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 3.2](#openssl-32)
14 - [OpenSSL 3.1](#openssl-31)
15 - [OpenSSL 3.0](#openssl-30)
16 - [OpenSSL 1.1.1](#openssl-111)
17 - [OpenSSL 1.1.0](#openssl-110)
18 - [OpenSSL 1.0.2](#openssl-102)
19 - [OpenSSL 1.0.1](#openssl-101)
20 - [OpenSSL 1.0.0](#openssl-100)
21 - [OpenSSL 0.9.x](#openssl-09x)
26 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
28 * TLS round-trip time calculation was added by a Brigham Young University
29 Capstone team partnering with Sandia National Laboratories. A new function
30 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
35 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
36 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
37 option. Use of the "advanced" s_client command command via the "-adv" option
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 * Added an "advanced" command mode to s_client. Use this with the "-adv"
54 option. The old "basic" command mode recognises certain letters that must
55 always appear at the start of a line and cannot be escaped. The advanced
56 command mode enables commands to be entered anywhere and there is an
57 escaping mechanism. After starting s_client with "-adv" type "{help}"
58 to show a list of available commands.
62 * Add Raw Public Key (RFC7250) support. Authentication is supported
63 by matching keys against either local policy (TLSA records synthesised
64 from the expected keys) or DANE (TLSA records obtained by the
65 application from DNS). TLSA records will also match the same key in
66 the server certificate, should RPK use not happen to be negotiated.
70 * Added EC_GROUP_to_params which creates an OSSL_PARAM array
71 from a given EC_GROUP.
75 * Added support for pluggable (provider-based) TLS signature algorithms.
76 This enables TLS 1.3 authentication operations with algorithms embedded
77 in providers not included by default in OpenSSL. In combination with
78 the already available pluggable KEM and X.509 support, this enables
79 for example suitable providers to deliver post-quantum or quantum-safe
80 cryptography to OpenSSL users.
84 * Added support for Hybrid Public Key Encryption (HPKE) as defined
85 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
86 Message Layer Security (MLS) and other IETF specifications.
87 HPKE can also be used by other applications that require
88 encrypting "to" an ECDH public key. External APIs are defined in
89 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
93 * Add support for certificate compression (RFC8879), including
94 library support for Brotli and Zstandard compression.
98 * Add the ability to add custom attributes to PKCS12 files. Add a new API
99 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
100 for a user specified callback and optional argument.
101 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
102 added to the existing STACK_OF attrs.
106 * Major refactor of the libssl record layer
110 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
114 * Add a mac salt length option for the pkcs12 command.
118 * Add more SRTP protection profiles from RFC8723 and RFC8269.
122 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
124 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
126 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
127 supported and enabled.
131 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
132 to the list of ciphersuites providing Perfect Forward Secrecy as
133 required by SECLEVEL >= 3.
135 *Dmitry Belyavskiy, Nicola Tuveri*
137 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
138 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
139 SSL_get0_iana_groups() function-like macro, retrieves the list of
140 supported groups sent by the peer.
141 The function SSL_client_hello_get_extension_order() populates
142 a caller-supplied array with the list of extension types present in the
143 ClientHello, in order of appearance.
147 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
148 to make it possible to use empty passphrase strings.
152 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
153 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
154 of 160 bits and above and less than 224 bits were previously accepted by
155 default but are now no longer allowed. By default TLS compression was
156 already disabled in previous OpenSSL versions. At security level 2 it cannot
161 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
166 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
167 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
168 will need to load the legacy crypto provider.
171 * CCM8 cipher suites in TLS have been downgraded to security level zero
172 because they use a short authentication tag which lowers their strength.
176 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
181 * Add X.509 certificate codeSigning purpose and related checks on key usage and
182 extended key usage of the leaf certificate according to the CA/Browser Forum.
186 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
187 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
188 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
189 X.509 version 3 if the certificate information includes X.509 extensions.
193 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
194 such as adding a trace facility for debugging certificate chain building.
198 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
199 in particular supporting requests for central key generation, generalized
200 polling, and various types of genm/genp exchanges defined in CMP Updates.
204 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
205 like correcting the TLS and proxy support and adding tracing for debugging.
209 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
213 * `CMS_add0_cert()` and `CMS_add1_cert()` no more throw an error
214 if a certificate to be added is already present.
215 * `CMS_sign_ex()` and `CMS_sign()` now ignore any duplicate certificates
216 in their `certs` argument and no longer throw an error for them.
220 * Fixed and extended `util/check-format.pl` for checking adherence to the
221 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
222 The checks are meanwhile more complete and yield fewer false positives.
226 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
227 sending and receiving multiple messages in a single call. An implementation
228 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
232 * Support for loading root certificates from the Windows certificate store
233 has been added. The support is in the form of a store which recognises the
234 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
235 arguments. This store is built by default and can be disabled using the new
236 compile-time option `no-winstore`. This store is not currently used by
237 default and must be loaded explicitly using the above store URI. It is
238 expected to be loaded by default in the future.
