersuite (same bitmap) from *other* protocol versions.
10935 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10936 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10938 Since SSL 2.0 does not have any ciphersuites for which the
10939 128/256 bit distinction would be relevant, this works for now.
10940 The proper fix will be to use different bits for AES128 and
10941 AES256, which would have avoided the problems from the beginning;
10942 however, bits are scarce, so we can only do this in a new release
10943 (not just a patchlevel) when we can change the SSL_CIPHER
10944 definition to split the single 'unsigned long mask' bitmap into
10945 multiple values to extend the available space.
10949 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10951 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10952 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10954 * Add AES IGE and biIGE modes.
10958 * Change the Unix randomness entropy gathering to use poll() when
10959 possible instead of select(), since the latter has some
10960 undesirable limitations.
10962 *Darryl Miles via Richard Levitte and Bodo Moeller*
10964 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10965 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10966 cannot be implicitly activated as part of, e.g., the "AES" alias.
10967 However, please upgrade to OpenSSL 0.9.9[-dev] for
10968 non-experimental use of the ECC ciphersuites to get TLS extension
10969 support, which is required for curve and point format negotiation
10970 to avoid potential handshake problems.
10974 * Disable rogue ciphersuites:
10976 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10977 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10978 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10980 The latter two were purportedly from
10981 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10984 Also deactivate the remaining ciphersuites from
10985 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10986 unofficial, and the ID has long expired.
10990 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10991 dual-core machines) and other potential thread-safety issues.
10995 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10996 versions), which is now available for royalty-free use
10997 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10998 Also, add Camellia TLS ciphersuites from RFC 4132.
11000 To minimize changes between patchlevels in the OpenSSL 0.9.8
11001 series, Camellia remains excluded from compilation unless OpenSSL
11002 is configured with 'enable-camellia'.
11006 * Disable the padding bug check when compression is in use. The padding
11007 bug check assumes the first packet is of even length, this is not
11008 necessarily true if compression is enabled and can result in false
11009 positives causing handshake failure. The actual bug test is ancient
11010 code so it is hoped that implementations will either have fixed it by
11011 now or any which still have the bug do not support compression.
11015 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11017 * When applying a cipher rule check to see if string match is an explicit
11018 cipher suite and only match that one cipher suite if it is.
11022 * Link in manifests for VC++ if needed.
11024 *Austin Ziegler <halostatue@gmail.com>*
11026 * Update support for ECC-based TLS ciphersuites according to
11027 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11028 TLS extensions, which are supported starting with the 0.9.9
11029 branch, not in the OpenSSL 0.9.8 branch).
11033 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11034 opaque EVP_CIPHER_CTX handling.
11038 * Fixes and enhancements to zlib compression code. We now only use
11039 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11040 to conform with the standards mentioned here:
11041 <http://www.zlib.net/DLL_FAQ.txt>
11042 Static zlib linking now works on Windows and the new --with-zlib-include
11043 --with-zlib-lib options to Configure can be used to supply the location
11044 of the headers and library. Gracefully handle case where zlib library
11049 * Several fixes and enhancements to the OID generation code. The old code
11050 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11051 handle numbers larger than ULONG_MAX, truncated printing and had a
11052 non standard OBJ_obj2txt() behaviour.
11056 * Add support for building of engines under engine/ as shared libraries
11057 under VC++ build system.
11061 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11062 Hopefully, we will not see any false combination of paths any more.
11066 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11068 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11069 (part of SSL_OP_ALL). This option used to disable the
11070 countermeasure against man-in-the-middle protocol-version
11071 rollback in the SSL 2.0 server implementation, which is a bad
11072 idea. ([CVE-2005-2969])
11074 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11075 for Information Security, National Institute of Advanced Industrial
11076 Science and Technology [AIST], Japan)*
11078 * Add two function to clear and return the verify parameter flags.
11082 * Keep cipherlists sorted in the source instead of sorting them at
11083 runtime, thus removing the need for a lock.
11087 * Avoid some small subgroup attacks in Diffie-Hellman.
11089 *Nick Mathewson and Ben Laurie*
11091 * Add functions for well-known primes.
11095 * Extended Windows CE support.
11097 *Satoshi Nakamura and Andy Polyakov*
11099 * Initialize SSL_METHOD structures at compile time instead of during
11100 runtime, thus removing the need for a lock.
11104 * Make PKCS7_decrypt() work even if no certificate is supplied by
11105 attempting to decrypt each encrypted key in turn. Add support to
11110 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11112 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11115 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11119 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11120 key into the same file any more.
11124 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11128 * Add -utf8 command line and config file option to 'ca'.
11130 *Stefan <stf@udoma.org*
11132 * Removed the macro des_crypt(), as it seems to conflict with some
11133 libraries. Use DES_crypt().
11137 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11138 involves renaming the source and generated shared-libs for
11139 both. The engines will accept the corrected or legacy ids
11140 ('ncipher' and '4758_cca' respectively) when binding. NB,
11141 this only applies when building 'shared'.
11143 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11145 * Add attribute functions to EVP_PKEY structure. Modify
11146 PKCS12_create() to recognize a CSP name attribute and
11147 use it. Make -CSP option work again in pkcs12 utility.
11151 * Add new functionality to the bn blinding code:
11152 - automatic re-creation of the BN_BLINDING parameters after
11153 a fixed number of uses (currently 32)
11154 - add new function for parameter creation
11155 - introduce flags to control the update behaviour of the
11156 BN_BLINDING parameters
11157 - hide BN_BLINDING structure
11158 Add a second BN_BLINDING slot to the RSA structure to improve
11159 performance when a single RSA object is shared among several
11164 * Add support for DTLS.
11166 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11168 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11169 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11173 * Remove buggy and incomplete DH cert support from
11174 ssl/ssl_rsa.c and ssl/s3_both.c
11178 * Use SHA-1 instead of MD5 as the default digest algorithm for
11179 the `apps/openssl` commands.
11183 * Compile clean with "-Wall -Wmissing-prototypes
11184 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11185 DEBUG_SAFESTACK must also be set.
11189 * Change ./Configure so that certain algorithms can be disabled by default.
11190 The new counterpiece to "no-xxx" is "enable-xxx".
11192 The patented RC5 and MDC2 algorithms will now be disabled unless
11193 "enable-rc5" and "enable-mdc2", respectively, are specified.
11195 (IDEA remains enabled despite being patented. This is because IDEA
11196 is frequently required for interoperability, and there is no license
11197 fee for non-commercial use. As before, "no-idea" can be used to
11198 avoid this algorithm.)
11202 * Add processing of proxy certificates (see RFC 3820). This work was
11203 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11204 EGEE (Enabling Grids for E-science in Europe).
11208 * RC4 performance overhaul on modern architectures/implementations, such
11209 as Intel P4, IA-64 and AMD64.
11213 * New utility extract-section.pl. This can be used specify an alternative
11214 section number in a pod file instead of having to treat each file as
11215 a separate case in Makefile. This can be done by adding two lines to the
11218 =for comment openssl_section:XXX
11220 The blank line is mandatory.
11224 * New arguments -certform, -keyform and -pass for s_client and s_server
11225 to allow alternative format key and certificate files and passphrase
11230 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11231 update associated structures and add various utility functions.
11233 Add new policy related verify parameters, include policy checking in
11234 standard verify code. Enhance 'smime' application with extra parameters
11235 to support policy checking and print out.
11239 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11240 Nehemiah processors. These extensions support AES encryption in hardware
11241 as well as RNG (though RNG support is currently disabled).
11243 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11245 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11249 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11251 *Andy Polyakov and a number of other people*
11253 * Improved PowerPC platform support. Most notably BIGNUM assembler
11254 implementation contributed by IBM.
11256 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11258 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11259 exponent rather than 'unsigned long'. There is a corresponding change to
11260 the new 'rsa_keygen' element of the RSA_METHOD structure.
11262 *Jelte Jansen, Geoff Thorpe*
11264 * Functionality for creating the initial serial number file is now
11265 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11267 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11268 number file to 1, which is bound to cause problems. To avoid
11269 the problems while respecting compatibility between different 0.9.7
11270 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11271 CA.pl for serial number initialization. With the new release 0.9.8,
11272 we can fix the problem directly in the 'ca' utility.)
11276 * Reduced header interdependencies by declaring more opaque objects in
11277 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11278 give fewer recursive includes, which could break lazy source code - so
11279 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11280 developers should define this symbol when building and using openssl to
11281 ensure they track the recommended behaviour, interfaces, [etc], but
11282 backwards-compatible behaviour prevails when this isn't defined.
11286 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11290 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11291 This will generate a random key of the appropriate length based on the
11292 cipher context. The EVP_CIPHER can provide its own random key generation
11293 routine to support keys of a specific form. This is used in the des and
11294 3des routines to generate a key of the correct parity. Update S/MIME
11295 code to use new functions and hence generate correct parity DES keys.
11296 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11297 valid (weak or incorrect parity).
11301 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11302 as looking them up. This is useful when the verified structure may contain
11303 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11304 present unless the new PKCS7_NO_CRL flag is asserted.
11308 * Extend ASN1 oid configuration module. It now additionally accepts the
11311 shortName = some long name, 1.2.3.4
11315 * Reimplemented the BN_CTX implementation. There is now no more static
11316 limitation on the number of variables it can handle nor the depth of the
11317 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11318 information can now expand as required, and rather than having a single
11319 static array of bignums, BN_CTX now uses a linked-list of such arrays
11320 allowing it to expand on demand whilst maintaining the usefulness of
11321 BN_CTX's "bundling".
11325 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11326 to allow all RSA operations to function using a single BN_CTX.
11330 * Preliminary support for certificate policy evaluation and checking. This
11331 is initially intended to pass the tests outlined in "Conformance Testing
11332 of Relying Party Client Certificate Path Processing Logic" v1.07.
11336 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11337 remained unused and not that useful. A variety of other little bignum
11338 tweaks and fixes have also been made continuing on from the audit (see
11343 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11344 associated ASN1, EVP and SSL functions and old ASN1 macros.
11348 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11349 and this should never fail. So the return value from the use of
11350 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11351 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11355 * BN_CTX_get() should return zero-valued bignums, providing the same
11356 initialised value as BN_new().
11358 *Geoff Thorpe, suggested by Ulf Möller*
11360 * Support for inhibitAnyPolicy certificate extension.
11364 * An audit of the BIGNUM code is underway, for which debugging code is
11365 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11366 is considered valid when processing BIGNUMs, and causes execution to
11367 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11368 further steps are taken to deliberately pollute unused data in BIGNUM
11369 structures to try and expose faulty code further on. For now, openssl will
11370 (in its default mode of operation) continue to tolerate the inconsistent
11371 forms that it has tolerated in the past, but authors and packagers should
11372 consider trying openssl and their own applications when compiled with
11373 these debugging symbols defined. It will help highlight potential bugs in
11374 their own code, and will improve the test coverage for OpenSSL itself. At
11375 some point, these tighter rules will become openssl's default to improve
11376 maintainability, though the assert()s and other overheads will remain only
11377 in debugging configurations. See bn.h for more details.
11379 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11381 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11382 that can only be obtained through BN_CTX_new() (which implicitly
11383 initialises it). The presence of this function only made it possible
11384 to overwrite an existing structure (and cause memory leaks).
11388 * Because of the callback-based approach for implementing LHASH as a
11389 template type, lh_insert() adds opaque objects to hash-tables and
11390 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11391 to clean up those corresponding objects before destroying the hash table
11392 (and losing the object pointers). So some over-zealous constifications in
11393 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11394 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11395 prototyped to have "const" restrictions on the object pointers they are
11396 given (and so aren't required to cast them away any more).
11400 * The tmdiff.h API was so ugly and minimal that our own timing utility
11401 (speed) prefers to use its own implementation. The two implementations
11402 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11403 its object type properly exposed (MS_TM) instead of casting to/from
11404 `char *`. This may still change yet if someone realises MS_TM and
11406 aren't necessarily the greatest nomenclatures - but this is what was used
11407 internally to the implementation so I've used that for now.
11411 * Ensure that deprecated functions do not get compiled when
11412 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11413 the self-tests were still using deprecated key-generation functions so
11414 these have been updated also.
11418 * Reorganise PKCS#7 code to separate the digest location functionality
11419 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11420 New function PKCS7_set_digest() to set the digest type for PKCS#7
11421 digestedData type. Add additional code to correctly generate the
11422 digestedData type and add support for this type in PKCS7 initialization
11427 * New function PKCS7_set0_type_other() this initializes a PKCS7
11428 structure of type "other".
11432 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11433 sure the loop does correctly stop and breaking ("division by zero")
11434 modulus operations are not performed. The (pre-generated) prime
11435 table crypto/bn/bn_prime.h was already correct, but it could not be
11436 re-generated on some platforms because of the "division by zero"
11437 situation in the script.
11439 *Ralf S. Engelschall*
11441 * Update support for ECC-based TLS ciphersuites according to
11442 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11443 SHA-1 now is only used for "small" curves (where the
11444 representation of a field element takes up to 24 bytes); for
11445 larger curves, the field element resulting from ECDH is directly
11446 used as premaster secret.
11448 *Douglas Stebila (Sun Microsystems Laboratories)*
11450 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11451 curve secp160r1 to the tests.
11453 *Douglas Stebila (Sun Microsystems Laboratories)*
11455 * Add the possibility to load symbols globally with DSO.
11457 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11459 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11460 control of the error stack.
11464 * Add support for STORE in ENGINE.
11468 * Add the STORE type. The intention is to provide a common interface
11469 to certificate and key stores, be they simple file-based stores, or
11470 HSM-type store, or LDAP stores, or...
11471 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11475 * Add a generic structure called OPENSSL_ITEM. This can be used to
11476 pass a list of arguments to any function as well as provide a way
11477 for a function to pass data back to the caller.
11481 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11482 works like BUF_strdup() but can be used to duplicate a portion of
11483 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11488 * Add the function sk_find_ex() which works like sk_find(), but will
11489 return an index to an element even if an exact match couldn't be
11490 found. The index is guaranteed to point at the element where the
11491 searched-for key would be inserted to preserve sorting order.
11495 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11496 takes an extra flags argument for optional functionality. Currently,
11497 the following flags are defined:
11499 OBJ_BSEARCH_VALUE_ON_NOMATCH
11500 This one gets OBJ_bsearch_ex() to return a pointer to the first
11501 element where the comparing function returns a negative or zero
11504 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11505 This one gets OBJ_bsearch_ex() to return a pointer to the first
11506 element where the comparing function returns zero. This is useful
11507 if there are more than one element where the comparing function
11512 * Make it possible to create self-signed certificates with 'openssl ca'
11513 in such a way that the self-signed certificate becomes part of the
11514 CA database and uses the same mechanisms for serial number generation
11515 as all other certificate signing. The new flag '-selfsign' enables
11516 this functionality. Adapt CA.sh and CA.pl.in.
11520 * Add functionality to check the public key of a certificate request
11521 against a given private. This is useful to check that a certificate
11522 request can be signed by that key (self-signing).
11526 * Make it possible to have multiple active certificates with the same
11527 subject in the CA index file. This is done only if the keyword
11528 'unique_subject' is set to 'no' in the main CA section (default
11529 if 'CA_default') of the configuration file. The value is saved
11530 with the database itself in a separate index attribute file,
11531 named like the index file with '.attr' appended to the name.
11535 * Generate multi-valued AVAs using '+' notation in config files for
11540 * Support for nameConstraints certificate extension.
11544 * Support for policyConstraints certificate extension.
11548 * Support for policyMappings certificate extension.
11552 * Make sure the default DSA_METHOD implementation only uses its
11553 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11554 and change its own handlers to be NULL so as to remove unnecessary
11555 indirection. This lets alternative implementations fallback to the
11556 default implementation more easily.
11560 * Support for directoryName in GeneralName related extensions
11565 * Make it possible to link applications using Makefile.shared.
11566 Make that possible even when linking against static libraries!
11570 * Support for single pass processing for S/MIME signing. This now
11571 means that S/MIME signing can be done from a pipe, in addition
11572 cleartext signing (multipart/signed type) is effectively streaming
11573 and the signed data does not need to be all held in memory.
11575 This is done with a new flag PKCS7_STREAM. When this flag is set
11576 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11577 is done after the data is output (and digests calculated) in
11578 SMIME_write_PKCS7().
11582 * Add full support for -rpath/-R, both in shared libraries and
11583 applications, at least on the platforms where it's known how
11588 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11589 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11590 will now compute a table of multiples of the generator that
11591 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11592 faster (notably in the case of a single point multiplication,
11593 scalar * generator).
11595 *Nils Larsch, Bodo Moeller*
11597 * IPv6 support for certificate extensions. The various extensions
11598 which use the IP:a.b.c.d can now take IPv6 addresses using the
11599 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11604 * Added an ENGINE that implements RSA by performing private key
11605 exponentiations with the GMP library. The conversions to and from
11606 GMP's mpz_t format aren't optimised nor are any montgomery forms
11607 cached, and on x86 it appears OpenSSL's own performance has caught up.
11608 However there are likely to be other architectures where GMP could
11609 provide a boost. This ENGINE is not built in by default, but it can be
11610 specified at Configure time and should be accompanied by the necessary
11611 linker additions, eg;
11612 ./config -DOPENSSL_USE_GMP -lgmp
11616 * "openssl engine" will not display ENGINE/DSO load failure errors when
11617 testing availability of engines with "-t" - the old behaviour is
11618 produced by increasing the feature's verbosity with "-tt".
11622 * ECDSA routines: under certain error conditions uninitialized BN objects
11623 could be freed. Solution: make sure initialization is performed early
11624 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11629 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11630 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11631 software implementations. For DSA and DH, parameter generation can
11632 also be overridden by providing the appropriate method callbacks.
11636 * Change the "progress" mechanism used in key-generation and
11637 primality testing to functions that take a new BN_GENCB pointer in
11638 place of callback/argument pairs. The new API functions have `_ex`
11639 postfixes and the older functions are reimplemented as wrappers for
11640 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11641 declarations of the old functions to help (graceful) attempts to
11642 migrate to the new functions. Also, the new key-generation API
11643 functions operate on a caller-supplied key-structure and return
11644 success/failure rather than returning a key or NULL - this is to
11645 help make "keygen" another member function of RSA_METHOD etc.
11647 Example for using the new callback interface:
11649 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11650 void *my_arg = ...;
11653 BN_GENCB_set(&my_cb, my_callback, my_arg);
11655 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11656 /* For the meaning of a, b in calls to my_callback(), see the
11657 * documentation of the function that calls the callback.
11658 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11659 * my_callback should return 1 if it wants BN_is_prime_ex()
11660 * to continue, or 0 to stop.
11665 * Change the ZLIB compression method to be stateful, and make it
11666 available to TLS with the number defined in
11667 draft-ietf-tls-compression-04.txt.
11671 * Add the ASN.1 structures and functions for CertificatePair, which
11672 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11674 CertificatePair ::= SEQUENCE {
11675 forward [0] Certificate OPTIONAL,
11676 reverse [1] Certificate OPTIONAL,
11677 -- at least one of the pair shall be present -- }
11679 Also implement the PEM functions to read and write certificate
11680 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11682 This needed to be defined, mostly for the sake of the LDAP
11683 attribute crossCertificatePair, but may prove useful elsewhere as
11688 * Make it possible to inhibit symlinking of shared libraries in
11689 Makefile.shared, for Cygwin's sake.
11693 * Extend the BIGNUM API by creating a function
11694 void BN_set_negative(BIGNUM *a, int neg);
11695 and a macro that behave like
11696 int BN_is_negative(const BIGNUM *a);
11698 to avoid the need to access 'a->neg' directly in applications.
11702 * Implement fast modular reduction for pseudo-Mersenne primes
11703 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11704 EC_GROUP_new_curve_GFp() will now automatically use this
11707 *Nils Larsch <nla@trustcenter.de>*
11709 * Add new lock type (CRYPTO_LOCK_BN).
11713 * Change the ENGINE framework to automatically load engines
11714 dynamically from specific directories unless they could be
11715 found to already be built in or loaded. Move all the
11716 current engines except for the cryptodev one to a new
11717 directory engines/.
11718 The engines in engines/ are built as shared libraries if
11719 the "shared" options was given to ./Configure or ./config.
11720 Otherwise, they are inserted in libcrypto.a.
11721 /usr/local/ssl/engines is the default directory for dynamic
11722 engines, but that can be overridden at configure time through
11723 the usual use of --prefix and/or --openssldir, and at run
11724 time with the environment variable OPENSSL_ENGINES.
11726 *Geoff Thorpe and Richard Levitte*
11728 * Add Makefile.shared, a helper makefile to build shared
11729 libraries. Adapt Makefile.org.
11733 * Add version info to Win32 DLLs.
11735 *Peter 'Luna' Runestig" <peter@runestig.com>*
11737 * Add new 'medium level' PKCS#12 API. Certificates and keys
11738 can be added using this API to created arbitrary PKCS#12
11739 files while avoiding the low-level API.
11741 New options to PKCS12_create(), key or cert can be NULL and
11742 will then be omitted from the output file. The encryption
11743 algorithm NIDs can be set to -1 for no encryption, the mac
11744 iteration count can be set to 0 to omit the mac.
11746 Enhance pkcs12 utility by making the -nokeys and -nocerts
11747 options work when creating a PKCS#12 file. New option -nomac
11748 to omit the mac, NONE can be set for an encryption algorithm.
11749 New code is modified to use the enhanced PKCS12_create()
11750 instead of the low-level API.
11754 * Extend ASN1 encoder to support indefinite length constructed
11755 encoding. This can output sequences tags and octet strings in
11756 this form. Modify pk7_asn1.c to support indefinite length
11757 encoding. This is experimental and needs additional code to
11758 be useful, such as an ASN1 bio and some enhanced streaming
11761 Extend template encode functionality so that tagging is passed
11762 down to the template encoder.
11766 * Let 'openssl req' fail if an argument to '-newkey' is not
11767 recognized instead of using RSA as a default.
11771 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11772 As these are not official, they are not included in "ALL";
11773 the "ECCdraft" ciphersuite group alias can be used to select them.
11775 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11777 * Add ECDH engine support.
11779 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11781 * Add ECDH in new directory crypto/ecdh/.
11783 *Douglas Stebila (Sun Microsystems Laboratories)*
11785 * Let BN_rand_range() abort with an error after 100 iterations
11786 without success (which indicates a broken PRNG).
11790 * Change BN_mod_sqrt() so that it verifies that the input value
11791 is really the square of the return value. (Previously,
11792 BN_mod_sqrt would show GIGO behaviour.)
11796 * Add named elliptic curves over binary fields from X9.62, SECG,
11797 and WAP/WTLS; add OIDs that were still missing.
11799 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11801 * Extend the EC library for elliptic curves over binary fields
11802 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11805 EC_GF2m_simple_method
11809 EC_GROUP_new_curve_GF2m
11810 EC_GROUP_set_curve_GF2m
11811 EC_GROUP_get_curve_GF2m
11812 EC_POINT_set_affine_coordinates_GF2m
11813 EC_POINT_get_affine_coordinates_GF2m
11814 EC_POINT_set_compressed_coordinates_GF2m
11816 Point compression for binary fields is disabled by default for
11817 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11820 As binary polynomials are represented as BIGNUMs, various members
11821 of the EC_GROUP and EC_POINT data structures can be shared
11822 between the implementations for prime fields and binary fields;
11823 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11824 are essentially identical to their `..._GFp` counterparts.
11825 (For simplicity, the `..._GFp` prefix has been dropped from
11826 various internal method names.)
11828 An internal 'field_div' method (similar to 'field_mul' and
11829 'field_sqr') has been added; this is used only for binary fields.
11831 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11833 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11834 through methods ('mul', 'precompute_mult').
11836 The generic implementations (now internally called 'ec_wNAF_mul'
11837 and 'ec_wNAF_precomputed_mult') remain the default if these
11838 methods are undefined.
11840 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11842 * New function EC_GROUP_get_degree, which is defined through
11843 EC_METHOD. For curves over prime fields, this returns the bit
11844 length of the modulus.
11846 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11848 * New functions EC_GROUP_dup, EC_POINT_dup.
11849 (These simply call ..._new and ..._copy).
11851 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11853 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11854 Polynomials are represented as BIGNUMs (where the sign bit is not
11855 used) in the following functions [macros]:
11858 BN_GF2m_sub [= BN_GF2m_add]
11859 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11860 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11861 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11863 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11864 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11865 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11866 BN_GF2m_cmp [= BN_ucmp]
11868 (Note that only the 'mod' functions are actually for fields GF(2^m).
11869 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11871 For some functions, an the irreducible polynomial defining a
11872 field can be given as an 'unsigned int[]' with strictly
11873 decreasing elements giving the indices of those bits that are set;
11874 i.e., p[] represents the polynomial
11875 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11877 p[0] > p[1] > ... > p[k] = 0.
11878 This applies to the following functions:
11881 BN_GF2m_mod_mul_arr
11882 BN_GF2m_mod_sqr_arr
11883 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11884 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11885 BN_GF2m_mod_exp_arr
11886 BN_GF2m_mod_sqrt_arr
11887 BN_GF2m_mod_solve_quad_arr
11891 Conversion can be performed by the following functions:
11896 bntest.c has additional tests for binary polynomial arithmetic.
11898 Two implementations for BN_GF2m_mod_div() are available.
11899 The default algorithm simply uses BN_GF2m_mod_inv() and
11900 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11901 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11902 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11904 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11906 * Add new error code 'ERR_R_DISABLED' that can be used when some
11907 functionality is disabled at compile-time.
11909 *Douglas Stebila <douglas.stebila@sun.com>*
11911 * Change default behaviour of 'openssl asn1parse' so that more
11912 information is visible when viewing, e.g., a certificate:
11914 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11915 mode the content of non-printable OCTET STRINGs is output in a
11916 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11917 avoid the appearance of a printable string.
11919 *Nils Larsch <nla@trustcenter.de>*
11921 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11923 EC_GROUP_set_asn1_flag()
11924 EC_GROUP_get_asn1_flag()
11925 EC_GROUP_set_point_conversion_form()
11926 EC_GROUP_get_point_conversion_form()
11927 These control ASN1 encoding details:
11928 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11929 has been set to OPENSSL_EC_NAMED_CURVE.
11930 - Points are encoded in uncompressed form by default; options for
11931 asn1_for are as for point2oct, namely
11932 POINT_CONVERSION_COMPRESSED
11933 POINT_CONVERSION_UNCOMPRESSED
11934 POINT_CONVERSION_HYBRID
11936 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11938 EC_GROUP_set_seed()
11939 EC_GROUP_get0_seed()
11940 EC_GROUP_get_seed_len()
11941 This is used only for ASN1 purposes (so far).
11943 *Nils Larsch <nla@trustcenter.de>*
11945 * Add 'field_type' member to EC_METHOD, which holds the NID
11946 of the appropriate field type OID. The new function
11947 EC_METHOD_get_field_type() returns this value.
11949 *Nils Larsch <nla@trustcenter.de>*
11952 EC_POINT_point2bn()
11953 EC_POINT_bn2point()
11954 EC_POINT_point2hex()
11955 EC_POINT_hex2point()
11956 providing useful interfaces to EC_POINT_point2oct() and
11957 EC_POINT_oct2point().
11959 *Nils Larsch <nla@trustcenter.de>*
11961 * Change internals of the EC library so that the functions
11962 EC_GROUP_set_generator()
11963 EC_GROUP_get_generator()
11964 EC_GROUP_get_order()
11965 EC_GROUP_get_cofactor()
11966 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11967 to methods, which would lead to unnecessary code duplication when
11968 adding different types of curves.
11970 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11972 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11973 arithmetic, and such that modified wNAFs are generated
11974 (which avoid length expansion in many cases).
11978 * Add a function EC_GROUP_check_discriminant() (defined via
11979 EC_METHOD) that verifies that the curve discriminant is non-zero.
11981 Add a function EC_GROUP_check() that makes some sanity tests
11982 on a EC_GROUP, its generator and order. This includes
11983 EC_GROUP_check_discriminant().
11985 *Nils Larsch <nla@trustcenter.de>*
11987 * Add ECDSA in new directory crypto/ecdsa/.
