4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.8 and 3.0.9 [xx XXX xxxx]
33 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
34 happens if the buffer size is 4 mod 5. This can trigger a crash of an
35 application using AES-XTS decryption if the memory just after the buffer
36 being decrypted is not mapped.
37 Thanks to Anton Romanov (Amazon) for discovering the issue.
42 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
43 The previous fix for this timing side channel turned out to cause
44 a severe 2-3x performance regression in the typical use case
45 compared to 3.0.7. The new fix uses existing constant time
46 code paths, and restores the previous performance level while
47 fully eliminating all existing timing side channels.
48 The fix was developed by Bernd Edlinger with testing support
53 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
54 that it does not enable policy checking. Thanks to David Benjamin for
55 discovering this issue.
60 * Fixed an issue where invalid certificate policies in leaf certificates are
61 silently ignored by OpenSSL and other certificate policy checks are skipped
62 for that certificate. A malicious CA could use this to deliberately assert
63 invalid certificate policies in order to circumvent policy checking on the
64 certificate altogether.
69 * Limited the number of nodes created in a policy tree to mitigate
70 against CVE-2023-0464. The default limit is set to 1000 nodes, which
71 should be sufficient for most installations. If required, the limit
72 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
73 time define to a desired maximum number of nodes or zero to allow
79 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
81 * Fixed NULL dereference during PKCS7 data verification.
83 A NULL pointer can be dereferenced when signatures are being
84 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
85 algorithm used for the signature is known to the OpenSSL library but
86 the implementation of the hash algorithm is not available the digest
87 initialization will fail. There is a missing check for the return
88 value from the initialization function which later leads to invalid
89 usage of the digest API most likely leading to a crash.
92 PKCS7 data is processed by the SMIME library calls and also by the
93 time stamp (TS) library calls. The TLS implementation in OpenSSL does
94 not call these functions however third party applications would be
95 affected if they call these functions to verify signatures on untrusted
100 * Fixed X.400 address type confusion in X.509 GeneralName.
102 There is a type confusion vulnerability relating to X.400 address processing
103 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
104 but the public structure definition for GENERAL_NAME incorrectly specified
105 the type of the x400Address field as ASN1_TYPE. This field is subsequently
106 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
109 When CRL checking is enabled (i.e. the application sets the
110 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
111 pass arbitrary pointers to a memcmp call, enabling them to read memory
112 contents or enact a denial of service.
117 * Fixed NULL dereference validating DSA public key.
119 An invalid pointer dereference on read can be triggered when an
120 application tries to check a malformed DSA public key by the
121 EVP_PKEY_public_check() function. This will most likely lead
122 to an application crash. This function can be called on public
123 keys supplied from untrusted sources which could allow an attacker
124 to cause a denial of service attack.
126 The TLS implementation in OpenSSL does not call this function
127 but applications might call the function if there are additional
128 security requirements imposed by standards such as FIPS 140-3.
131 *Shane Lontis, Tomáš Mráz*
133 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
135 An invalid pointer dereference on read can be triggered when an
136 application tries to load malformed PKCS7 data with the
137 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
139 The result of the dereference is an application crash which could
140 lead to a denial of service attack. The TLS implementation in OpenSSL
141 does not call this function however third party applications might
142 call these functions on untrusted data.
147 * Fixed Use-after-free following BIO_new_NDEF.
149 The public API function BIO_new_NDEF is a helper function used for
150 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
151 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
152 be called directly by end user applications.
154 The function receives a BIO from the caller, prepends a new BIO_f_asn1
155 filter BIO onto the front of it to form a BIO chain, and then returns
156 the new head of the BIO chain to the caller. Under certain conditions,
157 for example if a CMS recipient public key is invalid, the new filter BIO
158 is freed and the function returns a NULL result indicating a failure.
159 However, in this case, the BIO chain is not properly cleaned up and the
160 BIO passed by the caller still retains internal pointers to the previously
161 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
162 then a use-after-free will occur. This will most likely result in a crash.
165 *Viktor Dukhovni, Matt Caswell*
167 * Fixed Double free after calling PEM_read_bio_ex.
169 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
170 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
171 data. If the function succeeds then the "name_out", "header" and "data"
172 arguments are populated with pointers to buffers containing the relevant
173 decoded data. The caller is responsible for freeing those buffers. It is
174 possible to construct a PEM file that results in 0 bytes of payload data.
175 In this case PEM_read_bio_ex() will return a failure code but will populate
176 the header argument with a pointer to a buffer that has already been freed.
177 If the caller also frees this buffer then a double free will occur. This
178 will most likely lead to a crash.
180 The functions PEM_read_bio() and PEM_read() are simple wrappers around
181 PEM_read_bio_ex() and therefore these functions are also directly affected.
183 These functions are also called indirectly by a number of other OpenSSL
184 functions including PEM_X509_INFO_read_bio_ex() and
185 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
186 internal uses of these functions are not vulnerable because the caller does
187 not free the header argument if PEM_read_bio_ex() returns a failure code.
190 *Kurt Roeckx, Matt Caswell*
192 * Fixed Timing Oracle in RSA Decryption.
194 A timing based side channel exists in the OpenSSL RSA Decryption
195 implementation which could be sufficient to recover a plaintext across
196 a network in a Bleichenbacher style attack. To achieve a successful
197 decryption an attacker would have to be able to send a very large number
198 of trial messages for decryption. The vulnerability affects all RSA padding
199 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
202 *Dmitry Belyavsky, Hubert Kario*
204 * Fixed X.509 Name Constraints Read Buffer Overflow.
206 A read buffer overrun can be triggered in X.509 certificate verification,
207 specifically in name constraint checking. The read buffer overrun might
208 result in a crash which could lead to a denial of service attack.
209 In a TLS client, this can be triggered by connecting to a malicious
210 server. In a TLS server, this can be triggered if the server requests
211 client authentication and a malicious client connects.
216 * Fixed X.509 Policy Constraints Double Locking security issue.
218 If an X.509 certificate contains a malformed policy constraint and
219 policy processing is enabled, then a write lock will be taken twice
220 recursively. On some operating systems (most widely: Windows) this
221 results in a denial of service when the affected process hangs. Policy
222 processing being enabled on a publicly facing server is not considered
223 to be a common setup.
228 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
229 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
230 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
231 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
232 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
233 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
234 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
235 for legacy EC and SM2 keys is also changed similarly to honor the
236 equivalent conversion format flag as specified in the underlying
237 `EC_KEY` object being exported to a provider, when this function is
238 called through `EVP_PKEY_export()`.
242 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
244 * Fixed two buffer overflows in punycode decoding functions.
246 A buffer overrun can be triggered in X.509 certificate verification,
247 specifically in name constraint checking. Note that this occurs after
248 certificate chain signature verification and requires either a CA to
249 have signed the malicious certificate or for the application to continue
250 certificate verification despite failure to construct a path to a trusted
253 In a TLS client, this can be triggered by connecting to a malicious
254 server. In a TLS server, this can be triggered if the server requests
255 client authentication and a malicious client connects.
257 An attacker can craft a malicious email address to overflow
258 an arbitrary number of bytes containing the `.` character (decimal 46)
259 on the stack. This buffer overflow could result in a crash (causing a
263 An attacker can craft a malicious email address to overflow four
264 attacker-controlled bytes on the stack. This buffer overflow could
265 result in a crash (causing a denial of service) or potentially remote code
266 execution depending on stack layout for any given platform/compiler.
271 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
272 parameters in OpenSSL code.
273 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
274 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
275 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
276 Using these invalid names may cause algorithms to use slower methods
277 that ignore the CRT parameters.
281 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
286 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
287 data to be signed before signing the certificate.
291 * Added RIPEMD160 to the default provider.
295 * Ensured that the key share group sent or accepted for the key exchange
296 is allowed for the protocol version.
300 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
302 * OpenSSL supports creating a custom cipher via the legacy
303 EVP_CIPHER_meth_new() function and associated function calls. This function
304 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
305 to use the new provider mechanism in order to implement custom ciphers.
307 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
308 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
309 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
310 and decryption initialisation functions). Instead of using the custom cipher
311 directly it incorrectly tries to fetch an equivalent cipher from the
312 available providers. An equivalent cipher is found based on the NID passed to
313 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
314 given cipher. However it is possible for an application to incorrectly pass
315 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
316 is used in this way the OpenSSL encryption/decryption initialisation function
317 will match the NULL cipher as being equivalent and will fetch this from the
318 available providers. This will succeed if the default provider has been
319 loaded (or if a third party provider has been loaded that offers this
320 cipher). Using the NULL cipher means that the plaintext is emitted as the
323 Applications are only affected by this issue if they call
324 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
325 encryption/decryption initialisation function. Applications that only use
326 SSL/TLS are not impacted by this issue.
331 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
336 * Fixed the linux-mips64 Configure target which was missing the
337 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
342 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
347 * Correctly handle a retransmitted ClientHello in DTLS
351 * Fixed detection of ktls support in cross-compile environment on Linux
355 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
360 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
361 report correct results in some cases
365 * Fix UWP builds by defining VirtualLock
369 * For known safe primes use the minimum key length according to RFC 7919.
370 Longer private key sizes unnecessarily raise the cycles needed to compute the
371 shared secret without any increase of the real security. This fixes a
372 regression from 1.1.1 where these shorter keys were generated for the known
377 * Added the loongarch64 target
381 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
382 only passed to the FIPS provider and not to the default or legacy provider.
386 * Fixed reported performance degradation on aarch64. Restored the
387 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
388 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
389 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
390 The new algorithm is still used for 32 bit targets.
394 * Added a missing header for memcmp that caused compilation failure on some
399 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
401 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
402 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
403 This issue makes the RSA implementation with 2048 bit private keys
404 incorrect on such machines and memory corruption will happen during
405 the computation. As a consequence of the memory corruption an attacker
406 may be able to trigger a remote code execution on the machine performing
409 SSL/TLS servers or other servers using 2048 bit RSA private keys running
410 on machines supporting AVX512IFMA instructions of the X86_64 architecture
411 are affected by this issue.
416 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
417 implementation would not encrypt the entirety of the data under some
418 circumstances. This could reveal sixteen bytes of data that was
419 preexisting in the memory that wasn't written. In the special case of
420 "in place" encryption, sixteen bytes of the plaintext would be revealed.
422 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
423 they are both unaffected.
426 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
428 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
430 * In addition to the c_rehash shell command injection identified in
431 CVE-2022-1292, further bugs where the c_rehash script does not
432 properly sanitise shell metacharacters to prevent command injection have been
435 When the CVE-2022-1292 was fixed it was not discovered that there
436 are other places in the script where the file names of certificates
437 being hashed were possibly passed to a command executed through the shell.
439 This script is distributed by some operating systems in a manner where
440 it is automatically executed. On such operating systems, an attacker
441 could execute arbitrary commands with the privileges of the script.
443 Use of the c_rehash script is considered obsolete and should be replaced
444 by the OpenSSL rehash command line tool.
447 *Daniel Fiala, Tomáš Mráz*
449 * Case insensitive string comparison no longer uses locales. It has instead
450 been directly implemented.
454 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
456 * Case insensitive string comparison is reimplemented via new locale-agnostic
457 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
458 comparison. The previous implementation had problems when the Turkish locale
463 * Fixed a bug in the c_rehash script which was not properly sanitising shell
464 metacharacters to prevent command injection. This script is distributed by
465 some operating systems in a manner where it is automatically executed. On
466 such operating systems, an attacker could execute arbitrary commands with the
467 privileges of the script.
469 Use of the c_rehash script is considered obsolete and should be replaced
470 by the OpenSSL rehash command line tool.
475 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
476 certificate on an OCSP response. The bug caused the function in the case
477 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
478 response (meaning a successful verification) even in the case where the
479 response signing certificate fails to verify.
481 It is anticipated that most users of `OCSP_basic_verify` will not use the
482 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
483 a negative value (indicating a fatal error) in the case of a certificate
484 verification failure. The normal expected return value in this case would be
487 This issue also impacts the command line OpenSSL "ocsp" application. When
488 verifying an ocsp response with the "-no_cert_checks" option the command line
489 application will report that the verification is successful even though it
490 has in fact failed. In this case the incorrect successful response will also
491 be accompanied by error messages showing the failure and contradicting the
492 apparently successful result.
497 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
498 AAD data as the MAC key. This made the MAC key trivially predictable.
500 An attacker could exploit this issue by performing a man-in-the-middle attack
501 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
502 that the modified data would still pass the MAC integrity check.
504 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
505 endpoint will always be rejected by the recipient and the connection will
506 fail at that point. Many application protocols require data to be sent from
507 the client to the server first. Therefore, in such a case, only an OpenSSL
508 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
510 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
511 sent in both directions. In this case both clients and servers could be
512 affected, regardless of the application protocol.
514 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
515 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
516 the handshake when using this ciphersuite.
518 The confidentiality of data is not impacted by this issue, i.e. an attacker
519 cannot decrypt data that has been encrypted using this ciphersuite - they can
522 In order for this attack to work both endpoints must legitimately negotiate
523 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
524 OpenSSL 3.0, and is not available within the default provider or the default
525 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
526 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
527 following must have occurred:
529 1) OpenSSL must have been compiled with the (non-default) compile time option
530 enable-weak-ssl-ciphers
532 2) OpenSSL must have had the legacy provider explicitly loaded (either
533 through application code or via configuration)
535 3) The ciphersuite must have been explicitly added to the ciphersuite list
537 4) The libssl security level must have been set to 0 (default is 1)
539 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
541 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
542 others that both endpoints have in common
547 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
548 occuppied by the removed hash table entries.
550 This function is used when decoding certificates or keys. If a long lived
551 process periodically decodes certificates or keys its memory usage will
552 expand without bounds and the process might be terminated by the operating
553 system causing a denial of service. Also traversing the empty hash table
554 entries will take increasingly more time.
556 Typically such long lived processes might be TLS clients or TLS servers
557 configured to accept client certificate authentication.
560 *Hugo Landau, Aliaksei Levin*
562 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
563 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
564 statistics are no longer supported. For compatibility, these statistics are
565 still listed in the output but are now always reported as zero.
569 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
571 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
572 for non-prime moduli.
574 Internally this function is used when parsing certificates that contain
575 elliptic curve public keys in compressed form or explicit elliptic curve
576 parameters with a base point encoded in compressed form.
578 It is possible to trigger the infinite loop by crafting a certificate that
579 has invalid explicit curve parameters.
581 Since certificate parsing happens prior to verification of the certificate
582 signature, any process that parses an externally supplied certificate may thus
583 be subject to a denial of service attack. The infinite loop can also be
584 reached when parsing crafted private keys as they can contain explicit
585 elliptic curve parameters.
587 Thus vulnerable situations include:
589 - TLS clients consuming server certificates
590 - TLS servers consuming client certificates
591 - Hosting providers taking certificates or private keys from customers
592 - Certificate authorities parsing certification requests from subscribers
593 - Anything else which parses ASN.1 elliptic curve parameters
595 Also any other applications that use the BN_mod_sqrt() where the attacker
596 can control the parameter values are vulnerable to this DoS issue.
601 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
602 to the list of ciphersuites providing Perfect Forward Secrecy as
603 required by SECLEVEL >= 3.
605 *Dmitry Belyavskiy, Nicola Tuveri*
607 * Made the AES constant time code for no-asm configurations
608 optional due to the resulting 95% performance degradation.
609 The AES constant time code can be enabled, for no assembly
610 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
614 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
619 * The negative return value handling of the certificate verification callback
620 was reverted. The replacement is to set the verification retry state with
621 the SSL_set_retry_verify() function.
625 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
627 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
628 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
629 verify a certificate supplied by a server. That function may return a
630 negative return value to indicate an internal error (for example out of
631 memory). Such a negative return value is mishandled by OpenSSL and will cause
632 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
633 success and a subsequent call to SSL_get_error() to return the value
634 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
635 returned by OpenSSL if the application has previously called
636 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
637 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
638 totally unexpected and applications may not behave correctly as a result. The
639 exact behaviour will depend on the application but it could result in
640 crashes, infinite loops or other similar incorrect responses.
642 This issue is made more serious in combination with a separate bug in OpenSSL
643 3.0 that will cause X509_verify_cert() to indicate an internal error when
644 processing a certificate chain. This will occur where a certificate does not
645 include the Subject Alternative Name extension but where a Certificate
646 Authority has enforced name constraints. This issue can occur even with valid
652 * Corrected a few file name and file reference bugs in the build,
653 installation and setup scripts, which lead to installation verification
654 failures. Slightly enhanced the installation verification script.
658 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
663 * Fixed PVK encoder to properly query for the passphrase.
667 * Multiple fixes in the OSSL_HTTP API functions.
671 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
672 OSSL_PARAM_INTEGER data type and return error on negative numbers
673 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
674 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
678 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
682 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
686 * Multiple threading fixes.
690 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
694 * Allow fetching an operation from the provider that owns an unexportable key
695 as a fallback if that is still allowed by the property query.
699 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
701 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
706 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
707 S390X capability vector to zero. This simplifies testing of different code
708 paths on S390X architecture.
712 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
713 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
714 SP 800-38D". The communication will fail at this point.
718 * The EC_GROUP_clear_free() function is deprecated as there is nothing
719 confidential in EC_GROUP data.
723 * The byte order mark (BOM) character is ignored if encountered at the
724 beginning of a PEM-formatted file.
728 * Added CMS support for the Russian GOST algorithms.
732 * Due to move of the implementation of cryptographic operations
733 to the providers, validation of various operation parameters can
734 be postponed until the actual operation is executed where previously
735 it happened immediately when an operation parameter was set.
737 For example when setting an unsupported curve with
738 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
739 fail but later keygen operations with the EVP_PKEY_CTX will fail.
741 *OpenSSL team members and many third party contributors*
743 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
744 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
745 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
746 instead to retrieve these algorithms from a provider.
750 * On build targets where the multilib postfix is set in the build
751 configuration the libdir directory was changing based on whether
752 the lib directory with the multilib postfix exists on the system
753 or not. This unpredictable behavior was removed and eventual
754 multilib postfix is now always added to the default libdir. Use
755 `--libdir=lib` to override the libdir if adding the postfix is
760 * The triple DES key wrap functionality now conforms to RFC 3217 but is
761 no longer interoperable with OpenSSL 1.1.1.
765 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
766 function codes, this function can only cause problems for calling
771 * Add a configurable flag to output date formats as ISO 8601. Does not
772 change the default date format.
776 * Version of MSVC earlier than 1300 could get link warnings, which could
777 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
778 Support for this flag has been removed.
782 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
783 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
784 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
785 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
786 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
790 * The signatures of the functions to get and set options on SSL and
791 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
792 Some source code changes may be required.
796 * The public definitions of conf_method_st and conf_st have been
797 deprecated. They will be made opaque in a future release.
799 *Rich Salz and Tomáš Mráz*
801 * Client-initiated renegotiation is disabled by default. To allow it, use
802 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
803 flag, or the "ClientRenegotiation" config parameter as appropriate.
807 * Add "abspath" and "includedir" pragma's to config files, to prevent,
808 or modify relative pathname inclusion.
812 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
813 validated. Please consult the README-FIPS and
814 README-PROVIDERS files, as well as the migration guide.
816 *OpenSSL team members and many third party contributors*
818 * For the key types DH and DHX the allowed settable parameters are now different.
822 * The openssl commands that read keys, certificates, and CRLs now
823 automatically detect the PEM or DER format of the input files.
825 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
827 * Added enhanced PKCS#12 APIs which accept a library context.
831 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
835 * Added support for Kernel TLS (KTLS).
837 *Boris Pismenny, John Baldwin and Andrew Gallatin*
839 * Support for RFC 5746 secure renegotiation is now required by default for
840 SSL or TLS connections to succeed.
844 * The signature of the `copy` functional parameter of the
845 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
846 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
847 the signature of the `pub_decode` functional parameter of the
848 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
849 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
853 * The error return values from some control calls (ctrl) have changed.
857 * A public key check is now performed during EVP_PKEY_derive_set_peer().
861 * Many functions in the EVP_ namespace that are getters of values from
862 implementations or contexts were renamed to include get or get0 in their
863 names. Old names are provided as macro aliases for compatibility and
868 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
869 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
870 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
875 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
878 * The output from the command line applications may have minor
883 * The output from numerous "printing" may have minor changes.
887 * Windows thread synchronization uses read/write primitives (SRWLock) when
888 supported by the OS, otherwise CriticalSection continues to be used.
892 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
893 work on read only BIO source/sinks that do not support these functions.
894 This allows piping or redirection of a file BIO using stdin to be buffered
895 into memory. This is used internally in OSSL_DECODER_from_bio().
899 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
900 this function would return one of the values OSSL_STORE_INFO_NAME,
901 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
902 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
903 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
904 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
905 using this function should be amended to handle the changed return value.
909 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
910 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
911 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
912 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
913 contains more than one certificate identifier: This means that all
914 certificates referenced there MUST be part of the validation chain.
918 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
919 RC5, DESX and DES have been moved to the legacy provider.
923 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
924 RIPEMD-160 have been moved to the legacy provider.
928 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
933 * The deprecated functions EVP_PKEY_get0_RSA(),
934 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
935 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
936 well as the similarly named "get1" functions behave differently in
941 * A number of functions handling low-level keys or engines were deprecated
942 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
943 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
944 EVP_PKEY_get0_siphash().
948 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
949 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
950 will need to load the legacy crypto provider. This includes these PBE
951 algorithms which use this KDF:
952 - NID_pbeWithMD2AndDES_CBC
953 - NID_pbeWithMD5AndDES_CBC
954 - NID_pbeWithSHA1AndRC2_CBC
955 - NID_pbeWithMD2AndRC2_CBC
956 - NID_pbeWithMD5AndRC2_CBC
957 - NID_pbeWithSHA1AndDES_CBC
961 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
962 BIO_debug_callback() functions.
966 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
967 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
971 * The RAND_METHOD APIs have been deprecated.
975 * The SRP APIs have been deprecated.
979 * Add a compile time option to prevent the caching of provider fetched
980 algorithms. This is enabled by including the no-cached-fetch option
981 at configuration time.
985 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
986 count of PKCS12_DEFAULT_ITER.
988 *Tomáš Mráz and Sahana Prasad*
990 * The openssl speed command does not use low-level API calls anymore.
994 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
997 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
999 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1003 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1004 providers may supply their own group implementations (using either the "key
1005 exchange" or the "key encapsulation" methods) which will automatically be
1006 detected and used by libssl.
1008 *Matt Caswell, Nicola Tuveri*
1010 * The undocumented function X509_certificate_type() has been deprecated;
1014 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1018 * Removed RSA padding mode for SSLv23 (which was only used for
1019 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1020 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1025 * Deprecated the obsolete X9.31 RSA key generation related functions.
1027 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1028 is not allowed to return a value > 1, this is no more taken as failure.
1030 *Viktor Dukhovni and David von Oheimb*
1032 * Deprecated the obsolete X9.31 RSA key generation related functions
1033 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1034 BN_X931_generate_prime_ex().
1038 * The default key generation method for the regular 2-prime RSA keys was
1039 changed to the FIPS 186-4 B.3.6 method.
1043 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1047 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1051 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1052 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1054 *Rich Salz, Richard Levitte, and David von Oheimb*
1056 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1060 * Deprecated `OCSP_parse_url()`.
1064 * Validation of SM2 keys has been separated from the validation of regular EC
1069 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1070 switches: a validation failure triggers an early exit, returning a failure
1071 exit status to the parent process.
1075 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1076 to ignore unknown ciphers.
1080 * The `-cipher-commands` and `-digest-commands` options
1081 of the command line utility `list` have been deprecated.
1082 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1086 * Added convenience functions for generating asymmetric key pairs:
1087 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1088 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1092 * All of the low level EC_KEY functions have been deprecated.
1094 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1096 * Deprecated all the libcrypto and libssl error string loading
1101 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1102 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1107 * The `-crypt` option to the `passwd` command line tool has been removed.
1111 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1116 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1120 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1121 EVP_PKEY_get1_tls_encodedpoint().
1125 * The security callback, which can be customised by application code, supports
1126 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1127 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1131 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1132 interface. Their functionality remains unchanged.
1136 * Added new option for 'openssl list', '-providers', which will display the
1137 list of loaded providers, their names, version and status. It optionally
1138 displays their gettable parameters.
1142 * Removed EVP_PKEY_set_alias_type().
1146 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1147 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1151 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1152 parameters are still checked for type safety at compile time via helper
1157 * Remove the RAND_DRBG API
1159 *Paul Dale and Matthias St. Pierre*
1161 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1162 as well as actual hostnames.
1166 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1167 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1168 conversely, silently ignore DTLS protocol version bounds when configuring
1169 TLS-based contexts. The commands can be repeated to set bounds of both
1170 types. The same applies with the corresponding "min_protocol" and
1171 "max_protocol" command-line switches, in case some application uses both TLS
1174 SSL_CTX instances that are created for a fixed protocol version (e.g.
1175 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1176 attempts to apply bounds to these protocol versions would result in an
1177 error. Now only the "version-flexible" SSL_CTX instances are subject to
1178 limits in configuration files in command-line options.
1182 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1187 * Reworked the recorded ERR codes to make better space for system errors.
1188 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1189 given code is a system error (true) or an OpenSSL error (false).
1193 * Reworked the test perl framework to better allow parallel testing.
1195 *Nicola Tuveri and David von Oheimb*
1197 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1198 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1202 * 'Configure' has been changed to figure out the configuration target if
1203 none is given on the command line. Consequently, the 'config' script is
1204 now only a mere wrapper. All documentation is changed to only mention
1207 *Rich Salz and Richard Levitte*
1209 * Added a library context `OSSL_LIB_CTX` that applications as well as
1210 other libraries can use to form a separate context within which
1211 libcrypto operations are performed.
1215 * Added various `_ex` functions to the OpenSSL API that support using
1216 a non-default `OSSL_LIB_CTX`.
1220 * Handshake now fails if Extended Master Secret extension is dropped
1225 * Dropped interactive mode from the `openssl` program.
1229 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1231 *David von Oheimb and Shane Lontis*
1233 * Deprecated `EC_METHOD_get_field_type()`.
1237 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1238 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1239 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1243 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1247 * Add CAdES-BES signature verification support, mostly derived
1248 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1250 *Filipe Raimundo da Silva*
1252 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1256 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1257 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1261 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1265 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1266 EC_KEY_precompute_mult().
1270 * Deprecated EC_POINTs_mul().
1274 * Removed FIPS_mode() and FIPS_mode_set().
1278 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1282 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1283 EC_POINT_get_Jprojective_coordinates_GFp().
1287 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1288 arrays to be more easily constructed via a series of utility functions.
1289 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1290 the various push functions and finally convert to a passable OSSL_PARAM
1291 array using OSSL_PARAM_BLD_to_param().
1295 * The security strength of SHA1 and MD5 based signatures in TLS has been
1300 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1301 contain a provider side internal key.
1305 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1309 * Project text documents not yet having a proper file name extension
1310 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1311 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1315 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1316 have been converted to Markdown with the goal to produce documents
1317 which not only look pretty when viewed online in the browser, but
1318 remain well readable inside a plain text editor.
1320 To achieve this goal, a 'minimalistic' Markdown style has been applied
1321 which avoids formatting elements that interfere too much with the
1322 reading flow in the text file. For example, it
1324 * avoids [ATX headings][] and uses [setext headings][] instead
1325 (which works for `<h1>` and `<h2>` headings only).
1326 * avoids [inline links][] and uses [reference links][] instead.
1327 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1329 [ATX headings]: https://github.github.com/gfm/#atx-headings
1330 [setext headings]: https://github.github.com/gfm/#setext-headings
1331 [inline links]: https://github.github.com/gfm/#inline-link
1332 [reference links]: https://github.github.com/gfm/#reference-link
1333 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1334 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1336 *Matthias St. Pierre*
1338 * The test suite is changed to preserve results of each test recipe.
1339 A new directory test-runs/ with subdirectories named like the
1340 test recipes are created in the build tree for this purpose.
1344 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1345 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1346 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1348 *David von Oheimb, Martin Peylo*
1350 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1351 It supports arbitrary request and response content types, GET redirection,
1352 TLS, connections via HTTP(S) proxies, connections and exchange via
1353 user-defined BIOs (allowing implicit connections), persistent connections,
1354 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1355 The legacy OCSP-focused (and only partly documented) API
1356 is retained for backward compatibility, while most of it is deprecated.
1360 * Added `util/check-format.pl`, a tool for checking adherence to the
1361 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1362 The checks performed are incomplete and yield some false positives.
1363 Still the tool should be useful for detecting most typical glitches.
1367 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1368 If domain name resolution yields multiple IP addresses all of them are tried
1369 after `connect()` failures.
1373 * All of the low level RSA functions have been deprecated.
1377 * X509 certificates signed using SHA1 are no longer allowed at security
1382 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1383 modified to use PKEY APIs. These commands are now in maintenance mode
1384 and no new features will be added to them.
1388 * The command line utility rsautl has been deprecated.
1392 * The command line utilities genrsa and rsa have been modified to use PKEY
1393 APIs. They now write PKCS#8 keys by default. These commands are now in
1394 maintenance mode and no new features will be added to them.
1398 * All of the low level DH functions have been deprecated.
1400 *Paul Dale and Matt Caswell*
1402 * All of the low level DSA functions have been deprecated.
1406 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1407 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1411 * Deprecated low level ECDH and ECDSA functions.
1415 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1419 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1420 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1421 a new formulation to include all the things it can be used for,
1422 as well as words of caution.
1426 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1430 * All of the low level HMAC functions have been deprecated.
1432 *Paul Dale and David von Oheimb*
1434 * Over two thousand fixes were made to the documentation, including:
1435 - Common options (such as -rand/-writerand, TLS version control, etc)
1436 were refactored and point to newly-enhanced descriptions in openssl.pod.
1437 - Added style conformance for all options (with help from Richard Levitte),
1438 documented all reported missing options, added a CI build to check
1439 that all options are documented and that no unimplemented options
1441 - Documented some internals, such as all use of environment variables.
1442 - Addressed all internal broken L<> references.
1446 * All of the low level CMAC functions have been deprecated.
1450 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1451 functions have been deprecated.
1453 *Paul Dale and David von Oheimb*
1455 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1456 set of functions. The documentation mentioned negative values for some
1457 errors, but this was never the case, so the mention of negative values
1460 Code that followed the documentation and thereby check with something
1461 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1465 * All of the low level cipher functions have been deprecated.
1467 *Matt Caswell and Paul Dale*
1469 * Removed include/openssl/opensslconf.h.in and replaced it with
1470 include/openssl/configuration.h.in, which differs in not including
1471 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1472 was added to include both.
1474 This allows internal hacks where one might need to modify the set
1475 of configured macros, for example this if deprecated symbols are
1476 still supposed to be available internally:
1478 #include <openssl/configuration.h>
1480 #undef OPENSSL_NO_DEPRECATED
1481 #define OPENSSL_SUPPRESS_DEPRECATED
1483 #include <openssl/macros.h>
1485 This should not be used by applications that use the exported
1486 symbols, as that will lead to linking errors.
1490 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1491 used in exponentiation with 512-bit moduli. No EC algorithms are
1492 affected. Analysis suggests that attacks against 2-prime RSA1024,
1493 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1494 difficult to perform and are not believed likely. Attacks against DH512
1495 are considered just feasible. However, for an attack the target would
1496 have to re-use the DH512 private key, which is not recommended anyway.
1497 Also applications directly using the low-level API BN_mod_exp may be
1498 affected if they use BN_FLG_CONSTTIME.
1503 * Most memory-debug features have been deprecated, and the functionality
1504 replaced with no-ops.
1508 * Added documentation for the STACK API.
1512 * Introduced a new method type and API, OSSL_ENCODER, to represent
1513 generic encoders. These do the same sort of job that PEM writers
1514 and d2i functions do, but with support for methods supplied by
1515 providers, and the possibility for providers to support other
1520 * Introduced a new method type and API, OSSL_DECODER, to represent
1521 generic decoders. These do the same sort of job that PEM readers
1522 and i2d functions do, but with support for methods supplied by
1523 providers, and the possibility for providers to support other
1528 * Added a .pragma directive to the syntax of configuration files, to
1529 allow varying behavior in a supported and predictable manner.
1530 Currently added pragma:
1534 This allows dollar signs to be a keyword character unless it's
1535 followed by a opening brace or parenthesis. This is useful for
1536 platforms where dollar signs are commonly used in names, such as
1537 volume names and system directory names on VMS.
1541 * Added functionality to create an EVP_PKEY from user data.
1545 * Change the interpretation of the '--api' configuration option to
1546 mean that this is a desired API compatibility level with no
1547 further meaning. The previous interpretation, that this would
1548 also mean to remove all deprecated symbols up to and including
1549 the given version, no requires that 'no-deprecated' is also used
1550 in the configuration.
1552 When building applications, the desired API compatibility level
1553 can be set with the OPENSSL_API_COMPAT macro like before. For
1554 API compatibility version below 3.0, the old style numerical
1555 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1556 For version 3.0 and on, the value is expected to be the decimal
1557 value calculated from the major and minor version like this:
1559 MAJOR * 10000 + MINOR * 100
1563 -DOPENSSL_API_COMPAT=30000 For 3.0
1564 -DOPENSSL_API_COMPAT=30200 For 3.2
1566 To hide declarations that are deprecated up to and including the
1567 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1568 given when building the application as well.
1572 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1573 access to certificate and CRL stores via URIs and OSSL_STORE
1576 This adds the following functions:
1578 - X509_LOOKUP_store()
1579 - X509_STORE_load_file()
1580 - X509_STORE_load_path()
1581 - X509_STORE_load_store()
1582 - SSL_add_store_cert_subjects_to_stack()
1583 - SSL_CTX_set_default_verify_store()
1584 - SSL_CTX_load_verify_file()
1585 - SSL_CTX_load_verify_dir()
1586 - SSL_CTX_load_verify_store()
1590 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1591 The presence of this system service is determined at run-time.
1595 * Added functionality to create an EVP_PKEY context based on data
1596 for methods from providers. This takes an algorithm name and a
1597 property query string and simply stores them, with the intent
1598 that any operation that uses this context will use those strings
1599 to fetch the needed methods implicitly, thereby making the port
1600 of application written for pre-3.0 OpenSSL easier.
1604 * The undocumented function NCONF_WIN32() has been deprecated; for
1605 conversion details see the HISTORY section of doc/man5/config.pod
1609 * Introduced the new functions EVP_DigestSignInit_ex() and
1610 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1611 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1612 pages for further details.
1616 * Over two thousand fixes were made to the documentation, including:
1617 adding missing command flags, better style conformance, documentation
1620 *Rich Salz, Richard Levitte*
1622 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1623 X25519, X448, Ed25519 and Ed448.
1627 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1632 * Deprecated the public definition of `ERR_STATE` as well as the function
1633 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1638 * Added ERR functionality to give callers access to the stored function
1639 names that have replaced the older function code based functions.
1641 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1642 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1643 ERR_peek_error_all() and ERR_peek_last_error_all().
1645 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1646 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1647 ERR_func_error_string().
1651 * Extended testing to be verbose for failing tests only. The make variables
1652 VERBOSE_FAILURE or VF can be used to enable this:
1654 $ make VF=1 test # Unix
1655 $ mms /macro=(VF=1) test ! OpenVMS
1656 $ nmake VF=1 test # Windows
1660 * Added the `-copy_extensions` option to the `x509` command for use with
1661 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1662 all extensions in the request are copied to the certificate or vice versa.
1664 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1666 * Added the `-copy_extensions` option to the `req` command for use with
1667 `-x509`. When given with the `copy` or `copyall` argument,
1668 all extensions in the certification request are copied to the certificate.
1672 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1673 they generate are by default RFC 5280 compliant in the following sense:
1674 There is a subjectKeyIdentifier extension with a hash value of the public key
1675 and for not self-signed certs there is an authorityKeyIdentifier extension
1676 with a keyIdentifier field or issuer information identifying the signing key.
1677 This is done unless some configuration overrides the new default behavior,
1678 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1682 * Added several checks to `X509_verify_cert()` according to requirements in
1683 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1684 (which may be done by using the CLI option `-x509_strict`):
1685 * The basicConstraints of CA certificates must be marked critical.
1686 * CA certificates must explicitly include the keyUsage extension.
1687 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1688 * The issuer name of any certificate must not be empty.
1689 * The subject name of CA certs, certs with keyUsage crlSign,
1690 and certs without subjectAlternativeName must not be empty.
1691 * If a subjectAlternativeName extension is given it must not be empty.
1692 * The signatureAlgorithm field and the cert signature must be consistent.
1693 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1694 must not be marked critical.
1695 * The authorityKeyIdentifier must be given for X.509v3 certs
1696 unless they are self-signed.
1697 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1701 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1702 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1706 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1707 used even when parsing explicit parameters, when loading a encoded key
1708 or calling `EC_GROUP_new_from_ecpkparameters()`/
1709 `EC_GROUP_new_from_ecparameters()`.
1710 This prevents bypass of security hardening and performance gains,
1711 especially for curves with specialized EC_METHODs.
1712 By default, if a key encoded with explicit parameters is loaded and later
1713 encoded, the output is still encoded with explicit parameters, even if
1714 internally a "named" EC_GROUP is used for computation.
1718 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1719 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1720 NULL. After this change, only the cofactor parameter can be NULL. It also
1721 does some minimal sanity checks on the passed order.
1726 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1727 An attack is simple, if the first CMS_recipientInfo is valid but the
1728 second CMS_recipientInfo is chosen ciphertext. If the second
1729 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1730 encryption key will be replaced by garbage, and the message cannot be
1731 decoded, but if the RSA decryption fails, the correct encryption key is
1732 used and the recipient will not notice the attack.
1733 As a work around for this potential attack the length of the decrypted
1734 key must be equal to the cipher default key length, in case the
1735 certifiate is not given and all recipientInfo are tried out.
1736 The old behaviour can be re-enabled in the CMS code by setting the
1737 CMS_DEBUG_DECRYPT flag.
1741 * Early start up entropy quality from the DEVRANDOM seed source has been
1742 improved for older Linux systems. The RAND subsystem will wait for
1743 /dev/random to be producing output before seeding from /dev/urandom.
1744 The seeded state is stored for future library initialisations using
1745 a system global shared memory segment. The shared memory identifier
1746 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1747 the desired value. The default identifier is 114.
1751 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1752 when primes for RSA keys are computed.
1753 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1754 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1755 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1756 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1757 This avoids possible fingerprinting of newly generated RSA modules.
1761 * Correct the extended master secret constant on EBCDIC systems. Without this
1762 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1763 negotiate EMS will fail. Unfortunately this also means that TLS connections
1764 between EBCDIC systems with this fix, and EBCDIC systems without this
1765 fix will fail if they negotiate EMS.
1769 * Changed the library initialisation so that the config file is now loaded
1770 by default. This was already the case for libssl. It now occurs for both
1771 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1772 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1776 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1777 where the former acts as a replacement for `ERR_put_error()`, and the
1778 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1779 `ERR_raise_data()` adds more flexibility by taking a format string and
1780 an arbitrary number of arguments following it, to be processed with
1785 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1786 to check if a named provider is loaded and available. When called, it
1787 will also activate all fallback providers if such are still present.
1791 * Enforce a minimum DH modulus size of 512 bits.
1795 * Changed DH parameters to generate the order q subgroup instead of 2q.
1796 Previously generated DH parameters are still accepted by DH_check
1797 but DH_generate_key works around that by clearing bit 0 of the
1798 private key for those. This avoids leaking bit 0 of the private key.
1802 * Significantly reduce secure memory usage by the randomness pools.
1806 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1811 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1812 algorithms. An implementation of a key exchange algorithm can be obtained
1813 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1814 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1815 the older EVP_PKEY_derive_init() function. See the man pages for the new
1816 functions for further details.
1820 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1824 * Removed the function names from error messages and deprecated the
1829 * Removed NextStep support and the macro OPENSSL_UNISTD
1833 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1834 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1835 Also removed "export var as function" capability; we do not export
1836 variables, only functions.
1840 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1841 an error and 1 indicating success. In previous versions of OpenSSL this
1842 was a void type. If a key was set longer than the maximum possible this
1847 * Support SM2 signing and verification schemes with X509 certificate.
1851 * Use SHA256 as the default digest for TS query in the `ts` app.
1855 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1859 * Default cipher lists/suites are now available via a function, the
1860 #defines are deprecated.
1864 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1865 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1866 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1870 * Join the directories crypto/x509 and crypto/x509v3
1874 * Added command 'openssl kdf' that uses the EVP_KDF API.
1878 * Added command 'openssl mac' that uses the EVP_MAC API.
1882 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1883 as default directories. Also added the command 'openssl info'
1884 for scripting purposes.
1888 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1893 * Add prediction resistance to the DRBG reseeding process.
1897 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1898 mandated by IEEE Std 1619-2018.
1902 * Added newline escaping functionality to a filename when using openssl dgst.
1903 This output format is to replicate the output format found in the `*sum`
1904 checksum programs. This aims to preserve backward compatibility.
1906 *Matt Eaton, Richard Levitte, and Paul Dale*
1908 * Removed the heartbeat message in DTLS feature, as it has very
1909 little usage and doesn't seem to fulfill a valuable purpose.
1910 The configuration option is now deprecated.
1914 * Changed the output of 'openssl {digestname} < file' to display the
1915 digest name in its output.
1919 * Added a new generic trace API which provides support for enabling
1920 instrumentation through trace output.
1922 *Richard Levitte & Matthias St. Pierre*
1924 * Added build tests for C++. These are generated files that only do one
1925 thing, to include one public OpenSSL head file each. This tests that
1926 the public header files can be usefully included in a C++ application.
1928 This test isn't enabled by default. It can be enabled with the option
1929 'enable-buildtest-c++'.
1933 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1937 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1941 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1945 * Added KMAC to EVP_MAC.
1949 * Added property based algorithm implementation selection framework to
1954 * Added SCA hardening for modular field inversion in EC_GROUP through
1955 a new dedicated field_inv() pointer in EC_METHOD.
1956 This also addresses a leakage affecting conversions from projective
1957 to affine coordinates.
1959 *Billy Bob Brumley, Nicola Tuveri*
1961 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1962 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1963 those algorithms that were already supported through the EVP_PKEY API
1964 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1965 and scrypt are now wrappers that call EVP_KDF.
1969 * Build devcrypto engine as a dynamic engine.
1971 *Eneas U de Queiroz*
1973 * Add keyed BLAKE2 to EVP_MAC.
1977 * Fix a bug in the computation of the endpoint-pair shared secret used
1978 by DTLS over SCTP. This breaks interoperability with older versions
1979 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1980 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1981 interoperability with such broken implementations. However, enabling
1982 this switch breaks interoperability with correct implementations.
1984 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1985 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1989 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1993 * Changed the license to the Apache License v2.0.
1997 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1999 - Major releases (indicated by incrementing the MAJOR release number)
2000 may introduce incompatible API/ABI changes.
2001 - Minor releases (indicated by incrementing the MINOR release number)
2002 may introduce new features but retain API/ABI compatibility.
2003 - Patch releases (indicated by incrementing the PATCH number)
2004 are intended for bug fixes and other improvements of existing
2005 features only (like improving performance or adding documentation)
2006 and retain API/ABI compatibility.
2010 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2014 * Remove the 'dist' target and add a tarball building script. The
2015 'dist' target has fallen out of use, and it shouldn't be
2016 necessary to configure just to create a source distribution.
2020 * Recreate the OS390-Unix config target. It no longer relies on a
2021 special script like it did for OpenSSL pre-1.1.0.
2025 * Instead of having the source directories listed in Configure, add
2026 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2031 * Add GMAC to EVP_MAC.
2035 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2039 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2040 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2041 to facilitate the continued use of MACs through raw private keys in
2042 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2046 * Deprecate ECDH_KDF_X9_62().
2050 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2051 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2052 are retained for backwards compatibility.
2056 * AES-XTS mode now enforces that its two keys are different to mitigate
2057 the attacked described in "Efficient Instantiations of Tweakable
2058 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2059 Details of this attack can be obtained from:
2060 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2064 * Rename the object files, i.e. give them other names than in previous
2065 versions. Their names now include the name of the final product, as
2066 well as its type mnemonic (bin, lib, shlib).
2070 * Added new option for 'openssl list', '-objects', which will display the
2071 list of built in objects, i.e. OIDs with names.
2075 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2076 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2081 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2082 improves application performance by removing data copies and providing
2083 applications with zero-copy system calls such as sendfile and splice.
2087 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2091 * `PKCS12_parse` now maintains the order of the parsed certificates
2092 when outputting them via `*ca` (rather than reversing it).
2096 * Deprecated pthread fork support methods.
2100 * Added support for FFDHE key exchange in TLS 1.3.
2104 * Added a new concept for OpenSSL plugability: providers. This
2105 functionality is designed to replace the ENGINE API and ENGINE
2106 implementations, and to be much more dynamic, allowing provider
2107 authors to introduce new algorithms among other things, as long as
2108 there's an API that supports the algorithm type.
2110 With this concept comes a new core API for interaction between
2111 libcrypto and provider implementations. Public libcrypto functions
2112 that want to use providers do so through this core API.
2114 The main documentation for this core API is found in
2115 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2116 refer to other manuals describing the API specific for supported
2117 algorithm types (also called operations).
2124 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
2126 * Avoid loading of a dynamic engine twice.
2130 * Prioritise DANE TLSA issuer certs over peer certs
2134 * Fixed random API for MacOS prior to 10.12
2136 These MacOS versions don't support the CommonCrypto APIs
2140 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2142 * Fixed an SM2 Decryption Buffer Overflow.
2144 In order to decrypt SM2 encrypted data an application is expected to
2145 call the API function EVP_PKEY_decrypt(). Typically an application will
2146 call this function twice. The first time, on entry, the "out" parameter
2147 can be NULL and, on exit, the "outlen" parameter is populated with the
2148 buffer size required to hold the decrypted plaintext. The application
2149 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2150 again, but this time passing a non-NULL value for the "out" parameter.
2152 A bug in the implementation of the SM2 decryption code means that the
2153 calculation of the buffer size required to hold the plaintext returned
2154 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2155 size required by the second call. This can lead to a buffer overflow
2156 when EVP_PKEY_decrypt() is called by the application a second time with
2157 a buffer that is too small.
2159 A malicious attacker who is able present SM2 content for decryption to
2160 an application could cause attacker chosen data to overflow the buffer
2161 by up to a maximum of 62 bytes altering the contents of other data held
2162 after the buffer, possibly changing application behaviour or causing
2163 the application to crash. The location of the buffer is application
2164 dependent but is typically heap allocated.
2169 * Fixed various read buffer overruns processing ASN.1 strings
2171 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2172 structure which contains a buffer holding the string data and a field
2173 holding the buffer length. This contrasts with normal C strings which
2174 are repesented as a buffer for the string data which is terminated
2175 with a NUL (0) byte.
2177 Although not a strict requirement, ASN.1 strings that are parsed using
2178 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2179 well as any string whose value has been set with the ASN1_STRING_set()
2180 function will additionally NUL terminate the byte array in the
2181 ASN1_STRING structure.
2183 However, it is possible for applications to directly construct valid
2184 ASN1_STRING structures which do not NUL terminate the byte array by
2185 directly setting the "data" and "length" fields in the ASN1_STRING
2186 array. This can also happen by using the ASN1_STRING_set0() function.
2188 Numerous OpenSSL functions that print ASN.1 data have been found to
2189 assume that the ASN1_STRING byte array will be NUL terminated, even
2190 though this is not guaranteed for strings that have been directly
2191 constructed. Where an application requests an ASN.1 structure to be
2192 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2193 been directly constructed by the application without NUL terminating
2194 the "data" field, then a read buffer overrun can occur.
2196 The same thing can also occur during name constraints processing
2197 of certificates (for example if a certificate has been directly
2198 constructed by the application instead of loading it via the OpenSSL
2199 parsing functions, and the certificate contains non NUL terminated
2200 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2201 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2203 If a malicious actor can cause an application to directly construct an
2204 ASN1_STRING and then process it through one of the affected OpenSSL
2205 functions then this issue could be hit. This might result in a crash
2206 (causing a Denial of Service attack). It could also result in the
2207 disclosure of private memory contents (such as private keys, or
2208 sensitive plaintext).
2213 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2215 * Fixed a problem with verifying a certificate chain when using the
2216 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2217 the certificates present in a certificate chain. It is not set by default.
2219 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2220 the chain that have explicitly encoded elliptic curve parameters was added
2221 as an additional strict check.
2223 An error in the implementation of this check meant that the result of a
2224 previous check to confirm that certificates in the chain are valid CA
2225 certificates was overwritten. This effectively bypasses the check
2226 that non-CA certificates must not be able to issue other certificates.
2228 If a "purpose" has been configured then there is a subsequent opportunity
2229 for checks that the certificate is a valid CA. All of the named "purpose"
2230 values implemented in libcrypto perform this check. Therefore, where
2231 a purpose is set the certificate chain will still be rejected even when the
2232 strict flag has been used. A purpose is set by default in libssl client and
2233 server certificate verification routines, but it can be overridden or
2234 removed by an application.
2236 In order to be affected, an application must explicitly set the
2237 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2238 for the certificate verification or, in the case of TLS client or server
2239 applications, override the default purpose.
2244 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2245 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2246 renegotiation ClientHello omits the signature_algorithms extension (where it
2247 was present in the initial ClientHello), but includes a
2248 signature_algorithms_cert extension then a NULL pointer dereference will
2249 result, leading to a crash and a denial of service attack.
2251 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2252 (which is the default configuration). OpenSSL TLS clients are not impacted by
2256 *Peter Kästle and Samuel Sapalski*
2258 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2260 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2261 create a unique hash value based on the issuer and serial number data
2262 contained within an X509 certificate. However it was failing to correctly
2263 handle any errors that may occur while parsing the issuer field (which might
2264 occur if the issuer field is maliciously constructed). This may subsequently
2265 result in a NULL pointer deref and a crash leading to a potential denial of
2271 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2272 padding mode to correctly check for rollback attacks. This is considered a
2273 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2278 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2279 functions. Previously they could overflow the output length argument in some
2280 cases where the input length is close to the maximum permissable length for
2281 an integer on the platform. In such cases the return value from the function
2282 call would be 1 (indicating success), but the output length value would be
2283 negative. This could cause applications to behave incorrectly or crash.
2288 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2289 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2290 could be exploited in a side channel attack to recover the password. Since
2291 the attack is local host only this is outside of the current OpenSSL
2292 threat model and therefore no CVE is assigned.
2294 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2299 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2301 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2302 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2303 If an attacker can control both items being compared then this could lead
2304 to a possible denial of service attack. OpenSSL itself uses the
2305 GENERAL_NAME_cmp function for two purposes:
2306 1) Comparing CRL distribution point names between an available CRL and a
2307 CRL distribution point embedded in an X509 certificate
2308 2) When verifying that a timestamp response token signer matches the
2309 timestamp authority name (exposed via the API functions
2310 TS_RESP_verify_response and TS_RESP_verify_token)
2315 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2317 * Certificates with explicit curve parameters are now disallowed in
2318 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2322 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2323 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2324 conversely, silently ignore DTLS protocol version bounds when configuring
2325 TLS-based contexts. The commands can be repeated to set bounds of both
2326 types. The same applies with the corresponding "min_protocol" and
2327 "max_protocol" command-line switches, in case some application uses both TLS
2330 SSL_CTX instances that are created for a fixed protocol version (e.g.
2331 TLSv1_server_method()) also silently ignore version bounds. Previously
2332 attempts to apply bounds to these protocol versions would result in an
2333 error. Now only the "version-flexible" SSL_CTX instances are subject to
2334 limits in configuration files in command-line options.
2338 * Handshake now fails if Extended Master Secret extension is dropped
2343 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2345 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2347 * Fixed segmentation fault in SSL_check_chain()
2348 Server or client applications that call the SSL_check_chain() function
2349 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2350 dereference as a result of incorrect handling of the
2351 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2352 or unrecognised signature algorithm is received from the peer. This could
2353 be exploited by a malicious peer in a Denial of Service attack.
2358 * Added AES consttime code for no-asm configurations
2359 an optional constant time support for AES was added
2360 when building openssl for no-asm.
2361 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2362 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2363 At this time this feature is by default disabled.
2364 It will be enabled by default in 3.0.
2368 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2370 * Revert the change of EOF detection while reading in libssl to avoid
2371 regressions in applications depending on the current way of reporting
2372 the EOF. As the existing method is not fully accurate the change to
2373 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2374 branch and will be present in the 3.0 release.
2378 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2379 when primes for RSA keys are computed.
2380 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2381 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2382 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2383 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2384 This avoids possible fingerprinting of newly generated RSA modules.
2388 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2390 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2391 while reading in libssl then we would report an error back to the
2392 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2393 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2394 therefore give a hint as to what went wrong.
2398 * Check that ed25519 and ed448 are allowed by the security level. Previously
2399 signature algorithms not using an MD were not being checked that they were
2400 allowed by the security level.
2404 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2405 was not quite right. The behaviour was not consistent between resumption
2406 and normal handshakes, and also not quite consistent with historical
2407 behaviour. The behaviour in various scenarios has been clarified and
2408 it has been updated to make it match historical behaviour as closely as
2413 * *[VMS only]* The header files that the VMS compilers include automatically,
2414 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2415 that the C++ compiler doesn't understand. This is a shortcoming in the
2416 compiler, but can be worked around with `__cplusplus` guards.
2418 C++ applications that use OpenSSL libraries must be compiled using the
2419 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2420 functions. Otherwise, only functions with symbols of less than 31
2421 characters can be used, as the linker will not be able to successfully
2422 resolve symbols with longer names.
2426 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2427 The presence of this system service is determined at run-time.
2431 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2436 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2438 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2439 number generator (RNG). This was intended to include protection in the
2440 event of a fork() system call in order to ensure that the parent and child
2441 processes did not share the same RNG state. However this protection was not
2442 being used in the default case.
2444 A partial mitigation for this issue is that the output from a high
2445 precision timer is mixed into the RNG state so the likelihood of a parent
2446 and child process sharing state is significantly reduced.
2448 If an application already calls OPENSSL_init_crypto() explicitly using
2449 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2452 *Matthias St. Pierre*
2454 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2455 used even when parsing explicit parameters, when loading a encoded key
2456 or calling `EC_GROUP_new_from_ecpkparameters()`/
2457 `EC_GROUP_new_from_ecparameters()`.
2458 This prevents bypass of security hardening and performance gains,
2459 especially for curves with specialized EC_METHODs.
2460 By default, if a key encoded with explicit parameters is loaded and later
2461 encoded, the output is still encoded with explicit parameters, even if
2462 internally a "named" EC_GROUP is used for computation.
2466 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2467 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2468 NULL. After this change, only the cofactor parameter can be NULL. It also
2469 does some minimal sanity checks on the passed order.
2474 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2475 An attack is simple, if the first CMS_recipientInfo is valid but the
2476 second CMS_recipientInfo is chosen ciphertext. If the second
2477 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2478 encryption key will be replaced by garbage, and the message cannot be
2479 decoded, but if the RSA decryption fails, the correct encryption key is
2480 used and the recipient will not notice the attack.
2481 As a work around for this potential attack the length of the decrypted
2482 key must be equal to the cipher default key length, in case the
2483 certifiate is not given and all recipientInfo are tried out.
2484 The old behaviour can be re-enabled in the CMS code by setting the
2485 CMS_DEBUG_DECRYPT flag.
2490 * Early start up entropy quality from the DEVRANDOM seed source has been
2491 improved for older Linux systems. The RAND subsystem will wait for
2492 /dev/random to be producing output before seeding from /dev/urandom.
2493 The seeded state is stored for future library initialisations using
2494 a system global shared memory segment. The shared memory identifier
2495 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2496 the desired value. The default identifier is 114.
2500 * Correct the extended master secret constant on EBCDIC systems. Without this
2501 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2502 negotiate EMS will fail. Unfortunately this also means that TLS connections
2503 between EBCDIC systems with this fix, and EBCDIC systems without this
2504 fix will fail if they negotiate EMS.
2508 * Use Windows installation paths in the mingw builds
2510 Mingw isn't a POSIX environment per se, which means that Windows
2511 paths should be used for installation.
2516 * Changed DH_check to accept parameters with order q and 2q subgroups.
2517 With order 2q subgroups the bit 0 of the private key is not secret
2518 but DH_generate_key works around that by clearing bit 0 of the
2519 private key for those. This avoids leaking bit 0 of the private key.
2523 * Significantly reduce secure memory usage by the randomness pools.
2527 * Revert the DEVRANDOM_WAIT feature for Linux systems
2529 The DEVRANDOM_WAIT feature added a select() call to wait for the
2530 /dev/random device to become readable before reading from the
2531 /dev/urandom device.
2533 It turned out that this change had negative side effects on
2534 performance which were not acceptable. After some discussion it
2535 was decided to revert this feature and leave it up to the OS
2536 resp. the platform maintainer to ensure a proper initialization
2537 during early boot time.
2539 *Matthias St. Pierre*
2541 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2543 * Add build tests for C++. These are generated files that only do one
2544 thing, to include one public OpenSSL head file each. This tests that
2545 the public header files can be usefully included in a C++ application.
2547 This test isn't enabled by default. It can be enabled with the option
2548 'enable-buildtest-c++'.
2552 * Enable SHA3 pre-hashing for ECDSA and DSA.
2556 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2557 This changes the size when using the `genpkey` command when no size is given.
2558 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2559 generation commands to use 2048 bits by default.
2563 * Reorganize the manual pages to consistently have RETURN VALUES,
2564 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2565 util/fix-doc-nits accordingly.
2567 *Paul Yang, Joshua Lock*
2569 * Add the missing accessor EVP_PKEY_get0_engine()
2573 * Have commands like `s_client` and `s_server` output the signature scheme
2574 along with other cipher suite parameters when debugging.
2578 * Make OPENSSL_config() error agnostic again.
2582 * Do the error handling in RSA decryption constant time.
2586 * Prevent over long nonces in ChaCha20-Poly1305.
2588 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2589 for every encryption operation. RFC 7539 specifies that the nonce value
2590 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2591 and front pads the nonce with 0 bytes if it is less than 12
2592 bytes. However it also incorrectly allows a nonce to be set of up to 16
2593 bytes. In this case only the last 12 bytes are significant and any
2594 additional leading bytes are ignored.
2596 It is a requirement of using this cipher that nonce values are
2597 unique. Messages encrypted using a reused nonce value are susceptible to
2598 serious confidentiality and integrity attacks. If an application changes
2599 the default nonce length to be longer than 12 bytes and then makes a
2600 change to the leading bytes of the nonce expecting the new value to be a
2601 new unique nonce then such an application could inadvertently encrypt
2602 messages with a reused nonce.
2604 Additionally the ignored bytes in a long nonce are not covered by the
2605 integrity guarantee of this cipher. Any application that relies on the
2606 integrity of these ignored leading bytes of a long nonce may be further
2607 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2608 is safe because no such use sets such a long nonce value. However user
2609 applications that use this cipher directly and set a non-default nonce
2610 length to be longer than 12 bytes may be vulnerable.
2612 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2618 * Add DEVRANDOM_WAIT feature for Linux systems
2620 On older Linux systems where the getrandom() system call is not available,
2621 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2622 Contrary to getrandom(), the /dev/urandom device will not block during
2623 early boot when the kernel CSPRNG has not been seeded yet.
2625 To mitigate this known weakness, use select() to wait for /dev/random to
2626 become readable before reading from /dev/urandom.