242 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
243 kernel versions that support KTLS have a known bug in CCM processing. That
244 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
245 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
250 * Zerocopy KTLS sendfile() support on Linux.
254 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
255 decryption as a protection against Bleichenbacher-like attacks.
256 The RSA decryption API will now return a randomly generated deterministic
257 message instead of an error in case it detects an error when checking
258 padding during PKCS#1 v1.5 decryption. This is a general protection against
259 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
261 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
262 on the RSA decryption context.
266 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
267 basic thread pool implementation for select platforms.
274 ### Changes between 3.1.0 and 3.1.1 [xx XXX xxxx]
276 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
277 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
279 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
280 numeric text form. For gigantic sub-identifiers, this would take a very
281 long time, the time complexity being O(n^2) where n is the size of that
282 sub-identifier. ([CVE-2023-2650])
284 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
285 IDENTIFIER to canonical numeric text form if the size of that OBJECT
286 IDENTIFIER is 586 bytes or less, and fail otherwise.
288 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
289 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
290 most 128 sub-identifiers, and that the maximum value that each sub-
291 identifier may have is 2^32-1 (4294967295 decimal).
293 For each byte of every sub-identifier, only the 7 lower bits are part of
294 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
295 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
300 * Multiple algorithm implementation fixes for ARM BE platforms.
304 * Added a -pedantic option to fipsinstall that adjusts the various
305 settings to ensure strict FIPS compliance rather than backwards
310 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
311 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
312 trigger a crash of an application using AES-XTS decryption if the memory
313 just after the buffer being decrypted is not mapped.
314 Thanks to Anton Romanov (Amazon) for discovering the issue.
319 * Add FIPS provider configuration option to disallow the use of
320 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
321 The option '-no_drbg_truncated_digests' can optionally be
322 supplied to 'openssl fipsinstall'.
326 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
327 that it does not enable policy checking. Thanks to David Benjamin for
328 discovering this issue.
333 * Fixed an issue where invalid certificate policies in leaf certificates are
334 silently ignored by OpenSSL and other certificate policy checks are skipped
335 for that certificate. A malicious CA could use this to deliberately assert
336 invalid certificate policies in order to circumvent policy checking on the
337 certificate altogether.
342 * Limited the number of nodes created in a policy tree to mitigate
343 against CVE-2023-0464. The default limit is set to 1000 nodes, which
344 should be sufficient for most installations. If required, the limit
345 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
346 time define to a desired maximum number of nodes or zero to allow
352 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
354 * Add FIPS provider configuration option to enforce the
355 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
356 The option '-ems-check' can optionally be supplied to
357 'openssl fipsinstall'.
361 * The FIPS provider includes a few non-approved algorithms for
362 backward compatibility purposes and the "fips=yes" property query
363 must be used for all algorithm fetches to ensure FIPS compliance.
365 The algorithms that are included but not approved are Triple DES ECB,
366 Triple DES CBC and EdDSA.
370 * Added support for KMAC in KBKDF.
374 * RNDR and RNDRRS support in provider functions to provide
375 random number generation for Arm CPUs (aarch64).
379 * s_client and s_server apps now explicitly say when the TLS version
380 does not include the renegotiation mechanism. This avoids confusion
381 between that scenario versus when the TLS version includes secure
382 renegotiation but the peer lacks support for it.
386 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
388 *Tomasz Kantecki, Andrey Matyukov*
390 * The various OBJ_* functions have been made thread safe.
394 * Parallel dual-prime 1536/2048-bit modular exponentiation for
395 AVX512_IFMA capable processors.
397 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
399 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
400 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
401 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
402 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
403 `OPENSSL_NO_DEPRECATED_3_1`.
405 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
406 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
407 definitions for these functions regardless of whether
408 `OPENSSL_NO_DEPRECATED_3_1` is defined.
410 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
411 functions regardless of whether they are using them. It is recommended that
412 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
416 * When generating safe-prime DH parameters set the recommended private key
417 length equivalent to minimum key lengths as in RFC 7919.
421 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
422 maximum size that is smaller or equal to the digest length to comply with
423 FIPS 186-4 section 5. This is implemented by a new option
424 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
425 `rsa_pss_saltlen` parameter, which is now the default. Signature
426 verification is not affected by this change and continues to work as before.
433 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
434 listed here are only a brief description.
435 The migration guide contains more detailed information related to new features,
436 breaking changes, and mappings for the large list of deprecated functions.
438 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
440 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
442 * Fixed NULL dereference during PKCS7 data verification.