11989 Add applications 'openssl ecparam' and 'openssl ecdsa'
11990 (these are based on 'openssl dsaparam' and 'openssl dsa').
11992 ECDSA support is also included in various other files across the
11993 library. Most notably,
11994 - 'openssl req' now has a '-newkey ecdsa:file' option;
11995 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11996 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11997 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11998 them suitable for ECDSA where domain parameters must be
11999 extracted before the specific public key;
12000 - ECDSA engine support has been added.
12002 *Nils Larsch <nla@trustcenter.de>*
12004 * Include some named elliptic curves, and add OIDs from X9.62,
12005 SECG, and WAP/WTLS. Each curve can be obtained from the new
12007 EC_GROUP_new_by_curve_name(),
12008 and the list of available named curves can be obtained with
12009 EC_get_builtin_curves().
12010 Also add a 'curve_name' member to EC_GROUP objects, which can be
12012 EC_GROUP_set_curve_name()
12013 EC_GROUP_get_curve_name()
12015 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12017 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12018 was actually never needed) and in BN_mul(). The removal in BN_mul()
12019 required a small change in bn_mul_part_recursive() and the addition
12020 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12021 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12022 bn_sub_words() and bn_add_words() except they take arrays with
12027 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12029 * Cleanse PEM buffers before freeing them since they may contain
12032 *Benjamin Bennett <ben@psc.edu>*
12034 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12035 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12036 authentication-only ciphersuites.
12040 * Since AES128 and AES256 share a single mask bit in the logic of
12041 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12042 kludge to work properly if AES128 is available and AES256 isn't.
12046 * Expand security boundary to match 1.1.1 module.
12050 * Remove redundant features: hash file source, editing of test vectors
12051 modify fipsld to use external fips_premain.c signature.
12055 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12056 run algorithm test programs.
12060 * Make algorithm test programs more tolerant of whitespace.
12064 * Have SSL/TLS server implementation tolerate "mismatched" record
12065 protocol version while receiving ClientHello even if the
12066 ClientHello is fragmented. (The server can't insist on the
12067 particular protocol version it has chosen before the ServerHello
12068 message has informed the client about his choice.)
12072 * Load error codes if they are not already present instead of using a
12073 static variable. This allows them to be cleanly unloaded and reloaded.
12077 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12079 * Introduce limits to prevent malicious keys being able to
12080 cause a denial of service. ([CVE-2006-2940])
12082 *Steve Henson, Bodo Moeller*
12084 * Fix ASN.1 parsing of certain invalid structures that can result
12085 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12087 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12088 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12090 * Fix SSL client code which could crash if connecting to a
12091 malicious SSLv2 server. ([CVE-2006-4343])
12093 *Tavis Ormandy and Will Drewry, Google Security Team*
12095 * Change ciphersuite string processing so that an explicit
12096 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12097 will no longer include "AES128-SHA"), and any other similar
12098 ciphersuite (same bitmap) from *other* protocol versions (so that
12099 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12100 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12101 changes from 0.9.8b and 0.9.8d.
12105 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12107 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12108 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12110 * Change the Unix randomness entropy gathering to use poll() when
12111 possible instead of select(), since the latter has some
12112 undesirable limitations.
12114 *Darryl Miles via Richard Levitte and Bodo Moeller*
12116 * Disable rogue ciphersuites:
12118 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12119 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12120 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12122 The latter two were purportedly from
12123 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12126 Also deactivate the remaining ciphersuites from
12127 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12128 unofficial, and the ID has long expired.
12132 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12133 dual-core machines) and other potential thread-safety issues.
12137 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12139 * Adapt fipsld and the build system to link against the validated FIPS
12140 module in FIPS mode.
12144 * Fixes for VC++ 2005 build under Windows.
12148 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12149 from a Windows bash shell such as MSYS. It is autodetected from the
12150 "config" script when run from a VC++ environment. Modify standard VC++
12151 build to use fipscanister.o from the GNU make build.
12155 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12157 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12158 The value now differs depending on if you build for FIPS or not.
12159 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12160 safely run with a non-FIPSed libcrypto, as it may crash because of
12161 the difference induced by this change.
12165 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12167 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12168 (part of SSL_OP_ALL). This option used to disable the
12169 countermeasure against man-in-the-middle protocol-version
12170 rollback in the SSL 2.0 server implementation, which is a bad
12171 idea. ([CVE-2005-2969])
12173 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12174 for Information Security, National Institute of Advanced Industrial
12175 Science and Technology [AIST, Japan)]*
12177 * Minimal support for X9.31 signatures and PSS padding modes. This is
12178 mainly for FIPS compliance and not fully integrated at this stage.
12182 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12183 the exponentiation using a fixed-length exponent. (Otherwise,
12184 the information leaked through timing could expose the secret key
12185 after many signatures; cf. Bleichenbacher's attack on DSA with
12190 * Make a new fixed-window mod_exp implementation the default for
12191 RSA, DSA, and DH private-key operations so that the sequence of
12192 squares and multiplies and the memory access pattern are
12193 independent of the particular secret key. This will mitigate
12194 cache-timing and potential related attacks.
12196 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12197 and this is automatically used by BN_mod_exp_mont() if the new flag
12198 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12199 will use this BN flag for private exponents unless the flag
12200 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12201 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12203 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12205 * Change the client implementation for SSLv23_method() and
12206 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12207 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12208 (Previously, the SSL 2.0 backwards compatible Client Hello
12209 message format would be used even with SSL_OP_NO_SSLv2.)
12213 * Add support for smime-type MIME parameter in S/MIME messages which some
12218 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12219 a threadsafe manner. Modify rsa code to use new function and add calls
12220 to dsa and dh code (which had race conditions before).
12224 * Include the fixed error library code in the C error file definitions
12225 instead of fixing them up at runtime. This keeps the error code
12226 structures constant.
12230 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12232 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12235 * Fixes for newer kerberos headers. NB: the casts are needed because
12236 the 'length' field is signed on one version and unsigned on another
12237 with no (?) obvious way to tell the difference, without these VC++
12238 complains. Also the "definition" of FAR (blank) is no longer included
12239 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12240 some needed definitions.
12244 * Undo Cygwin change.
12248 * Added support for proxy certificates according to RFC 3820.
12249 Because they may be a security thread to unaware applications,
12250 they must be explicitly allowed in run-time. See
12251 docs/HOWTO/proxy_certificates.txt for further information.
12255 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12257 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12258 server and client random values. Previously
12259 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12260 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12262 This change has negligible security impact because:
12264 1. Server and client random values still have 24 bytes of pseudo random
12267 2. Server and client random values are sent in the clear in the initial
12270 3. The master secret is derived using the premaster secret (48 bytes in
12271 size for static RSA ciphersuites) as well as client server and random
12274 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12277 *Stephen Henson, reported by UK NISCC*
12279 * Use Windows randomness collection on Cygwin.
12283 * Fix hang in EGD/PRNGD query when communication socket is closed
12284 prematurely by EGD/PRNGD.
12286 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12288 * Prompt for pass phrases when appropriate for PKCS12 input format.
12292 * Back-port of selected performance improvements from development
12293 branch, as well as improved support for PowerPC platforms.
12297 * Add lots of checks for memory allocation failure, error codes to indicate
12298 failure and freeing up memory if a failure occurs.
12300 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12302 * Add new -passin argument to dgst.
12306 * Perform some character comparisons of different types in X509_NAME_cmp:
12307 this is needed for some certificates that re-encode DNs into UTF8Strings
12308 (in violation of RFC3280) and can't or won't issue name rollover
12313 * Make an explicit check during certificate validation to see that
12314 the CA setting in each certificate on the chain is correct. As a
12315 side effect always do the following basic checks on extensions,
12316 not just when there's an associated purpose to the check:
12318 - if there is an unhandled critical extension (unless the user
12319 has chosen to ignore this fault)
12320 - if the path length has been exceeded (if one is set at all)
12321 - that certain extensions fit the associated purpose (if one has
12326 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12328 * Avoid a race condition when CRLs are checked in a multi threaded
12329 environment. This would happen due to the reordering of the revoked
12330 entries during signature checking and serial number lookup. Now the
12331 encoding is cached and the serial number sort performed under a lock.
12332 Add new STACK function sk_is_sorted().
12336 * Add Delta CRL to the extension code.
12340 * Various fixes to s3_pkt.c so alerts are sent properly.
12342 *David Holmes <d.holmes@f5.com>*
12344 * Reduce the chances of duplicate issuer name and serial numbers (in
12345 violation of RFC3280) using the OpenSSL certificate creation utilities.
12346 This is done by creating a random 64 bit value for the initial serial
12347 number when a serial number file is created or when a self signed
12348 certificate is created using 'openssl req -x509'. The initial serial
12349 number file is created using 'openssl x509 -next_serial' in CA.pl
12350 rather than being initialized to 1.
12354 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12356 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12357 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12359 *Joe Orton, Steve Henson*
12361 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12364 *Joe Orton, Steve Henson*
12366 * Make it possible to have multiple active certificates with the same
12367 subject in the CA index file. This is done only if the keyword
12368 'unique_subject' is set to 'no' in the main CA section (default
12369 if 'CA_default') of the configuration file. The value is saved
12370 with the database itself in a separate index attribute file,
12371 named like the index file with '.attr' appended to the name.
12375 * X509 verify fixes. Disable broken certificate workarounds when
12376 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12377 keyUsage extension present. Don't accept CRLs with unhandled critical
12378 extensions: since verify currently doesn't process CRL extensions this
12379 rejects a CRL with *any* critical extensions. Add new verify error codes
12384 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12385 A clarification of RFC2560 will require the use of OCTET STRINGs and
12386 some implementations cannot handle the current raw format. Since OpenSSL
12387 copies and compares OCSP nonces as opaque blobs without any attempt at
12388 parsing them this should not create any compatibility issues.
12392 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12393 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12394 this HMAC (and other) operations are several times slower than OpenSSL
12399 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12401 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12403 * Use the correct content when signing type "other".
12407 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12409 * Fix various bugs revealed by running the NISCC test suite:
12411 Stop out of bounds reads in the ASN1 code when presented with
12412 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12414 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12416 If verify callback ignores invalid public key errors don't try to check
12417 certificate signature with the NULL public key.
12421 * New -ignore_err option in ocsp application to stop the server
12422 exiting on the first error in a request.
12426 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12427 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12432 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12433 extra data after the compression methods not only for TLS 1.0
12434 but also for SSL 3.0 (as required by the specification).
12436 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12438 * Change X509_certificate_type() to mark the key as exported/exportable
12439 when it's 512 *bits* long, not 512 bytes.
12443 * Change AES_cbc_encrypt() so it outputs exact multiple of
12444 blocks during encryption.
12448 * Various fixes to base64 BIO and non blocking I/O. On write
12449 flushes were not handled properly if the BIO retried. On read
12450 data was not being buffered properly and had various logic bugs.
12451 This also affects blocking I/O when the data being decoded is a
12456 * Various S/MIME bugfixes and compatibility changes:
12457 output correct application/pkcs7 MIME type if
12458 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12459 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12460 of files as .eml work). Correctly handle very long lines in MIME
12465 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12467 * Countermeasure against the Klima-Pokorny-Rosa extension of
12468 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12469 a protocol version number mismatch like a decryption error
12470 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12474 * Turn on RSA blinding by default in the default implementation
12475 to avoid a timing attack. Applications that don't want it can call
12476 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12477 They would be ill-advised to do so in most cases.
12479 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12481 * Change RSA blinding code so that it works when the PRNG is not
12482 seeded (in this case, the secret RSA exponent is abused as
12483 an unpredictable seed -- if it is not unpredictable, there
12484 is no point in blinding anyway). Make RSA blinding thread-safe
12485 by remembering the creator's thread ID in rsa->blinding and
12486 having all other threads use local one-time blinding factors
12487 (this requires more computation than sharing rsa->blinding, but
12488 avoids excessive locking; and if an RSA object is not shared
12489 between threads, blinding will still be very fast).
12493 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12494 ENGINE as defaults for all supported algorithms irrespective of
12495 the 'flags' parameter. 'flags' is now honoured, so applications
12496 should make sure they are passing it correctly.
12500 * Target "mingw" now allows native Windows code to be generated in
12501 the Cygwin environment as well as with the MinGW compiler.
12505 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12507 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12508 via timing by performing a MAC computation even if incorrect
12509 block cipher padding has been found. This is a countermeasure
12510 against active attacks where the attacker has to distinguish
12511 between bad padding and a MAC verification error. ([CVE-2003-0078])
12513 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12514 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12515 Martin Vuagnoux (EPFL, Ilion)*
12517 * Make the no-err option work as intended. The intention with no-err
12518 is not to have the whole error stack handling routines removed from
12519 libcrypto, it's only intended to remove all the function name and
12520 reason texts, thereby removing some of the footprint that may not
12521 be interesting if those errors aren't displayed anyway.
12523 NOTE: it's still possible for any application or module to have its
12524 own set of error texts inserted. The routines are there, just not
12525 used by default when no-err is given.
12529 * Add support for FreeBSD on IA64.
12531 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12533 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12534 Kerberos function mit_des_cbc_cksum(). Before this change,
12535 the value returned by DES_cbc_cksum() was like the one from
12536 mit_des_cbc_cksum(), except the bytes were swapped.
12538 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12540 * Allow an application to disable the automatic SSL chain building.
12541 Before this a rather primitive chain build was always performed in
12542 ssl3_output_cert_chain(): an application had no way to send the
12543 correct chain if the automatic operation produced an incorrect result.
12545 Now the chain builder is disabled if either:
12547 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12549 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12551 The reasoning behind this is that an application would not want the
12552 auto chain building to take place if extra chain certificates are
12553 present and it might also want a means of sending no additional
12554 certificates (for example the chain has two certificates and the
12559 * Add the possibility to build without the ENGINE framework.
12561 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12563 * Under Win32 gmtime() can return NULL: check return value in
12564 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12568 * DSA routines: under certain error conditions uninitialized BN objects
12569 could be freed. Solution: make sure initialization is performed early
12570 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12571 Nils Larsch <nla@trustcenter.de> via PR#459)
12575 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12576 checked on reconnect on the client side, therefore session resumption
12577 could still fail with a "ssl session id is different" error. This
12578 behaviour is masked when SSL_OP_ALL is used due to
12579 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12580 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12581 followup to PR #377.
12585 * IA-32 assembler support enhancements: unified ELF targets, support
12586 for SCO/Caldera platforms, fix for Cygwin shared build.
12590 * Add support for FreeBSD on sparc64. As a consequence, support for
12591 FreeBSD on non-x86 processors is separate from x86 processors on
12592 the config script, much like the NetBSD support.
12594 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12596 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12598 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12601 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12602 code (06) was taken as the first octet of the session ID and the last
12603 octet was ignored consequently. As a result SSLv2 client side session
12604 caching could not have worked due to the session ID mismatch between
12606 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12611 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12612 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12617 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12618 seems that in spite of existing for more than a year, many application
12619 author have done nothing to provide the necessary callbacks, which
12620 means that this particular engine will not work properly anywhere.
12621 This is a very unfortunate situation which forces us, in the name
12622 of usability, to give the hw_ncipher.c a static lock, which is part
12624 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12625 appear in 0.9.8 or later. We EXPECT application authors to have
12626 dealt properly with this when 0.9.8 is released (unless we actually
12627 make such changes in the libcrypto locking code that changes will
12628 have to be made anyway).
12632 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12633 octets have been read, EOF or an error occurs. Without this change
12634 some truncated ASN1 structures will not produce an error.
12638 * Disable Heimdal support, since it hasn't been fully implemented.
12639 Still give the possibility to force the use of Heimdal, but with
12640 warnings and a request that patches get sent to openssl-dev.
12644 * Add the VC-CE target, introduce the WINCE sysname, and add
12645 INSTALL.WCE and appropriate conditionals to make it build.
12647 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12649 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12650 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12651 edit numbers of the version.
12653 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12655 * Introduce safe string copy and catenation functions
12656 (BUF_strlcpy() and BUF_strlcat()).
12658 *Ben Laurie (CHATS) and Richard Levitte*
12660 * Avoid using fixed-size buffers for one-line DNs.
12662 *Ben Laurie (CHATS)*
12664 * Add BUF_MEM_grow_clean() to avoid information leakage when
12665 resizing buffers containing secrets, and use where appropriate.
12667 *Ben Laurie (CHATS)*
12669 * Avoid using fixed size buffers for configuration file location.
12671 *Ben Laurie (CHATS)*
12673 * Avoid filename truncation for various CA files.
12675 *Ben Laurie (CHATS)*
12677 * Use sizeof in preference to magic numbers.
12679 *Ben Laurie (CHATS)*
12681 * Avoid filename truncation in cert requests.
12683 *Ben Laurie (CHATS)*
12685 * Add assertions to check for (supposedly impossible) buffer
12688 *Ben Laurie (CHATS)*
12690 * Don't cache truncated DNS entries in the local cache (this could
12691 potentially lead to a spoofing attack).
12693 *Ben Laurie (CHATS)*
12695 * Fix various buffers to be large enough for hex/decimal
12696 representations in a platform independent manner.
12698 *Ben Laurie (CHATS)*
12700 * Add CRYPTO_realloc_clean() to avoid information leakage when
12701 resizing buffers containing secrets, and use where appropriate.
12703 *Ben Laurie (CHATS)*
12705 * Add BIO_indent() to avoid much slightly worrying code to do
12708 *Ben Laurie (CHATS)*
12710 * Convert sprintf()/BIO_puts() to BIO_printf().
12712 *Ben Laurie (CHATS)*
12714 * buffer_gets() could terminate with the buffer only half
12717 *Ben Laurie (CHATS)*
12719 * Add assertions to prevent user-supplied crypto functions from
12720 overflowing internal buffers by having large block sizes, etc.
12722 *Ben Laurie (CHATS)*
12724 * New OPENSSL_assert() macro (similar to assert(), but enabled
12727 *Ben Laurie (CHATS)*
12729 * Eliminate unused copy of key in RC4.
12731 *Ben Laurie (CHATS)*
12733 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12735 *Ben Laurie (CHATS)*
12737 * Fix off-by-one error in EGD path.
12739 *Ben Laurie (CHATS)*
12741 * If RANDFILE path is too long, ignore instead of truncating.
12743 *Ben Laurie (CHATS)*
12745 * Eliminate unused and incorrectly sized X.509 structure
12748 *Ben Laurie (CHATS)*
12750 * Eliminate unused and dangerous function knumber().
12752 *Ben Laurie (CHATS)*
12754 * Eliminate unused and dangerous structure, KSSL_ERR.
12756 *Ben Laurie (CHATS)*
12758 * Protect against overlong session ID context length in an encoded
12759 session object. Since these are local, this does not appear to be
12762 *Ben Laurie (CHATS)*
12764 * Change from security patch (see 0.9.6e below) that did not affect
12765 the 0.9.6 release series:
12767 Remote buffer overflow in SSL3 protocol - an attacker could
12768 supply an oversized master key in Kerberos-enabled versions.
12771 *Ben Laurie (CHATS)*
12773 * Change the SSL kerb5 codes to match RFC 2712.
12777 * Make -nameopt work fully for req and add -reqopt switch.
12779 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12781 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12783 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12785 * Make sure tests can be performed even if the corresponding algorithms
12786 have been removed entirely. This was also the last step to make
12787 OpenSSL compilable with DJGPP under all reasonable conditions.
12789 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12791 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12792 to allow version independent disabling of normally unselected ciphers,
12793 which may be activated as a side-effect of selecting a single cipher.
12795 (E.g., cipher list string "RSA" enables ciphersuites that are left
12796 out of "ALL" because they do not provide symmetric encryption.
12797 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12799 *Lutz Jaenicke, Bodo Moeller*
12801 * Add appropriate support for separate platform-dependent build
12802 directories. The recommended way to make a platform-dependent
12803 build directory is the following (tested on Linux), maybe with
12806 # Place yourself outside of the OpenSSL source tree. In
12807 # this example, the environment variable OPENSSL_SOURCE
12808 # is assumed to contain the absolute OpenSSL source directory.
12809 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12810 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12811 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12812 mkdir -p `dirname $F`
12813 ln -s $OPENSSL_SOURCE/$F $F
12816 To be absolutely sure not to disturb the source tree, a "make clean"
12817 is a good thing. If it isn't successful, don't worry about it,
12818 it probably means the source directory is very clean.
12822 * Make sure any ENGINE control commands make local copies of string
12823 pointers passed to them whenever necessary. Otherwise it is possible
12824 the caller may have overwritten (or deallocated) the original string
12825 data when a later ENGINE operation tries to use the stored values.
12827 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12829 * Improve diagnostics in file reading and command-line digests.
12831 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12833 * Add AES modes CFB and OFB to the object database. Correct an
12834 error in AES-CFB decryption.
12838 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12839 allows existing EVP_CIPHER_CTX structures to be reused after
12840 calling `EVP_*Final()`. This behaviour is used by encryption
12841 BIOs and some applications. This has the side effect that
12842 applications must explicitly clean up cipher contexts with
12843 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12847 * Check the values of dna and dnb in bn_mul_recursive before calling
12848 bn_mul_comba (a non zero value means the a or b arrays do not contain
12849 n2 elements) and fallback to bn_mul_normal if either is not zero.
12853 * Fix escaping of non-ASCII characters when using the -subj option
12854 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12858 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12859 form for "surname", serialNumber has no short form.
12860 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12861 therefore remove "mail" short name for "internet 7".
12862 The OID for unique identifiers in X509 certificates is
12863 x500UniqueIdentifier, not uniqueIdentifier.
12864 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12868 * Add an "init" command to the ENGINE config module and auto initialize
12869 ENGINEs. Without any "init" command the ENGINE will be initialized
12870 after all ctrl commands have been executed on it. If init=1 the
12871 ENGINE is initialized at that point (ctrls before that point are run
12872 on the uninitialized ENGINE and after on the initialized one). If
12873 init=0 then the ENGINE will not be initialized at all.
12877 * Fix the 'app_verify_callback' interface so that the user-defined
12878 argument is actually passed to the callback: In the
12879 SSL_CTX_set_cert_verify_callback() prototype, the callback
12880 declaration has been changed from
12883 int (*cb)(X509_STORE_CTX *,void *);
12884 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12885 i=s->ctx->app_verify_callback(&ctx)
12886 has been changed into
12887 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12889 To update applications using SSL_CTX_set_cert_verify_callback(),
12890 a dummy argument can be added to their callback functions.
12892 *D. K. Smetters <smetters@parc.xerox.com>*
12894 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12896 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12898 * Add and OPENSSL_LOAD_CONF define which will cause
12899 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12900 This allows older applications to transparently support certain
12901 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12902 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12903 load the config file and OPENSSL_add_all_algorithms_conf() which will
12904 always load it have also been added.
12908 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12909 Adjust NIDs and EVP layer.
12911 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12913 * Config modules support in openssl utility.
12915 Most commands now load modules from the config file,
12916 though in a few (such as version) this isn't done
12917 because it couldn't be used for anything.
12919 In the case of ca and req the config file used is
12920 the same as the utility itself: that is the -config
12921 command line option can be used to specify an
12926 * Move default behaviour from OPENSSL_config(). If appname is NULL
12927 use "openssl_conf" if filename is NULL use default openssl config file.
12931 * Add an argument to OPENSSL_config() to allow the use of an alternative
12932 config section name. Add a new flag to tolerate a missing config file
12933 and move code to CONF_modules_load_file().
12937 * Support for crypto accelerator cards from Accelerated Encryption
12938 Processing, www.aep.ie. (Use engine 'aep')
12939 The support was copied from 0.9.6c [engine] and adapted/corrected
12940 to work with the new engine framework.
12942 *AEP Inc. and Richard Levitte*
12944 * Support for SureWare crypto accelerator cards from Baltimore
12945 Technologies. (Use engine 'sureware')
12946 The support was copied from 0.9.6c [engine] and adapted
12947 to work with the new engine framework.
12951 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12952 make the newer ENGINE framework commands for the CHIL engine work.
12954 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12956 * Make it possible to produce shared libraries on ReliantUNIX.
12958 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12960 * Add the configuration target debug-linux-ppro.
12961 Make 'openssl rsa' use the general key loading routines
12962 implemented in `apps.c`, and make those routines able to
12963 handle the key format FORMAT_NETSCAPE and the variant
12966 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12968 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12970 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12972 * Add -keyform to rsautl, and document -engine.
12974 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12976 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12977 BIO_R_NO_SUCH_FILE error code rather than the generic
12978 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12982 * Add new functions
12983 ERR_peek_last_error
12984 ERR_peek_last_error_line
12985 ERR_peek_last_error_line_data.
12986 These are similar to
12988 ERR_peek_error_line
12989 ERR_peek_error_line_data,
12990 but report on the latest error recorded rather than the first one
12991 still in the error queue.
12993 *Ben Laurie, Bodo Moeller*
12995 * default_algorithms option in ENGINE config module. This allows things
12997 default_algorithms = ALL
12998 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13002 * Preliminary ENGINE config module.
13006 * New experimental application configuration code.
13010 * Change the AES code to follow the same name structure as all other
13011 symmetric ciphers, and behave the same way. Move everything to
13012 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13014 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13016 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13018 *Ben Laurie and Theo de Raadt*
13020 * Add option to output public keys in req command.
13022 *Massimiliano Pala madwolf@openca.org*
13024 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13025 (up to about 10% better than before for P-192 and P-224).
13029 * New functions/macros
13031 SSL_CTX_set_msg_callback(ctx, cb)
13032 SSL_CTX_set_msg_callback_arg(ctx, arg)
13033 SSL_set_msg_callback(ssl, cb)
13034 SSL_set_msg_callback_arg(ssl, arg)
13036 to request calling a callback function
13038 void cb(int write_p, int version, int content_type,
13039 const void *buf, size_t len, SSL *ssl, void *arg)
13041 whenever a protocol message has been completely received
13042 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13043 protocol version according to which the SSL library interprets
13044 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13045 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13046 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13047 specification (change_cipher_spec(20), alert(21), handshake(22)).
13048 'buf' and 'len' point to the actual message, 'ssl' to the
13049 SSL object, and 'arg' is the application-defined value set by
13050 SSL[_CTX]_set_msg_callback_arg().
13052 'openssl s_client' and 'openssl s_server' have new '-msg' options
13053 to enable a callback that displays all protocol messages.
13057 * Change the shared library support so shared libraries are built as
13058 soon as the corresponding static library is finished, and thereby get
13059 openssl and the test programs linked against the shared library.
13060 This still only happens when the keyword "shard" has been given to
13061 the configuration scripts.
13063 NOTE: shared library support is still an experimental thing, and
13064 backward binary compatibility is still not guaranteed.
13066 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13068 * Add support for Subject Information Access extension.
13070 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13072 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13073 additional bytes when new memory had to be allocated, not just
13074 when reusing an existing buffer.
13078 * New command line and configuration option 'utf8' for the req command.
13079 This allows field values to be specified as UTF8 strings.
13083 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13084 runs for the former and machine-readable output for the latter.
13088 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13089 of the e-mail address in the DN (i.e., it will go into a certificate
13090 extension only). The new configuration file option 'email_in_dn = no'
13091 has the same effect.
13093 *Massimiliano Pala madwolf@openca.org*
13095 * Change all functions with names starting with `des_` to be starting
13096 with `DES_` instead. Add wrappers that are compatible with libdes,
13097 but are named `_ossl_old_des_*`. Finally, add macros that map the
13098 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13099 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13100 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13103 Since we provide two compatibility mappings, the user needs to
13104 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13105 compatibility is desired. The default (i.e., when that macro
13106 isn't defined) is OpenSSL 0.9.6c compatibility.
13108 There are also macros that enable and disable the support of old
13109 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13110 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13111 are defined, the default will apply: to support the old des routines.
13113 In either case, one must include openssl/des.h to get the correct
13114 definitions. Do not try to just include openssl/des_old.h, that
13117 NOTE: This is a major break of an old API into a new one. Software
13118 authors are encouraged to switch to the `DES_` style functions. Some
13119 time in the future, des_old.h and the libdes compatibility functions
13120 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13121 default), and then completely removed.
13125 * Test for certificates which contain unsupported critical extensions.