2628 * Ensure that SM2 only uses SM3 as digest algorithm
2632 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2634 * Change the info callback signals for the start and end of a post-handshake
2635 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2636 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2637 confused by this and assume that a TLSv1.2 renegotiation has started. This
2638 can break KeyUpdate handling. Instead we no longer signal the start and end
2639 of a post handshake message exchange (although the messages themselves are
2640 still signalled). This could break some applications that were expecting
2641 the old signals. However without this KeyUpdate is not usable for many
2646 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2648 * Timing vulnerability in DSA signature generation
2650 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2651 timing side channel attack. An attacker could use variations in the signing
2652 algorithm to recover the private key.
2654 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2659 * Timing vulnerability in ECDSA signature generation
2661 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2662 timing side channel attack. An attacker could use variations in the signing
2663 algorithm to recover the private key.
2665 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2670 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2671 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2672 of two gigabytes and the error handling improved.
2674 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2675 categorized as a normal bug, not a security issue, because the DRBG reseeds
2676 automatically and is fully functional even without additional randomness
2677 provided by the application.
2679 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2681 * Add a new ClientHello callback. Provides a callback interface that gives
2682 the application the ability to adjust the nascent SSL object at the
2683 earliest stage of ClientHello processing, immediately after extensions have
2684 been collected but before they have been processed. In particular, this
2685 callback can adjust the supported TLS versions in response to the contents
2690 * Add SM2 base algorithm support.
2694 * s390x assembly pack: add (improved) hardware-support for the following
2695 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2696 aes-cfb/cfb8, aes-ecb.
2700 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2701 parameter is no longer accepted, as it leads to a corrupt table. NULL
2702 pem_str is reserved for alias entries only.
2706 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2707 step for prime curves. The new implementation is based on formulae from
2708 differential addition-and-doubling in homogeneous projective coordinates
2709 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2710 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2711 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2712 to work in projective coordinates.
2714 *Billy Bob Brumley, Nicola Tuveri*
2716 * Change generating and checking of primes so that the error rate of not
2717 being prime depends on the intended use based on the size of the input.
2718 For larger primes this will result in more rounds of Miller-Rabin.
2719 The maximal error rate for primes with more than 1080 bits is lowered
2722 *Kurt Roeckx, Annie Yousar*
2724 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2728 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2729 moving between systems, and to avoid confusion when a Windows build is
2730 done with mingw vs with MSVC. For POSIX installs, there's still a
2731 symlink or copy named 'tsget' to avoid that confusion as well.
2735 * Revert blinding in ECDSA sign and instead make problematic addition
2736 length-invariant. Switch even to fixed-length Montgomery multiplication.
2740 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2741 step for binary curves. The new implementation is based on formulae from
2742 differential addition-and-doubling in mixed Lopez-Dahab projective
2743 coordinates, modified to independently blind the operands.
2745 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2747 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2748 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2749 EC_METHODs to implement their own specialized "ladder step", to take
2750 advantage of more favorable coordinate systems or more efficient
2751 differential addition-and-doubling algorithms.
2753 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2755 * Modified the random device based seed sources to keep the relevant
2756 file descriptors open rather than reopening them on each access.
2757 This allows such sources to operate in a chroot() jail without
2758 the associated device nodes being available. This behaviour can be
2759 controlled using RAND_keep_random_devices_open().
2763 * Numerous side-channel attack mitigations have been applied. This may have
2764 performance impacts for some algorithms for the benefit of improved
2765 security. Specific changes are noted in this change log by their respective
2770 * AIX shared library support overhaul. Switch to AIX "natural" way of
2771 handling shared libraries, which means collecting shared objects of
2772 different versions and bitnesses in one common archive. This allows to
2773 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2774 doesn't affect the way 3rd party applications are linked, only how
2775 multi-version installation is managed.
2779 * Make ec_group_do_inverse_ord() more robust and available to other
2780 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2781 mitigations are applied to the fallback BN_mod_inverse().
2782 When using this function rather than BN_mod_inverse() directly, new
2783 EC cryptosystem implementations are then safer-by-default.
2787 * Add coordinate blinding for EC_POINT and implement projective
2788 coordinate blinding for generic prime curves as a countermeasure to
2789 chosen point SCA attacks.
2791 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2793 * Add blinding to ECDSA and DSA signatures to protect against side channel
2794 attacks discovered by Keegan Ryan (NCC Group).
2798 * Enforce checking in the `pkeyutl` command to ensure that the input
2799 length does not exceed the maximum supported digest length when performing
2800 a sign, verify or verifyrecover operation.
2804 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2805 I/O in combination with something like select() or poll() will hang. This
2806 can be turned off again using SSL_CTX_clear_mode().
2807 Many applications do not properly handle non-application data records, and
2808 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2809 around the problems in those applications, but can also break some.
2810 It's recommended to read the manpages about SSL_read(), SSL_write(),
2811 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2812 SSL_CTX_set_read_ahead() again.
2816 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2817 now allow empty (zero character) pass phrases.
2821 * Apply blinding to binary field modular inversion and remove patent
2822 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2826 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2827 binary and prime elliptic curves.
2831 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2832 constant time fixed point multiplication.
2836 * Revise elliptic curve scalar multiplication with timing attack
2837 defenses: ec_wNAF_mul redirects to a constant time implementation
2838 when computing fixed point and variable point multiplication (which
2839 in OpenSSL are mostly used with secret scalars in keygen, sign,
2840 ECDH derive operations).
2841 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2844 * Updated CONTRIBUTING
2848 * Updated DRBG / RAND to request nonce and additional low entropy
2849 randomness from the system.
2851 *Matthias St. Pierre*
2853 * Updated 'openssl rehash' to use OpenSSL consistent default.
2857 * Moved the load of the ssl_conf module to libcrypto, which helps
2858 loading engines that libssl uses before libssl is initialised.
2862 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2866 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2868 *Ingo Schwarze, Rich Salz*
2870 * Added output of accepting IP address and port for 'openssl s_server'
2874 * Added a new API for TLSv1.3 ciphersuites:
2875 SSL_CTX_set_ciphersuites()
2876 SSL_set_ciphersuites()
2880 * Memory allocation failures consistently add an error to the error
2885 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2886 in libcrypto when run as setuid/setgid.
2890 * Load any config file by default when libssl is used.
2894 * Added new public header file <openssl/rand_drbg.h> and documentation
2895 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2897 *Matthias St. Pierre*
2899 * QNX support removed (cannot find contributors to get their approval
2900 for the license change).
2904 * TLSv1.3 replay protection for early data has been implemented. See the
2905 SSL_read_early_data() man page for further details.
2909 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2910 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2911 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2912 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2913 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2914 configuration has been separated out. See the ciphers man page or the
2915 SSL_CTX_set_ciphersuites() man page for more information.
2919 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2920 in responder mode now supports the new "-multi" option, which
2921 spawns the specified number of child processes to handle OCSP
2922 requests. The "-timeout" option now also limits the OCSP
2923 responder's patience to wait to receive the full client request
2924 on a newly accepted connection. Child processes are respawned
2925 as needed, and the CA index file is automatically reloaded
2926 when changed. This makes it possible to run the "ocsp" responder
2927 as a long-running service, making the OpenSSL CA somewhat more
2928 feature-complete. In this mode, most diagnostic messages logged
2929 after entering the event loop are logged via syslog(3) rather than
2934 * Added support for X448 and Ed448. Heavily based on original work by
2939 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2940 objects loaded. This adds the functions OSSL_STORE_expect() and
2941 OSSL_STORE_find() as well as needed tools to construct searches and
2942 get the search data out of them.
2946 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2947 version of OpenSSL should review their configuration settings to ensure
2948 that they are still appropriate for TLSv1.3. For further information see:
2949 <https://wiki.openssl.org/index.php/TLS1.3>
2953 * Grand redesign of the OpenSSL random generator
2955 The default RAND method now utilizes an AES-CTR DRBG according to
2956 NIST standard SP 800-90Ar1. The new random generator is essentially
2957 a port of the default random generator from the OpenSSL FIPS 2.0
2958 object module. It is a hybrid deterministic random bit generator
2959 using an AES-CTR bit stream and which seeds and reseeds itself
2960 automatically using trusted system entropy sources.
2962 Some of its new features are:
2963 - Support for multiple DRBG instances with seed chaining.
2964 - The default RAND method makes use of a DRBG.
2965 - There is a public and private DRBG instance.
2966 - The DRBG instances are fork-safe.
2967 - Keep all global DRBG instances on the secure heap if it is enabled.
2968 - The public and private DRBG instance are per thread for lock free
2971 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2973 * Changed Configure so it only says what it does and doesn't dump
2974 so much data. Instead, ./configdata.pm should be used as a script
2975 to display all sorts of configuration data.
2979 * Added processing of "make variables" to Configure.
2983 * Added SHA512/224 and SHA512/256 algorithm support.
2987 * The last traces of Netware support, first removed in 1.1.0, have
2992 * Get rid of Makefile.shared, and in the process, make the processing
2993 of certain files (rc.obj, or the .def/.map/.opt files produced from
2994 the ordinal files) more visible and hopefully easier to trace and
2995 debug (or make silent).
2999 * Make it possible to have environment variable assignments as
3000 arguments to config / Configure.
3004 * Add multi-prime RSA (RFC 8017) support.
3008 * Add SM3 implemented according to GB/T 32905-2016
3009 *Jack Lloyd <jack.lloyd@ribose.com>,*
3010 *Ronald Tse <ronald.tse@ribose.com>,*
3011 *Erick Borsboom <erick.borsboom@ribose.com>*
3013 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3014 as documented in RFC6066.
3015 Based on a patch from Tomasz Moń
3017 *Filipe Raimundo da Silva*
3019 * Add SM4 implemented according to GB/T 32907-2016.
3020 *Jack Lloyd <jack.lloyd@ribose.com>,*
3021 *Ronald Tse <ronald.tse@ribose.com>,*
3022 *Erick Borsboom <erick.borsboom@ribose.com>*
3024 * Reimplement -newreq-nodes and ERR_error_string_n; the
3025 original author does not agree with the license change.
3029 * Add ARIA AEAD TLS support.
3033 * Some macro definitions to support VS6 have been removed. Visual
3034 Studio 6 has not worked since 1.1.0
3038 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3039 without clearing the errors.
3043 * Add "atfork" functions. If building on a system that without
3044 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3045 requirements. The RAND facility now uses/requires this.
3053 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3054 not possible to disable entirely. However, it's still possible to
3055 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3058 To disable, configure with 'no-ui-console'. 'no-ui' is still
3059 possible to use as an alias. Check at compile time with the
3060 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3061 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3065 * Add a STORE module, which implements a uniform and URI based reader of
3066 stores that can contain keys, certificates, CRLs and numerous other
3067 objects. The main API is loosely based on a few stdio functions,
3068 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3069 OSSL_STORE_error and OSSL_STORE_close.
3070 The implementation uses backends called "loaders" to implement arbitrary
3071 URI schemes. There is one built in "loader" for the 'file' scheme.
3075 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3076 then adjusted to work on FreeBSD 8.4 as well.
3077 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3078 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3082 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3083 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3084 error code calls like this:
3086 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3088 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3089 that can be encoded in C. For the foreseeable future, this will only
3092 *Richard Levitte and Tim Hudson*
3094 * Removed BSD cryptodev engine.
3098 * Add a build target 'build_all_generated', to build all generated files
3099 and only that. This can be used to prepare everything that requires
3100 things like perl for a system that lacks perl and then move everything
3101 to that system and do the rest of the build there.
3105 * In the UI interface, make it possible to duplicate the user data. This
3106 can be used by engines that need to retain the data for a longer time
3107 than just the call where this user data is passed.
3111 * Ignore the '-named_curve auto' value for compatibility of applications
3114 *Tomáš Mráz <tmraz@fedoraproject.org>*
3116 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3117 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3118 alerts across multiple records (some of which could be empty). In practice
3119 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3120 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3121 support this at all. Supporting it adds significant complexity to the
3122 record layer, and its removal is unlikely to cause interoperability
3127 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3128 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3129 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3134 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3135 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3137 *Richard Levitte, Andy Polyakov*
3139 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3144 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3145 platform rather than 'mingw'.
3149 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3150 success if they are asked to add an object which already exists
3151 in the store. This change cascades to other functions which load
3152 certificates and CRLs.
3156 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3157 facilitate stack unwinding even from assembly subroutines.
3161 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3162 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3166 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3167 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3168 which is the minimum version we support.
3172 * Certificate time validation (X509_cmp_time) enforces stricter
3173 compliance with RFC 5280. Fractional seconds and timezone offsets
3174 are no longer allowed.
3178 * Add support for ARIA
3182 * s_client will now send the Server Name Indication (SNI) extension by
3183 default unless the new "-noservername" option is used. The server name is
3184 based on the host provided to the "-connect" option unless overridden by
3185 using "-servername".
3189 * Add support for SipHash
3193 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3194 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3195 prevent issues where no progress is being made and the peer continually
3196 sends unrecognised record types, using up resources processing them.
3200 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3201 using the algorithm defined in
3202 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3206 * Heartbeat support has been removed; the ABI is changed for now.
3208 *Richard Levitte, Rich Salz*
3210 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3214 * The RSA "null" method, which was partially supported to avoid patent
3215 issues, has been replaced to always returns NULL.
3222 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3224 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3225 used even when parsing explicit parameters, when loading a encoded key
3226 or calling `EC_GROUP_new_from_ecpkparameters()`/
3227 `EC_GROUP_new_from_ecparameters()`.
3228 This prevents bypass of security hardening and performance gains,
3229 especially for curves with specialized EC_METHODs.
3230 By default, if a key encoded with explicit parameters is loaded and later
3231 encoded, the output is still encoded with explicit parameters, even if
3232 internally a "named" EC_GROUP is used for computation.
3236 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3237 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3238 NULL. After this change, only the cofactor parameter can be NULL. It also
3239 does some minimal sanity checks on the passed order.
3244 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3245 An attack is simple, if the first CMS_recipientInfo is valid but the
3246 second CMS_recipientInfo is chosen ciphertext. If the second
3247 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3248 encryption key will be replaced by garbage, and the message cannot be
3249 decoded, but if the RSA decryption fails, the correct encryption key is
3250 used and the recipient will not notice the attack.
3251 As a work around for this potential attack the length of the decrypted
3252 key must be equal to the cipher default key length, in case the
3253 certifiate is not given and all recipientInfo are tried out.
3254 The old behaviour can be re-enabled in the CMS code by setting the
3255 CMS_DEBUG_DECRYPT flag.
3260 * Use Windows installation paths in the mingw builds
3262 Mingw isn't a POSIX environment per se, which means that Windows
3263 paths should be used for installation.
3268 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3270 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3271 This changes the size when using the `genpkey` command when no size is given.
3272 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3273 generation commands to use 2048 bits by default.
3277 * Prevent over long nonces in ChaCha20-Poly1305.
3279 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3280 for every encryption operation. RFC 7539 specifies that the nonce value
3281 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3282 and front pads the nonce with 0 bytes if it is less than 12
3283 bytes. However it also incorrectly allows a nonce to be set of up to 16
3284 bytes. In this case only the last 12 bytes are significant and any
3285 additional leading bytes are ignored.
3287 It is a requirement of using this cipher that nonce values are
3288 unique. Messages encrypted using a reused nonce value are susceptible to
3289 serious confidentiality and integrity attacks. If an application changes
3290 the default nonce length to be longer than 12 bytes and then makes a
3291 change to the leading bytes of the nonce expecting the new value to be a
3292 new unique nonce then such an application could inadvertently encrypt
3293 messages with a reused nonce.
3295 Additionally the ignored bytes in a long nonce are not covered by the
3296 integrity guarantee of this cipher. Any application that relies on the
3297 integrity of these ignored leading bytes of a long nonce may be further
3298 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3299 is safe because no such use sets such a long nonce value. However user
3300 applications that use this cipher directly and set a non-default nonce
3301 length to be longer than 12 bytes may be vulnerable.
3303 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3309 * Added SCA hardening for modular field inversion in EC_GROUP through
3310 a new dedicated field_inv() pointer in EC_METHOD.
3311 This also addresses a leakage affecting conversions from projective
3312 to affine coordinates.
3314 *Billy Bob Brumley, Nicola Tuveri*
3316 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3317 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3321 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3325 * Remove the 'dist' target and add a tarball building script. The
3326 'dist' target has fallen out of use, and it shouldn't be
3327 necessary to configure just to create a source distribution.
3331 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3333 * Timing vulnerability in DSA signature generation
3335 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3336 timing side channel attack. An attacker could use variations in the signing
3337 algorithm to recover the private key.
3339 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3344 * Timing vulnerability in ECDSA signature generation
3346 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3347 timing side channel attack. An attacker could use variations in the signing
3348 algorithm to recover the private key.
3350 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3355 * Add coordinate blinding for EC_POINT and implement projective
3356 coordinate blinding for generic prime curves as a countermeasure to
3357 chosen point SCA attacks.
3359 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3361 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3363 * Client DoS due to large DH parameter
3365 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3366 malicious server can send a very large prime value to the client. This will
3367 cause the client to spend an unreasonably long period of time generating a
3368 key for this prime resulting in a hang until the client has finished. This
3369 could be exploited in a Denial Of Service attack.
3371 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3376 * Cache timing vulnerability in RSA Key Generation
3378 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3379 a cache timing side channel attack. An attacker with sufficient access to
3380 mount cache timing attacks during the RSA key generation process could
3381 recover the private key.
3383 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3384 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3389 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3390 parameter is no longer accepted, as it leads to a corrupt table. NULL
3391 pem_str is reserved for alias entries only.
3395 * Revert blinding in ECDSA sign and instead make problematic addition
3396 length-invariant. Switch even to fixed-length Montgomery multiplication.
3400 * Change generating and checking of primes so that the error rate of not
3401 being prime depends on the intended use based on the size of the input.
3402 For larger primes this will result in more rounds of Miller-Rabin.
3403 The maximal error rate for primes with more than 1080 bits is lowered
3406 *Kurt Roeckx, Annie Yousar*
3408 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3412 * Add blinding to ECDSA and DSA signatures to protect against side channel
3413 attacks discovered by Keegan Ryan (NCC Group).
3417 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3418 now allow empty (zero character) pass phrases.
3422 * Certificate time validation (X509_cmp_time) enforces stricter
3423 compliance with RFC 5280. Fractional seconds and timezone offsets
3424 are no longer allowed.
3428 * Fixed a text canonicalisation bug in CMS
3430 Where a CMS detached signature is used with text content the text goes
3431 through a canonicalisation process first prior to signing or verifying a
3432 signature. This process strips trailing space at the end of lines, converts
3433 line terminators to CRLF and removes additional trailing line terminators
3434 at the end of a file. A bug in the canonicalisation process meant that
3435 some characters, such as form-feed, were incorrectly treated as whitespace
3436 and removed. This is contrary to the specification (RFC5485). This fix
3437 could mean that detached text data signed with an earlier version of
3438 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3439 signed with a fixed OpenSSL may fail to verify with an earlier version of
3440 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3441 and use the "-binary" flag (for the "cms" command line application) or set
3442 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3446 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3448 * Constructed ASN.1 types with a recursive definition could exceed the stack
3450 Constructed ASN.1 types with a recursive definition (such as can be found
3451 in PKCS7) could eventually exceed the stack given malicious input with
3452 excessive recursion. This could result in a Denial Of Service attack. There
3453 are no such structures used within SSL/TLS that come from untrusted sources
3454 so this is considered safe.
3456 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3462 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3464 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3465 effectively reduced to only comparing the least significant bit of each
3466 byte. This allows an attacker to forge messages that would be considered as
3467 authenticated in an amount of tries lower than that guaranteed by the
3468 security claims of the scheme. The module can only be compiled by the
3469 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3471 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3477 * Add a build target 'build_all_generated', to build all generated files
3478 and only that. This can be used to prepare everything that requires
3479 things like perl for a system that lacks perl and then move everything
3480 to that system and do the rest of the build there.
3484 * Backport SSL_OP_NO_RENGOTIATION
3486 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3487 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3488 changes this is no longer possible in 1.1.0. Therefore the new
3489 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3490 1.1.0 to provide equivalent functionality.
3492 Note that if an application built against 1.1.0h headers (or above) is run
3493 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3494 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3498 * Removed the OS390-Unix config target. It relied on a script that doesn't
3503 * rsaz_1024_mul_avx2 overflow bug on x86_64
3505 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3506 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3507 Analysis suggests that attacks against RSA and DSA as a result of this
3508 defect would be very difficult to perform and are not believed likely.
3509 Attacks against DH1024 are considered just feasible, because most of the
3510 work necessary to deduce information about a private key may be performed
3511 offline. The amount of resources required for such an attack would be
3512 significant. However, for an attack on TLS to be meaningful, the server
3513 would have to share the DH1024 private key among multiple clients, which is
3514 no longer an option since CVE-2016-0701.
3516 This only affects processors that support the AVX2 but not ADX extensions
3517 like Intel Haswell (4th generation).
3519 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3520 was originally found via the OSS-Fuzz project.
3525 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3527 * bn_sqrx8x_internal carry bug on x86_64
3529 There is a carry propagating bug in the x86_64 Montgomery squaring
3530 procedure. No EC algorithms are affected. Analysis suggests that attacks
3531 against RSA and DSA as a result of this defect would be very difficult to
3532 perform and are not believed likely. Attacks against DH are considered just
3533 feasible (although very difficult) because most of the work necessary to
3534 deduce information about a private key may be performed offline. The amount
3535 of resources required for such an attack would be very significant and
3536 likely only accessible to a limited number of attackers. An attacker would
3537 additionally need online access to an unpatched system using the target
3538 private key in a scenario with persistent DH parameters and a private
3539 key that is shared between multiple clients.
3541 This only affects processors that support the BMI1, BMI2 and ADX extensions
3542 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3544 This issue was reported to OpenSSL by the OSS-Fuzz project.
3549 * Malformed X.509 IPAddressFamily could cause OOB read
3551 If an X.509 certificate has a malformed IPAddressFamily extension,
3552 OpenSSL could do a one-byte buffer overread. The most likely result
3553 would be an erroneous display of the certificate in text format.
3555 This issue was reported to OpenSSL by the OSS-Fuzz project.
3560 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3562 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3563 platform rather than 'mingw'.
3567 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3568 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3569 which is the minimum version we support.
3573 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3575 * Encrypt-Then-Mac renegotiation crash
3577 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3578 negotiated where it was not in the original handshake (or vice-versa) then
3579 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3580 and servers are affected.
3582 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3587 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3589 * Truncated packet could crash via OOB read
3591 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3592 cipher is being used, then a truncated packet can cause that host to
3593 perform an out-of-bounds read, usually resulting in a crash.
3595 This issue was reported to OpenSSL by Robert Święcki of Google.
3600 * Bad (EC)DHE parameters cause a client crash
3602 If a malicious server supplies bad parameters for a DHE or ECDHE key
3603 exchange then this can result in the client attempting to dereference a
3604 NULL pointer leading to a client crash. This could be exploited in a Denial
3607 This issue was reported to OpenSSL by Guido Vranken.
3612 * BN_mod_exp may produce incorrect results on x86_64
3614 There is a carry propagating bug in the x86_64 Montgomery squaring
3615 procedure. No EC algorithms are affected. Analysis suggests that attacks
3616 against RSA and DSA as a result of this defect would be very difficult to
3617 perform and are not believed likely. Attacks against DH are considered just
3618 feasible (although very difficult) because most of the work necessary to
3619 deduce information about a private key may be performed offline. The amount
3620 of resources required for such an attack would be very significant and
3621 likely only accessible to a limited number of attackers. An attacker would
3622 additionally need online access to an unpatched system using the target
3623 private key in a scenario with persistent DH parameters and a private
3624 key that is shared between multiple clients. For example this can occur by
3625 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3626 similar to CVE-2015-3193 but must be treated as a separate problem.
3628 This issue was reported to OpenSSL by the OSS-Fuzz project.
3633 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3635 * ChaCha20/Poly1305 heap-buffer-overflow
3637 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3638 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3639 crash. This issue is not considered to be exploitable beyond a DoS.
3641 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3646 * CMS Null dereference
3648 Applications parsing invalid CMS structures can crash with a NULL pointer
3649 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3650 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3651 structure callback if an attempt is made to free certain invalid encodings.
3652 Only CHOICE structures using a callback which do not handle NULL value are
3655 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3660 * Montgomery multiplication may produce incorrect results
3662 There is a carry propagating bug in the Broadwell-specific Montgomery
3663 multiplication procedure that handles input lengths divisible by, but
3664 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3665 and DH private keys are impossible. This is because the subroutine in
3666 question is not used in operations with the private key itself and an input
3667 of the attacker's direct choice. Otherwise the bug can manifest itself as
3668 transient authentication and key negotiation failures or reproducible
3669 erroneous outcome of public-key operations with specially crafted input.
3670 Among EC algorithms only Brainpool P-512 curves are affected and one
3671 presumably can attack ECDH key negotiation. Impact was not analyzed in
3672 detail, because pre-requisites for attack are considered unlikely. Namely
3673 multiple clients have to choose the curve in question and the server has to
3674 share the private key among them, neither of which is default behaviour.
3675 Even then only clients that chose the curve will be affected.
3677 This issue was publicly reported as transient failures and was not
3678 initially recognized as a security issue. Thanks to Richard Morgan for
3679 providing reproducible case.
3684 * Removed automatic addition of RPATH in shared libraries and executables,
3685 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3689 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3691 * Fix Use After Free for large message sizes
3693 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3694 message larger than approx 16k is received then the underlying buffer to
3695 store the incoming message is reallocated and moved. Unfortunately a
3696 dangling pointer to the old location is left which results in an attempt to
3697 write to the previously freed location. This is likely to result in a
3698 crash, however it could potentially lead to execution of arbitrary code.
3700 This issue only affects OpenSSL 1.1.0a.
3702 This issue was reported to OpenSSL by Robert Święcki.
3707 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3709 * OCSP Status Request extension unbounded memory growth
3711 A malicious client can send an excessively large OCSP Status Request
3712 extension. If that client continually requests renegotiation, sending a
3713 large OCSP Status Request extension each time, then there will be unbounded
3714 memory growth on the server. This will eventually lead to a Denial Of
3715 Service attack through memory exhaustion. Servers with a default
3716 configuration are vulnerable even if they do not support OCSP. Builds using
3717 the "no-ocsp" build time option are not affected.
3719 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3724 * SSL_peek() hang on empty record
3726 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3727 sends an empty record. This could be exploited by a malicious peer in a
3728 Denial Of Service attack.
3730 This issue was reported to OpenSSL by Alex Gaynor.
3735 * Excessive allocation of memory in tls_get_message_header() and
3736 dtls1_preprocess_fragment()
3738 A (D)TLS message includes 3 bytes for its length in the header for the
3739 message. This would allow for messages up to 16Mb in length. Messages of
3740 this length are excessive and OpenSSL includes a check to ensure that a
3741 peer is sending reasonably sized messages in order to avoid too much memory
3742 being consumed to service a connection. A flaw in the logic of version
3743 1.1.0 means that memory for the message is allocated too early, prior to
3744 the excessive message length check. Due to way memory is allocated in
3745 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3746 to service a connection. This could lead to a Denial of Service through
3747 memory exhaustion. However, the excessive message length check still takes
3748 place, and this would cause the connection to immediately fail. Assuming
3749 that the application calls SSL_free() on the failed connection in a timely
3750 manner then the 21Mb of allocated memory will then be immediately freed
3751 again. Therefore the excessive memory allocation will be transitory in
3752 nature. This then means that there is only a security impact if:
3754 1) The application does not call SSL_free() in a timely manner in the event
3755 that the connection fails
3757 2) The application is working in a constrained environment where there is
3758 very little free memory
3760 3) The attacker initiates multiple connection attempts such that there are
3761 multiple connections in a state where memory has been allocated for the
3762 connection; SSL_free() has not yet been called; and there is insufficient
3763 memory to service the multiple requests.
3765 Except in the instance of (1) above any Denial Of Service is likely to be
3766 transitory because as soon as the connection fails the memory is
3767 subsequently freed again in the SSL_free() call. However there is an
3768 increased risk during this period of application crashes due to the lack of
3769 memory - which would then mean a more serious Denial of Service.
3771 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3772 (CVE-2016-6307 and CVE-2016-6308)
3776 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3777 had to be removed. Primary reason is that vendor assembler can't
3778 assemble our modules with -KPIC flag. As result it, assembly
3779 support, was not even available as option. But its lack means
3780 lack of side-channel resistant code, which is incompatible with
3781 security by todays standards. Fortunately gcc is readily available
3782 prepackaged option, which we firmly point at...
3786 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3788 * Windows command-line tool supports UTF-8 opt-in option for arguments
3789 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3790 (to any value) allows Windows user to access PKCS#12 file generated
3791 with Windows CryptoAPI and protected with non-ASCII password, as well
3792 as files generated under UTF-8 locale on Linux also protected with
3797 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3798 have been disabled by default and removed from DEFAULT, just like RC4.
3799 See the RC4 item below to re-enable both.
3803 * The method for finding the storage location for the Windows RAND seed file
3804 has changed. First we check %RANDFILE%. If that is not set then we check
3805 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3806 all else fails we fall back to C:\.
3810 * The EVP_EncryptUpdate() function has had its return type changed from void
3811 to int. A return of 0 indicates and error while a return of 1 indicates
3816 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3817 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3818 off the constant time implementation for RSA, DSA and DH have been made
3819 no-ops and deprecated.
3823 * Windows RAND implementation was simplified to only get entropy by
3824 calling CryptGenRandom(). Various other RAND-related tickets
3827 *Joseph Wylie Yandle, Rich Salz*
3829 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3830 and `OPENSSL_LH_`, respectively. The old names are available
3831 with API compatibility. They new names are now completely documented.
3835 * Unify TYPE_up_ref(obj) methods signature.
3836 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3837 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3838 int (instead of void) like all others TYPE_up_ref() methods.
3839 So now these methods also check the return value of CRYPTO_atomic_add(),
3840 and the validity of object reference counter.
3842 *fdasilvayy@gmail.com*
3844 * With Windows Visual Studio builds, the .pdb files are installed
3845 alongside the installed libraries and executables. For a static
3846 library installation, ossl_static.pdb is the associate compiler
3847 generated .pdb file to be used when linking programs.
3851 * Remove openssl.spec. Packaging files belong with the packagers.
3855 * Automatic Darwin/OSX configuration has had a refresh, it will now
3856 recognise x86_64 architectures automatically. You can still decide
3857 to build for a different bitness with the environment variable
3858 KERNEL_BITS (can be 32 or 64), for example:
3860 KERNEL_BITS=32 ./config
3864 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3865 256 bit AES and HMAC with SHA256.
3869 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3873 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3877 * To enable users to have their own config files and build file templates,
3878 Configure looks in the directory indicated by the environment variable
3879 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3880 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3881 name and is used as is.
3885 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3886 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3887 X509_CERT_FILE_CTX was removed.
3891 * "shared" builds are now the default. To create only static libraries use
3892 the "no-shared" Configure option.
3896 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3897 All of these option have not worked for some while and are fundamental
3902 * Make various cleanup routines no-ops and mark them as deprecated. Most
3903 global cleanup functions are no longer required because they are handled
3904 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3905 Explicitly de-initing can cause problems (e.g. where a library that uses
3906 OpenSSL de-inits, but an application is still using it). The affected
3907 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3908 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3909 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3910 COMP_zlib_cleanup().
3914 * --strict-warnings no longer enables runtime debugging options
3915 such as REF_DEBUG. Instead, debug options are automatically
3916 enabled with '--debug' builds.
3918 *Andy Polyakov, Emilia Käsper*
3920 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3921 have been moved out of the public header files. New functions for managing
3922 these have been added.
3926 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3927 objects have been moved out of the public header files. New
3928 functions for managing these have been added.
3932 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3933 have been moved out of the public header files. New functions for managing
3934 these have been added.
3938 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3939 moved out of the public header files. New functions for managing these
3944 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3948 * Removed the mk1mf build scripts.
3952 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3953 it is always safe to #include a header now.
3957 * Removed the aged BC-32 config and all its supporting scripts
3961 * Removed support for Ultrix, Netware, and OS/2.
3965 * Add support for HKDF.
3967 *Alessandro Ghedini*
3969 * Add support for blake2b and blake2s
3973 * Added support for "pipelining". Ciphers that have the
3974 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3975 encryptions/decryptions simultaneously. There are currently no built-in
3976 ciphers with this property but the expectation is that engines will be able
3977 to offer it to significantly improve throughput. Support has been extended
3978 into libssl so that multiple records for a single connection can be
3979 processed in one go (for >=TLS 1.1).
3983 * Added the AFALG engine. This is an async capable engine which is able to
3984 offload work to the Linux kernel. In this initial version it only supports
3985 AES128-CBC. The kernel must be version 4.1.0 or greater.
3989 * OpenSSL now uses a new threading API. It is no longer necessary to
3990 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3991 are two supported threading models: pthreads and windows threads. It is
3992 also possible to configure OpenSSL at compile time for "no-threads". The
3993 old threading API should no longer be used. The functions have been
3994 replaced with "no-op" compatibility macros.
3996 *Alessandro Ghedini, Matt Caswell*
3998 * Modify behavior of ALPN to invoke callback after SNI/servername
3999 callback, such that updates to the SSL_CTX affect ALPN.
4003 * Add SSL_CIPHER queries for authentication and key-exchange.
4007 * Changes to the DEFAULT cipherlist:
4008 - Prefer (EC)DHE handshakes over plain RSA.
4009 - Prefer AEAD ciphers over legacy ciphers.
4010 - Prefer ECDSA over RSA when both certificates are available.
4011 - Prefer TLSv1.2 ciphers/PRF.
4012 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4017 * Change the ECC default curve list to be this, in order: x25519,
4018 secp256r1, secp521r1, secp384r1.
4022 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4023 disabled by default. They can be re-enabled using the
4024 enable-weak-ssl-ciphers option to Configure.
4028 * If the server has ALPN configured, but supports no protocols that the
4029 client advertises, send a fatal "no_application_protocol" alert.
4030 This behaviour is SHALL in RFC 7301, though it isn't universally
4031 implemented by other servers.
4035 * Add X25519 support.
4036 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4037 for public and private key encoding using the format documented in
4038 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4039 key generation and key derivation.
4041 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4046 * Deprecate SRP_VBASE_get_by_user.
4047 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4048 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4049 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4050 seed, even if the seed is configured.
4052 Users should use SRP_VBASE_get1_by_user instead. Note that in
4053 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4054 also that even though configuring the SRP seed attempts to hide
4055 invalid usernames by continuing the handshake with fake
4056 credentials, this behaviour is not constant time and no strong
4057 guarantees are made that the handshake is indistinguishable from
4058 that of a valid user.
4062 * Configuration change; it's now possible to build dynamic engines
4063 without having to build shared libraries and vice versa. This
4064 only applies to the engines in `engines/`, those in `crypto/engine/`
4065 will always be built into libcrypto (i.e. "static").
4067 Building dynamic engines is enabled by default; to disable, use
4068 the configuration option "disable-dynamic-engine".
4070 The only requirements for building dynamic engines are the
4071 presence of the DSO module and building with position independent
4072 code, so they will also automatically be disabled if configuring
4073 with "disable-dso" or "disable-pic".
4075 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4076 are also taken away from openssl/opensslconf.h, as they are
4081 * Configuration change; if there is a known flag to compile
4082 position independent code, it will always be applied on the
4083 libcrypto and libssl object files, and never on the application
4084 object files. This means other libraries that use routines from
4085 libcrypto / libssl can be made into shared libraries regardless
4086 of how OpenSSL was configured.
4088 If this isn't desirable, the configuration options "disable-pic"
4089 or "no-pic" can be used to disable the use of PIC. This will
4090 also disable building shared libraries and dynamic engines.
4094 * Removed JPAKE code. It was experimental and has no wide use.
4098 * The INSTALL_PREFIX Makefile variable has been renamed to
4099 DESTDIR. That makes for less confusion on what this variable
4100 is for. Also, the configuration option --install_prefix is
4105 * Heartbeat for TLS has been removed and is disabled by default
4106 for DTLS; configure with enable-heartbeats. Code that uses the
4107 old #define's might need to be updated.
4109 *Emilia Käsper, Rich Salz*
4111 * Rename REF_CHECK to REF_DEBUG.
4115 * New "unified" build system
4117 The "unified" build system is aimed to be a common system for all
4118 platforms we support. With it comes new support for VMS.
4120 This system builds supports building in a different directory tree
4121 than the source tree. It produces one Makefile (for unix family
4122 or lookalikes), or one descrip.mms (for VMS).
4124 The source of information to make the Makefile / descrip.mms is
4125 small files called 'build.info', holding the necessary
4126 information for each directory with source to compile, and a
4127 template in Configurations, like unix-Makefile.tmpl or
4130 With this change, the library names were also renamed on Windows
4131 and on VMS. They now have names that are closer to the standard
4132 on Unix, and include the major version number, and in certain
4133 cases, the architecture they are built for. See "Notes on shared
4134 libraries" in INSTALL.
4136 We rely heavily on the perl module Text::Template.
4140 * Added support for auto-initialisation and de-initialisation of the library.
4141 OpenSSL no longer requires explicit init or deinit routines to be called,
4142 except in certain circumstances. See the OPENSSL_init_crypto() and
4143 OPENSSL_init_ssl() man pages for further information.
4147 * The arguments to the DTLSv1_listen function have changed. Specifically the
4148 "peer" argument is now expected to be a BIO_ADDR object.
4150 * Rewrite of BIO networking library. The BIO library lacked consistent
4151 support of IPv6, and adding it required some more extensive
4152 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4153 which hold all types of addresses and chains of address information.
4154 It also introduces a new API, with functions like BIO_socket,
4155 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4156 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4157 have been adapted accordingly.
4161 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4166 * CRIME protection: disable compression by default, even if OpenSSL is
4167 compiled with zlib enabled. Applications can still enable compression
4168 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4169 using the SSL_CONF library to configure compression.
4173 * The signature of the session callback configured with
4174 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4175 was explicitly marked as `const unsigned char*` instead of
4180 * Always DPURIFY. Remove the use of uninitialized memory in the
4181 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4185 * Removed many obsolete configuration items, including
4186 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4187 MD2_CHAR, MD2_INT, MD2_LONG
4189 IDEA_SHORT, IDEA_LONG
4190 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4192 *Rich Salz, with advice from Andy Polyakov*
4194 * Many BN internals have been moved to an internal header file.
4196 *Rich Salz with help from Andy Polyakov*
4198 * Configuration and writing out the results from it has changed.
4199 Files such as Makefile include/openssl/opensslconf.h and are now
4200 produced through general templates, such as Makefile.in and
4201 crypto/opensslconf.h.in and some help from the perl module
4204 Also, the center of configuration information is no longer
4205 Makefile. Instead, Configure produces a perl module in
4206 configdata.pm which holds most of the config data (in the hash
4207 table %config), the target data that comes from the target
4208 configuration in one of the `Configurations/*.conf` files (in
4213 * To clarify their intended purposes, the Configure options
4214 --prefix and --openssldir change their semantics, and become more
4215 straightforward and less interdependent.
4217 --prefix shall be used exclusively to give the location INSTALLTOP
4218 where programs, scripts, libraries, include files and manuals are
4219 going to be installed. The default is now /usr/local.
4221 --openssldir shall be used exclusively to give the default
4222 location OPENSSLDIR where certificates, private keys, CRLs are
4223 managed. This is also where the default openssl.cnf gets
4225 If the directory given with this option is a relative path, the
4226 values of both the --prefix value and the --openssldir value will
4227 be combined to become OPENSSLDIR.
4228 The default for --openssldir is INSTALLTOP/ssl.
4230 Anyone who uses --openssldir to specify where OpenSSL is to be
4231 installed MUST change to use --prefix instead.
4235 * The GOST engine was out of date and therefore it has been removed. An up
4236 to date GOST engine is now being maintained in an external repository.
4237 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4238 support for GOST ciphersuites (these are only activated if a GOST engine
4243 * EGD is no longer supported by default; use enable-egd when
4246 *Ben Kaduk and Rich Salz*
4248 * The distribution now has Makefile.in files, which are used to
4249 create Makefile's when Configure is run. *Configure must be run
4250 before trying to build now.*
4254 * The return value for SSL_CIPHER_description() for error conditions
4259 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4261 Obtaining and performing DNSSEC validation of TLSA records is
4262 the application's responsibility. The application provides
4263 the TLSA records of its choice to OpenSSL, and these are then
4264 used to authenticate the peer.
4266 The TLSA records need not even come from DNS. They can, for
4267 example, be used to implement local end-entity certificate or
4268 trust-anchor "pinning", where the "pin" data takes the form
4269 of TLSA records, which can augment or replace verification
4270 based on the usual WebPKI public certification authorities.
4274 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4275 continues to support deprecated interfaces in default builds.
4276 However, applications are strongly advised to compile their
4277 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4278 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4279 or the 1.1.0 releases.
4281 In environments in which all applications have been ported to
4282 not use any deprecated interfaces OpenSSL's Configure script
4283 should be used with the --api=1.1.0 option to entirely remove
4284 support for the deprecated features from the library and
4285 unconditionally disable them in the installed headers.
4286 Essentially the same effect can be achieved with the "no-deprecated"
4287 argument to Configure, except that this will always restrict
4288 the build to just the latest API, rather than a fixed API
4291 As applications are ported to future revisions of the API,
4292 they should update their compile-time OPENSSL_API_COMPAT define
4293 accordingly, but in most cases should be able to continue to
4294 compile with later releases.
4296 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4297 0x10000000L and 0x00908000L, respectively. However those
4298 versions did not support the OPENSSL_API_COMPAT feature, and
4299 so applications are not typically tested for explicit support
4300 of just the undeprecated features of either release.
4304 * Add support for setting the minimum and maximum supported protocol.
4305 It can bet set via the SSL_set_min_proto_version() and
4306 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4307 MaxProtocol. It's recommended to use the new APIs to disable
4308 protocols instead of disabling individual protocols using
4309 SSL_set_options() or SSL_CONF's Protocol. This change also
4310 removes support for disabling TLS 1.2 in the OpenSSL TLS
4311 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4315 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4319 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4320 and integrates ECDSA and ECDH functionality into EC. Implementations can
4321 now redirect key generation and no longer need to convert to or from
4324 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4325 include the ec.h header file instead.
4329 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4330 ciphers who are no longer supported and drops support the ephemeral RSA key
4331 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4335 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4336 opaque. For HMAC_CTX, the following constructors and destructors
4339 HMAC_CTX *HMAC_CTX_new(void);
4340 void HMAC_CTX_free(HMAC_CTX *ctx);
4342 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4343 destroy such methods has been added. See EVP_MD_meth_new(3) and
4344 EVP_CIPHER_meth_new(3) for documentation.
4347 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4348 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4349 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4350 an already created structure.
4351 2) For consistency with the majority of our object creators and
4352 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4353 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4354 for deprecated builds.
4358 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4359 cryptographic operations to be performed asynchronously as long as an
4360 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4361 further details. Libssl has also had this capability integrated with the
4362 introduction of the new mode SSL_MODE_ASYNC and associated error
4363 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4364 pages. This work was developed in partnership with Intel Corp.
4368 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4369 always enabled now. If you want to disable the support you should
4370 exclude it using the list of supported ciphers. This also means that the
4371 "-no_ecdhe" option has been removed from s_server.
4375 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4376 SSL_{CTX_}set1_curves() which can set a list.
4380 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4381 curve you want to support using SSL_{CTX_}set1_curves().
4385 * State machine rewrite. The state machine code has been significantly
4386 refactored in order to remove much duplication of code and solve issues
4387 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4388 further details). This change does have some associated API changes.
4389 Notably the SSL_state() function has been removed and replaced by
4390 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4391 SSL_set_state() has been removed altogether. The previous handshake states
4392 defined in ssl.h and ssl3.h have also been removed.
4396 * All instances of the string "ssleay" in the public API were replaced
4397 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4398 Some error codes related to internal RSA_eay API's were renamed.
4402 * The demo files in crypto/threads were moved to demo/threads.
4406 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4409 *Matt Caswell, Rich Salz*
4411 * New ASN.1 embed macro.
4413 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4414 structure is not allocated: it is part of the parent. That is instead of
4422 This reduces memory fragmentation and make it impossible to accidentally
4423 set a mandatory field to NULL.
4425 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4426 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4427 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4432 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4436 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4437 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4438 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4439 DES and RC4 ciphersuites.
4443 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4444 This changes the decoding behaviour for some invalid messages,
4445 though the change is mostly in the more lenient direction, and
4446 legacy behaviour is preserved as much as possible.
4450 * Fix no-stdio build.
4451 *David Woodhouse <David.Woodhouse@intel.com> and also*
4452 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4454 * New testing framework
4455 The testing framework has been largely rewritten and is now using
4456 perl and the perl modules Test::Harness and an extended variant of
4457 Test::More called OpenSSL::Test to do its work. All test scripts in
4458 test/ have been rewritten into test recipes, and all direct calls to
4459 executables in test/Makefile have become individual recipes using the
4460 simplified testing OpenSSL::Test::Simple.
4462 For documentation on our testing modules, do:
4464 perldoc test/testlib/OpenSSL/Test/Simple.pm
4465 perldoc test/testlib/OpenSSL/Test.pm
4469 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4470 are used; the latter aborts on memory leaks (usually checked on exit).
4471 Some undocumented "set malloc, etc., hooks" functions were removed
4472 and others were changed. All are now documented.
4476 * In DSA_generate_parameters_ex, if the provided seed is too short,
4479 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4481 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4482 from RFC4279, RFC4785, RFC5487, RFC5489.
4484 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4485 original RSA_PSK patch.
4489 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4490 era flag was never set throughout the codebase (only read). Also removed
4491 SSL3_FLAGS_POP_BUFFER which was only used if
4492 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4496 * Changed the default name options in the "ca", "crl", "req" and "x509"
4497 to be "oneline" instead of "compat".
4501 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4502 not aware of clients that still exhibit this bug, and the workaround
4503 hasn't been working properly for a while.
4507 * The return type of BIO_number_read() and BIO_number_written() as well as
4508 the corresponding num_read and num_write members in the BIO structure has
4509 changed from unsigned long to uint64_t. On platforms where an unsigned
4510 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4515 * Given the pervasive nature of TLS extensions it is inadvisable to run
4516 OpenSSL without support for them. It also means that maintaining
4517 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4518 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4522 * Removed support for the two export grade static DH ciphersuites
4523 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4524 were newly added (along with a number of other static DH ciphersuites) to
4525 1.0.2. However the two export ones have *never* worked since they were
4526 introduced. It seems strange in any case to be adding new export
4527 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4531 * Version negotiation has been rewritten. In particular SSLv23_method(),
4532 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4533 and turned into macros which simply call the new preferred function names
4534 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4535 should use the new names instead. Also as part of this change the ssl23.h
4536 header file has been removed.
4540 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4541 code and the associated standard is no longer considered fit-for-purpose.
4545 * RT2547 was closed. When generating a private key, try to make the
4546 output file readable only by the owner. This behavior change might
4547 be noticeable when interacting with other software.
4549 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4554 * Added HTTP GET support to the ocsp command.
4558 * Changed default digest for the dgst and enc commands from MD5 to
4563 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4567 * Added support for TLS extended master secret from
4568 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4569 initial patch which was a great help during development.
4573 * All libssl internal structures have been removed from the public header
4574 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4575 now redundant). Users should not attempt to access internal structures
4576 directly. Instead they should use the provided API functions.
4580 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4581 Access to deprecated functions can be re-enabled by running config with
4582 "enable-deprecated". In addition applications wishing to use deprecated
4583 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4584 will, by default, disable some transitive includes that previously existed
4585 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4589 * Added support for OCB mode. OpenSSL has been granted a patent license
4590 compatible with the OpenSSL license for use of OCB. Details are available
4591 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4592 for OCB can be removed by calling config with no-ocb.
4596 * SSLv2 support has been removed. It still supports receiving a SSLv2
4597 compatible client hello.
4601 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4602 done while fixing the error code for the key-too-small case.
4604 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4606 * CA.sh has been removed; use CA.pl instead.
4610 * Removed old DES API.
4614 * Remove various unsupported platforms:
4620 Sinix/ReliantUNIX RM400
4625 16-bit platforms such as WIN16
4629 * Clean up OPENSSL_NO_xxx #define's
4630 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4631 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4632 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4633 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4634 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4635 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4636 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4637 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4638 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4639 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4643 * Cleaned up dead code
4644 Remove all but one '#ifdef undef' which is to be looked at.
4648 * Clean up calling of xxx_free routines.
4649 Just like free(), fix most of the xxx_free routines to accept
4650 NULL. Remove the non-null checks from callers. Save much code.
4654 * Add secure heap for storage of private keys (when possible).
4655 Add BIO_s_secmem(), CBIGNUM, etc.
4656 Contributed by Akamai Technologies under our Corporate CLA.
4660 * Experimental support for a new, fast, unbiased prime candidate generator,
4661 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4663 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4665 * New output format NSS in the sess_id command line tool. This allows
4666 exporting the session id and the master key in NSS keylog format.
4668 *Martin Kaiser <martin@kaiser.cx>*
4670 * Harmonize version and its documentation. -f flag is used to display
4673 *mancha <mancha1@zoho.com>*
4675 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4676 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4678 *mancha <mancha1@zoho.com>*
4680 * Fix some double frees. These are not thought to be exploitable.
4682 *mancha <mancha1@zoho.com>*
4684 * A missing bounds check in the handling of the TLS heartbeat extension
4685 can be used to reveal up to 64k of memory to a connected client or
4688 Thanks for Neel Mehta of Google Security for discovering this bug and to
4689 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4690 preparing the fix ([CVE-2014-0160])
4692 *Adam Langley, Bodo Moeller*
4694 * Fix for the attack described in the paper "Recovering OpenSSL
4695 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4696 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4697 <http://eprint.iacr.org/2014/140>
4699 Thanks to Yuval Yarom and Naomi Benger for discovering this
4700 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4702 *Yuval Yarom and Naomi Benger*
4704 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4705 this fixes a limitation in previous versions of OpenSSL.
4709 * Experimental encrypt-then-mac support.
4711 Experimental support for encrypt then mac from
4712 draft-gutmann-tls-encrypt-then-mac-02.txt
4714 To enable it set the appropriate extension number (0x42 for the test
4715 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4717 For non-compliant peers (i.e. just about everything) this should have no
4720 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4724 * Add EVP support for key wrapping algorithms, to avoid problems with
4725 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4726 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4727 algorithms and include tests cases.
4731 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4736 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4737 MGF1 digest and OAEP label.
4741 * Make openssl verify return errors.
4743 *Chris Palmer <palmer@google.com> and Ben Laurie*
4745 * New function ASN1_TIME_diff to calculate the difference between two
4746 ASN1_TIME structures or one structure and the current time.
4750 * Update fips_test_suite to support multiple command line options. New
4751 test to induce all self test errors in sequence and check expected
4756 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4757 sign or verify all in one operation.
4761 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4762 test programs and fips_test_suite. Includes functionality to parse
4763 the minimal script output of fipsalgest.pl directly.
4767 * Add authorisation parameter to FIPS_module_mode_set().
4771 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4775 * Use separate DRBG fields for internal and external flags. New function
4776 FIPS_drbg_health_check() to perform on demand health checking. Add
4777 generation tests to fips_test_suite with reduced health check interval to
4778 demonstrate periodic health checking. Add "nodh" option to
4779 fips_test_suite to skip very slow DH test.
4783 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4788 * More extensive health check for DRBG checking many more failure modes.
4789 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4790 combination: call this in fips_test_suite.
4794 * Add support for canonical generation of DSA parameter 'g'. See
4797 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4798 POST to handle HMAC cases.
4802 * Add functions FIPS_module_version() and FIPS_module_version_text()
4803 to return numerical and string versions of the FIPS module number.
4807 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4808 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4809 outside the validated module in the FIPS capable OpenSSL.
4813 * Minor change to DRBG entropy callback semantics. In some cases
4814 there is no multiple of the block length between min_len and
4815 max_len. Allow the callback to return more than max_len bytes
4816 of entropy but discard any extra: it is the callback's responsibility
4817 to ensure that the extra data discarded does not impact the
4818 requested amount of entropy.
4822 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4823 information in FIPS186-3, SP800-57 and SP800-131A.
4827 * CCM support via EVP. Interface is very similar to GCM case except we
4828 must supply all data in one chunk (i.e. no update, final) and the
4829 message length must be supplied if AAD is used. Add algorithm test
4834 * Initial version of POST overhaul. Add POST callback to allow the status
4835 of POST to be monitored and/or failures induced. Modify fips_test_suite
4836 to use callback. Always run all selftests even if one fails.
4840 * XTS support including algorithm test driver in the fips_gcmtest program.
4841 Note: this does increase the maximum key length from 32 to 64 bytes but
4842 there should be no binary compatibility issues as existing applications
4843 will never use XTS mode.
4847 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4848 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4849 performs algorithm blocking for unapproved PRNG types. Also do not
4850 set PRNG type in FIPS_mode_set(): leave this to the application.
4851 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4852 the standard OpenSSL PRNG: set additional data to a date time vector.
4856 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4857 This shouldn't present any incompatibility problems because applications
4858 shouldn't be using these directly and any that are will need to rethink
4859 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4863 * Extensive self tests and health checking required by SP800-90 DRBG.
4864 Remove strength parameter from FIPS_drbg_instantiate and always
4865 instantiate at maximum supported strength.
4869 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4873 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4877 * New function DH_compute_key_padded() to compute a DH key and pad with
4878 leading zeroes if needed: this complies with SP800-56A et al.
4882 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4883 anything, incomplete, subject to change and largely untested at present.
4887 * Modify fipscanisteronly build option to only build the necessary object
4888 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4892 * Add experimental option FIPSSYMS to give all symbols in
4893 fipscanister.o and FIPS or fips prefix. This will avoid
4894 conflicts with future versions of OpenSSL. Add perl script
4895 util/fipsas.pl to preprocess assembly language source files
4896 and rename any affected symbols.
4900 * Add selftest checks and algorithm block of non-fips algorithms in
4901 FIPS mode. Remove DES2 from selftests.
4905 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4906 return internal method without any ENGINE dependencies. Add new
4907 tiny fips sign and verify functions.
4911 * New build option no-ec2m to disable characteristic 2 code.
4915 * New build option "fipscanisteronly". This only builds fipscanister.o
4916 and (currently) associated fips utilities. Uses the file Makefile.fips
4917 instead of Makefile.org as the prototype.
4921 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4922 Update fips_gcmtest to use IV generator.
4926 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4927 setting output buffer to NULL. The `*Final` function must be
4928 called although it will not retrieve any additional data. The tag
4929 can be set or retrieved with a ctrl. The IV length is by default 12
4930 bytes (96 bits) but can be set to an alternative value. If the IV
4931 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4936 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4937 underlying do_cipher function handles all cipher semantics itself
4938 including padding and finalisation. This is useful if (for example)
4939 an ENGINE cipher handles block padding itself. The behaviour of
4940 do_cipher is subtly changed if this flag is set: the return value
4941 is the number of characters written to the output buffer (zero is
4942 no longer an error code) or a negative error code. Also if the
4943 input buffer is NULL and length 0 finalisation should be performed.
4947 * If a candidate issuer certificate is already part of the constructed
4948 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4952 * Improve forward-security support: add functions
4954 void SSL_CTX_set_not_resumable_session_callback(
4955 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4956 void SSL_set_not_resumable_session_callback(
4957 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4959 for use by SSL/TLS servers; the callback function will be called whenever a
4960 new session is created, and gets to decide whether the session may be
4961 cached to make it resumable (return 0) or not (return 1). (As by the
4962 SSL/TLS protocol specifications, the session_id sent by the server will be
4963 empty to indicate that the session is not resumable; also, the server will
4964 not generate RFC 4507 (RFC 5077) session tickets.)
4966 A simple reasonable callback implementation is to return is_forward_secure.
4967 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4968 by the SSL/TLS server library, indicating whether it can provide forward
4971 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4973 * New -verify_name option in command line utilities to set verification
4978 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4979 Add CMAC pkey methods.
4983 * Experimental renegotiation in s_server -www mode. If the client
4984 browses /reneg connection is renegotiated. If /renegcert it is
4985 renegotiated requesting a certificate.
4989 * Add an "external" session cache for debugging purposes to s_server. This
4990 should help trace issues which normally are only apparent in deployed
4991 multi-process servers.
4995 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4996 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4997 BIO_set_cipher() and some obscure PEM functions were changed so they
4998 can now return an error. The RAND changes required a change to the
4999 RAND_METHOD structure.
5003 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5004 a gcc attribute to warn if the result of a function is ignored. This
5005 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5006 whose return value is often ignored.
5010 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5011 These allow SCTs (signed certificate timestamps) to be requested and
5012 validated when establishing a connection.
5014 *Rob Percival <robpercival@google.com>*
5019 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5021 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5022 used even when parsing explicit parameters, when loading a encoded key
5023 or calling `EC_GROUP_new_from_ecpkparameters()`/
5024 `EC_GROUP_new_from_ecparameters()`.
5025 This prevents bypass of security hardening and performance gains,
5026 especially for curves with specialized EC_METHODs.
5027 By default, if a key encoded with explicit parameters is loaded and later
5028 encoded, the output is still encoded with explicit parameters, even if
5029 internally a "named" EC_GROUP is used for computation.
5033 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5034 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5035 NULL. After this change, only the cofactor parameter can be NULL. It also
5036 does some minimal sanity checks on the passed order.
5041 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5042 An attack is simple, if the first CMS_recipientInfo is valid but the
5043 second CMS_recipientInfo is chosen ciphertext. If the second
5044 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5045 encryption key will be replaced by garbage, and the message cannot be
5046 decoded, but if the RSA decryption fails, the correct encryption key is
5047 used and the recipient will not notice the attack.
5048 As a work around for this potential attack the length of the decrypted
5049 key must be equal to the cipher default key length, in case the
5050 certifiate is not given and all recipientInfo are tried out.
5051 The old behaviour can be re-enabled in the CMS code by setting the
5052 CMS_DEBUG_DECRYPT flag.
5057 * Document issue with installation paths in diverse Windows builds
5059 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5060 binaries and run-time config file.
5065 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5067 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5068 This changes the size when using the `genpkey` command when no size is given.
5069 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5070 generation commands to use 2048 bits by default.
5074 * Add FIPS support for Android Arm 64-bit
5076 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5077 Module in Version 2.0.10. For some reason, the corresponding target
5078 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5079 built with FIPS support on Android Arm 64-bit. This omission has been
5082 *Matthias St. Pierre*
5084 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5086 * 0-byte record padding oracle
5088 If an application encounters a fatal protocol error and then calls
5089 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5090 then OpenSSL can respond differently to the calling application if a 0 byte
5091 record is received with invalid padding compared to if a 0 byte record is
5092 received with an invalid MAC. If the application then behaves differently
5093 based on that in a way that is detectable to the remote peer, then this
5094 amounts to a padding oracle that could be used to decrypt data.
5096 In order for this to be exploitable "non-stitched" ciphersuites must be in
5097 use. Stitched ciphersuites are optimised implementations of certain
5098 commonly used ciphersuites. Also the application must call SSL_shutdown()
5099 twice even if a protocol error has occurred (applications should not do
5100 this but some do anyway).
5102 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5103 Aviram, with additional investigation by Steven Collison and Andrew
5104 Hourselt. It was reported to OpenSSL on 10th December 2018.
5109 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5113 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5115 * Microarchitecture timing vulnerability in ECC scalar multiplication
5117 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5118 shown to be vulnerable to a microarchitecture timing side channel attack.
5119 An attacker with sufficient access to mount local timing attacks during
5120 ECDSA signature generation could recover the private key.
5122 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5123 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5129 * Timing vulnerability in DSA signature generation
5131 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5132 timing side channel attack. An attacker could use variations in the signing
5133 algorithm to recover the private key.