444 A NULL pointer can be dereferenced when signatures are being
445 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
446 algorithm used for the signature is known to the OpenSSL library but
447 the implementation of the hash algorithm is not available the digest
448 initialization will fail. There is a missing check for the return
449 value from the initialization function which later leads to invalid
450 usage of the digest API most likely leading to a crash.
453 PKCS7 data is processed by the SMIME library calls and also by the
454 time stamp (TS) library calls. The TLS implementation in OpenSSL does
455 not call these functions however third party applications would be
456 affected if they call these functions to verify signatures on untrusted
461 * Fixed X.400 address type confusion in X.509 GeneralName.
463 There is a type confusion vulnerability relating to X.400 address processing
464 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
465 but the public structure definition for GENERAL_NAME incorrectly specified
466 the type of the x400Address field as ASN1_TYPE. This field is subsequently
467 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
470 When CRL checking is enabled (i.e. the application sets the
471 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
472 pass arbitrary pointers to a memcmp call, enabling them to read memory
473 contents or enact a denial of service.
478 * Fixed NULL dereference validating DSA public key.
480 An invalid pointer dereference on read can be triggered when an
481 application tries to check a malformed DSA public key by the
482 EVP_PKEY_public_check() function. This will most likely lead
483 to an application crash. This function can be called on public
484 keys supplied from untrusted sources which could allow an attacker
485 to cause a denial of service attack.
487 The TLS implementation in OpenSSL does not call this function
488 but applications might call the function if there are additional
489 security requirements imposed by standards such as FIPS 140-3.
492 *Shane Lontis, Tomáš Mráz*
494 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
496 An invalid pointer dereference on read can be triggered when an
497 application tries to load malformed PKCS7 data with the
498 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
500 The result of the dereference is an application crash which could
501 lead to a denial of service attack. The TLS implementation in OpenSSL
502 does not call this function however third party applications might
503 call these functions on untrusted data.
508 * Fixed Use-after-free following BIO_new_NDEF.
510 The public API function BIO_new_NDEF is a helper function used for
511 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
512 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
513 be called directly by end user applications.
515 The function receives a BIO from the caller, prepends a new BIO_f_asn1
516 filter BIO onto the front of it to form a BIO chain, and then returns
517 the new head of the BIO chain to the caller. Under certain conditions,
518 for example if a CMS recipient public key is invalid, the new filter BIO
519 is freed and the function returns a NULL result indicating a failure.
520 However, in this case, the BIO chain is not properly cleaned up and the
521 BIO passed by the caller still retains internal pointers to the previously
522 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
523 then a use-after-free will occur. This will most likely result in a crash.
526 *Viktor Dukhovni, Matt Caswell*
528 * Fixed Double free after calling PEM_read_bio_ex.
530 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
531 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
532 data. If the function succeeds then the "name_out", "header" and "data"
533 arguments are populated with pointers to buffers containing the relevant
534 decoded data. The caller is responsible for freeing those buffers. It is
535 possible to construct a PEM file that results in 0 bytes of payload data.
536 In this case PEM_read_bio_ex() will return a failure code but will populate
537 the header argument with a pointer to a buffer that has already been freed.
538 If the caller also frees this buffer then a double free will occur. This
539 will most likely lead to a crash.
541 The functions PEM_read_bio() and PEM_read() are simple wrappers around
542 PEM_read_bio_ex() and therefore these functions are also directly affected.
544 These functions are also called indirectly by a number of other OpenSSL
545 functions including PEM_X509_INFO_read_bio_ex() and
546 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
547 internal uses of these functions are not vulnerable because the caller does
548 not free the header argument if PEM_read_bio_ex() returns a failure code.
551 *Kurt Roeckx, Matt Caswell*
553 * Fixed Timing Oracle in RSA Decryption.
555 A timing based side channel exists in the OpenSSL RSA Decryption
556 implementation which could be sufficient to recover a plaintext across
557 a network in a Bleichenbacher style attack. To achieve a successful
558 decryption an attacker would have to be able to send a very large number
559 of trial messages for decryption. The vulnerability affects all RSA padding
560 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
563 *Dmitry Belyavsky, Hubert Kario*
565 * Fixed X.509 Name Constraints Read Buffer Overflow.
567 A read buffer overrun can be triggered in X.509 certificate verification,
568 specifically in name constraint checking. The read buffer overrun might
569 result in a crash which could lead to a denial of service attack.
570 In a TLS client, this can be triggered by connecting to a malicious
571 server. In a TLS server, this can be triggered if the server requests
572 client authentication and a malicious client connects.
577 * Fixed X.509 Policy Constraints Double Locking security issue.
579 If an X.509 certificate contains a malformed policy constraint and
580 policy processing is enabled, then a write lock will be taken twice
581 recursively. On some operating systems (most widely: Windows) this
582 results in a denial of service when the affected process hangs. Policy
583 processing being enabled on a publicly facing server is not considered
584 to be a common setup.