13126 If such a certificate is found during a verify operation it is
13127 rejected by default: this behaviour can be overridden by either
13128 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13129 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13130 X509_supported_extension() has also been added which returns 1 if a
13131 particular extension is supported.
13135 * Modify the behaviour of EVP cipher functions in similar way to digests
13136 to retain compatibility with existing code.
13140 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13141 compatibility with existing code. In particular the 'ctx' parameter does
13142 not have to be to be initialized before the call to EVP_DigestInit() and
13143 it is tidied up after a call to EVP_DigestFinal(). New function
13144 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13145 EVP_MD_CTX_copy() changed to not require the destination to be
13146 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13147 requires the destination to be valid.
13149 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13150 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13154 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13155 so that complete 'Handshake' protocol structures are kept in memory
13156 instead of overwriting 'msg_type' and 'length' with 'body' data.
13160 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13162 *Massimo Santin via Richard Levitte*
13164 * Major restructuring to the underlying ENGINE code. This includes
13165 reduction of linker bloat, separation of pure "ENGINE" manipulation
13166 (initialisation, etc) from functionality dealing with implementations
13167 of specific crypto interfaces. This change also introduces integrated
13168 support for symmetric ciphers and digest implementations - so ENGINEs
13169 can now accelerate these by providing EVP_CIPHER and EVP_MD
13170 implementations of their own. This is detailed in
13171 [crypto/engine/README.md](crypto/engine/README.md)
13172 as it couldn't be adequately described here. However, there are a few
13173 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13174 were changed in the original introduction of ENGINE code have now
13175 reverted back - the hooking from this code to ENGINE is now a good
13176 deal more passive and at run-time, operations deal directly with
13177 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13178 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13179 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13180 they were not being used by the framework as there is no concept of a
13181 BIGNUM_METHOD and they could not be generalised to the new
13182 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13183 ENGINE_cpy() has been removed as it cannot be consistently defined in
13188 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13192 * Change mkdef.pl to sort symbols that get the same entry number,
13193 and make sure the automatically generated functions `ERR_load_*`
13194 become part of libeay.num as well.
13198 * New function SSL_renegotiate_pending(). This returns true once
13199 renegotiation has been requested (either SSL_renegotiate() call
13200 or HelloRequest/ClientHello received from the peer) and becomes
13201 false once a handshake has been completed.
13202 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13203 sends a HelloRequest, but does not ensure that a handshake takes
13204 place. SSL_renegotiate_pending() is useful for checking if the
13205 client has followed the request.)
13209 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13210 By default, clients may request session resumption even during
13211 renegotiation (if session ID contexts permit); with this option,
13212 session resumption is possible only in the first handshake.
13214 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13215 more bits available for options that should not be part of
13216 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13220 * Add some demos for certificate and certificate request creation.
13224 * Make maximum certificate chain size accepted from the peer application
13225 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13226 "Douglas E. Engert" <deengert@anl.gov>.
13230 * Add support for shared libraries for Unixware-7
13231 (Boyd Lynn Gerber <gerberb@zenez.com>).
13235 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13236 be done prior to destruction. Use this to unload error strings from
13237 ENGINEs that load their own error strings. NB: This adds two new API
13238 functions to "get" and "set" this destroy handler in an ENGINE.
13242 * Alter all existing ENGINE implementations (except "openssl" and
13243 "openbsd") to dynamically instantiate their own error strings. This
13244 makes them more flexible to be built both as statically-linked ENGINEs
13245 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13246 Also, add stub code to each that makes building them as self-contained
13247 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13251 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13252 implementations into applications that are completely implemented in
13253 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13254 commands that can be used to configure what shared-library to load and
13255 to control aspects of the way it is handled. Also, made an update to
13256 the [README-Engine.md](README-Engine.md) file
13257 that brings its information up-to-date and
13258 provides some information and instructions on the "dynamic" ENGINE
13259 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13263 * Make it possible to unload ranges of ERR strings with a new
13264 "ERR_unload_strings" function.
13268 * Add a copy() function to EVP_MD.
13272 * Make EVP_MD routines take a context pointer instead of just the
13273 md_data void pointer.
13277 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13278 that the digest can only process a single chunk of data
13279 (typically because it is provided by a piece of
13280 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13281 is only going to provide a single chunk of data, and hence the
13282 framework needn't accumulate the data for oneshot drivers.
13286 * As with "ERR", make it possible to replace the underlying "ex_data"
13287 functions. This change also alters the storage and management of global
13288 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13289 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13290 index counters. The API functions that use this state have been changed
13291 to take a "class_index" rather than pointers to the class's local STACK
13292 and counter, and there is now an API function to dynamically create new
13293 classes. This centralisation allows us to (a) plug a lot of the
13294 thread-safety problems that existed, and (b) makes it possible to clean
13295 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13296 such data would previously have always leaked in application code and
13297 workarounds were in place to make the memory debugging turn a blind eye
13298 to it. Application code that doesn't use this new function will still
13299 leak as before, but their memory debugging output will announce it now
13300 rather than letting it slide.
13302 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13303 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13304 has a return value to indicate success or failure.
13308 * Make it possible to replace the underlying "ERR" functions such that the
13309 global state (2 LHASH tables and 2 locks) is only used by the "default"
13310 implementation. This change also adds two functions to "get" and "set"
13311 the implementation prior to it being automatically set the first time
13312 any other ERR function takes place. Ie. an application can call "get",
13313 pass the return value to a module it has just loaded, and that module
13314 can call its own "set" function using that value. This means the
13315 module's "ERR" operations will use (and modify) the error state in the
13316 application and not in its own statically linked copy of OpenSSL code.
13320 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13321 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13322 the operation, and provides a more encapsulated way for external code
13323 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13324 to use these functions rather than manually incrementing the counts.
13326 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13330 * Add EVP test program.
13334 * Add symmetric cipher support to ENGINE. Expect the API to change!
13338 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13339 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13340 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13341 These allow a CRL to be built without having to access X509_CRL fields
13342 directly. Modify 'ca' application to use new functions.
13346 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13347 bug workarounds. Rollback attack detection is a security feature.
13348 The problem will only arise on OpenSSL servers when TLSv1 is not
13349 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13350 Software authors not wanting to support TLSv1 will have special reasons
13351 for their choice and can explicitly enable this option.
13353 *Bodo Moeller, Lutz Jaenicke*
13355 * Rationalise EVP so it can be extended: don't include a union of
13356 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13357 (similar to those existing for EVP_CIPHER_CTX).
13362 EVP_MD_CTX_init(&md); /* new function call */
13363 EVP_DigestInit(&md, EVP_sha1());
13364 EVP_DigestUpdate(&md, in, len);
13365 EVP_DigestFinal(&md, out, NULL);
13366 EVP_MD_CTX_cleanup(&md); /* new function call */
13370 * Make DES key schedule conform to the usual scheme, as well as
13371 correcting its structure. This means that calls to DES functions
13372 now have to pass a pointer to a des_key_schedule instead of a
13373 plain des_key_schedule (which was actually always a pointer
13376 des_key_schedule ks;
13378 des_set_key_checked(..., &ks);
13379 des_ncbc_encrypt(..., &ks, ...);
13381 (Note that a later change renames 'des_...' into 'DES_...'.)
13385 * Initial reduction of linker bloat: the use of some functions, such as
13386 PEM causes large amounts of unused functions to be linked in due to
13387 poor organisation. For example pem_all.c contains every PEM function
13388 which has a knock on effect of linking in large amounts of (unused)
13389 ASN1 code. Grouping together similar functions and splitting unrelated
13390 functions prevents this.
13394 * Cleanup of EVP macros.
13398 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13399 correct `_ecb suffix`.
13403 * Add initial OCSP responder support to ocsp application. The
13404 revocation information is handled using the text based index
13405 use by the ca application. The responder can either handle
13406 requests generated internally, supplied in files (for example
13407 via a CGI script) or using an internal minimal server.
13411 * Add configuration choices to get zlib compression for TLS.
13415 * Changes to Kerberos SSL for RFC 2712 compliance:
13416 1. Implemented real KerberosWrapper, instead of just using
13417 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13418 2. Implemented optional authenticator field of KerberosWrapper.
13420 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13421 and authenticator structs; see crypto/krb5/.
13423 Generalized Kerberos calls to support multiple Kerberos libraries.
13424 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13425 via Richard Levitte*
13427 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13428 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13429 values for each of the key sizes rather than having just
13430 parameters (and 'speed' generating keys each time).
13434 * Speed up EVP routines.
13437 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13438 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13439 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13440 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13442 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13443 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13444 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13447 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13449 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13453 * Added the OS2-EMX target.
13455 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13457 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13458 New functions to support `NCONF` routines in extension code.
13459 New function `CONF_set_nconf()`
13460 to allow functions which take an `NCONF` to also handle the old `LHASH`
13461 structure: this means that the old `CONF` compatible routines can be
13462 retained (in particular w.rt. extensions) without having to duplicate the
13463 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13467 * Enhance the general user interface with mechanisms for inner control
13468 and with possibilities to have yes/no kind of prompts.
13472 * Change all calls to low-level digest routines in the library and
13473 applications to use EVP. Add missing calls to HMAC_cleanup() and
13474 don't assume HMAC_CTX can be copied using memcpy().
13476 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13478 * Add the possibility to control engines through control names but with
13479 arbitrary arguments instead of just a string.
13480 Change the key loaders to take a UI_METHOD instead of a callback
13481 function pointer. NOTE: this breaks binary compatibility with earlier
13482 versions of OpenSSL [engine].
13483 Adapt the nCipher code for these new conditions and add a card insertion
13488 * Enhance the general user interface with mechanisms to better support
13489 dialog box interfaces, application-defined prompts, the possibility
13490 to use defaults (for example default passwords from somewhere else)
13491 and interrupts/cancellations.
13495 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13496 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13500 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13501 tidy up some unnecessarily weird code in 'sk_new()').
13503 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13505 * Change the key loading routines for ENGINEs to use the same kind
13506 callback (pem_password_cb) as all other routines that need this
13511 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13512 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13513 than this minimum value is recommended.
13517 * New random seeder for OpenVMS, using the system process statistics
13518 that are easily reachable.
13522 * Windows apparently can't transparently handle global
13523 variables defined in DLLs. Initialisations such as:
13525 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13527 won't compile. This is used by the any applications that need to
13528 declare their own ASN1 modules. This was fixed by adding the option
13529 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13530 needed for static libraries under Win32.
13534 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13535 setting of purpose and trust fields. New X509_STORE trust and
13536 purpose functions and tidy up setting in other SSL functions.
13540 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13541 structure. These are inherited by X509_STORE_CTX when it is
13542 initialised. This allows various defaults to be set in the
13543 X509_STORE structure (such as flags for CRL checking and custom
13544 purpose or trust settings) for functions which only use X509_STORE_CTX
13545 internally such as S/MIME.
13547 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13548 trust settings if they are not set in X509_STORE. This allows X509_STORE
13549 purposes and trust (in S/MIME for example) to override any set by default.
13551 Add command line options for CRL checking to smime, s_client and s_server
13556 * Initial CRL based revocation checking. If the CRL checking flag(s)
13557 are set then the CRL is looked up in the X509_STORE structure and
13558 its validity and signature checked, then if the certificate is found
13559 in the CRL the verify fails with a revoked error.
13561 Various new CRL related callbacks added to X509_STORE_CTX structure.
13563 Command line options added to 'verify' application to support this.
13565 This needs some additional work, such as being able to handle multiple
13566 CRLs with different times, extension based lookup (rather than just
13567 by subject name) and ultimately more complete V2 CRL extension
13572 * Add a general user interface API (crypto/ui/). This is designed
13573 to replace things like des_read_password and friends (backward
13574 compatibility functions using this new API are provided).
13575 The purpose is to remove prompting functions from the DES code
13576 section as well as provide for prompting through dialog boxes in
13577 a window system and the like.
13581 * Add "ex_data" support to ENGINE so implementations can add state at a
13582 per-structure level rather than having to store it globally.
13586 * Make it possible for ENGINE structures to be copied when retrieved by
13587 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13588 This causes the "original" ENGINE structure to act like a template,
13589 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13590 operational state can be localised to each ENGINE structure, despite the
13591 fact they all share the same "methods". New ENGINE structures returned in
13592 this case have no functional references and the return value is the single
13593 structural reference. This matches the single structural reference returned
13594 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13599 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13600 needs to match any other type at all we need to manually clear the
13605 * Changes to the "openssl engine" utility to include;
13606 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13607 about an ENGINE's available control commands.
13608 - executing control commands from command line arguments using the
13609 '-pre' and '-post' switches. '-post' is only used if '-t' is
13610 specified and the ENGINE is successfully initialised. The syntax for
13611 the individual commands are colon-separated, for example;
13612 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13616 * New dynamic control command support for ENGINEs. ENGINEs can now
13617 declare their own commands (numbers), names (strings), descriptions,
13618 and input types for run-time discovery by calling applications. A
13619 subset of these commands are implicitly classed as "executable"
13620 depending on their input type, and only these can be invoked through
13621 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13622 can be based on user input, config files, etc). The distinction is
13623 that "executable" commands cannot return anything other than a boolean
13624 result and can only support numeric or string input, whereas some
13625 discoverable commands may only be for direct use through
13626 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13627 pointers, or other custom uses. The "executable" commands are to
13628 support parameterisations of ENGINE behaviour that can be
13629 unambiguously defined by ENGINEs and used consistently across any
13630 OpenSSL-based application. Commands have been added to all the
13631 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13632 control over shared-library paths without source code alterations.
13636 * Changed all ENGINE implementations to dynamically allocate their
13637 ENGINEs rather than declaring them statically. Apart from this being
13638 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13639 this also allows the implementations to compile without using the
13640 internal engine_int.h header.
13644 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13645 'const' value. Any code that should be able to modify a RAND_METHOD
13646 should already have non-const pointers to it (ie. they should only
13647 modify their own ones).
13651 * Made a variety of little tweaks to the ENGINE code.
13652 - "atalla" and "ubsec" string definitions were moved from header files
13653 to C code. "nuron" string definitions were placed in variables
13654 rather than hard-coded - allowing parameterisation of these values
13655 later on via ctrl() commands.
13656 - Removed unused "#if 0"'d code.
13657 - Fixed engine list iteration code so it uses ENGINE_free() to release
13658 structural references.
13659 - Constified the RAND_METHOD element of ENGINE structures.
13660 - Constified various get/set functions as appropriate and added
13661 missing functions (including a catch-all ENGINE_cpy that duplicates
13662 all ENGINE values onto a new ENGINE except reference counts/state).
13663 - Removed NULL parameter checks in get/set functions. Setting a method
13664 or function to NULL is a way of cancelling out a previously set
13665 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13666 and doesn't justify the extra error symbols and code.
13667 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13668 flags from engine_int.h to engine.h.
13669 - Changed prototypes for ENGINE handler functions (init(), finish(),
13670 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13674 * Implement binary inversion algorithm for BN_mod_inverse in addition
13675 to the algorithm using long division. The binary algorithm can be
13676 used only if the modulus is odd. On 32-bit systems, it is faster
13677 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13678 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13679 up to 450 bits. In 64-bit environments, the binary algorithm
13680 appears to be advantageous for much longer moduli; here we use it
13681 for moduli up to 2048 bits.
13685 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13686 could not support the combine flag in choice fields.
13690 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13691 extensions from a certificate request to the certificate.
13695 * Allow multiple 'certopt' and 'nameopt' options to be separated
13696 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13697 file: this allows the display of the certificate about to be
13698 signed to be customised, to allow certain fields to be included
13699 or excluded and extension details. The old system didn't display
13700 multicharacter strings properly, omitted fields not in the policy
13701 and couldn't display additional details such as extensions.
13705 * Function EC_POINTs_mul for multiple scalar multiplication
13706 of an arbitrary number of elliptic curve points
13707 \sum scalars[i]*points[i],
13708 optionally including the generator defined for the EC_GROUP:
13709 scalar*generator + \sum scalars[i]*points[i].
13711 EC_POINT_mul is a simple wrapper function for the typical case
13712 that the point list has just one item (besides the optional
13717 * First EC_METHODs for curves over GF(p):
13719 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13720 operations and provides various method functions that can also
13721 operate with faster implementations of modular arithmetic.
13723 EC_GFp_mont_method() reuses most functions that are part of
13724 EC_GFp_simple_method, but uses Montgomery arithmetic.
13726 *Bodo Moeller; point addition and point doubling
13727 implementation directly derived from source code provided by
13728 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13730 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13731 crypto/ec/ec_lib.c):
13733 Curves are EC_GROUP objects (with an optional group generator)
13734 based on EC_METHODs that are built into the library.
13736 Points are EC_POINT objects based on EC_GROUP objects.
13738 Most of the framework would be able to handle curves over arbitrary
13739 finite fields, but as there are no obvious types for fields other
13740 than GF(p), some functions are limited to that for now.
13744 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13745 that the file contains a complete HTTP response.
13749 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13750 change the def and num file printf format specifier from "%-40sXXX"
13751 to "%-39s XXX". The latter will always guarantee a space after the
13752 field while the former will cause them to run together if the field
13753 is 40 of more characters long.
13757 * Constify the cipher and digest 'method' functions and structures
13758 and modify related functions to take constant EVP_MD and EVP_CIPHER
13763 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13764 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13768 * Modify `EVP_Digest*()` routines so they now return values. Although the
13769 internal software routines can never fail additional hardware versions
13774 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13776 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13777 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13780 ERR_R_NESTED_ASN1_ERROR
13782 ERR_R_MISSING_ASN1_EOS
13783 were 4 .. 9, conflicting with
13784 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13786 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13787 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13789 Add new error code 'ERR_R_INTERNAL_ERROR'.
13793 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13798 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13799 sets the subject name for a new request or supersedes the
13800 subject name in a given request. Formats that can be parsed are
13801 'CN=Some Name, OU=myOU, C=IT'
13803 'CN=Some Name/OU=myOU/C=IT'.
13805 Add options '-batch' and '-verbose' to 'openssl req'.
13807 *Massimiliano Pala <madwolf@hackmasters.net>*
13809 * Introduce the possibility to access global variables through
13810 functions on platform were that's the best way to handle exporting
13811 global variables in shared libraries. To enable this functionality,
13812 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13813 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13814 is normally done by Configure or something similar).
13816 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13817 in the source file (foo.c) like this:
13819 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13820 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13822 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13823 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13825 OPENSSL_DECLARE_GLOBAL(int,foo);
13826 #define foo OPENSSL_GLOBAL_REF(foo)
13827 OPENSSL_DECLARE_GLOBAL(double,bar);
13828 #define bar OPENSSL_GLOBAL_REF(bar)
13830 The #defines are very important, and therefore so is including the
13831 header file everywhere where the defined globals are used.
13833 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13834 of ASN.1 items, but that structure is a bit different.
13836 The largest change is in util/mkdef.pl which has been enhanced with
13837 better and easier to understand logic to choose which symbols should
13838 go into the Windows .def files as well as a number of fixes and code
13839 cleanup (among others, algorithm keywords are now sorted
13840 lexicographically to avoid constant rewrites).
13844 * In BN_div() keep a copy of the sign of 'num' before writing the
13845 result to 'rm' because if rm==num the value will be overwritten
13846 and produce the wrong result if 'num' is negative: this caused
13847 problems with BN_mod() and BN_nnmod().
13851 * Function OCSP_request_verify(). This checks the signature on an
13852 OCSP request and verifies the signer certificate. The signer
13853 certificate is just checked for a generic purpose and OCSP request
13858 * Add OCSP_check_validity() function to check the validity of OCSP
13859 responses. OCSP responses are prepared in real time and may only
13860 be a few seconds old. Simply checking that the current time lies
13861 between thisUpdate and nextUpdate max reject otherwise valid responses
13862 caused by either OCSP responder or client clock inaccuracy. Instead
13863 we allow thisUpdate and nextUpdate to fall within a certain period of
13864 the current time. The age of the response can also optionally be
13865 checked. Two new options -validity_period and -status_age added to
13870 * If signature or public key algorithm is unrecognized print out its
13871 OID rather that just UNKNOWN.
13875 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13876 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13877 ID to be generated from the issuer certificate alone which can then be
13878 passed to OCSP_id_issuer_cmp().
13882 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13883 ASN1 modules to export functions returning ASN1_ITEM pointers
13884 instead of the ASN1_ITEM structures themselves. This adds several
13885 new macros which allow the underlying ASN1 function/structure to
13886 be accessed transparently. As a result code should not use ASN1_ITEM
13887 references directly (such as &X509_it) but instead use the relevant
13888 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13889 use of the new ASN1 code on platforms where exporting structures
13890 is problematical (for example in shared libraries) but exporting
13891 functions returning pointers to structures is not.
13895 * Add support for overriding the generation of SSL/TLS session IDs.
13896 These callbacks can be registered either in an SSL_CTX or per SSL.
13897 The purpose of this is to allow applications to control, if they wish,
13898 the arbitrary values chosen for use as session IDs, particularly as it
13899 can be useful for session caching in multiple-server environments. A
13900 command-line switch for testing this (and any client code that wishes
13901 to use such a feature) has been added to "s_server".
13903 *Geoff Thorpe, Lutz Jaenicke*
13905 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13906 of the form `#if defined(...) || defined(...) || ...` and
13907 `#if !defined(...) && !defined(...) && ...`. This also avoids
13908 the growing number of special cases it was previously handling.
13912 * Make all configuration macros available for application by making
13913 sure they are available in opensslconf.h, by giving them names starting
13914 with `OPENSSL_` to avoid conflicts with other packages and by making
13915 sure e_os2.h will cover all platform-specific cases together with
13917 Additionally, it is now possible to define configuration/platform-
13918 specific names (called "system identities"). In the C code, these
13919 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13920 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13921 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13926 * New option -set_serial to 'req' and 'x509' this allows the serial
13927 number to use to be specified on the command line. Previously self
13928 signed certificates were hard coded with serial number 0 and the
13929 CA options of 'x509' had to use a serial number in a file which was
13934 * New options to 'ca' utility to support V2 CRL entry extensions.
13935 Currently CRL reason, invalidity date and hold instruction are
13936 supported. Add new CRL extensions to V3 code and some new objects.
13940 * New function EVP_CIPHER_CTX_set_padding() this is used to
13941 disable standard block padding (aka PKCS#5 padding) in the EVP
13942 API, which was previously mandatory. This means that the data is
13943 not padded in any way and so the total length much be a multiple
13944 of the block size, otherwise an error occurs.
13948 * Initial (incomplete) OCSP SSL support.
13952 * New function OCSP_parse_url(). This splits up a URL into its host,
13953 port and path components: primarily to parse OCSP URLs. New -url
13954 option to ocsp utility.
13958 * New nonce behavior. The return value of OCSP_check_nonce() now
13959 reflects the various checks performed. Applications can decide
13960 whether to tolerate certain situations such as an absent nonce
13961 in a response when one was present in a request: the ocsp application
13962 just prints out a warning. New function OCSP_add1_basic_nonce()
13963 this is to allow responders to include a nonce in a response even if
13964 the request is nonce-less.
13968 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13969 skipped when using openssl x509 multiple times on a single input file,
13970 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13974 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13975 set string type: to handle setting ASN1_TIME structures. Fix ca
13976 utility to correctly initialize revocation date of CRLs.
13980 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13981 the clients preferred ciphersuites and rather use its own preferences.
13982 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13983 Internet Explorer by ensuring unchanged hash method during stepup.
13984 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13988 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13989 to aes and add a new 'exist' option to print out symbols that don't
13994 * Additional options to ocsp utility to allow flags to be set and
13995 additional certificates supplied.
13999 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14000 OCSP client a number of certificate to only verify the response
14005 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14006 handle the new API. Currently only ECB, CBC modes supported. Add new
14009 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14010 Encryption Standard (AES) Ciphersuites for Transport Layer
14011 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14012 not enabled by default and were not part of the "ALL" ciphersuite
14013 alias because they were not yet official; they could be
14014 explicitly requested by specifying the "AESdraft" ciphersuite
14015 group alias. In the final release of OpenSSL 0.9.7, the group
14016 alias is called "AES" and is part of "ALL".)
14018 *Ben Laurie, Steve Henson, Bodo Moeller*
14020 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14021 request to response.
14025 * Functions for OCSP responders. OCSP_request_onereq_count(),
14026 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14027 extract information from a certificate request. OCSP_response_create()
14028 creates a response and optionally adds a basic response structure.
14029 OCSP_basic_add1_status() adds a complete single response to a basic
14030 response and returns the OCSP_SINGLERESP structure just added (to allow
14031 extensions to be included for example). OCSP_basic_add1_cert() adds a
14032 certificate to a basic response and OCSP_basic_sign() signs a basic
14033 response with various flags. New helper functions ASN1_TIME_check()
14034 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14035 (converts ASN1_TIME to GeneralizedTime).
14039 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14040 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14041 structure from a certificate. X509_pubkey_digest() digests the public_key
14042 contents: this is used in various key identifiers.
14046 * Make sk_sort() tolerate a NULL argument.
14048 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14050 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14051 passed by the function are trusted implicitly. If any of them signed the
14052 response then it is assumed to be valid and is not verified.
14056 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14057 to data. This was previously part of the PKCS7 ASN1 code. This
14058 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14059 *Steve Henson, reported by Kenneth R. Robinette
14060 <support@securenetterm.com>*
14062 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14063 routines: without these tracing memory leaks is very painful.
14064 Fix leaks in PKCS12 and PKCS7 routines.
14068 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14069 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14070 effectively meant GeneralizedTime would never be used. Now it
14071 is initialised to -1 but X509_time_adj() now has to check the value
14072 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14073 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14074 *Steve Henson, reported by Kenneth R. Robinette
14075 <support@securenetterm.com>*
14077 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14078 result in a zero length in the ASN1_INTEGER structure which was
14079 not consistent with the structure when d2i_ASN1_INTEGER() was used
14080 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14081 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14082 where it did not print out a minus for negative ASN1_INTEGER.
14086 * Add summary printout to ocsp utility. The various functions which
14087 convert status values to strings have been renamed to:
14088 OCSP_response_status_str(), OCSP_cert_status_str() and
14089 OCSP_crl_reason_str() and are no longer static. New options
14090 to verify nonce values and to disable verification. OCSP response
14091 printout format cleaned up.
14095 * Add additional OCSP certificate checks. These are those specified
14096 in RFC2560. This consists of two separate checks: the CA of the
14097 certificate being checked must either be the OCSP signer certificate
14098 or the issuer of the OCSP signer certificate. In the latter case the
14099 OCSP signer certificate must contain the OCSP signing extended key
14100 usage. This check is performed by attempting to match the OCSP
14101 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14102 in the OCSP_CERTID structures of the response.
14106 * Initial OCSP certificate verification added to OCSP_basic_verify()
14107 and related routines. This uses the standard OpenSSL certificate
14108 verify routines to perform initial checks (just CA validity) and
14109 to obtain the certificate chain. Then additional checks will be
14110 performed on the chain. Currently the root CA is checked to see
14111 if it is explicitly trusted for OCSP signing. This is used to set
14112 a root CA as a global signing root: that is any certificate that
14113 chains to that CA is an acceptable OCSP signing certificate.
14117 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14118 extensions from a separate configuration file.
14119 As when reading extensions from the main configuration file,
14120 the '-extensions ...' option may be used for specifying the
14123 *Massimiliano Pala <madwolf@comune.modena.it>*
14125 * New OCSP utility. Allows OCSP requests to be generated or
14126 read. The request can be sent to a responder and the output
14127 parsed, outputted or printed in text form. Not complete yet:
14128 still needs to check the OCSP response validity.
14132 * New subcommands for 'openssl ca':
14133 `openssl ca -status <serial>` prints the status of the cert with
14134 the given serial number (according to the index file).
14135 `openssl ca -updatedb` updates the expiry status of certificates
14138 *Massimiliano Pala <madwolf@comune.modena.it>*
14140 * New '-newreq-nodes' command option to CA.pl. This is like
14141 '-newreq', but calls 'openssl req' with the '-nodes' option
14142 so that the resulting key is not encrypted.
14144 *Damien Miller <djm@mindrot.org>*
14146 * New configuration for the GNU Hurd.