5135 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5140 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5141 Module, accidentally introduced while backporting security fixes from the
5142 development branch and hindering the use of ECC in FIPS mode.
5146 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5148 * Client DoS due to large DH parameter
5150 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5151 malicious server can send a very large prime value to the client. This will
5152 cause the client to spend an unreasonably long period of time generating a
5153 key for this prime resulting in a hang until the client has finished. This
5154 could be exploited in a Denial Of Service attack.
5156 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5161 * Cache timing vulnerability in RSA Key Generation
5163 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5164 a cache timing side channel attack. An attacker with sufficient access to
5165 mount cache timing attacks during the RSA key generation process could
5166 recover the private key.
5168 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5169 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5174 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5175 parameter is no longer accepted, as it leads to a corrupt table. NULL
5176 pem_str is reserved for alias entries only.
5180 * Revert blinding in ECDSA sign and instead make problematic addition
5181 length-invariant. Switch even to fixed-length Montgomery multiplication.
5185 * Change generating and checking of primes so that the error rate of not
5186 being prime depends on the intended use based on the size of the input.
5187 For larger primes this will result in more rounds of Miller-Rabin.
5188 The maximal error rate for primes with more than 1080 bits is lowered
5191 *Kurt Roeckx, Annie Yousar*
5193 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5197 * Add blinding to ECDSA and DSA signatures to protect against side channel
5198 attacks discovered by Keegan Ryan (NCC Group).
5202 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5203 now allow empty (zero character) pass phrases.
5207 * Certificate time validation (X509_cmp_time) enforces stricter
5208 compliance with RFC 5280. Fractional seconds and timezone offsets
5209 are no longer allowed.
5213 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5215 * Constructed ASN.1 types with a recursive definition could exceed the stack
5217 Constructed ASN.1 types with a recursive definition (such as can be found
5218 in PKCS7) could eventually exceed the stack given malicious input with
5219 excessive recursion. This could result in a Denial Of Service attack. There
5220 are no such structures used within SSL/TLS that come from untrusted sources
5221 so this is considered safe.
5223 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5229 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5231 * Read/write after SSL object in error state
5233 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5234 mechanism. The intent was that if a fatal error occurred during a handshake
5235 then OpenSSL would move into the error state and would immediately fail if
5236 you attempted to continue the handshake. This works as designed for the
5237 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5238 SSL_connect()), however due to a bug it does not work correctly if
5239 SSL_read() or SSL_write() is called directly. In that scenario, if the
5240 handshake fails then a fatal error will be returned in the initial function
5241 call. If SSL_read()/SSL_write() is subsequently called by the application
5242 for the same SSL object then it will succeed and the data is passed without
5243 being decrypted/encrypted directly from the SSL/TLS record layer.
5245 In order to exploit this issue an application bug would have to be present
5246 that resulted in a call to SSL_read()/SSL_write() being issued after having
5247 already received a fatal error.
5249 This issue was reported to OpenSSL by David Benjamin (Google).
5254 * rsaz_1024_mul_avx2 overflow bug on x86_64
5256 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5257 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5258 Analysis suggests that attacks against RSA and DSA as a result of this
5259 defect would be very difficult to perform and are not believed likely.
5260 Attacks against DH1024 are considered just feasible, because most of the
5261 work necessary to deduce information about a private key may be performed
5262 offline. The amount of resources required for such an attack would be
5263 significant. However, for an attack on TLS to be meaningful, the server
5264 would have to share the DH1024 private key among multiple clients, which is
5265 no longer an option since CVE-2016-0701.
5267 This only affects processors that support the AVX2 but not ADX extensions
5268 like Intel Haswell (4th generation).
5270 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5271 was originally found via the OSS-Fuzz project.
5276 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5278 * bn_sqrx8x_internal carry bug on x86_64
5280 There is a carry propagating bug in the x86_64 Montgomery squaring
5281 procedure. No EC algorithms are affected. Analysis suggests that attacks
5282 against RSA and DSA as a result of this defect would be very difficult to
5283 perform and are not believed likely. Attacks against DH are considered just
5284 feasible (although very difficult) because most of the work necessary to
5285 deduce information about a private key may be performed offline. The amount
5286 of resources required for such an attack would be very significant and
5287 likely only accessible to a limited number of attackers. An attacker would
5288 additionally need online access to an unpatched system using the target
5289 private key in a scenario with persistent DH parameters and a private
5290 key that is shared between multiple clients.
5292 This only affects processors that support the BMI1, BMI2 and ADX extensions
5293 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5295 This issue was reported to OpenSSL by the OSS-Fuzz project.
5300 * Malformed X.509 IPAddressFamily could cause OOB read
5302 If an X.509 certificate has a malformed IPAddressFamily extension,
5303 OpenSSL could do a one-byte buffer overread. The most likely result
5304 would be an erroneous display of the certificate in text format.
5306 This issue was reported to OpenSSL by the OSS-Fuzz project.
5310 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5312 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5313 platform rather than 'mingw'.
5317 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5319 * Truncated packet could crash via OOB read
5321 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5322 cipher is being used, then a truncated packet can cause that host to
5323 perform an out-of-bounds read, usually resulting in a crash.
5325 This issue was reported to OpenSSL by Robert Święcki of Google.
5330 * BN_mod_exp may produce incorrect results on x86_64
5332 There is a carry propagating bug in the x86_64 Montgomery squaring
5333 procedure. No EC algorithms are affected. Analysis suggests that attacks
5334 against RSA and DSA as a result of this defect would be very difficult to
5335 perform and are not believed likely. Attacks against DH are considered just
5336 feasible (although very difficult) because most of the work necessary to
5337 deduce information about a private key may be performed offline. The amount
5338 of resources required for such an attack would be very significant and
5339 likely only accessible to a limited number of attackers. An attacker would
5340 additionally need online access to an unpatched system using the target
5341 private key in a scenario with persistent DH parameters and a private
5342 key that is shared between multiple clients. For example this can occur by
5343 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5344 similar to CVE-2015-3193 but must be treated as a separate problem.
5346 This issue was reported to OpenSSL by the OSS-Fuzz project.
5351 * Montgomery multiplication may produce incorrect results
5353 There is a carry propagating bug in the Broadwell-specific Montgomery
5354 multiplication procedure that handles input lengths divisible by, but
5355 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5356 and DH private keys are impossible. This is because the subroutine in
5357 question is not used in operations with the private key itself and an input
5358 of the attacker's direct choice. Otherwise the bug can manifest itself as
5359 transient authentication and key negotiation failures or reproducible
5360 erroneous outcome of public-key operations with specially crafted input.
5361 Among EC algorithms only Brainpool P-512 curves are affected and one
5362 presumably can attack ECDH key negotiation. Impact was not analyzed in
5363 detail, because pre-requisites for attack are considered unlikely. Namely
5364 multiple clients have to choose the curve in question and the server has to
5365 share the private key among them, neither of which is default behaviour.
5366 Even then only clients that chose the curve will be affected.
5368 This issue was publicly reported as transient failures and was not
5369 initially recognized as a security issue. Thanks to Richard Morgan for
5370 providing reproducible case.
5375 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5376 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5377 prevent issues where no progress is being made and the peer continually
5378 sends unrecognised record types, using up resources processing them.
5382 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5384 * Missing CRL sanity check
5386 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5387 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5388 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5390 This issue only affects the OpenSSL 1.0.2i
5395 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5397 * OCSP Status Request extension unbounded memory growth
5399 A malicious client can send an excessively large OCSP Status Request
5400 extension. If that client continually requests renegotiation, sending a
5401 large OCSP Status Request extension each time, then there will be unbounded
5402 memory growth on the server. This will eventually lead to a Denial Of
5403 Service attack through memory exhaustion. Servers with a default
5404 configuration are vulnerable even if they do not support OCSP. Builds using
5405 the "no-ocsp" build time option are not affected.
5407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5412 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5415 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5421 * OOB write in MDC2_Update()
5423 An overflow can occur in MDC2_Update() either if called directly or
5424 through the EVP_DigestUpdate() function using MDC2. If an attacker
5425 is able to supply very large amounts of input data after a previous
5426 call to EVP_EncryptUpdate() with a partial block then a length check
5427 can overflow resulting in a heap corruption.
5429 The amount of data needed is comparable to SIZE_MAX which is impractical
5432 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5437 * Malformed SHA512 ticket DoS
5439 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5440 DoS attack where a malformed ticket will result in an OOB read which will
5443 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5444 a custom server callback and ticket lookup mechanism.
5446 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5451 * OOB write in BN_bn2dec()
5453 The function BN_bn2dec() does not check the return value of BN_div_word().
5454 This can cause an OOB write if an application uses this function with an
5455 overly large BIGNUM. This could be a problem if an overly large certificate
5456 or CRL is printed out from an untrusted source. TLS is not affected because
5457 record limits will reject an oversized certificate before it is parsed.
5459 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5464 * OOB read in TS_OBJ_print_bio()
5466 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5467 the total length the OID text representation would use and not the amount
5468 of data written. This will result in OOB reads when large OIDs are
5471 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5476 * Pointer arithmetic undefined behaviour
5478 Avoid some undefined pointer arithmetic
5480 A common idiom in the codebase is to check limits in the following manner:
5483 Where "p" points to some malloc'd data of SIZE bytes and
5486 "len" here could be from some externally supplied data (e.g. from a TLS
5489 The rules of C pointer arithmetic are such that "p + len" is only well
5490 defined where len <= SIZE. Therefore the above idiom is actually
5491 undefined behaviour.
5493 For example this could cause problems if some malloc implementation
5494 provides an address for "p" such that "p + len" actually overflows for
5495 values of len that are too big and therefore p + len < limit.
5497 This issue was reported to OpenSSL by Guido Vranken
5502 * Constant time flag not preserved in DSA signing
5504 Operations in the DSA signing algorithm should run in constant time in
5505 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5506 implementation means that a non-constant time codepath is followed for
5507 certain operations. This has been demonstrated through a cache-timing
5508 attack to be sufficient for an attacker to recover the private DSA key.
5510 This issue was reported by César Pereida (Aalto University), Billy Brumley
5511 (Tampere University of Technology), and Yuval Yarom (The University of
5512 Adelaide and NICTA).
5517 * DTLS buffered message DoS
5519 In a DTLS connection where handshake messages are delivered out-of-order
5520 those messages that OpenSSL is not yet ready to process will be buffered
5521 for later use. Under certain circumstances, a flaw in the logic means that
5522 those messages do not get removed from the buffer even though the handshake
5523 has been completed. An attacker could force up to approx. 15 messages to
5524 remain in the buffer when they are no longer required. These messages will
5525 be cleared when the DTLS connection is closed. The default maximum size for
5526 a message is 100k. Therefore the attacker could force an additional 1500k
5527 to be consumed per connection. By opening many simulataneous connections an
5528 attacker could cause a DoS attack through memory exhaustion.
5530 This issue was reported to OpenSSL by Quan Luo.
5535 * DTLS replay protection DoS
5537 A flaw in the DTLS replay attack protection mechanism means that records
5538 that arrive for future epochs update the replay protection "window" before
5539 the MAC for the record has been validated. This could be exploited by an
5540 attacker by sending a record for the next epoch (which does not have to
5541 decrypt or have a valid MAC), with a very large sequence number. This means
5542 that all subsequent legitimate packets are dropped causing a denial of
5543 service for a specific DTLS connection.
5545 This issue was reported to OpenSSL by the OCAP audit team.
5550 * Certificate message OOB reads
5552 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5553 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5554 theoretical DoS risk but this has not been observed in practice on common
5557 The messages affected are client certificate, client certificate request
5558 and server certificate. As a result the attack can only be performed
5559 against a client or a server which enables client authentication.
5561 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5566 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5568 * Prevent padding oracle in AES-NI CBC MAC check
5570 A MITM attacker can use a padding oracle attack to decrypt traffic
5571 when the connection uses an AES CBC cipher and the server support
5574 This issue was introduced as part of the fix for Lucky 13 padding
5575 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5576 constant time by making sure that always the same bytes are read and
5577 compared against either the MAC or padding bytes. But it no longer
5578 checked that there was enough data to have both the MAC and padding
5581 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5585 * Fix EVP_EncodeUpdate overflow
5587 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5588 Base64 encoding of binary data. If an attacker is able to supply very large
5589 amounts of input data then a length check can overflow resulting in a heap
5592 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5593 the `PEM_write_bio*` family of functions. These are mainly used within the
5594 OpenSSL command line applications, so any application which processes data
5595 from an untrusted source and outputs it as a PEM file should be considered
5596 vulnerable to this issue. User applications that call these APIs directly
5597 with large amounts of untrusted data may also be vulnerable.
5599 This issue was reported by Guido Vranken.
5604 * Fix EVP_EncryptUpdate overflow
5606 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5607 is able to supply very large amounts of input data after a previous call to
5608 EVP_EncryptUpdate() with a partial block then a length check can overflow
5609 resulting in a heap corruption. Following an analysis of all OpenSSL
5610 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5611 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5612 the first called function after an EVP_EncryptInit(), and therefore that
5613 specific call must be safe. The second form is where the length passed to
5614 EVP_EncryptUpdate() can be seen from the code to be some small value and
5615 therefore there is no possibility of an overflow. Since all instances are
5616 one of these two forms, it is believed that there can be no overflows in
5617 internal code due to this problem. It should be noted that
5618 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5619 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5620 of these calls have also been analysed too and it is believed there are no
5621 instances in internal usage where an overflow could occur.
5623 This issue was reported by Guido Vranken.
5628 * Prevent ASN.1 BIO excessive memory allocation
5630 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5631 a short invalid encoding can cause allocation of large amounts of memory
5632 potentially consuming excessive resources or exhausting memory.
5634 Any application parsing untrusted data through d2i BIO functions is
5635 affected. The memory based functions such as d2i_X509() are *not* affected.
5636 Since the memory based functions are used by the TLS library, TLS
5637 applications are not affected.
5639 This issue was reported by Brian Carpenter.
5646 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5647 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5648 in arbitrary stack data being returned in the buffer.
5650 This issue was reported by Guido Vranken.
5655 * Modify behavior of ALPN to invoke callback after SNI/servername
5656 callback, such that updates to the SSL_CTX affect ALPN.
5660 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5665 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5666 methods are enabled and ssl2 is disabled the methods return NULL.
5670 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5672 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5673 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5674 provide any "EXPORT" or "LOW" strength ciphers.
5678 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5679 is by default disabled at build-time. Builds that are not configured with
5680 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5681 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5682 will need to explicitly call either of:
5684 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5686 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5688 as appropriate. Even if either of those is used, or the application
5689 explicitly uses the version-specific SSLv2_method() or its client and
5690 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5691 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5692 ciphers, and SSLv2 56-bit DES are no longer available.
5697 * Fix a double-free in DSA code
5699 A double free bug was discovered when OpenSSL parses malformed DSA private
5700 keys and could lead to a DoS attack or memory corruption for applications
5701 that receive DSA private keys from untrusted sources. This scenario is
5704 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5710 * Disable SRP fake user seed to address a server memory leak.
5712 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5714 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5715 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5716 was changed to ignore the "fake user" SRP seed, even if the seed
5719 Users should use SRP_VBASE_get1_by_user instead. Note that in
5720 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5721 also that even though configuring the SRP seed attempts to hide
5722 invalid usernames by continuing the handshake with fake
5723 credentials, this behaviour is not constant time and no strong
5724 guarantees are made that the handshake is indistinguishable from
5725 that of a valid user.
5730 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5732 In the BN_hex2bn function the number of hex digits is calculated using an
5733 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5734 large values of `i` this can result in `bn_expand` not allocating any
5735 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5736 field as NULL leading to a subsequent NULL ptr deref. For very large values
5737 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5738 In this case memory is allocated to the internal BIGNUM data field, but it
5739 is insufficiently sized leading to heap corruption. A similar issue exists
5740 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5741 is ever called by user applications with very large untrusted hex/dec data.
5742 This is anticipated to be a rare occurrence.
5744 All OpenSSL internal usage of these functions use data that is not expected
5745 to be untrusted, e.g. config file data or application command line
5746 arguments. If user developed applications generate config file data based
5747 on untrusted data then it is possible that this could also lead to security
5748 consequences. This is also anticipated to be rare.
5750 This issue was reported to OpenSSL by Guido Vranken.
5755 * Fix memory issues in `BIO_*printf` functions
5757 The internal `fmtstr` function used in processing a "%s" format string in
5758 the `BIO_*printf` functions could overflow while calculating the length of a
5759 string and cause an OOB read when printing very long strings.
5761 Additionally the internal `doapr_outch` function can attempt to write to an
5762 OOB memory location (at an offset from the NULL pointer) in the event of a
5763 memory allocation failure. In 1.0.2 and below this could be caused where
5764 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5765 could be in processing a very long "%s" format string. Memory leaks can
5768 The first issue may mask the second issue dependent on compiler behaviour.
5769 These problems could enable attacks where large amounts of untrusted data
5770 is passed to the `BIO_*printf` functions. If applications use these functions
5771 in this way then they could be vulnerable. OpenSSL itself uses these
5772 functions when printing out human-readable dumps of ASN.1 data. Therefore
5773 applications that print this data could be vulnerable if the data is from
5774 untrusted sources. OpenSSL command line applications could also be
5775 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5776 as command line arguments.
5778 Libssl is not considered directly vulnerable. Additionally certificates etc
5779 received via remote connections via libssl are also unlikely to be able to
5780 trigger these issues because of message size limits enforced within libssl.
5782 This issue was reported to OpenSSL Guido Vranken.
5787 * Side channel attack on modular exponentiation
5789 A side-channel attack was found which makes use of cache-bank conflicts on
5790 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5791 of RSA keys. The ability to exploit this issue is limited as it relies on
5792 an attacker who has control of code in a thread running on the same
5793 hyper-threaded core as the victim thread which is performing decryptions.
5795 This issue was reported to OpenSSL by Yuval Yarom, The University of
5796 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5797 Nadia Heninger, University of Pennsylvania with more information at
5798 <http://cachebleed.info>.
5803 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5804 if no keysize is specified with default_bits. This fixes an
5805 omission in an earlier change that changed all RSA/DSA key generation
5806 commands to use 2048 bits by default.
5810 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5812 * DH small subgroups
5814 Historically OpenSSL only ever generated DH parameters based on "safe"
5815 primes. More recently (in version 1.0.2) support was provided for
5816 generating X9.42 style parameter files such as those required for RFC 5114
5817 support. The primes used in such files may not be "safe". Where an
5818 application is using DH configured with parameters based on primes that are
5819 not "safe" then an attacker could use this fact to find a peer's private
5820 DH exponent. This attack requires that the attacker complete multiple
5821 handshakes in which the peer uses the same private DH exponent. For example
5822 this could be used to discover a TLS server's private DH exponent if it's
5823 reusing the private DH exponent or it's using a static DH ciphersuite.
5825 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5826 TLS. It is not on by default. If the option is not set then the server
5827 reuses the same private DH exponent for the life of the server process and
5828 would be vulnerable to this attack. It is believed that many popular
5829 applications do set this option and would therefore not be at risk.
5831 The fix for this issue adds an additional check where a "q" parameter is
5832 available (as is the case in X9.42 based parameters). This detects the
5833 only known attack, and is the only possible defense for static DH
5834 ciphersuites. This could have some performance impact.
5836 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5837 default and cannot be disabled. This could have some performance impact.
5839 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5844 * SSLv2 doesn't block disabled ciphers
5846 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5847 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5848 been disabled, provided that the SSLv2 protocol was not also disabled via
5851 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5852 and Sebastian Schinzel.
5857 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5859 * BN_mod_exp may produce incorrect results on x86_64
5861 There is a carry propagating bug in the x86_64 Montgomery squaring
5862 procedure. No EC algorithms are affected. Analysis suggests that attacks
5863 against RSA and DSA as a result of this defect would be very difficult to
5864 perform and are not believed likely. Attacks against DH are considered just
5865 feasible (although very difficult) because most of the work necessary to
5866 deduce information about a private key may be performed offline. The amount
5867 of resources required for such an attack would be very significant and
5868 likely only accessible to a limited number of attackers. An attacker would
5869 additionally need online access to an unpatched system using the target
5870 private key in a scenario with persistent DH parameters and a private
5871 key that is shared between multiple clients. For example this can occur by
5872 default in OpenSSL DHE based SSL/TLS ciphersuites.
5874 This issue was reported to OpenSSL by Hanno Böck.
5879 * Certificate verify crash with missing PSS parameter
5881 The signature verification routines will crash with a NULL pointer
5882 dereference if presented with an ASN.1 signature using the RSA PSS
5883 algorithm and absent mask generation function parameter. Since these
5884 routines are used to verify certificate signature algorithms this can be
5885 used to crash any certificate verification operation and exploited in a
5886 DoS attack. Any application which performs certificate verification is
5887 vulnerable including OpenSSL clients and servers which enable client
5890 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5895 * X509_ATTRIBUTE memory leak
5897 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5898 memory. This structure is used by the PKCS#7 and CMS routines so any
5899 application which reads PKCS#7 or CMS data from untrusted sources is
5900 affected. SSL/TLS is not affected.
5902 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5908 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5909 This changes the decoding behaviour for some invalid messages,
5910 though the change is mostly in the more lenient direction, and
5911 legacy behaviour is preserved as much as possible.
5915 * In DSA_generate_parameters_ex, if the provided seed is too short,
5918 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5920 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5922 * Alternate chains certificate forgery
5924 During certificate verification, OpenSSL will attempt to find an
5925 alternative certificate chain if the first attempt to build such a chain
5926 fails. An error in the implementation of this logic can mean that an
5927 attacker could cause certain checks on untrusted certificates to be
5928 bypassed, such as the CA flag, enabling them to use a valid leaf
5929 certificate to act as a CA and "issue" an invalid certificate.
5931 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5936 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5938 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5939 incompatibility in the handling of HMAC. The previous ABI has now been
5944 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5946 * Malformed ECParameters causes infinite loop
5948 When processing an ECParameters structure OpenSSL enters an infinite loop
5949 if the curve specified is over a specially malformed binary polynomial
5952 This can be used to perform denial of service against any
5953 system which processes public keys, certificate requests or
5954 certificates. This includes TLS clients and TLS servers with
5955 client authentication enabled.
5957 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5962 * Exploitable out-of-bounds read in X509_cmp_time
5964 X509_cmp_time does not properly check the length of the ASN1_TIME
5965 string and can read a few bytes out of bounds. In addition,
5966 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5969 An attacker can use this to craft malformed certificates and CRLs of
5970 various sizes and potentially cause a segmentation fault, resulting in
5971 a DoS on applications that verify certificates or CRLs. TLS clients
5972 that verify CRLs are affected. TLS clients and servers with client
5973 authentication enabled may be affected if they use custom verification
5976 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5977 independently by Hanno Böck.
5982 * PKCS7 crash with missing EnvelopedContent
5984 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5985 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5986 with missing content and trigger a NULL pointer dereference on parsing.
5988 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5989 structures from untrusted sources are affected. OpenSSL clients and
5990 servers are not affected.
5992 This issue was reported to OpenSSL by Michal Zalewski (Google).
5997 * CMS verify infinite loop with unknown hash function
5999 When verifying a signedData message the CMS code can enter an infinite loop
6000 if presented with an unknown hash function OID. This can be used to perform
6001 denial of service against any system which verifies signedData messages using
6003 This issue was reported to OpenSSL by Johannes Bauer.
6008 * Race condition handling NewSessionTicket
6010 If a NewSessionTicket is received by a multi-threaded client when attempting to
6011 reuse a previous ticket then a race condition can occur potentially leading to
6012 a double free of the ticket data.
6017 * Only support 256-bit or stronger elliptic curves with the
6018 'ecdh_auto' setting (server) or by default (client). Of supported
6019 curves, prefer P-256 (both).
6023 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6025 * ClientHello sigalgs DoS fix
6027 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6028 invalid signature algorithms extension a NULL pointer dereference will
6029 occur. This can be exploited in a DoS attack against the server.
6031 This issue was was reported to OpenSSL by David Ramos of Stanford
6035 *Stephen Henson and Matt Caswell*
6037 * Multiblock corrupted pointer fix
6039 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6040 feature only applies on 64 bit x86 architecture platforms that support AES
6041 NI instructions. A defect in the implementation of "multiblock" can cause
6042 OpenSSL's internal write buffer to become incorrectly set to NULL when
6043 using non-blocking IO. Typically, when the user application is using a
6044 socket BIO for writing, this will only result in a failed connection.
6045 However if some other BIO is used then it is likely that a segmentation
6046 fault will be triggered, thus enabling a potential DoS attack.
6048 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6053 * Segmentation fault in DTLSv1_listen fix
6055 The DTLSv1_listen function is intended to be stateless and processes the
6056 initial ClientHello from many peers. It is common for user code to loop
6057 over the call to DTLSv1_listen until a valid ClientHello is received with
6058 an associated cookie. A defect in the implementation of DTLSv1_listen means
6059 that state is preserved in the SSL object from one invocation to the next
6060 that can lead to a segmentation fault. Errors processing the initial
6061 ClientHello can trigger this scenario. An example of such an error could be
6062 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6065 This issue was reported to OpenSSL by Per Allansson.
6070 * Segmentation fault in ASN1_TYPE_cmp fix
6072 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6073 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6074 certificate signature algorithm consistency this can be used to crash any
6075 certificate verification operation and exploited in a DoS attack. Any
6076 application which performs certificate verification is vulnerable including
6077 OpenSSL clients and servers which enable client authentication.
6082 * Segmentation fault for invalid PSS parameters fix
6084 The signature verification routines will crash with a NULL pointer
6085 dereference if presented with an ASN.1 signature using the RSA PSS
6086 algorithm and invalid parameters. Since these routines are used to verify
6087 certificate signature algorithms this can be used to crash any
6088 certificate verification operation and exploited in a DoS attack. Any
6089 application which performs certificate verification is vulnerable including
6090 OpenSSL clients and servers which enable client authentication.
6092 This issue was was reported to OpenSSL by Brian Carpenter.
6097 * ASN.1 structure reuse memory corruption fix
6099 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6100 memory corruption via an invalid write. Such reuse is and has been
6101 strongly discouraged and is believed to be rare.
6103 Applications that parse structures containing CHOICE or ANY DEFINED BY
6104 components may be affected. Certificate parsing (d2i_X509 and related
6105 functions) are however not affected. OpenSSL clients and servers are
6111 * PKCS7 NULL pointer dereferences fix
6113 The PKCS#7 parsing code does not handle missing outer ContentInfo
6114 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6115 missing content and trigger a NULL pointer dereference on parsing.
6117 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6118 otherwise parse PKCS#7 structures from untrusted sources are
6119 affected. OpenSSL clients and servers are not affected.
6121 This issue was reported to OpenSSL by Michal Zalewski (Google).
6126 * DoS via reachable assert in SSLv2 servers fix
6128 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6129 servers that both support SSLv2 and enable export cipher suites by sending
6130 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6132 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6133 (OpenSSL development team).
6138 * Empty CKE with client auth and DHE fix
6140 If client auth is used then a server can seg fault in the event of a DHE
6141 ciphersuite being selected and a zero length ClientKeyExchange message
6142 being sent by the client. This could be exploited in a DoS attack.
6147 * Handshake with unseeded PRNG fix
6149 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6150 with an unseeded PRNG. The conditions are:
6151 - The client is on a platform where the PRNG has not been seeded
6152 automatically, and the user has not seeded manually
6153 - A protocol specific client method version has been used (i.e. not
6154 SSL_client_methodv23)
6155 - A ciphersuite is used that does not require additional random data from
6156 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6158 If the handshake succeeds then the client random that has been used will
6159 have been generated from a PRNG with insufficient entropy and therefore the
6160 output may be predictable.
6162 For example using the following command with an unseeded openssl will
6163 succeed on an unpatched platform:
6165 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6170 * Use After Free following d2i_ECPrivatekey error fix
6172 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6173 could cause a use after free condition. This, in turn, could cause a double
6174 free in several private key parsing functions (such as d2i_PrivateKey
6175 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6176 for applications that receive EC private keys from untrusted
6177 sources. This scenario is considered rare.
6179 This issue was discovered by the BoringSSL project and fixed in their
6185 * X509_to_X509_REQ NULL pointer deref fix
6187 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6188 the certificate key is invalid. This function is rarely used in practice.
6190 This issue was discovered by Brian Carpenter.
6195 * Removed the export ciphers from the DEFAULT ciphers
6199 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6201 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6202 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6203 So far those who have to target multiple platforms would compromise
6204 and argue that binary targeting say ARMv5 would still execute on
6205 ARMv8. "Universal" build resolves this compromise by providing
6206 near-optimal performance even on newer platforms.
6210 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6211 (other platforms pending).
6213 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6215 * Add support for the SignedCertificateTimestampList certificate and
6216 OCSP response extensions from RFC6962.
6220 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6221 for corner cases. (Certain input points at infinity could lead to
6222 bogus results, with non-infinity inputs mapped to infinity too.)
6226 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6227 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6228 common cases are optimized and there still is room for further
6229 improvements. Vector Permutation AES for Altivec is also added.
6233 * Add support for little-endian ppc64 Linux target.
6235 *Marcelo Cerri (IBM)*
6237 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6238 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6239 are optimized and there still is room for further improvements.
6240 Both 32- and 64-bit modes are supported.
6242 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6244 * Improved ARMv7 NEON support.
6248 * Support for SPARC Architecture 2011 crypto extensions, first
6249 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6250 SHA256/512, MD5, GHASH and modular exponentiation.
6252 *Andy Polyakov, David Miller*
6254 * Accelerated modular exponentiation for Intel processors, a.k.a.
6257 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6259 * Support for new and upcoming Intel processors, including AVX2,
6260 BMI and SHA ISA extensions. This includes additional "stitched"
6261 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6264 This work was sponsored by Intel Corp.
6268 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6269 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6270 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6274 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6275 this fixes a limitation in previous versions of OpenSSL.
6279 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6280 MGF1 digest and OAEP label.
6284 * Add EVP support for key wrapping algorithms, to avoid problems with
6285 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6286 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6287 algorithms and include tests cases.
6291 * Add functions to allocate and set the fields of an ECDSA_METHOD
6294 *Douglas E. Engert, Steve Henson*
6296 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6297 difference in days and seconds between two tm or ASN1_TIME structures.
6301 * Add -rev test option to s_server to just reverse order of characters
6302 received by client and send back to server. Also prints an abbreviated
6303 summary of the connection parameters.
6307 * New option -brief for s_client and s_server to print out a brief summary
6308 of connection parameters.
6312 * Add callbacks for arbitrary TLS extensions.
6314 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6316 * New option -crl_download in several openssl utilities to download CRLs
6317 from CRLDP extension in certificates.
6321 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6325 * New function X509_CRL_diff to generate a delta CRL from the difference
6326 of two full CRLs. Add support to "crl" utility.
6330 * New functions to set lookup_crls function and to retrieve
6331 X509_STORE from X509_STORE_CTX.
6335 * Print out deprecated issuer and subject unique ID fields in
6340 * Extend OCSP I/O functions so they can be used for simple general purpose
6341 HTTP as well as OCSP. New wrapper function which can be used to download
6342 CRLs using the OCSP API.
6346 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6350 * `SSL_CONF*` functions. These provide a common framework for application
6351 configuration using configuration files or command lines.
6355 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6356 message callback and prints the results. Needs compile time option
6357 "enable-ssl-trace". New options to s_client and s_server to enable
6362 * New ctrl and macro to retrieve supported points extensions.
6363 Print out extension in s_server and s_client.
6367 * New functions to retrieve certificate signature and signature
6372 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6377 * New Suite B modes for TLS code. These use and enforce the requirements
6378 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6379 only use Suite B curves. The Suite B modes can be set by using the
6380 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6384 * New chain verification flags for Suite B levels of security. Check
6385 algorithms are acceptable when flags are set in X509_verify_cert.
6389 * Make tls1_check_chain return a set of flags indicating checks passed
6390 by a certificate chain. Add additional tests to handle client
6391 certificates: checks for matching certificate type and issuer name
6396 * If an attempt is made to use a signature algorithm not in the peer
6397 preference list abort the handshake. If client has no suitable
6398 signature algorithms in response to a certificate request do not
6399 use the certificate.
6403 * If server EC tmp key is not in client preference list abort handshake.
6407 * Add support for certificate stores in CERT structure. This makes it
6408 possible to have different stores per SSL structure or one store in
6409 the parent SSL_CTX. Include distinct stores for certificate chain
6410 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6411 to build and store a certificate chain in CERT structure: returning
6412 an error if the chain cannot be built: this will allow applications
6413 to test if a chain is correctly configured.
6415 Note: if the CERT based stores are not set then the parent SSL_CTX
6416 store is used to retain compatibility with existing behaviour.
6420 * New function ssl_set_client_disabled to set a ciphersuite disabled
6421 mask based on the current session, check mask when sending client
6422 hello and checking the requested ciphersuite.
6426 * New ctrls to retrieve and set certificate types in a certificate
6427 request message. Print out received values in s_client. If certificate
6428 types is not set with custom values set sensible values based on
6429 supported signature algorithms.
6433 * Support for distinct client and server supported signature algorithms.
6437 * Add certificate callback. If set this is called whenever a certificate
6438 is required by client or server. An application can decide which
6439 certificate chain to present based on arbitrary criteria: for example
6440 supported signature algorithms. Add very simple example to s_server.
6441 This fixes many of the problems and restrictions of the existing client
6442 certificate callback: for example you can now clear an existing
6443 certificate and specify the whole chain.
6447 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6448 the certificate can be used for (if anything). Set valid_flags field
6449 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6450 to have similar checks in it.
6452 Add new "cert_flags" field to CERT structure and include a "strict mode".
6453 This enforces some TLS certificate requirements (such as only permitting
6454 certificate signature algorithms contained in the supported algorithms
6455 extension) which some implementations ignore: this option should be used
6456 with caution as it could cause interoperability issues.
6460 * Update and tidy signature algorithm extension processing. Work out
6461 shared signature algorithms based on preferences and peer algorithms
6462 and print them out in s_client and s_server. Abort handshake if no
6463 shared signature algorithms.
6467 * Add new functions to allow customised supported signature algorithms
6468 for SSL and SSL_CTX structures. Add options to s_client and s_server
6473 * New function SSL_certs_clear() to delete all references to certificates
6474 from an SSL structure. Before this once a certificate had been added
6475 it couldn't be removed.
6479 * Integrate hostname, email address and IP address checking with certificate
6480 verification. New verify options supporting checking in openssl utility.
6484 * Fixes and wildcard matching support to hostname and email checking
6485 functions. Add manual page.
6487 *Florian Weimer (Red Hat Product Security Team)*
6489 * New functions to check a hostname email or IP address against a
6490 certificate. Add options x509 utility to print results of checks against
6495 * Fix OCSP checking.
6497 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6499 * Initial experimental support for explicitly trusted non-root CAs.
6500 OpenSSL still tries to build a complete chain to a root but if an
6501 intermediate CA has a trust setting included that is used. The first
6502 setting is used: whether to trust (e.g., -addtrust option to the x509
6507 * Add -trusted_first option which attempts to find certificates in the
6508 trusted store even if an untrusted chain is also supplied.
6512 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6513 platform support for Linux and Android.
6517 * Support for linux-x32, ILP32 environment in x86_64 framework.
6521 * Experimental multi-implementation support for FIPS capable OpenSSL.
6522 When in FIPS mode the approved implementations are used as normal,
6523 when not in FIPS mode the internal unapproved versions are used instead.
6524 This means that the FIPS capable OpenSSL isn't forced to use the
6525 (often lower performance) FIPS implementations outside FIPS mode.
6529 * Transparently support X9.42 DH parameters when calling
6530 PEM_read_bio_DHparameters. This means existing applications can handle
6531 the new parameter format automatically.
6535 * Initial experimental support for X9.42 DH parameter format: mainly
6536 to support use of 'q' parameter for RFC5114 parameters.
6540 * Add DH parameters from RFC5114 including test data to dhtest.
6544 * Support for automatic EC temporary key parameter selection. If enabled
6545 the most preferred EC parameters are automatically used instead of
6546 hardcoded fixed parameters. Now a server just has to call:
6547 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6548 support ECDH and use the most appropriate parameters.
6552 * Enhance and tidy EC curve and point format TLS extension code. Use
6553 static structures instead of allocation if default values are used.
6554 New ctrls to set curves we wish to support and to retrieve shared curves.
6555 Print out shared curves in s_server. New options to s_server and s_client
6556 to set list of supported curves.
6560 * New ctrls to retrieve supported signature algorithms and
6561 supported curve values as an array of NIDs. Extend openssl utility
6562 to print out received values.
6566 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6567 between NIDs and the more common NIST names such as "P-256". Enhance
6568 ecparam utility and ECC method to recognise the NIST names for curves.
6572 * Enhance SSL/TLS certificate chain handling to support different
6573 chains for each certificate instead of one chain in the parent SSL_CTX.
6577 * Support for fixed DH ciphersuite client authentication: where both
6578 server and client use DH certificates with common parameters.
6582 * Support for fixed DH ciphersuites: those requiring DH server
6587 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6589 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6590 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6591 X509_CINF_get_signature were reverted post internal team review.
6596 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6598 * OCSP Status Request extension unbounded memory growth
6600 A malicious client can send an excessively large OCSP Status Request
6601 extension. If that client continually requests renegotiation, sending a
6602 large OCSP Status Request extension each time, then there will be unbounded
6603 memory growth on the server. This will eventually lead to a Denial Of
6604 Service attack through memory exhaustion. Servers with a default
6605 configuration are vulnerable even if they do not support OCSP. Builds using
6606 the "no-ocsp" build time option are not affected.
6608 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6613 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6616 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6622 * OOB write in MDC2_Update()
6624 An overflow can occur in MDC2_Update() either if called directly or
6625 through the EVP_DigestUpdate() function using MDC2. If an attacker
6626 is able to supply very large amounts of input data after a previous
6627 call to EVP_EncryptUpdate() with a partial block then a length check
6628 can overflow resulting in a heap corruption.
6630 The amount of data needed is comparable to SIZE_MAX which is impractical
6633 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6638 * Malformed SHA512 ticket DoS
6640 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6641 DoS attack where a malformed ticket will result in an OOB read which will
6644 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6645 a custom server callback and ticket lookup mechanism.
6647 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6652 * OOB write in BN_bn2dec()
6654 The function BN_bn2dec() does not check the return value of BN_div_word().
6655 This can cause an OOB write if an application uses this function with an
6656 overly large BIGNUM. This could be a problem if an overly large certificate
6657 or CRL is printed out from an untrusted source. TLS is not affected because
6658 record limits will reject an oversized certificate before it is parsed.
6660 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6665 * OOB read in TS_OBJ_print_bio()
6667 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6668 the total length the OID text representation would use and not the amount
6669 of data written. This will result in OOB reads when large OIDs are
6672 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6677 * Pointer arithmetic undefined behaviour
6679 Avoid some undefined pointer arithmetic
6681 A common idiom in the codebase is to check limits in the following manner:
6684 Where "p" points to some malloc'd data of SIZE bytes and
6687 "len" here could be from some externally supplied data (e.g. from a TLS
6690 The rules of C pointer arithmetic are such that "p + len" is only well
6691 defined where len <= SIZE. Therefore the above idiom is actually
6692 undefined behaviour.
6694 For example this could cause problems if some malloc implementation
6695 provides an address for "p" such that "p + len" actually overflows for
6696 values of len that are too big and therefore p + len < limit.
6698 This issue was reported to OpenSSL by Guido Vranken
6703 * Constant time flag not preserved in DSA signing
6705 Operations in the DSA signing algorithm should run in constant time in
6706 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6707 implementation means that a non-constant time codepath is followed for
6708 certain operations. This has been demonstrated through a cache-timing
6709 attack to be sufficient for an attacker to recover the private DSA key.
6711 This issue was reported by César Pereida (Aalto University), Billy Brumley
6712 (Tampere University of Technology), and Yuval Yarom (The University of
6713 Adelaide and NICTA).
6718 * DTLS buffered message DoS
6720 In a DTLS connection where handshake messages are delivered out-of-order
6721 those messages that OpenSSL is not yet ready to process will be buffered
6722 for later use. Under certain circumstances, a flaw in the logic means that
6723 those messages do not get removed from the buffer even though the handshake
6724 has been completed. An attacker could force up to approx. 15 messages to
6725 remain in the buffer when they are no longer required. These messages will
6726 be cleared when the DTLS connection is closed. The default maximum size for
6727 a message is 100k. Therefore the attacker could force an additional 1500k
6728 to be consumed per connection. By opening many simulataneous connections an
6729 attacker could cause a DoS attack through memory exhaustion.
6731 This issue was reported to OpenSSL by Quan Luo.
6736 * DTLS replay protection DoS
6738 A flaw in the DTLS replay attack protection mechanism means that records
6739 that arrive for future epochs update the replay protection "window" before
6740 the MAC for the record has been validated. This could be exploited by an
6741 attacker by sending a record for the next epoch (which does not have to
6742 decrypt or have a valid MAC), with a very large sequence number. This means
6743 that all subsequent legitimate packets are dropped causing a denial of
6744 service for a specific DTLS connection.
6746 This issue was reported to OpenSSL by the OCAP audit team.
6751 * Certificate message OOB reads
6753 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6754 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6755 theoretical DoS risk but this has not been observed in practice on common
6758 The messages affected are client certificate, client certificate request
6759 and server certificate. As a result the attack can only be performed
6760 against a client or a server which enables client authentication.
6762 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6767 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6769 * Prevent padding oracle in AES-NI CBC MAC check
6771 A MITM attacker can use a padding oracle attack to decrypt traffic
6772 when the connection uses an AES CBC cipher and the server support
6775 This issue was introduced as part of the fix for Lucky 13 padding
6776 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6777 constant time by making sure that always the same bytes are read and
6778 compared against either the MAC or padding bytes. But it no longer
6779 checked that there was enough data to have both the MAC and padding
6782 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6787 * Fix EVP_EncodeUpdate overflow
6789 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6790 Base64 encoding of binary data. If an attacker is able to supply very large
6791 amounts of input data then a length check can overflow resulting in a heap
6794 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6795 the `PEM_write_bio*` family of functions. These are mainly used within the
6796 OpenSSL command line applications, so any application which processes data
6797 from an untrusted source and outputs it as a PEM file should be considered
6798 vulnerable to this issue. User applications that call these APIs directly
6799 with large amounts of untrusted data may also be vulnerable.
6801 This issue was reported by Guido Vranken.
6806 * Fix EVP_EncryptUpdate overflow
6808 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6809 is able to supply very large amounts of input data after a previous call to
6810 EVP_EncryptUpdate() with a partial block then a length check can overflow
6811 resulting in a heap corruption. Following an analysis of all OpenSSL
6812 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6813 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6814 the first called function after an EVP_EncryptInit(), and therefore that
6815 specific call must be safe. The second form is where the length passed to
6816 EVP_EncryptUpdate() can be seen from the code to be some small value and
6817 therefore there is no possibility of an overflow. Since all instances are
6818 one of these two forms, it is believed that there can be no overflows in
6819 internal code due to this problem. It should be noted that
6820 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6821 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6822 of these calls have also been analysed too and it is believed there are no
6823 instances in internal usage where an overflow could occur.
6825 This issue was reported by Guido Vranken.
6830 * Prevent ASN.1 BIO excessive memory allocation
6832 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6833 a short invalid encoding can casuse allocation of large amounts of memory
6834 potentially consuming excessive resources or exhausting memory.
6836 Any application parsing untrusted data through d2i BIO functions is
6837 affected. The memory based functions such as d2i_X509() are *not* affected.
6838 Since the memory based functions are used by the TLS library, TLS
6839 applications are not affected.
6841 This issue was reported by Brian Carpenter.
6848 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6849 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6850 in arbitrary stack data being returned in the buffer.
6852 This issue was reported by Guido Vranken.
6857 * Modify behavior of ALPN to invoke callback after SNI/servername
6858 callback, such that updates to the SSL_CTX affect ALPN.
6862 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6867 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6868 methods are enabled and ssl2 is disabled the methods return NULL.
6872 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6874 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6875 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6876 provide any "EXPORT" or "LOW" strength ciphers.
6880 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6881 is by default disabled at build-time. Builds that are not configured with
6882 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6883 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6884 will need to explicitly call either of:
6886 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6888 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6890 as appropriate. Even if either of those is used, or the application
6891 explicitly uses the version-specific SSLv2_method() or its client and
6892 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6893 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6894 ciphers, and SSLv2 56-bit DES are no longer available.
6899 * Fix a double-free in DSA code
6901 A double free bug was discovered when OpenSSL parses malformed DSA private
6902 keys and could lead to a DoS attack or memory corruption for applications
6903 that receive DSA private keys from untrusted sources. This scenario is
6906 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6912 * Disable SRP fake user seed to address a server memory leak.
6914 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6916 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6917 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6918 was changed to ignore the "fake user" SRP seed, even if the seed
6921 Users should use SRP_VBASE_get1_by_user instead. Note that in
6922 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6923 also that even though configuring the SRP seed attempts to hide
6924 invalid usernames by continuing the handshake with fake
6925 credentials, this behaviour is not constant time and no strong
6926 guarantees are made that the handshake is indistinguishable from
6927 that of a valid user.
6932 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6934 In the BN_hex2bn function the number of hex digits is calculated using an
6935 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6936 large values of `i` this can result in `bn_expand` not allocating any
6937 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6938 field as NULL leading to a subsequent NULL ptr deref. For very large values
6939 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6940 In this case memory is allocated to the internal BIGNUM data field, but it
6941 is insufficiently sized leading to heap corruption. A similar issue exists
6942 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6943 is ever called by user applications with very large untrusted hex/dec data.
6944 This is anticipated to be a rare occurrence.
6946 All OpenSSL internal usage of these functions use data that is not expected
6947 to be untrusted, e.g. config file data or application command line
6948 arguments. If user developed applications generate config file data based
6949 on untrusted data then it is possible that this could also lead to security
6950 consequences. This is also anticipated to be rare.
6952 This issue was reported to OpenSSL by Guido Vranken.
6957 * Fix memory issues in `BIO_*printf` functions
6959 The internal `fmtstr` function used in processing a "%s" format string in
6960 the `BIO_*printf` functions could overflow while calculating the length of a
6961 string and cause an OOB read when printing very long strings.
6963 Additionally the internal `doapr_outch` function can attempt to write to an
6964 OOB memory location (at an offset from the NULL pointer) in the event of a
6965 memory allocation failure. In 1.0.2 and below this could be caused where
6966 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6967 could be in processing a very long "%s" format string. Memory leaks can
6970 The first issue may mask the second issue dependent on compiler behaviour.
6971 These problems could enable attacks where large amounts of untrusted data
6972 is passed to the `BIO_*printf` functions. If applications use these functions
6973 in this way then they could be vulnerable. OpenSSL itself uses these
6974 functions when printing out human-readable dumps of ASN.1 data. Therefore
6975 applications that print this data could be vulnerable if the data is from
6976 untrusted sources. OpenSSL command line applications could also be
6977 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6978 as command line arguments.
6980 Libssl is not considered directly vulnerable. Additionally certificates etc
6981 received via remote connections via libssl are also unlikely to be able to
6982 trigger these issues because of message size limits enforced within libssl.
6984 This issue was reported to OpenSSL Guido Vranken.
6989 * Side channel attack on modular exponentiation
6991 A side-channel attack was found which makes use of cache-bank conflicts on
6992 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6993 of RSA keys. The ability to exploit this issue is limited as it relies on
6994 an attacker who has control of code in a thread running on the same
6995 hyper-threaded core as the victim thread which is performing decryptions.
6997 This issue was reported to OpenSSL by Yuval Yarom, The University of
6998 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6999 Nadia Heninger, University of Pennsylvania with more information at
7000 <http://cachebleed.info>.
7005 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7006 if no keysize is specified with default_bits. This fixes an
7007 omission in an earlier change that changed all RSA/DSA key generation
7008 commands to use 2048 bits by default.
7012 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7014 * Protection for DH small subgroup attacks
7016 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7017 switched on by default and cannot be disabled. This could have some
7022 * SSLv2 doesn't block disabled ciphers
7024 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7025 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7026 been disabled, provided that the SSLv2 protocol was not also disabled via
7029 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7030 and Sebastian Schinzel.
7035 * Reject DH handshakes with parameters shorter than 1024 bits.
7039 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7041 * Certificate verify crash with missing PSS parameter
7043 The signature verification routines will crash with a NULL pointer
7044 dereference if presented with an ASN.1 signature using the RSA PSS
7045 algorithm and absent mask generation function parameter. Since these
7046 routines are used to verify certificate signature algorithms this can be
7047 used to crash any certificate verification operation and exploited in a
7048 DoS attack. Any application which performs certificate verification is
7049 vulnerable including OpenSSL clients and servers which enable client
7052 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7057 * X509_ATTRIBUTE memory leak
7059 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7060 memory. This structure is used by the PKCS#7 and CMS routines so any
7061 application which reads PKCS#7 or CMS data from untrusted sources is
7062 affected. SSL/TLS is not affected.
7064 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7070 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7071 This changes the decoding behaviour for some invalid messages,
7072 though the change is mostly in the more lenient direction, and
7073 legacy behaviour is preserved as much as possible.
7077 * In DSA_generate_parameters_ex, if the provided seed is too short,
7078 use a random seed, as already documented.
7080 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7082 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7084 * Alternate chains certificate forgery
7086 During certificate verfification, OpenSSL will attempt to find an
7087 alternative certificate chain if the first attempt to build such a chain
7088 fails. An error in the implementation of this logic can mean that an
7089 attacker could cause certain checks on untrusted certificates to be
7090 bypassed, such as the CA flag, enabling them to use a valid leaf
7091 certificate to act as a CA and "issue" an invalid certificate.
7093 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7099 * Race condition handling PSK identify hint
7101 If PSK identity hints are received by a multi-threaded client then
7102 the values are wrongly updated in the parent SSL_CTX structure. This can
7103 result in a race condition potentially leading to a double free of the
7109 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7111 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7112 incompatibility in the handling of HMAC. The previous ABI has now been
7115 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7117 * Malformed ECParameters causes infinite loop
7119 When processing an ECParameters structure OpenSSL enters an infinite loop
7120 if the curve specified is over a specially malformed binary polynomial
7123 This can be used to perform denial of service against any
7124 system which processes public keys, certificate requests or
7125 certificates. This includes TLS clients and TLS servers with
7126 client authentication enabled.
7128 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7133 * Exploitable out-of-bounds read in X509_cmp_time
7135 X509_cmp_time does not properly check the length of the ASN1_TIME
7136 string and can read a few bytes out of bounds. In addition,
7137 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7140 An attacker can use this to craft malformed certificates and CRLs of
7141 various sizes and potentially cause a segmentation fault, resulting in
7142 a DoS on applications that verify certificates or CRLs. TLS clients
7143 that verify CRLs are affected. TLS clients and servers with client
7144 authentication enabled may be affected if they use custom verification
7147 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7148 independently by Hanno Böck.
7153 * PKCS7 crash with missing EnvelopedContent
7155 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7156 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7157 with missing content and trigger a NULL pointer dereference on parsing.
7159 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7160 structures from untrusted sources are affected. OpenSSL clients and
7161 servers are not affected.
7163 This issue was reported to OpenSSL by Michal Zalewski (Google).
7168 * CMS verify infinite loop with unknown hash function
7170 When verifying a signedData message the CMS code can enter an infinite loop
7171 if presented with an unknown hash function OID. This can be used to perform
7172 denial of service against any system which verifies signedData messages using
7174 This issue was reported to OpenSSL by Johannes Bauer.
7179 * Race condition handling NewSessionTicket
7181 If a NewSessionTicket is received by a multi-threaded client when attempting to
7182 reuse a previous ticket then a race condition can occur potentially leading to
7183 a double free of the ticket data.
7188 * Reject DH handshakes with parameters shorter than 768 bits.
7190 *Kurt Roeckx and Emilia Kasper*
7192 * dhparam: generate 2048-bit parameters by default.
7194 *Kurt Roeckx and Emilia Kasper*
7196 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7198 * Segmentation fault in ASN1_TYPE_cmp fix
7200 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7201 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7202 certificate signature algorithm consistency this can be used to crash any
7203 certificate verification operation and exploited in a DoS attack. Any
7204 application which performs certificate verification is vulnerable including
7205 OpenSSL clients and servers which enable client authentication.
7210 * ASN.1 structure reuse memory corruption fix
7212 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7213 memory corruption via an invalid write. Such reuse is and has been
7214 strongly discouraged and is believed to be rare.
7216 Applications that parse structures containing CHOICE or ANY DEFINED BY
7217 components may be affected. Certificate parsing (d2i_X509 and related
7218 functions) are however not affected. OpenSSL clients and servers are
7224 * PKCS7 NULL pointer dereferences fix
7226 The PKCS#7 parsing code does not handle missing outer ContentInfo
7227 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7228 missing content and trigger a NULL pointer dereference on parsing.
7230 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7231 otherwise parse PKCS#7 structures from untrusted sources are
7232 affected. OpenSSL clients and servers are not affected.
7234 This issue was reported to OpenSSL by Michal Zalewski (Google).
7239 * DoS via reachable assert in SSLv2 servers fix
7241 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7242 servers that both support SSLv2 and enable export cipher suites by sending
7243 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7245 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7246 (OpenSSL development team).
7251 * Use After Free following d2i_ECPrivatekey error fix
7253 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7254 could cause a use after free condition. This, in turn, could cause a double
7255 free in several private key parsing functions (such as d2i_PrivateKey
7256 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7257 for applications that receive EC private keys from untrusted
7258 sources. This scenario is considered rare.
7260 This issue was discovered by the BoringSSL project and fixed in their
7266 * X509_to_X509_REQ NULL pointer deref fix
7268 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7269 the certificate key is invalid. This function is rarely used in practice.
7271 This issue was discovered by Brian Carpenter.
7276 * Removed the export ciphers from the DEFAULT ciphers
7280 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7282 * Build fixes for the Windows and OpenVMS platforms
7284 *Matt Caswell and Richard Levitte*
7286 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7288 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7289 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7290 dereference. This could lead to a Denial Of Service attack. Thanks to
7291 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7296 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7297 dtls1_buffer_record function under certain conditions. In particular this
7298 could occur if an attacker sent repeated DTLS records with the same
7299 sequence number but for the next epoch. The memory leak could be exploited
7300 by an attacker in a Denial of Service attack through memory exhaustion.
7301 Thanks to Chris Mueller for reporting this issue.
7306 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7307 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7308 method would be set to NULL which could later result in a NULL pointer
7309 dereference. Thanks to Frank Schmirler for reporting this issue.
7314 * Abort handshake if server key exchange message is omitted for ephemeral
7317 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7318 reporting this issue.
7323 * Remove non-export ephemeral RSA code on client and server. This code
7324 violated the TLS standard by allowing the use of temporary RSA keys in
7325 non-export ciphersuites and could be used by a server to effectively
7326 downgrade the RSA key length used to a value smaller than the server
7327 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7328 INRIA or reporting this issue.
7333 * Fixed issue where DH client certificates are accepted without verification.
7334 An OpenSSL server will accept a DH certificate for client authentication
7335 without the certificate verify message. This effectively allows a client to
7336 authenticate without the use of a private key. This only affects servers
7337 which trust a client certificate authority which issues certificates
7338 containing DH keys: these are extremely rare and hardly ever encountered.
7339 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7345 * Ensure that the session ID context of an SSL is updated when its
7346 SSL_CTX is updated via SSL_set_SSL_CTX.
7348 The session ID context is typically set from the parent SSL_CTX,
7349 and can vary with the CTX.
7353 * Fix various certificate fingerprint issues.
7355 By using non-DER or invalid encodings outside the signed portion of a
7356 certificate the fingerprint can be changed without breaking the signature.
7357 Although no details of the signed portion of the certificate can be changed
7358 this can cause problems with some applications: e.g. those using the
7359 certificate fingerprint for blacklists.
7361 1. Reject signatures with non zero unused bits.
7363 If the BIT STRING containing the signature has non zero unused bits reject
7364 the signature. All current signature algorithms require zero unused bits.
7366 2. Check certificate algorithm consistency.
7368 Check the AlgorithmIdentifier inside TBS matches the one in the
7369 certificate signature. NB: this will result in signature failure
7370 errors for some broken certificates.
7372 Thanks to Konrad Kraszewski from Google for reporting this issue.
7374 3. Check DSA/ECDSA signatures use DER.
7376 Re-encode DSA/ECDSA signatures and compare with the original received
7377 signature. Return an error if there is a mismatch.
7379 This will reject various cases including garbage after signature
7380 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7381 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7382 (negative or with leading zeroes).
7384 Further analysis was conducted and fixes were developed by Stephen Henson
7385 of the OpenSSL core team.
7391 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7392 results on some platforms, including x86_64. This bug occurs at random
7393 with a very low probability, and is not known to be exploitable in any
7394 way, though its exact impact is difficult to determine. Thanks to Pieter
7395 Wuille (Blockstream) who reported this issue and also suggested an initial
7396 fix. Further analysis was conducted by the OpenSSL development team and
7397 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7398 the OpenSSL core team.
7403 * Do not resume sessions on the server if the negotiated protocol
7404 version does not match the session's version. Resuming with a different
7405 version, while not strictly forbidden by the RFC, is of questionable
7406 sanity and breaks all known clients.
7408 *David Benjamin, Emilia Käsper*
7410 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7411 early CCS messages during renegotiation. (Note that because
7412 renegotiation is encrypted, this early CCS was not exploitable.)
7416 * Tighten client-side session ticket handling during renegotiation:
7417 ensure that the client only accepts a session ticket if the server sends
7418 the extension anew in the ServerHello. Previously, a TLS client would
7419 reuse the old extension state and thus accept a session ticket if one was
7420 announced in the initial ServerHello.
7422 Similarly, ensure that the client requires a session ticket if one
7423 was advertised in the ServerHello. Previously, a TLS client would
7424 ignore a missing NewSessionTicket message.
7428 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7432 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7433 sends a carefully crafted handshake message, to cause OpenSSL to fail
7434 to free up to 64k of memory causing a memory leak. This could be
7435 exploited in a Denial Of Service attack. This issue affects OpenSSL
7436 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7437 whether SRTP is used or configured. Implementations of OpenSSL that
7438 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7440 The fix was developed by the OpenSSL team.
7445 * Session Ticket Memory Leak.
7447 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7448 integrity of that ticket is first verified. In the event of a session
7449 ticket integrity check failing, OpenSSL will fail to free memory
7450 causing a memory leak. By sending a large number of invalid session
7451 tickets an attacker could exploit this issue in a Denial Of Service
7457 * Build option no-ssl3 is incomplete.
7459 When OpenSSL is configured with "no-ssl3" as a build option, servers
7460 could accept and complete a SSL 3.0 handshake, and clients could be
7461 configured to send them.
7464 *Akamai and the OpenSSL team*
7466 * Add support for TLS_FALLBACK_SCSV.
7467 Client applications doing fallback retries should call
7468 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7471 *Adam Langley, Bodo Moeller*
7473 * Add additional DigestInfo checks.
7475 Re-encode DigestInto in DER and check against the original when
7476 verifying RSA signature: this will reject any improperly encoded
7477 DigestInfo structures.
7479 Note: this is a precautionary measure and no attacks are currently known.
7483 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7485 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7486 SRP code can be overrun an internal buffer. Add sanity check that
7487 g, A, B < N to SRP code.
7489 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7490 Group for discovering this issue.