14148 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14150 * Initial code to implement OCSP basic response verify. This
14151 is currently incomplete. Currently just finds the signer's
14152 certificate and verifies the signature on the response.
14156 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14157 value of OPENSSLDIR. This is available via the new '-d' option
14158 to 'openssl version', and is also included in 'openssl version -a'.
14162 * Allowing defining memory allocation callbacks that will be given
14163 file name and line number information in additional arguments
14164 (a `const char*` and an int). The basic functionality remains, as
14165 well as the original possibility to just replace malloc(),
14166 realloc() and free() by functions that do not know about these
14167 additional arguments. To register and find out the current
14168 settings for extended allocation functions, the following
14169 functions are provided:
14171 CRYPTO_set_mem_ex_functions
14172 CRYPTO_set_locked_mem_ex_functions
14173 CRYPTO_get_mem_ex_functions
14174 CRYPTO_get_locked_mem_ex_functions
14176 These work the same way as CRYPTO_set_mem_functions and friends.
14177 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14178 extended allocation function is enabled.
14179 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14180 a conventional allocation function is enabled.
14182 *Richard Levitte, Bodo Moeller*
14184 * Finish off removing the remaining LHASH function pointer casts.
14185 There should no longer be any prototype-casting required when using
14186 the LHASH abstraction, and any casts that remain are "bugs". See
14187 the callback types and macros at the head of lhash.h for details
14188 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14192 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14193 If /dev/[u]random devices are not available or do not return enough
14194 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14196 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14197 /etc/entropy will be queried once each in this sequence, querying stops
14198 when enough entropy was collected without querying more sockets.
14202 * Change the Unix RAND_poll() variant to be able to poll several
14203 random devices, as specified by DEVRANDOM, until a sufficient amount
14204 of data has been collected. We spend at most 10 ms on each file
14205 (select timeout) and read in non-blocking mode. DEVRANDOM now
14206 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14207 (previously it was just the string "/dev/urandom"), so on typical
14208 platforms the 10 ms delay will never occur.
14209 Also separate out the Unix variant to its own file, rand_unix.c.
14210 For VMS, there's a currently-empty rand_vms.c.
14214 * Move OCSP client related routines to ocsp_cl.c. These
14215 provide utility functions which an application needing
14216 to issue a request to an OCSP responder and analyse the
14217 response will typically need: as opposed to those which an
14218 OCSP responder itself would need which will be added later.
14220 OCSP_request_sign() signs an OCSP request with an API similar
14221 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14222 response. OCSP_response_get1_basic() extracts basic response
14223 from response. OCSP_resp_find_status(): finds and extracts status
14224 information from an OCSP_CERTID structure (which will be created
14225 when the request structure is built). These are built from lower
14226 level functions which work on OCSP_SINGLERESP structures but
14227 won't normally be used unless the application wishes to examine
14228 extensions in the OCSP response for example.
14230 Replace nonce routines with a pair of functions.
14231 OCSP_request_add1_nonce() adds a nonce value and optionally
14232 generates a random value. OCSP_check_nonce() checks the
14233 validity of the nonce in an OCSP response.
14237 * Change function OCSP_request_add() to OCSP_request_add0_id().
14238 This doesn't copy the supplied OCSP_CERTID and avoids the
14239 need to free up the newly created id. Change return type
14240 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14241 This can then be used to add extensions to the request.
14242 Deleted OCSP_request_new(), since most of its functionality
14243 is now in OCSP_REQUEST_new() (and the case insensitive name
14244 clash) apart from the ability to set the request name which
14245 will be added elsewhere.
14249 * Update OCSP API. Remove obsolete extensions argument from
14250 various functions. Extensions are now handled using the new
14251 OCSP extension code. New simple OCSP HTTP function which
14252 can be used to send requests and parse the response.
14256 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14257 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14258 uses the special reorder version of SET OF to sort the attributes
14259 and reorder them to match the encoded order. This resolves a long
14260 standing problem: a verify on a PKCS7 structure just after signing
14261 it used to fail because the attribute order did not match the
14262 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14263 it uses the received order. This is necessary to tolerate some broken
14264 software that does not order SET OF. This is handled by encoding
14265 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14266 to produce the required SET OF.
14270 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14271 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14272 files to get correct declarations of the ASN.1 item variables.
14276 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14277 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14278 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14279 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14280 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14281 ASN1_ITEM and no wrapper functions.
14285 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14286 replace the old function pointer based I/O routines. Change most of
14287 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14291 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14292 lines, recognize more "algorithms" that can be deselected, and make
14293 it complain about algorithm deselection that isn't recognised.
14297 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14298 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14299 to use new functions. Add NO_ASN1_OLD which can be set to remove
14300 some old style ASN1 functions: this can be used to determine if old
14301 code will still work when these eventually go away.
14305 * New extension functions for OCSP structures, these follow the
14306 same conventions as certificates and CRLs.
14310 * New function X509V3_add1_i2d(). This automatically encodes and
14311 adds an extension. Its behaviour can be customised with various
14312 flags to append, replace or delete. Various wrappers added for
14313 certificates and CRLs.
14317 * Fix to avoid calling the underlying ASN1 print routine when
14318 an extension cannot be parsed. Correct a typo in the
14319 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14323 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14324 entries for variables.
14328 * Add functionality to `apps/openssl.c` for detecting locking
14329 problems: As the program is single-threaded, all we have
14330 to do is register a locking callback using an array for
14331 storing which locks are currently held by the program.
14335 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14336 SSL_get_ex_data_X509_STORE_idx(), which is used in
14337 ssl_verify_cert_chain() and thus can be called at any time
14338 during TLS/SSL handshakes so that thread-safety is essential.
14339 Unfortunately, the ex_data design is not at all suited
14340 for multi-threaded use, so it probably should be abolished.
14344 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14346 *Broadcom, tweaked and integrated by Geoff Thorpe*
14348 * Move common extension printing code to new function
14349 X509V3_print_extensions(). Reorganise OCSP print routines and
14350 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14354 * New function X509_signature_print() to remove duplication in some
14359 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14360 set (this was treated exactly the same as SET OF previously). This
14361 is used to reorder the STACK representing the structure to match the
14362 encoding. This will be used to get round a problem where a PKCS7
14363 structure which was signed could not be verified because the STACK
14364 order did not reflect the encoded order.
14368 * Reimplement the OCSP ASN1 module using the new code.
14372 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14373 for its ASN1 operations. The old style function pointers still exist
14374 for now but they will eventually go away.
14378 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14379 completely replaces the old ASN1 functionality with a table driven
14380 encoder and decoder which interprets an ASN1_ITEM structure describing
14381 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14382 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14383 has also been converted to the new form.
14387 * Change BN_mod_exp_recp so that negative moduli are tolerated
14388 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14389 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14390 for negative moduli.
14394 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14395 of not touching the result's sign bit.
14399 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14404 * Changed the LHASH code to use prototypes for callbacks, and created
14405 macros to declare and implement thin (optionally static) functions
14406 that provide type-safety and avoid function pointer casting for the
14407 type-specific callbacks.
14411 * Added Kerberos Cipher Suites to be used with TLS, as written in
14413 *Veers Staats <staatsvr@asc.hpc.mil>,
14414 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14416 * Reformat the FAQ so the different questions and answers can be divided
14417 in sections depending on the subject.
14421 * Have the zlib compression code load ZLIB.DLL dynamically under
14426 * New function BN_mod_sqrt for computing square roots modulo a prime
14427 (using the probabilistic Tonelli-Shanks algorithm unless
14428 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14429 be handled deterministically).
14431 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14433 * Make BN_mod_inverse faster by explicitly handling small quotients
14434 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14435 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14439 * New function BN_kronecker.
14443 * Fix BN_gcd so that it works on negative inputs; the result is
14444 positive unless both parameters are zero.
14445 Previously something reasonably close to an infinite loop was
14446 possible because numbers could be growing instead of shrinking
14447 in the implementation of Euclid's algorithm.
14451 * Fix BN_is_word() and BN_is_one() macros to take into account the
14452 sign of the number in question.
14454 Fix BN_is_word(a,w) to work correctly for w == 0.
14456 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14457 because its test if the absolute value of 'a' equals 'w'.
14458 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14459 it exists mostly for use in the implementations of BN_is_zero(),
14460 BN_is_one(), and BN_is_word().
14464 * New function BN_swap.
14468 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14469 the exponentiation functions are more likely to produce reasonable
14470 results on negative inputs.
14474 * Change BN_mod_mul so that the result is always non-negative.
14475 Previously, it could be negative if one of the factors was negative;
14476 I don't think anyone really wanted that behaviour.
14480 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14481 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14482 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14483 and add new functions:
14492 BN_mod_lshift1_quick
14494 BN_mod_lshift_quick
14496 These functions always generate non-negative results.
14498 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14499 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14501 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14502 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14503 be reduced modulo `m`.
14505 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14508 The following entry accidentally appeared in the CHANGES file
14509 distributed with OpenSSL 0.9.7. The modifications described in
14510 it do *not* apply to OpenSSL 0.9.7.
14512 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14513 was actually never needed) and in BN_mul(). The removal in BN_mul()
14514 required a small change in bn_mul_part_recursive() and the addition
14515 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14516 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14517 bn_sub_words() and bn_add_words() except they take arrays with
14523 * In 'openssl passwd', verify passwords read from the terminal
14524 unless the '-salt' option is used (which usually means that
14525 verification would just waste user's time since the resulting
14526 hash is going to be compared with some given password hash)
14527 or the new '-noverify' option is used.
14529 This is an incompatible change, but it does not affect
14530 non-interactive use of 'openssl passwd' (passwords on the command
14531 line, '-stdin' option, '-in ...' option) and thus should not
14532 cause any problems.
14536 * Remove all references to RSAref, since there's no more need for it.
14540 * Make DSO load along a path given through an environment variable
14541 (SHLIB_PATH) with shl_load().
14545 * Constify the ENGINE code as a result of BIGNUM constification.
14546 Also constify the RSA code and most things related to it. In a
14547 few places, most notable in the depth of the ASN.1 code, ugly
14548 casts back to non-const were required (to be solved at a later
14553 * Make it so the openssl application has all engines loaded by default.
14557 * Constify the BIGNUM routines a little more.
14561 * Add the following functions:
14563 ENGINE_load_cswift()
14565 ENGINE_load_atalla()
14566 ENGINE_load_nuron()
14567 ENGINE_load_builtin_engines()
14569 That way, an application can itself choose if external engines that
14570 are built-in in OpenSSL shall ever be used or not. The benefit is
14571 that applications won't have to be linked with libdl or other dso
14572 libraries unless it's really needed.
14574 Changed 'openssl engine' to load all engines on demand.
14575 Changed the engine header files to avoid the duplication of some
14576 declarations (they differed!).
14580 * 'openssl engine' can now list capabilities.
14584 * Better error reporting in 'openssl engine'.
14588 * Never call load_dh_param(NULL) in s_server.
14592 * Add engine application. It can currently list engines by name and
14593 identity, and test if they are actually available.
14597 * Improve RPM specification file by forcing symbolic linking and making
14598 sure the installed documentation is also owned by root.root.
14600 *Damien Miller <djm@mindrot.org>*
14602 * Give the OpenSSL applications more possibilities to make use of
14603 keys (public as well as private) handled by engines.
14607 * Add OCSP code that comes from CertCo.
14611 * Add VMS support for the Rijndael code.
14615 * Added untested support for Nuron crypto accelerator.
14619 * Add support for external cryptographic devices. This code was
14620 previously distributed separately as the "engine" branch.
14622 *Geoff Thorpe, Richard Levitte*
14624 * Rework the filename-translation in the DSO code. It is now possible to
14625 have far greater control over how a "name" is turned into a filename
14626 depending on the operating environment and any oddities about the
14627 different shared library filenames on each system.
14631 * Support threads on FreeBSD-elf in Configure.
14635 * Fix for SHA1 assembly problem with MASM: it produces
14636 warnings about corrupt line number information when assembling
14637 with debugging information. This is caused by the overlapping
14640 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14643 NCONF_get_number() has no error checking at all. As a replacement,
14644 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14645 promoted strongly. The old NCONF_get_number is kept around for
14646 binary backward compatibility.
14647 Make it possible for methods to load from something other than a BIO,
14648 by providing a function pointer that is given a name instead of a BIO.
14649 For example, this could be used to load configuration data from an
14654 * Fix for non blocking accept BIOs. Added new I/O special reason
14655 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14656 with non blocking I/O was not possible because no retry code was
14657 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14662 * Added the beginnings of Rijndael support.
14666 * Fix for bug in DirectoryString mask setting. Add support for
14667 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14668 to allow certificate printing to more controllable, additional
14669 'certopt' option to 'x509' to allow new printing options to be
14674 * Clean old EAY MD5 hack from e_os.h.
14678 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14680 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14681 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14683 *Joe Orton, Steve Henson*
14685 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14687 * Fix additional bug revealed by the NISCC test suite:
14689 Stop bug triggering large recursion when presented with
14690 certain ASN.1 tags ([CVE-2003-0851])
14694 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14696 * Fix various bugs revealed by running the NISCC test suite:
14698 Stop out of bounds reads in the ASN1 code when presented with
14699 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14701 If verify callback ignores invalid public key errors don't try to check
14702 certificate signature with the NULL public key.
14706 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14707 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14712 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14713 extra data after the compression methods not only for TLS 1.0
14714 but also for SSL 3.0 (as required by the specification).
14716 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14718 * Change X509_certificate_type() to mark the key as exported/exportable
14719 when it's 512 *bits* long, not 512 bytes.
14723 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14725 * Countermeasure against the Klima-Pokorny-Rosa extension of
14726 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14727 a protocol version number mismatch like a decryption error
14728 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14732 * Turn on RSA blinding by default in the default implementation
14733 to avoid a timing attack. Applications that don't want it can call
14734 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14735 They would be ill-advised to do so in most cases.
14737 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14739 * Change RSA blinding code so that it works when the PRNG is not
14740 seeded (in this case, the secret RSA exponent is abused as
14741 an unpredictable seed -- if it is not unpredictable, there
14742 is no point in blinding anyway). Make RSA blinding thread-safe
14743 by remembering the creator's thread ID in rsa->blinding and
14744 having all other threads use local one-time blinding factors
14745 (this requires more computation than sharing rsa->blinding, but
14746 avoids excessive locking; and if an RSA object is not shared
14747 between threads, blinding will still be very fast).
14751 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14753 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14754 via timing by performing a MAC computation even if incorrect
14755 block cipher padding has been found. This is a countermeasure
14756 against active attacks where the attacker has to distinguish
14757 between bad padding and a MAC verification error. ([CVE-2003-0078])
14759 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14760 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14761 Martin Vuagnoux (EPFL, Ilion)*
14763 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14765 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14766 memory from its contents. This is done with a counter that will
14767 place alternating values in each byte. This can be used to solve
14768 two issues: 1) the removal of calls to memset() by highly optimizing
14769 compilers, and 2) cleansing with other values than 0, since those can
14770 be read through on certain media, for example a swap space on disk.
14774 * Bugfix: client side session caching did not work with external caching,
14775 because the session->cipher setting was not restored when reloading
14776 from the external cache. This problem was masked, when
14777 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14778 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14782 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14783 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14785 *Zeev Lieber <zeev-l@yahoo.com>*
14787 * Undo an undocumented change introduced in 0.9.6e which caused
14788 repeated calls to OpenSSL_add_all_ciphers() and
14789 OpenSSL_add_all_digests() to be ignored, even after calling
14794 * Change the default configuration reader to deal with last line not
14795 being properly terminated.
14799 * Change X509_NAME_cmp() so it applies the special rules on handling
14800 DN values that are of type PrintableString, as well as RDNs of type
14801 emailAddress where the value has the type ia5String.
14803 *stefank@valicert.com via Richard Levitte*
14805 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14806 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14807 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14808 the bitwise-OR of the two for use by the majority of applications
14809 wanting this behaviour, and update the docs. The documented
14810 behaviour and actual behaviour were inconsistent and had been
14811 changing anyway, so this is more a bug-fix than a behavioural
14814 *Geoff Thorpe, diagnosed by Nadav Har'El*
14816 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14817 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14821 * Fix initialization code race conditions in
14822 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14823 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14824 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14825 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14826 ssl2_get_cipher_by_char(),
14827 ssl3_get_cipher_by_char().
14829 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14831 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14832 the cached sessions are flushed, as the remove_cb() might use ex_data
14833 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14834 (see [openssl.org #212]).
14836 *Geoff Thorpe, Lutz Jaenicke*
14838 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14839 length, instead of the encoding length to d2i_ASN1_OBJECT.
14843 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14845 * [In 0.9.6g-engine release:]
14846 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14848 *Lynn Gazis <lgazis@rainbow.com>*
14850 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14852 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14853 and get fix the header length calculation.
14854 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14855 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14857 * Use proper error handling instead of 'assertions' in buffer
14858 overflow checks added in 0.9.6e. This prevents DoS (the
14859 assertions could call abort()).
14861 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14863 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14865 * Add various sanity checks to asn1_get_length() to reject
14866 the ASN1 length bytes if they exceed sizeof(long), will appear
14867 negative or the content length exceeds the length of the
14870 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14872 * Fix cipher selection routines: ciphers without encryption had no flags
14873 for the cipher strength set and where therefore not handled correctly
14874 by the selection routines (PR #130).
14878 * Fix EVP_dsa_sha macro.
14883 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14884 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14885 that was added in OpenSSL 0.9.6d.
14887 As the countermeasure turned out to be incompatible with some
14888 broken SSL implementations, the new option is part of SSL_OP_ALL.
14889 SSL_OP_ALL is usually employed when compatibility with weird SSL
14890 implementations is desired (e.g. '-bugs' option to 's_client' and
14891 's_server'), so the new option is automatically set in many
14896 * Changes in security patch:
14898 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14899 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14900 Air Force Materiel Command, USAF, under agreement number
14903 * Add various sanity checks to asn1_get_length() to reject
14904 the ASN1 length bytes if they exceed sizeof(long), will appear
14905 negative or the content length exceeds the length of the
14906 supplied buffer. ([CVE-2002-0659])
14908 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14910 * Assertions for various potential buffer overflows, not known to
14911 happen in practice.
14913 *Ben Laurie (CHATS)*
14915 * Various temporary buffers to hold ASCII versions of integers were
14916 too small for 64 bit platforms. ([CVE-2002-0655])
14917 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14919 * Remote buffer overflow in SSL3 protocol - an attacker could
14920 supply an oversized session ID to a client. ([CVE-2002-0656])
14922 *Ben Laurie (CHATS)*
14924 * Remote buffer overflow in SSL2 protocol - an attacker could
14925 supply an oversized client master key. ([CVE-2002-0656])
14927 *Ben Laurie (CHATS)*
14929 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14931 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14932 encoded as NULL) with id-dsa-with-sha1.
14934 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14936 * Check various `X509_...()` return values in `apps/req.c`.
14938 *Nils Larsch <nla@trustcenter.de>*
14940 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14941 an end-of-file condition would erroneously be flagged, when the CRLF
14942 was just at the end of a processed block. The bug was discovered when
14943 processing data through a buffering memory BIO handing the data to a
14944 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14945 <ptsekov@syntrex.com> and Nedelcho Stanev.
14949 * Implement a countermeasure against a vulnerability recently found
14950 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14951 before application data chunks to avoid the use of known IVs
14952 with data potentially chosen by the attacker.
14956 * Fix length checks in ssl3_get_client_hello().
14960 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14961 to prevent ssl3_read_internal() from incorrectly assuming that
14962 ssl3_read_bytes() found application data while handshake
14963 processing was enabled when in fact s->s3->in_read_app_data was
14964 merely automatically cleared during the initial handshake.
14966 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14968 * Fix object definitions for Private and Enterprise: they were not
14969 recognized in their shortname (=lowercase) representation. Extend
14970 obj_dat.pl to issue an error when using undefined keywords instead
14971 of silently ignoring the problem (Svenning Sorensen
14972 <sss@sss.dnsalias.net>).
14976 * Fix DH_generate_parameters() so that it works for 'non-standard'
14977 generators, i.e. generators other than 2 and 5. (Previously, the
14978 code did not properly initialise the 'add' and 'rem' values to
14979 BN_generate_prime().)
14981 In the new general case, we do not insist that 'generator' is
14982 actually a primitive root: This requirement is rather pointless;
14983 a generator of the order-q subgroup is just as good, if not
14988 * Map new X509 verification errors to alerts. Discovered and submitted by
14989 Tom Wu <tom@arcot.com>.
14993 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14994 returning non-zero before the data has been completely received
14995 when using non-blocking I/O.
14997 *Bodo Moeller; problem pointed out by John Hughes*
14999 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15001 *Ben Laurie, Lutz Jaenicke*
15003 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15004 Yoram Zahavi <YoramZ@gilian.com>).
15008 * Add information about CygWin 1.3 and on, and preserve proper
15009 configuration for the versions before that.
15011 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15013 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15014 check whether we deal with a copy of a session and do not delete from
15015 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15016 <izhar@checkpoint.com>.
15020 * Do not store session data into the internal session cache, if it
15021 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15022 flag is set). Proposed by Aslam <aslam@funk.com>.
15026 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15031 * [In 0.9.6d-engine release:]
15032 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15034 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15036 * Add the configuration target linux-s390x.
15038 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15040 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15041 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15042 variable as an indication that a ClientHello message has been
15043 received. As the flag value will be lost between multiple
15044 invocations of ssl3_accept when using non-blocking I/O, the
15045 function may not be aware that a handshake has actually taken
15046 place, thus preventing a new session from being added to the
15049 To avoid this problem, we now set s->new_session to 2 instead of
15050 using a local variable.
15052 *Lutz Jaenicke, Bodo Moeller*
15054 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15055 if the SSL_R_LENGTH_MISMATCH error is detected.
15057 *Geoff Thorpe, Bodo Moeller*
15059 * New 'shared_ldflag' column in Configure platform table.
15063 * Fix EVP_CIPHER_mode macro.
15065 *"Dan S. Camper" <dan@bti.net>*
15067 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15068 type, we must throw them away by setting rr->length to 0.
15070 *D P Chang <dpc@qualys.com>*
15072 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15074 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15075 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15076 worked incorrectly for those cases where range = `10..._2` and
15077 `3*range` is two bits longer than range.)
15081 * Only add signing time to PKCS7 structures if it is not already
15086 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15087 OBJ_ld_ce should be OBJ_id_ce.
15088 Also some ip-pda OIDs in crypto/objects/objects.txt were
15089 incorrect (cf. RFC 3039).
15091 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15093 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15094 returns early because it has nothing to do.
15096 *Andy Schneider <andy.schneider@bjss.co.uk>*
15098 * [In 0.9.6c-engine release:]
15099 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15101 *Andy Schneider <andy.schneider@bjss.co.uk>*
15103 * [In 0.9.6c-engine release:]
15104 Add support for Cryptographic Appliance's keyserver technology.
15105 (Use engine 'keyclient')
15107 *Cryptographic Appliances and Geoff Thorpe*
15109 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15110 is called via tools/c89.sh because arguments have to be
15111 rearranged (all '-L' options must appear before the first object
15114 *Richard Shapiro <rshapiro@abinitio.com>*
15116 * [In 0.9.6c-engine release:]
15117 Add support for Broadcom crypto accelerator cards, backported
15120 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15122 * [In 0.9.6c-engine release:]
15123 Add support for SureWare crypto accelerator cards from
15124 Baltimore Technologies. (Use engine 'sureware')
15126 *Baltimore Technologies and Mark Cox*
15128 * [In 0.9.6c-engine release:]
15129 Add support for crypto accelerator cards from Accelerated
15130 Encryption Processing, www.aep.ie. (Use engine 'aep')
15132 *AEP Inc. and Mark Cox*
15134 * Add a configuration entry for gcc on UnixWare.
15136 *Gary Benson <gbenson@redhat.com>*
15138 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15139 messages are stored in a single piece (fixed-length part and
15140 variable-length part combined) and fix various bugs found on the way.
15144 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15145 instead. BIO_gethostbyname() does not know what timeouts are
15146 appropriate, so entries would stay in cache even when they have
15148 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15150 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15151 faced with a pathologically small ClientHello fragment that does
15152 not contain client_version: Instead of aborting with an error,
15153 simply choose the highest available protocol version (i.e.,
15154 TLS 1.0 unless it is disabled). In practice, ClientHello
15155 messages are never sent like this, but this change gives us
15156 strictly correct behaviour at least for TLS.
15160 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15161 never resets s->method to s->ctx->method when called from within
15162 one of the SSL handshake functions.
15164 *Bodo Moeller; problem pointed out by Niko Baric*
15166 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15167 (sent using the client's version number) if client_version is
15168 smaller than the protocol version in use. Also change
15169 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15170 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15171 the client will at least see that alert.
15175 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15180 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15181 client receives HelloRequest while in a handshake.
15183 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15185 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15186 should end in 'break', not 'goto end' which circumvents various
15187 cleanups done in state SSL_ST_OK. But session related stuff
15188 must be disabled for SSL_ST_OK in the case that we just sent a
15191 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15192 before just sending a HelloRequest.
15194 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15196 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15197 reveal whether illegal block cipher padding was found or a MAC
15198 verification error occurred. (Neither SSLerr() codes nor alerts
15199 are directly visible to potential attackers, but the information
15200 may leak via logfiles.)
15202 Similar changes are not required for the SSL 2.0 implementation
15203 because the number of padding bytes is sent in clear for SSL 2.0,
15204 and the extra bytes are just ignored. However ssl/s2_pkt.c
15205 failed to verify that the purported number of padding bytes is in
15210 * Add OpenUNIX-8 support including shared libraries
15211 (Boyd Lynn Gerber <gerberb@zenez.com>).
15215 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15216 'wristwatch attack' using huge encoding parameters (cf.
15217 James H. Manger's CRYPTO 2001 paper). Note that the
15218 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15219 encoding parameters and hence was not vulnerable.
15223 * BN_sqr() bug fix.
15225 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15227 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15228 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15229 followed by modular reduction.
15231 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15233 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15234 equivalent based on BN_pseudo_rand() instead of BN_rand().
15238 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15239 This function was broken, as the check for a new client hello message
15240 to handle SGC did not allow these large messages.
15241 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15245 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15249 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15250 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15254 * Rework the configuration and shared library support for Tru64 Unix.
15255 The configuration part makes use of modern compiler features and
15256 still retains old compiler behavior for those that run older versions
15257 of the OS. The shared library support part includes a variant that
15258 uses the RPATH feature, and is available through the special
15259 configuration target "alpha-cc-rpath", which will never be selected
15262 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15264 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15265 with the same message size as in ssl3_get_certificate_request().
15266 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15267 messages might inadvertently be reject as too long.
15269 *Petr Lampa <lampa@fee.vutbr.cz>*
15271 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15275 * Modified SSL library such that the verify_callback that has been set
15276 specifically for an SSL object with SSL_set_verify() is actually being
15277 used. Before the change, a verify_callback set with this function was
15278 ignored and the verify_callback() set in the SSL_CTX at the time of
15279 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15280 to allow the necessary settings.
15284 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15285 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15286 done automatically (in contradiction to the requirements of the C
15287 standard). This made problems when used from OpenSSH.
15291 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15292 dh->length and always used
15294 BN_rand_range(priv_key, dh->p).
15296 BN_rand_range() is not necessary for Diffie-Hellman, and this
15297 specific range makes Diffie-Hellman unnecessarily inefficient if
15298 dh->length (recommended exponent length) is much smaller than the
15299 length of dh->p. We could use BN_rand_range() if the order of
15300 the subgroup was stored in the DH structure, but we only have
15305 BN_rand(priv_key, l, ...)
15307 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15314 RSA_eay_public_encrypt
15315 RSA_eay_private_decrypt
15316 RSA_eay_private_encrypt (signing)
15317 RSA_eay_public_decrypt (signature verification)
15319 (default implementations for RSA_public_encrypt,
15320 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15321 always reject numbers >= n.
15325 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15326 to synchronize access to 'locking_thread'. This is necessary on
15327 systems where access to 'locking_thread' (an 'unsigned long'
15328 variable) is not atomic.
15332 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15333 *before* setting the 'crypto_lock_rand' flag. The previous code had
15334 a race condition if 0 is a valid thread ID.
15336 *Travis Vitek <vitek@roguewave.com>*
15338 * Add support for shared libraries under Irix.
15340 *Albert Chin-A-Young <china@thewrittenword.com>*
15342 * Add configuration option to build on Linux on both big-endian and
15343 little-endian MIPS.
15345 *Ralf Baechle <ralf@uni-koblenz.de>*
15347 * Add the possibility to create shared libraries on HP-UX.
15351 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15353 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15354 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15355 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15356 PRNG state recovery was possible based on the output of
15357 one PRNG request appropriately sized to gain knowledge on
15358 'md' followed by enough consecutive 1-byte PRNG requests
15359 to traverse all of 'state'.
15361 1. When updating 'md_local' (the current thread's copy of 'md')
15362 during PRNG output generation, hash all of the previous
15363 'md_local' value, not just the half used for PRNG output.
15365 2. Make the number of bytes from 'state' included into the hash
15366 independent from the number of PRNG bytes requested.
15368 The first measure alone would be sufficient to avoid
15369 Markku-Juhani's attack. (Actually it had never occurred
15370 to me that the half of 'md_local' used for chaining was the
15371 half from which PRNG output bytes were taken -- I had always
15372 assumed that the secret half would be used.) The second
15373 measure makes sure that additional data from 'state' is never
15374 mixed into 'md_local' in small portions; this heuristically
15375 further strengthens the PRNG.
15379 * Fix crypto/bn/asm/mips3.s.
15383 * When only the key is given to "enc", the IV is undefined. Print out
15384 an error message in this case.
15388 * Handle special case when X509_NAME is empty in X509 printing routines.
15392 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15393 positive and less than q.
15397 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15398 used: it isn't thread safe and the add_lock_callback should handle
15401 *Paul Rose <Paul.Rose@bridge.com>*
15403 * Verify that incoming data obeys the block size in
15404 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15410 *Ulf Möller, Bodo Möller*
15412 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15413 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15414 when fixing the server behaviour for backwards-compatible 'client
15415 hello' messages. (Note that the attack is impractical against
15416 SSL 3.0 and TLS 1.0 anyway because length and version checking
15417 means that the probability of guessing a valid ciphertext is
15418 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15421 Before 0.9.5, the countermeasure (hide the error by generating a
15422 random 'decryption result') did not work properly because
15423 ERR_clear_error() was missing, meaning that SSL_get_error() would
15424 detect the supposedly ignored error.
15426 Both problems are now fixed.
15430 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15431 (previously it was 1024).
15435 * Fix for compatibility mode trust settings: ignore trust settings
15436 unless some valid trust or reject settings are present.
15440 * Fix for blowfish EVP: its a variable length cipher.
15444 * Fix various bugs related to DSA S/MIME verification. Handle missing
15445 parameters in DSA public key structures and return an error in the
15446 DSA routines if parameters are absent.
15450 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15451 in the current directory if neither $RANDFILE nor $HOME was set.
15452 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15453 caused some confusion to Windows users who haven't defined $HOME.
15454 Thus RAND_file_name() is changed again: e_os.h can define a
15455 DEFAULT_HOME, which will be used if $HOME is not set.
15456 For Windows, we use "C:"; on other platforms, we still require
15457 environment variables.
15459 * Move 'if (!initialized) RAND_poll()' into regions protected by
15460 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15461 having multiple threads call RAND_poll() concurrently.
15465 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15466 combination of a flag and a thread ID variable.
15467 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15468 flag), *other* threads can enter ssleay_add_bytes without obeying
15469 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15470 that they do not hold after the first thread unsets add_do_not_lock).
15474 * Change bctest again: '-x' expressions are not available in all
15475 versions of 'test'.
15479 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15481 * Fix a couple of memory leaks in PKCS7_dataDecode()
15483 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15485 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15486 the default extension for executables, if any. Also, make the perl
15487 scripts that use symlink() to test if it really exists and use "cp"
15488 if it doesn't. All this made OpenSSL compilable and installable in
15493 * Fix for asn1_GetSequence() for indefinite length constructed data.
15494 If SEQUENCE is length is indefinite just set c->slen to the total
15495 amount of data available.
15497 *Steve Henson, reported by shige@FreeBSD.org*
15499 *This change does not apply to 0.9.7.*
15501 * Change bctest to avoid here-documents inside command substitution
15502 (workaround for FreeBSD /bin/sh bug).
15503 For compatibility with Ultrix, avoid shell functions (introduced
15504 in the bctest version that searches along $PATH).
15508 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15509 with des_encrypt() defined on some operating systems, like Solaris
15514 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15515 On the Importance of Eliminating Errors in Cryptographic
15516 Computations, J. Cryptology 14 (2001) 2, 101-119,
15517 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15521 * MIPS assembler BIGNUM division bug fix.
15525 * Disabled incorrect Alpha assembler code.
15529 * Fix PKCS#7 decode routines so they correctly update the length
15530 after reading an EOC for the EXPLICIT tag.
15534 *This change does not apply to 0.9.7.*
15536 * Fix bug in PKCS#12 key generation routines. This was triggered
15537 if a 3DES key was generated with a 0 initial byte. Include
15538 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15539 (but broken) behaviour.
15543 * Enhance bctest to search for a working bc along $PATH and print
15546 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15548 * Fix memory leaks in err.c: free err_data string if necessary;
15549 don't write to the wrong index in ERR_set_error_data.
15553 * Implement ssl23_peek (analogous to ssl23_read), which previously
15558 * Replace rdtsc with `_emit` statements for VC++ version 5.
15560 *Jeremy Cooper <jeremy@baymoo.org>*
15562 * Make it possible to reuse SSLv2 sessions.
15566 * In copy_email() check for >= 0 as a return value for
15567 X509_NAME_get_index_by_NID() since 0 is a valid index.
15569 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15571 * Avoid coredump with unsupported or invalid public keys by checking if
15572 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15573 PKCS7_verify() fails with non detached data.
15577 * Don't use getenv in library functions when run as setuid/setgid.
15578 New function OPENSSL_issetugid().
15582 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15583 due to incorrect handling of multi-threading:
15585 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15587 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15589 3. Count how many times MemCheck_off() has been called so that
15590 nested use can be treated correctly. This also avoids
15591 inband-signalling in the previous code (which relied on the
15592 assumption that thread ID 0 is impossible).
15596 * Add "-rand" option also to s_client and s_server.
15600 * Fix CPU detection on Irix 6.x.
15601 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15602 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15604 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15609 *This change does not apply to 0.9.7.*
15611 * Use the cached encoding of an X509_NAME structure rather than
15612 copying it. This is apparently the reason for the libsafe "errors"
15613 but the code is actually correct.
15617 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15618 Bleichenbacher's DSA attack.
15619 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15620 to be set and top=0 forces the highest bit to be set; top=-1 is new
15621 and leaves the highest bit random.
15623 *Ulf Moeller, Bodo Moeller*
15625 * In the `NCONF_...`-based implementations for `CONF_...` queries
15626 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15627 a temporary CONF structure with the data component set to NULL
15628 (which gives segmentation faults in lh_retrieve).
15629 Instead, use NULL for the CONF pointer in CONF_get_string and
15630 CONF_get_number (which may use environment variables) and directly
15631 return NULL from CONF_get_section.
15635 * Fix potential buffer overrun for EBCDIC.
15639 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15640 keyUsage if basicConstraints absent for a CA.
15644 * Make SMIME_write_PKCS7() write mail header values with a format that
15645 is more generally accepted (no spaces before the semicolon), since
15646 some programs can't parse those values properly otherwise. Also make
15647 sure BIO's that break lines after each write do not create invalid
15652 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15653 macros previously used would not encode an empty SEQUENCE OF
15654 and break the signature.
15658 *This change does not apply to 0.9.7.*
15660 * Zero the premaster secret after deriving the master secret in
15665 * Add some EVP_add_digest_alias registrations (as found in
15666 OpenSSL_add_all_digests()) to SSL_library_init()
15667 aka OpenSSL_add_ssl_algorithms(). This provides improved
15668 compatibility with peers using X.509 certificates
15669 with unconventional AlgorithmIdentifier OIDs.
15673 * Fix for Irix with NO_ASM.
15675 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15677 * ./config script fixes.
15679 *Ulf Moeller, Richard Levitte*
15681 * Fix 'openssl passwd -1'.
15685 * Change PKCS12_key_gen_asc() so it can cope with non null
15686 terminated strings whose length is passed in the passlen
15687 parameter, for example from PEM callbacks. This was done
15688 by adding an extra length parameter to asc2uni().
15690 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15692 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15693 call failed, free the DSA structure.
15697 * Fix to uni2asc() to cope with zero length Unicode strings.
15698 These are present in some PKCS#12 files.
15702 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15703 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15704 when writing a 32767 byte record.
15706 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15708 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15709 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15711 (RSA objects have a reference count access to which is protected
15712 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15713 so they are meant to be shared between threads.)
15714 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15715 "Reddie, Steven" <Steven.Reddie@ca.com>*
15717 * Fix a deadlock in CRYPTO_mem_leaks().
15721 * Use better test patterns in bntest.
15725 * rand_win.c fix for Borland C.
15729 * BN_rshift bugfix for n == 0.
15733 * Add a 'bctest' script that checks for some known 'bc' bugs
15734 so that 'make test' does not abort just because 'bc' is broken.
15738 * Store verify_result within SSL_SESSION also for client side to
15739 avoid potential security hole. (Re-used sessions on the client side
15740 always resulted in verify_result==X509_V_OK, not using the original
15741 result of the server certificate verification.)
15745 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15746 SSL3_RT_APPLICATION_DATA, return 0.
15747 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15752 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15753 releases, have been re-implemented by renaming the previous
15754 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15755 and ssl3_read_internal, respectively, and adding 'peek' parameters
15756 to them. The new ssl[23]_{read,peek} functions are calls to
15757 ssl[23]_read_internal with the 'peek' flag set appropriately.
15758 A 'peek' parameter has also been added to ssl3_read_bytes, which
15759 does the actual work for ssl3_read_internal.
15763 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15764 the method-specific "init()" handler. Also clean up ex_data after
15765 calling the method-specific "finish()" handler. Previously, this was
15766 happening the other way round.
15770 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15771 The previous value, 12, was not always sufficient for BN_mod_exp().
15775 * Make sure that shared libraries get the internal name engine with
15776 the full version number and not just 0. This should mark the
15777 shared libraries as not backward compatible. Of course, this should
15778 be changed again when we can guarantee backward binary compatibility.
15782 * Fix typo in get_cert_by_subject() in by_dir.c
15784 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15786 * Rework the system to generate shared libraries:
15788 - Make note of the expected extension for the shared libraries and
15789 if there is a need for symbolic links from for example libcrypto.so.0
15790 to libcrypto.so.0.9.7. There is extended info in Configure for
15793 - Make as few rebuilds of the shared libraries as possible.
15795 - Still avoid linking the OpenSSL programs with the shared libraries.
15797 - When installing, install the shared libraries separately from the
15802 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15804 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15805 and not in SSL_clear because the latter is also used by the
15806 accept/connect functions; previously, the settings made by
15807 SSL_set_read_ahead would be lost during the handshake.
15809 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15811 * Correct util/mkdef.pl to be selective about disabled algorithms.
15812 Previously, it would create entries for disabled algorithms no
15817 * Added several new manual pages for SSL_* function.
15821 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15823 * In ssl23_get_client_hello, generate an error message when faced
15824 with an initial SSL 3.0/TLS record that is too small to contain the
15825 first two bytes of the ClientHello message, i.e. client_version.
15826 (Note that this is a pathologic case that probably has never happened
15827 in real life.) The previous approach was to use the version number
15828 from the record header as a substitute; but our protocol choice
15829 should not depend on that one because it is not authenticated
15830 by the Finished messages.
15834 * More robust randomness gathering functions for Windows.
15836 *Jeffrey Altman <jaltman@columbia.edu>*
15838 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15839 not set then we don't setup the error code for issuer check errors
15840 to avoid possibly overwriting other errors which the callback does
15841 handle. If an application does set the flag then we assume it knows
15842 what it is doing and can handle the new informational codes
15847 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15848 a general "ANY" type, as such it should be able to decode anything
15849 including tagged types. However it didn't check the class so it would
15850 wrongly interpret tagged types in the same way as their universal
15851 counterpart and unknown types were just rejected. Changed so that the
15852 tagged and unknown types are handled in the same way as a SEQUENCE:
15853 that is the encoding is stored intact. There is also a new type
15854 "V_ASN1_OTHER" which is used when the class is not universal, in this
15855 case we have no idea what the actual type is so we just lump them all
15860 * On VMS, stdout may very well lead to a file that is written to
15861 in a record-oriented fashion. That means that every write() will
15862 write a separate record, which will be read separately by the
15863 programs trying to read from it. This can be very confusing.
15865 The solution is to put a BIO filter in the way that will buffer
15866 text until a linefeed is reached, and then write everything a
15867 line at a time, so every record written will be an actual line,
15868 not chunks of lines and not (usually doesn't happen, but I've
15869 seen it once) several lines in one record. BIO_f_linebuffer() is
15872 Currently, it's a VMS-only method, because that's where it has
15873 been tested well enough.
15877 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15878 it can return incorrect results.
15879 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15880 but it was in 0.9.6-beta[12].)
15884 * Disable the check for content being present when verifying detached
15885 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15886 include zero length content when signing messages.
15890 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15891 BIO_ctrl (for BIO pairs).
15895 * Add DSO method for VMS.
15899 * Bug fix: Montgomery multiplication could produce results with the
15904 * Add RPM specification openssl.spec and modify it to build three
15905 packages. The default package contains applications, application
15906 documentation and run-time libraries. The devel package contains
15907 include files, static libraries and function documentation. The
15908 doc package contains the contents of the doc directory. The original
15909 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15913 * Add a large number of documentation files for many SSL routines.
15915 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15917 * Add a configuration entry for Sony News 4.
15919 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15921 * Don't set the two most significant bits to one when generating a
15922 random number < q in the DSA library.
15926 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15927 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15928 the underlying transport is blocking) if a handshake took place.
15929 (The default behaviour is needed by applications such as s_client
15930 and s_server that use select() to determine when to use SSL_read;
15931 but for applications that know in advance when to expect data, it
15932 just makes things more complicated.)
15936 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15941 * Add a few more EBCDIC conditionals that make `req` and `x509`
15942 work better on such systems.
15944 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15946 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15947 Update PKCS12_parse() so it copies the friendlyName and the
15948 keyid to the certificates aux info.
15952 * Fix bug in PKCS7_verify() which caused an infinite loop
15953 if there was more than one signature.
15955 *Sven Uszpelkat <su@celocom.de>*
15957 * Major change in util/mkdef.pl to include extra information
15958 about each symbol, as well as presenting variables as well
15959 as functions. This change means that there's n more need
15960 to rebuild the .num files when some algorithms are excluded.
15964 * Allow the verify time to be set by an application,
15965 rather than always using the current time.
15969 * Phase 2 verify code reorganisation. The certificate
15970 verify code now looks up an issuer certificate by a
15971 number of criteria: subject name, authority key id
15972 and key usage. It also verifies self signed certificates
15973 by the same criteria. The main comparison function is
15974 X509_check_issued() which performs these checks.
15976 Lot of changes were necessary in order to support this
15977 without completely rewriting the lookup code.
15979 Authority and subject key identifier are now cached.
15981 The LHASH 'certs' is X509_STORE has now been replaced
15982 by a STACK_OF(X509_OBJECT). This is mainly because an
15983 LHASH can't store or retrieve multiple objects with
15984 the same hash value.
15986 As a result various functions (which were all internal
15987 use only) have changed to handle the new X509_STORE
15988 structure. This will break anything that messed round
15989 with X509_STORE internally.
15991 The functions X509_STORE_add_cert() now checks for an
15992 exact match, rather than just subject name.
15994 The X509_STORE API doesn't directly support the retrieval
15995 of multiple certificates matching a given criteria, however
15996 this can be worked round by performing a lookup first
15997 (which will fill the cache with candidate certificates)
15998 and then examining the cache for matches. This is probably
15999 the best we can do without throwing out X509_LOOKUP
16000 entirely (maybe later...).
16002 The X509_VERIFY_CTX structure has been enhanced considerably.
16004 All certificate lookup operations now go via a get_issuer()
16005 callback. Although this currently uses an X509_STORE it
16006 can be replaced by custom lookups. This is a simple way
16007 to bypass the X509_STORE hackery necessary to make this
16008 work and makes it possible to use more efficient techniques
16009 in future. A very simple version which uses a simple
16010 STACK for its trusted certificate store is also provided
16011 using X509_STORE_CTX_trusted_stack().
16013 The verify_cb() and verify() callbacks now have equivalents
16014 in the X509_STORE_CTX structure.
16016 X509_STORE_CTX also has a 'flags' field which can be used
16017 to customise the verify behaviour.
16021 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16022 excludes S/MIME capabilities.
16026 * When a certificate request is read in keep a copy of the
16027 original encoding of the signed data and use it when outputting
16028 again. Signatures then use the original encoding rather than
16029 a decoded, encoded version which may cause problems if the
16030 request is improperly encoded.
16034 * For consistency with other BIO_puts implementations, call
16035 buffer_write(b, ...) directly in buffer_puts instead of calling
16038 In BIO_puts, increment b->num_write as in BIO_write.
16040 *Peter.Sylvester@EdelWeb.fr*
16042 * Fix BN_mul_word for the case where the word is 0. (We have to use
16043 BN_zero, we may not return a BIGNUM with an array consisting of
16044 words set to zero.)
16048 * Avoid calling abort() from within the library when problems are
16049 detected, except if preprocessor symbols have been defined
16050 (such as REF_CHECK, BN_DEBUG etc.).
16054 * New openssl application 'rsautl'. This utility can be
16055 used for low-level RSA operations. DER public key
16056 BIO/fp routines also added.
16060 * New Configure entry and patches for compiling on QNX 4.
16062 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16064 * A demo state-machine implementation was sponsored by
16065 Nuron (<http://www.nuron.com/>) and is now available in
16066 demos/state_machine.
16070 * New options added to the 'dgst' utility for signature
16071 generation and verification.
16075 * Unrecognized PKCS#7 content types are now handled via a
16076 catch all ASN1_TYPE structure. This allows unsupported
16077 types to be stored as a "blob" and an application can
16078 encode and decode it manually.
16082 * Fix various signed/unsigned issues to make a_strex.c
16083 compile under VC++.
16085 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16087 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16088 length if passed a buffer. ASN1_INTEGER_to_BN failed
16089 if passed a NULL BN and its argument was negative.
16091 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16093 * Modification to PKCS#7 encoding routines to output definite
16094 length encoding. Since currently the whole structures are in
16095 memory there's not real point in using indefinite length
16096 constructed encoding. However if OpenSSL is compiled with
16097 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16101 * Added BIO_vprintf() and BIO_vsnprintf().
16105 * Added more prefixes to parse for in the strings written
16106 through a logging bio, to cover all the levels that are available
16107 through syslog. The prefixes are now:
16109 PANIC, EMERG, EMR => LOG_EMERG
16110 ALERT, ALR => LOG_ALERT
16111 CRIT, CRI => LOG_CRIT
16112 ERROR, ERR => LOG_ERR
16113 WARNING, WARN, WAR => LOG_WARNING
16114 NOTICE, NOTE, NOT => LOG_NOTICE
16115 INFO, INF => LOG_INFO
16116 DEBUG, DBG => LOG_DEBUG
16118 and as before, if none of those prefixes are present at the
16119 beginning of the string, LOG_ERR is chosen.
16121 On Win32, the `LOG_*` levels are mapped according to this:
16123 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16124 LOG_WARNING => EVENTLOG_WARNING_TYPE
16125 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16129 * Made it possible to reconfigure with just the configuration
16130 argument "reconf" or "reconfigure". The command line arguments
16131 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16132 and are retrieved from there when reconfiguring.
16138 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16140 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16144 * The obj_dat.pl script was messing up the sorting of object
16145 names. The reason was that it compared the quoted version
16146 of strings as a result "OCSP" > "OCSP Signing" because
16147 " > SPACE. Changed script to store unquoted versions of
16148 names and add quotes on output. It was also omitting some
16149 names from the lookup table if they were given a default
16150 value (that is if SN is missing it is given the same
16151 value as LN and vice versa), these are now added on the
16152 grounds that if an object has a name we should be able to
16153 look it up. Finally added warning output when duplicate
16154 short or long names are found.
16158 * Changes needed for Tandem NSK.
16160 *Scott Uroff <scott@xypro.com>*
16162 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16163 RSA_padding_check_SSLv23(), special padding was never detected
16164 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16165 version rollback attacks was not effective.
16167 In s23_clnt.c, don't use special rollback-attack detection padding
16168 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16169 client; similarly, in s23_srvr.c, don't do the rollback check if
16170 SSL 2.0 is the only protocol enabled in the server.
16174 * Make it possible to get hexdumps of unprintable data with 'openssl
16175 asn1parse'. By implication, the functions ASN1_parse_dump() and
16176 BIO_dump_indent() are added.
16180 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16181 these print out strings and name structures based on various
16182 flags including RFC2253 support and proper handling of
16183 multibyte characters. Added options to the 'x509' utility
16184 to allow the various flags to be set.
16188 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16189 Also change the functions X509_cmp_current_time() and
16190 X509_gmtime_adj() work with an ASN1_TIME structure,
16191 this will enable certificates using GeneralizedTime in validity
16192 dates to be checked.
16196 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16197 negative public key encodings) on by default,
16198 NO_NEG_PUBKEY_BUG can be set to disable it.
16202 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16203 content octets. An i2c_ASN1_OBJECT is unnecessary because
16204 the encoding can be trivially obtained from the structure.
16208 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16209 not read locks (`CRYPTO_r_[un]lock`).
16213 * A first attempt at creating official support for shared
16214 libraries through configuration. I've kept it so the
16215 default is static libraries only, and the OpenSSL programs
16216 are always statically linked for now, but there are
16217 preparations for dynamic linking in place.
16218 This has been tested on Linux and Tru64.
16222 * Randomness polling function for Win9x, as described in:
16223 Peter Gutmann, Software Generation of Practically Strong
16228 * Fix so PRNG is seeded in req if using an already existing
16233 * New options to smime application. -inform and -outform
16234 allow alternative formats for the S/MIME message including
16235 PEM and DER. The -content option allows the content to be
16236 specified separately. This should allow things like Netscape
16237 form signing output easier to verify.
16241 * Fix the ASN1 encoding of tags using the 'long form'.
16245 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16246 STRING types. These convert content octets to and from the
16247 underlying type. The actual tag and length octets are
16248 already assumed to have been read in and checked. These
16249 are needed because all other string types have virtually
16250 identical handling apart from the tag. By having versions
16251 of the ASN1 functions that just operate on content octets
16252 IMPLICIT tagging can be handled properly. It also allows
16253 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16254 and ASN1_INTEGER are identical apart from the tag.
16258 * Change the handling of OID objects as follows:
16260 - New object identifiers are inserted in objects.txt, following
16261 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16262 - objects.pl is used to process obj_mac.num and create a new
16264 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16267 This is currently kind of a hack, and the perl code in objects.pl
16268 isn't very elegant, but it works as I intended. The simplest way
16269 to check that it worked correctly is to look in obj_dat.h and
16270 check the array nid_objs and make sure the objects haven't moved
16271 around (this is important!). Additions are OK, as well as
16272 consistent name changes.
16276 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16280 * Addition of the command line parameter '-rand file' to 'openssl req'.
16281 The given file adds to whatever has already been seeded into the
16282 random pool through the RANDFILE configuration file option or
16283 environment variable, or the default random state file.
16287 * mkstack.pl now sorts each macro group into lexical order.
16288 Previously the output order depended on the order the files
16289 appeared in the directory, resulting in needless rewriting
16294 * Patches to make OpenSSL compile under Win32 again. Mostly
16295 work arounds for the VC++ problem that it treats func() as
16296 func(void). Also stripped out the parts of mkdef.pl that
16297 added extra typesafe functions: these no longer exist.
16301 * Reorganisation of the stack code. The macros are now all
16302 collected in safestack.h . Each macro is defined in terms of
16303 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16304 DEBUG_SAFESTACK is now handled in terms of function casts,
16305 this has the advantage of retaining type safety without the
16306 use of additional functions. If DEBUG_SAFESTACK is not defined
16307 then the non typesafe macros are used instead. Also modified the
16308 mkstack.pl script to handle the new form. Needs testing to see
16309 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16310 the default if no major problems. Similar behaviour for ASN1_SET_OF
16311 and PKCS12_STACK_OF.
16315 * When some versions of IIS use the 'NET' form of private key the
16316 key derivation algorithm is different. Normally MD5(password) is
16317 used as a 128 bit RC4 key. In the modified case
16318 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16319 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16320 as the old Netscape_RSA functions except they have an additional
16321 'sgckey' parameter which uses the modified algorithm. Also added
16322 an -sgckey command line option to the rsa utility. Thanks to
16323 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16324 algorithm to openssl-dev.
16328 * The evp_local.h macros were using 'c.##kname' which resulted in
16329 invalid expansion on some systems (SCO 5.0.5 for example).
16330 Corrected to 'c.kname'.
16332 *Phillip Porch <root@theporch.com>*
16334 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16335 a STACK of email addresses from a certificate or request, these look
16336 in the subject name and the subject alternative name extensions and
16337 omit any duplicate addresses.
16341 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16342 This makes DSA verification about 2 % faster.
16346 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16347 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16348 plus overhead for 1024 bit moduli).
16349 This makes exponentiations about 0.5 % faster for 1024 bit
16350 exponents (as measured by "openssl speed rsa2048").
16354 * Rename memory handling macros to avoid conflicts with other
16356 Malloc => OPENSSL_malloc
16357 Malloc_locked => OPENSSL_malloc_locked
16358 Realloc => OPENSSL_realloc
16359 Free => OPENSSL_free
16363 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16364 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16368 * CygWin32 support.
16370 *John Jarvie <jjarvie@newsguy.com>*
16372 * The type-safe stack code has been rejigged. It is now only compiled
16373 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16374 by default all type-specific stack functions are "#define"d back to
16375 standard stack functions. This results in more streamlined output
16376 but retains the type-safety checking possibilities of the original
16381 * The STACK code has been cleaned up, and certain type declarations
16382 that didn't make a lot of sense have been brought in line. This has
16383 also involved a cleanup of sorts in safestack.h to more correctly
16384 map type-safe stack functions onto their plain stack counterparts.
16385 This work has also resulted in a variety of "const"ifications of
16386 lots of the code, especially `_cmp` operations which should normally
16387 be prototyped with "const" parameters anyway.
16391 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16392 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16393 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16394 where all of 'md' is used each time the PRNG is used, but 'state'
16395 is used only indexed by a cyclic counter. As entropy may not be
16396 well distributed from the beginning, 'md' is important as a
16397 chaining variable. However, the output function chains only half
16398 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16399 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16400 in all of 'state' being rewritten, with the new values depending
16401 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16405 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16406 the handshake is continued after ssl_verify_cert_chain();
16407 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16408 can lead to 'unexplainable' connection aborts later.
16410 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16412 * Major EVP API cipher revision.
16413 Add hooks for extra EVP features. This allows various cipher
16414 parameters to be set in the EVP interface. Support added for variable
16415 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16416 setting of RC2 and RC5 parameters.
16418 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16421 Remove lots of duplicated code from the EVP library. For example *every*
16422 cipher init() function handles the 'iv' in the same way according to the
16423 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16424 for CFB and OFB modes they zero ctx->num.
16426 New functionality allows removal of S/MIME code RC2 hack.
16428 Most of the routines have the same form and so can be declared in terms
16431 By shifting this to the top level EVP_CipherInit() it can be removed from
16432 all individual ciphers. If the cipher wants to handle IVs or keys
16433 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16436 Change lots of functions like EVP_EncryptUpdate() to now return a
16437 value: although software versions of the algorithms cannot fail
16438 any installed hardware versions can.
16442 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16443 this option is set, tolerate broken clients that send the negotiated
16444 protocol version number instead of the requested protocol version
16449 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16450 i.e. non-zero for export ciphersuites, zero otherwise.