7495 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7496 TLS 1.0 instead of higher protocol versions when the ClientHello message
7497 is badly fragmented. This allows a man-in-the-middle attacker to force a
7498 downgrade to TLS 1.0 even if both the server and the client support a
7499 higher protocol version, by modifying the client's TLS records.
7501 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7502 researching this issue.
7507 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7508 to a denial of service attack. A malicious server can crash the client
7509 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7510 ciphersuite and sending carefully crafted handshake messages.
7512 Thanks to Felix Gröbert (Google) for discovering and researching this
7518 * By sending carefully crafted DTLS packets an attacker could cause openssl
7519 to leak memory. This can be exploited through a Denial of Service attack.
7520 Thanks to Adam Langley for discovering and researching this issue.
7525 * An attacker can force openssl to consume large amounts of memory whilst
7526 processing DTLS handshake messages. This can be exploited through a
7527 Denial of Service attack.
7528 Thanks to Adam Langley for discovering and researching this issue.
7533 * An attacker can force an error condition which causes openssl to crash
7534 whilst processing DTLS packets due to memory being freed twice. This
7535 can be exploited through a Denial of Service attack.
7536 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7542 * If a multithreaded client connects to a malicious server using a resumed
7543 session and the server sends an ec point format extension it could write
7544 up to 255 bytes to freed memory.
7546 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7552 * A malicious server can crash an OpenSSL client with a null pointer
7553 dereference (read) by specifying an SRP ciphersuite even though it was not
7554 properly negotiated with the client. This can be exploited through a
7555 Denial of Service attack.
7557 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7558 discovering and researching this issue.
7563 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7564 X509_name_oneline, X509_name_print_ex et al. to leak some information
7565 from the stack. Applications may be affected if they echo pretty printing
7566 output to the attacker.
7568 Thanks to Ivan Fratric (Google) for discovering this issue.
7571 *Emilia Käsper, and Steve Henson*
7573 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7574 for corner cases. (Certain input points at infinity could lead to
7575 bogus results, with non-infinity inputs mapped to infinity too.)
7579 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7581 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7582 handshake can force the use of weak keying material in OpenSSL
7583 SSL/TLS clients and servers.
7585 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7586 researching this issue. ([CVE-2014-0224])
7588 *KIKUCHI Masashi, Steve Henson*
7590 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7591 OpenSSL DTLS client the code can be made to recurse eventually crashing
7594 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7597 *Imre Rad, Steve Henson*
7599 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7600 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7601 client or server. This is potentially exploitable to run arbitrary
7602 code on a vulnerable client or server.
7604 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7606 *Jüri Aedla, Steve Henson*
7608 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7609 are subject to a denial of service attack.
7611 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7612 this issue. ([CVE-2014-3470])
7614 *Felix Gröbert, Ivan Fratric, Steve Henson*
7616 * Harmonize version and its documentation. -f flag is used to display
7619 *mancha <mancha1@zoho.com>*
7621 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7622 in i2d_ECPrivateKey.
7624 *mancha <mancha1@zoho.com>*
7626 * Fix some double frees. These are not thought to be exploitable.
7628 *mancha <mancha1@zoho.com>*
7630 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7632 * A missing bounds check in the handling of the TLS heartbeat extension
7633 can be used to reveal up to 64k of memory to a connected client or
7636 Thanks for Neel Mehta of Google Security for discovering this bug and to
7637 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7638 preparing the fix ([CVE-2014-0160])
7640 *Adam Langley, Bodo Moeller*
7642 * Fix for the attack described in the paper "Recovering OpenSSL
7643 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7644 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7645 <http://eprint.iacr.org/2014/140>
7647 Thanks to Yuval Yarom and Naomi Benger for discovering this
7648 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7650 *Yuval Yarom and Naomi Benger*
7652 * TLS pad extension: draft-agl-tls-padding-03
7654 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7655 TLS client Hello record length value would otherwise be > 255 and
7656 less that 512 pad with a dummy extension containing zeroes so it
7657 is at least 512 bytes long.
7659 *Adam Langley, Steve Henson*
7661 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7663 * Fix for TLS record tampering bug. A carefully crafted invalid
7664 handshake could crash OpenSSL with a NULL pointer exception.
7665 Thanks to Anton Johansson for reporting this issues.
7668 * Keep original DTLS digest and encryption contexts in retransmission
7669 structures so we can use the previous session parameters if they need
7670 to be resent. ([CVE-2013-6450])
7674 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7675 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7676 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7677 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7678 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7679 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7681 *Rob Stradling, Adam Langley*
7683 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7685 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7686 supporting platforms or when small records were transferred.
7688 *Andy Polyakov, Steve Henson*
7690 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7692 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7694 This addresses the flaw in CBC record processing discovered by
7695 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7696 at: <http://www.isg.rhul.ac.uk/tls/>
7698 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7699 Security Group at Royal Holloway, University of London
7700 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7701 Emilia Käsper for the initial patch.
7704 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7706 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7707 ciphersuites which can be exploited in a denial of service attack.
7708 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7709 and detecting this bug and to Wolfgang Ettlinger
7710 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7715 * Return an error when checking OCSP signatures when key is NULL.
7716 This fixes a DoS attack. ([CVE-2013-0166])
7720 * Make openssl verify return errors.
7722 *Chris Palmer <palmer@google.com> and Ben Laurie*
7724 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7725 the right response is stapled. Also change SSL_get_certificate()
7726 so it returns the certificate actually sent.
7727 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7729 *Rob Stradling <rob.stradling@comodo.com>*
7731 * Fix possible deadlock when decoding public keys.
7735 * Don't use TLS 1.0 record version number in initial client hello
7740 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7742 * Sanity check record length before skipping explicit IV in TLS
7743 1.2, 1.1 and DTLS to fix DoS attack.
7745 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7746 fuzzing as a service testing platform.
7751 * Initialise tkeylen properly when encrypting CMS messages.
7752 Thanks to Solar Designer of Openwall for reporting this issue.
7756 * In FIPS mode don't try to use composite ciphers as they are not
7761 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7763 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7764 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7765 mean any application compiled against OpenSSL 1.0.0 headers setting
7766 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7767 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7768 0x10000000L Any application which was previously compiled against
7769 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7770 will need to be recompiled as a result. Letting be results in
7771 inability to disable specifically TLS 1.1 and in client context,
7772 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7776 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7777 disable just protocol X, but all protocols above X *if* there are
7778 protocols *below* X still enabled. In more practical terms it means
7779 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7780 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7781 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7786 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7788 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7789 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7790 in CRYPTO_realloc_clean.
7792 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7793 issue and to Adam Langley <agl@chromium.org> for fixing it.
7796 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7798 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7802 * Workarounds for some broken servers that "hang" if a client hello
7803 record length exceeds 255 bytes.
7805 1. Do not use record version number > TLS 1.0 in initial client
7806 hello: some (but not all) hanging servers will now work.
7807 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7808 the number of ciphers sent in the client hello. This should be
7809 set to an even number, such as 50, for example by passing:
7810 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7811 Most broken servers should now work.
7812 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7813 TLS 1.2 client support entirely.
7817 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7821 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7823 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7824 STRING form instead of a DigestInfo.
7828 * The format used for MDC2 RSA signatures is inconsistent between EVP
7829 and the RSA_sign/RSA_verify functions. This was made more apparent when
7830 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7831 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7832 the correct format in RSA_verify so both forms transparently work.
7836 * Some servers which support TLS 1.0 can choke if we initially indicate
7837 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7838 encrypted premaster secret. As a workaround use the maximum permitted
7839 client version in client hello, this should keep such servers happy
7840 and still work with previous versions of OpenSSL.
7844 * Add support for TLS/DTLS heartbeats.
7846 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7848 * Add support for SCTP.
7850 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7852 * Improved PRNG seeding for VOS.
7854 *Paul Green <Paul.Green@stratus.com>*
7856 * Extensive assembler packs updates, most notably:
7858 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7859 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7860 - x86_64: bit-sliced AES implementation;
7861 - ARM: NEON support, contemporary platforms optimizations;
7862 - s390x: z196 support;
7863 - `*`: GHASH and GF(2^m) multiplication implementations;
7867 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7868 (removal of unnecessary code)
7870 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7872 * Add TLS key material exporter from RFC 5705.
7876 * Add DTLS-SRTP negotiation from RFC 5764.
7880 * Add Next Protocol Negotiation,
7881 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7882 disabled with a no-npn flag to config or Configure. Code donated
7885 *Adam Langley <agl@google.com> and Ben Laurie*
7887 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7888 NIST-P256, NIST-P521, with constant-time single point multiplication on
7889 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7890 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7891 Code made available under Apache License version 2.0.
7893 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7894 line to include this in your build of OpenSSL, and run "make depend" (or
7895 "make update"). This enables the following EC_METHODs:
7897 EC_GFp_nistp224_method()
7898 EC_GFp_nistp256_method()
7899 EC_GFp_nistp521_method()
7901 EC_GROUP_new_by_curve_name() will automatically use these (while
7902 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7905 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7907 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7908 all platforms. Move ssize_t definition from e_os.h to the public
7909 header file e_os2.h as it now appears in public header file cms.h
7913 * New -sigopt option to the ca, req and x509 utilities. Additional
7914 signature parameters can be passed using this option and in
7919 * Add RSA PSS signing function. This will generate and set the
7920 appropriate AlgorithmIdentifiers for PSS based on those in the
7921 corresponding EVP_MD_CTX structure. No application support yet.
7925 * Support for companion algorithm specific ASN1 signing routines.
7926 New function ASN1_item_sign_ctx() signs a pre-initialised
7927 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7928 the appropriate parameters.
7932 * Add new algorithm specific ASN1 verification initialisation function
7933 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7934 handling will be the same no matter what EVP_PKEY_METHOD is used.
7935 Add a PSS handler to support verification of PSS signatures: checked
7936 against a number of sample certificates.
7940 * Add signature printing for PSS. Add PSS OIDs.
7942 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7944 * Add algorithm specific signature printing. An individual ASN1 method
7945 can now print out signatures instead of the standard hex dump.
7947 More complex signatures (e.g. PSS) can print out more meaningful
7948 information. Include DSA version that prints out the signature
7953 * Password based recipient info support for CMS library: implementing
7958 * Split password based encryption into PBES2 and PBKDF2 functions. This
7959 neatly separates the code into cipher and PBE sections and is required
7960 for some algorithms that split PBES2 into separate pieces (such as
7961 password based CMS).
7965 * Session-handling fixes:
7966 - Fix handling of connections that are resuming with a session ID,
7967 but also support Session Tickets.
7968 - Fix a bug that suppressed issuing of a new ticket if the client
7969 presented a ticket with an expired session.
7970 - Try to set the ticket lifetime hint to something reasonable.
7971 - Make tickets shorter by excluding irrelevant information.
7972 - On the client side, don't ignore renewed tickets.
7974 *Adam Langley, Bodo Moeller (Google)*
7976 * Fix PSK session representation.
7980 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7982 This work was sponsored by Intel.
7986 * Add GCM support to TLS library. Some custom code is needed to split
7987 the IV between the fixed (from PRF) and explicit (from TLS record)
7988 portions. This adds all GCM ciphersuites supported by RFC5288 and
7989 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7990 add a special AESGCM string for GCM only.
7994 * Expand range of ctrls for AES GCM. Permit setting invocation
7995 field on decrypt and retrieval of invocation field only on encrypt.
7999 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8000 As required by RFC5289 these ciphersuites cannot be used if for
8001 versions of TLS earlier than 1.2.
8005 * For FIPS capable OpenSSL interpret a NULL default public key method
8006 as unset and return the appropriate default but do *not* set the default.
8007 This means we can return the appropriate method in applications that
8008 switch between FIPS and non-FIPS modes.
8012 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8013 ENGINE is used then we cannot handle that in the FIPS module so we
8014 keep original code iff non-FIPS operations are allowed.
8018 * Add -attime option to openssl utilities.
8020 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8022 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8026 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8027 FIPS EC methods unconditionally for now.
8031 * New build option no-ec2m to disable characteristic 2 code.
8035 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8036 all cases can be covered as some introduce binary incompatibilities.
8040 * Redirect RSA operations to FIPS module including keygen,
8041 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8045 * Add similar low-level API blocking to ciphers.
8049 * low-level digest APIs are not approved in FIPS mode: any attempt
8050 to use these will cause a fatal error. Applications that *really* want
8051 to use them can use the `private_*` version instead.
8055 * Redirect cipher operations to FIPS module for FIPS builds.
8059 * Redirect digest operations to FIPS module for FIPS builds.
8063 * Update build system to add "fips" flag which will link in fipscanister.o
8064 for static and shared library builds embedding a signature if needed.
8068 * Output TLS supported curves in preference order instead of numerical
8069 order. This is currently hardcoded for the highest order curves first.
8070 This should be configurable so applications can judge speed vs strength.
8074 * Add TLS v1.2 server support for client authentication.
8078 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8083 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8084 FIPS modules versions.
8088 * Add TLS v1.2 client side support for client authentication. Keep cache
8089 of handshake records longer as we don't know the hash algorithm to use
8090 until after the certificate request message is received.
8094 * Initial TLS v1.2 client support. Add a default signature algorithms
8095 extension including all the algorithms we support. Parse new signature
8096 format in client key exchange. Relax some ECC signing restrictions for
8097 TLS v1.2 as indicated in RFC5246.
8101 * Add server support for TLS v1.2 signature algorithms extension. Switch
8102 to new signature format when needed using client digest preference.
8103 All server ciphersuites should now work correctly in TLS v1.2. No client
8104 support yet and no support for client certificates.
8108 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8109 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8110 ciphersuites. At present only RSA key exchange ciphersuites work with
8111 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8112 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8113 and version checking.
8117 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8118 with this defined it will not be affected by any changes to ssl internal
8119 structures. Add several utility functions to allow openssl application
8120 to work with OPENSSL_NO_SSL_INTERN defined.
8124 * A long standing patch to add support for SRP from EdelWeb (Peter
8125 Sylvester and Christophe Renou) was integrated.
8126 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8127 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8130 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8134 * Permit abbreviated handshakes when renegotiating using the function
8135 SSL_renegotiate_abbreviated().
8137 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8139 * Add call to ENGINE_register_all_complete() to
8140 ENGINE_load_builtin_engines(), so some implementations get used
8141 automatically instead of needing explicit application support.
8145 * Add support for TLS key exporter as described in RFC5705.
8147 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8149 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8150 a few changes are required:
8152 Add SSL_OP_NO_TLSv1_1 flag.
8153 Add TLSv1_1 methods.
8154 Update version checking logic to handle version 1.1.
8155 Add explicit IV handling (ported from DTLS code).
8156 Add command line options to s_client/s_server.
8163 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8165 * X509_ATTRIBUTE memory leak
8167 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8168 memory. This structure is used by the PKCS#7 and CMS routines so any
8169 application which reads PKCS#7 or CMS data from untrusted sources is
8170 affected. SSL/TLS is not affected.
8172 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8178 * Race condition handling PSK identify hint
8180 If PSK identity hints are received by a multi-threaded client then
8181 the values are wrongly updated in the parent SSL_CTX structure. This can
8182 result in a race condition potentially leading to a double free of the
8188 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8190 * Malformed ECParameters causes infinite loop
8192 When processing an ECParameters structure OpenSSL enters an infinite loop
8193 if the curve specified is over a specially malformed binary polynomial
8196 This can be used to perform denial of service against any
8197 system which processes public keys, certificate requests or
8198 certificates. This includes TLS clients and TLS servers with
8199 client authentication enabled.
8201 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8206 * Exploitable out-of-bounds read in X509_cmp_time
8208 X509_cmp_time does not properly check the length of the ASN1_TIME
8209 string and can read a few bytes out of bounds. In addition,
8210 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8213 An attacker can use this to craft malformed certificates and CRLs of
8214 various sizes and potentially cause a segmentation fault, resulting in
8215 a DoS on applications that verify certificates or CRLs. TLS clients
8216 that verify CRLs are affected. TLS clients and servers with client
8217 authentication enabled may be affected if they use custom verification
8220 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8221 independently by Hanno Böck.
8226 * PKCS7 crash with missing EnvelopedContent
8228 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8229 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8230 with missing content and trigger a NULL pointer dereference on parsing.
8232 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8233 structures from untrusted sources are affected. OpenSSL clients and
8234 servers are not affected.
8236 This issue was reported to OpenSSL by Michal Zalewski (Google).
8241 * CMS verify infinite loop with unknown hash function
8243 When verifying a signedData message the CMS code can enter an infinite loop
8244 if presented with an unknown hash function OID. This can be used to perform
8245 denial of service against any system which verifies signedData messages using
8247 This issue was reported to OpenSSL by Johannes Bauer.
8252 * Race condition handling NewSessionTicket
8254 If a NewSessionTicket is received by a multi-threaded client when attempting to
8255 reuse a previous ticket then a race condition can occur potentially leading to
8256 a double free of the ticket data.
8261 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8263 * Segmentation fault in ASN1_TYPE_cmp fix
8265 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8266 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8267 certificate signature algorithm consistency this can be used to crash any
8268 certificate verification operation and exploited in a DoS attack. Any
8269 application which performs certificate verification is vulnerable including
8270 OpenSSL clients and servers which enable client authentication.
8275 * ASN.1 structure reuse memory corruption fix
8277 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8278 memory corruption via an invalid write. Such reuse is and has been
8279 strongly discouraged and is believed to be rare.
8281 Applications that parse structures containing CHOICE or ANY DEFINED BY
8282 components may be affected. Certificate parsing (d2i_X509 and related
8283 functions) are however not affected. OpenSSL clients and servers are
8289 * PKCS7 NULL pointer dereferences fix
8291 The PKCS#7 parsing code does not handle missing outer ContentInfo
8292 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8293 missing content and trigger a NULL pointer dereference on parsing.
8295 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8296 otherwise parse PKCS#7 structures from untrusted sources are
8297 affected. OpenSSL clients and servers are not affected.
8299 This issue was reported to OpenSSL by Michal Zalewski (Google).
8304 * DoS via reachable assert in SSLv2 servers fix
8306 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8307 servers that both support SSLv2 and enable export cipher suites by sending
8308 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8310 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8311 (OpenSSL development team).
8316 * Use After Free following d2i_ECPrivatekey error fix
8318 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8319 could cause a use after free condition. This, in turn, could cause a double
8320 free in several private key parsing functions (such as d2i_PrivateKey
8321 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8322 for applications that receive EC private keys from untrusted
8323 sources. This scenario is considered rare.
8325 This issue was discovered by the BoringSSL project and fixed in their
8331 * X509_to_X509_REQ NULL pointer deref fix
8333 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8334 the certificate key is invalid. This function is rarely used in practice.
8336 This issue was discovered by Brian Carpenter.
8341 * Removed the export ciphers from the DEFAULT ciphers
8345 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8347 * Build fixes for the Windows and OpenVMS platforms
8349 *Matt Caswell and Richard Levitte*
8351 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8353 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8354 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8355 dereference. This could lead to a Denial Of Service attack. Thanks to
8356 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8361 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8362 dtls1_buffer_record function under certain conditions. In particular this
8363 could occur if an attacker sent repeated DTLS records with the same
8364 sequence number but for the next epoch. The memory leak could be exploited
8365 by an attacker in a Denial of Service attack through memory exhaustion.
8366 Thanks to Chris Mueller for reporting this issue.
8371 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8372 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8373 method would be set to NULL which could later result in a NULL pointer
8374 dereference. Thanks to Frank Schmirler for reporting this issue.
8379 * Abort handshake if server key exchange message is omitted for ephemeral
8382 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8383 reporting this issue.
8388 * Remove non-export ephemeral RSA code on client and server. This code
8389 violated the TLS standard by allowing the use of temporary RSA keys in
8390 non-export ciphersuites and could be used by a server to effectively
8391 downgrade the RSA key length used to a value smaller than the server
8392 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8393 INRIA or reporting this issue.
8398 * Fixed issue where DH client certificates are accepted without verification.
8399 An OpenSSL server will accept a DH certificate for client authentication
8400 without the certificate verify message. This effectively allows a client to
8401 authenticate without the use of a private key. This only affects servers
8402 which trust a client certificate authority which issues certificates
8403 containing DH keys: these are extremely rare and hardly ever encountered.
8404 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8410 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8411 results on some platforms, including x86_64. This bug occurs at random
8412 with a very low probability, and is not known to be exploitable in any
8413 way, though its exact impact is difficult to determine. Thanks to Pieter
8414 Wuille (Blockstream) who reported this issue and also suggested an initial
8415 fix. Further analysis was conducted by the OpenSSL development team and
8416 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8417 the OpenSSL core team.
8422 * Fix various certificate fingerprint issues.
8424 By using non-DER or invalid encodings outside the signed portion of a
8425 certificate the fingerprint can be changed without breaking the signature.
8426 Although no details of the signed portion of the certificate can be changed
8427 this can cause problems with some applications: e.g. those using the
8428 certificate fingerprint for blacklists.
8430 1. Reject signatures with non zero unused bits.
8432 If the BIT STRING containing the signature has non zero unused bits reject
8433 the signature. All current signature algorithms require zero unused bits.
8435 2. Check certificate algorithm consistency.
8437 Check the AlgorithmIdentifier inside TBS matches the one in the
8438 certificate signature. NB: this will result in signature failure
8439 errors for some broken certificates.
8441 Thanks to Konrad Kraszewski from Google for reporting this issue.
8443 3. Check DSA/ECDSA signatures use DER.
8445 Reencode DSA/ECDSA signatures and compare with the original received
8446 signature. Return an error if there is a mismatch.
8448 This will reject various cases including garbage after signature
8449 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8450 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8451 (negative or with leading zeroes).
8453 Further analysis was conducted and fixes were developed by Stephen Henson
8454 of the OpenSSL core team.
8460 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8462 * Session Ticket Memory Leak.
8464 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8465 integrity of that ticket is first verified. In the event of a session
8466 ticket integrity check failing, OpenSSL will fail to free memory
8467 causing a memory leak. By sending a large number of invalid session
8468 tickets an attacker could exploit this issue in a Denial Of Service
8474 * Build option no-ssl3 is incomplete.
8476 When OpenSSL is configured with "no-ssl3" as a build option, servers
8477 could accept and complete a SSL 3.0 handshake, and clients could be
8478 configured to send them.
8481 *Akamai and the OpenSSL team*
8483 * Add support for TLS_FALLBACK_SCSV.
8484 Client applications doing fallback retries should call
8485 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8488 *Adam Langley, Bodo Moeller*
8490 * Add additional DigestInfo checks.
8492 Reencode DigestInto in DER and check against the original when
8493 verifying RSA signature: this will reject any improperly encoded
8494 DigestInfo structures.
8496 Note: this is a precautionary measure and no attacks are currently known.
8500 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8502 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8503 to a denial of service attack. A malicious server can crash the client
8504 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8505 ciphersuite and sending carefully crafted handshake messages.
8507 Thanks to Felix Gröbert (Google) for discovering and researching this
8513 * By sending carefully crafted DTLS packets an attacker could cause openssl
8514 to leak memory. This can be exploited through a Denial of Service attack.
8515 Thanks to Adam Langley for discovering and researching this issue.
8520 * An attacker can force openssl to consume large amounts of memory whilst
8521 processing DTLS handshake messages. This can be exploited through a
8522 Denial of Service attack.
8523 Thanks to Adam Langley for discovering and researching this issue.
8528 * An attacker can force an error condition which causes openssl to crash
8529 whilst processing DTLS packets due to memory being freed twice. This
8530 can be exploited through a Denial of Service attack.
8531 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8537 * If a multithreaded client connects to a malicious server using a resumed
8538 session and the server sends an ec point format extension it could write
8539 up to 255 bytes to freed memory.
8541 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8547 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8548 X509_name_oneline, X509_name_print_ex et al. to leak some information
8549 from the stack. Applications may be affected if they echo pretty printing
8550 output to the attacker.
8552 Thanks to Ivan Fratric (Google) for discovering this issue.
8555 *Emilia Käsper, and Steve Henson*
8557 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8558 for corner cases. (Certain input points at infinity could lead to
8559 bogus results, with non-infinity inputs mapped to infinity too.)
8563 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8565 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8566 handshake can force the use of weak keying material in OpenSSL
8567 SSL/TLS clients and servers.
8569 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8570 researching this issue. ([CVE-2014-0224])
8572 *KIKUCHI Masashi, Steve Henson*
8574 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8575 OpenSSL DTLS client the code can be made to recurse eventually crashing
8578 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8581 *Imre Rad, Steve Henson*
8583 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8584 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8585 client or server. This is potentially exploitable to run arbitrary
8586 code on a vulnerable client or server.
8588 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8590 *Jüri Aedla, Steve Henson*
8592 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8593 are subject to a denial of service attack.
8595 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8596 this issue. ([CVE-2014-3470])
8598 *Felix Gröbert, Ivan Fratric, Steve Henson*
8600 * Harmonize version and its documentation. -f flag is used to display
8603 *mancha <mancha1@zoho.com>*
8605 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8606 in i2d_ECPrivateKey.
8608 *mancha <mancha1@zoho.com>*
8610 * Fix some double frees. These are not thought to be exploitable.
8612 *mancha <mancha1@zoho.com>*
8614 * Fix for the attack described in the paper "Recovering OpenSSL
8615 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8616 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8617 <http://eprint.iacr.org/2014/140>
8619 Thanks to Yuval Yarom and Naomi Benger for discovering this
8620 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8622 *Yuval Yarom and Naomi Benger*
8624 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8626 * Keep original DTLS digest and encryption contexts in retransmission
8627 structures so we can use the previous session parameters if they need
8628 to be resent. ([CVE-2013-6450])
8632 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8633 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8634 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8635 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8636 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8637 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8639 *Rob Stradling, Adam Langley*
8641 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8643 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8645 This addresses the flaw in CBC record processing discovered by
8646 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8647 at: <http://www.isg.rhul.ac.uk/tls/>
8649 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8650 Security Group at Royal Holloway, University of London
8651 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8652 Emilia Käsper for the initial patch.
8655 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8657 * Return an error when checking OCSP signatures when key is NULL.
8658 This fixes a DoS attack. ([CVE-2013-0166])
8662 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8663 the right response is stapled. Also change SSL_get_certificate()
8664 so it returns the certificate actually sent.
8665 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8666 (This is a backport)
8668 *Rob Stradling <rob.stradling@comodo.com>*
8670 * Fix possible deadlock when decoding public keys.
8674 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8676 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8679 * Sanity check record length before skipping explicit IV in DTLS
8682 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8683 fuzzing as a service testing platform.
8688 * Initialise tkeylen properly when encrypting CMS messages.
8689 Thanks to Solar Designer of Openwall for reporting this issue.
8693 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8695 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8696 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8697 in CRYPTO_realloc_clean.
8699 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8700 issue and to Adam Langley <agl@chromium.org> for fixing it.
8703 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8705 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8707 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8708 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8709 content decryption and always return the same error. Note: this attack
8710 needs on average 2^20 messages so it only affects automated senders. The
8711 old behaviour can be re-enabled in the CMS code by setting the
8712 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8713 an MMA defence is not necessary.
8714 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8715 this issue. ([CVE-2012-0884])
8719 * Fix CVE-2011-4619: make sure we really are receiving a
8720 client hello before rejecting multiple SGC restarts. Thanks to
8721 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8725 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8727 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8728 Thanks to Antonio Martin, Enterprise Secure Access Research and
8729 Development, Cisco Systems, Inc. for discovering this bug and
8730 preparing a fix. ([CVE-2012-0050])
8734 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8736 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8737 of the Vaudenay padding oracle attack on CBC mode encryption
8738 which enables an efficient plaintext recovery attack against
8739 the OpenSSL implementation of DTLS. Their attack exploits timing
8740 differences arising during decryption processing. A research
8741 paper describing this attack can be found at:
8742 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8743 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8744 Security Group at Royal Holloway, University of London
8745 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8746 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8747 for preparing the fix. ([CVE-2011-4108])
8749 *Robin Seggelmann, Michael Tuexen*
8751 * Clear bytes used for block padding of SSL 3.0 records.
8754 *Adam Langley (Google)*
8756 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8757 Kadianakis <desnacked@gmail.com> for discovering this issue and
8758 Adam Langley for preparing the fix. ([CVE-2011-4619])
8760 *Adam Langley (Google)*
8762 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8764 *Andrey Kulikov <amdeich@gmail.com>*
8766 * Prevent malformed RFC3779 data triggering an assertion failure.
8767 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8768 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8770 *Rob Austein <sra@hactrn.net>*
8772 * Improved PRNG seeding for VOS.
8774 *Paul Green <Paul.Green@stratus.com>*
8776 * Fix ssl_ciph.c set-up race.
8778 *Adam Langley (Google)*
8780 * Fix spurious failures in ecdsatest.c.
8782 *Emilia Käsper (Google)*
8784 * Fix the BIO_f_buffer() implementation (which was mixing different
8785 interpretations of the `..._len` fields).
8787 *Adam Langley (Google)*
8789 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8790 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8791 threads won't reuse the same blinding coefficients.
8793 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8794 lock to call BN_BLINDING_invert_ex, and avoids one use of
8795 BN_BLINDING_update for each BN_BLINDING structure (previously,
8796 the last update always remained unused).
8798 *Emilia Käsper (Google)*
8800 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8802 *Bob Buckholz (Google)*
8804 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8806 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8807 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8809 *Kaspar Brand <ossl@velox.ch>*
8811 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8812 for multi-threaded use of ECDH. ([CVE-2011-3210])
8814 *Adam Langley (Google)*
8816 * Fix x509_name_ex_d2i memory leak on bad inputs.
8820 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8821 signature public key algorithm by using OID xref utilities instead.
8822 Before this you could only use some ECC ciphersuites with SHA1 only.
8826 * Add protection against ECDSA timing attacks as mentioned in the paper
8827 by Billy Bob Brumley and Nicola Tuveri, see:
8828 <http://eprint.iacr.org/2011/232.pdf>
8830 *Billy Bob Brumley and Nicola Tuveri*
8832 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8834 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8836 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8838 * Fix bug in string printing code: if *any* escaping is enabled we must
8839 escape the escape character (backslash) or the resulting string is
8844 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8846 * Disable code workaround for ancient and obsolete Netscape browsers
8847 and servers: an attacker can use it in a ciphersuite downgrade attack.
8848 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8852 * Fixed J-PAKE implementation error, originally discovered by
8853 Sebastien Martini, further info and confirmation from Stefan
8854 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8858 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8860 * Fix extension code to avoid race conditions which can result in a buffer
8861 overrun vulnerability: resumed sessions must not be modified as they can
8862 be shared by multiple threads. CVE-2010-3864
8866 * Fix WIN32 build system to correctly link an ENGINE directory into
8871 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8873 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8876 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8878 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8880 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8881 context. The operation can be customised via the ctrl mechanism in
8882 case ENGINEs want to include additional functionality.
8886 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8890 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8891 output hashes compatible with older versions of OpenSSL.
8893 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8895 * Fix compression algorithm handling: if resuming a session use the
8896 compression algorithm of the resumed session instead of determining
8897 it from client hello again. Don't allow server to change algorithm.
8901 * Add load_crls() function to commands tidying load_certs() too. Add option
8902 to verify utility to allow additional CRLs to be included.
8906 * Update OCSP request code to permit adding custom headers to the request:
8907 some responders need this.
8911 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8914 *Julia Lawall <julia@diku.dk>*
8916 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8917 needlessly dereferenced structures, used obsolete functions and
8918 didn't handle all updated verify codes correctly.
8922 * Disable MD2 in the default configuration.
8926 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8927 indicate the initial BIO being pushed or popped. This makes it possible
8928 to determine whether the BIO is the one explicitly called or as a result
8929 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8930 it handles reference counts correctly and doesn't zero out the I/O bio
8931 when it is not being explicitly popped. WARNING: applications which
8932 included workarounds for the old buggy behaviour will need to be modified
8933 or they could free up already freed BIOs.
8937 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8938 renaming to all platforms (within the 0.9.8 branch, this was
8939 done conditionally on Netware platforms to avoid a name clash).
8941 *Guenter <lists@gknw.net>*
8943 * Add ECDHE and PSK support to DTLS.
8945 *Michael Tuexen <tuexen@fh-muenster.de>*
8947 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8952 * Add "missing" function EVP_MD_flags() (without this the only way to
8953 retrieve a digest flags is by accessing the structure directly. Update
8954 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8955 or cipher is registered as in the "from" argument. Print out all
8956 registered digests in the dgst usage message instead of manually
8957 attempting to work them out.
8961 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8962 this allows the use of compression and extensions. Change default cipher
8963 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8964 by default unless an application cipher string requests it.
8968 * Alter match criteria in PKCS12_parse(). It used to try to use local
8969 key ids to find matching certificates and keys but some PKCS#12 files
8970 don't follow the (somewhat unwritten) rules and this strategy fails.
8971 Now just gather all certificates together and the first private key
8972 then look for the first certificate that matches the key.
8976 * Support use of registered digest and cipher names for dgst and cipher
8977 commands instead of having to add each one as a special case. So now
8984 openssl dgst -sha256 foo
8986 and this works for ENGINE based algorithms too.
8990 * Update Gost ENGINE to support parameter files.
8992 *Victor B. Wagner <vitus@cryptocom.ru>*
8994 * Support GeneralizedTime in ca utility.
8996 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8998 * Enhance the hash format used for certificate directory links. The new
8999 form uses the canonical encoding (meaning equivalent names will work
9000 even if they aren't identical) and uses SHA1 instead of MD5. This form
9001 is incompatible with the older format and as a result c_rehash should
9002 be used to rebuild symbolic links.
9006 * Make PKCS#8 the default write format for private keys, replacing the
9007 traditional format. This form is standardised, more secure and doesn't
9008 include an implicit MD5 dependency.
9012 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9013 committed to OpenSSL should pass this lot as a minimum.
9017 * Add session ticket override functionality for use by EAP-FAST.
9019 *Jouni Malinen <j@w1.fi>*
9021 * Modify HMAC functions to return a value. Since these can be implemented
9022 in an ENGINE errors can occur.
9026 * Type-checked OBJ_bsearch_ex.
9030 * Type-checked OBJ_bsearch. Also some constification necessitated
9031 by type-checking. Still to come: TXT_DB, bsearch(?),
9032 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9037 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9038 seconds to a tm structure directly, instead of going through OS
9039 specific date routines. This avoids any issues with OS routines such
9040 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9041 and X509_time_adj_ex() to cover the extended range. The existing
9042 X509_time_adj() is still usable and will no longer have any date issues.
9046 * Delta CRL support. New use deltas option which will attempt to locate
9047 and search any appropriate delta CRLs available.
9049 This work was sponsored by Google.
9053 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9054 code and add additional score elements. Validate alternate CRL paths
9055 as part of the CRL checking and indicate a new error "CRL path validation
9056 error" in this case. Applications wanting additional details can use
9057 the verify callback and check the new "parent" field. If this is not
9058 NULL CRL path validation is taking place. Existing applications won't
9059 see this because it requires extended CRL support which is off by
9062 This work was sponsored by Google.
9066 * Support for freshest CRL extension.
9068 This work was sponsored by Google.
9072 * Initial indirect CRL support. Currently only supported in the CRLs
9073 passed directly and not via lookup. Process certificate issuer
9074 CRL entry extension and lookup CRL entries by bother issuer name
9075 and serial number. Check and process CRL issuer entry in IDP extension.
9077 This work was sponsored by Google.
9081 * Add support for distinct certificate and CRL paths. The CRL issuer
9082 certificate is validated separately in this case. Only enabled if
9083 an extended CRL support flag is set: this flag will enable additional
9084 CRL functionality in future.
9086 This work was sponsored by Google.
9090 * Add support for policy mappings extension.
9092 This work was sponsored by Google.
9096 * Fixes to pathlength constraint, self issued certificate handling,
9097 policy processing to align with RFC3280 and PKITS tests.
9099 This work was sponsored by Google.
9103 * Support for name constraints certificate extension. DN, email, DNS
9104 and URI types are currently supported.
9106 This work was sponsored by Google.
9110 * To cater for systems that provide a pointer-based thread ID rather
9111 than numeric, deprecate the current numeric thread ID mechanism and
9112 replace it with a structure and associated callback type. This
9113 mechanism allows a numeric "hash" to be extracted from a thread ID in
9114 either case, and on platforms where pointers are larger than 'long',
9115 mixing is done to help ensure the numeric 'hash' is usable even if it
9116 can't be guaranteed unique. The default mechanism is to use "&errno"
9117 as a pointer-based thread ID to distinguish between threads.
9119 Applications that want to provide their own thread IDs should now use
9120 CRYPTO_THREADID_set_callback() to register a callback that will call
9121 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9123 Note that ERR_remove_state() is now deprecated, because it is tied
9124 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9125 to free the current thread's error state should be replaced by
9126 ERR_remove_thread_state(NULL).
9128 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9129 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9130 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9131 application was previously providing a numeric thread callback that
9132 was inappropriate for distinguishing threads, then uniqueness might
9133 have been obtained with &errno that happened immediately in the
9134 intermediate development versions of OpenSSL; this is no longer the
9135 case, the numeric thread callback will now override the automatic use
9138 *Geoff Thorpe, with help from Bodo Moeller*
9140 * Initial support for different CRL issuing certificates. This covers a
9141 simple case where the self issued certificates in the chain exist and
9142 the real CRL issuer is higher in the existing chain.
9144 This work was sponsored by Google.
9148 * Removed effectively defunct crypto/store from the build.
9152 * Revamp of STACK to provide stronger type-checking. Still to come:
9153 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9154 ASN1_STRING, CONF_VALUE.
9158 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9159 RAM on SSL connections. This option can save about 34k per idle SSL.
9163 * Revamp of LHASH to provide stronger type-checking. Still to come:
9164 STACK, TXT_DB, bsearch, qsort.
9168 * Initial support for Cryptographic Message Syntax (aka CMS) based
9169 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9170 support for data, signedData, compressedData, digestedData and
9171 encryptedData, envelopedData types included. Scripts to check against
9172 RFC4134 examples draft and interop and consistency checks of many
9173 content types and variants.
9177 * Add options to enc utility to support use of zlib compression BIO.
9181 * Extend mk1mf to support importing of options and assembly language
9182 files from Configure script, currently only included in VC-WIN32.
9183 The assembly language rules can now optionally generate the source
9184 files from the associated perl scripts.
9188 * Implement remaining functionality needed to support GOST ciphersuites.
9189 Interop testing has been performed using CryptoPro implementations.
9191 *Victor B. Wagner <vitus@cryptocom.ru>*
9193 * s390x assembler pack.
9197 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9202 * Implement Opaque PRF Input TLS extension as specified in
9203 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9204 official specification yet and no extension type assignment by
9205 IANA exists, this extension (for now) will have to be explicitly
9206 enabled when building OpenSSL by providing the extension number
9207 to use. For example, specify an option
9209 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9211 to the "config" or "Configure" script to enable the extension,
9212 assuming extension number 0x9527 (which is a completely arbitrary
9213 and unofficial assignment based on the MD5 hash of the Internet
9214 Draft). Note that by doing so, you potentially lose
9215 interoperability with other TLS implementations since these might
9216 be using the same extension number for other purposes.
9218 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9219 opaque PRF input value to use in the handshake. This will create
9220 an internal copy of the length-'len' string at 'src', and will
9221 return non-zero for success.
9223 To get more control and flexibility, provide a callback function
9226 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9227 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9231 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9234 Callback function 'cb' will be called in handshakes, and is
9235 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9236 Argument 'arg' is for application purposes (the value as given to
9237 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9238 be provided to the callback function). The callback function
9239 has to return non-zero to report success: usually 1 to use opaque
9240 PRF input just if possible, or 2 to enforce use of the opaque PRF
9241 input. In the latter case, the library will abort the handshake
9242 if opaque PRF input is not successfully negotiated.
9244 Arguments 'peerinput' and 'len' given to the callback function
9245 will always be NULL and 0 in the case of a client. A server will
9246 see the client's opaque PRF input through these variables if
9247 available (NULL and 0 otherwise). Note that if the server
9248 provides an opaque PRF input, the length must be the same as the
9249 length of the client's opaque PRF input.
9251 Note that the callback function will only be called when creating
9252 a new session (session resumption can resume whatever was
9253 previously negotiated), and will not be called in SSL 2.0
9254 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9255 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9256 for applications that need to enforce opaque PRF input.
9260 * Update ssl code to support digests other than SHA1+MD5 for handshake
9263 *Victor B. Wagner <vitus@cryptocom.ru>*
9265 * Add RFC4507 support to OpenSSL. This includes the corrections in
9266 RFC4507bis. The encrypted ticket format is an encrypted encoded
9267 SSL_SESSION structure, that way new session features are automatically
9270 If a client application caches session in an SSL_SESSION structure
9271 support is transparent because tickets are now stored in the encoded
9274 The SSL_CTX structure automatically generates keys for ticket
9275 protection in servers so again support should be possible
9276 with no application modification.
9278 If a client or server wishes to disable RFC4507 support then the option
9279 SSL_OP_NO_TICKET can be set.
9281 Add a TLS extension debugging callback to allow the contents of any client
9282 or server extensions to be examined.
9284 This work was sponsored by Google.
9288 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9289 OpenSSL should now compile cleanly on gcc 4.2
9291 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9293 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9294 support including streaming MAC support: this is required for GOST
9295 ciphersuite support.
9297 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9299 * Add option -stream to use PKCS#7 streaming in smime utility. New
9300 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9301 to output in BER and PEM format.
9305 * Experimental support for use of HMAC via EVP_PKEY interface. This
9306 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9307 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9308 ENGINE support for HMAC keys which are unextractable. New -mac and
9309 -macopt options to dgst utility.
9313 * New option -sigopt to dgst utility. Update dgst to use
9314 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9315 alternative signing parameters such as X9.31 or PSS in the dgst
9320 * Change ssl_cipher_apply_rule(), the internal function that does
9321 the work each time a ciphersuite string requests enabling
9322 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9323 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9324 the order of disabled ciphersuites such that those ciphersuites
9325 that most recently went from enabled to disabled not only stay
9326 in order with respect to each other, but also have higher priority
9327 than other disabled ciphersuites the next time ciphersuites are
9330 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9331 the same ciphersuites as with "HIGH" alone, but in a specific
9332 order where the PSK ciphersuites come first (since they are the
9333 most recently disabled ciphersuites when "HIGH" is parsed).
9335 Also, change ssl_create_cipher_list() (using this new
9336 functionality) such that between otherwise identical
9337 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9342 * Change ssl_create_cipher_list() so that it automatically
9343 arranges the ciphersuites in reasonable order before starting
9344 to process the rule string. Thus, the definition for "DEFAULT"
9345 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9346 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9347 This makes it much easier to arrive at a reasonable default order
9348 in applications for which anonymous ciphers are OK (meaning
9349 that you can't actually use DEFAULT).
9351 *Bodo Moeller; suggested by Victor Duchovni*
9353 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9354 processing) into multiple integers instead of setting
9355 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9356 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9357 (These masks as well as the individual bit definitions are hidden
9358 away into the non-exported interface ssl/ssl_locl.h, so this
9359 change to the definition of the SSL_CIPHER structure shouldn't
9360 affect applications.) This give us more bits for each of these
9361 categories, so there is no longer a need to coagulate AES128 and
9362 AES256 into a single algorithm bit, and to coagulate Camellia128
9363 and Camellia256 into a single algorithm bit, which has led to all
9366 Thus, among other things, the kludge introduced in 0.9.7m and
9367 0.9.8e for masking out AES256 independently of AES128 or masking
9368 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9370 With the change, we also introduce new ciphersuite aliases that
9371 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9376 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9377 Use the leftmost N bytes of the signature input if the input is
9378 larger than the prime q (with N being the size in bytes of q).
9382 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9383 it yet and it is largely untested.
9387 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9391 * Initial incomplete changes to avoid need for function casts in OpenSSL
9392 some compilers (gcc 4.2 and later) reject their use. Safestack is
9393 reimplemented. Update ASN1 to avoid use of legacy functions.
9397 * Win32/64 targets are linked with Winsock2.
9401 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9402 to external functions. This can be used to increase CRL handling
9403 efficiency especially when CRLs are very large by (for example) storing
9404 the CRL revoked certificates in a database.
9408 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9409 new CRLs added to a directory can be used. New command line option
9410 -verify_return_error to s_client and s_server. This causes real errors
9411 to be returned by the verify callback instead of carrying on no matter
9412 what. This reflects the way a "real world" verify callback would behave.
9416 * GOST engine, supporting several GOST algorithms and public key formats.
9417 Kindly donated by Cryptocom.
9421 * Partial support for Issuing Distribution Point CRL extension. CRLs
9422 partitioned by DP are handled but no indirect CRL or reason partitioning
9423 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9424 selected via a scoring technique which handles IDP and AKID in CRLs.
9428 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9429 will ultimately be used for all verify operations: this will remove the
9430 X509_STORE dependency on certificate verification and allow alternative
9431 lookup methods. X509_STORE based implementations of these two callbacks.
9435 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9436 Modify get_crl() to find a valid (unexpired) CRL if possible.
9440 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9441 this would be called X509_CRL_cmp() but that name is already used by
9442 a function that just compares CRL issuer names. Cache several CRL
9443 extensions in X509_CRL structure and cache CRLDP in X509.
9447 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9448 this maps equivalent X509_NAME structures into a consistent structure.
9449 Name comparison can then be performed rapidly using memcmp().
9453 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9458 * Allow digests to supply their own micalg string for S/MIME type using
9459 the ctrl EVP_MD_CTRL_MICALG.
9463 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9464 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9465 ctrl. It can then customise the structure before and/or after signing
9470 * New function OBJ_add_sigid() to allow application defined signature OIDs
9471 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9472 to free up any added signature OIDs.
9476 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9477 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9478 digest and cipher tables. New options added to openssl utility:
9479 list-message-digest-algorithms and list-cipher-algorithms.
9483 * Change the array representation of binary polynomials: the list
9484 of degrees of non-zero coefficients is now terminated with -1.
9485 Previously it was terminated with 0, which was also part of the
9486 value; thus, the array representation was not applicable to
9487 polynomials where t^0 has coefficient zero. This change makes
9488 the array representation useful in a more general context.
9492 * Various modifications and fixes to SSL/TLS cipher string
9493 handling. For ECC, the code now distinguishes between fixed ECDH
9494 with RSA certificates on the one hand and with ECDSA certificates
9495 on the other hand, since these are separate ciphersuites. The
9496 unused code for Fortezza ciphersuites has been removed.
9498 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9499 (not "ECDHE"). For consistency with the code for DH
9500 certificates, use of ECDH certificates is now considered ECDH
9501 authentication, not RSA or ECDSA authentication (the latter is
9502 merely the CA's signing algorithm and not actively used in the
9505 The temporary ciphersuite alias "ECCdraft" is no longer
9506 available, and ECC ciphersuites are no longer excluded from "ALL"
9507 and "DEFAULT". The following aliases now exist for RFC 4492
9508 ciphersuites, most of these by analogy with the DH case:
9510 kECDHr - ECDH cert, signed with RSA
9511 kECDHe - ECDH cert, signed with ECDSA
9512 kECDH - ECDH cert (signed with either RSA or ECDSA)
9513 kEECDH - ephemeral ECDH
9514 ECDH - ECDH cert or ephemeral ECDH
9520 AECDH - anonymous ECDH
9521 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9525 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9526 Use correct micalg parameters depending on digest(s) in signed message.
9530 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9531 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9535 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9536 an engine to register a method. Add ENGINE lookups for methods and
9537 functional reference processing.
9541 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9542 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9547 * New -resign option to smime utility. This adds one or more signers
9548 to an existing PKCS#7 signedData structure. Also -md option to use an
9549 alternative message digest algorithm for signing.
9553 * Tidy up PKCS#7 routines and add new functions to make it easier to
9554 create PKCS7 structures containing multiple signers. Update smime
9555 application to support multiple signers.
9559 * New -macalg option to pkcs12 utility to allow setting of an alternative
9564 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9565 Reorganize PBE internals to lookup from a static table using NIDs,
9566 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9567 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9568 PRF which will be automatically used with PBES2.
9572 * Replace the algorithm specific calls to generate keys in "req" with the
9577 * Update PKCS#7 enveloped data routines to use new API. This is now
9578 supported by any public key method supporting the encrypt operation. A
9579 ctrl is added to allow the public key algorithm to examine or modify
9580 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9585 * Add a ctrl to asn1 method to allow a public key algorithm to express
9586 a default digest type to use. In most cases this will be SHA1 but some
9587 algorithms (such as GOST) need to specify an alternative digest. The
9588 return value indicates how strong the preference is 1 means optional and
9589 2 is mandatory (that is it is the only supported type). Modify
9590 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9591 use the default md. Update openssl utilities to use the default digest
9592 type for signing if it is not explicitly indicated.
9596 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9597 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9598 signing method from the key type. This effectively removes the link
9599 between digests and public key types.
9603 * Add an OID cross reference table and utility functions. Its purpose is to
9604 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9605 rsaEncryption. This will allow some of the algorithm specific hackery
9606 needed to use the correct OID to be removed.
9610 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9611 structures for PKCS7_sign(). They are now set up by the relevant public
9616 * Add provisional EC pkey method with support for ECDSA and ECDH.
9620 * Add support for key derivation (agreement) in the API, DH method and
9625 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9626 public and private key formats. As a side effect these add additional
9627 command line functionality not previously available: DSA signatures can be
9628 generated and verified using pkeyutl and DH key support and generation in
9635 *Oliver Tappe <zooey@hirschkaefer.de>*
9637 * New make target "install_html_docs" installs HTML renditions of the
9640 *Oliver Tappe <zooey@hirschkaefer.de>*
9642 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9643 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9644 support key and parameter generation and add initial key generation
9645 functionality for RSA.
9649 * Add functions for main EVP_PKEY_method operations. The undocumented
9650 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9651 `EVP_PKEY_{encrypt,decrypt}_old`.
9655 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9656 key API, doesn't do much yet.
9660 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9661 public key algorithms. New option to openssl utility:
9662 "list-public-key-algorithms" to print out info.
9666 * Implement the Supported Elliptic Curves Extension for
9667 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9671 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9672 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9676 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9677 utilities such as rsa, dsa, dsaparam etc except they process any key
9682 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9683 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9684 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9689 * Initial support for pluggable public key ASN1.
9690 De-spaghettify the public key ASN1 handling. Move public and private
9691 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9692 algorithm specific handling to a single module within the relevant
9693 algorithm directory. Add functions to allow (near) opaque processing
9694 of public and private key structures.
9698 * Implement the Supported Point Formats Extension for
9699 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9703 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9704 for the psk identity [hint] and the psk callback functions to the
9705 SSL_SESSION, SSL and SSL_CTX structure.
9708 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9712 SSL_CTX_use_psk_identity_hint
9713 SSL_get_psk_identity_hint
9714 SSL_get_psk_identity
9715 SSL_use_psk_identity_hint
9717 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9719 * Add RFC 3161 compliant time stamp request creation, response generation
9720 and response verification functionality.
9722 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9724 * Add initial support for TLS extensions, specifically for the server_name
9725 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9726 have new members for a host name. The SSL data structure has an
9727 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9728 stored in that context to allow for session resumption, even after the
9729 SSL has been switched to a new SSL_CTX in reaction to a client's
9730 server_name extension.
9732 New functions (subject to change):
9734 SSL_get_servername()
9735 SSL_get_servername_type()
9738 New CTRL codes and macros (subject to change):
9740 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9741 - SSL_CTX_set_tlsext_servername_callback()
9742 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9743 - SSL_CTX_set_tlsext_servername_arg()
9744 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9746 openssl s_client has a new '-servername ...' option.
9748 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9749 '-key2 ...', '-servername_fatal' (subject to change). This allows
9750 testing the HostName extension for a specific single host name ('-cert'
9751 and '-key' remain fallbacks for handshakes without HostName
9752 negotiation). If the unrecognized_name alert has to be sent, this by
9753 default is a warning; it becomes fatal with the '-servername_fatal'
9756 *Peter Sylvester, Remy Allais, Christophe Renou*
9758 * Whirlpool hash implementation is added.
9762 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9763 bn(64,32). Because of instruction set limitations it doesn't have
9764 any negative impact on performance. This was done mostly in order
9765 to make it possible to share assembler modules, such as bn_mul_mont
9766 implementations, between 32- and 64-bit builds without hassle.
9770 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9771 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9776 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9777 dedicated Montgomery multiplication procedure, is introduced.
9778 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9779 "64-bit" performance on certain 32-bit targets.
9783 * New option SSL_OP_NO_COMP to disable use of compression selectively
9784 in SSL structures. New SSL ctrl to set maximum send fragment size.
9785 Save memory by setting the I/O buffer sizes dynamically instead of
9786 using the maximum available value.
9790 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9791 in addition to the text details.
9795 * Very, very preliminary EXPERIMENTAL support for printing of general
9796 ASN1 structures. This currently produces rather ugly output and doesn't
9797 handle several customised structures at all.
9801 * Integrated support for PVK file format and some related formats such
9802 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9803 these in the 'rsa' and 'dsa' utilities.
9807 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9811 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9812 place for the (very old) "NETSCAPE" format certificates which are now
9813 handled using new ASN1 code equivalents.
9817 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9818 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9819 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9823 * Modify CRL distribution points extension code to print out previously
9824 unsupported fields. Enhance extension setting code to allow setting of
9829 * Add print and set support for Issuing Distribution Point CRL extension.
9833 * Change 'Configure' script to enable Camellia by default.
9840 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9842 * When rejecting SSL/TLS records due to an incorrect version number, never
9843 update s->server with a new major version number. As of
9844 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9845 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9846 the previous behavior could result in a read attempt at NULL when
9847 receiving specific incorrect SSL/TLS records once record payload
9848 protection is active. ([CVE-2010-0740])
9850 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9852 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9853 could be crashed if the relevant tables were not present (e.g. chrooted).
9855 *Tomas Hoger <thoger@redhat.com>*
9857 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9859 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9861 *Martin Olsson, Neel Mehta*
9863 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9864 accommodate for stack sorting, always a write lock!).
9868 * On some versions of WIN32 Heap32Next is very slow. This can cause
9869 excessive delays in the RAND_poll(): over a minute. As a workaround
9870 include a time check in the inner Heap32Next loop too.
9874 * The code that handled flushing of data in SSL/TLS originally used the
9875 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9876 the problem outlined in PR#1949. The fix suggested there however can
9877 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9878 of Apache). So instead simplify the code to flush unconditionally.
9879 This should be fine since flushing with no data to flush is a no op.
9883 * Handle TLS versions 2.0 and later properly and correctly use the
9884 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9885 off ancient servers have a habit of sticking around for a while...
9889 * Modify compression code so it frees up structures without using the
9890 ex_data callbacks. This works around a problem where some applications
9891 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9892 restarting) then use compression (e.g. SSL with compression) later.
9893 This results in significant per-connection memory leaks and
9894 has caused some security issues including CVE-2008-1678 and
9899 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9900 change when encrypting or decrypting.
9904 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9905 connect and renegotiate with servers which do not support RI.
9906 Until RI is more widely deployed this option is enabled by default.
9910 * Add "missing" ssl ctrls to clear options and mode.
9914 * If client attempts to renegotiate and doesn't support RI respond with
9915 a no_renegotiation alert as required by RFC5746. Some renegotiating
9916 TLS clients will continue a connection gracefully when they receive
9917 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9918 waiting for a server hello which it will never receive. Now we treat a
9919 received no_renegotiation alert as a fatal error. This is because
9920 applications requesting a renegotiation might well expect it to succeed
9921 and would have no code in place to handle the server denying it so the
9922 only safe thing to do is to terminate the connection.
9926 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9927 peer supports secure renegotiation and 0 otherwise. Print out peer
9928 renegotiation support in s_client/s_server.
9932 * Replace the highly broken and deprecated SPKAC certification method with
9933 the updated NID creation version. This should correctly handle UTF8.
9937 * Implement RFC5746. Re-enable renegotiation but require the extension
9938 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9939 turns out to be a bad idea. It has been replaced by
9940 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9941 SSL_CTX_set_options(). This is really not recommended unless you
9942 know what you are doing.
9944 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9946 * Fixes to stateless session resumption handling. Use initial_ctx when
9947 issuing and attempting to decrypt tickets in case it has changed during
9948 servername handling. Use a non-zero length session ID when attempting
9949 stateless session resumption: this makes it possible to determine if
9950 a resumption has occurred immediately after receiving server hello
9951 (several places in OpenSSL subtly assume this) instead of later in
9956 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9957 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9958 fixes for a few places where the return code is not checked
9961 *Julia Lawall <julia@diku.dk>*
9963 * Add --strict-warnings option to Configure script to include devteam
9964 warnings in other configurations.
9968 * Add support for --libdir option and LIBDIR variable in makefiles. This
9969 makes it possible to install openssl libraries in locations which
9970 have names other than "lib", for example "/usr/lib64" which some
9973 *Steve Henson, based on patch from Jeremy Utley*
9975 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9976 X690 8.9.12 and can produce some misleading textual output of OIDs.
9978 *Steve Henson, reported by Dan Kaminsky*
9980 * Delete MD2 from algorithm tables. This follows the recommendation in
9981 several standards that it is not used in new applications due to
9982 several cryptographic weaknesses. For binary compatibility reasons
9983 the MD2 API is still compiled in by default.
9987 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9992 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9993 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9996 *Guenter <lists@gknw.net>*
9998 * Fix the server certificate chain building code to use X509_verify_cert(),
9999 it used to have an ad-hoc builder which was unable to cope with anything
10000 other than a simple chain.
10002 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10004 * Don't check self signed certificate signatures in X509_verify_cert()
10005 by default (a flag can override this): it just wastes time without
10006 adding any security. As a useful side effect self signed root CAs
10007 with non-FIPS digests are now usable in FIPS mode.
10011 * In dtls1_process_out_of_seq_message() the check if the current message
10012 is already buffered was missing. For every new message was memory
10013 allocated, allowing an attacker to perform an denial of service attack
10014 with sending out of seq handshake messages until there is no memory
10015 left. Additionally every future message was buffered, even if the
10016 sequence number made no sense and would be part of another handshake.
10017 So only messages with sequence numbers less than 10 in advance will be
10018 buffered. ([CVE-2009-1378])
10020 *Robin Seggelmann, discovered by Daniel Mentz*
10022 * Records are buffered if they arrive with a future epoch to be
10023 processed after finishing the corresponding handshake. There is
10024 currently no limitation to this buffer allowing an attacker to perform
10025 a DOS attack with sending records with future epochs until there is no
10026 memory left. This patch adds the pqueue_size() function to determine
10027 the size of a buffer and limits the record buffer to 100 entries.
10030 *Robin Seggelmann, discovered by Daniel Mentz*
10032 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10033 parent structure is freed. ([CVE-2009-1379])
10037 * Handle non-blocking I/O properly in SSL_shutdown() call.
10039 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10041 * Add `2.5.4.*` OIDs
10043 *Ilya O. <vrghost@gmail.com>*
10045 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10047 * Disable renegotiation completely - this fixes a severe security
10048 problem ([CVE-2009-3555]) at the cost of breaking all
10049 renegotiation. Renegotiation can be re-enabled by setting
10050 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10051 run-time. This is really not recommended unless you know what
10056 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10058 * Don't set val to NULL when freeing up structures, it is freed up by
10059 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10060 zeroing past the valid field. ([CVE-2009-0789])
10062 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10064 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10065 checked correctly. This would allow some invalid signed attributes to
10066 appear to verify correctly. ([CVE-2009-0591])
10068 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10070 * Reject UniversalString and BMPString types with invalid lengths. This
10071 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10072 a legal length. ([CVE-2009-0590])
10076 * Set S/MIME signing as the default purpose rather than setting it
10077 unconditionally. This allows applications to override it at the store
10082 * Permit restricted recursion of ASN1 strings. This is needed in practice
10083 to handle some structures.