16451 Previous versions had this flag inverted, inconsistent with
16452 rsa_tmp_cb (..._TMP_RSA_CB).
16454 *Bodo Moeller; problem reported by Amit Chopra*
16456 * Add missing DSA library text string. Work around for some IIS
16457 key files with invalid SEQUENCE encoding.
16461 * Add a document (doc/standards.txt) that list all kinds of standards
16462 and so on that are implemented in OpenSSL.
16466 * Enhance c_rehash script. Old version would mishandle certificates
16467 with the same subject name hash and wouldn't handle CRLs at all.
16468 Added -fingerprint option to crl utility, to support new c_rehash
16473 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16477 * Fix for SSL server purpose checking. Server checking was
16478 rejecting certificates which had extended key usage present
16479 but no ssl client purpose.
16481 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16483 * Make PKCS#12 code work with no password. The PKCS#12 spec
16484 is a little unclear about how a blank password is handled.
16485 Since the password in encoded as a BMPString with terminating
16486 double NULL a zero length password would end up as just the
16487 double NULL. However no password at all is different and is
16488 handled differently in the PKCS#12 key generation code. NS
16489 treats a blank password as zero length. MSIE treats it as no
16490 password on export: but it will try both on import. We now do
16491 the same: PKCS12_parse() tries zero length and no password if
16492 the password is set to "" or NULL (NULL is now a valid password:
16493 it wasn't before) as does the pkcs12 application.
16497 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16498 perror when PEM_read_bio_X509_REQ fails, the error message must
16499 be obtained from the error queue.
16503 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16504 it in ERR_remove_state if appropriate, and change ERR_get_state
16505 accordingly to avoid race conditions (this is necessary because
16506 thread_hash is no longer constant once set).
16510 * Bugfix for linux-elf makefile.one.
16514 * RSA_get_default_method() will now cause a default
16515 RSA_METHOD to be chosen if one doesn't exist already.
16516 Previously this was only set during a call to RSA_new()
16517 or RSA_new_method(NULL) meaning it was possible for
16518 RSA_get_default_method() to return NULL.
16522 * Added native name translation to the existing DSO code
16523 that will convert (if the flag to do so is set) filenames
16524 that are sufficiently small and have no path information
16525 into a canonical native form. Eg. "blah" converted to
16526 "libblah.so" or "blah.dll" etc.
16530 * New function ERR_error_string_n(e, buf, len) which is like
16531 ERR_error_string(e, buf), but writes at most 'len' bytes
16532 including the 0 terminator. For ERR_error_string_n, 'buf'
16535 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16537 * CONF library reworked to become more general. A new CONF
16538 configuration file reader "class" is implemented as well as a
16539 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16540 old `CONF_*` functions are still there, but are reimplemented to
16541 work in terms of the new functions. Also, a set of functions
16542 to handle the internal storage of the configuration data is
16543 provided to make it easier to write new configuration file
16544 reader "classes" (I can definitely see something reading a
16545 configuration file in XML format, for example), called `_CONF_*`,
16546 or "the configuration storage API"...
16548 The new configuration file reading functions are:
16550 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16551 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16553 NCONF_default, NCONF_WIN32
16555 NCONF_dump_fp, NCONF_dump_bio
16557 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16558 NCONF_new creates a new CONF object. This works in the same way
16559 as other interfaces in OpenSSL, like the BIO interface.
16560 `NCONF_dump_*` dump the internal storage of the configuration file,
16561 which is useful for debugging. All other functions take the same
16562 arguments as the old `CONF_*` functions with the exception of the
16563 first that must be a `CONF *` instead of a `LHASH *`.
16565 To make it easier to use the new classes with the old `CONF_*` functions,
16566 the function CONF_set_default_method is provided.
16570 * Add '-tls1' option to 'openssl ciphers', which was already
16571 mentioned in the documentation but had not been implemented.
16572 (This option is not yet really useful because even the additional
16573 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16577 * Initial DSO code added into libcrypto for letting OpenSSL (and
16578 OpenSSL-based applications) load shared libraries and bind to
16579 them in a portable way.
16581 *Geoff Thorpe, with contributions from Richard Levitte*
16583 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16585 * Make sure _lrotl and _lrotr are only used with MSVC.
16587 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16588 (the default implementation of RAND_status).
16590 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16591 to '-clrext' (= clear extensions), as intended and documented.
16592 *Bodo Moeller; inconsistency pointed out by Michael Attili
16593 <attili@amaxo.com>*
16595 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16596 was larger than the MD block size.
16598 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16600 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16601 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16602 using the passed key: if the passed key was a private key the result
16603 of X509_print(), for example, would be to print out all the private key
16608 * des_quad_cksum() byte order bug fix.
16609 *Ulf Möller, using the problem description in krb4-0.9.7, where
16610 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16612 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16615 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16617 * For easily testing in shell scripts whether some command
16618 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16619 returns with exit code 0 iff no command of the given name is available.
16620 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16621 the output goes to stdout and nothing is printed to stderr.
16622 Additional arguments are always ignored.
16624 Since for each cipher there is a command of the same name,
16625 the 'no-cipher' compilation switches can be tested this way.
16627 ('openssl no-XXX' is not able to detect pseudo-commands such
16628 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16632 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16636 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16637 is set; it will be thrown away anyway because each handshake creates
16639 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16640 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16641 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16642 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16646 * New s_client option -ign_eof: EOF at stdin is ignored, and
16647 'Q' and 'R' lose their special meanings (quit/renegotiate).
16648 This is part of what -quiet does; unlike -quiet, -ign_eof
16649 does not suppress any output.
16653 * Add compatibility options to the purpose and trust code. The
16654 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16655 accepts a certificate or CA, this was the previous behaviour,
16656 with all the associated security issues.
16658 X509_TRUST_COMPAT is the old trust behaviour: only and
16659 automatically trust self signed roots in certificate store. A
16660 new trust setting X509_TRUST_DEFAULT is used to specify that
16661 a purpose has no associated trust setting and it should instead
16662 use the value in the default purpose.
16666 * Fix the PKCS#8 DSA private key code so it decodes keys again
16667 and fix a memory leak.
16671 * In util/mkerr.pl (which implements 'make errors'), preserve
16672 reason strings from the previous version of the .c file, as
16673 the default to have only downcase letters (and digits) in
16674 automatically generated reasons codes is not always appropriate.
16678 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16679 using strerror. Previously, ERR_reason_error_string() returned
16680 library names as reason strings for SYSerr; but SYSerr is a special
16681 case where small numbers are errno values, not library numbers.
16685 * Add '-dsaparam' option to 'openssl dhparam' application. This
16686 converts DSA parameters into DH parameters. (When creating parameters,
16687 DSA_generate_parameters is used.)
16691 * Include 'length' (recommended exponent length) in C code generated
16692 by 'openssl dhparam -C'.
16696 * The second argument to set_label in perlasm was already being used
16697 so couldn't be used as a "file scope" flag. Moved to third argument
16702 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16703 instead of RAND_bytes for encryption IVs and salts.
16707 * Include RAND_status() into RAND_METHOD instead of implementing
16708 it only for md_rand.c Otherwise replacing the PRNG by calling
16709 RAND_set_rand_method would be impossible.
16713 * Don't let DSA_generate_key() enter an infinite loop if the random
16714 number generation fails.
16718 * New 'rand' application for creating pseudo-random output.
16722 * Added configuration support for Linux/IA64
16724 *Rolf Haberrecker <rolf@suse.de>*
16726 * Assembler module support for Mingw32.
16730 * Shared library support for HPUX (in shlib/).
16732 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16734 * Shared library support for Solaris gcc.
16736 *Lutz Behnke <behnke@trustcenter.de>*
16738 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16740 * PKCS7_encrypt() was adding text MIME headers twice because they
16741 were added manually and by SMIME_crlf_copy().
16745 * In bntest.c don't call BN_rand with zero bits argument.
16747 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16749 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16750 case was implemented. This caused BN_div_recp() to fail occasionally.
16754 * Add an optional second argument to the set_label() in the perl
16755 assembly language builder. If this argument exists and is set
16756 to 1 it signals that the assembler should use a symbol whose
16757 scope is the entire file, not just the current function. This
16758 is needed with MASM which uses the format label:: for this scope.
16760 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16762 * Change the ASN1 types so they are typedefs by default. Before
16763 almost all types were #define'd to ASN1_STRING which was causing
16764 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16769 * Change names of new functions to the new get1/get0 naming
16770 convention: After 'get1', the caller owns a reference count
16771 and has to call `..._free`; 'get0' returns a pointer to some
16772 data structure without incrementing reference counters.
16773 (Some of the existing 'get' functions increment a reference
16774 counter, some don't.)
16775 Similarly, 'set1' and 'add1' functions increase reference
16776 counters or duplicate objects.
16780 * Allow for the possibility of temp RSA key generation failure:
16781 the code used to assume it always worked and crashed on failure.
16785 * Fix potential buffer overrun problem in BIO_printf().
16786 *Ulf Möller, using public domain code by Patrick Powell; problem
16787 pointed out by David Sacerdote <das33@cornell.edu>*
16789 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16790 RAND_egd() and RAND_status(). In the command line application,
16791 the EGD socket can be specified like a seed file using RANDFILE
16796 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16797 Some CAs (e.g. Verisign) distribute certificates in this form.
16801 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16802 list to exclude them. This means that no special compilation option
16803 is needed to use anonymous DH: it just needs to be included in the
16808 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16809 EVP_MD_type. The old functionality is available in a new macro called
16810 EVP_MD_md(). Change code that uses it and update docs.
16814 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16815 where the `void *` argument is replaced by a function pointer argument.
16816 Previously `void *` was abused to point to functions, which works on
16817 many platforms, but is not correct. As these functions are usually
16818 called by macros defined in OpenSSL header files, most source code
16819 should work without changes.
16823 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16824 sections with information on -D... compiler switches used for
16825 compiling the library so that applications can see them. To enable
16826 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16827 must be defined. E.g.,
16828 #define OPENSSL_ALGORITHM_DEFINES
16829 #include <openssl/opensslconf.h>
16830 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16832 *Richard Levitte, Ulf and Bodo Möller*
16834 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16839 * Change the 'other' type in certificate aux info to a STACK_OF
16840 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16841 the required ASN1 format: arbitrary types determined by an OID.
16845 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16846 argument to 'req'. This is not because the function is newer or
16847 better than others it just uses the work 'NEW' in the certificate
16848 request header lines. Some software needs this.
16852 * Reorganise password command line arguments: now passwords can be
16853 obtained from various sources. Delete the PEM_cb function and make
16854 it the default behaviour: i.e. if the callback is NULL and the
16855 usrdata argument is not NULL interpret it as a null terminated pass
16856 phrase. If usrdata and the callback are NULL then the pass phrase
16857 is prompted for as usual.
16861 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16862 the support is automatically enabled. The resulting binaries will
16863 autodetect the card and use it if present.
16865 *Ben Laurie and Compaq Inc.*
16867 * Work around for Netscape hang bug. This sends certificate request
16868 and server done in one record. Since this is perfectly legal in the
16869 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16870 the bugs/SSLv3 entry for more info.
16874 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16878 * Add -rand argument to smime and pkcs12 applications and read/write
16883 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16887 * Add command line password options to the remaining applications.
16891 * Bug fix for BN_div_recp() for numerators with an even number of
16896 * More tests in bntest.c, and changed test_bn output.
16900 * ./config recognizes MacOS X now.
16904 * Bug fix for BN_div() when the first words of num and divisor are
16905 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16909 * Add support for various broken PKCS#8 formats, and command line
16910 options to produce them.
16914 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16915 get temporary BIGNUMs from a BN_CTX.
16919 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16924 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16925 include a #define from the old name to the new. The original intent
16926 was that statically linked binaries could for example just call
16927 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16928 link with digests. This never worked because SSLeay_add_all_digests()
16929 and SSLeay_add_all_ciphers() were in the same source file so calling
16930 one would link with the other. They are now in separate source files.
16934 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16938 * Use a less unusual form of the Miller-Rabin primality test (it used
16939 a binary algorithm for exponentiation integrated into the Miller-Rabin
16940 loop, our standard modexp algorithms are faster).
16944 * Support for the EBCDIC character set completed.
16946 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16948 * Source code cleanups: use const where appropriate, eliminate casts,
16949 use `void *` instead of `char *` in lhash.
16953 * Bugfix: ssl3_send_server_key_exchange was not restartable
16954 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16955 this the server could overwrite ephemeral keys that the client
16960 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16961 using 50 iterations of the Rabin-Miller test.
16963 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16964 iterations of the Rabin-Miller test as required by the appendix
16965 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16966 As BN_is_prime_fasttest includes trial division, DSA parameter
16967 generation becomes much faster.
16969 This implies a change for the callback functions in DSA_is_prime
16970 and DSA_generate_parameters: The callback function is called once
16971 for each positive witness in the Rabin-Miller test, not just
16972 occasionally in the inner loop; and the parameters to the
16973 callback function now provide an iteration count for the outer
16974 loop rather than for the current invocation of the inner loop.
16975 DSA_generate_parameters additionally can call the callback
16976 function with an 'iteration count' of -1, meaning that a
16977 candidate has passed the trial division test (when q is generated
16978 from an application-provided seed, trial division is skipped).
16982 * New function BN_is_prime_fasttest that optionally does trial
16983 division before starting the Rabin-Miller test and has
16984 an additional BN_CTX * argument (whereas BN_is_prime always
16985 has to allocate at least one BN_CTX).
16986 'callback(1, -1, cb_arg)' is called when a number has passed the
16987 trial division stage.
16991 * Fix for bug in CRL encoding. The validity dates weren't being handled
16996 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17000 * New function BN_pseudo_rand().
17004 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17005 bignum version of BN_from_montgomery() with the working code from
17006 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17011 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17012 made it impossible to use the same SSL_SESSION data structure in
17013 SSL2 clients in multiple threads.
17017 * The return value of RAND_load_file() no longer counts bytes obtained
17018 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17019 to seed the PRNG (previously an explicit byte count was required).
17021 *Ulf Möller, Bodo Möller*
17023 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17024 used `char *` instead of `void *` and had casts all over the place.
17028 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17032 * Retain source code compatibility for BN_prime_checks macro:
17033 BN_is_prime(..., BN_prime_checks, ...) now uses
17034 BN_prime_checks_for_size to determine the appropriate number of
17035 Rabin-Miller iterations.
17039 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17040 DH_CHECK_P_NOT_SAFE_PRIME.
17041 (Check if this is true? OpenPGP calls them "strong".)
17045 * Merge the functionality of "dh" and "gendh" programs into a new program
17046 "dhparam". The old programs are retained for now but will handle DH keys
17047 (instead of parameters) in future.
17051 * Make the ciphers, s_server and s_client programs check the return values
17052 when a new cipher list is set.
17056 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17057 ciphers. Before when the 56bit ciphers were enabled the sorting was
17060 The syntax for the cipher sorting has been extended to support sorting by
17061 cipher-strength (using the strength_bits hard coded in the tables).
17062 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17064 Fix a bug in the cipher-command parser: when supplying a cipher command
17065 string with an "undefined" symbol (neither command nor alphanumeric
17066 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17067 an error is flagged.
17069 Due to the strength-sorting extension, the code of the
17070 ssl_create_cipher_list() function was completely rearranged. I hope that
17071 the readability was also increased :-)
17073 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17075 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17076 for the first serial number and places 2 in the serial number file. This
17077 avoids problems when the root CA is created with serial number zero and
17078 the first user certificate has the same issuer name and serial number
17083 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17084 the new code. Add documentation for this stuff.
17088 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17089 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17090 structures and behave in an analogous way to the X509v3 functions:
17091 they shouldn't be called directly but wrapper functions should be used
17094 So we also now have some wrapper functions that call the X509at functions
17095 when passed certificate requests. (TO DO: similar things can be done with
17096 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17097 things. Some of these need some d2i or i2d and print functionality
17098 because they handle more complex structures.)
17102 * Add missing #ifndefs that caused missing symbols when building libssl
17103 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17104 NO_RSA in `ssl/s2*.c`.
17106 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17108 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17109 has a return value which indicates the quality of the random data
17110 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17111 error queue. New function RAND_pseudo_bytes() generates output that is
17112 guaranteed to be unique but not unpredictable. RAND_add is like
17113 RAND_seed, but takes an extra argument for an entropy estimate
17114 (RAND_seed always assumes full entropy).
17118 * Do more iterations of Rabin-Miller probable prime test (specifically,
17119 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17120 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17121 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17122 false-positive rate of at most 2^-80 for random input.
17126 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17130 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17131 in the 0.9.5 release), this returns the chain
17132 from an X509_CTX structure with a dup of the stack and all
17133 the X509 reference counts upped: so the stack will exist
17134 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17137 Also make SSL_SESSION_print() print out the verify return
17142 * Add manpage for the pkcs12 command. Also change the default
17143 behaviour so MAC iteration counts are used unless the new
17144 -nomaciter option is used. This improves file security and
17145 only older versions of MSIE (4.0 for example) need it.
17149 * Honor the no-xxx Configure options when creating .DEF files.
17153 * Add PKCS#10 attributes to field table: challengePassword,
17154 unstructuredName and unstructuredAddress. These are taken from
17155 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17156 international characters are used.
17158 More changes to X509_ATTRIBUTE code: allow the setting of types
17159 based on strings. Remove the 'loc' parameter when adding
17160 attributes because these will be a SET OF encoding which is sorted
17165 * Initial changes to the 'req' utility to allow request generation
17166 automation. This will allow an application to just generate a template
17167 file containing all the field values and have req construct the
17170 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17171 used all over the place including certificate requests and PKCS#7
17172 structures. They are currently handled manually where necessary with
17173 some primitive wrappers for PKCS#7. The new functions behave in a
17174 manner analogous to the X509 extension functions: they allow
17175 attributes to be looked up by NID and added.
17177 Later something similar to the X509V3 code would be desirable to
17178 automatically handle the encoding, decoding and printing of the
17179 more complex types. The string types like challengePassword can
17180 be handled by the string table functions.
17182 Also modified the multi byte string table handling. Now there is
17183 a 'global mask' which masks out certain types. The table itself
17184 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17185 is useful when for example there is only one permissible type
17186 (as in countryName) and using the mask might result in no valid
17191 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17192 SSL_get_peer_finished to allow applications to obtain the latest
17193 Finished messages sent to the peer or expected from the peer,
17194 respectively. (SSL_get_peer_finished is usually the Finished message
17195 actually received from the peer, otherwise the protocol will be aborted.)
17197 As the Finished message are message digests of the complete handshake
17198 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17199 be used for external authentication procedures when the authentication
17200 provided by SSL/TLS is not desired or is not enough.
17204 * Enhanced support for Alpha Linux is added. Now ./config checks if
17205 the host supports BWX extension and if Compaq C is present on the
17206 $PATH. Just exploiting of the BWX extension results in 20-30%
17207 performance kick for some algorithms, e.g. DES and RC4 to mention
17208 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17213 * Add support for MS "fast SGC". This is arguably a violation of the
17214 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17215 weak crypto and after checking the certificate is SGC a second one
17216 with strong crypto. MS SGC stops the first handshake after receiving
17217 the server certificate message and sends a second client hello. Since
17218 a server will typically do all the time consuming operations before
17219 expecting any further messages from the client (server key exchange
17220 is the most expensive) there is little difference between the two.
17222 To get OpenSSL to support MS SGC we have to permit a second client
17223 hello message after we have sent server done. In addition we have to
17224 reset the MAC if we do get this second client hello.
17228 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17229 if a DER encoded private key is RSA or DSA traditional format. Changed
17230 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17231 format DER encoded private key. Newer code should use PKCS#8 format which
17232 has the key type encoded in the ASN1 structure. Added DER private key
17233 support to pkcs8 application.
17237 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17238 ciphersuites has been selected (as required by the SSL 3/TLS 1
17239 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17240 is set, we interpret this as a request to violate the specification
17241 (the worst that can happen is a handshake failure, and 'correct'
17242 behaviour would result in a handshake failure anyway).
17246 * In SSL_CTX_add_session, take into account that there might be multiple
17247 SSL_SESSION structures with the same session ID (e.g. when two threads
17248 concurrently obtain them from an external cache).
17249 The internal cache can handle only one SSL_SESSION with a given ID,
17250 so if there's a conflict, we now throw out the old one to achieve
17255 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17256 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17257 some routines that use cipher OIDs: some ciphers do not have OIDs
17258 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17263 * Simplify the trust setting structure and code. Now we just have
17264 two sequences of OIDs for trusted and rejected settings. These will
17265 typically have values the same as the extended key usage extension
17266 and any application specific purposes.
17268 The trust checking code now has a default behaviour: it will just
17269 check for an object with the same NID as the passed id. Functions can
17270 be provided to override either the default behaviour or the behaviour
17271 for a given id. SSL client, server and email already have functions
17272 in place for compatibility: they check the NID and also return "trusted"
17273 if the certificate is self signed.
17277 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17278 traditional format into an EVP_PKEY structure.
17282 * Add a password callback function PEM_cb() which either prompts for
17283 a password if usr_data is NULL or otherwise assumes it is a null
17284 terminated password. Allow passwords to be passed on command line
17285 environment or config files in a few more utilities.
17289 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17290 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17291 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17292 Update documentation.
17296 * Support for ASN1 "NULL" type. This could be handled before by using
17297 ASN1_TYPE but there wasn't any function that would try to read a NULL
17298 and produce an error if it couldn't. For compatibility we also have
17299 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17300 don't allocate anything because they don't need to.
17304 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17307 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17309 * Rebuild of the memory allocation routines used by OpenSSL code and
17310 possibly others as well. The purpose is to make an interface that
17311 provide hooks so anyone can build a separate set of allocation and
17312 deallocation routines to be used by OpenSSL, for example memory
17313 pool implementations, or something else, which was previously hard
17314 since Malloc(), Realloc() and Free() were defined as macros having
17315 the values malloc, realloc and free, respectively (except for Win32
17316 compilations). The same is provided for memory debugging code.
17317 OpenSSL already comes with functionality to find memory leaks, but
17318 this gives people a chance to debug other memory problems.
17320 With these changes, a new set of functions and macros have appeared:
17322 CRYPTO_set_mem_debug_functions() [F]
17323 CRYPTO_get_mem_debug_functions() [F]
17324 CRYPTO_dbg_set_options() [F]
17325 CRYPTO_dbg_get_options() [F]
17326 CRYPTO_malloc_debug_init() [M]
17328 The memory debug functions are NULL by default, unless the library
17329 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17330 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17331 gives the standard debugging functions that come with OpenSSL) or
17332 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17333 provided by the library user) must be used. When the standard
17334 debugging functions are used, CRYPTO_dbg_set_options can be used to
17335 request additional information:
17336 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17337 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17339 Also, things like CRYPTO_set_mem_functions will always give the
17340 expected result (the new set of functions is used for allocation
17341 and deallocation) at all times, regardless of platform and compiler
17344 To finish it up, some functions that were never use in any other
17345 way than through macros have a new API and new semantic:
17347 CRYPTO_dbg_malloc()
17348 CRYPTO_dbg_realloc()
17351 All macros of value have retained their old syntax.
17353 *Richard Levitte and Bodo Moeller*
17355 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17356 ordering of SMIMECapabilities wasn't in "strength order" and there
17357 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17362 * Some ASN1 types with illegal zero length encoding (INTEGER,
17363 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17365 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17367 * Merge in my S/MIME library for OpenSSL. This provides a simple
17368 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17369 functionality to handle multipart/signed properly) and a utility
17370 called 'smime' to call all this stuff. This is based on code I
17371 originally wrote for Celo who have kindly allowed it to be
17372 included in OpenSSL.
17376 * Add variants des_set_key_checked and des_set_key_unchecked of
17377 des_set_key (aka des_key_sched). Global variable des_check_key
17378 decides which of these is called by des_set_key; this way
17379 des_check_key behaves as it always did, but applications and
17380 the library itself, which was buggy for des_check_key == 1,
17381 have a cleaner way to pick the version they need.
17385 * New function PKCS12_newpass() which changes the password of a
17390 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17391 dynamic mix. In both cases the ids can be used as an index into the
17392 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17393 functions so they accept a list of the field values and the
17394 application doesn't need to directly manipulate the X509_TRUST
17399 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17404 * Modify the way the V3 extension code looks up extensions. This now
17405 works in a similar way to the object code: we have some "standard"
17406 extensions in a static table which is searched with OBJ_bsearch()
17407 and the application can add dynamic ones if needed. The file
17408 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17409 updated whenever a new extension is added to the core code and kept
17410 in ext_nid order. There is a simple program 'tabtest.c' which checks
17411 this. New extensions are not added too often so this file can readily
17412 be maintained manually.
17414 There are two big advantages in doing things this way. The extensions
17415 can be looked up immediately and no longer need to be "added" using
17416 X509V3_add_standard_extensions(): this function now does nothing.
17417 Side note: I get *lots* of email saying the extension code doesn't
17418 work because people forget to call this function.
17419 Also no dynamic allocation is done unless new extensions are added:
17420 so if we don't add custom extensions there is no need to call
17421 X509V3_EXT_cleanup().
17425 * Modify enc utility's salting as follows: make salting the default. Add a
17426 magic header, so unsalted files fail gracefully instead of just decrypting
17427 to garbage. This is because not salting is a big security hole, so people
17428 should be discouraged from doing it.
17432 * Fixes and enhancements to the 'x509' utility. It allowed a message
17433 digest to be passed on the command line but it only used this
17434 parameter when signing a certificate. Modified so all relevant
17435 operations are affected by the digest parameter including the
17436 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17437 DSA key was used because it didn't fix the digest.
17441 * Initial certificate chain verify code. Currently tests the untrusted
17442 certificates for consistency with the verify purpose (which is set
17443 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17445 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17446 this is because it will reject chains with invalid extensions whereas
17447 every previous version of OpenSSL and SSLeay made no checks at all.
17449 Trust code: checks the root CA for the relevant trust settings. Trust
17450 settings have an initial value consistent with the verify purpose: e.g.
17451 if the verify purpose is for SSL client use it expects the CA to be
17452 trusted for SSL client use. However the default value can be changed to
17453 permit custom trust settings: one example of this would be to only trust
17454 certificates from a specific "secure" set of CAs.
17456 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17457 which should be used for version portability: especially since the
17458 verify structure is likely to change more often now.
17460 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17461 to set them. If not set then assume SSL clients will verify SSL servers
17464 Two new options to the verify program: -untrusted allows a set of
17465 untrusted certificates to be passed in and -purpose which sets the
17466 intended purpose of the certificate. If a purpose is set then the
17467 new chain verify code is used to check extension consistency.
17471 * Support for the authority information access extension.
17475 * Modify RSA and DSA PEM read routines to transparently handle
17476 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17477 public keys in a format compatible with certificate
17478 SubjectPublicKeyInfo structures. Unfortunately there were already
17479 functions called *_PublicKey_* which used various odd formats so
17480 these are retained for compatibility: however the DSA variants were
17481 never in a public release so they have been deleted. Changed dsa/rsa
17482 utilities to handle the new format: note no releases ever handled public
17483 keys so we should be OK.
17485 The primary motivation for this change is to avoid the same fiasco
17486 that dogs private keys: there are several incompatible private key
17487 formats some of which are standard and some OpenSSL specific and
17488 require various evil hacks to allow partial transparent handling and
17489 even then it doesn't work with DER formats. Given the option anything
17490 other than PKCS#8 should be dumped: but the other formats have to
17491 stay in the name of compatibility.
17493 With public keys and the benefit of hindsight one standard format
17494 is used which works with EVP_PKEY, RSA or DSA structures: though
17495 it clearly returns an error if you try to read the wrong kind of key.
17497 Added a -pubkey option to the 'x509' utility to output the public key.
17498 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17499 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17500 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17501 that do the same as the `EVP_PKEY_assign_*()` except they up the
17502 reference count of the added key (they don't "swallow" the
17507 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17508 CRLs would fail if the file contained no certificates or no CRLs:
17509 added a new function to read in both types and return the number
17510 read: this means that if none are read it will be an error. The
17511 DER versions of the certificate and CRL reader would always fail
17512 because it isn't possible to mix certificates and CRLs in DER format
17513 without choking one or the other routine. Changed this to just read
17514 a certificate: this is the best we can do. Also modified the code
17515 in `apps/verify.c` to take notice of return codes: it was previously
17516 attempting to read in certificates from NULL pointers and ignoring
17517 any errors: this is one reason why the cert and CRL reader seemed
17518 to work. It doesn't check return codes from the default certificate
17519 routines: these may well fail if the certificates aren't installed.