10087 * Improve efficiency of mem_gets: don't search whole buffer each time
10090 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10092 * New -hex option for openssl rand.
10096 * Print out UTF8String and NumericString when parsing ASN1.
10100 * Support NumericString type for name components.
10104 * Allow CC in the environment to override the automatically chosen
10105 compiler. Note that nothing is done to ensure flags work with the
10110 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10112 * Properly check EVP_VerifyFinal() and similar return values
10115 *Ben Laurie, Bodo Moeller, Google Security Team*
10117 * Enable TLS extensions by default.
10121 * Allow the CHIL engine to be loaded, whether the application is
10122 multithreaded or not. (This does not release the developer from the
10123 obligation to set up the dynamic locking callbacks.)
10125 *Sander Temme <sander@temme.net>*
10127 * Use correct exit code if there is an error in dgst command.
10129 *Steve Henson; problem pointed out by Roland Dirlewanger*
10131 * Tweak Configure so that you need to say "experimental-jpake" to enable
10132 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10136 * Add experimental JPAKE support, including demo authentication in
10137 s_client and s_server.
10141 * Set the comparison function in v3_addr_canonize().
10143 *Rob Austein <sra@hactrn.net>*
10145 * Add support for XMPP STARTTLS in s_client.
10147 *Philip Paeps <philip@freebsd.org>*
10149 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10150 to ensure that even with this option, only ciphersuites in the
10151 server's preference list will be accepted. (Note that the option
10152 applies only when resuming a session, so the earlier behavior was
10153 just about the algorithm choice for symmetric cryptography.)
10157 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10159 * Fix NULL pointer dereference if a DTLS server received
10160 ChangeCipherSpec as first record ([CVE-2009-1386]).
10164 * Fix a state transition in s3_srvr.c and d1_srvr.c
10165 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10167 *Nagendra Modadugu*
10169 * The fix in 0.9.8c that supposedly got rid of unsafe
10170 double-checked locking was incomplete for RSA blinding,
10171 addressing just one layer of what turns out to have been
10172 doubly unsafe triple-checked locking.
10174 So now fix this for real by retiring the MONT_HELPER macro
10175 in crypto/rsa/rsa_eay.c.
10177 *Bodo Moeller; problem pointed out by Marius Schilder*
10179 * Various precautionary measures:
10181 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10183 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10184 (NB: This would require knowledge of the secret session ticket key
10185 to exploit, in which case you'd be SOL either way.)
10187 - Change bn_nist.c so that it will properly handle input BIGNUMs
10188 outside the expected range.
10190 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10193 *Neel Mehta, Bodo Moeller*
10195 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10196 the load fails. Useful for distros.
10198 *Ben Laurie and the FreeBSD team*
10200 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10204 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10208 * Expand ENGINE to support engine supplied SSL client certificate functions.
10210 This work was sponsored by Logica.
10214 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10215 keystores. Support for SSL/TLS client authentication too.
10216 Not compiled unless enable-capieng specified to Configure.
10218 This work was sponsored by Logica.
10222 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10223 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10224 attribute creation routines such as certificate requests and PKCS#12
10229 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10231 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10232 handshake which could lead to a client crash as found using the
10233 Codenomicon TLS test suite ([CVE-2008-1672])
10235 *Steve Henson, Mark Cox*
10237 * Fix double free in TLS server name extensions which could lead to
10238 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10242 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10244 Clear the error queue to ensure that error entries left from
10245 older function calls do not interfere with the correct operation.
10247 *Lutz Jaenicke, Erik de Castro Lopo*
10249 * Remove root CA certificates of commercial CAs:
10251 The OpenSSL project does not recommend any specific CA and does not
10252 have any policy with respect to including or excluding any CA.
10253 Therefore it does not make any sense to ship an arbitrary selection
10254 of root CA certificates with the OpenSSL software.
10258 * RSA OAEP patches to fix two separate invalid memory reads.
10259 The first one involves inputs when 'lzero' is greater than
10260 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10261 before the beginning of from). The second one involves inputs where
10262 the 'db' section contains nothing but zeroes (there is a one-byte
10263 invalid read after the end of 'db').
10265 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10267 * Partial backport from 0.9.9-dev:
10269 Introduce bn_mul_mont (dedicated Montgomery multiplication
10270 procedure) as a candidate for BIGNUM assembler implementation.
10271 While 0.9.9-dev uses assembler for various architectures, only
10272 x86_64 is available by default here in the 0.9.8 branch, and
10273 32-bit x86 is available through a compile-time setting.
10275 To try the 32-bit x86 assembler implementation, use Configure
10276 option "enable-montasm" (which exists only for this backport).
10278 As "enable-montasm" for 32-bit x86 disclaims code stability
10279 anyway, in this constellation we activate additional code
10280 backported from 0.9.9-dev for further performance improvements,
10281 namely BN_from_montgomery_word. (To enable this otherwise,
10282 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10284 *Andy Polyakov (backport partially by Bodo Moeller)*
10286 * Add TLS session ticket callback. This allows an application to set
10287 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10288 values. This is useful for key rollover for example where several key
10289 sets may exist with different names.
10293 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10294 This was broken until now in 0.9.8 releases, such that the only way
10295 a registered ENGINE could be used (assuming it initialises
10296 successfully on the host) was to explicitly set it as the default
10297 for the relevant algorithms. This is in contradiction with 0.9.7
10298 behaviour and the documentation. With this fix, when an ENGINE is
10299 registered into a given algorithm's table of implementations, the
10300 'uptodate' flag is reset so that auto-discovery will be used next
10301 time a new context for that algorithm attempts to select an
10304 *Ian Lister (tweaked by Geoff Thorpe)*
10306 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10307 implementation in the following ways:
10309 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10312 Lack of BER streaming support means one pass streaming processing is
10313 only supported if data is detached: setting the streaming flag is
10314 ignored for embedded content.
10316 CMS support is disabled by default and must be explicitly enabled
10317 with the enable-cms configuration option.
10321 * Update the GMP engine glue to do direct copies between BIGNUM and
10322 mpz_t when openssl and GMP use the same limb size. Otherwise the
10323 existing "conversion via a text string export" trick is still used.
10325 *Paul Sheer <paulsheer@gmail.com>*
10327 * Zlib compression BIO. This is a filter BIO which compressed and
10328 uncompresses any data passed through it.
10332 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10333 RFC3394 compatible AES key wrapping.
10337 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10338 sets string data without copying. X509_ALGOR_set0() and
10339 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10340 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10341 from an X509_ATTRIBUTE structure optionally checking it occurs only
10342 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10347 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10348 to get the expected BN_FLG_CONSTTIME behavior.
10350 *Bodo Moeller (Google)*
10354 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10355 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10356 - added some more tests to do_tests.pl
10357 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10358 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10359 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10360 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10361 - various changes to netware.pl to enable gcc-cross builds on Win32
10363 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10364 - various changes to fix missing prototype warnings
10365 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10366 - added AES, WHIRLPOOL and CPUID assembler code to build files
10367 - added missing AES assembler make rules to mk1mf.pl
10368 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10370 *Guenter Knauf <eflash@gmx.net>*
10372 * Implement certificate status request TLS extension defined in RFC3546.
10373 A client can set the appropriate parameters and receive the encoded
10374 OCSP response via a callback. A server can query the supplied parameters
10375 and set the encoded OCSP response in the callback. Add simplified examples
10376 to s_client and s_server.
10380 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10382 * Fix various bugs:
10383 + Binary incompatibility of ssl_ctx_st structure
10384 + DTLS interoperation with non-compliant servers
10385 + Don't call get_session_cb() without proposed session
10386 + Fix ia64 assembler code
10388 *Andy Polyakov, Steve Henson*
10390 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10392 * DTLS Handshake overhaul. There were longstanding issues with
10393 OpenSSL DTLS implementation, which were making it impossible for
10394 RFC 4347 compliant client to communicate with OpenSSL server.
10395 Unfortunately just fixing these incompatibilities would "cut off"
10396 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10397 server keeps tolerating non RFC compliant syntax. The opposite is
10398 not true, 0.9.8f client can not communicate with earlier server.
10399 This update even addresses CVE-2007-4995.
10403 * Changes to avoid need for function casts in OpenSSL: some compilers
10404 (gcc 4.2 and later) reject their use.
10405 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10408 * Add RFC4507 support to OpenSSL. This includes the corrections in
10409 RFC4507bis. The encrypted ticket format is an encrypted encoded
10410 SSL_SESSION structure, that way new session features are automatically
10413 If a client application caches session in an SSL_SESSION structure
10414 support is transparent because tickets are now stored in the encoded
10417 The SSL_CTX structure automatically generates keys for ticket
10418 protection in servers so again support should be possible
10419 with no application modification.
10421 If a client or server wishes to disable RFC4507 support then the option
10422 SSL_OP_NO_TICKET can be set.
10424 Add a TLS extension debugging callback to allow the contents of any client
10425 or server extensions to be examined.
10427 This work was sponsored by Google.
10431 * Add initial support for TLS extensions, specifically for the server_name
10432 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10433 have new members for a host name. The SSL data structure has an
10434 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10435 stored in that context to allow for session resumption, even after the
10436 SSL has been switched to a new SSL_CTX in reaction to a client's
10437 server_name extension.
10439 New functions (subject to change):
10441 SSL_get_servername()
10442 SSL_get_servername_type()
10445 New CTRL codes and macros (subject to change):
10447 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10448 - SSL_CTX_set_tlsext_servername_callback()
10449 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10450 - SSL_CTX_set_tlsext_servername_arg()
10451 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10453 openssl s_client has a new '-servername ...' option.
10455 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10456 '-key2 ...', '-servername_fatal' (subject to change). This allows
10457 testing the HostName extension for a specific single host name ('-cert'
10458 and '-key' remain fallbacks for handshakes without HostName
10459 negotiation). If the unrecognized_name alert has to be sent, this by
10460 default is a warning; it becomes fatal with the '-servername_fatal'
10463 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10465 * Add AES and SSE2 assembly language support to VC++ build.
10469 * Mitigate attack on final subtraction in Montgomery reduction.
10473 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10474 (which previously caused an internal error).
10478 * Squeeze another 10% out of IGE mode when in != out.
10482 * AES IGE mode speedup.
10484 *Dean Gaudet (Google)*
10486 * Add the Korean symmetric 128-bit cipher SEED (see
10487 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10488 add SEED ciphersuites from RFC 4162:
10490 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10491 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10492 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10493 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10495 To minimize changes between patchlevels in the OpenSSL 0.9.8
10496 series, SEED remains excluded from compilation unless OpenSSL
10497 is configured with 'enable-seed'.
10499 *KISA, Bodo Moeller*
10501 * Mitigate branch prediction attacks, which can be practical if a
10502 single processor is shared, allowing a spy process to extract
10503 information. For detailed background information, see
10504 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10505 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10506 and Necessary Software Countermeasures"). The core of the change
10507 are new versions BN_div_no_branch() and
10508 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10509 respectively, which are slower, but avoid the security-relevant
10510 conditional branches. These are automatically called by BN_div()
10511 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10512 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10513 remove a conditional branch.
10515 BN_FLG_CONSTTIME is the new name for the previous
10516 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10517 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10518 in the exponent causes BN_mod_exp_mont() to use the alternative
10519 implementation in BN_mod_exp_mont_consttime().) The old name
10520 remains as a deprecated alias.
10522 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10523 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10524 constant-time implementations for more than just exponentiation.
10525 Here too the old name is kept as a deprecated alias.
10527 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10528 the BN_BLINDING structure gets an independent copy of the
10529 modulus. This means that the previous `BIGNUM *m` argument to
10530 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10531 essentially becomes `const BIGNUM *m`, although we can't actually
10532 change this in the header file before 0.9.9. It allows
10533 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10534 enable BN_FLG_CONSTTIME.
10536 *Matthew D Wood (Intel Corp)*
10538 * In the SSL/TLS server implementation, be strict about session ID
10539 context matching (which matters if an application uses a single
10540 external cache for different purposes). Previously,
10541 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10542 set. This did ensure strict client verification, but meant that,
10543 with applications using a single external cache for quite
10544 different requirements, clients could circumvent ciphersuite
10545 restrictions for a given session ID context by starting a session
10546 in a different context.
10550 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10551 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10552 authentication-only ciphersuites.
10556 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10557 not complete and could lead to a possible single byte overflow
10558 ([CVE-2007-5135]) [Ben Laurie]
10560 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10562 * Since AES128 and AES256 (and similarly Camellia128 and
10563 Camellia256) share a single mask bit in the logic of
10564 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10565 kludge to work properly if AES128 is available and AES256 isn't
10566 (or if Camellia128 is available and Camellia256 isn't).
10570 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10571 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10572 When a point or a seed is encoded in a BIT STRING, we need to
10573 prevent the removal of trailing zero bits to get the proper DER
10574 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10575 of a NamedBitList, for which trailing 0 bits need to be removed.)
10579 * Have SSL/TLS server implementation tolerate "mismatched" record
10580 protocol version while receiving ClientHello even if the
10581 ClientHello is fragmented. (The server can't insist on the
10582 particular protocol version it has chosen before the ServerHello
10583 message has informed the client about his choice.)
10587 * Add RFC 3779 support.
10589 *Rob Austein for ARIN, Ben Laurie*
10591 * Load error codes if they are not already present instead of using a
10592 static variable. This allows them to be cleanly unloaded and reloaded.
10593 Improve header file function name parsing.
10597 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10598 or CAPABILITY handshake as required by RFCs.
10600 *Goetz Babin-Ebell*
10602 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10604 * Introduce limits to prevent malicious keys being able to
10605 cause a denial of service. ([CVE-2006-2940])
10607 *Steve Henson, Bodo Moeller*
10609 * Fix ASN.1 parsing of certain invalid structures that can result
10610 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10612 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10613 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10615 * Fix SSL client code which could crash if connecting to a
10616 malicious SSLv2 server. ([CVE-2006-4343])
10618 *Tavis Ormandy and Will Drewry, Google Security Team*
10620 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10621 match only those. Before that, "AES256-SHA" would be interpreted
10622 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10623 the same strength classification in 0.9.7h) as we currently only
10624 have a single AES bit in the ciphersuite description bitmap.
10625 That change, however, also applied to ciphersuite strings such as
10626 "RC4-MD5" that intentionally matched multiple ciphersuites --
10627 namely, SSL 2.0 ciphersuites in addition to the more common ones
10628 from SSL 3.0/TLS 1.0.
10630 So we change the selection algorithm again: Naming an explicit
10631 ciphersuite selects this one ciphersuite, and any other similar
10632 ciphersuite (same bitmap) from *other* protocol versions.
10633 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10634 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10636 Since SSL 2.0 does not have any ciphersuites for which the
10637 128/256 bit distinction would be relevant, this works for now.
10638 The proper fix will be to use different bits for AES128 and
10639 AES256, which would have avoided the problems from the beginning;
10640 however, bits are scarce, so we can only do this in a new release
10641 (not just a patchlevel) when we can change the SSL_CIPHER
10642 definition to split the single 'unsigned long mask' bitmap into
10643 multiple values to extend the available space.
10647 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10649 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10650 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10652 * Add AES IGE and biIGE modes.
10656 * Change the Unix randomness entropy gathering to use poll() when
10657 possible instead of select(), since the latter has some
10658 undesirable limitations.
10660 *Darryl Miles via Richard Levitte and Bodo Moeller*
10662 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10663 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10664 cannot be implicitly activated as part of, e.g., the "AES" alias.
10665 However, please upgrade to OpenSSL 0.9.9[-dev] for
10666 non-experimental use of the ECC ciphersuites to get TLS extension
10667 support, which is required for curve and point format negotiation
10668 to avoid potential handshake problems.
10672 * Disable rogue ciphersuites:
10674 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10675 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10676 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10678 The latter two were purportedly from
10679 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10682 Also deactivate the remaining ciphersuites from
10683 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10684 unofficial, and the ID has long expired.
10688 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10689 dual-core machines) and other potential thread-safety issues.
10693 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10694 versions), which is now available for royalty-free use
10695 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10696 Also, add Camellia TLS ciphersuites from RFC 4132.
10698 To minimize changes between patchlevels in the OpenSSL 0.9.8
10699 series, Camellia remains excluded from compilation unless OpenSSL
10700 is configured with 'enable-camellia'.
10704 * Disable the padding bug check when compression is in use. The padding
10705 bug check assumes the first packet is of even length, this is not
10706 necessarily true if compression is enabled and can result in false
10707 positives causing handshake failure. The actual bug test is ancient
10708 code so it is hoped that implementations will either have fixed it by
10709 now or any which still have the bug do not support compression.
10713 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10715 * When applying a cipher rule check to see if string match is an explicit
10716 cipher suite and only match that one cipher suite if it is.
10720 * Link in manifests for VC++ if needed.
10722 *Austin Ziegler <halostatue@gmail.com>*
10724 * Update support for ECC-based TLS ciphersuites according to
10725 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10726 TLS extensions, which are supported starting with the 0.9.9
10727 branch, not in the OpenSSL 0.9.8 branch).
10731 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10732 opaque EVP_CIPHER_CTX handling.
10736 * Fixes and enhancements to zlib compression code. We now only use
10737 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10738 to conform with the standards mentioned here:
10739 <http://www.zlib.net/DLL_FAQ.txt>
10740 Static zlib linking now works on Windows and the new --with-zlib-include
10741 --with-zlib-lib options to Configure can be used to supply the location
10742 of the headers and library. Gracefully handle case where zlib library
10747 * Several fixes and enhancements to the OID generation code. The old code
10748 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10749 handle numbers larger than ULONG_MAX, truncated printing and had a
10750 non standard OBJ_obj2txt() behaviour.
10754 * Add support for building of engines under engine/ as shared libraries
10755 under VC++ build system.
10759 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10760 Hopefully, we will not see any false combination of paths any more.
10764 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10766 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10767 (part of SSL_OP_ALL). This option used to disable the
10768 countermeasure against man-in-the-middle protocol-version
10769 rollback in the SSL 2.0 server implementation, which is a bad
10770 idea. ([CVE-2005-2969])
10772 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10773 for Information Security, National Institute of Advanced Industrial
10774 Science and Technology [AIST], Japan)*
10776 * Add two function to clear and return the verify parameter flags.
10780 * Keep cipherlists sorted in the source instead of sorting them at
10781 runtime, thus removing the need for a lock.
10785 * Avoid some small subgroup attacks in Diffie-Hellman.
10787 *Nick Mathewson and Ben Laurie*
10789 * Add functions for well-known primes.
10793 * Extended Windows CE support.
10795 *Satoshi Nakamura and Andy Polyakov*
10797 * Initialize SSL_METHOD structures at compile time instead of during
10798 runtime, thus removing the need for a lock.
10802 * Make PKCS7_decrypt() work even if no certificate is supplied by
10803 attempting to decrypt each encrypted key in turn. Add support to
10808 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10810 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10813 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10817 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10818 key into the same file any more.
10822 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10826 * Add -utf8 command line and config file option to 'ca'.
10828 *Stefan <stf@udoma.org*
10830 * Removed the macro des_crypt(), as it seems to conflict with some
10831 libraries. Use DES_crypt().
10835 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10836 involves renaming the source and generated shared-libs for
10837 both. The engines will accept the corrected or legacy ids
10838 ('ncipher' and '4758_cca' respectively) when binding. NB,
10839 this only applies when building 'shared'.
10841 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10843 * Add attribute functions to EVP_PKEY structure. Modify
10844 PKCS12_create() to recognize a CSP name attribute and
10845 use it. Make -CSP option work again in pkcs12 utility.
10849 * Add new functionality to the bn blinding code:
10850 - automatic re-creation of the BN_BLINDING parameters after
10851 a fixed number of uses (currently 32)
10852 - add new function for parameter creation
10853 - introduce flags to control the update behaviour of the
10854 BN_BLINDING parameters
10855 - hide BN_BLINDING structure
10856 Add a second BN_BLINDING slot to the RSA structure to improve
10857 performance when a single RSA object is shared among several
10862 * Add support for DTLS.
10864 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10866 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10867 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10871 * Remove buggy and incomplete DH cert support from
10872 ssl/ssl_rsa.c and ssl/s3_both.c
10876 * Use SHA-1 instead of MD5 as the default digest algorithm for
10877 the `apps/openssl` commands.
10881 * Compile clean with "-Wall -Wmissing-prototypes
10882 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10883 DEBUG_SAFESTACK must also be set.
10887 * Change ./Configure so that certain algorithms can be disabled by default.
10888 The new counterpiece to "no-xxx" is "enable-xxx".
10890 The patented RC5 and MDC2 algorithms will now be disabled unless
10891 "enable-rc5" and "enable-mdc2", respectively, are specified.
10893 (IDEA remains enabled despite being patented. This is because IDEA
10894 is frequently required for interoperability, and there is no license
10895 fee for non-commercial use. As before, "no-idea" can be used to
10896 avoid this algorithm.)
10900 * Add processing of proxy certificates (see RFC 3820). This work was
10901 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10902 EGEE (Enabling Grids for E-science in Europe).
10906 * RC4 performance overhaul on modern architectures/implementations, such
10907 as Intel P4, IA-64 and AMD64.
10911 * New utility extract-section.pl. This can be used specify an alternative
10912 section number in a pod file instead of having to treat each file as
10913 a separate case in Makefile. This can be done by adding two lines to the
10916 =for comment openssl_section:XXX
10918 The blank line is mandatory.
10922 * New arguments -certform, -keyform and -pass for s_client and s_server
10923 to allow alternative format key and certificate files and passphrase
10928 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10929 update associated structures and add various utility functions.
10931 Add new policy related verify parameters, include policy checking in
10932 standard verify code. Enhance 'smime' application with extra parameters
10933 to support policy checking and print out.
10937 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10938 Nehemiah processors. These extensions support AES encryption in hardware
10939 as well as RNG (though RNG support is currently disabled).
10941 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10943 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10947 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10949 *Andy Polyakov and a number of other people*
10951 * Improved PowerPC platform support. Most notably BIGNUM assembler
10952 implementation contributed by IBM.
10954 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10956 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10957 exponent rather than 'unsigned long'. There is a corresponding change to
10958 the new 'rsa_keygen' element of the RSA_METHOD structure.
10960 *Jelte Jansen, Geoff Thorpe*
10962 * Functionality for creating the initial serial number file is now
10963 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10965 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10966 number file to 1, which is bound to cause problems. To avoid
10967 the problems while respecting compatibility between different 0.9.7
10968 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10969 CA.pl for serial number initialization. With the new release 0.9.8,
10970 we can fix the problem directly in the 'ca' utility.)
10974 * Reduced header interdependencies by declaring more opaque objects in
10975 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10976 give fewer recursive includes, which could break lazy source code - so
10977 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10978 developers should define this symbol when building and using openssl to
10979 ensure they track the recommended behaviour, interfaces, [etc], but
10980 backwards-compatible behaviour prevails when this isn't defined.
10984 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10988 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10989 This will generate a random key of the appropriate length based on the
10990 cipher context. The EVP_CIPHER can provide its own random key generation
10991 routine to support keys of a specific form. This is used in the des and
10992 3des routines to generate a key of the correct parity. Update S/MIME
10993 code to use new functions and hence generate correct parity DES keys.
10994 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10995 valid (weak or incorrect parity).
10999 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11000 as looking them up. This is useful when the verified structure may contain
11001 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11002 present unless the new PKCS7_NO_CRL flag is asserted.
11006 * Extend ASN1 oid configuration module. It now additionally accepts the
11009 shortName = some long name, 1.2.3.4
11013 * Reimplemented the BN_CTX implementation. There is now no more static
11014 limitation on the number of variables it can handle nor the depth of the
11015 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11016 information can now expand as required, and rather than having a single
11017 static array of bignums, BN_CTX now uses a linked-list of such arrays
11018 allowing it to expand on demand whilst maintaining the usefulness of
11019 BN_CTX's "bundling".
11023 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11024 to allow all RSA operations to function using a single BN_CTX.
11028 * Preliminary support for certificate policy evaluation and checking. This
11029 is initially intended to pass the tests outlined in "Conformance Testing
11030 of Relying Party Client Certificate Path Processing Logic" v1.07.
11034 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11035 remained unused and not that useful. A variety of other little bignum
11036 tweaks and fixes have also been made continuing on from the audit (see
11041 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11042 associated ASN1, EVP and SSL functions and old ASN1 macros.
11046 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11047 and this should never fail. So the return value from the use of
11048 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11049 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11053 * BN_CTX_get() should return zero-valued bignums, providing the same
11054 initialised value as BN_new().
11056 *Geoff Thorpe, suggested by Ulf Möller*
11058 * Support for inhibitAnyPolicy certificate extension.
11062 * An audit of the BIGNUM code is underway, for which debugging code is
11063 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11064 is considered valid when processing BIGNUMs, and causes execution to
11065 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11066 further steps are taken to deliberately pollute unused data in BIGNUM
11067 structures to try and expose faulty code further on. For now, openssl will
11068 (in its default mode of operation) continue to tolerate the inconsistent
11069 forms that it has tolerated in the past, but authors and packagers should
11070 consider trying openssl and their own applications when compiled with
11071 these debugging symbols defined. It will help highlight potential bugs in
11072 their own code, and will improve the test coverage for OpenSSL itself. At
11073 some point, these tighter rules will become openssl's default to improve
11074 maintainability, though the assert()s and other overheads will remain only
11075 in debugging configurations. See bn.h for more details.
11077 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11079 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11080 that can only be obtained through BN_CTX_new() (which implicitly
11081 initialises it). The presence of this function only made it possible
11082 to overwrite an existing structure (and cause memory leaks).
11086 * Because of the callback-based approach for implementing LHASH as a
11087 template type, lh_insert() adds opaque objects to hash-tables and
11088 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11089 to clean up those corresponding objects before destroying the hash table
11090 (and losing the object pointers). So some over-zealous constifications in
11091 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11092 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11093 prototyped to have "const" restrictions on the object pointers they are
11094 given (and so aren't required to cast them away any more).
11098 * The tmdiff.h API was so ugly and minimal that our own timing utility
11099 (speed) prefers to use its own implementation. The two implementations
11100 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11101 its object type properly exposed (MS_TM) instead of casting to/from
11102 `char *`. This may still change yet if someone realises MS_TM and
11104 aren't necessarily the greatest nomenclatures - but this is what was used
11105 internally to the implementation so I've used that for now.
11109 * Ensure that deprecated functions do not get compiled when
11110 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11111 the self-tests were still using deprecated key-generation functions so
11112 these have been updated also.
11116 * Reorganise PKCS#7 code to separate the digest location functionality
11117 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11118 New function PKCS7_set_digest() to set the digest type for PKCS#7
11119 digestedData type. Add additional code to correctly generate the
11120 digestedData type and add support for this type in PKCS7 initialization
11125 * New function PKCS7_set0_type_other() this initializes a PKCS7
11126 structure of type "other".
11130 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11131 sure the loop does correctly stop and breaking ("division by zero")
11132 modulus operations are not performed. The (pre-generated) prime
11133 table crypto/bn/bn_prime.h was already correct, but it could not be
11134 re-generated on some platforms because of the "division by zero"
11135 situation in the script.
11137 *Ralf S. Engelschall*
11139 * Update support for ECC-based TLS ciphersuites according to
11140 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11141 SHA-1 now is only used for "small" curves (where the
11142 representation of a field element takes up to 24 bytes); for
11143 larger curves, the field element resulting from ECDH is directly
11144 used as premaster secret.
11146 *Douglas Stebila (Sun Microsystems Laboratories)*
11148 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11149 curve secp160r1 to the tests.
11151 *Douglas Stebila (Sun Microsystems Laboratories)*
11153 * Add the possibility to load symbols globally with DSO.
11155 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11157 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11158 control of the error stack.
11162 * Add support for STORE in ENGINE.
11166 * Add the STORE type. The intention is to provide a common interface
11167 to certificate and key stores, be they simple file-based stores, or
11168 HSM-type store, or LDAP stores, or...
11169 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11173 * Add a generic structure called OPENSSL_ITEM. This can be used to
11174 pass a list of arguments to any function as well as provide a way
11175 for a function to pass data back to the caller.
11179 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11180 works like BUF_strdup() but can be used to duplicate a portion of
11181 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11186 * Add the function sk_find_ex() which works like sk_find(), but will
11187 return an index to an element even if an exact match couldn't be
11188 found. The index is guaranteed to point at the element where the
11189 searched-for key would be inserted to preserve sorting order.
11193 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11194 takes an extra flags argument for optional functionality. Currently,
11195 the following flags are defined:
11197 OBJ_BSEARCH_VALUE_ON_NOMATCH
11198 This one gets OBJ_bsearch_ex() to return a pointer to the first
11199 element where the comparing function returns a negative or zero
11202 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11203 This one gets OBJ_bsearch_ex() to return a pointer to the first
11204 element where the comparing function returns zero. This is useful
11205 if there are more than one element where the comparing function
11210 * Make it possible to create self-signed certificates with 'openssl ca'
11211 in such a way that the self-signed certificate becomes part of the
11212 CA database and uses the same mechanisms for serial number generation
11213 as all other certificate signing. The new flag '-selfsign' enables
11214 this functionality. Adapt CA.sh and CA.pl.in.
11218 * Add functionality to check the public key of a certificate request
11219 against a given private. This is useful to check that a certificate
11220 request can be signed by that key (self-signing).
11224 * Make it possible to have multiple active certificates with the same
11225 subject in the CA index file. This is done only if the keyword
11226 'unique_subject' is set to 'no' in the main CA section (default
11227 if 'CA_default') of the configuration file. The value is saved
11228 with the database itself in a separate index attribute file,
11229 named like the index file with '.attr' appended to the name.
11233 * Generate multi-valued AVAs using '+' notation in config files for
11238 * Support for nameConstraints certificate extension.
11242 * Support for policyConstraints certificate extension.
11246 * Support for policyMappings certificate extension.
11250 * Make sure the default DSA_METHOD implementation only uses its
11251 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11252 and change its own handlers to be NULL so as to remove unnecessary
11253 indirection. This lets alternative implementations fallback to the
11254 default implementation more easily.
11258 * Support for directoryName in GeneralName related extensions
11263 * Make it possible to link applications using Makefile.shared.
11264 Make that possible even when linking against static libraries!
11268 * Support for single pass processing for S/MIME signing. This now
11269 means that S/MIME signing can be done from a pipe, in addition
11270 cleartext signing (multipart/signed type) is effectively streaming
11271 and the signed data does not need to be all held in memory.
11273 This is done with a new flag PKCS7_STREAM. When this flag is set
11274 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11275 is done after the data is output (and digests calculated) in
11276 SMIME_write_PKCS7().
11280 * Add full support for -rpath/-R, both in shared libraries and
11281 applications, at least on the platforms where it's known how
11286 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11287 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11288 will now compute a table of multiples of the generator that
11289 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11290 faster (notably in the case of a single point multiplication,
11291 scalar * generator).
11293 *Nils Larsch, Bodo Moeller*
11295 * IPv6 support for certificate extensions. The various extensions
11296 which use the IP:a.b.c.d can now take IPv6 addresses using the
11297 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11302 * Added an ENGINE that implements RSA by performing private key
11303 exponentiations with the GMP library. The conversions to and from
11304 GMP's mpz_t format aren't optimised nor are any montgomery forms
11305 cached, and on x86 it appears OpenSSL's own performance has caught up.
11306 However there are likely to be other architectures where GMP could
11307 provide a boost. This ENGINE is not built in by default, but it can be
11308 specified at Configure time and should be accompanied by the necessary
11309 linker additions, eg;
11310 ./config -DOPENSSL_USE_GMP -lgmp
11314 * "openssl engine" will not display ENGINE/DSO load failure errors when
11315 testing availability of engines with "-t" - the old behaviour is
11316 produced by increasing the feature's verbosity with "-tt".
11320 * ECDSA routines: under certain error conditions uninitialized BN objects
11321 could be freed. Solution: make sure initialization is performed early
11322 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11327 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11328 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11329 software implementations. For DSA and DH, parameter generation can
11330 also be overridden by providing the appropriate method callbacks.
11334 * Change the "progress" mechanism used in key-generation and
11335 primality testing to functions that take a new BN_GENCB pointer in
11336 place of callback/argument pairs. The new API functions have `_ex`
11337 postfixes and the older functions are reimplemented as wrappers for
11338 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11339 declarations of the old functions to help (graceful) attempts to
11340 migrate to the new functions. Also, the new key-generation API
11341 functions operate on a caller-supplied key-structure and return
11342 success/failure rather than returning a key or NULL - this is to
11343 help make "keygen" another member function of RSA_METHOD etc.
11345 Example for using the new callback interface:
11347 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11348 void *my_arg = ...;
11351 BN_GENCB_set(&my_cb, my_callback, my_arg);
11353 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11354 /* For the meaning of a, b in calls to my_callback(), see the
11355 * documentation of the function that calls the callback.
11356 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11357 * my_callback should return 1 if it wants BN_is_prime_ex()
11358 * to continue, or 0 to stop.
11363 * Change the ZLIB compression method to be stateful, and make it
11364 available to TLS with the number defined in
11365 draft-ietf-tls-compression-04.txt.
11369 * Add the ASN.1 structures and functions for CertificatePair, which
11370 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11372 CertificatePair ::= SEQUENCE {
11373 forward [0] Certificate OPTIONAL,
11374 reverse [1] Certificate OPTIONAL,
11375 -- at least one of the pair shall be present -- }
11377 Also implement the PEM functions to read and write certificate
11378 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11380 This needed to be defined, mostly for the sake of the LDAP
11381 attribute crossCertificatePair, but may prove useful elsewhere as
11386 * Make it possible to inhibit symlinking of shared libraries in
11387 Makefile.shared, for Cygwin's sake.
11391 * Extend the BIGNUM API by creating a function
11392 void BN_set_negative(BIGNUM *a, int neg);
11393 and a macro that behave like
11394 int BN_is_negative(const BIGNUM *a);
11396 to avoid the need to access 'a->neg' directly in applications.
11400 * Implement fast modular reduction for pseudo-Mersenne primes
11401 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11402 EC_GROUP_new_curve_GFp() will now automatically use this
11405 *Nils Larsch <nla@trustcenter.de>*
11407 * Add new lock type (CRYPTO_LOCK_BN).
11411 * Change the ENGINE framework to automatically load engines
11412 dynamically from specific directories unless they could be
11413 found to already be built in or loaded. Move all the
11414 current engines except for the cryptodev one to a new
11415 directory engines/.
11416 The engines in engines/ are built as shared libraries if
11417 the "shared" options was given to ./Configure or ./config.
11418 Otherwise, they are inserted in libcrypto.a.
11419 /usr/local/ssl/engines is the default directory for dynamic
11420 engines, but that can be overridden at configure time through
11421 the usual use of --prefix and/or --openssldir, and at run
11422 time with the environment variable OPENSSL_ENGINES.
11424 *Geoff Thorpe and Richard Levitte*
11426 * Add Makefile.shared, a helper makefile to build shared
11427 libraries. Adapt Makefile.org.
11431 * Add version info to Win32 DLLs.
11433 *Peter 'Luna' Runestig" <peter@runestig.com>*
11435 * Add new 'medium level' PKCS#12 API. Certificates and keys
11436 can be added using this API to created arbitrary PKCS#12
11437 files while avoiding the low-level API.
11439 New options to PKCS12_create(), key or cert can be NULL and
11440 will then be omitted from the output file. The encryption
11441 algorithm NIDs can be set to -1 for no encryption, the mac
11442 iteration count can be set to 0 to omit the mac.
11444 Enhance pkcs12 utility by making the -nokeys and -nocerts
11445 options work when creating a PKCS#12 file. New option -nomac
11446 to omit the mac, NONE can be set for an encryption algorithm.
11447 New code is modified to use the enhanced PKCS12_create()
11448 instead of the low-level API.
11452 * Extend ASN1 encoder to support indefinite length constructed
11453 encoding. This can output sequences tags and octet strings in
11454 this form. Modify pk7_asn1.c to support indefinite length
11455 encoding. This is experimental and needs additional code to
11456 be useful, such as an ASN1 bio and some enhanced streaming
11459 Extend template encode functionality so that tagging is passed
11460 down to the template encoder.
11464 * Let 'openssl req' fail if an argument to '-newkey' is not
11465 recognized instead of using RSA as a default.
11469 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11470 As these are not official, they are not included in "ALL";
11471 the "ECCdraft" ciphersuite group alias can be used to select them.
11473 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11475 * Add ECDH engine support.
11477 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11479 * Add ECDH in new directory crypto/ecdh/.
11481 *Douglas Stebila (Sun Microsystems Laboratories)*
11483 * Let BN_rand_range() abort with an error after 100 iterations
11484 without success (which indicates a broken PRNG).
11488 * Change BN_mod_sqrt() so that it verifies that the input value
11489 is really the square of the return value. (Previously,
11490 BN_mod_sqrt would show GIGO behaviour.)
11494 * Add named elliptic curves over binary fields from X9.62, SECG,
11495 and WAP/WTLS; add OIDs that were still missing.
11497 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11499 * Extend the EC library for elliptic curves over binary fields
11500 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11503 EC_GF2m_simple_method
11507 EC_GROUP_new_curve_GF2m
11508 EC_GROUP_set_curve_GF2m
11509 EC_GROUP_get_curve_GF2m
11510 EC_POINT_set_affine_coordinates_GF2m
11511 EC_POINT_get_affine_coordinates_GF2m
11512 EC_POINT_set_compressed_coordinates_GF2m
11514 Point compression for binary fields is disabled by default for
11515 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11518 As binary polynomials are represented as BIGNUMs, various members
11519 of the EC_GROUP and EC_POINT data structures can be shared
11520 between the implementations for prime fields and binary fields;
11521 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11522 are essentially identical to their `..._GFp` counterparts.
11523 (For simplicity, the `..._GFp` prefix has been dropped from
11524 various internal method names.)
11526 An internal 'field_div' method (similar to 'field_mul' and
11527 'field_sqr') has been added; this is used only for binary fields.
11529 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11531 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11532 through methods ('mul', 'precompute_mult').
11534 The generic implementations (now internally called 'ec_wNAF_mul'
11535 and 'ec_wNAF_precomputed_mult') remain the default if these
11536 methods are undefined.
11538 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11540 * New function EC_GROUP_get_degree, which is defined through
11541 EC_METHOD. For curves over prime fields, this returns the bit
11542 length of the modulus.
11544 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11546 * New functions EC_GROUP_dup, EC_POINT_dup.
11547 (These simply call ..._new and ..._copy).
11549 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11551 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11552 Polynomials are represented as BIGNUMs (where the sign bit is not
11553 used) in the following functions [macros]:
11556 BN_GF2m_sub [= BN_GF2m_add]
11557 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11558 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11559 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11561 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11562 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11563 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11564 BN_GF2m_cmp [= BN_ucmp]
11566 (Note that only the 'mod' functions are actually for fields GF(2^m).
11567 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11569 For some functions, an the irreducible polynomial defining a
11570 field can be given as an 'unsigned int[]' with strictly
11571 decreasing elements giving the indices of those bits that are set;
11572 i.e., p[] represents the polynomial
11573 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11575 p[0] > p[1] > ... > p[k] = 0.
11576 This applies to the following functions:
11579 BN_GF2m_mod_mul_arr
11580 BN_GF2m_mod_sqr_arr
11581 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11582 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11583 BN_GF2m_mod_exp_arr
11584 BN_GF2m_mod_sqrt_arr
11585 BN_GF2m_mod_solve_quad_arr
11589 Conversion can be performed by the following functions:
11594 bntest.c has additional tests for binary polynomial arithmetic.
11596 Two implementations for BN_GF2m_mod_div() are available.
11597 The default algorithm simply uses BN_GF2m_mod_inv() and
11598 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11599 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11600 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11602 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11604 * Add new error code 'ERR_R_DISABLED' that can be used when some
11605 functionality is disabled at compile-time.
11607 *Douglas Stebila <douglas.stebila@sun.com>*
11609 * Change default behaviour of 'openssl asn1parse' so that more
11610 information is visible when viewing, e.g., a certificate:
11612 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11613 mode the content of non-printable OCTET STRINGs is output in a
11614 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11615 avoid the appearance of a printable string.
11617 *Nils Larsch <nla@trustcenter.de>*
11619 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11621 EC_GROUP_set_asn1_flag()
11622 EC_GROUP_get_asn1_flag()
11623 EC_GROUP_set_point_conversion_form()
11624 EC_GROUP_get_point_conversion_form()
11625 These control ASN1 encoding details:
11626 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11627 has been set to OPENSSL_EC_NAMED_CURVE.
11628 - Points are encoded in uncompressed form by default; options for
11629 asn1_for are as for point2oct, namely
11630 POINT_CONVERSION_COMPRESSED
11631 POINT_CONVERSION_UNCOMPRESSED
11632 POINT_CONVERSION_HYBRID
11634 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11636 EC_GROUP_set_seed()
11637 EC_GROUP_get0_seed()
11638 EC_GROUP_get_seed_len()
11639 This is used only for ASN1 purposes (so far).
11641 *Nils Larsch <nla@trustcenter.de>*
11643 * Add 'field_type' member to EC_METHOD, which holds the NID
11644 of the appropriate field type OID. The new function
11645 EC_METHOD_get_field_type() returns this value.
11647 *Nils Larsch <nla@trustcenter.de>*
11650 EC_POINT_point2bn()
11651 EC_POINT_bn2point()
11652 EC_POINT_point2hex()
11653 EC_POINT_hex2point()
11654 providing useful interfaces to EC_POINT_point2oct() and
11655 EC_POINT_oct2point().
11657 *Nils Larsch <nla@trustcenter.de>*
11659 * Change internals of the EC library so that the functions
11660 EC_GROUP_set_generator()
11661 EC_GROUP_get_generator()
11662 EC_GROUP_get_order()
11663 EC_GROUP_get_cofactor()
11664 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11665 to methods, which would lead to unnecessary code duplication when
11666 adding different types of curves.
11668 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11670 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11671 arithmetic, and such that modified wNAFs are generated
11672 (which avoid length expansion in many cases).
11676 * Add a function EC_GROUP_check_discriminant() (defined via
11677 EC_METHOD) that verifies that the curve discriminant is non-zero.
11679 Add a function EC_GROUP_check() that makes some sanity tests
11680 on a EC_GROUP, its generator and order. This includes
11681 EC_GROUP_check_discriminant().
11683 *Nils Larsch <nla@trustcenter.de>*
11685 * Add ECDSA in new directory crypto/ecdsa/.
11687 Add applications 'openssl ecparam' and 'openssl ecdsa'
11688 (these are based on 'openssl dsaparam' and 'openssl dsa').
11690 ECDSA support is also included in various other files across the
11691 library. Most notably,
11692 - 'openssl req' now has a '-newkey ecdsa:file' option;
11693 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11694 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11695 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11696 them suitable for ECDSA where domain parameters must be
11697 extracted before the specific public key;
11698 - ECDSA engine support has been added.
11700 *Nils Larsch <nla@trustcenter.de>*
11702 * Include some named elliptic curves, and add OIDs from X9.62,
11703 SECG, and WAP/WTLS. Each curve can be obtained from the new
11705 EC_GROUP_new_by_curve_name(),
11706 and the list of available named curves can be obtained with
11707 EC_get_builtin_curves().
11708 Also add a 'curve_name' member to EC_GROUP objects, which can be
11710 EC_GROUP_set_curve_name()
11711 EC_GROUP_get_curve_name()
11713 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11715 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11716 was actually never needed) and in BN_mul(). The removal in BN_mul()
11717 required a small change in bn_mul_part_recursive() and the addition
11718 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11719 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11720 bn_sub_words() and bn_add_words() except they take arrays with
11725 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11727 * Cleanse PEM buffers before freeing them since they may contain
11730 *Benjamin Bennett <ben@psc.edu>*
11732 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11733 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11734 authentication-only ciphersuites.
11738 * Since AES128 and AES256 share a single mask bit in the logic of
11739 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11740 kludge to work properly if AES128 is available and AES256 isn't.
11744 * Expand security boundary to match 1.1.1 module.
11748 * Remove redundant features: hash file source, editing of test vectors
11749 modify fipsld to use external fips_premain.c signature.
11753 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11754 run algorithm test programs.
11758 * Make algorithm test programs more tolerant of whitespace.
11762 * Have SSL/TLS server implementation tolerate "mismatched" record
11763 protocol version while receiving ClientHello even if the
11764 ClientHello is fragmented. (The server can't insist on the
11765 particular protocol version it has chosen before the ServerHello
11766 message has informed the client about his choice.)
11770 * Load error codes if they are not already present instead of using a
11771 static variable. This allows them to be cleanly unloaded and reloaded.
11775 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11777 * Introduce limits to prevent malicious keys being able to
11778 cause a denial of service. ([CVE-2006-2940])
11780 *Steve Henson, Bodo Moeller*
11782 * Fix ASN.1 parsing of certain invalid structures that can result
11783 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11785 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11786 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11788 * Fix SSL client code which could crash if connecting to a
11789 malicious SSLv2 server. ([CVE-2006-4343])
11791 *Tavis Ormandy and Will Drewry, Google Security Team*
11793 * Change ciphersuite string processing so that an explicit
11794 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11795 will no longer include "AES128-SHA"), and any other similar
11796 ciphersuite (same bitmap) from *other* protocol versions (so that
11797 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11798 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11799 changes from 0.9.8b and 0.9.8d.
11803 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11805 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11806 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11808 * Change the Unix randomness entropy gathering to use poll() when
11809 possible instead of select(), since the latter has some
11810 undesirable limitations.
11812 *Darryl Miles via Richard Levitte and Bodo Moeller*
11814 * Disable rogue ciphersuites:
11816 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11817 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11818 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11820 The latter two were purportedly from
11821 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11824 Also deactivate the remaining ciphersuites from
11825 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11826 unofficial, and the ID has long expired.
11830 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11831 dual-core machines) and other potential thread-safety issues.
11835 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11837 * Adapt fipsld and the build system to link against the validated FIPS
11838 module in FIPS mode.
11842 * Fixes for VC++ 2005 build under Windows.
11846 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11847 from a Windows bash shell such as MSYS. It is autodetected from the
11848 "config" script when run from a VC++ environment. Modify standard VC++
11849 build to use fipscanister.o from the GNU make build.
11853 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11855 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11856 The value now differs depending on if you build for FIPS or not.
11857 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11858 safely run with a non-FIPSed libcrypto, as it may crash because of
11859 the difference induced by this change.
11863 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11865 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11866 (part of SSL_OP_ALL). This option used to disable the
11867 countermeasure against man-in-the-middle protocol-version
11868 rollback in the SSL 2.0 server implementation, which is a bad
11869 idea. ([CVE-2005-2969])
11871 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11872 for Information Security, National Institute of Advanced Industrial
11873 Science and Technology [AIST, Japan)]*
11875 * Minimal support for X9.31 signatures and PSS padding modes. This is
11876 mainly for FIPS compliance and not fully integrated at this stage.
11880 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11881 the exponentiation using a fixed-length exponent. (Otherwise,
11882 the information leaked through timing could expose the secret key
11883 after many signatures; cf. Bleichenbacher's attack on DSA with
11888 * Make a new fixed-window mod_exp implementation the default for
11889 RSA, DSA, and DH private-key operations so that the sequence of
11890 squares and multiplies and the memory access pattern are
11891 independent of the particular secret key. This will mitigate
11892 cache-timing and potential related attacks.
11894 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11895 and this is automatically used by BN_mod_exp_mont() if the new flag
11896 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11897 will use this BN flag for private exponents unless the flag
11898 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11899 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11901 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11903 * Change the client implementation for SSLv23_method() and
11904 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11905 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11906 (Previously, the SSL 2.0 backwards compatible Client Hello
11907 message format would be used even with SSL_OP_NO_SSLv2.)
11911 * Add support for smime-type MIME parameter in S/MIME messages which some
11916 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11917 a threadsafe manner. Modify rsa code to use new function and add calls
11918 to dsa and dh code (which had race conditions before).
11922 * Include the fixed error library code in the C error file definitions
11923 instead of fixing them up at runtime. This keeps the error code
11924 structures constant.
11928 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11930 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11933 * Fixes for newer kerberos headers. NB: the casts are needed because
11934 the 'length' field is signed on one version and unsigned on another
11935 with no (?) obvious way to tell the difference, without these VC++
11936 complains. Also the "definition" of FAR (blank) is no longer included
11937 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11938 some needed definitions.
11942 * Undo Cygwin change.
11946 * Added support for proxy certificates according to RFC 3820.
11947 Because they may be a security thread to unaware applications,
11948 they must be explicitly allowed in run-time. See
11949 docs/HOWTO/proxy_certificates.txt for further information.
11953 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11955 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11956 server and client random values. Previously
11957 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11958 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11960 This change has negligible security impact because:
11962 1. Server and client random values still have 24 bytes of pseudo random
11965 2. Server and client random values are sent in the clear in the initial
11968 3. The master secret is derived using the premaster secret (48 bytes in
11969 size for static RSA ciphersuites) as well as client server and random
11972 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11975 *Stephen Henson, reported by UK NISCC*
11977 * Use Windows randomness collection on Cygwin.
11981 * Fix hang in EGD/PRNGD query when communication socket is closed
11982 prematurely by EGD/PRNGD.
11984 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11986 * Prompt for pass phrases when appropriate for PKCS12 input format.
11990 * Back-port of selected performance improvements from development
11991 branch, as well as improved support for PowerPC platforms.
11995 * Add lots of checks for memory allocation failure, error codes to indicate
11996 failure and freeing up memory if a failure occurs.
11998 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12000 * Add new -passin argument to dgst.
12004 * Perform some character comparisons of different types in X509_NAME_cmp:
12005 this is needed for some certificates that re-encode DNs into UTF8Strings
12006 (in violation of RFC3280) and can't or won't issue name rollover
12011 * Make an explicit check during certificate validation to see that
12012 the CA setting in each certificate on the chain is correct. As a
12013 side effect always do the following basic checks on extensions,
12014 not just when there's an associated purpose to the check:
12016 - if there is an unhandled critical extension (unless the user
12017 has chosen to ignore this fault)
12018 - if the path length has been exceeded (if one is set at all)
12019 - that certain extensions fit the associated purpose (if one has
12024 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12026 * Avoid a race condition when CRLs are checked in a multi threaded
12027 environment. This would happen due to the reordering of the revoked
12028 entries during signature checking and serial number lookup. Now the
12029 encoding is cached and the serial number sort performed under a lock.
12030 Add new STACK function sk_is_sorted().
12034 * Add Delta CRL to the extension code.
12038 * Various fixes to s3_pkt.c so alerts are sent properly.
12040 *David Holmes <d.holmes@f5.com>*
12042 * Reduce the chances of duplicate issuer name and serial numbers (in
12043 violation of RFC3280) using the OpenSSL certificate creation utilities.
12044 This is done by creating a random 64 bit value for the initial serial
12045 number when a serial number file is created or when a self signed
12046 certificate is created using 'openssl req -x509'. The initial serial
12047 number file is created using 'openssl x509 -next_serial' in CA.pl
12048 rather than being initialized to 1.
12052 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12054 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12055 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12057 *Joe Orton, Steve Henson*
12059 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12062 *Joe Orton, Steve Henson*
12064 * Make it possible to have multiple active certificates with the same
12065 subject in the CA index file. This is done only if the keyword
12066 'unique_subject' is set to 'no' in the main CA section (default
12067 if 'CA_default') of the configuration file. The value is saved
12068 with the database itself in a separate index attribute file,
12069 named like the index file with '.attr' appended to the name.
12073 * X509 verify fixes. Disable broken certificate workarounds when
12074 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12075 keyUsage extension present. Don't accept CRLs with unhandled critical
12076 extensions: since verify currently doesn't process CRL extensions this
12077 rejects a CRL with *any* critical extensions. Add new verify error codes
12082 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12083 A clarification of RFC2560 will require the use of OCTET STRINGs and
12084 some implementations cannot handle the current raw format. Since OpenSSL
12085 copies and compares OCSP nonces as opaque blobs without any attempt at
12086 parsing them this should not create any compatibility issues.
12090 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12091 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12092 this HMAC (and other) operations are several times slower than OpenSSL
12097 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12099 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12101 * Use the correct content when signing type "other".
12105 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12107 * Fix various bugs revealed by running the NISCC test suite:
12109 Stop out of bounds reads in the ASN1 code when presented with
12110 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12112 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12114 If verify callback ignores invalid public key errors don't try to check
12115 certificate signature with the NULL public key.
12119 * New -ignore_err option in ocsp application to stop the server
12120 exiting on the first error in a request.
12124 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12125 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12130 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12131 extra data after the compression methods not only for TLS 1.0
12132 but also for SSL 3.0 (as required by the specification).
12134 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12136 * Change X509_certificate_type() to mark the key as exported/exportable
12137 when it's 512 *bits* long, not 512 bytes.
12141 * Change AES_cbc_encrypt() so it outputs exact multiple of
12142 blocks during encryption.
12146 * Various fixes to base64 BIO and non blocking I/O. On write
12147 flushes were not handled properly if the BIO retried. On read
12148 data was not being buffered properly and had various logic bugs.
12149 This also affects blocking I/O when the data being decoded is a
12154 * Various S/MIME bugfixes and compatibility changes:
12155 output correct application/pkcs7 MIME type if
12156 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12157 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12158 of files as .eml work). Correctly handle very long lines in MIME
12163 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12165 * Countermeasure against the Klima-Pokorny-Rosa extension of
12166 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12167 a protocol version number mismatch like a decryption error
12168 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12172 * Turn on RSA blinding by default in the default implementation
12173 to avoid a timing attack. Applications that don't want it can call
12174 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12175 They would be ill-advised to do so in most cases.
12177 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12179 * Change RSA blinding code so that it works when the PRNG is not
12180 seeded (in this case, the secret RSA exponent is abused as
12181 an unpredictable seed -- if it is not unpredictable, there
12182 is no point in blinding anyway). Make RSA blinding thread-safe
12183 by remembering the creator's thread ID in rsa->blinding and
12184 having all other threads use local one-time blinding factors
12185 (this requires more computation than sharing rsa->blinding, but
12186 avoids excessive locking; and if an RSA object is not shared
12187 between threads, blinding will still be very fast).
12191 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12192 ENGINE as defaults for all supported algorithms irrespective of
12193 the 'flags' parameter. 'flags' is now honoured, so applications
12194 should make sure they are passing it correctly.
12198 * Target "mingw" now allows native Windows code to be generated in
12199 the Cygwin environment as well as with the MinGW compiler.
12203 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12205 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12206 via timing by performing a MAC computation even if incorrect
12207 block cipher padding has been found. This is a countermeasure
12208 against active attacks where the attacker has to distinguish
12209 between bad padding and a MAC verification error. ([CVE-2003-0078])
12211 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12212 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12213 Martin Vuagnoux (EPFL, Ilion)*
12215 * Make the no-err option work as intended. The intention with no-err
12216 is not to have the whole error stack handling routines removed from
12217 libcrypto, it's only intended to remove all the function name and
12218 reason texts, thereby removing some of the footprint that may not
12219 be interesting if those errors aren't displayed anyway.
12221 NOTE: it's still possible for any application or module to have its
12222 own set of error texts inserted. The routines are there, just not
12223 used by default when no-err is given.
12227 * Add support for FreeBSD on IA64.
12229 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12231 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12232 Kerberos function mit_des_cbc_cksum(). Before this change,
12233 the value returned by DES_cbc_cksum() was like the one from
12234 mit_des_cbc_cksum(), except the bytes were swapped.
12236 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12238 * Allow an application to disable the automatic SSL chain building.
12239 Before this a rather primitive chain build was always performed in
12240 ssl3_output_cert_chain(): an application had no way to send the
12241 correct chain if the automatic operation produced an incorrect result.
12243 Now the chain builder is disabled if either:
12245 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12247 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12249 The reasoning behind this is that an application would not want the
12250 auto chain building to take place if extra chain certificates are
12251 present and it might also want a means of sending no additional
12252 certificates (for example the chain has two certificates and the
12257 * Add the possibility to build without the ENGINE framework.
12259 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12261 * Under Win32 gmtime() can return NULL: check return value in
12262 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12266 * DSA routines: under certain error conditions uninitialized BN objects
12267 could be freed. Solution: make sure initialization is performed early
12268 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12269 Nils Larsch <nla@trustcenter.de> via PR#459)
12273 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12274 checked on reconnect on the client side, therefore session resumption
12275 could still fail with a "ssl session id is different" error. This
12276 behaviour is masked when SSL_OP_ALL is used due to
12277 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12278 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12279 followup to PR #377.
12283 * IA-32 assembler support enhancements: unified ELF targets, support
12284 for SCO/Caldera platforms, fix for Cygwin shared build.
12288 * Add support for FreeBSD on sparc64. As a consequence, support for
12289 FreeBSD on non-x86 processors is separate from x86 processors on
12290 the config script, much like the NetBSD support.
12292 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12294 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12296 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12299 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12300 code (06) was taken as the first octet of the session ID and the last
12301 octet was ignored consequently. As a result SSLv2 client side session
12302 caching could not have worked due to the session ID mismatch between
12304 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12309 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12310 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12315 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12316 seems that in spite of existing for more than a year, many application
12317 author have done nothing to provide the necessary callbacks, which
12318 means that this particular engine will not work properly anywhere.
12319 This is a very unfortunate situation which forces us, in the name
12320 of usability, to give the hw_ncipher.c a static lock, which is part
12322 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12323 appear in 0.9.8 or later. We EXPECT application authors to have
12324 dealt properly with this when 0.9.8 is released (unless we actually
12325 make such changes in the libcrypto locking code that changes will
12326 have to be made anyway).
12330 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12331 octets have been read, EOF or an error occurs. Without this change
12332 some truncated ASN1 structures will not produce an error.
12336 * Disable Heimdal support, since it hasn't been fully implemented.
12337 Still give the possibility to force the use of Heimdal, but with
12338 warnings and a request that patches get sent to openssl-dev.
12342 * Add the VC-CE target, introduce the WINCE sysname, and add
12343 INSTALL.WCE and appropriate conditionals to make it build.
12345 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12347 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12348 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12349 edit numbers of the version.
12351 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12353 * Introduce safe string copy and catenation functions
12354 (BUF_strlcpy() and BUF_strlcat()).
12356 *Ben Laurie (CHATS) and Richard Levitte*
12358 * Avoid using fixed-size buffers for one-line DNs.
12360 *Ben Laurie (CHATS)*
12362 * Add BUF_MEM_grow_clean() to avoid information leakage when
12363 resizing buffers containing secrets, and use where appropriate.
12365 *Ben Laurie (CHATS)*
12367 * Avoid using fixed size buffers for configuration file location.
12369 *Ben Laurie (CHATS)*
12371 * Avoid filename truncation for various CA files.
12373 *Ben Laurie (CHATS)*
12375 * Use sizeof in preference to magic numbers.
12377 *Ben Laurie (CHATS)*
12379 * Avoid filename truncation in cert requests.
12381 *Ben Laurie (CHATS)*
12383 * Add assertions to check for (supposedly impossible) buffer
12386 *Ben Laurie (CHATS)*
12388 * Don't cache truncated DNS entries in the local cache (this could
12389 potentially lead to a spoofing attack).
12391 *Ben Laurie (CHATS)*
12393 * Fix various buffers to be large enough for hex/decimal
12394 representations in a platform independent manner.
12396 *Ben Laurie (CHATS)*
12398 * Add CRYPTO_realloc_clean() to avoid information leakage when
12399 resizing buffers containing secrets, and use where appropriate.
12401 *Ben Laurie (CHATS)*
12403 * Add BIO_indent() to avoid much slightly worrying code to do
12406 *Ben Laurie (CHATS)*
12408 * Convert sprintf()/BIO_puts() to BIO_printf().
12410 *Ben Laurie (CHATS)*
12412 * buffer_gets() could terminate with the buffer only half
12415 *Ben Laurie (CHATS)*
12417 * Add assertions to prevent user-supplied crypto functions from
12418 overflowing internal buffers by having large block sizes, etc.
12420 *Ben Laurie (CHATS)*
12422 * New OPENSSL_assert() macro (similar to assert(), but enabled
12425 *Ben Laurie (CHATS)*
12427 * Eliminate unused copy of key in RC4.
12429 *Ben Laurie (CHATS)*
12431 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12433 *Ben Laurie (CHATS)*
12435 * Fix off-by-one error in EGD path.
12437 *Ben Laurie (CHATS)*
12439 * If RANDFILE path is too long, ignore instead of truncating.
12441 *Ben Laurie (CHATS)*
12443 * Eliminate unused and incorrectly sized X.509 structure
12446 *Ben Laurie (CHATS)*
12448 * Eliminate unused and dangerous function knumber().
12450 *Ben Laurie (CHATS)*
12452 * Eliminate unused and dangerous structure, KSSL_ERR.
12454 *Ben Laurie (CHATS)*
12456 * Protect against overlong session ID context length in an encoded
12457 session object. Since these are local, this does not appear to be
12460 *Ben Laurie (CHATS)*
12462 * Change from security patch (see 0.9.6e below) that did not affect
12463 the 0.9.6 release series:
12465 Remote buffer overflow in SSL3 protocol - an attacker could
12466 supply an oversized master key in Kerberos-enabled versions.
12469 *Ben Laurie (CHATS)*
12471 * Change the SSL kerb5 codes to match RFC 2712.
12475 * Make -nameopt work fully for req and add -reqopt switch.
12477 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12479 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12481 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12483 * Make sure tests can be performed even if the corresponding algorithms
12484 have been removed entirely. This was also the last step to make
12485 OpenSSL compilable with DJGPP under all reasonable conditions.
12487 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12489 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12490 to allow version independent disabling of normally unselected ciphers,
12491 which may be activated as a side-effect of selecting a single cipher.
12493 (E.g., cipher list string "RSA" enables ciphersuites that are left
12494 out of "ALL" because they do not provide symmetric encryption.
12495 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12497 *Lutz Jaenicke, Bodo Moeller*
12499 * Add appropriate support for separate platform-dependent build
12500 directories. The recommended way to make a platform-dependent
12501 build directory is the following (tested on Linux), maybe with
12504 # Place yourself outside of the OpenSSL source tree. In
12505 # this example, the environment variable OPENSSL_SOURCE
12506 # is assumed to contain the absolute OpenSSL source directory.
12507 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12508 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12509 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12510 mkdir -p `dirname $F`
12511 ln -s $OPENSSL_SOURCE/$F $F
12514 To be absolutely sure not to disturb the source tree, a "make clean"
12515 is a good thing. If it isn't successful, don't worry about it,
12516 it probably means the source directory is very clean.
12520 * Make sure any ENGINE control commands make local copies of string
12521 pointers passed to them whenever necessary. Otherwise it is possible
12522 the caller may have overwritten (or deallocated) the original string
12523 data when a later ENGINE operation tries to use the stored values.
12525 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12527 * Improve diagnostics in file reading and command-line digests.
12529 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12531 * Add AES modes CFB and OFB to the object database. Correct an
12532 error in AES-CFB decryption.
12536 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12537 allows existing EVP_CIPHER_CTX structures to be reused after
12538 calling `EVP_*Final()`. This behaviour is used by encryption
12539 BIOs and some applications. This has the side effect that
12540 applications must explicitly clean up cipher contexts with
12541 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12545 * Check the values of dna and dnb in bn_mul_recursive before calling
12546 bn_mul_comba (a non zero value means the a or b arrays do not contain
12547 n2 elements) and fallback to bn_mul_normal if either is not zero.
12551 * Fix escaping of non-ASCII characters when using the -subj option
12552 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12556 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12557 form for "surname", serialNumber has no short form.
12558 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12559 therefore remove "mail" short name for "internet 7".
12560 The OID for unique identifiers in X509 certificates is
12561 x500UniqueIdentifier, not uniqueIdentifier.
12562 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12566 * Add an "init" command to the ENGINE config module and auto initialize
12567 ENGINEs. Without any "init" command the ENGINE will be initialized
12568 after all ctrl commands have been executed on it. If init=1 the
12569 ENGINE is initialized at that point (ctrls before that point are run
12570 on the uninitialized ENGINE and after on the initialized one). If
12571 init=0 then the ENGINE will not be initialized at all.
12575 * Fix the 'app_verify_callback' interface so that the user-defined
12576 argument is actually passed to the callback: In the
12577 SSL_CTX_set_cert_verify_callback() prototype, the callback
12578 declaration has been changed from
12581 int (*cb)(X509_STORE_CTX *,void *);
12582 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12583 i=s->ctx->app_verify_callback(&ctx)
12584 has been changed into
12585 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12587 To update applications using SSL_CTX_set_cert_verify_callback(),
12588 a dummy argument can be added to their callback functions.
12590 *D. K. Smetters <smetters@parc.xerox.com>*
12592 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12594 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12596 * Add and OPENSSL_LOAD_CONF define which will cause
12597 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12598 This allows older applications to transparently support certain
12599 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12600 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12601 load the config file and OPENSSL_add_all_algorithms_conf() which will
12602 always load it have also been added.
12606 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12607 Adjust NIDs and EVP layer.
12609 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12611 * Config modules support in openssl utility.
12613 Most commands now load modules from the config file,
12614 though in a few (such as version) this isn't done
12615 because it couldn't be used for anything.
12617 In the case of ca and req the config file used is
12618 the same as the utility itself: that is the -config
12619 command line option can be used to specify an
12624 * Move default behaviour from OPENSSL_config(). If appname is NULL
12625 use "openssl_conf" if filename is NULL use default openssl config file.
12629 * Add an argument to OPENSSL_config() to allow the use of an alternative
12630 config section name. Add a new flag to tolerate a missing config file
12631 and move code to CONF_modules_load_file().
12635 * Support for crypto accelerator cards from Accelerated Encryption
12636 Processing, www.aep.ie. (Use engine 'aep')
12637 The support was copied from 0.9.6c [engine] and adapted/corrected
12638 to work with the new engine framework.
12640 *AEP Inc. and Richard Levitte*
12642 * Support for SureWare crypto accelerator cards from Baltimore
12643 Technologies. (Use engine 'sureware')
12644 The support was copied from 0.9.6c [engine] and adapted
12645 to work with the new engine framework.
12649 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12650 make the newer ENGINE framework commands for the CHIL engine work.
12652 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12654 * Make it possible to produce shared libraries on ReliantUNIX.
12656 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12658 * Add the configuration target debug-linux-ppro.
12659 Make 'openssl rsa' use the general key loading routines
12660 implemented in `apps.c`, and make those routines able to
12661 handle the key format FORMAT_NETSCAPE and the variant
12664 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12666 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12668 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12670 * Add -keyform to rsautl, and document -engine.
12672 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12674 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12675 BIO_R_NO_SUCH_FILE error code rather than the generic
12676 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12680 * Add new functions
12681 ERR_peek_last_error
12682 ERR_peek_last_error_line
12683 ERR_peek_last_error_line_data.
12684 These are similar to
12686 ERR_peek_error_line
12687 ERR_peek_error_line_data,
12688 but report on the latest error recorded rather than the first one
12689 still in the error queue.
12691 *Ben Laurie, Bodo Moeller*
12693 * default_algorithms option in ENGINE config module. This allows things
12695 default_algorithms = ALL
12696 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12700 * Preliminary ENGINE config module.
12704 * New experimental application configuration code.
12708 * Change the AES code to follow the same name structure as all other
12709 symmetric ciphers, and behave the same way. Move everything to
12710 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12712 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12714 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12716 *Ben Laurie and Theo de Raadt*
12718 * Add option to output public keys in req command.
12720 *Massimiliano Pala madwolf@openca.org*
12722 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12723 (up to about 10% better than before for P-192 and P-224).
12727 * New functions/macros
12729 SSL_CTX_set_msg_callback(ctx, cb)
12730 SSL_CTX_set_msg_callback_arg(ctx, arg)
12731 SSL_set_msg_callback(ssl, cb)
12732 SSL_set_msg_callback_arg(ssl, arg)
12734 to request calling a callback function
12736 void cb(int write_p, int version, int content_type,
12737 const void *buf, size_t len, SSL *ssl, void *arg)
12739 whenever a protocol message has been completely received
12740 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12741 protocol version according to which the SSL library interprets
12742 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12743 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12744 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12745 specification (change_cipher_spec(20), alert(21), handshake(22)).
12746 'buf' and 'len' point to the actual message, 'ssl' to the
12747 SSL object, and 'arg' is the application-defined value set by
12748 SSL[_CTX]_set_msg_callback_arg().
12750 'openssl s_client' and 'openssl s_server' have new '-msg' options
12751 to enable a callback that displays all protocol messages.
12755 * Change the shared library support so shared libraries are built as
12756 soon as the corresponding static library is finished, and thereby get
12757 openssl and the test programs linked against the shared library.
12758 This still only happens when the keyword "shard" has been given to
12759 the configuration scripts.
12761 NOTE: shared library support is still an experimental thing, and
12762 backward binary compatibility is still not guaranteed.
12764 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12766 * Add support for Subject Information Access extension.
12768 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12770 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12771 additional bytes when new memory had to be allocated, not just
12772 when reusing an existing buffer.
12776 * New command line and configuration option 'utf8' for the req command.
12777 This allows field values to be specified as UTF8 strings.
12781 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12782 runs for the former and machine-readable output for the latter.
12786 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12787 of the e-mail address in the DN (i.e., it will go into a certificate
12788 extension only). The new configuration file option 'email_in_dn = no'
12789 has the same effect.
12791 *Massimiliano Pala madwolf@openca.org*
12793 * Change all functions with names starting with `des_` to be starting
12794 with `DES_` instead. Add wrappers that are compatible with libdes,
12795 but are named `_ossl_old_des_*`. Finally, add macros that map the
12796 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12797 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12798 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12801 Since we provide two compatibility mappings, the user needs to
12802 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12803 compatibility is desired. The default (i.e., when that macro
12804 isn't defined) is OpenSSL 0.9.6c compatibility.
12806 There are also macros that enable and disable the support of old
12807 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12808 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12809 are defined, the default will apply: to support the old des routines.
12811 In either case, one must include openssl/des.h to get the correct
12812 definitions. Do not try to just include openssl/des_old.h, that
12815 NOTE: This is a major break of an old API into a new one. Software
12816 authors are encouraged to switch to the `DES_` style functions. Some
12817 time in the future, des_old.h and the libdes compatibility functions
12818 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12819 default), and then completely removed.
12823 * Test for certificates which contain unsupported critical extensions.
12824 If such a certificate is found during a verify operation it is
12825 rejected by default: this behaviour can be overridden by either
12826 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12827 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12828 X509_supported_extension() has also been added which returns 1 if a
12829 particular extension is supported.
12833 * Modify the behaviour of EVP cipher functions in similar way to digests
12834 to retain compatibility with existing code.
12838 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12839 compatibility with existing code. In particular the 'ctx' parameter does
12840 not have to be to be initialized before the call to EVP_DigestInit() and
12841 it is tidied up after a call to EVP_DigestFinal(). New function
12842 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12843 EVP_MD_CTX_copy() changed to not require the destination to be
12844 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12845 requires the destination to be valid.
12847 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12848 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12852 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12853 so that complete 'Handshake' protocol structures are kept in memory
12854 instead of overwriting 'msg_type' and 'length' with 'body' data.
12858 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12860 *Massimo Santin via Richard Levitte*
12862 * Major restructuring to the underlying ENGINE code. This includes
12863 reduction of linker bloat, separation of pure "ENGINE" manipulation
12864 (initialisation, etc) from functionality dealing with implementations
12865 of specific crypto interfaces. This change also introduces integrated
12866 support for symmetric ciphers and digest implementations - so ENGINEs
12867 can now accelerate these by providing EVP_CIPHER and EVP_MD
12868 implementations of their own. This is detailed in
12869 [crypto/engine/README.md](crypto/engine/README.md)
12870 as it couldn't be adequately described here. However, there are a few
12871 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12872 were changed in the original introduction of ENGINE code have now
12873 reverted back - the hooking from this code to ENGINE is now a good
12874 deal more passive and at run-time, operations deal directly with
12875 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12876 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12877 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12878 they were not being used by the framework as there is no concept of a
12879 BIGNUM_METHOD and they could not be generalised to the new
12880 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12881 ENGINE_cpy() has been removed as it cannot be consistently defined in
12886 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12890 * Change mkdef.pl to sort symbols that get the same entry number,
12891 and make sure the automatically generated functions `ERR_load_*`
12892 become part of libeay.num as well.
12896 * New function SSL_renegotiate_pending(). This returns true once
12897 renegotiation has been requested (either SSL_renegotiate() call
12898 or HelloRequest/ClientHello received from the peer) and becomes
12899 false once a handshake has been completed.
12900 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12901 sends a HelloRequest, but does not ensure that a handshake takes
12902 place. SSL_renegotiate_pending() is useful for checking if the
12903 client has followed the request.)
12907 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12908 By default, clients may request session resumption even during
12909 renegotiation (if session ID contexts permit); with this option,
12910 session resumption is possible only in the first handshake.
12912 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12913 more bits available for options that should not be part of
12914 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12918 * Add some demos for certificate and certificate request creation.
12922 * Make maximum certificate chain size accepted from the peer application
12923 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12924 "Douglas E. Engert" <deengert@anl.gov>.
12928 * Add support for shared libraries for Unixware-7
12929 (Boyd Lynn Gerber <gerberb@zenez.com>).
12933 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12934 be done prior to destruction. Use this to unload error strings from
12935 ENGINEs that load their own error strings. NB: This adds two new API
12936 functions to "get" and "set" this destroy handler in an ENGINE.
12940 * Alter all existing ENGINE implementations (except "openssl" and
12941 "openbsd") to dynamically instantiate their own error strings. This
12942 makes them more flexible to be built both as statically-linked ENGINEs
12943 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12944 Also, add stub code to each that makes building them as self-contained
12945 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12949 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12950 implementations into applications that are completely implemented in
12951 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12952 commands that can be used to configure what shared-library to load and
12953 to control aspects of the way it is handled. Also, made an update to
12954 the [README-Engine.md](README-Engine.md) file
12955 that brings its information up-to-date and
12956 provides some information and instructions on the "dynamic" ENGINE
12957 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12961 * Make it possible to unload ranges of ERR strings with a new
12962 "ERR_unload_strings" function.
12966 * Add a copy() function to EVP_MD.
12970 * Make EVP_MD routines take a context pointer instead of just the
12971 md_data void pointer.
12975 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12976 that the digest can only process a single chunk of data
12977 (typically because it is provided by a piece of
12978 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12979 is only going to provide a single chunk of data, and hence the
12980 framework needn't accumulate the data for oneshot drivers.
12984 * As with "ERR", make it possible to replace the underlying "ex_data"
12985 functions. This change also alters the storage and management of global
12986 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12987 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12988 index counters. The API functions that use this state have been changed
12989 to take a "class_index" rather than pointers to the class's local STACK
12990 and counter, and there is now an API function to dynamically create new
12991 classes. This centralisation allows us to (a) plug a lot of the
12992 thread-safety problems that existed, and (b) makes it possible to clean
12993 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12994 such data would previously have always leaked in application code and
12995 workarounds were in place to make the memory debugging turn a blind eye
12996 to it. Application code that doesn't use this new function will still
12997 leak as before, but their memory debugging output will announce it now
12998 rather than letting it slide.
13000 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13001 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13002 has a return value to indicate success or failure.
13006 * Make it possible to replace the underlying "ERR" functions such that the
13007 global state (2 LHASH tables and 2 locks) is only used by the "default"
13008 implementation. This change also adds two functions to "get" and "set"
13009 the implementation prior to it being automatically set the first time
13010 any other ERR function takes place. Ie. an application can call "get",
13011 pass the return value to a module it has just loaded, and that module
13012 can call its own "set" function using that value. This means the
13013 module's "ERR" operations will use (and modify) the error state in the
13014 application and not in its own statically linked copy of OpenSSL code.
13018 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13019 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13020 the operation, and provides a more encapsulated way for external code
13021 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13022 to use these functions rather than manually incrementing the counts.
13024 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13028 * Add EVP test program.
13032 * Add symmetric cipher support to ENGINE. Expect the API to change!
13036 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13037 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13038 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13039 These allow a CRL to be built without having to access X509_CRL fields
13040 directly. Modify 'ca' application to use new functions.
13044 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13045 bug workarounds. Rollback attack detection is a security feature.
13046 The problem will only arise on OpenSSL servers when TLSv1 is not
13047 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13048 Software authors not wanting to support TLSv1 will have special reasons
13049 for their choice and can explicitly enable this option.
13051 *Bodo Moeller, Lutz Jaenicke*
13053 * Rationalise EVP so it can be extended: don't include a union of
13054 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13055 (similar to those existing for EVP_CIPHER_CTX).
13060 EVP_MD_CTX_init(&md); /* new function call */
13061 EVP_DigestInit(&md, EVP_sha1());
13062 EVP_DigestUpdate(&md, in, len);
13063 EVP_DigestFinal(&md, out, NULL);
13064 EVP_MD_CTX_cleanup(&md); /* new function call */
13068 * Make DES key schedule conform to the usual scheme, as well as
13069 correcting its structure. This means that calls to DES functions
13070 now have to pass a pointer to a des_key_schedule instead of a
13071 plain des_key_schedule (which was actually always a pointer
13074 des_key_schedule ks;
13076 des_set_key_checked(..., &ks);
13077 des_ncbc_encrypt(..., &ks, ...);
13079 (Note that a later change renames 'des_...' into 'DES_...'.)
13083 * Initial reduction of linker bloat: the use of some functions, such as
13084 PEM causes large amounts of unused functions to be linked in due to
13085 poor organisation. For example pem_all.c contains every PEM function
13086 which has a knock on effect of linking in large amounts of (unused)
13087 ASN1 code. Grouping together similar functions and splitting unrelated
13088 functions prevents this.
13092 * Cleanup of EVP macros.
13096 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13097 correct `_ecb suffix`.
13101 * Add initial OCSP responder support to ocsp application. The
13102 revocation information is handled using the text based index
13103 use by the ca application. The responder can either handle
13104 requests generated internally, supplied in files (for example
13105 via a CGI script) or using an internal minimal server.
13109 * Add configuration choices to get zlib compression for TLS.
13113 * Changes to Kerberos SSL for RFC 2712 compliance:
13114 1. Implemented real KerberosWrapper, instead of just using
13115 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13116 2. Implemented optional authenticator field of KerberosWrapper.
13118 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13119 and authenticator structs; see crypto/krb5/.
13121 Generalized Kerberos calls to support multiple Kerberos libraries.
13122 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13123 via Richard Levitte*
13125 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13126 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13127 values for each of the key sizes rather than having just
13128 parameters (and 'speed' generating keys each time).
13132 * Speed up EVP routines.
13135 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13136 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13137 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13138 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13140 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13141 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13142 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13145 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13147 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13151 * Added the OS2-EMX target.
13153 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13155 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13156 New functions to support `NCONF` routines in extension code.
13157 New function `CONF_set_nconf()`
13158 to allow functions which take an `NCONF` to also handle the old `LHASH`
13159 structure: this means that the old `CONF` compatible routines can be
13160 retained (in particular w.rt. extensions) without having to duplicate the
13161 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13165 * Enhance the general user interface with mechanisms for inner control
13166 and with possibilities to have yes/no kind of prompts.
13170 * Change all calls to low-level digest routines in the library and
13171 applications to use EVP. Add missing calls to HMAC_cleanup() and
13172 don't assume HMAC_CTX can be copied using memcpy().
13174 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13176 * Add the possibility to control engines through control names but with
13177 arbitrary arguments instead of just a string.
13178 Change the key loaders to take a UI_METHOD instead of a callback
13179 function pointer. NOTE: this breaks binary compatibility with earlier
13180 versions of OpenSSL [engine].
13181 Adapt the nCipher code for these new conditions and add a card insertion
13186 * Enhance the general user interface with mechanisms to better support
13187 dialog box interfaces, application-defined prompts, the possibility
13188 to use defaults (for example default passwords from somewhere else)
13189 and interrupts/cancellations.
13193 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13194 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13198 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13199 tidy up some unnecessarily weird code in 'sk_new()').
13201 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13203 * Change the key loading routines for ENGINEs to use the same kind
13204 callback (pem_password_cb) as all other routines that need this
13209 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13210 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13211 than this minimum value is recommended.
13215 * New random seeder for OpenVMS, using the system process statistics
13216 that are easily reachable.
13220 * Windows apparently can't transparently handle global
13221 variables defined in DLLs. Initialisations such as:
13223 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13225 won't compile. This is used by the any applications that need to
13226 declare their own ASN1 modules. This was fixed by adding the option
13227 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13228 needed for static libraries under Win32.
13232 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13233 setting of purpose and trust fields. New X509_STORE trust and
13234 purpose functions and tidy up setting in other SSL functions.
13238 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13239 structure. These are inherited by X509_STORE_CTX when it is
13240 initialised. This allows various defaults to be set in the
13241 X509_STORE structure (such as flags for CRL checking and custom
13242 purpose or trust settings) for functions which only use X509_STORE_CTX
13243 internally such as S/MIME.
13245 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13246 trust settings if they are not set in X509_STORE. This allows X509_STORE
13247 purposes and trust (in S/MIME for example) to override any set by default.
13249 Add command line options for CRL checking to smime, s_client and s_server
13254 * Initial CRL based revocation checking. If the CRL checking flag(s)
13255 are set then the CRL is looked up in the X509_STORE structure and
13256 its validity and signature checked, then if the certificate is found
13257 in the CRL the verify fails with a revoked error.
13259 Various new CRL related callbacks added to X509_STORE_CTX structure.
13261 Command line options added to 'verify' application to support this.
13263 This needs some additional work, such as being able to handle multiple
13264 CRLs with different times, extension based lookup (rather than just
13265 by subject name) and ultimately more complete V2 CRL extension
13270 * Add a general user interface API (crypto/ui/). This is designed
13271 to replace things like des_read_password and friends (backward
13272 compatibility functions using this new API are provided).
13273 The purpose is to remove prompting functions from the DES code
13274 section as well as provide for prompting through dialog boxes in
13275 a window system and the like.
13279 * Add "ex_data" support to ENGINE so implementations can add state at a
13280 per-structure level rather than having to store it globally.
13284 * Make it possible for ENGINE structures to be copied when retrieved by
13285 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13286 This causes the "original" ENGINE structure to act like a template,
13287 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13288 operational state can be localised to each ENGINE structure, despite the
13289 fact they all share the same "methods". New ENGINE structures returned in
13290 this case have no functional references and the return value is the single
13291 structural reference. This matches the single structural reference returned
13292 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13297 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13298 needs to match any other type at all we need to manually clear the
13303 * Changes to the "openssl engine" utility to include;
13304 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13305 about an ENGINE's available control commands.
13306 - executing control commands from command line arguments using the
13307 '-pre' and '-post' switches. '-post' is only used if '-t' is
13308 specified and the ENGINE is successfully initialised. The syntax for
13309 the individual commands are colon-separated, for example;
13310 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13314 * New dynamic control command support for ENGINEs. ENGINEs can now
13315 declare their own commands (numbers), names (strings), descriptions,
13316 and input types for run-time discovery by calling applications. A
13317 subset of these commands are implicitly classed as "executable"
13318 depending on their input type, and only these can be invoked through
13319 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13320 can be based on user input, config files, etc). The distinction is
13321 that "executable" commands cannot return anything other than a boolean
13322 result and can only support numeric or string input, whereas some
13323 discoverable commands may only be for direct use through
13324 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13325 pointers, or other custom uses. The "executable" commands are to
13326 support parameterisations of ENGINE behaviour that can be
13327 unambiguously defined by ENGINEs and used consistently across any
13328 OpenSSL-based application. Commands have been added to all the
13329 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13330 control over shared-library paths without source code alterations.
13334 * Changed all ENGINE implementations to dynamically allocate their
13335 ENGINEs rather than declaring them statically. Apart from this being
13336 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13337 this also allows the implementations to compile without using the
13338 internal engine_int.h header.
13342 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13343 'const' value. Any code that should be able to modify a RAND_METHOD
13344 should already have non-const pointers to it (ie. they should only
13345 modify their own ones).
13349 * Made a variety of little tweaks to the ENGINE code.
13350 - "atalla" and "ubsec" string definitions were moved from header files
13351 to C code. "nuron" string definitions were placed in variables
13352 rather than hard-coded - allowing parameterisation of these values
13353 later on via ctrl() commands.
13354 - Removed unused "#if 0"'d code.
13355 - Fixed engine list iteration code so it uses ENGINE_free() to release
13356 structural references.
13357 - Constified the RAND_METHOD element of ENGINE structures.
13358 - Constified various get/set functions as appropriate and added
13359 missing functions (including a catch-all ENGINE_cpy that duplicates
13360 all ENGINE values onto a new ENGINE except reference counts/state).
13361 - Removed NULL parameter checks in get/set functions. Setting a method
13362 or function to NULL is a way of cancelling out a previously set
13363 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13364 and doesn't justify the extra error symbols and code.
13365 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13366 flags from engine_int.h to engine.h.
13367 - Changed prototypes for ENGINE handler functions (init(), finish(),
13368 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13372 * Implement binary inversion algorithm for BN_mod_inverse in addition
13373 to the algorithm using long division. The binary algorithm can be
13374 used only if the modulus is odd. On 32-bit systems, it is faster
13375 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13376 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13377 up to 450 bits. In 64-bit environments, the binary algorithm
13378 appears to be advantageous for much longer moduli; here we use it
13379 for moduli up to 2048 bits.
13383 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13384 could not support the combine flag in choice fields.
13388 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13389 extensions from a certificate request to the certificate.
13393 * Allow multiple 'certopt' and 'nameopt' options to be separated
13394 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13395 file: this allows the display of the certificate about to be
13396 signed to be customised, to allow certain fields to be included
13397 or excluded and extension details. The old system didn't display
13398 multicharacter strings properly, omitted fields not in the policy
13399 and couldn't display additional details such as extensions.
13403 * Function EC_POINTs_mul for multiple scalar multiplication
13404 of an arbitrary number of elliptic curve points
13405 \sum scalars[i]*points[i],
13406 optionally including the generator defined for the EC_GROUP:
13407 scalar*generator + \sum scalars[i]*points[i].
13409 EC_POINT_mul is a simple wrapper function for the typical case
13410 that the point list has just one item (besides the optional
13415 * First EC_METHODs for curves over GF(p):
13417 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13418 operations and provides various method functions that can also
13419 operate with faster implementations of modular arithmetic.
13421 EC_GFp_mont_method() reuses most functions that are part of
13422 EC_GFp_simple_method, but uses Montgomery arithmetic.
13424 *Bodo Moeller; point addition and point doubling
13425 implementation directly derived from source code provided by
13426 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13428 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13429 crypto/ec/ec_lib.c):
13431 Curves are EC_GROUP objects (with an optional group generator)
13432 based on EC_METHODs that are built into the library.
13434 Points are EC_POINT objects based on EC_GROUP objects.
13436 Most of the framework would be able to handle curves over arbitrary
13437 finite fields, but as there are no obvious types for fields other
13438 than GF(p), some functions are limited to that for now.
13442 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13443 that the file contains a complete HTTP response.
13447 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13448 change the def and num file printf format specifier from "%-40sXXX"
13449 to "%-39s XXX". The latter will always guarantee a space after the
13450 field while the former will cause them to run together if the field
13451 is 40 of more characters long.
13455 * Constify the cipher and digest 'method' functions and structures
13456 and modify related functions to take constant EVP_MD and EVP_CIPHER
13461 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13462 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13466 * Modify `EVP_Digest*()` routines so they now return values. Although the
13467 internal software routines can never fail additional hardware versions
13472 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13474 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13475 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13478 ERR_R_NESTED_ASN1_ERROR
13480 ERR_R_MISSING_ASN1_EOS
13481 were 4 .. 9, conflicting with
13482 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13484 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13485 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13487 Add new error code 'ERR_R_INTERNAL_ERROR'.
13491 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13496 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13497 sets the subject name for a new request or supersedes the
13498 subject name in a given request. Formats that can be parsed are
13499 'CN=Some Name, OU=myOU, C=IT'
13501 'CN=Some Name/OU=myOU/C=IT'.
13503 Add options '-batch' and '-verbose' to 'openssl req'.
13505 *Massimiliano Pala <madwolf@hackmasters.net>*
13507 * Introduce the possibility to access global variables through
13508 functions on platform were that's the best way to handle exporting
13509 global variables in shared libraries. To enable this functionality,
13510 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13511 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13512 is normally done by Configure or something similar).
13514 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13515 in the source file (foo.c) like this:
13517 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13518 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13520 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13521 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13523 OPENSSL_DECLARE_GLOBAL(int,foo);
13524 #define foo OPENSSL_GLOBAL_REF(foo)
13525 OPENSSL_DECLARE_GLOBAL(double,bar);
13526 #define bar OPENSSL_GLOBAL_REF(bar)
13528 The #defines are very important, and therefore so is including the
13529 header file everywhere where the defined globals are used.
13531 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13532 of ASN.1 items, but that structure is a bit different.
13534 The largest change is in util/mkdef.pl which has been enhanced with
13535 better and easier to understand logic to choose which symbols should
13536 go into the Windows .def files as well as a number of fixes and code
13537 cleanup (among others, algorithm keywords are now sorted
13538 lexicographically to avoid constant rewrites).
13542 * In BN_div() keep a copy of the sign of 'num' before writing the
13543 result to 'rm' because if rm==num the value will be overwritten
13544 and produce the wrong result if 'num' is negative: this caused
13545 problems with BN_mod() and BN_nnmod().
13549 * Function OCSP_request_verify(). This checks the signature on an
13550 OCSP request and verifies the signer certificate. The signer
13551 certificate is just checked for a generic purpose and OCSP request
13556 * Add OCSP_check_validity() function to check the validity of OCSP
13557 responses. OCSP responses are prepared in real time and may only
13558 be a few seconds old. Simply checking that the current time lies
13559 between thisUpdate and nextUpdate max reject otherwise valid responses
13560 caused by either OCSP responder or client clock inaccuracy. Instead
13561 we allow thisUpdate and nextUpdate to fall within a certain period of
13562 the current time. The age of the response can also optionally be
13563 checked. Two new options -validity_period and -status_age added to
13568 * If signature or public key algorithm is unrecognized print out its
13569 OID rather that just UNKNOWN.
13573 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13574 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13575 ID to be generated from the issuer certificate alone which can then be
13576 passed to OCSP_id_issuer_cmp().
13580 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13581 ASN1 modules to export functions returning ASN1_ITEM pointers
13582 instead of the ASN1_ITEM structures themselves. This adds several
13583 new macros which allow the underlying ASN1 function/structure to
13584 be accessed transparently. As a result code should not use ASN1_ITEM
13585 references directly (such as &X509_it) but instead use the relevant
13586 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13587 use of the new ASN1 code on platforms where exporting structures
13588 is problematical (for example in shared libraries) but exporting
13589 functions returning pointers to structures is not.
13593 * Add support for overriding the generation of SSL/TLS session IDs.
13594 These callbacks can be registered either in an SSL_CTX or per SSL.
13595 The purpose of this is to allow applications to control, if they wish,
13596 the arbitrary values chosen for use as session IDs, particularly as it
13597 can be useful for session caching in multiple-server environments. A
13598 command-line switch for testing this (and any client code that wishes
13599 to use such a feature) has been added to "s_server".
13601 *Geoff Thorpe, Lutz Jaenicke*
13603 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13604 of the form `#if defined(...) || defined(...) || ...` and
13605 `#if !defined(...) && !defined(...) && ...`. This also avoids
13606 the growing number of special cases it was previously handling.
13610 * Make all configuration macros available for application by making
13611 sure they are available in opensslconf.h, by giving them names starting
13612 with `OPENSSL_` to avoid conflicts with other packages and by making
13613 sure e_os2.h will cover all platform-specific cases together with
13615 Additionally, it is now possible to define configuration/platform-
13616 specific names (called "system identities"). In the C code, these
13617 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13618 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13619 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13624 * New option -set_serial to 'req' and 'x509' this allows the serial
13625 number to use to be specified on the command line. Previously self
13626 signed certificates were hard coded with serial number 0 and the
13627 CA options of 'x509' had to use a serial number in a file which was
13632 * New options to 'ca' utility to support V2 CRL entry extensions.
13633 Currently CRL reason, invalidity date and hold instruction are
13634 supported. Add new CRL extensions to V3 code and some new objects.
13638 * New function EVP_CIPHER_CTX_set_padding() this is used to
13639 disable standard block padding (aka PKCS#5 padding) in the EVP
13640 API, which was previously mandatory. This means that the data is
13641 not padded in any way and so the total length much be a multiple
13642 of the block size, otherwise an error occurs.
13646 * Initial (incomplete) OCSP SSL support.
13650 * New function OCSP_parse_url(). This splits up a URL into its host,
13651 port and path components: primarily to parse OCSP URLs. New -url
13652 option to ocsp utility.
13656 * New nonce behavior. The return value of OCSP_check_nonce() now
13657 reflects the various checks performed. Applications can decide
13658 whether to tolerate certain situations such as an absent nonce
13659 in a response when one was present in a request: the ocsp application
13660 just prints out a warning. New function OCSP_add1_basic_nonce()
13661 this is to allow responders to include a nonce in a response even if
13662 the request is nonce-less.
13666 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13667 skipped when using openssl x509 multiple times on a single input file,
13668 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13672 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13673 set string type: to handle setting ASN1_TIME structures. Fix ca
13674 utility to correctly initialize revocation date of CRLs.
13678 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13679 the clients preferred ciphersuites and rather use its own preferences.
13680 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13681 Internet Explorer by ensuring unchanged hash method during stepup.
13682 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13686 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13687 to aes and add a new 'exist' option to print out symbols that don't
13692 * Additional options to ocsp utility to allow flags to be set and
13693 additional certificates supplied.
13697 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13698 OCSP client a number of certificate to only verify the response
13703 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13704 handle the new API. Currently only ECB, CBC modes supported. Add new
13707 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13708 Encryption Standard (AES) Ciphersuites for Transport Layer
13709 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13710 not enabled by default and were not part of the "ALL" ciphersuite
13711 alias because they were not yet official; they could be
13712 explicitly requested by specifying the "AESdraft" ciphersuite
13713 group alias. In the final release of OpenSSL 0.9.7, the group
13714 alias is called "AES" and is part of "ALL".)
13716 *Ben Laurie, Steve Henson, Bodo Moeller*
13718 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13719 request to response.
13723 * Functions for OCSP responders. OCSP_request_onereq_count(),
13724 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13725 extract information from a certificate request. OCSP_response_create()
13726 creates a response and optionally adds a basic response structure.
13727 OCSP_basic_add1_status() adds a complete single response to a basic
13728 response and returns the OCSP_SINGLERESP structure just added (to allow
13729 extensions to be included for example). OCSP_basic_add1_cert() adds a
13730 certificate to a basic response and OCSP_basic_sign() signs a basic
13731 response with various flags. New helper functions ASN1_TIME_check()
13732 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13733 (converts ASN1_TIME to GeneralizedTime).
13737 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13738 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13739 structure from a certificate. X509_pubkey_digest() digests the public_key
13740 contents: this is used in various key identifiers.
13744 * Make sk_sort() tolerate a NULL argument.
13746 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13748 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13749 passed by the function are trusted implicitly. If any of them signed the
13750 response then it is assumed to be valid and is not verified.
13754 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13755 to data. This was previously part of the PKCS7 ASN1 code. This
13756 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13757 *Steve Henson, reported by Kenneth R. Robinette
13758 <support@securenetterm.com>*
13760 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13761 routines: without these tracing memory leaks is very painful.
13762 Fix leaks in PKCS12 and PKCS7 routines.
13766 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13767 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13768 effectively meant GeneralizedTime would never be used. Now it
13769 is initialised to -1 but X509_time_adj() now has to check the value
13770 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13771 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13772 *Steve Henson, reported by Kenneth R. Robinette
13773 <support@securenetterm.com>*
13775 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13776 result in a zero length in the ASN1_INTEGER structure which was
13777 not consistent with the structure when d2i_ASN1_INTEGER() was used
13778 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13779 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13780 where it did not print out a minus for negative ASN1_INTEGER.
13784 * Add summary printout to ocsp utility. The various functions which
13785 convert status values to strings have been renamed to:
13786 OCSP_response_status_str(), OCSP_cert_status_str() and
13787 OCSP_crl_reason_str() and are no longer static. New options
13788 to verify nonce values and to disable verification. OCSP response
13789 printout format cleaned up.
13793 * Add additional OCSP certificate checks. These are those specified
13794 in RFC2560. This consists of two separate checks: the CA of the
13795 certificate being checked must either be the OCSP signer certificate
13796 or the issuer of the OCSP signer certificate. In the latter case the
13797 OCSP signer certificate must contain the OCSP signing extended key
13798 usage. This check is performed by attempting to match the OCSP
13799 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13800 in the OCSP_CERTID structures of the response.
13804 * Initial OCSP certificate verification added to OCSP_basic_verify()
13805 and related routines. This uses the standard OpenSSL certificate
13806 verify routines to perform initial checks (just CA validity) and
13807 to obtain the certificate chain. Then additional checks will be
13808 performed on the chain. Currently the root CA is checked to see
13809 if it is explicitly trusted for OCSP signing. This is used to set
13810 a root CA as a global signing root: that is any certificate that
13811 chains to that CA is an acceptable OCSP signing certificate.
13815 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13816 extensions from a separate configuration file.
13817 As when reading extensions from the main configuration file,
13818 the '-extensions ...' option may be used for specifying the
13821 *Massimiliano Pala <madwolf@comune.modena.it>*
13823 * New OCSP utility. Allows OCSP requests to be generated or
13824 read. The request can be sent to a responder and the output
13825 parsed, outputted or printed in text form. Not complete yet:
13826 still needs to check the OCSP response validity.
13830 * New subcommands for 'openssl ca':
13831 `openssl ca -status <serial>` prints the status of the cert with
13832 the given serial number (according to the index file).
13833 `openssl ca -updatedb` updates the expiry status of certificates
13836 *Massimiliano Pala <madwolf@comune.modena.it>*
13838 * New '-newreq-nodes' command option to CA.pl. This is like
13839 '-newreq', but calls 'openssl req' with the '-nodes' option
13840 so that the resulting key is not encrypted.
13842 *Damien Miller <djm@mindrot.org>*
13844 * New configuration for the GNU Hurd.
13846 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13848 * Initial code to implement OCSP basic response verify. This
13849 is currently incomplete. Currently just finds the signer's
13850 certificate and verifies the signature on the response.
13854 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13855 value of OPENSSLDIR. This is available via the new '-d' option
13856 to 'openssl version', and is also included in 'openssl version -a'.
13860 * Allowing defining memory allocation callbacks that will be given
13861 file name and line number information in additional arguments
13862 (a `const char*` and an int). The basic functionality remains, as
13863 well as the original possibility to just replace malloc(),
13864 realloc() and free() by functions that do not know about these
13865 additional arguments. To register and find out the current
13866 settings for extended allocation functions, the following
13867 functions are provided:
13869 CRYPTO_set_mem_ex_functions
13870 CRYPTO_set_locked_mem_ex_functions
13871 CRYPTO_get_mem_ex_functions
13872 CRYPTO_get_locked_mem_ex_functions
13874 These work the same way as CRYPTO_set_mem_functions and friends.
13875 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13876 extended allocation function is enabled.
13877 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13878 a conventional allocation function is enabled.
13880 *Richard Levitte, Bodo Moeller*
13882 * Finish off removing the remaining LHASH function pointer casts.
13883 There should no longer be any prototype-casting required when using
13884 the LHASH abstraction, and any casts that remain are "bugs". See
13885 the callback types and macros at the head of lhash.h for details
13886 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13890 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13891 If /dev/[u]random devices are not available or do not return enough
13892 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13894 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13895 /etc/entropy will be queried once each in this sequence, querying stops
13896 when enough entropy was collected without querying more sockets.
13900 * Change the Unix RAND_poll() variant to be able to poll several
13901 random devices, as specified by DEVRANDOM, until a sufficient amount
13902 of data has been collected. We spend at most 10 ms on each file
13903 (select timeout) and read in non-blocking mode. DEVRANDOM now
13904 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13905 (previously it was just the string "/dev/urandom"), so on typical
13906 platforms the 10 ms delay will never occur.
13907 Also separate out the Unix variant to its own file, rand_unix.c.
13908 For VMS, there's a currently-empty rand_vms.c.
13912 * Move OCSP client related routines to ocsp_cl.c. These
13913 provide utility functions which an application needing
13914 to issue a request to an OCSP responder and analyse the
13915 response will typically need: as opposed to those which an
13916 OCSP responder itself would need which will be added later.
13918 OCSP_request_sign() signs an OCSP request with an API similar
13919 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13920 response. OCSP_response_get1_basic() extracts basic response
13921 from response. OCSP_resp_find_status(): finds and extracts status
13922 information from an OCSP_CERTID structure (which will be created
13923 when the request structure is built). These are built from lower
13924 level functions which work on OCSP_SINGLERESP structures but
13925 won't normally be used unless the application wishes to examine
13926 extensions in the OCSP response for example.
13928 Replace nonce routines with a pair of functions.
13929 OCSP_request_add1_nonce() adds a nonce value and optionally
13930 generates a random value. OCSP_check_nonce() checks the
13931 validity of the nonce in an OCSP response.
13935 * Change function OCSP_request_add() to OCSP_request_add0_id().
13936 This doesn't copy the supplied OCSP_CERTID and avoids the
13937 need to free up the newly created id. Change return type
13938 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13939 This can then be used to add extensions to the request.
13940 Deleted OCSP_request_new(), since most of its functionality
13941 is now in OCSP_REQUEST_new() (and the case insensitive name
13942 clash) apart from the ability to set the request name which
13943 will be added elsewhere.
13947 * Update OCSP API. Remove obsolete extensions argument from
13948 various functions. Extensions are now handled using the new
13949 OCSP extension code. New simple OCSP HTTP function which
13950 can be used to send requests and parse the response.
13954 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13955 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13956 uses the special reorder version of SET OF to sort the attributes
13957 and reorder them to match the encoded order. This resolves a long
13958 standing problem: a verify on a PKCS7 structure just after signing
13959 it used to fail because the attribute order did not match the
13960 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13961 it uses the received order. This is necessary to tolerate some broken
13962 software that does not order SET OF. This is handled by encoding
13963 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13964 to produce the required SET OF.
13968 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13969 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13970 files to get correct declarations of the ASN.1 item variables.
13974 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13975 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13976 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13977 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13978 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13979 ASN1_ITEM and no wrapper functions.
13983 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13984 replace the old function pointer based I/O routines. Change most of
13985 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13989 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13990 lines, recognize more "algorithms" that can be deselected, and make
13991 it complain about algorithm deselection that isn't recognised.
13995 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13996 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13997 to use new functions. Add NO_ASN1_OLD which can be set to remove
13998 some old style ASN1 functions: this can be used to determine if old
13999 code will still work when these eventually go away.
14003 * New extension functions for OCSP structures, these follow the
14004 same conventions as certificates and CRLs.
14008 * New function X509V3_add1_i2d(). This automatically encodes and
14009 adds an extension. Its behaviour can be customised with various
14010 flags to append, replace or delete. Various wrappers added for
14011 certificates and CRLs.
14015 * Fix to avoid calling the underlying ASN1 print routine when
14016 an extension cannot be parsed. Correct a typo in the
14017 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14021 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14022 entries for variables.
14026 * Add functionality to `apps/openssl.c` for detecting locking
14027 problems: As the program is single-threaded, all we have
14028 to do is register a locking callback using an array for
14029 storing which locks are currently held by the program.
14033 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14034 SSL_get_ex_data_X509_STORE_idx(), which is used in
14035 ssl_verify_cert_chain() and thus can be called at any time
14036 during TLS/SSL handshakes so that thread-safety is essential.
14037 Unfortunately, the ex_data design is not at all suited
14038 for multi-threaded use, so it probably should be abolished.
14042 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14044 *Broadcom, tweaked and integrated by Geoff Thorpe*
14046 * Move common extension printing code to new function
14047 X509V3_print_extensions(). Reorganise OCSP print routines and
14048 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14052 * New function X509_signature_print() to remove duplication in some
14057 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14058 set (this was treated exactly the same as SET OF previously). This
14059 is used to reorder the STACK representing the structure to match the
14060 encoding. This will be used to get round a problem where a PKCS7
14061 structure which was signed could not be verified because the STACK
14062 order did not reflect the encoded order.
14066 * Reimplement the OCSP ASN1 module using the new code.
14070 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14071 for its ASN1 operations. The old style function pointers still exist
14072 for now but they will eventually go away.
14076 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14077 completely replaces the old ASN1 functionality with a table driven
14078 encoder and decoder which interprets an ASN1_ITEM structure describing
14079 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14080 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14081 has also been converted to the new form.
14085 * Change BN_mod_exp_recp so that negative moduli are tolerated
14086 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14087 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14088 for negative moduli.
14092 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14093 of not touching the result's sign bit.
14097 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14102 * Changed the LHASH code to use prototypes for callbacks, and created
14103 macros to declare and implement thin (optionally static) functions
14104 that provide type-safety and avoid function pointer casting for the
14105 type-specific callbacks.
14109 * Added Kerberos Cipher Suites to be used with TLS, as written in
14111 *Veers Staats <staatsvr@asc.hpc.mil>,
14112 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14114 * Reformat the FAQ so the different questions and answers can be divided
14115 in sections depending on the subject.
14119 * Have the zlib compression code load ZLIB.DLL dynamically under
14124 * New function BN_mod_sqrt for computing square roots modulo a prime
14125 (using the probabilistic Tonelli-Shanks algorithm unless
14126 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14127 be handled deterministically).
14129 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14131 * Make BN_mod_inverse faster by explicitly handling small quotients
14132 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14133 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14137 * New function BN_kronecker.
14141 * Fix BN_gcd so that it works on negative inputs; the result is
14142 positive unless both parameters are zero.
14143 Previously something reasonably close to an infinite loop was
14144 possible because numbers could be growing instead of shrinking
14145 in the implementation of Euclid's algorithm.
14149 * Fix BN_is_word() and BN_is_one() macros to take into account the
14150 sign of the number in question.
14152 Fix BN_is_word(a,w) to work correctly for w == 0.
14154 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14155 because its test if the absolute value of 'a' equals 'w'.
14156 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14157 it exists mostly for use in the implementations of BN_is_zero(),
14158 BN_is_one(), and BN_is_word().
14162 * New function BN_swap.
14166 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14167 the exponentiation functions are more likely to produce reasonable
14168 results on negative inputs.
14172 * Change BN_mod_mul so that the result is always non-negative.
14173 Previously, it could be negative if one of the factors was negative;
14174 I don't think anyone really wanted that behaviour.
14178 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14179 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14180 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14181 and add new functions:
14190 BN_mod_lshift1_quick
14192 BN_mod_lshift_quick
14194 These functions always generate non-negative results.
14196 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14197 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14199 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14200 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14201 be reduced modulo `m`.
14203 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14206 The following entry accidentally appeared in the CHANGES file
14207 distributed with OpenSSL 0.9.7. The modifications described in
14208 it do *not* apply to OpenSSL 0.9.7.
14210 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14211 was actually never needed) and in BN_mul(). The removal in BN_mul()
14212 required a small change in bn_mul_part_recursive() and the addition
14213 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14214 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14215 bn_sub_words() and bn_add_words() except they take arrays with
14221 * In 'openssl passwd', verify passwords read from the terminal
14222 unless the '-salt' option is used (which usually means that
14223 verification would just waste user's time since the resulting
14224 hash is going to be compared with some given password hash)
14225 or the new '-noverify' option is used.
14227 This is an incompatible change, but it does not affect
14228 non-interactive use of 'openssl passwd' (passwords on the command
14229 line, '-stdin' option, '-in ...' option) and thus should not
14230 cause any problems.
14234 * Remove all references to RSAref, since there's no more need for it.
14238 * Make DSO load along a path given through an environment variable
14239 (SHLIB_PATH) with shl_load().
14243 * Constify the ENGINE code as a result of BIGNUM constification.
14244 Also constify the RSA code and most things related to it. In a
14245 few places, most notable in the depth of the ASN.1 code, ugly
14246 casts back to non-const were required (to be solved at a later
14251 * Make it so the openssl application has all engines loaded by default.
14255 * Constify the BIGNUM routines a little more.
14259 * Add the following functions:
14261 ENGINE_load_cswift()
14263 ENGINE_load_atalla()
14264 ENGINE_load_nuron()
14265 ENGINE_load_builtin_engines()
14267 That way, an application can itself choose if external engines that
14268 are built-in in OpenSSL shall ever be used or not. The benefit is
14269 that applications won't have to be linked with libdl or other dso
14270 libraries unless it's really needed.
14272 Changed 'openssl engine' to load all engines on demand.
14273 Changed the engine header files to avoid the duplication of some
14274 declarations (they differed!).
14278 * 'openssl engine' can now list capabilities.
14282 * Better error reporting in 'openssl engine'.
14286 * Never call load_dh_param(NULL) in s_server.
14290 * Add engine application. It can currently list engines by name and
14291 identity, and test if they are actually available.
14295 * Improve RPM specification file by forcing symbolic linking and making
14296 sure the installed documentation is also owned by root.root.
14298 *Damien Miller <djm@mindrot.org>*
14300 * Give the OpenSSL applications more possibilities to make use of
14301 keys (public as well as private) handled by engines.
14305 * Add OCSP code that comes from CertCo.
14309 * Add VMS support for the Rijndael code.
14313 * Added untested support for Nuron crypto accelerator.
14317 * Add support for external cryptographic devices. This code was
14318 previously distributed separately as the "engine" branch.
14320 *Geoff Thorpe, Richard Levitte*
14322 * Rework the filename-translation in the DSO code. It is now possible to
14323 have far greater control over how a "name" is turned into a filename
14324 depending on the operating environment and any oddities about the
14325 different shared library filenames on each system.
14329 * Support threads on FreeBSD-elf in Configure.
14333 * Fix for SHA1 assembly problem with MASM: it produces
14334 warnings about corrupt line number information when assembling
14335 with debugging information. This is caused by the overlapping
14338 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14341 NCONF_get_number() has no error checking at all. As a replacement,
14342 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14343 promoted strongly. The old NCONF_get_number is kept around for
14344 binary backward compatibility.
14345 Make it possible for methods to load from something other than a BIO,
14346 by providing a function pointer that is given a name instead of a BIO.
14347 For example, this could be used to load configuration data from an
14352 * Fix for non blocking accept BIOs. Added new I/O special reason
14353 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14354 with non blocking I/O was not possible because no retry code was
14355 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14360 * Added the beginnings of Rijndael support.
14364 * Fix for bug in DirectoryString mask setting. Add support for
14365 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14366 to allow certificate printing to more controllable, additional
14367 'certopt' option to 'x509' to allow new printing options to be
14372 * Clean old EAY MD5 hack from e_os.h.
14376 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14378 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14379 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14381 *Joe Orton, Steve Henson*
14383 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14385 * Fix additional bug revealed by the NISCC test suite:
14387 Stop bug triggering large recursion when presented with
14388 certain ASN.1 tags ([CVE-2003-0851])
14392 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14394 * Fix various bugs revealed by running the NISCC test suite:
14396 Stop out of bounds reads in the ASN1 code when presented with
14397 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14399 If verify callback ignores invalid public key errors don't try to check
14400 certificate signature with the NULL public key.
14404 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14405 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14410 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14411 extra data after the compression methods not only for TLS 1.0
14412 but also for SSL 3.0 (as required by the specification).
14414 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14416 * Change X509_certificate_type() to mark the key as exported/exportable
14417 when it's 512 *bits* long, not 512 bytes.
14421 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14423 * Countermeasure against the Klima-Pokorny-Rosa extension of
14424 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14425 a protocol version number mismatch like a decryption error
14426 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14430 * Turn on RSA blinding by default in the default implementation
14431 to avoid a timing attack. Applications that don't want it can call
14432 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14433 They would be ill-advised to do so in most cases.
14435 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14437 * Change RSA blinding code so that it works when the PRNG is not
14438 seeded (in this case, the secret RSA exponent is abused as
14439 an unpredictable seed -- if it is not unpredictable, there
14440 is no point in blinding anyway). Make RSA blinding thread-safe
14441 by remembering the creator's thread ID in rsa->blinding and
14442 having all other threads use local one-time blinding factors
14443 (this requires more computation than sharing rsa->blinding, but
14444 avoids excessive locking; and if an RSA object is not shared
14445 between threads, blinding will still be very fast).
14449 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14451 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14452 via timing by performing a MAC computation even if incorrect
14453 block cipher padding has been found. This is a countermeasure
14454 against active attacks where the attacker has to distinguish
14455 between bad padding and a MAC verification error. ([CVE-2003-0078])
14457 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14458 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14459 Martin Vuagnoux (EPFL, Ilion)*
14461 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14463 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14464 memory from its contents. This is done with a counter that will
14465 place alternating values in each byte. This can be used to solve
14466 two issues: 1) the removal of calls to memset() by highly optimizing
14467 compilers, and 2) cleansing with other values than 0, since those can
14468 be read through on certain media, for example a swap space on disk.
14472 * Bugfix: client side session caching did not work with external caching,
14473 because the session->cipher setting was not restored when reloading
14474 from the external cache. This problem was masked, when
14475 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14476 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14480 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14481 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14483 *Zeev Lieber <zeev-l@yahoo.com>*
14485 * Undo an undocumented change introduced in 0.9.6e which caused
14486 repeated calls to OpenSSL_add_all_ciphers() and
14487 OpenSSL_add_all_digests() to be ignored, even after calling
14492 * Change the default configuration reader to deal with last line not
14493 being properly terminated.
14497 * Change X509_NAME_cmp() so it applies the special rules on handling
14498 DN values that are of type PrintableString, as well as RDNs of type
14499 emailAddress where the value has the type ia5String.
14501 *stefank@valicert.com via Richard Levitte*
14503 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14504 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14505 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14506 the bitwise-OR of the two for use by the majority of applications
14507 wanting this behaviour, and update the docs. The documented
14508 behaviour and actual behaviour were inconsistent and had been
14509 changing anyway, so this is more a bug-fix than a behavioural
14512 *Geoff Thorpe, diagnosed by Nadav Har'El*
14514 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14515 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14519 * Fix initialization code race conditions in
14520 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14521 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14522 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14523 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14524 ssl2_get_cipher_by_char(),
14525 ssl3_get_cipher_by_char().
14527 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14529 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14530 the cached sessions are flushed, as the remove_cb() might use ex_data
14531 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14532 (see [openssl.org #212]).
14534 *Geoff Thorpe, Lutz Jaenicke*
14536 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14537 length, instead of the encoding length to d2i_ASN1_OBJECT.
14541 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14543 * [In 0.9.6g-engine release:]
14544 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14546 *Lynn Gazis <lgazis@rainbow.com>*
14548 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14550 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14551 and get fix the header length calculation.
14552 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14553 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14555 * Use proper error handling instead of 'assertions' in buffer
14556 overflow checks added in 0.9.6e. This prevents DoS (the
14557 assertions could call abort()).
14559 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14561 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14563 * Add various sanity checks to asn1_get_length() to reject
14564 the ASN1 length bytes if they exceed sizeof(long), will appear
14565 negative or the content length exceeds the length of the
14568 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14570 * Fix cipher selection routines: ciphers without encryption had no flags
14571 for the cipher strength set and where therefore not handled correctly
14572 by the selection routines (PR #130).
14576 * Fix EVP_dsa_sha macro.
14581 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14582 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14583 that was added in OpenSSL 0.9.6d.
14585 As the countermeasure turned out to be incompatible with some
14586 broken SSL implementations, the new option is part of SSL_OP_ALL.
14587 SSL_OP_ALL is usually employed when compatibility with weird SSL
14588 implementations is desired (e.g. '-bugs' option to 's_client' and
14589 's_server'), so the new option is automatically set in many
14594 * Changes in security patch:
14596 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14597 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14598 Air Force Materiel Command, USAF, under agreement number
14601 * Add various sanity checks to asn1_get_length() to reject
14602 the ASN1 length bytes if they exceed sizeof(long), will appear
14603 negative or the content length exceeds the length of the
14604 supplied buffer. ([CVE-2002-0659])
14606 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14608 * Assertions for various potential buffer overflows, not known to
14609 happen in practice.
14611 *Ben Laurie (CHATS)*
14613 * Various temporary buffers to hold ASCII versions of integers were
14614 too small for 64 bit platforms. ([CVE-2002-0655])
14615 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14617 * Remote buffer overflow in SSL3 protocol - an attacker could
14618 supply an oversized session ID to a client. ([CVE-2002-0656])
14620 *Ben Laurie (CHATS)*
14622 * Remote buffer overflow in SSL2 protocol - an attacker could
14623 supply an oversized client master key. ([CVE-2002-0656])
14625 *Ben Laurie (CHATS)*
14627 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14629 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14630 encoded as NULL) with id-dsa-with-sha1.
14632 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14634 * Check various `X509_...()` return values in `apps/req.c`.
14636 *Nils Larsch <nla@trustcenter.de>*
14638 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14639 an end-of-file condition would erroneously be flagged, when the CRLF
14640 was just at the end of a processed block. The bug was discovered when
14641 processing data through a buffering memory BIO handing the data to a
14642 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14643 <ptsekov@syntrex.com> and Nedelcho Stanev.
14647 * Implement a countermeasure against a vulnerability recently found
14648 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14649 before application data chunks to avoid the use of known IVs
14650 with data potentially chosen by the attacker.
14654 * Fix length checks in ssl3_get_client_hello().
14658 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14659 to prevent ssl3_read_internal() from incorrectly assuming that
14660 ssl3_read_bytes() found application data while handshake
14661 processing was enabled when in fact s->s3->in_read_app_data was
14662 merely automatically cleared during the initial handshake.
14664 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14666 * Fix object definitions for Private and Enterprise: they were not
14667 recognized in their shortname (=lowercase) representation. Extend
14668 obj_dat.pl to issue an error when using undefined keywords instead
14669 of silently ignoring the problem (Svenning Sorensen
14670 <sss@sss.dnsalias.net>).
14674 * Fix DH_generate_parameters() so that it works for 'non-standard'
14675 generators, i.e. generators other than 2 and 5. (Previously, the
14676 code did not properly initialise the 'add' and 'rem' values to
14677 BN_generate_prime().)
14679 In the new general case, we do not insist that 'generator' is
14680 actually a primitive root: This requirement is rather pointless;
14681 a generator of the order-q subgroup is just as good, if not
14686 * Map new X509 verification errors to alerts. Discovered and submitted by
14687 Tom Wu <tom@arcot.com>.
14691 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14692 returning non-zero before the data has been completely received
14693 when using non-blocking I/O.
14695 *Bodo Moeller; problem pointed out by John Hughes*
14697 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14699 *Ben Laurie, Lutz Jaenicke*
14701 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14702 Yoram Zahavi <YoramZ@gilian.com>).
14706 * Add information about CygWin 1.3 and on, and preserve proper
14707 configuration for the versions before that.