17523 * Code to support otherName option in GeneralName.
17527 * First update to verify code. Change the verify utility
17528 so it warns if it is passed a self signed certificate:
17529 for consistency with the normal behaviour. X509_verify
17530 has been modified to it will now verify a self signed
17531 certificate if *exactly* the same certificate appears
17532 in the store: it was previously impossible to trust a
17533 single self signed certificate. This means that:
17534 openssl verify ss.pem
17535 now gives a warning about a self signed certificate but
17536 openssl verify -CAfile ss.pem ss.pem
17541 * For servers, store verify_result in SSL_SESSION data structure
17542 (and add it to external session representation).
17543 This is needed when client certificate verifications fails,
17544 but an application-provided verification callback (set by
17545 SSL_CTX_set_cert_verify_callback) allows accepting the session
17546 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17547 but returns 1): When the session is reused, we have to set
17548 ssl->verify_result to the appropriate error code to avoid
17551 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17553 * Fix a bug in the new PKCS#7 code: it didn't consider the
17554 case in PKCS7_dataInit() where the signed PKCS7 structure
17555 didn't contain any existing data because it was being created.
17557 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17559 * Add a salt to the key derivation routines in enc.c. This
17560 forms the first 8 bytes of the encrypted file. Also add a
17561 -S option to allow a salt to be input on the command line.
17565 * New function X509_cmp(). Oddly enough there wasn't a function
17566 to compare two certificates. We do this by working out the SHA1
17567 hash and comparing that. X509_cmp() will be needed by the trust
17572 * SSL_get1_session() is like SSL_get_session(), but increments
17573 the reference count in the SSL_SESSION returned.
17575 *Geoff Thorpe <geoff@eu.c2.net>*
17577 * Fix for 'req': it was adding a null to request attributes.
17578 Also change the X509_LOOKUP and X509_INFO code to handle
17579 certificate auxiliary information.
17583 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17588 * Add the possibility to add extra information to the memory leak
17589 detecting output, to form tracebacks, showing from where each
17590 allocation was originated: CRYPTO_push_info("constant string") adds
17591 the string plus current file name and line number to a per-thread
17592 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17593 is like calling CYRPTO_pop_info() until the stack is empty.
17594 Also updated memory leak detection code to be multi-thread-safe.
17598 * Add options -text and -noout to pkcs7 utility and delete the
17599 encryption options which never did anything. Update docs.
17603 * Add options to some of the utilities to allow the pass phrase
17604 to be included on either the command line (not recommended on
17605 OSes like Unix) or read from the environment. Update the
17606 manpages and fix a few bugs.
17610 * Add a few manpages for some of the openssl commands.
17614 * Fix the -revoke option in ca. It was freeing up memory twice,
17615 leaking and not finding already revoked certificates.
17619 * Extensive changes to support certificate auxiliary information.
17620 This involves the use of X509_CERT_AUX structure and X509_AUX
17621 functions. An X509_AUX function such as PEM_read_X509_AUX()
17622 can still read in a certificate file in the usual way but it
17623 will also read in any additional "auxiliary information". By
17624 doing things this way a fair degree of compatibility can be
17625 retained: existing certificates can have this information added
17626 using the new 'x509' options.
17628 Current auxiliary information includes an "alias" and some trust
17629 settings. The trust settings will ultimately be used in enhanced
17630 certificate chain verification routines: currently a certificate
17631 can only be trusted if it is self signed and then it is trusted
17636 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17637 The problem was that one of the replacement routines had not been working
17638 since SSLeay releases. For now the offending routine has been replaced
17639 with non-optimised assembler. Even so, this now gives around 95%
17640 performance improvement for 1024 bit RSA signs.
17644 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17645 handling. Most clients have the effective key size in bits equal to
17646 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17647 A few however don't do this and instead use the size of the decrypted key
17648 to determine the RC2 key length and the AlgorithmIdentifier to determine
17649 the effective key length. In this case the effective key length can still
17650 be 40 bits but the key length can be 168 bits for example. This is fixed
17651 by manually forcing an RC2 key into the EVP_PKEY structure because the
17652 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17653 the key length and effective key length are equal.
17657 * Add a bunch of functions that should simplify the creation of
17658 X509_NAME structures. Now you should be able to do:
17659 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17660 and have it automatically work out the correct field type and fill in
17661 the structures. The more adventurous can try:
17662 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17663 and it will (hopefully) work out the correct multibyte encoding.
17667 * Change the 'req' utility to use the new field handling and multibyte
17668 copy routines. Before the DN field creation was handled in an ad hoc
17669 way in req, ca, and x509 which was rather broken and didn't support
17670 BMPStrings or UTF8Strings. Since some software doesn't implement
17671 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17672 using the dirstring_type option. See the new comment in the default
17673 openssl.cnf for more info.
17677 * Make crypto/rand/md_rand.c more robust:
17678 - Assure unique random numbers after fork().
17679 - Make sure that concurrent threads access the global counter and
17680 md serializably so that we never lose entropy in them
17681 or use exactly the same state in multiple threads.
17682 Access to the large state is not always serializable because
17683 the additional locking could be a performance killer, and
17684 md should be large enough anyway.
17688 * New file `apps/app_rand.c` with commonly needed functionality
17689 for handling the random seed file.
17691 Use the random seed file in some applications that previously did not:
17693 dsaparam -genkey (which also ignored its '-rand' option),
17696 x509 (when signing).
17697 Except on systems with /dev/urandom, it is crucial to have a random
17698 seed file at least for key creation, DSA signing, and for DH exchanges;
17699 for RSA signatures we could do without one.
17701 gendh and gendsa (unlike genrsa) used to read only the first byte
17702 of each file listed in the '-rand' option. The function as previously
17703 found in genrsa is now in app_rand.c and is used by all programs
17704 that support '-rand'.
17708 * In RAND_write_file, use mode 0600 for creating files;
17709 don't just chmod when it may be too late.
17713 * Report an error from X509_STORE_load_locations
17714 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17718 * New function ASN1_mbstring_copy() this copies a string in either
17719 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17720 into an ASN1_STRING type. A mask of permissible types is passed
17721 and it chooses the "minimal" type to use or an error if not type
17726 * Add function equivalents to the various macros in asn1.h. The old
17727 macros are retained with an `M_` prefix. Code inside the library can
17728 use the `M_` macros. External code (including the openssl utility)
17729 should *NOT* in order to be "shared library friendly".
17733 * Add various functions that can check a certificate's extensions
17734 to see if it usable for various purposes such as SSL client,
17735 server or S/MIME and CAs of these types. This is currently
17736 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17737 verification. Also added a -purpose flag to x509 utility to
17738 print out all the purposes.
17742 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17747 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17748 for, obtain and decode and extension and obtain its critical flag.
17749 This allows all the necessary extension code to be handled in a
17750 single function call.
17754 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17755 platforms. See crypto/rc4/rc4_enc.c for further details.
17759 * New -noout option to asn1parse. This causes no output to be produced
17760 its main use is when combined with -strparse and -out to extract data
17761 from a file (which may not be in ASN.1 format).
17765 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17766 when producing the local key id.
17768 *Richard Levitte <levitte@stacken.kth.se>*
17770 * New option -dhparam in s_server. This allows a DH parameter file to be
17771 stated explicitly. If it is not stated then it tries the first server
17772 certificate file. The previous behaviour hard coded the filename
17777 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17778 a public key to be input or output. For example:
17779 openssl rsa -in key.pem -pubout -out pubkey.pem
17780 Also added necessary DSA public key functions to handle this.
17784 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17785 in the message. This was handled by allowing
17786 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17788 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17790 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17791 to the end of the strings whereas this didn't. This would cause problems
17792 if strings read with d2i_ASN1_bytes() were later modified.
17794 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17796 * Fix for base64 decode bug. When a base64 bio reads only one line of
17797 data and it contains EOF it will end up returning an error. This is
17798 caused by input 46 bytes long. The cause is due to the way base64
17799 BIOs find the start of base64 encoded data. They do this by trying a
17800 trial decode on each line until they find one that works. When they
17801 do a flag is set and it starts again knowing it can pass all the
17802 data directly through the decoder. Unfortunately it doesn't reset
17803 the context it uses. This means that if EOF is reached an attempt
17804 is made to pass two EOFs through the context and this causes the
17805 resulting error. This can also cause other problems as well. As is
17806 usual with these problems it takes *ages* to find and the fix is
17807 trivial: move one line.
17809 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17811 * Ugly workaround to get s_client and s_server working under Windows. The
17812 old code wouldn't work because it needed to select() on sockets and the
17813 tty (for keypresses and to see if data could be written). Win32 only
17814 supports select() on sockets so we select() with a 1s timeout on the
17815 sockets and then see if any characters are waiting to be read, if none
17816 are present then we retry, we also assume we can always write data to
17817 the tty. This isn't nice because the code then blocks until we've
17818 received a complete line of data and it is effectively polling the
17819 keyboard at 1s intervals: however it's quite a bit better than not
17820 working at all :-) A dedicated Windows application might handle this
17821 with an event loop for example.
17825 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17826 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17827 will be called when RSA_sign() and RSA_verify() are used. This is useful
17828 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17829 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17830 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17831 This necessitated the support of an extra signature type NID_md5_sha1
17832 for SSL signatures and modifications to the SSL library to use it instead
17833 of calling RSA_public_decrypt() and RSA_private_encrypt().
17837 * Add new -verify -CAfile and -CApath options to the crl program, these
17838 will lookup a CRL issuers certificate and verify the signature in a
17839 similar way to the verify program. Tidy up the crl program so it
17840 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17841 less strict. It will now permit CRL extensions even if it is not
17842 a V2 CRL: this will allow it to tolerate some broken CRLs.
17846 * Initialize all non-automatic variables each time one of the openssl
17847 sub-programs is started (this is necessary as they may be started
17848 multiple times from the "OpenSSL>" prompt).
17850 *Lennart Bang, Bodo Moeller*
17852 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17853 removing all other RSA functionality (this is what NO_RSA does). This
17854 is so (for example) those in the US can disable those operations covered
17855 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17860 * Non-copying interface to BIO pairs.
17861 (still largely untested)
17865 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17866 ASCII string. This was handled independently in various places before.
17870 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17871 UTF8 strings a character at a time.
17875 * Use client_version from client hello to select the protocol
17876 (s23_srvr.c) and for RSA client key exchange verification
17877 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17881 * Add various utility functions to handle SPKACs, these were previously
17882 handled by poking round in the structure internals. Added new function
17883 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17884 print, verify and generate SPKACs. Based on an original idea from
17885 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17889 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17893 * Allow the config file extension section to be overwritten on the
17894 command line. Based on an original idea from Massimiliano Pala
17895 <madwolf@comune.modena.it>. The new option is called -extensions
17896 and can be applied to ca, req and x509. Also -reqexts to override
17897 the request extensions in req and -crlexts to override the crl extensions
17902 * Add new feature to the SPKAC handling in ca. Now you can include
17903 the same field multiple times by preceding it by "XXXX." for example:
17906 this is the same syntax as used in the req config file.
17910 * Allow certificate extensions to be added to certificate requests. These
17911 are specified in a 'req_extensions' option of the req section of the
17912 config file. They can be printed out with the -text option to req but
17913 are otherwise ignored at present.
17917 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17918 data read consists of only the final block it would not decrypted because
17919 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17920 A misplaced 'break' also meant the decrypted final block might not be
17921 copied until the next read.
17925 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17926 a few extra parameters to the DH structure: these will be useful if
17927 for example we want the value of 'q' or implement X9.42 DH.
17931 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17932 provides hooks that allow the default DSA functions or functions on a
17933 "per key" basis to be replaced. This allows hardware acceleration and
17934 hardware key storage to be handled without major modification to the
17935 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17936 associated functions.
17940 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17941 as "read only": it can't be written to and the buffer it points to will
17942 not be freed. Reading from a read only BIO is much more efficient than
17943 a normal memory BIO. This was added because there are several times when
17944 an area of memory needs to be read from a BIO. The previous method was
17945 to create a memory BIO and write the data to it, this results in two
17946 copies of the data and an O(n^2) reading algorithm. There is a new
17947 function BIO_new_mem_buf() which creates a read only memory BIO from
17948 an area of memory. Also modified the PKCS#7 routines to use read only
17953 * Bugfix: ssl23_get_client_hello did not work properly when called in
17954 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17955 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17956 but a retry condition occurred while trying to read the rest.
17960 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17961 NID_pkcs7_encrypted by default: this was wrong since this should almost
17962 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17963 the encrypted data type: this is a more sensible place to put it and it
17964 allows the PKCS#12 code to be tidied up that duplicated this
17969 * Changed obj_dat.pl script so it takes its input and output files on
17970 the command line. This should avoid shell escape redirection problems
17975 * Initial support for certificate extension requests, these are included
17976 in things like Xenroll certificate requests. Included functions to allow
17977 extensions to be obtained and added.
17981 * -crlf option to s_client and s_server for sending newlines as
17982 CRLF (as required by many protocols).
17986 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17988 * Install libRSAglue.a when OpenSSL is built with RSAref.
17990 *Ralf S. Engelschall*
17992 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17994 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17996 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18001 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18002 DH parameters/keys (q is lost during that conversion, but the resulting
18003 DH parameters contain its length).
18005 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18006 much faster than DH_generate_parameters (which creates parameters
18007 where `p = 2*q + 1`), and also the smaller q makes DH computations
18008 much more efficient (160-bit exponentiation instead of 1024-bit
18009 exponentiation); so this provides a convenient way to support DHE
18010 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18011 utter importance to use
18012 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18014 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18015 when such DH parameters are used, because otherwise small subgroup
18016 attacks may become possible!
18020 * Avoid memory leak in i2d_DHparams.
18024 * Allow the -k option to be used more than once in the enc program:
18025 this allows the same encrypted message to be read by multiple recipients.
18029 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18030 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18031 it will always use the numerical form of the OID, even if it has a short
18036 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18037 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18038 otherwise bn_mod_exp was called. In the case of hardware keys for example
18039 no private key components need be present and it might store extra data
18040 in the RSA structure, which cannot be accessed from bn_mod_exp.
18041 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18042 private key operations.
18046 * Added support for SPARC Linux.
18050 * pem_password_cb function type incompatibly changed from
18051 typedef int pem_password_cb(char *buf, int size, int rwflag);
18053 ....(char *buf, int size, int rwflag, void *userdata);
18054 so that applications can pass data to their callbacks:
18055 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18056 additional void * argument, which is just handed through whenever
18057 the password callback is called.
18059 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18061 New function SSL_CTX_set_default_passwd_cb_userdata.
18063 Compatibility note: As many C implementations push function arguments
18064 onto the stack in reverse order, the new library version is likely to
18065 interoperate with programs that have been compiled with the old
18066 pem_password_cb definition (PEM_whatever takes some data that
18067 happens to be on the stack as its last argument, and the callback
18068 just ignores this garbage); but there is no guarantee whatsoever that
18071 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18072 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18073 problems not only on Windows, but also on some Unix platforms.
18074 To avoid problematic command lines, these definitions are now in an
18075 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18076 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18080 * MIPS III/IV assembler module is reimplemented.
18084 * More DES library cleanups: remove references to srand/rand and
18085 delete an unused file.
18089 * Add support for the free Netwide assembler (NASM) under Win32,
18090 since not many people have MASM (ml) and it can be hard to obtain.
18091 This is currently experimental but it seems to work OK and pass all
18092 the tests. Check out INSTALL.W32 for info.
18096 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18097 without temporary keys kept an extra copy of the server key,
18098 and connections with temporary keys did not free everything in case
18103 * New function RSA_check_key and new openssl rsa option -check
18104 for verifying the consistency of RSA keys.
18106 *Ulf Moeller, Bodo Moeller*
18108 * Various changes to make Win32 compile work:
18109 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18110 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18111 comparison" warnings.
18112 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18116 * Add a debugging option to PKCS#5 v2 key generation function: when
18117 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18118 derived keys are printed to stderr.
18122 * Copy the flags in ASN1_STRING_dup().
18124 *Roman E. Pavlov <pre@mo.msk.ru>*
18126 * The x509 application mishandled signing requests containing DSA
18127 keys when the signing key was also DSA and the parameters didn't match.
18129 It was supposed to omit the parameters when they matched the signing key:
18130 the verifying software was then supposed to automatically use the CA's
18131 parameters if they were absent from the end user certificate.
18133 Omitting parameters is no longer recommended. The test was also
18134 the wrong way round! This was probably due to unusual behaviour in
18135 EVP_cmp_parameters() which returns 1 if the parameters match.
18136 This meant that parameters were omitted when they *didn't* match and
18137 the certificate was useless. Certificates signed with 'ca' didn't have
18140 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18142 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18143 The interface is as follows:
18144 Applications can use
18145 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18146 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18147 "off" is now the default.
18148 The library internally uses
18149 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18150 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18151 to disable memory-checking temporarily.
18153 Some inconsistent states that previously were possible (and were
18154 even the default) are now avoided.
18156 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18157 with each memory chunk allocated; this is occasionally more helpful
18158 than just having a counter.
18160 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18162 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18167 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18168 which largely parallels "options", but is for changing API behaviour,
18169 whereas "options" are about protocol behaviour.
18170 Initial "mode" flags are:
18172 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18173 a single record has been written.
18174 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18175 retries use the same buffer location.
18176 (But all of the contents must be
18181 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18184 * Fix problems with no-hmac etc.
18186 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18188 * New functions RSA_get_default_method(), RSA_set_method() and
18189 RSA_get_method(). These allows replacement of RSA_METHODs without having
18190 to mess around with the internals of an RSA structure.
18194 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18195 Also really enable memory leak checks in openssl.c and in some
18198 *Chad C. Mulligan, Bodo Moeller*
18200 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18201 up the length of negative integers. This has now been simplified to just
18202 store the length when it is first determined and use it later, rather
18203 than trying to keep track of where data is copied and updating it to
18205 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18207 * Add a new function PKCS7_signatureVerify. This allows the verification
18208 of a PKCS#7 signature but with the signing certificate passed to the
18209 function itself. This contrasts with PKCS7_dataVerify which assumes the
18210 certificate is present in the PKCS#7 structure. This isn't always the
18211 case: certificates can be omitted from a PKCS#7 structure and be
18212 distributed by "out of band" means (such as a certificate database).
18216 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18217 function prototypes in pem.h, also change util/mkdef.pl to add the
18218 necessary function names.
18222 * mk1mf.pl (used by Windows builds) did not properly read the
18223 options set by Configure in the top level Makefile, and Configure
18224 was not even able to write more than one option correctly.
18225 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18229 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18230 file to be loaded from a BIO or FILE pointer. The BIO version will
18231 for example allow memory BIOs to contain config info.
18235 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18236 Whoever hopes to achieve shared-library compatibility across versions
18237 must use this, not the compile-time macro.
18238 (Exercise 0.9.4: Which is the minimum library version required by
18240 Note: All this applies only to multi-threaded programs, others don't
18245 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18246 through a BIO pair triggered the default case, i.e.
18247 SSLerr(...,SSL_R_UNKNOWN_STATE).
18251 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18252 can use the SSL library even if none of the specific BIOs is
18257 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18258 for the encoded length.
18260 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18262 * Add initial documentation of the X509V3 functions.
18266 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18267 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18268 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18269 secure PKCS#8 private key format with a high iteration count.
18273 * Fix determination of Perl interpreter: A perl or perl5
18274 *directory* in $PATH was also accepted as the interpreter.
18276 *Ralf S. Engelschall*
18278 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18279 wrong with it but it was very old and did things like calling
18280 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18281 unusual formatting.
18285 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18286 to use the new extension code.
18290 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18291 with macros. This should make it easier to change their form, add extra
18292 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18297 * Add to configuration table a new entry that can specify an alternative
18298 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18299 according to Mark Crispin <MRC@Panda.COM>.
18303 * DES CBC did not update the IV. Weird.
18307 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18308 Changing the behaviour of the former might break existing programs --
18309 where IV updating is needed, des_ncbc_encrypt can be used.
18312 * When bntest is run from "make test" it drives bc to check its
18313 calculations, as well as internally checking them. If an internal check
18314 fails, it needs to cause bc to give a non-zero result or make test carries
18315 on without noticing the failure. Fixed.
18319 * DES library cleanups.
18323 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18324 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18325 ciphers. NOTE: although the key derivation function has been verified
18326 against some published test vectors it has not been extensively tested
18327 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18332 * Instead of "mkdir -p", which is not fully portable, use new
18333 Perl script "util/mkdir-p.pl".
18337 * Rewrite the way password based encryption (PBE) is handled. It used to
18338 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18339 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18340 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18341 the 'parameter' field of the AlgorithmIdentifier is passed to the
18342 underlying key generation function so it must do its own ASN1 parsing.
18343 This has also changed the EVP_PBE_CipherInit() function which now has a
18344 'parameter' argument instead of literal salt and iteration count values
18345 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18349 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18350 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18351 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18352 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18353 value was just used as a "magic string" and not used directly its
18354 value doesn't matter.
18358 * Introduce some semblance of const correctness to BN. Shame C doesn't
18363 * "linux-sparc64" configuration (ultrapenguin).
18365 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18366 "linux-sparc" configuration.
18368 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18370 * config now generates no-xxx options for missing ciphers.
18374 * Support the EBCDIC character set (work in progress).
18375 File ebcdic.c not yet included because it has a different license.
18377 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18379 * Support BS2000/OSD-POSIX.
18381 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18383 * Make callbacks for key generation use `void *` instead of `char *`.
18387 * Make S/MIME samples compile (not yet tested).
18391 * Additional typesafe stacks.
18395 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18399 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18401 * New configuration variant "sco5-gcc".
18403 * Updated some demos.
18405 *Sean O Riordain, Wade Scholine*
18407 * Add missing BIO_free at exit of pkcs12 application.
18411 * Fix memory leak in conf.c.
18415 * Updates for Win32 to assembler version of MD5.
18419 * Set #! path to perl in `apps/der_chop` to where we found it
18420 instead of using a fixed path.
18424 * SHA library changes for irix64-mips4-cc.
18428 * Improvements for VMS support.
18432 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18434 * Bignum library bug fix. IRIX 6 passes "make test" now!
18435 This also avoids the problems with SC4.2 and unpatched SC5.
18437 *Andy Polyakov <appro@fy.chalmers.se>*
18439 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18440 These are required because of the typesafe stack would otherwise break
18441 existing code. If old code used a structure member which used to be STACK
18442 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18443 sk_num or sk_value it would produce an error because the num, data members
18444 are not present in STACK_OF. Now it just produces a warning. sk_set
18445 replaces the old method of assigning a value to sk_value
18446 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18447 that does this will no longer work (and should use sk_set instead) but
18448 this could be regarded as a "questionable" behaviour anyway.
18452 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18453 correctly handle encrypted S/MIME data.
18457 * Change type of various DES function arguments from des_cblock
18458 (which means, in function argument declarations, pointer to char)
18459 to des_cblock * (meaning pointer to array with 8 char elements),
18460 which allows the compiler to do more typechecking; it was like
18461 that back in SSLeay, but with lots of ugly casts.
18463 Introduce new type const_des_cblock.
18467 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18468 problems: find RecipientInfo structure that matches recipient certificate
18469 and initialise the ASN1 structures properly based on passed cipher.
18473 * Belatedly make the BN tests actually check the results.
18477 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18478 to and from BNs: it was completely broken. New compilation option
18479 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18480 key elements as negative integers.
18484 * Reorganize and speed up MD5.
18486 *Andy Polyakov <appro@fy.chalmers.se>*
18490 *Richard Levitte <richard@levitte.org>*
18492 * New option -out to asn1parse to allow the parsed structure to be
18493 output to a file. This is most useful when combined with the -strparse
18494 option to examine the output of things like OCTET STRINGS.
18498 * Make SSL library a little more fool-proof by not requiring any longer
18499 that `SSL_set_{accept,connect}_state` be called before
18500 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18501 in many applications because usually everything *appeared* to work as
18502 intended anyway -- now it really works as intended).
18506 * Move openssl.cnf out of lib/.
18510 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18511 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18512 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18514 *Ralf S. Engelschall*
18516 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18517 handle PKCS#7 enveloped data properly.
18519 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18521 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18522 copying pointers. The cert_st handling is changed by this in
18523 various ways (and thus what used to be known as ctx->default_cert
18524 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18525 any longer when s->cert does not give us what we need).
18526 ssl_cert_instantiate becomes obsolete by this change.
18527 As soon as we've got the new code right (possibly it already is?),
18528 we have solved a couple of bugs of the earlier code where s->cert
18529 was used as if it could not have been shared with other SSL structures.
18531 Note that using the SSL API in certain dirty ways now will result
18532 in different behaviour than observed with earlier library versions:
18533 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18534 does not influence s as it used to.
18536 In order to clean up things more thoroughly, inside SSL_SESSION
18537 we don't use CERT any longer, but a new structure SESS_CERT
18538 that holds per-session data (if available); currently, this is
18539 the peer's certificate chain and, for clients, the server's certificate
18540 and temporary key. CERT holds only those values that can have
18541 meaningful defaults in an SSL_CTX.
18545 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18546 from the internal representation. Various PKCS#7 fixes: remove some
18547 evil casts and set the enc_dig_alg field properly based on the signing
18552 * Allow PKCS#12 password to be set from the command line or the
18553 environment. Let 'ca' get its config file name from the environment
18554 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18559 * Allow certificate policies extension to use an IA5STRING for the
18560 organization field. This is contrary to the PKIX definition but
18561 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18566 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18567 without disallowing inline assembler and the like for non-pedantic builds.
18571 * Support Borland C++ builder.
18573 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18579 * SHA-1 cleanups and performance enhancements.
18581 *Andy Polyakov <appro@fy.chalmers.se>*
18583 * Sparc v8plus assembler for the bignum library.
18585 *Andy Polyakov <appro@fy.chalmers.se>*
18587 * Accept any -xxx and +xxx compiler options in Configure.
18591 * Update HPUX configuration.
18595 * Add missing `sk_<type>_unshift()` function to safestack.h
18597 *Ralf S. Engelschall*
18599 * New function SSL_CTX_use_certificate_chain_file that sets the
18600 "extra_cert"s in addition to the certificate. (This makes sense
18601 only for "PEM" format files, as chains as a whole are not
18606 * Support verify_depth from the SSL API.
18607 x509_vfy.c had what can be considered an off-by-one-error:
18608 Its depth (which was not part of the external interface)
18609 was actually counting the number of certificates in a chain;
18610 now it really counts the depth.
18614 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18615 instead of X509err, which often resulted in confusing error
18616 messages since the error codes are not globally unique
18617 (e.g. an alleged error in ssl3_accept when a certificate
18618 didn't match the private key).
18620 * New function SSL_CTX_set_session_id_context that allows to set a default
18621 value (so that you don't need SSL_set_session_id_context for each
18622 connection using the SSL_CTX).
18626 * OAEP decoding bug fix.
18630 * Support INSTALL_PREFIX for package builders, as proposed by
18635 * New Configure options "threads" and "no-threads". For systems
18636 where the proper compiler options are known (currently Solaris
18637 and Linux), "threads" is the default.
18641 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18645 * Install various scripts to $(OPENSSLDIR)/misc, not to
18646 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18647 such as /usr/local/bin.
18651 * "make linux-shared" to build shared libraries.
18653 *Niels Poppe <niels@netbox.org>*
18655 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18659 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18660 extension adding in x509 utility.
18664 * Remove NOPROTO sections and error code comments.
18668 * Partial rewrite of the DEF file generator to now parse the ANSI
18673 * New Configure options --prefix=DIR and --openssldir=DIR.