14709 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14711 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14712 check whether we deal with a copy of a session and do not delete from
14713 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14714 <izhar@checkpoint.com>.
14718 * Do not store session data into the internal session cache, if it
14719 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14720 flag is set). Proposed by Aslam <aslam@funk.com>.
14724 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14729 * [In 0.9.6d-engine release:]
14730 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14732 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14734 * Add the configuration target linux-s390x.
14736 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14738 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14739 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14740 variable as an indication that a ClientHello message has been
14741 received. As the flag value will be lost between multiple
14742 invocations of ssl3_accept when using non-blocking I/O, the
14743 function may not be aware that a handshake has actually taken
14744 place, thus preventing a new session from being added to the
14747 To avoid this problem, we now set s->new_session to 2 instead of
14748 using a local variable.
14750 *Lutz Jaenicke, Bodo Moeller*
14752 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14753 if the SSL_R_LENGTH_MISMATCH error is detected.
14755 *Geoff Thorpe, Bodo Moeller*
14757 * New 'shared_ldflag' column in Configure platform table.
14761 * Fix EVP_CIPHER_mode macro.
14763 *"Dan S. Camper" <dan@bti.net>*
14765 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14766 type, we must throw them away by setting rr->length to 0.
14768 *D P Chang <dpc@qualys.com>*
14770 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14772 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14773 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14774 worked incorrectly for those cases where range = `10..._2` and
14775 `3*range` is two bits longer than range.)
14779 * Only add signing time to PKCS7 structures if it is not already
14784 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14785 OBJ_ld_ce should be OBJ_id_ce.
14786 Also some ip-pda OIDs in crypto/objects/objects.txt were
14787 incorrect (cf. RFC 3039).
14789 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14791 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14792 returns early because it has nothing to do.
14794 *Andy Schneider <andy.schneider@bjss.co.uk>*
14796 * [In 0.9.6c-engine release:]
14797 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14799 *Andy Schneider <andy.schneider@bjss.co.uk>*
14801 * [In 0.9.6c-engine release:]
14802 Add support for Cryptographic Appliance's keyserver technology.
14803 (Use engine 'keyclient')
14805 *Cryptographic Appliances and Geoff Thorpe*
14807 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14808 is called via tools/c89.sh because arguments have to be
14809 rearranged (all '-L' options must appear before the first object
14812 *Richard Shapiro <rshapiro@abinitio.com>*
14814 * [In 0.9.6c-engine release:]
14815 Add support for Broadcom crypto accelerator cards, backported
14818 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14820 * [In 0.9.6c-engine release:]
14821 Add support for SureWare crypto accelerator cards from
14822 Baltimore Technologies. (Use engine 'sureware')
14824 *Baltimore Technologies and Mark Cox*
14826 * [In 0.9.6c-engine release:]
14827 Add support for crypto accelerator cards from Accelerated
14828 Encryption Processing, www.aep.ie. (Use engine 'aep')
14830 *AEP Inc. and Mark Cox*
14832 * Add a configuration entry for gcc on UnixWare.
14834 *Gary Benson <gbenson@redhat.com>*
14836 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14837 messages are stored in a single piece (fixed-length part and
14838 variable-length part combined) and fix various bugs found on the way.
14842 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14843 instead. BIO_gethostbyname() does not know what timeouts are
14844 appropriate, so entries would stay in cache even when they have
14846 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14848 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14849 faced with a pathologically small ClientHello fragment that does
14850 not contain client_version: Instead of aborting with an error,
14851 simply choose the highest available protocol version (i.e.,
14852 TLS 1.0 unless it is disabled). In practice, ClientHello
14853 messages are never sent like this, but this change gives us
14854 strictly correct behaviour at least for TLS.
14858 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14859 never resets s->method to s->ctx->method when called from within
14860 one of the SSL handshake functions.
14862 *Bodo Moeller; problem pointed out by Niko Baric*
14864 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14865 (sent using the client's version number) if client_version is
14866 smaller than the protocol version in use. Also change
14867 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14868 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14869 the client will at least see that alert.
14873 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14878 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14879 client receives HelloRequest while in a handshake.
14881 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14883 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14884 should end in 'break', not 'goto end' which circumvents various
14885 cleanups done in state SSL_ST_OK. But session related stuff
14886 must be disabled for SSL_ST_OK in the case that we just sent a
14889 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14890 before just sending a HelloRequest.
14892 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14894 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14895 reveal whether illegal block cipher padding was found or a MAC
14896 verification error occurred. (Neither SSLerr() codes nor alerts
14897 are directly visible to potential attackers, but the information
14898 may leak via logfiles.)
14900 Similar changes are not required for the SSL 2.0 implementation
14901 because the number of padding bytes is sent in clear for SSL 2.0,
14902 and the extra bytes are just ignored. However ssl/s2_pkt.c
14903 failed to verify that the purported number of padding bytes is in
14908 * Add OpenUNIX-8 support including shared libraries
14909 (Boyd Lynn Gerber <gerberb@zenez.com>).
14913 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14914 'wristwatch attack' using huge encoding parameters (cf.
14915 James H. Manger's CRYPTO 2001 paper). Note that the
14916 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14917 encoding parameters and hence was not vulnerable.
14921 * BN_sqr() bug fix.
14923 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14925 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14926 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14927 followed by modular reduction.
14929 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14931 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14932 equivalent based on BN_pseudo_rand() instead of BN_rand().
14936 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14937 This function was broken, as the check for a new client hello message
14938 to handle SGC did not allow these large messages.
14939 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14943 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14947 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14948 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14952 * Rework the configuration and shared library support for Tru64 Unix.
14953 The configuration part makes use of modern compiler features and
14954 still retains old compiler behavior for those that run older versions
14955 of the OS. The shared library support part includes a variant that
14956 uses the RPATH feature, and is available through the special
14957 configuration target "alpha-cc-rpath", which will never be selected
14960 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14962 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14963 with the same message size as in ssl3_get_certificate_request().
14964 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14965 messages might inadvertently be reject as too long.
14967 *Petr Lampa <lampa@fee.vutbr.cz>*
14969 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14973 * Modified SSL library such that the verify_callback that has been set
14974 specifically for an SSL object with SSL_set_verify() is actually being
14975 used. Before the change, a verify_callback set with this function was
14976 ignored and the verify_callback() set in the SSL_CTX at the time of
14977 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14978 to allow the necessary settings.
14982 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14983 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14984 done automatically (in contradiction to the requirements of the C
14985 standard). This made problems when used from OpenSSH.
14989 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14990 dh->length and always used
14992 BN_rand_range(priv_key, dh->p).
14994 BN_rand_range() is not necessary for Diffie-Hellman, and this
14995 specific range makes Diffie-Hellman unnecessarily inefficient if
14996 dh->length (recommended exponent length) is much smaller than the
14997 length of dh->p. We could use BN_rand_range() if the order of
14998 the subgroup was stored in the DH structure, but we only have
15003 BN_rand(priv_key, l, ...)
15005 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15012 RSA_eay_public_encrypt
15013 RSA_eay_private_decrypt
15014 RSA_eay_private_encrypt (signing)
15015 RSA_eay_public_decrypt (signature verification)
15017 (default implementations for RSA_public_encrypt,
15018 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15019 always reject numbers >= n.
15023 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15024 to synchronize access to 'locking_thread'. This is necessary on
15025 systems where access to 'locking_thread' (an 'unsigned long'
15026 variable) is not atomic.
15030 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15031 *before* setting the 'crypto_lock_rand' flag. The previous code had
15032 a race condition if 0 is a valid thread ID.
15034 *Travis Vitek <vitek@roguewave.com>*
15036 * Add support for shared libraries under Irix.
15038 *Albert Chin-A-Young <china@thewrittenword.com>*
15040 * Add configuration option to build on Linux on both big-endian and
15041 little-endian MIPS.
15043 *Ralf Baechle <ralf@uni-koblenz.de>*
15045 * Add the possibility to create shared libraries on HP-UX.
15049 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15051 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15052 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15053 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15054 PRNG state recovery was possible based on the output of
15055 one PRNG request appropriately sized to gain knowledge on
15056 'md' followed by enough consecutive 1-byte PRNG requests
15057 to traverse all of 'state'.
15059 1. When updating 'md_local' (the current thread's copy of 'md')
15060 during PRNG output generation, hash all of the previous
15061 'md_local' value, not just the half used for PRNG output.
15063 2. Make the number of bytes from 'state' included into the hash
15064 independent from the number of PRNG bytes requested.
15066 The first measure alone would be sufficient to avoid
15067 Markku-Juhani's attack. (Actually it had never occurred
15068 to me that the half of 'md_local' used for chaining was the
15069 half from which PRNG output bytes were taken -- I had always
15070 assumed that the secret half would be used.) The second
15071 measure makes sure that additional data from 'state' is never
15072 mixed into 'md_local' in small portions; this heuristically
15073 further strengthens the PRNG.
15077 * Fix crypto/bn/asm/mips3.s.
15081 * When only the key is given to "enc", the IV is undefined. Print out
15082 an error message in this case.
15086 * Handle special case when X509_NAME is empty in X509 printing routines.
15090 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15091 positive and less than q.
15095 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15096 used: it isn't thread safe and the add_lock_callback should handle
15099 *Paul Rose <Paul.Rose@bridge.com>*
15101 * Verify that incoming data obeys the block size in
15102 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15108 *Ulf Möller, Bodo Möller*
15110 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15111 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15112 when fixing the server behaviour for backwards-compatible 'client
15113 hello' messages. (Note that the attack is impractical against
15114 SSL 3.0 and TLS 1.0 anyway because length and version checking
15115 means that the probability of guessing a valid ciphertext is
15116 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15119 Before 0.9.5, the countermeasure (hide the error by generating a
15120 random 'decryption result') did not work properly because
15121 ERR_clear_error() was missing, meaning that SSL_get_error() would
15122 detect the supposedly ignored error.
15124 Both problems are now fixed.
15128 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15129 (previously it was 1024).
15133 * Fix for compatibility mode trust settings: ignore trust settings
15134 unless some valid trust or reject settings are present.
15138 * Fix for blowfish EVP: its a variable length cipher.
15142 * Fix various bugs related to DSA S/MIME verification. Handle missing
15143 parameters in DSA public key structures and return an error in the
15144 DSA routines if parameters are absent.
15148 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15149 in the current directory if neither $RANDFILE nor $HOME was set.
15150 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15151 caused some confusion to Windows users who haven't defined $HOME.
15152 Thus RAND_file_name() is changed again: e_os.h can define a
15153 DEFAULT_HOME, which will be used if $HOME is not set.
15154 For Windows, we use "C:"; on other platforms, we still require
15155 environment variables.
15157 * Move 'if (!initialized) RAND_poll()' into regions protected by
15158 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15159 having multiple threads call RAND_poll() concurrently.
15163 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15164 combination of a flag and a thread ID variable.
15165 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15166 flag), *other* threads can enter ssleay_add_bytes without obeying
15167 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15168 that they do not hold after the first thread unsets add_do_not_lock).
15172 * Change bctest again: '-x' expressions are not available in all
15173 versions of 'test'.
15177 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15179 * Fix a couple of memory leaks in PKCS7_dataDecode()
15181 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15183 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15184 the default extension for executables, if any. Also, make the perl
15185 scripts that use symlink() to test if it really exists and use "cp"
15186 if it doesn't. All this made OpenSSL compilable and installable in
15191 * Fix for asn1_GetSequence() for indefinite length constructed data.
15192 If SEQUENCE is length is indefinite just set c->slen to the total
15193 amount of data available.
15195 *Steve Henson, reported by shige@FreeBSD.org*
15197 *This change does not apply to 0.9.7.*
15199 * Change bctest to avoid here-documents inside command substitution
15200 (workaround for FreeBSD /bin/sh bug).
15201 For compatibility with Ultrix, avoid shell functions (introduced
15202 in the bctest version that searches along $PATH).
15206 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15207 with des_encrypt() defined on some operating systems, like Solaris
15212 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15213 On the Importance of Eliminating Errors in Cryptographic
15214 Computations, J. Cryptology 14 (2001) 2, 101-119,
15215 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15219 * MIPS assembler BIGNUM division bug fix.
15223 * Disabled incorrect Alpha assembler code.
15227 * Fix PKCS#7 decode routines so they correctly update the length
15228 after reading an EOC for the EXPLICIT tag.
15232 *This change does not apply to 0.9.7.*
15234 * Fix bug in PKCS#12 key generation routines. This was triggered
15235 if a 3DES key was generated with a 0 initial byte. Include
15236 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15237 (but broken) behaviour.
15241 * Enhance bctest to search for a working bc along $PATH and print
15244 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15246 * Fix memory leaks in err.c: free err_data string if necessary;
15247 don't write to the wrong index in ERR_set_error_data.
15251 * Implement ssl23_peek (analogous to ssl23_read), which previously
15256 * Replace rdtsc with `_emit` statements for VC++ version 5.
15258 *Jeremy Cooper <jeremy@baymoo.org>*
15260 * Make it possible to reuse SSLv2 sessions.
15264 * In copy_email() check for >= 0 as a return value for
15265 X509_NAME_get_index_by_NID() since 0 is a valid index.
15267 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15269 * Avoid coredump with unsupported or invalid public keys by checking if
15270 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15271 PKCS7_verify() fails with non detached data.
15275 * Don't use getenv in library functions when run as setuid/setgid.
15276 New function OPENSSL_issetugid().
15280 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15281 due to incorrect handling of multi-threading:
15283 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15285 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15287 3. Count how many times MemCheck_off() has been called so that
15288 nested use can be treated correctly. This also avoids
15289 inband-signalling in the previous code (which relied on the
15290 assumption that thread ID 0 is impossible).
15294 * Add "-rand" option also to s_client and s_server.
15298 * Fix CPU detection on Irix 6.x.
15299 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15300 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15302 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15307 *This change does not apply to 0.9.7.*
15309 * Use the cached encoding of an X509_NAME structure rather than
15310 copying it. This is apparently the reason for the libsafe "errors"
15311 but the code is actually correct.
15315 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15316 Bleichenbacher's DSA attack.
15317 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15318 to be set and top=0 forces the highest bit to be set; top=-1 is new
15319 and leaves the highest bit random.
15321 *Ulf Moeller, Bodo Moeller*
15323 * In the `NCONF_...`-based implementations for `CONF_...` queries
15324 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15325 a temporary CONF structure with the data component set to NULL
15326 (which gives segmentation faults in lh_retrieve).
15327 Instead, use NULL for the CONF pointer in CONF_get_string and
15328 CONF_get_number (which may use environment variables) and directly
15329 return NULL from CONF_get_section.
15333 * Fix potential buffer overrun for EBCDIC.
15337 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15338 keyUsage if basicConstraints absent for a CA.
15342 * Make SMIME_write_PKCS7() write mail header values with a format that
15343 is more generally accepted (no spaces before the semicolon), since
15344 some programs can't parse those values properly otherwise. Also make
15345 sure BIO's that break lines after each write do not create invalid
15350 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15351 macros previously used would not encode an empty SEQUENCE OF
15352 and break the signature.
15356 *This change does not apply to 0.9.7.*
15358 * Zero the premaster secret after deriving the master secret in
15363 * Add some EVP_add_digest_alias registrations (as found in
15364 OpenSSL_add_all_digests()) to SSL_library_init()
15365 aka OpenSSL_add_ssl_algorithms(). This provides improved
15366 compatibility with peers using X.509 certificates
15367 with unconventional AlgorithmIdentifier OIDs.
15371 * Fix for Irix with NO_ASM.
15373 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15375 * ./config script fixes.
15377 *Ulf Moeller, Richard Levitte*
15379 * Fix 'openssl passwd -1'.
15383 * Change PKCS12_key_gen_asc() so it can cope with non null
15384 terminated strings whose length is passed in the passlen
15385 parameter, for example from PEM callbacks. This was done
15386 by adding an extra length parameter to asc2uni().
15388 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15390 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15391 call failed, free the DSA structure.
15395 * Fix to uni2asc() to cope with zero length Unicode strings.
15396 These are present in some PKCS#12 files.
15400 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15401 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15402 when writing a 32767 byte record.
15404 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15406 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15407 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15409 (RSA objects have a reference count access to which is protected
15410 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15411 so they are meant to be shared between threads.)
15412 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15413 "Reddie, Steven" <Steven.Reddie@ca.com>*
15415 * Fix a deadlock in CRYPTO_mem_leaks().
15419 * Use better test patterns in bntest.
15423 * rand_win.c fix for Borland C.
15427 * BN_rshift bugfix for n == 0.
15431 * Add a 'bctest' script that checks for some known 'bc' bugs
15432 so that 'make test' does not abort just because 'bc' is broken.
15436 * Store verify_result within SSL_SESSION also for client side to
15437 avoid potential security hole. (Re-used sessions on the client side
15438 always resulted in verify_result==X509_V_OK, not using the original
15439 result of the server certificate verification.)
15443 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15444 SSL3_RT_APPLICATION_DATA, return 0.
15445 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15450 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15451 releases, have been re-implemented by renaming the previous
15452 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15453 and ssl3_read_internal, respectively, and adding 'peek' parameters
15454 to them. The new ssl[23]_{read,peek} functions are calls to
15455 ssl[23]_read_internal with the 'peek' flag set appropriately.
15456 A 'peek' parameter has also been added to ssl3_read_bytes, which
15457 does the actual work for ssl3_read_internal.
15461 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15462 the method-specific "init()" handler. Also clean up ex_data after
15463 calling the method-specific "finish()" handler. Previously, this was
15464 happening the other way round.
15468 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15469 The previous value, 12, was not always sufficient for BN_mod_exp().
15473 * Make sure that shared libraries get the internal name engine with
15474 the full version number and not just 0. This should mark the
15475 shared libraries as not backward compatible. Of course, this should
15476 be changed again when we can guarantee backward binary compatibility.
15480 * Fix typo in get_cert_by_subject() in by_dir.c
15482 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15484 * Rework the system to generate shared libraries:
15486 - Make note of the expected extension for the shared libraries and
15487 if there is a need for symbolic links from for example libcrypto.so.0
15488 to libcrypto.so.0.9.7. There is extended info in Configure for
15491 - Make as few rebuilds of the shared libraries as possible.
15493 - Still avoid linking the OpenSSL programs with the shared libraries.
15495 - When installing, install the shared libraries separately from the
15500 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15502 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15503 and not in SSL_clear because the latter is also used by the
15504 accept/connect functions; previously, the settings made by
15505 SSL_set_read_ahead would be lost during the handshake.
15507 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15509 * Correct util/mkdef.pl to be selective about disabled algorithms.
15510 Previously, it would create entries for disabled algorithms no
15515 * Added several new manual pages for SSL_* function.
15519 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15521 * In ssl23_get_client_hello, generate an error message when faced
15522 with an initial SSL 3.0/TLS record that is too small to contain the
15523 first two bytes of the ClientHello message, i.e. client_version.
15524 (Note that this is a pathologic case that probably has never happened
15525 in real life.) The previous approach was to use the version number
15526 from the record header as a substitute; but our protocol choice
15527 should not depend on that one because it is not authenticated
15528 by the Finished messages.
15532 * More robust randomness gathering functions for Windows.
15534 *Jeffrey Altman <jaltman@columbia.edu>*
15536 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15537 not set then we don't setup the error code for issuer check errors
15538 to avoid possibly overwriting other errors which the callback does
15539 handle. If an application does set the flag then we assume it knows
15540 what it is doing and can handle the new informational codes
15545 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15546 a general "ANY" type, as such it should be able to decode anything
15547 including tagged types. However it didn't check the class so it would
15548 wrongly interpret tagged types in the same way as their universal
15549 counterpart and unknown types were just rejected. Changed so that the
15550 tagged and unknown types are handled in the same way as a SEQUENCE:
15551 that is the encoding is stored intact. There is also a new type
15552 "V_ASN1_OTHER" which is used when the class is not universal, in this
15553 case we have no idea what the actual type is so we just lump them all
15558 * On VMS, stdout may very well lead to a file that is written to
15559 in a record-oriented fashion. That means that every write() will
15560 write a separate record, which will be read separately by the
15561 programs trying to read from it. This can be very confusing.
15563 The solution is to put a BIO filter in the way that will buffer
15564 text until a linefeed is reached, and then write everything a
15565 line at a time, so every record written will be an actual line,
15566 not chunks of lines and not (usually doesn't happen, but I've
15567 seen it once) several lines in one record. BIO_f_linebuffer() is
15570 Currently, it's a VMS-only method, because that's where it has
15571 been tested well enough.
15575 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15576 it can return incorrect results.
15577 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15578 but it was in 0.9.6-beta[12].)
15582 * Disable the check for content being present when verifying detached
15583 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15584 include zero length content when signing messages.
15588 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15589 BIO_ctrl (for BIO pairs).
15593 * Add DSO method for VMS.
15597 * Bug fix: Montgomery multiplication could produce results with the
15602 * Add RPM specification openssl.spec and modify it to build three
15603 packages. The default package contains applications, application
15604 documentation and run-time libraries. The devel package contains
15605 include files, static libraries and function documentation. The
15606 doc package contains the contents of the doc directory. The original
15607 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15611 * Add a large number of documentation files for many SSL routines.
15613 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15615 * Add a configuration entry for Sony News 4.
15617 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15619 * Don't set the two most significant bits to one when generating a
15620 random number < q in the DSA library.
15624 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15625 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15626 the underlying transport is blocking) if a handshake took place.
15627 (The default behaviour is needed by applications such as s_client
15628 and s_server that use select() to determine when to use SSL_read;
15629 but for applications that know in advance when to expect data, it
15630 just makes things more complicated.)
15634 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15639 * Add a few more EBCDIC conditionals that make `req` and `x509`
15640 work better on such systems.
15642 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15644 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15645 Update PKCS12_parse() so it copies the friendlyName and the
15646 keyid to the certificates aux info.
15650 * Fix bug in PKCS7_verify() which caused an infinite loop
15651 if there was more than one signature.
15653 *Sven Uszpelkat <su@celocom.de>*
15655 * Major change in util/mkdef.pl to include extra information
15656 about each symbol, as well as presenting variables as well
15657 as functions. This change means that there's n more need
15658 to rebuild the .num files when some algorithms are excluded.
15662 * Allow the verify time to be set by an application,
15663 rather than always using the current time.
15667 * Phase 2 verify code reorganisation. The certificate
15668 verify code now looks up an issuer certificate by a
15669 number of criteria: subject name, authority key id
15670 and key usage. It also verifies self signed certificates
15671 by the same criteria. The main comparison function is
15672 X509_check_issued() which performs these checks.
15674 Lot of changes were necessary in order to support this
15675 without completely rewriting the lookup code.
15677 Authority and subject key identifier are now cached.
15679 The LHASH 'certs' is X509_STORE has now been replaced
15680 by a STACK_OF(X509_OBJECT). This is mainly because an
15681 LHASH can't store or retrieve multiple objects with
15682 the same hash value.
15684 As a result various functions (which were all internal
15685 use only) have changed to handle the new X509_STORE
15686 structure. This will break anything that messed round
15687 with X509_STORE internally.
15689 The functions X509_STORE_add_cert() now checks for an
15690 exact match, rather than just subject name.
15692 The X509_STORE API doesn't directly support the retrieval
15693 of multiple certificates matching a given criteria, however
15694 this can be worked round by performing a lookup first
15695 (which will fill the cache with candidate certificates)
15696 and then examining the cache for matches. This is probably
15697 the best we can do without throwing out X509_LOOKUP
15698 entirely (maybe later...).
15700 The X509_VERIFY_CTX structure has been enhanced considerably.
15702 All certificate lookup operations now go via a get_issuer()
15703 callback. Although this currently uses an X509_STORE it
15704 can be replaced by custom lookups. This is a simple way
15705 to bypass the X509_STORE hackery necessary to make this
15706 work and makes it possible to use more efficient techniques
15707 in future. A very simple version which uses a simple
15708 STACK for its trusted certificate store is also provided
15709 using X509_STORE_CTX_trusted_stack().
15711 The verify_cb() and verify() callbacks now have equivalents
15712 in the X509_STORE_CTX structure.
15714 X509_STORE_CTX also has a 'flags' field which can be used
15715 to customise the verify behaviour.
15719 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15720 excludes S/MIME capabilities.
15724 * When a certificate request is read in keep a copy of the
15725 original encoding of the signed data and use it when outputting
15726 again. Signatures then use the original encoding rather than
15727 a decoded, encoded version which may cause problems if the
15728 request is improperly encoded.
15732 * For consistency with other BIO_puts implementations, call
15733 buffer_write(b, ...) directly in buffer_puts instead of calling
15736 In BIO_puts, increment b->num_write as in BIO_write.
15738 *Peter.Sylvester@EdelWeb.fr*
15740 * Fix BN_mul_word for the case where the word is 0. (We have to use
15741 BN_zero, we may not return a BIGNUM with an array consisting of
15742 words set to zero.)
15746 * Avoid calling abort() from within the library when problems are
15747 detected, except if preprocessor symbols have been defined
15748 (such as REF_CHECK, BN_DEBUG etc.).
15752 * New openssl application 'rsautl'. This utility can be
15753 used for low-level RSA operations. DER public key
15754 BIO/fp routines also added.
15758 * New Configure entry and patches for compiling on QNX 4.
15760 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15762 * A demo state-machine implementation was sponsored by
15763 Nuron (<http://www.nuron.com/>) and is now available in
15764 demos/state_machine.
15768 * New options added to the 'dgst' utility for signature
15769 generation and verification.
15773 * Unrecognized PKCS#7 content types are now handled via a
15774 catch all ASN1_TYPE structure. This allows unsupported
15775 types to be stored as a "blob" and an application can
15776 encode and decode it manually.
15780 * Fix various signed/unsigned issues to make a_strex.c
15781 compile under VC++.
15783 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15785 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15786 length if passed a buffer. ASN1_INTEGER_to_BN failed
15787 if passed a NULL BN and its argument was negative.
15789 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15791 * Modification to PKCS#7 encoding routines to output definite
15792 length encoding. Since currently the whole structures are in
15793 memory there's not real point in using indefinite length
15794 constructed encoding. However if OpenSSL is compiled with
15795 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15799 * Added BIO_vprintf() and BIO_vsnprintf().
15803 * Added more prefixes to parse for in the strings written
15804 through a logging bio, to cover all the levels that are available
15805 through syslog. The prefixes are now:
15807 PANIC, EMERG, EMR => LOG_EMERG
15808 ALERT, ALR => LOG_ALERT
15809 CRIT, CRI => LOG_CRIT
15810 ERROR, ERR => LOG_ERR
15811 WARNING, WARN, WAR => LOG_WARNING
15812 NOTICE, NOTE, NOT => LOG_NOTICE
15813 INFO, INF => LOG_INFO
15814 DEBUG, DBG => LOG_DEBUG
15816 and as before, if none of those prefixes are present at the
15817 beginning of the string, LOG_ERR is chosen.
15819 On Win32, the `LOG_*` levels are mapped according to this:
15821 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15822 LOG_WARNING => EVENTLOG_WARNING_TYPE
15823 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15827 * Made it possible to reconfigure with just the configuration
15828 argument "reconf" or "reconfigure". The command line arguments
15829 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15830 and are retrieved from there when reconfiguring.
15836 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15838 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15842 * The obj_dat.pl script was messing up the sorting of object
15843 names. The reason was that it compared the quoted version
15844 of strings as a result "OCSP" > "OCSP Signing" because
15845 " > SPACE. Changed script to store unquoted versions of
15846 names and add quotes on output. It was also omitting some
15847 names from the lookup table if they were given a default
15848 value (that is if SN is missing it is given the same
15849 value as LN and vice versa), these are now added on the
15850 grounds that if an object has a name we should be able to
15851 look it up. Finally added warning output when duplicate
15852 short or long names are found.
15856 * Changes needed for Tandem NSK.
15858 *Scott Uroff <scott@xypro.com>*
15860 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15861 RSA_padding_check_SSLv23(), special padding was never detected
15862 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15863 version rollback attacks was not effective.
15865 In s23_clnt.c, don't use special rollback-attack detection padding
15866 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15867 client; similarly, in s23_srvr.c, don't do the rollback check if
15868 SSL 2.0 is the only protocol enabled in the server.
15872 * Make it possible to get hexdumps of unprintable data with 'openssl
15873 asn1parse'. By implication, the functions ASN1_parse_dump() and
15874 BIO_dump_indent() are added.
15878 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15879 these print out strings and name structures based on various
15880 flags including RFC2253 support and proper handling of
15881 multibyte characters. Added options to the 'x509' utility
15882 to allow the various flags to be set.
15886 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15887 Also change the functions X509_cmp_current_time() and
15888 X509_gmtime_adj() work with an ASN1_TIME structure,
15889 this will enable certificates using GeneralizedTime in validity
15890 dates to be checked.
15894 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15895 negative public key encodings) on by default,
15896 NO_NEG_PUBKEY_BUG can be set to disable it.
15900 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15901 content octets. An i2c_ASN1_OBJECT is unnecessary because
15902 the encoding can be trivially obtained from the structure.
15906 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15907 not read locks (`CRYPTO_r_[un]lock`).
15911 * A first attempt at creating official support for shared
15912 libraries through configuration. I've kept it so the
15913 default is static libraries only, and the OpenSSL programs
15914 are always statically linked for now, but there are
15915 preparations for dynamic linking in place.
15916 This has been tested on Linux and Tru64.
15920 * Randomness polling function for Win9x, as described in:
15921 Peter Gutmann, Software Generation of Practically Strong
15926 * Fix so PRNG is seeded in req if using an already existing
15931 * New options to smime application. -inform and -outform
15932 allow alternative formats for the S/MIME message including
15933 PEM and DER. The -content option allows the content to be
15934 specified separately. This should allow things like Netscape
15935 form signing output easier to verify.
15939 * Fix the ASN1 encoding of tags using the 'long form'.
15943 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15944 STRING types. These convert content octets to and from the
15945 underlying type. The actual tag and length octets are
15946 already assumed to have been read in and checked. These
15947 are needed because all other string types have virtually
15948 identical handling apart from the tag. By having versions
15949 of the ASN1 functions that just operate on content octets
15950 IMPLICIT tagging can be handled properly. It also allows
15951 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15952 and ASN1_INTEGER are identical apart from the tag.
15956 * Change the handling of OID objects as follows:
15958 - New object identifiers are inserted in objects.txt, following
15959 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15960 - objects.pl is used to process obj_mac.num and create a new
15962 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15965 This is currently kind of a hack, and the perl code in objects.pl
15966 isn't very elegant, but it works as I intended. The simplest way
15967 to check that it worked correctly is to look in obj_dat.h and
15968 check the array nid_objs and make sure the objects haven't moved
15969 around (this is important!). Additions are OK, as well as
15970 consistent name changes.
15974 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15978 * Addition of the command line parameter '-rand file' to 'openssl req'.
15979 The given file adds to whatever has already been seeded into the
15980 random pool through the RANDFILE configuration file option or
15981 environment variable, or the default random state file.
15985 * mkstack.pl now sorts each macro group into lexical order.
15986 Previously the output order depended on the order the files
15987 appeared in the directory, resulting in needless rewriting
15992 * Patches to make OpenSSL compile under Win32 again. Mostly
15993 work arounds for the VC++ problem that it treats func() as
15994 func(void). Also stripped out the parts of mkdef.pl that
15995 added extra typesafe functions: these no longer exist.
15999 * Reorganisation of the stack code. The macros are now all
16000 collected in safestack.h . Each macro is defined in terms of
16001 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16002 DEBUG_SAFESTACK is now handled in terms of function casts,
16003 this has the advantage of retaining type safety without the
16004 use of additional functions. If DEBUG_SAFESTACK is not defined
16005 then the non typesafe macros are used instead. Also modified the
16006 mkstack.pl script to handle the new form. Needs testing to see
16007 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16008 the default if no major problems. Similar behaviour for ASN1_SET_OF
16009 and PKCS12_STACK_OF.
16013 * When some versions of IIS use the 'NET' form of private key the
16014 key derivation algorithm is different. Normally MD5(password) is
16015 used as a 128 bit RC4 key. In the modified case
16016 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16017 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16018 as the old Netscape_RSA functions except they have an additional
16019 'sgckey' parameter which uses the modified algorithm. Also added
16020 an -sgckey command line option to the rsa utility. Thanks to
16021 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16022 algorithm to openssl-dev.
16026 * The evp_local.h macros were using 'c.##kname' which resulted in
16027 invalid expansion on some systems (SCO 5.0.5 for example).
16028 Corrected to 'c.kname'.
16030 *Phillip Porch <root@theporch.com>*
16032 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16033 a STACK of email addresses from a certificate or request, these look
16034 in the subject name and the subject alternative name extensions and
16035 omit any duplicate addresses.
16039 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16040 This makes DSA verification about 2 % faster.
16044 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16045 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16046 plus overhead for 1024 bit moduli).
16047 This makes exponentiations about 0.5 % faster for 1024 bit
16048 exponents (as measured by "openssl speed rsa2048").
16052 * Rename memory handling macros to avoid conflicts with other
16054 Malloc => OPENSSL_malloc
16055 Malloc_locked => OPENSSL_malloc_locked
16056 Realloc => OPENSSL_realloc
16057 Free => OPENSSL_free
16061 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16062 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16066 * CygWin32 support.
16068 *John Jarvie <jjarvie@newsguy.com>*
16070 * The type-safe stack code has been rejigged. It is now only compiled
16071 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16072 by default all type-specific stack functions are "#define"d back to
16073 standard stack functions. This results in more streamlined output
16074 but retains the type-safety checking possibilities of the original
16079 * The STACK code has been cleaned up, and certain type declarations
16080 that didn't make a lot of sense have been brought in line. This has
16081 also involved a cleanup of sorts in safestack.h to more correctly
16082 map type-safe stack functions onto their plain stack counterparts.
16083 This work has also resulted in a variety of "const"ifications of
16084 lots of the code, especially `_cmp` operations which should normally
16085 be prototyped with "const" parameters anyway.
16089 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16090 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16091 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16092 where all of 'md' is used each time the PRNG is used, but 'state'
16093 is used only indexed by a cyclic counter. As entropy may not be
16094 well distributed from the beginning, 'md' is important as a
16095 chaining variable. However, the output function chains only half
16096 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16097 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16098 in all of 'state' being rewritten, with the new values depending
16099 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16103 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16104 the handshake is continued after ssl_verify_cert_chain();
16105 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16106 can lead to 'unexplainable' connection aborts later.
16108 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16110 * Major EVP API cipher revision.
16111 Add hooks for extra EVP features. This allows various cipher
16112 parameters to be set in the EVP interface. Support added for variable
16113 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16114 setting of RC2 and RC5 parameters.
16116 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16119 Remove lots of duplicated code from the EVP library. For example *every*
16120 cipher init() function handles the 'iv' in the same way according to the
16121 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16122 for CFB and OFB modes they zero ctx->num.
16124 New functionality allows removal of S/MIME code RC2 hack.
16126 Most of the routines have the same form and so can be declared in terms
16129 By shifting this to the top level EVP_CipherInit() it can be removed from
16130 all individual ciphers. If the cipher wants to handle IVs or keys
16131 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16134 Change lots of functions like EVP_EncryptUpdate() to now return a
16135 value: although software versions of the algorithms cannot fail
16136 any installed hardware versions can.
16140 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16141 this option is set, tolerate broken clients that send the negotiated
16142 protocol version number instead of the requested protocol version
16147 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16148 i.e. non-zero for export ciphersuites, zero otherwise.
16149 Previous versions had this flag inverted, inconsistent with
16150 rsa_tmp_cb (..._TMP_RSA_CB).
16152 *Bodo Moeller; problem reported by Amit Chopra*
16154 * Add missing DSA library text string. Work around for some IIS
16155 key files with invalid SEQUENCE encoding.
16159 * Add a document (doc/standards.txt) that list all kinds of standards
16160 and so on that are implemented in OpenSSL.
16164 * Enhance c_rehash script. Old version would mishandle certificates
16165 with the same subject name hash and wouldn't handle CRLs at all.
16166 Added -fingerprint option to crl utility, to support new c_rehash
16171 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16175 * Fix for SSL server purpose checking. Server checking was
16176 rejecting certificates which had extended key usage present
16177 but no ssl client purpose.
16179 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16181 * Make PKCS#12 code work with no password. The PKCS#12 spec
16182 is a little unclear about how a blank password is handled.
16183 Since the password in encoded as a BMPString with terminating
16184 double NULL a zero length password would end up as just the
16185 double NULL. However no password at all is different and is
16186 handled differently in the PKCS#12 key generation code. NS
16187 treats a blank password as zero length. MSIE treats it as no
16188 password on export: but it will try both on import. We now do
16189 the same: PKCS12_parse() tries zero length and no password if
16190 the password is set to "" or NULL (NULL is now a valid password:
16191 it wasn't before) as does the pkcs12 application.
16195 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16196 perror when PEM_read_bio_X509_REQ fails, the error message must
16197 be obtained from the error queue.
16201 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16202 it in ERR_remove_state if appropriate, and change ERR_get_state
16203 accordingly to avoid race conditions (this is necessary because
16204 thread_hash is no longer constant once set).
16208 * Bugfix for linux-elf makefile.one.
16212 * RSA_get_default_method() will now cause a default
16213 RSA_METHOD to be chosen if one doesn't exist already.
16214 Previously this was only set during a call to RSA_new()
16215 or RSA_new_method(NULL) meaning it was possible for
16216 RSA_get_default_method() to return NULL.
16220 * Added native name translation to the existing DSO code
16221 that will convert (if the flag to do so is set) filenames
16222 that are sufficiently small and have no path information
16223 into a canonical native form. Eg. "blah" converted to
16224 "libblah.so" or "blah.dll" etc.
16228 * New function ERR_error_string_n(e, buf, len) which is like
16229 ERR_error_string(e, buf), but writes at most 'len' bytes
16230 including the 0 terminator. For ERR_error_string_n, 'buf'
16233 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16235 * CONF library reworked to become more general. A new CONF
16236 configuration file reader "class" is implemented as well as a
16237 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16238 old `CONF_*` functions are still there, but are reimplemented to
16239 work in terms of the new functions. Also, a set of functions
16240 to handle the internal storage of the configuration data is
16241 provided to make it easier to write new configuration file
16242 reader "classes" (I can definitely see something reading a
16243 configuration file in XML format, for example), called `_CONF_*`,
16244 or "the configuration storage API"...
16246 The new configuration file reading functions are:
16248 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16249 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16251 NCONF_default, NCONF_WIN32
16253 NCONF_dump_fp, NCONF_dump_bio
16255 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16256 NCONF_new creates a new CONF object. This works in the same way
16257 as other interfaces in OpenSSL, like the BIO interface.
16258 `NCONF_dump_*` dump the internal storage of the configuration file,
16259 which is useful for debugging. All other functions take the same
16260 arguments as the old `CONF_*` functions with the exception of the
16261 first that must be a `CONF *` instead of a `LHASH *`.
16263 To make it easier to use the new classes with the old `CONF_*` functions,
16264 the function CONF_set_default_method is provided.
16268 * Add '-tls1' option to 'openssl ciphers', which was already
16269 mentioned in the documentation but had not been implemented.
16270 (This option is not yet really useful because even the additional
16271 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16275 * Initial DSO code added into libcrypto for letting OpenSSL (and
16276 OpenSSL-based applications) load shared libraries and bind to
16277 them in a portable way.
16279 *Geoff Thorpe, with contributions from Richard Levitte*
16281 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16283 * Make sure _lrotl and _lrotr are only used with MSVC.
16285 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16286 (the default implementation of RAND_status).
16288 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16289 to '-clrext' (= clear extensions), as intended and documented.
16290 *Bodo Moeller; inconsistency pointed out by Michael Attili
16291 <attili@amaxo.com>*
16293 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16294 was larger than the MD block size.
16296 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16298 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16299 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16300 using the passed key: if the passed key was a private key the result
16301 of X509_print(), for example, would be to print out all the private key
16306 * des_quad_cksum() byte order bug fix.
16307 *Ulf Möller, using the problem description in krb4-0.9.7, where
16308 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16310 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16313 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16315 * For easily testing in shell scripts whether some command
16316 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16317 returns with exit code 0 iff no command of the given name is available.
16318 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16319 the output goes to stdout and nothing is printed to stderr.
16320 Additional arguments are always ignored.
16322 Since for each cipher there is a command of the same name,
16323 the 'no-cipher' compilation switches can be tested this way.
16325 ('openssl no-XXX' is not able to detect pseudo-commands such
16326 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16330 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16334 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16335 is set; it will be thrown away anyway because each handshake creates
16337 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16338 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16339 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16340 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16344 * New s_client option -ign_eof: EOF at stdin is ignored, and
16345 'Q' and 'R' lose their special meanings (quit/renegotiate).
16346 This is part of what -quiet does; unlike -quiet, -ign_eof
16347 does not suppress any output.
16351 * Add compatibility options to the purpose and trust code. The
16352 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16353 accepts a certificate or CA, this was the previous behaviour,
16354 with all the associated security issues.
16356 X509_TRUST_COMPAT is the old trust behaviour: only and
16357 automatically trust self signed roots in certificate store. A
16358 new trust setting X509_TRUST_DEFAULT is used to specify that
16359 a purpose has no associated trust setting and it should instead
16360 use the value in the default purpose.
16364 * Fix the PKCS#8 DSA private key code so it decodes keys again
16365 and fix a memory leak.
16369 * In util/mkerr.pl (which implements 'make errors'), preserve
16370 reason strings from the previous version of the .c file, as
16371 the default to have only downcase letters (and digits) in
16372 automatically generated reasons codes is not always appropriate.
16376 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16377 using strerror. Previously, ERR_reason_error_string() returned
16378 library names as reason strings for SYSerr; but SYSerr is a special
16379 case where small numbers are errno values, not library numbers.
16383 * Add '-dsaparam' option to 'openssl dhparam' application. This
16384 converts DSA parameters into DH parameters. (When creating parameters,
16385 DSA_generate_parameters is used.)
16389 * Include 'length' (recommended exponent length) in C code generated
16390 by 'openssl dhparam -C'.
16394 * The second argument to set_label in perlasm was already being used
16395 so couldn't be used as a "file scope" flag. Moved to third argument
16400 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16401 instead of RAND_bytes for encryption IVs and salts.
16405 * Include RAND_status() into RAND_METHOD instead of implementing
16406 it only for md_rand.c Otherwise replacing the PRNG by calling
16407 RAND_set_rand_method would be impossible.
16411 * Don't let DSA_generate_key() enter an infinite loop if the random
16412 number generation fails.
16416 * New 'rand' application for creating pseudo-random output.
16420 * Added configuration support for Linux/IA64
16422 *Rolf Haberrecker <rolf@suse.de>*
16424 * Assembler module support for Mingw32.
16428 * Shared library support for HPUX (in shlib/).
16430 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16432 * Shared library support for Solaris gcc.
16434 *Lutz Behnke <behnke@trustcenter.de>*
16436 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16438 * PKCS7_encrypt() was adding text MIME headers twice because they
16439 were added manually and by SMIME_crlf_copy().
16443 * In bntest.c don't call BN_rand with zero bits argument.
16445 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16447 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16448 case was implemented. This caused BN_div_recp() to fail occasionally.
16452 * Add an optional second argument to the set_label() in the perl
16453 assembly language builder. If this argument exists and is set
16454 to 1 it signals that the assembler should use a symbol whose
16455 scope is the entire file, not just the current function. This
16456 is needed with MASM which uses the format label:: for this scope.
16458 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16460 * Change the ASN1 types so they are typedefs by default. Before
16461 almost all types were #define'd to ASN1_STRING which was causing
16462 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16467 * Change names of new functions to the new get1/get0 naming
16468 convention: After 'get1', the caller owns a reference count
16469 and has to call `..._free`; 'get0' returns a pointer to some
16470 data structure without incrementing reference counters.
16471 (Some of the existing 'get' functions increment a reference
16472 counter, some don't.)
16473 Similarly, 'set1' and 'add1' functions increase reference
16474 counters or duplicate objects.
16478 * Allow for the possibility of temp RSA key generation failure:
16479 the code used to assume it always worked and crashed on failure.
16483 * Fix potential buffer overrun problem in BIO_printf().
16484 *Ulf Möller, using public domain code by Patrick Powell; problem
16485 pointed out by David Sacerdote <das33@cornell.edu>*
16487 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16488 RAND_egd() and RAND_status(). In the command line application,
16489 the EGD socket can be specified like a seed file using RANDFILE
16494 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16495 Some CAs (e.g. Verisign) distribute certificates in this form.
16499 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16500 list to exclude them. This means that no special compilation option
16501 is needed to use anonymous DH: it just needs to be included in the
16506 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16507 EVP_MD_type. The old functionality is available in a new macro called
16508 EVP_MD_md(). Change code that uses it and update docs.
16512 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16513 where the `void *` argument is replaced by a function pointer argument.
16514 Previously `void *` was abused to point to functions, which works on
16515 many platforms, but is not correct. As these functions are usually
16516 called by macros defined in OpenSSL header files, most source code
16517 should work without changes.
16521 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16522 sections with information on -D... compiler switches used for
16523 compiling the library so that applications can see them. To enable
16524 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16525 must be defined. E.g.,
16526 #define OPENSSL_ALGORITHM_DEFINES
16527 #include <openssl/opensslconf.h>
16528 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16530 *Richard Levitte, Ulf and Bodo Möller*
16532 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16537 * Change the 'other' type in certificate aux info to a STACK_OF
16538 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16539 the required ASN1 format: arbitrary types determined by an OID.
16543 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16544 argument to 'req'. This is not because the function is newer or
16545 better than others it just uses the work 'NEW' in the certificate
16546 request header lines. Some software needs this.
16550 * Reorganise password command line arguments: now passwords can be
16551 obtained from various sources. Delete the PEM_cb function and make
16552 it the default behaviour: i.e. if the callback is NULL and the
16553 usrdata argument is not NULL interpret it as a null terminated pass
16554 phrase. If usrdata and the callback are NULL then the pass phrase
16555 is prompted for as usual.
16559 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16560 the support is automatically enabled. The resulting binaries will
16561 autodetect the card and use it if present.
16563 *Ben Laurie and Compaq Inc.*
16565 * Work around for Netscape hang bug. This sends certificate request
16566 and server done in one record. Since this is perfectly legal in the
16567 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16568 the bugs/SSLv3 entry for more info.
16572 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16576 * Add -rand argument to smime and pkcs12 applications and read/write
16581 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16585 * Add command line password options to the remaining applications.
16589 * Bug fix for BN_div_recp() for numerators with an even number of
16594 * More tests in bntest.c, and changed test_bn output.
16598 * ./config recognizes MacOS X now.
16602 * Bug fix for BN_div() when the first words of num and divisor are
16603 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16607 * Add support for various broken PKCS#8 formats, and command line
16608 options to produce them.
16612 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16613 get temporary BIGNUMs from a BN_CTX.
16617 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16622 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16623 include a #define from the old name to the new. The original intent
16624 was that statically linked binaries could for example just call
16625 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16626 link with digests. This never worked because SSLeay_add_all_digests()
16627 and SSLeay_add_all_ciphers() were in the same source file so calling
16628 one would link with the other. They are now in separate source files.
16632 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16636 * Use a less unusual form of the Miller-Rabin primality test (it used
16637 a binary algorithm for exponentiation integrated into the Miller-Rabin
16638 loop, our standard modexp algorithms are faster).
16642 * Support for the EBCDIC character set completed.
16644 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16646 * Source code cleanups: use const where appropriate, eliminate casts,
16647 use `void *` instead of `char *` in lhash.
16651 * Bugfix: ssl3_send_server_key_exchange was not restartable
16652 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16653 this the server could overwrite ephemeral keys that the client
16658 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16659 using 50 iterations of the Rabin-Miller test.
16661 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16662 iterations of the Rabin-Miller test as required by the appendix
16663 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16664 As BN_is_prime_fasttest includes trial division, DSA parameter
16665 generation becomes much faster.
16667 This implies a change for the callback functions in DSA_is_prime
16668 and DSA_generate_parameters: The callback function is called once
16669 for each positive witness in the Rabin-Miller test, not just
16670 occasionally in the inner loop; and the parameters to the
16671 callback function now provide an iteration count for the outer
16672 loop rather than for the current invocation of the inner loop.
16673 DSA_generate_parameters additionally can call the callback
16674 function with an 'iteration count' of -1, meaning that a
16675 candidate has passed the trial division test (when q is generated
16676 from an application-provided seed, trial division is skipped).
16680 * New function BN_is_prime_fasttest that optionally does trial
16681 division before starting the Rabin-Miller test and has
16682 an additional BN_CTX * argument (whereas BN_is_prime always
16683 has to allocate at least one BN_CTX).
16684 'callback(1, -1, cb_arg)' is called when a number has passed the
16685 trial division stage.
16689 * Fix for bug in CRL encoding. The validity dates weren't being handled
16694 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16698 * New function BN_pseudo_rand().
16702 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16703 bignum version of BN_from_montgomery() with the working code from
16704 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16709 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16710 made it impossible to use the same SSL_SESSION data structure in
16711 SSL2 clients in multiple threads.
16715 * The return value of RAND_load_file() no longer counts bytes obtained
16716 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16717 to seed the PRNG (previously an explicit byte count was required).
16719 *Ulf Möller, Bodo Möller*
16721 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16722 used `char *` instead of `void *` and had casts all over the place.
16726 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16730 * Retain source code compatibility for BN_prime_checks macro:
16731 BN_is_prime(..., BN_prime_checks, ...) now uses
16732 BN_prime_checks_for_size to determine the appropriate number of
16733 Rabin-Miller iterations.
16737 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16738 DH_CHECK_P_NOT_SAFE_PRIME.
16739 (Check if this is true? OpenPGP calls them "strong".)
16743 * Merge the functionality of "dh" and "gendh" programs into a new program
16744 "dhparam". The old programs are retained for now but will handle DH keys
16745 (instead of parameters) in future.
16749 * Make the ciphers, s_server and s_client programs check the return values
16750 when a new cipher list is set.
16754 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16755 ciphers. Before when the 56bit ciphers were enabled the sorting was
16758 The syntax for the cipher sorting has been extended to support sorting by
16759 cipher-strength (using the strength_bits hard coded in the tables).
16760 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16762 Fix a bug in the cipher-command parser: when supplying a cipher command
16763 string with an "undefined" symbol (neither command nor alphanumeric
16764 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16765 an error is flagged.
16767 Due to the strength-sorting extension, the code of the
16768 ssl_create_cipher_list() function was completely rearranged. I hope that
16769 the readability was also increased :-)
16771 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16773 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16774 for the first serial number and places 2 in the serial number file. This
16775 avoids problems when the root CA is created with serial number zero and
16776 the first user certificate has the same issuer name and serial number
16781 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16782 the new code. Add documentation for this stuff.
16786 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16787 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16788 structures and behave in an analogous way to the X509v3 functions:
16789 they shouldn't be called directly but wrapper functions should be used
16792 So we also now have some wrapper functions that call the X509at functions
16793 when passed certificate requests. (TO DO: similar things can be done with
16794 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16795 things. Some of these need some d2i or i2d and print functionality
16796 because they handle more complex structures.)
16800 * Add missing #ifndefs that caused missing symbols when building libssl
16801 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16802 NO_RSA in `ssl/s2*.c`.
16804 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16806 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16807 has a return value which indicates the quality of the random data
16808 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16809 error queue. New function RAND_pseudo_bytes() generates output that is
16810 guaranteed to be unique but not unpredictable. RAND_add is like
16811 RAND_seed, but takes an extra argument for an entropy estimate
16812 (RAND_seed always assumes full entropy).
16816 * Do more iterations of Rabin-Miller probable prime test (specifically,
16817 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16818 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16819 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16820 false-positive rate of at most 2^-80 for random input.
16824 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16828 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16829 in the 0.9.5 release), this returns the chain
16830 from an X509_CTX structure with a dup of the stack and all
16831 the X509 reference counts upped: so the stack will exist
16832 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16835 Also make SSL_SESSION_print() print out the verify return
16840 * Add manpage for the pkcs12 command. Also change the default
16841 behaviour so MAC iteration counts are used unless the new
16842 -nomaciter option is used. This improves file security and
16843 only older versions of MSIE (4.0 for example) need it.
16847 * Honor the no-xxx Configure options when creating .DEF files.
16851 * Add PKCS#10 attributes to field table: challengePassword,
16852 unstructuredName and unstructuredAddress. These are taken from
16853 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16854 international characters are used.
16856 More changes to X509_ATTRIBUTE code: allow the setting of types
16857 based on strings. Remove the 'loc' parameter when adding
16858 attributes because these will be a SET OF encoding which is sorted
16863 * Initial changes to the 'req' utility to allow request generation
16864 automation. This will allow an application to just generate a template
16865 file containing all the field values and have req construct the
16868 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16869 used all over the place including certificate requests and PKCS#7
16870 structures. They are currently handled manually where necessary with
16871 some primitive wrappers for PKCS#7. The new functions behave in a
16872 manner analogous to the X509 extension functions: they allow
16873 attributes to be looked up by NID and added.
16875 Later something similar to the X509V3 code would be desirable to
16876 automatically handle the encoding, decoding and printing of the
16877 more complex types. The string types like challengePassword can
16878 be handled by the string table functions.
16880 Also modified the multi byte string table handling. Now there is
16881 a 'global mask' which masks out certain types. The table itself
16882 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16883 is useful when for example there is only one permissible type
16884 (as in countryName) and using the mask might result in no valid
16889 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16890 SSL_get_peer_finished to allow applications to obtain the latest
16891 Finished messages sent to the peer or expected from the peer,
16892 respectively. (SSL_get_peer_finished is usually the Finished message
16893 actually received from the peer, otherwise the protocol will be aborted.)
16895 As the Finished message are message digests of the complete handshake
16896 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16897 be used for external authentication procedures when the authentication
16898 provided by SSL/TLS is not desired or is not enough.
16902 * Enhanced support for Alpha Linux is added. Now ./config checks if
16903 the host supports BWX extension and if Compaq C is present on the
16904 $PATH. Just exploiting of the BWX extension results in 20-30%
16905 performance kick for some algorithms, e.g. DES and RC4 to mention
16906 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16911 * Add support for MS "fast SGC". This is arguably a violation of the
16912 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16913 weak crypto and after checking the certificate is SGC a second one
16914 with strong crypto. MS SGC stops the first handshake after receiving
16915 the server certificate message and sends a second client hello. Since
16916 a server will typically do all the time consuming operations before
16917 expecting any further messages from the client (server key exchange
16918 is the most expensive) there is little difference between the two.
16920 To get OpenSSL to support MS SGC we have to permit a second client
16921 hello message after we have sent server done. In addition we have to
16922 reset the MAC if we do get this second client hello.
16926 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16927 if a DER encoded private key is RSA or DSA traditional format. Changed
16928 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16929 format DER encoded private key. Newer code should use PKCS#8 format which
16930 has the key type encoded in the ASN1 structure. Added DER private key
16931 support to pkcs8 application.
16935 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16936 ciphersuites has been selected (as required by the SSL 3/TLS 1
16937 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16938 is set, we interpret this as a request to violate the specification
16939 (the worst that can happen is a handshake failure, and 'correct'
16940 behaviour would result in a handshake failure anyway).
16944 * In SSL_CTX_add_session, take into account that there might be multiple
16945 SSL_SESSION structures with the same session ID (e.g. when two threads
16946 concurrently obtain them from an external cache).
16947 The internal cache can handle only one SSL_SESSION with a given ID,
16948 so if there's a conflict, we now throw out the old one to achieve
16953 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16954 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16955 some routines that use cipher OIDs: some ciphers do not have OIDs
16956 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16961 * Simplify the trust setting structure and code. Now we just have
16962 two sequences of OIDs for trusted and rejected settings. These will
16963 typically have values the same as the extended key usage extension
16964 and any application specific purposes.
16966 The trust checking code now has a default behaviour: it will just
16967 check for an object with the same NID as the passed id. Functions can
16968 be provided to override either the default behaviour or the behaviour
16969 for a given id. SSL client, server and email already have functions
16970 in place for compatibility: they check the NID and also return "trusted"
16971 if the certificate is self signed.
16975 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16976 traditional format into an EVP_PKEY structure.
16980 * Add a password callback function PEM_cb() which either prompts for
16981 a password if usr_data is NULL or otherwise assumes it is a null
16982 terminated password. Allow passwords to be passed on command line
16983 environment or config files in a few more utilities.
16987 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16988 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16989 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16990 Update documentation.
16994 * Support for ASN1 "NULL" type. This could be handled before by using
16995 ASN1_TYPE but there wasn't any function that would try to read a NULL
16996 and produce an error if it couldn't. For compatibility we also have
16997 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16998 don't allocate anything because they don't need to.
17002 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17005 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17007 * Rebuild of the memory allocation routines used by OpenSSL code and
17008 possibly others as well. The purpose is to make an interface that
17009 provide hooks so anyone can build a separate set of allocation and
17010 deallocation routines to be used by OpenSSL, for example memory
17011 pool implementations, or something else, which was previously hard
17012 since Malloc(), Realloc() and Free() were defined as macros having
17013 the values malloc, realloc and free, respectively (except for Win32
17014 compilations). The same is provided for memory debugging code.
17015 OpenSSL already comes with functionality to find memory leaks, but
17016 this gives people a chance to debug other memory problems.
17018 With these changes, a new set of functions and macros have appeared:
17020 CRYPTO_set_mem_debug_functions() [F]
17021 CRYPTO_get_mem_debug_functions() [F]
17022 CRYPTO_dbg_set_options() [F]
17023 CRYPTO_dbg_get_options() [F]
17024 CRYPTO_malloc_debug_init() [M]
17026 The memory debug functions are NULL by default, unless the library
17027 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17028 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17029 gives the standard debugging functions that come with OpenSSL) or
17030 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17031 provided by the library user) must be used. When the standard
17032 debugging functions are used, CRYPTO_dbg_set_options can be used to
17033 request additional information:
17034 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17035 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17037 Also, things like CRYPTO_set_mem_functions will always give the
17038 expected result (the new set of functions is used for allocation
17039 and deallocation) at all times, regardless of platform and compiler
17042 To finish it up, some functions that were never use in any other
17043 way than through macros have a new API and new semantic:
17045 CRYPTO_dbg_malloc()
17046 CRYPTO_dbg_realloc()
17049 All macros of value have retained their old syntax.
17051 *Richard Levitte and Bodo Moeller*
17053 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17054 ordering of SMIMECapabilities wasn't in "strength order" and there
17055 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17060 * Some ASN1 types with illegal zero length encoding (INTEGER,
17061 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17063 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17065 * Merge in my S/MIME library for OpenSSL. This provides a simple
17066 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17067 functionality to handle multipart/signed properly) and a utility
17068 called 'smime' to call all this stuff. This is based on code I
17069 originally wrote for Celo who have kindly allowed it to be
17070 included in OpenSSL.
17074 * Add variants des_set_key_checked and des_set_key_unchecked of
17075 des_set_key (aka des_key_sched). Global variable des_check_key
17076 decides which of these is called by des_set_key; this way
17077 des_check_key behaves as it always did, but applications and
17078 the library itself, which was buggy for des_check_key == 1,
17079 have a cleaner way to pick the version they need.
17083 * New function PKCS12_newpass() which changes the password of a
17088 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17089 dynamic mix. In both cases the ids can be used as an index into the
17090 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17091 functions so they accept a list of the field values and the
17092 application doesn't need to directly manipulate the X509_TRUST
17097 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17102 * Modify the way the V3 extension code looks up extensions. This now
17103 works in a similar way to the object code: we have some "standard"
17104 extensions in a static table which is searched with OBJ_bsearch()
17105 and the application can add dynamic ones if needed. The file
17106 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17107 updated whenever a new extension is added to the core code and kept
17108 in ext_nid order. There is a simple program 'tabtest.c' which checks
17109 this. New extensions are not added too often so this file can readily
17110 be maintained manually.
17112 There are two big advantages in doing things this way. The extensions
17113 can be looked up immediately and no longer need to be "added" using
17114 X509V3_add_standard_extensions(): this function now does nothing.
17115 Side note: I get *lots* of email saying the extension code doesn't
17116 work because people forget to call this function.
17117 Also no dynamic allocation is done unless new extensions are added:
17118 so if we don't add custom extensions there is no need to call
17119 X509V3_EXT_cleanup().
17123 * Modify enc utility's salting as follows: make salting the default. Add a
17124 magic header, so unsalted files fail gracefully instead of just decrypting
17125 to garbage. This is because not salting is a big security hole, so people
17126 should be discouraged from doing it.
17130 * Fixes and enhancements to the 'x509' utility. It allowed a message
17131 digest to be passed on the command line but it only used this
17132 parameter when signing a certificate. Modified so all relevant
17133 operations are affected by the digest parameter including the
17134 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17135 DSA key was used because it didn't fix the digest.
17139 * Initial certificate chain verify code. Currently tests the untrusted
17140 certificates for consistency with the verify purpose (which is set
17141 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17143 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17144 this is because it will reject chains with invalid extensions whereas
17145 every previous version of OpenSSL and SSLeay made no checks at all.
17147 Trust code: checks the root CA for the relevant trust settings. Trust
17148 settings have an initial value consistent with the verify purpose: e.g.
17149 if the verify purpose is for SSL client use it expects the CA to be
17150 trusted for SSL client use. However the default value can be changed to
17151 permit custom trust settings: one example of this would be to only trust
17152 certificates from a specific "secure" set of CAs.
17154 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17155 which should be used for version portability: especially since the
17156 verify structure is likely to change more often now.
17158 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17159 to set them. If not set then assume SSL clients will verify SSL servers
17162 Two new options to the verify program: -untrusted allows a set of
17163 untrusted certificates to be passed in and -purpose which sets the
17164 intended purpose of the certificate. If a purpose is set then the
17165 new chain verify code is used to check extension consistency.
17169 * Support for the authority information access extension.
17173 * Modify RSA and DSA PEM read routines to transparently handle
17174 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17175 public keys in a format compatible with certificate
17176 SubjectPublicKeyInfo structures. Unfortunately there were already
17177 functions called *_PublicKey_* which used various odd formats so
17178 these are retained for compatibility: however the DSA variants were
17179 never in a public release so they have been deleted. Changed dsa/rsa
17180 utilities to handle the new format: note no releases ever handled public
17181 keys so we should be OK.
17183 The primary motivation for this change is to avoid the same fiasco
17184 that dogs private keys: there are several incompatible private key
17185 formats some of which are standard and some OpenSSL specific and
17186 require various evil hacks to allow partial transparent handling and
17187 even then it doesn't work with DER formats. Given the option anything
17188 other than PKCS#8 should be dumped: but the other formats have to
17189 stay in the name of compatibility.
17191 With public keys and the benefit of hindsight one standard format
17192 is used which works with EVP_PKEY, RSA or DSA structures: though
17193 it clearly returns an error if you try to read the wrong kind of key.
17195 Added a -pubkey option to the 'x509' utility to output the public key.
17196 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17197 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17198 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17199 that do the same as the `EVP_PKEY_assign_*()` except they up the
17200 reference count of the added key (they don't "swallow" the
17205 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17206 CRLs would fail if the file contained no certificates or no CRLs:
17207 added a new function to read in both types and return the number
17208 read: this means that if none are read it will be an error. The
17209 DER versions of the certificate and CRL reader would always fail
17210 because it isn't possible to mix certificates and CRLs in DER format
17211 without choking one or the other routine. Changed this to just read
17212 a certificate: this is the best we can do. Also modified the code
17213 in `apps/verify.c` to take notice of return codes: it was previously
17214 attempting to read in certificates from NULL pointers and ignoring
17215 any errors: this is one reason why the cert and CRL reader seemed
17216 to work. It doesn't check return codes from the default certificate
17217 routines: these may well fail if the certificates aren't installed.
17221 * Code to support otherName option in GeneralName.
17225 * First update to verify code. Change the verify utility
17226 so it warns if it is passed a self signed certificate:
17227 for consistency with the normal behaviour. X509_verify
17228 has been modified to it will now verify a self signed
17229 certificate if *exactly* the same certificate appears
17230 in the store: it was previously impossible to trust a
17231 single self signed certificate. This means that:
17232 openssl verify ss.pem
17233 now gives a warning about a self signed certificate but
17234 openssl verify -CAfile ss.pem ss.pem
17239 * For servers, store verify_result in SSL_SESSION data structure
17240 (and add it to external session representation).
17241 This is needed when client certificate verifications fails,
17242 but an application-provided verification callback (set by
17243 SSL_CTX_set_cert_verify_callback) allows accepting the session
17244 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17245 but returns 1): When the session is reused, we have to set
17246 ssl->verify_result to the appropriate error code to avoid
17249 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17251 * Fix a bug in the new PKCS#7 code: it didn't consider the
17252 case in PKCS7_dataInit() where the signed PKCS7 structure
17253 didn't contain any existing data because it was being created.
17255 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17257 * Add a salt to the key derivation routines in enc.c. This
17258 forms the first 8 bytes of the encrypted file. Also add a
17259 -S option to allow a salt to be input on the command line.
17263 * New function X509_cmp(). Oddly enough there wasn't a function
17264 to compare two certificates. We do this by working out the SHA1
17265 hash and comparing that. X509_cmp() will be needed by the trust
17270 * SSL_get1_session() is like SSL_get_session(), but increments
17271 the reference count in the SSL_SESSION returned.
17273 *Geoff Thorpe <geoff@eu.c2.net>*
17275 * Fix for 'req': it was adding a null to request attributes.
17276 Also change the X509_LOOKUP and X509_INFO code to handle
17277 certificate auxiliary information.
17281 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17286 * Add the possibility to add extra information to the memory leak
17287 detecting output, to form tracebacks, showing from where each
17288 allocation was originated: CRYPTO_push_info("constant string") adds
17289 the string plus current file name and line number to a per-thread
17290 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17291 is like calling CYRPTO_pop_info() until the stack is empty.
17292 Also updated memory leak detection code to be multi-thread-safe.
17296 * Add options -text and -noout to pkcs7 utility and delete the
17297 encryption options which never did anything. Update docs.
17301 * Add options to some of the utilities to allow the pass phrase
17302 to be included on either the command line (not recommended on
17303 OSes like Unix) or read from the environment. Update the
17304 manpages and fix a few bugs.
17308 * Add a few manpages for some of the openssl commands.
17312 * Fix the -revoke option in ca. It was freeing up memory twice,
17313 leaking and not finding already revoked certificates.
17317 * Extensive changes to support certificate auxiliary information.
17318 This involves the use of X509_CERT_AUX structure and X509_AUX
17319 functions. An X509_AUX function such as PEM_read_X509_AUX()
17320 can still read in a certificate file in the usual way but it
17321 will also read in any additional "auxiliary information". By
17322 doing things this way a fair degree of compatibility can be
17323 retained: existing certificates can have this information added
17324 using the new 'x509' options.
17326 Current auxiliary information includes an "alias" and some trust
17327 settings. The trust settings will ultimately be used in enhanced
17328 certificate chain verification routines: currently a certificate
17329 can only be trusted if it is self signed and then it is trusted
17334 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17335 The problem was that one of the replacement routines had not been working
17336 since SSLeay releases. For now the offending routine has been replaced
17337 with non-optimised assembler. Even so, this now gives around 95%
17338 performance improvement for 1024 bit RSA signs.
17342 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17343 handling. Most clients have the effective key size in bits equal to
17344 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17345 A few however don't do this and instead use the size of the decrypted key
17346 to determine the RC2 key length and the AlgorithmIdentifier to determine
17347 the effective key length. In this case the effective key length can still
17348 be 40 bits but the key length can be 168 bits for example. This is fixed
17349 by manually forcing an RC2 key into the EVP_PKEY structure because the
17350 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17351 the key length and effective key length are equal.
17355 * Add a bunch of functions that should simplify the creation of
17356 X509_NAME structures. Now you should be able to do:
17357 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17358 and have it automatically work out the correct field type and fill in
17359 the structures. The more adventurous can try:
17360 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17361 and it will (hopefully) work out the correct multibyte encoding.
17365 * Change the 'req' utility to use the new field handling and multibyte
17366 copy routines. Before the DN field creation was handled in an ad hoc
17367 way in req, ca, and x509 which was rather broken and didn't support
17368 BMPStrings or UTF8Strings. Since some software doesn't implement
17369 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17370 using the dirstring_type option. See the new comment in the default
17371 openssl.cnf for more info.
17375 * Make crypto/rand/md_rand.c more robust:
17376 - Assure unique random numbers after fork().
17377 - Make sure that concurrent threads access the global counter and
17378 md serializably so that we never lose entropy in them
17379 or use exactly the same state in multiple threads.
17380 Access to the large state is not always serializable because
17381 the additional locking could be a performance killer, and
17382 md should be large enough anyway.
17386 * New file `apps/app_rand.c` with commonly needed functionality
17387 for handling the random seed file.
17389 Use the random seed file in some applications that previously did not:
17391 dsaparam -genkey (which also ignored its '-rand' option),
17394 x509 (when signing).
17395 Except on systems with /dev/urandom, it is crucial to have a random
17396 seed file at least for key creation, DSA signing, and for DH exchanges;
17397 for RSA signatures we could do without one.
17399 gendh and gendsa (unlike genrsa) used to read only the first byte
17400 of each file listed in the '-rand' option. The function as previously
17401 found in genrsa is now in app_rand.c and is used by all programs
17402 that support '-rand'.
17406 * In RAND_write_file, use mode 0600 for creating files;
17407 don't just chmod when it may be too late.
17411 * Report an error from X509_STORE_load_locations
17412 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17416 * New function ASN1_mbstring_copy() this copies a string in either
17417 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17418 into an ASN1_STRING type. A mask of permissible types is passed
17419 and it chooses the "minimal" type to use or an error if not type
17424 * Add function equivalents to the various macros in asn1.h. The old
17425 macros are retained with an `M_` prefix. Code inside the library can
17426 use the `M_` macros. External code (including the openssl utility)
17427 should *NOT* in order to be "shared library friendly".
17431 * Add various functions that can check a certificate's extensions
17432 to see if it usable for various purposes such as SSL client,
17433 server or S/MIME and CAs of these types. This is currently
17434 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17435 verification. Also added a -purpose flag to x509 utility to
17436 print out all the purposes.
17440 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17445 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17446 for, obtain and decode and extension and obtain its critical flag.
17447 This allows all the necessary extension code to be handled in a
17448 single function call.
17452 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17453 platforms. See crypto/rc4/rc4_enc.c for further details.
17457 * New -noout option to asn1parse. This causes no output to be produced
17458 its main use is when combined with -strparse and -out to extract data
17459 from a file (which may not be in ASN.1 format).
17463 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17464 when producing the local key id.
17466 *Richard Levitte <levitte@stacken.kth.se>*
17468 * New option -dhparam in s_server. This allows a DH parameter file to be
17469 stated explicitly. If it is not stated then it tries the first server
17470 certificate file. The previous behaviour hard coded the filename
17475 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17476 a public key to be input or output. For example:
17477 openssl rsa -in key.pem -pubout -out pubkey.pem
17478 Also added necessary DSA public key functions to handle this.
17482 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17483 in the message. This was handled by allowing
17484 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17486 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17488 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17489 to the end of the strings whereas this didn't. This would cause problems
17490 if strings read with d2i_ASN1_bytes() were later modified.
17492 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17494 * Fix for base64 decode bug. When a base64 bio reads only one line of
17495 data and it contains EOF it will end up returning an error. This is
17496 caused by input 46 bytes long. The cause is due to the way base64
17497 BIOs find the start of base64 encoded data. They do this by trying a
17498 trial decode on each line until they find one that works. When they
17499 do a flag is set and it starts again knowing it can pass all the
17500 data directly through the decoder. Unfortunately it doesn't reset
17501 the context it uses. This means that if EOF is reached an attempt
17502 is made to pass two EOFs through the context and this causes the
17503 resulting error. This can also cause other problems as well. As is
17504 usual with these problems it takes *ages* to find and the fix is
17505 trivial: move one line.
17507 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17509 * Ugly workaround to get s_client and s_server working under Windows. The
17510 old code wouldn't work because it needed to select() on sockets and the
17511 tty (for keypresses and to see if data could be written). Win32 only
17512 supports select() on sockets so we select() with a 1s timeout on the
17513 sockets and then see if any characters are waiting to be read, if none
17514 are present then we retry, we also assume we can always write data to
17515 the tty. This isn't nice because the code then blocks until we've
17516 received a complete line of data and it is effectively polling the
17517 keyboard at 1s intervals: however it's quite a bit better than not
17518 working at all :-) A dedicated Windows application might handle this
17519 with an event loop for example.
17523 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17524 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17525 will be called when RSA_sign() and RSA_verify() are used. This is useful
17526 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17527 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17528 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17529 This necessitated the support of an extra signature type NID_md5_sha1
17530 for SSL signatures and modifications to the SSL library to use it instead
17531 of calling RSA_public_decrypt() and RSA_private_encrypt().
17535 * Add new -verify -CAfile and -CApath options to the crl program, these
17536 will lookup a CRL issuers certificate and verify the signature in a
17537 similar way to the verify program. Tidy up the crl program so it
17538 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17539 less strict. It will now permit CRL extensions even if it is not
17540 a V2 CRL: this will allow it to tolerate some broken CRLs.
17544 * Initialize all non-automatic variables each time one of the openssl
17545 sub-programs is started (this is necessary as they may be started
17546 multiple times from the "OpenSSL>" prompt).
17548 *Lennart Bang, Bodo Moeller*
17550 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17551 removing all other RSA functionality (this is what NO_RSA does). This
17552 is so (for example) those in the US can disable those operations covered
17553 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17558 * Non-copying interface to BIO pairs.
17559 (still largely untested)
17563 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17564 ASCII string. This was handled independently in various places before.
17568 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17569 UTF8 strings a character at a time.
17573 * Use client_version from client hello to select the protocol
17574 (s23_srvr.c) and for RSA client key exchange verification
17575 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17579 * Add various utility functions to handle SPKACs, these were previously
17580 handled by poking round in the structure internals. Added new function
17581 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17582 print, verify and generate SPKACs. Based on an original idea from
17583 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17587 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17591 * Allow the config file extension section to be overwritten on the
17592 command line. Based on an original idea from Massimiliano Pala
17593 <madwolf@comune.modena.it>. The new option is called -extensions
17594 and can be applied to ca, req and x509. Also -reqexts to override
17595 the request extensions in req and -crlexts to override the crl extensions
17600 * Add new feature to the SPKAC handling in ca. Now you can include
17601 the same field multiple times by preceding it by "XXXX." for example:
17604 this is the same syntax as used in the req config file.
17608 * Allow certificate extensions to be added to certificate requests. These
17609 are specified in a 'req_extensions' option of the req section of the
17610 config file. They can be printed out with the -text option to req but
17611 are otherwise ignored at present.
17615 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17616 data read consists of only the final block it would not decrypted because
17617 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17618 A misplaced 'break' also meant the decrypted final block might not be
17619 copied until the next read.
17623 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17624 a few extra parameters to the DH structure: these will be useful if
17625 for example we want the value of 'q' or implement X9.42 DH.
17629 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17630 provides hooks that allow the default DSA functions or functions on a
17631 "per key" basis to be replaced. This allows hardware acceleration and
17632 hardware key storage to be handled without major modification to the
17633 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17634 associated functions.
17638 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17639 as "read only": it can't be written to and the buffer it points to will
17640 not be freed. Reading from a read only BIO is much more efficient than
17641 a normal memory BIO. This was added because there are several times when
17642 an area of memory needs to be read from a BIO. The previous method was
17643 to create a memory BIO and write the data to it, this results in two
17644 copies of the data and an O(n^2) reading algorithm. There is a new
17645 function BIO_new_mem_buf() which creates a read only memory BIO from
17646 an area of memory. Also modified the PKCS#7 routines to use read only
17651 * Bugfix: ssl23_get_client_hello did not work properly when called in
17652 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17653 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17654 but a retry condition occurred while trying to read the rest.
17658 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17659 NID_pkcs7_encrypted by default: this was wrong since this should almost
17660 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17661 the encrypted data type: this is a more sensible place to put it and it
17662 allows the PKCS#12 code to be tidied up that duplicated this
17667 * Changed obj_dat.pl script so it takes its input and output files on
17668 the command line. This should avoid shell escape redirection problems
17673 * Initial support for certificate extension requests, these are included
17674 in things like Xenroll certificate requests. Included functions to allow
17675 extensions to be obtained and added.
17679 * -crlf option to s_client and s_server for sending newlines as
17680 CRLF (as required by many protocols).
17684 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17686 * Install libRSAglue.a when OpenSSL is built with RSAref.
17688 *Ralf S. Engelschall*
17690 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17692 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17694 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17699 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17700 DH parameters/keys (q is lost during that conversion, but the resulting
17701 DH parameters contain its length).
17703 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17704 much faster than DH_generate_parameters (which creates parameters
17705 where `p = 2*q + 1`), and also the smaller q makes DH computations
17706 much more efficient (160-bit exponentiation instead of 1024-bit
17707 exponentiation); so this provides a convenient way to support DHE
17708 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17709 utter importance to use
17710 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17712 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17713 when such DH parameters are used, because otherwise small subgroup
17714 attacks may become possible!
17718 * Avoid memory leak in i2d_DHparams.
17722 * Allow the -k option to be used more than once in the enc program:
17723 this allows the same encrypted message to be read by multiple recipients.
17727 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17728 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17729 it will always use the numerical form of the OID, even if it has a short
17734 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17735 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17736 otherwise bn_mod_exp was called. In the case of hardware keys for example
17737 no private key components need be present and it might store extra data
17738 in the RSA structure, which cannot be accessed from bn_mod_exp.
17739 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17740 private key operations.
17744 * Added support for SPARC Linux.
17748 * pem_password_cb function type incompatibly changed from
17749 typedef int pem_password_cb(char *buf, int size, int rwflag);
17751 ....(char *buf, int size, int rwflag, void *userdata);
17752 so that applications can pass data to their callbacks:
17753 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17754 additional void * argument, which is just handed through whenever
17755 the password callback is called.
17757 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17759 New function SSL_CTX_set_default_passwd_cb_userdata.
17761 Compatibility note: As many C implementations push function arguments
17762 onto the stack in reverse order, the new library version is likely to
17763 interoperate with programs that have been compiled with the old
17764 pem_password_cb definition (PEM_whatever takes some data that
17765 happens to be on the stack as its last argument, and the callback
17766 just ignores this garbage); but there is no guarantee whatsoever that
17769 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17770 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17771 problems not only on Windows, but also on some Unix platforms.
17772 To avoid problematic command lines, these definitions are now in an
17773 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17774 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17778 * MIPS III/IV assembler module is reimplemented.
17782 * More DES library cleanups: remove references to srand/rand and
17783 delete an unused file.
17787 * Add support for the free Netwide assembler (NASM) under Win32,
17788 since not many people have MASM (ml) and it can be hard to obtain.
17789 This is currently experimental but it seems to work OK and pass all
17790 the tests. Check out INSTALL.W32 for info.
17794 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17795 without temporary keys kept an extra copy of the server key,
17796 and connections with temporary keys did not free everything in case
17801 * New function RSA_check_key and new openssl rsa option -check
17802 for verifying the consistency of RSA keys.
17804 *Ulf Moeller, Bodo Moeller*
17806 * Various changes to make Win32 compile work:
17807 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17808 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17809 comparison" warnings.
17810 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17814 * Add a debugging option to PKCS#5 v2 key generation function: when
17815 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17816 derived keys are printed to stderr.
17820 * Copy the flags in ASN1_STRING_dup().
17822 *Roman E. Pavlov <pre@mo.msk.ru>*
17824 * The x509 application mishandled signing requests containing DSA
17825 keys when the signing key was also DSA and the parameters didn't match.
17827 It was supposed to omit the parameters when they matched the signing key:
17828 the verifying software was then supposed to automatically use the CA's
17829 parameters if they were absent from the end user certificate.
17831 Omitting parameters is no longer recommended. The test was also
17832 the wrong way round! This was probably due to unusual behaviour in
17833 EVP_cmp_parameters() which returns 1 if the parameters match.
17834 This meant that parameters were omitted when they *didn't* match and
17835 the certificate was useless. Certificates signed with 'ca' didn't have
17838 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17840 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17841 The interface is as follows:
17842 Applications can use
17843 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17844 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17845 "off" is now the default.
17846 The library internally uses
17847 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17848 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17849 to disable memory-checking temporarily.
17851 Some inconsistent states that previously were possible (and were
17852 even the default) are now avoided.
17854 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17855 with each memory chunk allocated; this is occasionally more helpful
17856 than just having a counter.
17858 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17860 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17865 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17866 which largely parallels "options", but is for changing API behaviour,
17867 whereas "options" are about protocol behaviour.
17868 Initial "mode" flags are:
17870 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17871 a single record has been written.
17872 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17873 retries use the same buffer location.
17874 (But all of the contents must be
17879 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17882 * Fix problems with no-hmac etc.
17884 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17886 * New functions RSA_get_default_method(), RSA_set_method() and
17887 RSA_get_method(). These allows replacement of RSA_METHODs without having
17888 to mess around with the internals of an RSA structure.
17892 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17893 Also really enable memory leak checks in openssl.c and in some
17896 *Chad C. Mulligan, Bodo Moeller*
17898 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17899 up the length of negative integers. This has now been simplified to just
17900 store the length when it is first determined and use it later, rather
17901 than trying to keep track of where data is copied and updating it to
17903 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17905 * Add a new function PKCS7_signatureVerify. This allows the verification
17906 of a PKCS#7 signature but with the signing certificate passed to the
17907 function itself. This contrasts with PKCS7_dataVerify which assumes the
17908 certificate is present in the PKCS#7 structure. This isn't always the
17909 case: certificates can be omitted from a PKCS#7 structure and be
17910 distributed by "out of band" means (such as a certificate database).
17914 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17915 function prototypes in pem.h, also change util/mkdef.pl to add the
17916 necessary function names.
17920 * mk1mf.pl (used by Windows builds) did not properly read the
17921 options set by Configure in the top level Makefile, and Configure
17922 was not even able to write more than one option correctly.
17923 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17927 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17928 file to be loaded from a BIO or FILE pointer. The BIO version will
17929 for example allow memory BIOs to contain config info.
17933 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17934 Whoever hopes to achieve shared-library compatibility across versions
17935 must use this, not the compile-time macro.
17936 (Exercise 0.9.4: Which is the minimum library version required by
17938 Note: All this applies only to multi-threaded programs, others don't
17943 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17944 through a BIO pair triggered the default case, i.e.
17945 SSLerr(...,SSL_R_UNKNOWN_STATE).
17949 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17950 can use the SSL library even if none of the specific BIOs is
17955 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17956 for the encoded length.
17958 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17960 * Add initial documentation of the X509V3 functions.
17964 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17965 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17966 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17967 secure PKCS#8 private key format with a high iteration count.
17971 * Fix determination of Perl interpreter: A perl or perl5
17972 *directory* in $PATH was also accepted as the interpreter.
17974 *Ralf S. Engelschall*
17976 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17977 wrong with it but it was very old and did things like calling
17978 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17979 unusual formatting.
17983 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17984 to use the new extension code.
17988 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17989 with macros. This should make it easier to change their form, add extra
17990 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17995 * Add to configuration table a new entry that can specify an alternative
17996 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17997 according to Mark Crispin <MRC@Panda.COM>.
18001 * DES CBC did not update the IV. Weird.
18005 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18006 Changing the behaviour of the former might break existing programs --
18007 where IV updating is needed, des_ncbc_encrypt can be used.
18010 * When bntest is run from "make test" it drives bc to check its
18011 calculations, as well as internally checking them. If an internal check
18012 fails, it needs to cause bc to give a non-zero result or make test carries
18013 on without noticing the failure. Fixed.
18017 * DES library cleanups.
18021 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18022 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18023 ciphers. NOTE: although the key derivation function has been verified
18024 against some published test vectors it has not been extensively tested
18025 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18030 * Instead of "mkdir -p", which is not fully portable, use new
18031 Perl script "util/mkdir-p.pl".
18035 * Rewrite the way password based encryption (PBE) is handled. It used to
18036 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18037 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18038 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18039 the 'parameter' field of the AlgorithmIdentifier is passed to the
18040 underlying key generation function so it must do its own ASN1 parsing.
18041 This has also changed the EVP_PBE_CipherInit() function which now has a
18042 'parameter' argument instead of literal salt and iteration count values
18043 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18047 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18048 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18049 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18050 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18051 value was just used as a "magic string" and not used directly its
18052 value doesn't matter.
18056 * Introduce some semblance of const correctness to BN. Shame C doesn't
18061 * "linux-sparc64" configuration (ultrapenguin).
18063 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18064 "linux-sparc" configuration.
18066 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18068 * config now generates no-xxx options for missing ciphers.
18072 * Support the EBCDIC character set (work in progress).
18073 File ebcdic.c not yet included because it has a different license.
18075 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18077 * Support BS2000/OSD-POSIX.
18079 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18081 * Make callbacks for key generation use `void *` instead of `char *`.
18085 * Make S/MIME samples compile (not yet tested).
18089 * Additional typesafe stacks.
18093 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18097 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18099 * New configuration variant "sco5-gcc".
18101 * Updated some demos.
18103 *Sean O Riordain, Wade Scholine*
18105 * Add missing BIO_free at exit of pkcs12 application.
18109 * Fix memory leak in conf.c.
18113 * Updates for Win32 to assembler version of MD5.
18117 * Set #! path to perl in `apps/der_chop` to where we found it
18118 instead of using a fixed path.
18122 * SHA library changes for irix64-mips4-cc.
18126 * Improvements for VMS support.
18130 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18132 * Bignum library bug fix. IRIX 6 passes "make test" now!
18133 This also avoids the problems with SC4.2 and unpatched SC5.
18135 *Andy Polyakov <appro@fy.chalmers.se>*
18137 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18138 These are required because of the typesafe stack would otherwise break
18139 existing code. If old code used a structure member which used to be STACK
18140 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18141 sk_num or sk_value it would produce an error because the num, data members
18142 are not present in STACK_OF. Now it just produces a warning. sk_set
18143 replaces the old method of assigning a value to sk_value
18144 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18145 that does this will no longer work (and should use sk_set instead) but
18146 this could be regarded as a "questionable" behaviour anyway.
18150 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18151 correctly handle encrypted S/MIME data.
18155 * Change type of various DES function arguments from des_cblock
18156 (which means, in function argument declarations, pointer to char)
18157 to des_cblock * (meaning pointer to array with 8 char elements),
18158 which allows the compiler to do more typechecking; it was like
18159 that back in SSLeay, but with lots of ugly casts.
18161 Introduce new type const_des_cblock.
18165 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18166 problems: find RecipientInfo structure that matches recipient certificate
18167 and initialise the ASN1 structures properly based on passed cipher.
18171 * Belatedly make the BN tests actually check the results.
18175 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18176 to and from BNs: it was completely broken. New compilation option
18177 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18178 key elements as negative integers.
18182 * Reorganize and speed up MD5.
18184 *Andy Polyakov <appro@fy.chalmers.se>*
18188 *Richard Levitte <richard@levitte.org>*
18190 * New option -out to asn1parse to allow the parsed structure to be
18191 output to a file. This is most useful when combined with the -strparse
18192 option to examine the output of things like OCTET STRINGS.
18196 * Make SSL library a little more fool-proof by not requiring any longer
18197 that `SSL_set_{accept,connect}_state` be called before
18198 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18199 in many applications because usually everything *appeared* to work as
18200 intended anyway -- now it really works as intended).
18204 * Move openssl.cnf out of lib/.
18208 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18209 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18210 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18212 *Ralf S. Engelschall*
18214 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18215 handle PKCS#7 enveloped data properly.
18217 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18219 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18220 copying pointers. The cert_st handling is changed by this in
18221 various ways (and thus what used to be known as ctx->default_cert
18222 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18223 any longer when s->cert does not give us what we need).
18224 ssl_cert_instantiate becomes obsolete by this change.
18225 As soon as we've got the new code right (possibly it already is?),
18226 we have solved a couple of bugs of the earlier code where s->cert
18227 was used as if it could not have been shared with other SSL structures.
18229 Note that using the SSL API in certain dirty ways now will result
18230 in different behaviour than observed with earlier library versions:
18231 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18232 does not influence s as it used to.
18234 In order to clean up things more thoroughly, inside SSL_SESSION
18235 we don't use CERT any longer, but a new structure SESS_CERT
18236 that holds per-session data (if available); currently, this is
18237 the peer's certificate chain and, for clients, the server's certificate
18238 and temporary key. CERT holds only those values that can have
18239 meaningful defaults in an SSL_CTX.
18243 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18244 from the internal representation. Various PKCS#7 fixes: remove some
18245 evil casts and set the enc_dig_alg field properly based on the signing
18250 * Allow PKCS#12 password to be set from the command line or the
18251 environment. Let 'ca' get its config file name from the environment
18252 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18257 * Allow certificate policies extension to use an IA5STRING for the
18258 organization field. This is contrary to the PKIX definition but
18259 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18264 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18265 without disallowing inline assembler and the like for non-pedantic builds.
18269 * Support Borland C++ builder.
18271 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18277 * SHA-1 cleanups and performance enhancements.
18279 *Andy Polyakov <appro@fy.chalmers.se>*
18281 * Sparc v8plus assembler for the bignum library.
18283 *Andy Polyakov <appro@fy.chalmers.se>*
18285 * Accept any -xxx and +xxx compiler options in Configure.
18289 * Update HPUX configuration.
18293 * Add missing `sk_<type>_unshift()` function to safestack.h
18295 *Ralf S. Engelschall*
18297 * New function SSL_CTX_use_certificate_chain_file that sets the
18298 "extra_cert"s in addition to the certificate. (This makes sense
18299 only for "PEM" format files, as chains as a whole are not
18304 * Support verify_depth from the SSL API.
18305 x509_vfy.c had what can be considered an off-by-one-error:
18306 Its depth (which was not part of the external interface)
18307 was actually counting the number of certificates in a chain;
18308 now it really counts the depth.
18312 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18313 instead of X509err, which often resulted in confusing error
18314 messages since the error codes are not globally unique
18315 (e.g. an alleged error in ssl3_accept when a certificate
18316 didn't match the private key).
18318 * New function SSL_CTX_set_session_id_context that allows to set a default
18319 value (so that you don't need SSL_set_session_id_context for each
18320 connection using the SSL_CTX).
18324 * OAEP decoding bug fix.
18328 * Support INSTALL_PREFIX for package builders, as proposed by
18333 * New Configure options "threads" and "no-threads". For systems
18334 where the proper compiler options are known (currently Solaris
18335 and Linux), "threads" is the default.
18339 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18343 * Install various scripts to $(OPENSSLDIR)/misc, not to
18344 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18345 such as /usr/local/bin.
18349 * "make linux-shared" to build shared libraries.
18351 *Niels Poppe <niels@netbox.org>*
18353 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18357 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18358 extension adding in x509 utility.
18362 * Remove NOPROTO sections and error code comments.
18366 * Partial rewrite of the DEF file generator to now parse the ANSI
18371 * New Configure options --prefix=DIR and --openssldir=DIR.
18375 * Complete rewrite of the error code script(s). It is all now handled
18376 by one script at the top level which handles error code gathering,
18377 header rewriting and C source file generation. It should be much better
18378 than the old method: it now uses a modified version of Ulf's parser to
18379 read the ANSI prototypes in all header files (thus the old K&R definitions
18380 aren't needed for error creation any more) and do a better job of
18381 translating function codes into names. The old 'ASN1 error code embedded
18382 in a comment' is no longer necessary and it doesn't use .err files which
18383 have now been deleted. Also the error code call doesn't have to appear all
18384 on one line (which resulted in some large lines...).
18388 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18392 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18393 0 (which usually indicates a closed connection), but continue reading.
18397 * Fix some race conditions.
18401 * Add support for CRL distribution points extension. Add Certificate
18402 Policies and CRL distribution points documentation.
18406 * Move the autogenerated header file parts to crypto/opensslconf.h.
18410 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18411 8 of keying material. Merlin has also confirmed interop with this fix
18412 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18414 *Merlin Hughes <merlin@baltimore.ie>*
18416 * Fix lots of warnings.
18418 *Richard Levitte <levitte@stacken.kth.se>*
18420 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18421 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18423 *Richard Levitte <levitte@stacken.kth.se>*
18425 * Fix problems with sizeof(long) == 8.
18427 *Andy Polyakov <appro@fy.chalmers.se>*
18429 * Change functions to ANSI C.
18433 * Fix typos in error codes.
18435 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18437 * Remove defunct assembler files from Configure.
18441 * SPARC v8 assembler BIGNUM implementation.
18443 *Andy Polyakov <appro@fy.chalmers.se>*
18445 * Support for Certificate Policies extension: both print and set.
18446 Various additions to support the r2i method this uses.
18450 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18451 return a const string when you are expecting an allocated buffer.
18455 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18456 types DirectoryString and DisplayText.
18460 * Add code to allow r2i extensions to access the configuration database,
18461 add an LHASH database driver and add several ctx helper functions.
18465 * Fix an evil bug in bn_expand2() which caused various BN functions to
18466 fail when they extended the size of a BIGNUM.
18470 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18471 support typesafe stack.
18475 * Fix typo in SSL_[gs]et_options().
18477 *Nils Frostberg <nils@medcom.se>*
18479 * Delete various functions and files that belonged to the (now obsolete)
18480 old X509V3 handling code.
18484 * New Configure option "rsaref".
18488 * Don't auto-generate pem.h.
18492 * Introduce type-safe ASN.1 SETs.
18496 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18498 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18500 * Introduce type-safe STACKs. This will almost certainly break lots of code
18501 that links with OpenSSL (well at least cause lots of warnings), but fear
18502 not: the conversion is trivial, and it eliminates loads of evil casts. A
18503 few STACKed things have been converted already. Feel free to convert more.
18504 In the fullness of time, I'll do away with the STACK type altogether.
18508 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18509 specified in `<certfile>` by updating the entry in the index.txt file.
18510 This way one no longer has to edit the index.txt file manually for
18511 revoking a certificate. The -revoke option does the gory details now.
18513 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18515 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18516 `-text` option at all and this way the `-noout -text` combination was
18517 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18519 *Ralf S. Engelschall*
18521 * Make sure a corresponding plain text error message exists for the
18522 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18523 verify callback function determined that a certificate was revoked.
18525 *Ralf S. Engelschall*
18527 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18528 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18529 all available ciphers including rc5, which was forgotten until now.
18530 In order to let the testing shell script know which algorithms
18531 are available, a new (up to now undocumented) command
18532 `openssl list-cipher-commands` is used.
18536 * Bugfix: s_client occasionally would sleep in select() when
18537 it should have checked SSL_pending() first.
18541 * New functions DSA_do_sign and DSA_do_verify to provide access to
18542 the raw DSA values prior to ASN.1 encoding.
18546 * Tweaks to Configure
18548 *Niels Poppe <niels@netbox.org>*
18550 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18555 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18559 * New config option to avoid instructions that are illegal on the 80386.
18560 The default code is faster, but requires at least a 486.
18564 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18565 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18566 same as SSL2_VERSION anyway.
18570 * New "-showcerts" option for s_client.
18574 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18575 application. Various cleanups and fixes.
18579 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18580 modify error routines to work internally. Add error codes and PBE init
18581 to library startup routines.
18585 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18586 packing functions to asn1 and evp. Changed function names and error
18587 codes along the way.
18591 * PKCS12 integration: and so it begins... First of several patches to
18592 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18593 objects to objects.h
18597 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18598 and display support for Thawte strong extranet extension.
18602 * Add LinuxPPC support.
18604 *Jeff Dubrule <igor@pobox.org>*
18606 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18607 bn_div_words in alpha.s.
18609 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18611 * Make sure the RSA OAEP test is skipped under -DRSAref because
18612 OAEP isn't supported when OpenSSL is built with RSAref.
18614 *Ulf Moeller <ulf@fitug.de>*
18616 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18617 so they no longer are missing under -DNOPROTO.
18619 *Soren S. Jorvang <soren@t.dk>*
18621 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18623 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18624 doesn't work when the session is reused. Coming soon!
18628 * Fix a security hole, that allows sessions to be reused in the wrong
18629 context thus bypassing client cert protection! All software that uses
18630 client certs and session caches in multiple contexts NEEDS PATCHING to
18631 allow session reuse! A fuller solution is in the works.
18633 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18635 * Some more source tree cleanups (removed obsolete files
18636 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18637 permission on "config" script to be executable) and a fix for the INSTALL
18640 *Ulf Moeller <ulf@fitug.de>*
18642 * Remove some legacy and erroneous uses of malloc, free instead of
18645 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18647 * Make rsa_oaep_test return non-zero on error.
18649 *Ulf Moeller <ulf@fitug.de>*
18651 * Add support for native Solaris shared libraries. Configure
18652 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18653 if someone would make that last step automatic.
18655 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18657 * ctx_size was not built with the right compiler during "make links". Fixed.
18661 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18662 except NULL ciphers". This means the default cipher list will no longer
18663 enable NULL ciphers. They need to be specifically enabled e.g. with
18664 the string "DEFAULT:eNULL".
18668 * Fix to RSA private encryption routines: if p < q then it would
18669 occasionally produce an invalid result. This will only happen with
18670 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18674 * Be less restrictive and allow also `perl util/perlpath.pl
18675 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18676 because this way one can also use an interpreter named `perl5` (which is
18677 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18678 installed as `perl`).
18680 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18682 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18684 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18686 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18687 advapi32.lib to Win32 build and change the pem test comparison
18688 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18689 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18690 and crypto/des/ede_cbcm_enc.c.
18694 * DES quad checksum was broken on big-endian architectures. Fixed.
18698 * Comment out two functions in bio.h that aren't implemented. Fix up the
18699 Win32 test batch file so it (might) work again. The Win32 test batch file
18700 is horrible: I feel ill....
18704 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18705 in e_os.h. Audit of header files to check ANSI and non ANSI
18706 sections: 10 functions were absent from non ANSI section and not exported
18707 from Windows DLLs. Fixed up libeay.num for new functions.
18711 * Make `openssl version` output lines consistent.
18713 *Ralf S. Engelschall*
18715 * Fix Win32 symbol export lists for BIO functions: Added
18716 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18717 to ms/libeay{16,32}.def.
18719 *Ralf S. Engelschall*
18721 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18722 fine under Unix and passes some trivial tests I've now added. But the
18723 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18724 added to make sure no one expects that this stuff really works in the
18725 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18726 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18729 *Ralf S. Engelschall*
18731 * Fix the generation of two part addresses in perl.
18733 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18735 * Add config entry for Linux on MIPS.
18737 *John Tobey <jtobey@channel1.com>*
18739 * Make links whenever Configure is run, unless we are on Windoze.
18743 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18744 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18749 * Add a useful kludge to allow package maintainers to specify compiler and
18750 other platforms details on the command line without having to patch the
18751 Configure script every time: One now can use
18752 `perl Configure <id>:<details>`,
18753 i.e. platform ids are allowed to have details appended
18754 to them (separated by colons). This is treated as there would be a static
18755 pre-configured entry in Configure's %table under key `<id>` with value
18756 `<details>` and `perl Configure <id>` is called. So, when you want to
18757 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18758 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18759 now, which overrides the FreeBSD-elf entry on-the-fly.
18761 *Ralf S. Engelschall*
18763 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18767 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18768 on the `perl Configure ...` command line. This way one can compile
18769 OpenSSL libraries with Position Independent Code (PIC) which is needed
18770 for linking it into DSOs.
18772 *Ralf S. Engelschall*
18774 * Remarkably, export ciphers were totally broken and no-one had noticed!
18779 * Cleaned up the LICENSE document: The official contact for any license
18780 questions now is the OpenSSL core team under openssl-core@openssl.org.
18781 And add a paragraph about the dual-license situation to make sure people
18782 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18783 to the OpenSSL toolkit.
18785 *Ralf S. Engelschall*
18787 * General source tree makefile cleanups: Made `making xxx in yyy...`
18788 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18789 Additionally cleaned up the `make links` target: Remove unnecessary
18790 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18791 to speed processing and no longer clutter the display with confusing
18792 stuff. Instead only the actually done links are displayed.
18794 *Ralf S. Engelschall*
18796 * Permit null encryption ciphersuites, used for authentication only. It used
18797 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18798 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18803 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18804 signed attributes when verifying signatures (this would break them),
18805 the detached data encoding was wrong and public keys obtained using
18806 X509_get_pubkey() weren't freed.
18810 * Add text documentation for the BUFFER functions. Also added a work around
18811 to a Win95 console bug. This was triggered by the password read stuff: the
18812 last character typed gets carried over to the next fread(). If you were
18813 generating a new cert request using 'req' for example then the last
18814 character of the passphrase would be CR which would then enter the first
18819 * Added the new 'Includes OpenSSL Cryptography Software' button as
18820 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18821 button and can be used by applications based on OpenSSL to show the
18822 relationship to the OpenSSL project.
18824 *Ralf S. Engelschall*
18826 * Remove confusing variables in function signatures in files
18827 ssl/ssl_lib.c and ssl/ssl.h.
18829 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18831 * Don't install bss_file.c under PREFIX/include/
18833 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18835 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18836 functions that return function pointers and has support for NT specific
18837 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18838 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18839 unsigned to signed types: this was killing the Win32 compile.
18843 * Add new certificate file to stack functions,
18844 SSL_add_dir_cert_subjects_to_stack() and
18845 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18846 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18847 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18848 This means that Apache-SSL and similar packages don't have to mess around
18849 to add as many CAs as they want to the preferred list.
18853 * Experiment with doxygen documentation. Currently only partially applied to
18855 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18856 openssl.doxy as the configuration file.
18860 * Get rid of remaining C++-style comments which strict C compilers hate.
18862 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18864 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18865 compiled in by default: it has problems with large keys.
18869 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18870 DH private keys and/or callback functions which directly correspond to
18871 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18872 is needed for applications which have to configure certificates on a
18873 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18875 For the RSA certificate situation is makes no difference, but
18876 for the DSA certificate situation this fixes the "no shared cipher"
18877 problem where the OpenSSL cipher selection procedure failed because the
18878 temporary keys were not overtaken from the context and the API provided
18879 no way to reconfigure them.
18880 The new functions now let applications reconfigure the stuff and they
18881 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18882 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18883 non-public-API function ssl_cert_instantiate() is used as a helper
18884 function and also to reduce code redundancy inside ssl_rsa.c.
18886 *Ralf S. Engelschall*
18888 * Move s_server -dcert and -dkey options out of the undocumented feature
18889 area because they are useful for the DSA situation and should be
18890 recognized by the users.
18892 *Ralf S. Engelschall*
18894 * Fix the cipher decision scheme for export ciphers: the export bits are
18895 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18896 SSL_EXP_MASK. So, the original variable has to be used instead of the
18897 already masked variable.
18899 *Richard Levitte <levitte@stacken.kth.se>*
18901 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18903 *Richard Levitte <levitte@stacken.kth.se>*
18905 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18906 from `int` to `unsigned int` because it is a length and initialized by
18907 EVP_DigestFinal() which expects an `unsigned int *`.
18909 *Richard Levitte <levitte@stacken.kth.se>*
18911 * Don't hard-code path to Perl interpreter on shebang line of Configure
18912 script. Instead use the usual Shell->Perl transition trick.
18914 *Ralf S. Engelschall*
18916 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18917 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18918 -noout -modulus` as it's already the case for `openssl rsa -noout
18919 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18920 currently the public key is printed (a decision which was already done by
18921 `openssl dsa -modulus` in the past) which serves a similar purpose.
18922 Additionally the NO_RSA no longer completely removes the whole -modulus
18923 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18926 *Ralf S. Engelschall*
18928 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18929 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18931 *Arne Ansper <arne@ats.cyber.ee>*
18933 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18934 to be added. Now both 'req' and 'ca' can use new objects defined in the
18939 * Add cool BIO that does syslog (or event log on NT).
18941 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18943 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18944 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18945 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18946 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18950 * Add preliminary config info for new extension code.
18954 * Make RSA_NO_PADDING really use no padding.
18956 *Ulf Moeller <ulf@fitug.de>*
18958 * Generate errors when private/public key check is done.
18962 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18963 for some CRL extensions and new objects added.
18967 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18968 key usage extension and fuller support for authority key id.
18972 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18973 padding method for RSA, which is recommended for new applications in PKCS
18974 #1 v2.0 (RFC 2437, October 1998).
18975 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18976 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18977 against Bleichbacher's attack on RSA.
18978 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18981 * Updates to the new SSL compression code
18983 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18985 * Fix so that the version number in the master secret, when passed
18986 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18987 (because the server will not accept higher), that the version number
18988 is 0x03,0x01, not 0x03,0x00
18990 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18992 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18993 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18994 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18998 * Support for RAW extensions where an arbitrary extension can be
18999 created by including its DER encoding. See `apps/openssl.cnf` for
19004 * Make sure latest Perl versions don't interpret some generated C array
19005 code as Perl array code in the crypto/err/err_genc.pl script.
19007 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19009 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19010 not many people have the assembler. Various Win32 compilation fixes and
19011 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19012 build instructions.
19016 * Modify configure script 'Configure' to automatically create crypto/date.h
19017 file under Win32 and also build pem.h from pem.org. New script
19018 util/mkfiles.pl to create the MINFO file on environments that can't do a
19019 'make files': perl util/mkfiles.pl >MINFO should work.
19023 * Major rework of DES function declarations, in the pursuit of correctness
19024 and purity. As a result, many evil casts evaporated, and some weirdness,
19025 too. You may find this causes warnings in your code. Zapping your evil
19026 casts will probably fix them. Mostly.
19030 * Fix for a typo in asn1.h. Bug fix to object creation script
19031 obj_dat.pl. It considered a zero in an object definition to mean
19032 "end of object": none of the objects in objects.h have any zeros
19033 so it wasn't spotted.
19035 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19037 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19038 Masking (CBCM). In the absence of test vectors, the best I have been able
19039 to do is check that the decrypt undoes the encrypt, so far. Send me test
19040 vectors if you have them.
19044 * Correct calculation of key length for export ciphers (too much space was
19045 allocated for null ciphers). This has not been tested!
19049 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19050 message is now correct (it understands "crypto" and "ssl" on its
19051 command line). There is also now an "update" option. This will update
19052 the util/ssleay.num and util/libeay.num files with any new functions.
19054 perl util/mkdef.pl crypto ssl update
19055 it will update them.
19059 * Overhauled the Perl interface:
19060 - ported BN stuff to OpenSSL's different BN library
19061 - made the perl/ source tree CVS-aware
19062 - renamed the package from SSLeay to OpenSSL (the files still contain
19063 their history because I've copied them in the repository)
19064 - removed obsolete files (the test scripts will be replaced
19065 by better Test::Harness variants in the future)
19067 *Ralf S. Engelschall*
19069 * First cut for a very conservative source tree cleanup:
19070 1. merge various obsolete readme texts into doc/ssleay.txt
19071 where we collect the old documents and readme texts.
19072 2. remove the first part of files where I'm already sure that we no
19073 longer need them because of three reasons: either they are just temporary
19074 files which were left by Eric or they are preserved original files where
19075 I've verified that the diff is also available in the CVS via "cvs diff
19076 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19077 the crypto/md/ stuff).
19079 *Ralf S. Engelschall*
19081 * More extension code. Incomplete support for subject and issuer alt
19082 name, issuer and authority key id. Change the i2v function parameters
19083 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19084 what that's for :-) Fix to ASN1 macro which messed up
19085 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19089 * Preliminary support for ENUMERATED type. This is largely copied from the
19094 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19096 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19098 * Make sure `make rehash` target really finds the `openssl` program.
19100 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19102 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19103 like to hear about it if this slows down other processors.
19107 * Add CygWin32 platform information to Configure script.
19109 *Alan Batie <batie@aahz.jf.intel.com>*
19111 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19113 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19115 * New program nseq to manipulate netscape certificate sequences
19119 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19124 * Fixes to BN code. Previously the default was to define BN_RECURSION
19125 but the BN code had some problems that would cause failures when
19126 doing certificate verification and some other functions.
19128 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19130 * Add ASN1 and PEM code to support netscape certificate sequences.
19134 * Add ASN1 and PEM code to support netscape certificate sequences.
19138 * Add several PKIX and private extended key usage OIDs.
19142 * Modify the 'ca' program to handle the new extension code. Modify
19143 openssl.cnf for new extension format, add comments.
19147 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19148 and add a sample to openssl.cnf so req -x509 now adds appropriate
19153 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19154 error code, add initial support to X509_print() and x509 application.
19158 * Takes a deep breath and start adding X509 V3 extension support code. Add
19159 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19160 stuff is currently isolated and isn't even compiled yet.
19164 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19165 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19166 Removed the versions check from X509 routines when loading extensions:
19167 this allows certain broken certificates that don't set the version
19168 properly to be processed.
19172 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19173 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19174 can still be regenerated with "make depend".
19178 * Spelling mistake in C version of CAST-128.
19180 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19182 * Changes to the error generation code. The perl script err-code.pl
19183 now reads in the old error codes and retains the old numbers, only
19184 adding new ones if necessary. It also only changes the .err files if new
19185 codes are added. The makefiles have been modified to only insert errors
19186 when needed (to avoid needlessly modifying header files). This is done
19187 by only inserting errors if the .err file is newer than the auto generated
19188 C file. To rebuild all the error codes from scratch (the old behaviour)
19189 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19190 or delete all the .err files.
19194 * CAST-128 was incorrectly implemented for short keys. The C version has
19195 been fixed, but is untested. The assembler versions are also fixed, but
19196 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19197 to regenerate it if needed.
19198 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19199 Hagino <itojun@kame.net>*
19201 * File was opened incorrectly in randfile.c.
19203 *Ulf Möller <ulf@fitug.de>*
19205 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19206 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19207 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19208 al: it's just almost always a UTCTime. Note this patch adds new error
19209 codes so do a "make errors" if there are problems.
19213 * Correct Linux 1 recognition in config.
19215 *Ulf Möller <ulf@fitug.de>*
19217 * Remove pointless MD5 hash when using DSA keys in ca.
19219 *Anonymous <nobody@replay.com>*
19221 * Generate an error if given an empty string as a cert directory. Also
19222 generate an error if handed NULL (previously returned 0 to indicate an
19223 error, but didn't set one).
19225 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19227 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19231 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19232 parameters. This was causing a warning which killed off the Win32 compile.
19236 * Remove C++ style comments from crypto/bn/bn_local.h.
19238 *Neil Costigan <neil.costigan@celocom.com>*
19240 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19241 based on a text string, looking up short and long names and finally
19242 "dot" format. The "dot" format stuff didn't work. Added new function
19243 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19244 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19245 OID is not part of the table.
19249 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19250 X509_LOOKUP_by_alias().
19254 * Sort openssl functions by name.
19258 * Get the `gendsa` command working and add it to the `list` command. Remove
19259 encryption from sample DSA keys (in case anyone is interested the password
19264 * Make *all* `*_free` functions accept a NULL pointer.
19266 *Frans Heymans <fheymans@isaserver.be>*
19268 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19271 *Anonymous <nobody@replay.com>*
19273 * s_server should send the CAfile as acceptable CAs, not its own cert.
19275 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19277 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19279 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19281 * Temp key "for export" tests were wrong in s3_srvr.c.
19283 *Anonymous <nobody@replay.com>*
19285 * Add prototype for temp key callback functions
19286 SSL_CTX_set_tmp_{rsa,dh}_callback().
19290 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19291 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19295 * X509_name_add_entry() freed the wrong thing after an error.
19297 *Arne Ansper <arne@ats.cyber.ee>*
19299 * rsa_eay.c would attempt to free a NULL context.
19301 *Arne Ansper <arne@ats.cyber.ee>*
19303 * BIO_s_socket() had a broken should_retry() on Windoze.
19305 *Arne Ansper <arne@ats.cyber.ee>*
19307 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19309 *Arne Ansper <arne@ats.cyber.ee>*
19311 * Make sure the already existing X509_STORE->depth variable is initialized
19312 in X509_STORE_new(), but document the fact that this variable is still
19313 unused in the certificate verification process.
19315 *Ralf S. Engelschall*
19317 * Fix the various library and `apps/` files to free up pkeys obtained from
19318 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19322 * Fix reference counting in X509_PUBKEY_get(). This makes
19323 demos/maurice/example2.c work, amongst others, probably.
19325 *Steve Henson and Ben Laurie*
19327 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19328 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19329 are no longer created. This way we have a single and consistent command
19330 line interface `openssl <command>`, similar to `cvs <command>`.
19332 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19334 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19335 BIT STRING wrapper always have zero unused bits.
19339 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19343 * Make the top-level INSTALL documentation easier to understand.
19347 * Makefiles updated to exit if an error occurs in a sub-directory
19348 make (including if user presses ^C) [Paul Sutton]
19350 * Make Montgomery context stuff explicit in RSA data structure.
19354 * Fix build order of pem and err to allow for generated pem.h.
19358 * Fix renumbering bug in X509_NAME_delete_entry().
19362 * Enhanced the err-ins.pl script so it makes the error library number
19363 global and can add a library name. This is needed for external ASN1 and
19364 other error libraries.
19368 * Fixed sk_insert which never worked properly.
19372 * Fix ASN1 macros so they can handle indefinite length constructed
19373 EXPLICIT tags. Some non standard certificates use these: they can now
19378 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19379 into a single doc/ssleay.txt bundle. This way the information is still
19380 preserved but no longer messes up this directory. Now it's new room for
19381 the new set of documentation files.
19383 *Ralf S. Engelschall*
19385 * SETs were incorrectly DER encoded. This was a major pain, because they
19386 shared code with SEQUENCEs, which aren't coded the same. This means that
19387 almost everything to do with SETs or SEQUENCEs has either changed name or
19388 number of arguments.
19390 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19392 * Fix test data to work with the above.
19396 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19397 was already fixed by Eric for 0.9.1 it seems.
19399 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19401 * Autodetect FreeBSD3.
19405 * Fix various bugs in Configure. This affects the following platforms:
19409 unixware-2.0-pentium
19414 * Eliminate generated files from CVS. Reorder tests to regenerate files
19415 before they are needed.
19419 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19423 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19425 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19426 changed SSLeay to OpenSSL in version strings.
19428 *Ralf S. Engelschall*
19430 * Some fixups to the top-level documents.
19434 * Fixed the nasty bug where rsaref.h was not found under compile-time
19435 because the symlink to include/ was missing.
19437 *Ralf S. Engelschall*
19439 * Incorporated the popular no-RSA/DSA-only patches
19440 which allow to compile an RSA-free SSLeay.
19442 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19444 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19445 when "ssleay" is still not found.
19447 *Ralf S. Engelschall*
19449 * Added more platforms to Configure: Cray T3E, HPUX 11,
19451 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19453 * Updated the README file.
19455 *Ralf S. Engelschall*
19457 * Added various .cvsignore files in the CVS repository subdirs
19458 to make a "cvs update" really silent.
19460 *Ralf S. Engelschall*
19462 * Recompiled the error-definition header files and added
19463 missing symbols to the Win32 linker tables.
19465 *Ralf S. Engelschall*
19467 * Cleaned up the top-level documents;
19468 o new files: CHANGES and LICENSE
19469 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19470 o merged COPYRIGHT into LICENSE
19471 o removed obsolete TODO file
19472 o renamed MICROSOFT to INSTALL.W32
19474 *Ralf S. Engelschall*
19476 * Removed dummy files from the 0.9.1b source tree:
19477 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19478 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19479 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19480 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19481 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19483 *Ralf S. Engelschall*
19485 * Added various platform portability fixes.
19489 * The Genesis of the OpenSSL rpject:
19490 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19491 Young and Tim J. Hudson created while they were working for C2Net until
19494 *The OpenSSL Project*
19496 ### Changes between 0.9.0b and 0.9.1b [not released]
19498 * Updated a few CA certificates under certs/
19502 * Changed some BIGNUM api stuff.
19506 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19507 DGUX x86, Linux Alpha, etc.
19511 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19512 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19517 * Add -strparse option to asn1pars program which parses nested
19520 *Dr Stephen Henson <shenson@bigfoot.com>*
19522 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19526 * DSA fix for "ca" program.
19530 * Added "-genkey" option to "dsaparam" program.
19534 * Added RIPE MD160 (rmd160) message digest.
19538 * Added -a (all) option to "ssleay version" command.
19542 * Added PLATFORM define which is the id given to Configure.
19546 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19550 * Extended the ASN.1 parser routines.
19554 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19558 * Added a BN_CTX to the BN library.
19562 * Fixed the weak key values in DES library
19566 * Changed API in EVP library for cipher aliases.
19570 * Added support for RC2/64bit cipher.
19574 * Converted the lhash library to the crypto/mem.c functions.
19578 * Added more recognized ASN.1 object ids.
19582 * Added more RSA padding checks for SSL/TLS.
19586 * Added BIO proxy/filter functionality.
19590 * Added extra_certs to SSL_CTX which can be used
19591 send extra CA certificates to the client in the CA cert chain sending
19592 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19596 * Now Fortezza is denied in the authentication phase because
19597 this is key exchange mechanism is not supported by SSLeay at all.
19601 * Additional PKCS1 checks.
19605 * Support the string "TLSv1" for all TLS v1 ciphers.
19609 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19610 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19614 * Fixed a few memory leaks.
19618 * Fixed various code and comment typos.
19622 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19623 bytes sent in the client random.
19625 *Edward Bishop <ebishop@spyglass.com>*
19629 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19630 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19631 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19632 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19633 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19634 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
19635 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
19636 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
19637 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
19638 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
19639 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
19640 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
19641 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
19642 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
19643 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
19644 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19645 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19646 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19647 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19648 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19649 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19650 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19651 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19652 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19653 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19654 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19655 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19656 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19657 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19658 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19659 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19660 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19661 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19662 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19663 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19664 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19665 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19666 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19667 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19668 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19669 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19670 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19671 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19672 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19673 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19674 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19675 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19676 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19677 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19678 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19679 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19680 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19681 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19682 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19683 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19684 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19685 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19686 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19687 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19688 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19689 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19690 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19691 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19692 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19693 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19694 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19695 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19696 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19697 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19698 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19699 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19700 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19701 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19702 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19703 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19704 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19705 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19706 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19707 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19708 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19709 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19710 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19711 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19712 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19713 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19714 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19715 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19716 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19717 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19718 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19719 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19720 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19721 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19722 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19723 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19724 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19725 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19726 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19727 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19728 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19729 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19730 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19731 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19732 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19733 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19734 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19735 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19736 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19737 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19738 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19739 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19740 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19741 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19742 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19743 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19744 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19745 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19746 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19747 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19748 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19749 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19750 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19751 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19752 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19753 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19754 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19755 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19756 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19757 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19758 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19759 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19760 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19761 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19762 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19763 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19764 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19765 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19766 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19767 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19768 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19769 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19770 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19771 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19772 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19773 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19774 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19775 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19776 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19777 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19778 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19779 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19780 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19781 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19782 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19783 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19784 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19785 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19786 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19787 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19788 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19789 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19790 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19791 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19792 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19793 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19794 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19795 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19796 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19797 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19798 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19799 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19800 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19801 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19802 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19803 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19804 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19805 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655