18677 * Complete rewrite of the error code script(s). It is all now handled
18678 by one script at the top level which handles error code gathering,
18679 header rewriting and C source file generation. It should be much better
18680 than the old method: it now uses a modified version of Ulf's parser to
18681 read the ANSI prototypes in all header files (thus the old K&R definitions
18682 aren't needed for error creation any more) and do a better job of
18683 translating function codes into names. The old 'ASN1 error code embedded
18684 in a comment' is no longer necessary and it doesn't use .err files which
18685 have now been deleted. Also the error code call doesn't have to appear all
18686 on one line (which resulted in some large lines...).
18690 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18694 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18695 0 (which usually indicates a closed connection), but continue reading.
18699 * Fix some race conditions.
18703 * Add support for CRL distribution points extension. Add Certificate
18704 Policies and CRL distribution points documentation.
18708 * Move the autogenerated header file parts to crypto/opensslconf.h.
18712 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18713 8 of keying material. Merlin has also confirmed interop with this fix
18714 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18716 *Merlin Hughes <merlin@baltimore.ie>*
18718 * Fix lots of warnings.
18720 *Richard Levitte <levitte@stacken.kth.se>*
18722 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18723 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18725 *Richard Levitte <levitte@stacken.kth.se>*
18727 * Fix problems with sizeof(long) == 8.
18729 *Andy Polyakov <appro@fy.chalmers.se>*
18731 * Change functions to ANSI C.
18735 * Fix typos in error codes.
18737 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18739 * Remove defunct assembler files from Configure.
18743 * SPARC v8 assembler BIGNUM implementation.
18745 *Andy Polyakov <appro@fy.chalmers.se>*
18747 * Support for Certificate Policies extension: both print and set.
18748 Various additions to support the r2i method this uses.
18752 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18753 return a const string when you are expecting an allocated buffer.
18757 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18758 types DirectoryString and DisplayText.
18762 * Add code to allow r2i extensions to access the configuration database,
18763 add an LHASH database driver and add several ctx helper functions.
18767 * Fix an evil bug in bn_expand2() which caused various BN functions to
18768 fail when they extended the size of a BIGNUM.
18772 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18773 support typesafe stack.
18777 * Fix typo in SSL_[gs]et_options().
18779 *Nils Frostberg <nils@medcom.se>*
18781 * Delete various functions and files that belonged to the (now obsolete)
18782 old X509V3 handling code.
18786 * New Configure option "rsaref".
18790 * Don't auto-generate pem.h.
18794 * Introduce type-safe ASN.1 SETs.
18798 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18800 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18802 * Introduce type-safe STACKs. This will almost certainly break lots of code
18803 that links with OpenSSL (well at least cause lots of warnings), but fear
18804 not: the conversion is trivial, and it eliminates loads of evil casts. A
18805 few STACKed things have been converted already. Feel free to convert more.
18806 In the fullness of time, I'll do away with the STACK type altogether.
18810 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18811 specified in `<certfile>` by updating the entry in the index.txt file.
18812 This way one no longer has to edit the index.txt file manually for
18813 revoking a certificate. The -revoke option does the gory details now.
18815 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18817 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18818 `-text` option at all and this way the `-noout -text` combination was
18819 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18821 *Ralf S. Engelschall*
18823 * Make sure a corresponding plain text error message exists for the
18824 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18825 verify callback function determined that a certificate was revoked.
18827 *Ralf S. Engelschall*
18829 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18830 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18831 all available ciphers including rc5, which was forgotten until now.
18832 In order to let the testing shell script know which algorithms
18833 are available, a new (up to now undocumented) command
18834 `openssl list-cipher-commands` is used.
18838 * Bugfix: s_client occasionally would sleep in select() when
18839 it should have checked SSL_pending() first.
18843 * New functions DSA_do_sign and DSA_do_verify to provide access to
18844 the raw DSA values prior to ASN.1 encoding.
18848 * Tweaks to Configure
18850 *Niels Poppe <niels@netbox.org>*
18852 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18857 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18861 * New config option to avoid instructions that are illegal on the 80386.
18862 The default code is faster, but requires at least a 486.
18866 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18867 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18868 same as SSL2_VERSION anyway.
18872 * New "-showcerts" option for s_client.
18876 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18877 application. Various cleanups and fixes.
18881 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18882 modify error routines to work internally. Add error codes and PBE init
18883 to library startup routines.
18887 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18888 packing functions to asn1 and evp. Changed function names and error
18889 codes along the way.
18893 * PKCS12 integration: and so it begins... First of several patches to
18894 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18895 objects to objects.h
18899 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18900 and display support for Thawte strong extranet extension.
18904 * Add LinuxPPC support.
18906 *Jeff Dubrule <igor@pobox.org>*
18908 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18909 bn_div_words in alpha.s.
18911 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18913 * Make sure the RSA OAEP test is skipped under -DRSAref because
18914 OAEP isn't supported when OpenSSL is built with RSAref.
18916 *Ulf Moeller <ulf@fitug.de>*
18918 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18919 so they no longer are missing under -DNOPROTO.
18921 *Soren S. Jorvang <soren@t.dk>*
18923 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18925 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18926 doesn't work when the session is reused. Coming soon!
18930 * Fix a security hole, that allows sessions to be reused in the wrong
18931 context thus bypassing client cert protection! All software that uses
18932 client certs and session caches in multiple contexts NEEDS PATCHING to
18933 allow session reuse! A fuller solution is in the works.
18935 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18937 * Some more source tree cleanups (removed obsolete files
18938 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18939 permission on "config" script to be executable) and a fix for the INSTALL
18942 *Ulf Moeller <ulf@fitug.de>*
18944 * Remove some legacy and erroneous uses of malloc, free instead of
18947 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18949 * Make rsa_oaep_test return non-zero on error.
18951 *Ulf Moeller <ulf@fitug.de>*
18953 * Add support for native Solaris shared libraries. Configure
18954 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18955 if someone would make that last step automatic.
18957 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18959 * ctx_size was not built with the right compiler during "make links". Fixed.
18963 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18964 except NULL ciphers". This means the default cipher list will no longer
18965 enable NULL ciphers. They need to be specifically enabled e.g. with
18966 the string "DEFAULT:eNULL".
18970 * Fix to RSA private encryption routines: if p < q then it would
18971 occasionally produce an invalid result. This will only happen with
18972 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18976 * Be less restrictive and allow also `perl util/perlpath.pl
18977 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18978 because this way one can also use an interpreter named `perl5` (which is
18979 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18980 installed as `perl`).
18982 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18984 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18986 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18988 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18989 advapi32.lib to Win32 build and change the pem test comparison
18990 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18991 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18992 and crypto/des/ede_cbcm_enc.c.
18996 * DES quad checksum was broken on big-endian architectures. Fixed.
19000 * Comment out two functions in bio.h that aren't implemented. Fix up the
19001 Win32 test batch file so it (might) work again. The Win32 test batch file
19002 is horrible: I feel ill....
19006 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19007 in e_os.h. Audit of header files to check ANSI and non ANSI
19008 sections: 10 functions were absent from non ANSI section and not exported
19009 from Windows DLLs. Fixed up libeay.num for new functions.
19013 * Make `openssl version` output lines consistent.
19015 *Ralf S. Engelschall*
19017 * Fix Win32 symbol export lists for BIO functions: Added
19018 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19019 to ms/libeay{16,32}.def.
19021 *Ralf S. Engelschall*
19023 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19024 fine under Unix and passes some trivial tests I've now added. But the
19025 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19026 added to make sure no one expects that this stuff really works in the
19027 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19028 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19031 *Ralf S. Engelschall*
19033 * Fix the generation of two part addresses in perl.
19035 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19037 * Add config entry for Linux on MIPS.
19039 *John Tobey <jtobey@channel1.com>*
19041 * Make links whenever Configure is run, unless we are on Windoze.
19045 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19046 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19051 * Add a useful kludge to allow package maintainers to specify compiler and
19052 other platforms details on the command line without having to patch the
19053 Configure script every time: One now can use
19054 `perl Configure <id>:<details>`,
19055 i.e. platform ids are allowed to have details appended
19056 to them (separated by colons). This is treated as there would be a static
19057 pre-configured entry in Configure's %table under key `<id>` with value
19058 `<details>` and `perl Configure <id>` is called. So, when you want to
19059 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19060 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19061 now, which overrides the FreeBSD-elf entry on-the-fly.
19063 *Ralf S. Engelschall*
19065 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19069 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19070 on the `perl Configure ...` command line. This way one can compile
19071 OpenSSL libraries with Position Independent Code (PIC) which is needed
19072 for linking it into DSOs.
19074 *Ralf S. Engelschall*
19076 * Remarkably, export ciphers were totally broken and no-one had noticed!
19081 * Cleaned up the LICENSE document: The official contact for any license
19082 questions now is the OpenSSL core team under openssl-core@openssl.org.
19083 And add a paragraph about the dual-license situation to make sure people
19084 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19085 to the OpenSSL toolkit.
19087 *Ralf S. Engelschall*
19089 * General source tree makefile cleanups: Made `making xxx in yyy...`
19090 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19091 Additionally cleaned up the `make links` target: Remove unnecessary
19092 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19093 to speed processing and no longer clutter the display with confusing
19094 stuff. Instead only the actually done links are displayed.
19096 *Ralf S. Engelschall*
19098 * Permit null encryption ciphersuites, used for authentication only. It used
19099 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19100 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19105 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19106 signed attributes when verifying signatures (this would break them),
19107 the detached data encoding was wrong and public keys obtained using
19108 X509_get_pubkey() weren't freed.
19112 * Add text documentation for the BUFFER functions. Also added a work around
19113 to a Win95 console bug. This was triggered by the password read stuff: the
19114 last character typed gets carried over to the next fread(). If you were
19115 generating a new cert request using 'req' for example then the last
19116 character of the passphrase would be CR which would then enter the first
19121 * Added the new 'Includes OpenSSL Cryptography Software' button as
19122 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19123 button and can be used by applications based on OpenSSL to show the
19124 relationship to the OpenSSL project.
19126 *Ralf S. Engelschall*
19128 * Remove confusing variables in function signatures in files
19129 ssl/ssl_lib.c and ssl/ssl.h.
19131 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19133 * Don't install bss_file.c under PREFIX/include/
19135 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19137 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19138 functions that return function pointers and has support for NT specific
19139 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19140 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19141 unsigned to signed types: this was killing the Win32 compile.
19145 * Add new certificate file to stack functions,
19146 SSL_add_dir_cert_subjects_to_stack() and
19147 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19148 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19149 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19150 This means that Apache-SSL and similar packages don't have to mess around
19151 to add as many CAs as they want to the preferred list.
19155 * Experiment with doxygen documentation. Currently only partially applied to
19157 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19158 openssl.doxy as the configuration file.
19162 * Get rid of remaining C++-style comments which strict C compilers hate.
19164 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19166 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19167 compiled in by default: it has problems with large keys.
19171 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19172 DH private keys and/or callback functions which directly correspond to
19173 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19174 is needed for applications which have to configure certificates on a
19175 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19177 For the RSA certificate situation is makes no difference, but
19178 for the DSA certificate situation this fixes the "no shared cipher"
19179 problem where the OpenSSL cipher selection procedure failed because the
19180 temporary keys were not overtaken from the context and the API provided
19181 no way to reconfigure them.
19182 The new functions now let applications reconfigure the stuff and they
19183 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19184 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19185 non-public-API function ssl_cert_instantiate() is used as a helper
19186 function and also to reduce code redundancy inside ssl_rsa.c.
19188 *Ralf S. Engelschall*
19190 * Move s_server -dcert and -dkey options out of the undocumented feature
19191 area because they are useful for the DSA situation and should be
19192 recognized by the users.
19194 *Ralf S. Engelschall*
19196 * Fix the cipher decision scheme for export ciphers: the export bits are
19197 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19198 SSL_EXP_MASK. So, the original variable has to be used instead of the
19199 already masked variable.
19201 *Richard Levitte <levitte@stacken.kth.se>*
19203 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19205 *Richard Levitte <levitte@stacken.kth.se>*
19207 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19208 from `int` to `unsigned int` because it is a length and initialized by
19209 EVP_DigestFinal() which expects an `unsigned int *`.
19211 *Richard Levitte <levitte@stacken.kth.se>*
19213 * Don't hard-code path to Perl interpreter on shebang line of Configure
19214 script. Instead use the usual Shell->Perl transition trick.
19216 *Ralf S. Engelschall*
19218 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19219 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19220 -noout -modulus` as it's already the case for `openssl rsa -noout
19221 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19222 currently the public key is printed (a decision which was already done by
19223 `openssl dsa -modulus` in the past) which serves a similar purpose.
19224 Additionally the NO_RSA no longer completely removes the whole -modulus
19225 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19228 *Ralf S. Engelschall*
19230 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19231 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19233 *Arne Ansper <arne@ats.cyber.ee>*
19235 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19236 to be added. Now both 'req' and 'ca' can use new objects defined in the
19241 * Add cool BIO that does syslog (or event log on NT).
19243 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19245 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19246 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19247 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19248 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19252 * Add preliminary config info for new extension code.
19256 * Make RSA_NO_PADDING really use no padding.
19258 *Ulf Moeller <ulf@fitug.de>*
19260 * Generate errors when private/public key check is done.
19264 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19265 for some CRL extensions and new objects added.
19269 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19270 key usage extension and fuller support for authority key id.
19274 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19275 padding method for RSA, which is recommended for new applications in PKCS
19276 #1 v2.0 (RFC 2437, October 1998).
19277 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19278 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19279 against Bleichbacher's attack on RSA.
19280 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19283 * Updates to the new SSL compression code
19285 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19287 * Fix so that the version number in the master secret, when passed
19288 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19289 (because the server will not accept higher), that the version number
19290 is 0x03,0x01, not 0x03,0x00
19292 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19294 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19295 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19296 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19300 * Support for RAW extensions where an arbitrary extension can be
19301 created by including its DER encoding. See `apps/openssl.cnf` for
19306 * Make sure latest Perl versions don't interpret some generated C array
19307 code as Perl array code in the crypto/err/err_genc.pl script.
19309 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19311 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19312 not many people have the assembler. Various Win32 compilation fixes and
19313 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19314 build instructions.
19318 * Modify configure script 'Configure' to automatically create crypto/date.h
19319 file under Win32 and also build pem.h from pem.org. New script
19320 util/mkfiles.pl to create the MINFO file on environments that can't do a
19321 'make files': perl util/mkfiles.pl >MINFO should work.
19325 * Major rework of DES function declarations, in the pursuit of correctness
19326 and purity. As a result, many evil casts evaporated, and some weirdness,
19327 too. You may find this causes warnings in your code. Zapping your evil
19328 casts will probably fix them. Mostly.
19332 * Fix for a typo in asn1.h. Bug fix to object creation script
19333 obj_dat.pl. It considered a zero in an object definition to mean
19334 "end of object": none of the objects in objects.h have any zeros
19335 so it wasn't spotted.
19337 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19339 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19340 Masking (CBCM). In the absence of test vectors, the best I have been able
19341 to do is check that the decrypt undoes the encrypt, so far. Send me test
19342 vectors if you have them.
19346 * Correct calculation of key length for export ciphers (too much space was
19347 allocated for null ciphers). This has not been tested!
19351 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19352 message is now correct (it understands "crypto" and "ssl" on its
19353 command line). There is also now an "update" option. This will update
19354 the util/ssleay.num and util/libeay.num files with any new functions.
19356 perl util/mkdef.pl crypto ssl update
19357 it will update them.
19361 * Overhauled the Perl interface:
19362 - ported BN stuff to OpenSSL's different BN library
19363 - made the perl/ source tree CVS-aware
19364 - renamed the package from SSLeay to OpenSSL (the files still contain
19365 their history because I've copied them in the repository)
19366 - removed obsolete files (the test scripts will be replaced
19367 by better Test::Harness variants in the future)
19369 *Ralf S. Engelschall*
19371 * First cut for a very conservative source tree cleanup:
19372 1. merge various obsolete readme texts into doc/ssleay.txt
19373 where we collect the old documents and readme texts.
19374 2. remove the first part of files where I'm already sure that we no
19375 longer need them because of three reasons: either they are just temporary
19376 files which were left by Eric or they are preserved original files where
19377 I've verified that the diff is also available in the CVS via "cvs diff
19378 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19379 the crypto/md/ stuff).
19381 *Ralf S. Engelschall*
19383 * More extension code. Incomplete support for subject and issuer alt
19384 name, issuer and authority key id. Change the i2v function parameters
19385 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19386 what that's for :-) Fix to ASN1 macro which messed up
19387 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19391 * Preliminary support for ENUMERATED type. This is largely copied from the
19396 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19398 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19400 * Make sure `make rehash` target really finds the `openssl` program.
19402 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19404 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19405 like to hear about it if this slows down other processors.
19409 * Add CygWin32 platform information to Configure script.
19411 *Alan Batie <batie@aahz.jf.intel.com>*
19413 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19415 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19417 * New program nseq to manipulate netscape certificate sequences
19421 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19426 * Fixes to BN code. Previously the default was to define BN_RECURSION
19427 but the BN code had some problems that would cause failures when
19428 doing certificate verification and some other functions.
19430 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19432 * Add ASN1 and PEM code to support netscape certificate sequences.
19436 * Add ASN1 and PEM code to support netscape certificate sequences.
19440 * Add several PKIX and private extended key usage OIDs.
19444 * Modify the 'ca' program to handle the new extension code. Modify
19445 openssl.cnf for new extension format, add comments.
19449 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19450 and add a sample to openssl.cnf so req -x509 now adds appropriate
19455 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19456 error code, add initial support to X509_print() and x509 application.
19460 * Takes a deep breath and start adding X509 V3 extension support code. Add
19461 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19462 stuff is currently isolated and isn't even compiled yet.
19466 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19467 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19468 Removed the versions check from X509 routines when loading extensions:
19469 this allows certain broken certificates that don't set the version
19470 properly to be processed.
19474 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19475 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19476 can still be regenerated with "make depend".
19480 * Spelling mistake in C version of CAST-128.
19482 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19484 * Changes to the error generation code. The perl script err-code.pl
19485 now reads in the old error codes and retains the old numbers, only
19486 adding new ones if necessary. It also only changes the .err files if new
19487 codes are added. The makefiles have been modified to only insert errors
19488 when needed (to avoid needlessly modifying header files). This is done
19489 by only inserting errors if the .err file is newer than the auto generated
19490 C file. To rebuild all the error codes from scratch (the old behaviour)
19491 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19492 or delete all the .err files.
19496 * CAST-128 was incorrectly implemented for short keys. The C version has
19497 been fixed, but is untested. The assembler versions are also fixed, but
19498 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19499 to regenerate it if needed.
19500 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19501 Hagino <itojun@kame.net>*
19503 * File was opened incorrectly in randfile.c.
19505 *Ulf Möller <ulf@fitug.de>*
19507 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19508 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19509 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19510 al: it's just almost always a UTCTime. Note this patch adds new error
19511 codes so do a "make errors" if there are problems.
19515 * Correct Linux 1 recognition in config.
19517 *Ulf Möller <ulf@fitug.de>*
19519 * Remove pointless MD5 hash when using DSA keys in ca.
19521 *Anonymous <nobody@replay.com>*
19523 * Generate an error if given an empty string as a cert directory. Also
19524 generate an error if handed NULL (previously returned 0 to indicate an
19525 error, but didn't set one).
19527 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19529 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19533 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19534 parameters. This was causing a warning which killed off the Win32 compile.
19538 * Remove C++ style comments from crypto/bn/bn_local.h.
19540 *Neil Costigan <neil.costigan@celocom.com>*
19542 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19543 based on a text string, looking up short and long names and finally
19544 "dot" format. The "dot" format stuff didn't work. Added new function
19545 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19546 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19547 OID is not part of the table.
19551 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19552 X509_LOOKUP_by_alias().
19556 * Sort openssl functions by name.
19560 * Get the `gendsa` command working and add it to the `list` command. Remove
19561 encryption from sample DSA keys (in case anyone is interested the password
19566 * Make *all* `*_free` functions accept a NULL pointer.
19568 *Frans Heymans <fheymans@isaserver.be>*
19570 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19573 *Anonymous <nobody@replay.com>*
19575 * s_server should send the CAfile as acceptable CAs, not its own cert.
19577 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19579 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19581 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19583 * Temp key "for export" tests were wrong in s3_srvr.c.
19585 *Anonymous <nobody@replay.com>*
19587 * Add prototype for temp key callback functions
19588 SSL_CTX_set_tmp_{rsa,dh}_callback().
19592 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19593 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19597 * X509_name_add_entry() freed the wrong thing after an error.
19599 *Arne Ansper <arne@ats.cyber.ee>*
19601 * rsa_eay.c would attempt to free a NULL context.
19603 *Arne Ansper <arne@ats.cyber.ee>*
19605 * BIO_s_socket() had a broken should_retry() on Windoze.
19607 *Arne Ansper <arne@ats.cyber.ee>*
19609 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19611 *Arne Ansper <arne@ats.cyber.ee>*
19613 * Make sure the already existing X509_STORE->depth variable is initialized
19614 in X509_STORE_new(), but document the fact that this variable is still
19615 unused in the certificate verification process.
19617 *Ralf S. Engelschall*
19619 * Fix the various library and `apps/` files to free up pkeys obtained from
19620 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19624 * Fix reference counting in X509_PUBKEY_get(). This makes
19625 demos/maurice/example2.c work, amongst others, probably.
19627 *Steve Henson and Ben Laurie*
19629 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19630 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19631 are no longer created. This way we have a single and consistent command
19632 line interface `openssl <command>`, similar to `cvs <command>`.
19634 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19636 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19637 BIT STRING wrapper always have zero unused bits.
19641 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19645 * Make the top-level INSTALL documentation easier to understand.
19649 * Makefiles updated to exit if an error occurs in a sub-directory
19650 make (including if user presses ^C) [Paul Sutton]
19652 * Make Montgomery context stuff explicit in RSA data structure.
19656 * Fix build order of pem and err to allow for generated pem.h.
19660 * Fix renumbering bug in X509_NAME_delete_entry().
19664 * Enhanced the err-ins.pl script so it makes the error library number
19665 global and can add a library name. This is needed for external ASN1 and
19666 other error libraries.
19670 * Fixed sk_insert which never worked properly.
19674 * Fix ASN1 macros so they can handle indefinite length constructed
19675 EXPLICIT tags. Some non standard certificates use these: they can now
19680 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19681 into a single doc/ssleay.txt bundle. This way the information is still
19682 preserved but no longer messes up this directory. Now it's new room for
19683 the new set of documentation files.
19685 *Ralf S. Engelschall*
19687 * SETs were incorrectly DER encoded. This was a major pain, because they
19688 shared code with SEQUENCEs, which aren't coded the same. This means that
19689 almost everything to do with SETs or SEQUENCEs has either changed name or
19690 number of arguments.
19692 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19694 * Fix test data to work with the above.
19698 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19699 was already fixed by Eric for 0.9.1 it seems.
19701 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19703 * Autodetect FreeBSD3.
19707 * Fix various bugs in Configure. This affects the following platforms:
19711 unixware-2.0-pentium
19716 * Eliminate generated files from CVS. Reorder tests to regenerate files
19717 before they are needed.
19721 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19725 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19727 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19728 changed SSLeay to OpenSSL in version strings.
19730 *Ralf S. Engelschall*
19732 * Some fixups to the top-level documents.
19736 * Fixed the nasty bug where rsaref.h was not found under compile-time
19737 because the symlink to include/ was missing.
19739 *Ralf S. Engelschall*
19741 * Incorporated the popular no-RSA/DSA-only patches
19742 which allow to compile an RSA-free SSLeay.
19744 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19746 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19747 when "ssleay" is still not found.
19749 *Ralf S. Engelschall*
19751 * Added more platforms to Configure: Cray T3E, HPUX 11,
19753 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19755 * Updated the README file.
19757 *Ralf S. Engelschall*
19759 * Added various .cvsignore files in the CVS repository subdirs
19760 to make a "cvs update" really silent.
19762 *Ralf S. Engelschall*
19764 * Recompiled the error-definition header files and added
19765 missing symbols to the Win32 linker tables.
19767 *Ralf S. Engelschall*
19769 * Cleaned up the top-level documents;
19770 o new files: CHANGES and LICENSE
19771 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19772 o merged COPYRIGHT into LICENSE
19773 o removed obsolete TODO file
19774 o renamed MICROSOFT to INSTALL.W32
19776 *Ralf S. Engelschall*
19778 * Removed dummy files from the 0.9.1b source tree:
19779 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19780 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19781 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19782 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19783 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19785 *Ralf S. Engelschall*
19787 * Added various platform portability fixes.
19791 * The Genesis of the OpenSSL rpject:
19792 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19793 Young and Tim J. Hudson created while they were working for C2Net until
19796 *The OpenSSL Project*
19798 ### Changes between 0.9.0b and 0.9.1b [not released]
19800 * Updated a few CA certificates under certs/
19804 * Changed some BIGNUM api stuff.
19808 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19809 DGUX x86, Linux Alpha, etc.
19813 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19814 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19819 * Add -strparse option to asn1pars program which parses nested
19822 *Dr Stephen Henson <shenson@bigfoot.com>*
19824 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19828 * DSA fix for "ca" program.
19832 * Added "-genkey" option to "dsaparam" program.
19836 * Added RIPE MD160 (rmd160) message digest.
19840 * Added -a (all) option to "ssleay version" command.
19844 * Added PLATFORM define which is the id given to Configure.
19848 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19852 * Extended the ASN.1 parser routines.
19856 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19860 * Added a BN_CTX to the BN library.
19864 * Fixed the weak key values in DES library
19868 * Changed API in EVP library for cipher aliases.
19872 * Added support for RC2/64bit cipher.
19876 * Converted the lhash library to the crypto/mem.c functions.
19880 * Added more recognized ASN.1 object ids.
19884 * Added more RSA padding checks for SSL/TLS.
19888 * Added BIO proxy/filter functionality.
19892 * Added extra_certs to SSL_CTX which can be used
19893 send extra CA certificates to the client in the CA cert chain sending
19894 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19898 * Now Fortezza is denied in the authentication phase because
19899 this is key exchange mechanism is not supported by SSLeay at all.
19903 * Additional PKCS1 checks.
19907 * Support the string "TLSv1" for all TLS v1 ciphers.
19911 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19912 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19916 * Fixed a few memory leaks.
19920 * Fixed various code and comment typos.
19924 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19925 bytes sent in the client random.
19927 *Edward Bishop <ebishop@spyglass.com>*
19931 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
19932 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
19933 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
19934 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
19935 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
19936 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
19937 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
19938 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
19939 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19940 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19941 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19942 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19943 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19944 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19945 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19946 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19947 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19948 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19949 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19950 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19951 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19952 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19953 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19954 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19955 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19956 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19957 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19958 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19959 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19960 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19961 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19962 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19963 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19964 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19965 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19966 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19967 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19968 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19969 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19970 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19971 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19972 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19973 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19974 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19975 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19976 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19977 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19978 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19979 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19980 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19981 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19982 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19983 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19984 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19985 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19986 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19987 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19988 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19989 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19990 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19991 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19992 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19993 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19994 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19995 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19996 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19997 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19998 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19999 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20000 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20001 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20002 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20003 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20004 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20005 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20006 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20007 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20008 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20009 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20010 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20011 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20012 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20013 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20014 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20015 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20016 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20017 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20018 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20019 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20020 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20021 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20022 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20023 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20024 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20025 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20026 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20027 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20028 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20029 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20030 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20031 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20032 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20033 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20034 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20035 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20036 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20037 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20038 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20039 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20040 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20041 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20042 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20043 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20044 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20045 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20046 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20047 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20048 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20049 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20050 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20051 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20052 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20053 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20054 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20055 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20056 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20057 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20058 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20059 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20060 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20061 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20062 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20063 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20064 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20065 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20066 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20067 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20068 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20069 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20070 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20071 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20072 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20073 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20074 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20075 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20076 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20077 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20078 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20079 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20080 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20081 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20082 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20083 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20084 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20085 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20086 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20087 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20088 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20089 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20090 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20091 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20092 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20093 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20094 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20095 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20096 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20097 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20098 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20099 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20100 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20101 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20102 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20103 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20104 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20105 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20106 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20107 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20108 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20109 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20110 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20111 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20112 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20113 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20114 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20115 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20116 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20117 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655