589 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
590 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
591 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
592 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
593 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
594 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
595 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
596 for legacy EC and SM2 keys is also changed similarly to honor the
597 equivalent conversion format flag as specified in the underlying
598 `EC_KEY` object being exported to a provider, when this function is
599 called through `EVP_PKEY_export()`.
603 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
605 * Fixed two buffer overflows in punycode decoding functions.
607 A buffer overrun can be triggered in X.509 certificate verification,
608 specifically in name constraint checking. Note that this occurs after
609 certificate chain signature verification and requires either a CA to
610 have signed the malicious certificate or for the application to continue
611 certificate verification despite failure to construct a path to a trusted
614 In a TLS client, this can be triggered by connecting to a malicious
615 server. In a TLS server, this can be triggered if the server requests
616 client authentication and a malicious client connects.
618 An attacker can craft a malicious email address to overflow
619 an arbitrary number of bytes containing the `.` character (decimal 46)
620 on the stack. This buffer overflow could result in a crash (causing a
624 An attacker can craft a malicious email address to overflow four
625 attacker-controlled bytes on the stack. This buffer overflow could
626 result in a crash (causing a denial of service) or potentially remote code
627 execution depending on stack layout for any given platform/compiler.
632 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
633 parameters in OpenSSL code.
634 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
635 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
636 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
637 Using these invalid names may cause algorithms to use slower methods
638 that ignore the CRT parameters.
642 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
647 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
648 data to be signed before signing the certificate.
652 * Added RIPEMD160 to the default provider.
656 * Ensured that the key share group sent or accepted for the key exchange
657 is allowed for the protocol version.
661 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
663 * OpenSSL supports creating a custom cipher via the legacy
664 EVP_CIPHER_meth_new() function and associated function calls. This function
665 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
666 to use the new provider mechanism in order to implement custom ciphers.
668 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
669 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
670 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
671 and decryption initialisation functions). Instead of using the custom cipher
672 directly it incorrectly tries to fetch an equivalent cipher from the
673 available providers. An equivalent cipher is found based on the NID passed to
674 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
675 given cipher. However it is possible for an application to incorrectly pass
676 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
677 is used in this way the OpenSSL encryption/decryption initialisation function
678 will match the NULL cipher as being equivalent and will fetch this from the
679 available providers. This will succeed if the default provider has been
680 loaded (or if a third party provider has been loaded that offers this
681 cipher). Using the NULL cipher means that the plaintext is emitted as the
684 Applications are only affected by this issue if they call
685 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
686 encryption/decryption initialisation function. Applications that only use
687 SSL/TLS are not impacted by this issue.
692 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
697 * Fixed the linux-mips64 Configure target which was missing the
698 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
703 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
708 * Correctly handle a retransmitted ClientHello in DTLS
712 * Fixed detection of ktls support in cross-compile environment on Linux
716 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
721 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
722 report correct results in some cases
726 * Fix UWP builds by defining VirtualLock
730 * For known safe primes use the minimum key length according to RFC 7919.
731 Longer private key sizes unnecessarily raise the cycles needed to compute the
732 shared secret without any increase of the real security. This fixes a
733 regression from 1.1.1 where these shorter keys were generated for the known
738 * Added the loongarch64 target
742 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
743 only passed to the FIPS provider and not to the default or legacy provider.
747 * Fixed reported performance degradation on aarch64. Restored the
748 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
749 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
750 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
751 The new algorithm is still used for 32 bit targets.
755 * Added a missing header for memcmp that caused compilation failure on some
760 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
762 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
763 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
764 This issue makes the RSA implementation with 2048 bit private keys
765 incorrect on such machines and memory corruption will happen during
766 the computation. As a consequence of the memory corruption an attacker
767 may be able to trigger a remote code execution on the machine performing
770 SSL/TLS servers or other servers using 2048 bit RSA private keys running
771 on machines supporting AVX512IFMA instructions of the X86_64 architecture
772 are affected by this issue.
777 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
778 implementation would not encrypt the entirety of the data under some
779 circumstances. This could reveal sixteen bytes of data that was
780 preexisting in the memory that wasn't written. In the special case of
781 "in place" encryption, sixteen bytes of the plaintext would be revealed.
783 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
784 they are both unaffected.
787 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
789 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
791 * In addition to the c_rehash shell command injection identified in
792 CVE-2022-1292, further bugs where the c_rehash script does not
793 properly sanitise shell metacharacters to prevent command injection have been
796 When the CVE-2022-1292 was fixed it was not discovered that there
797 are other places in the script where the file names of certificates
798 being hashed were possibly passed to a command executed through the shell.
800 This script is distributed by some operating systems in a manner where
801 it is automatically executed. On such operating systems, an attacker
802 could execute arbitrary commands with the privileges of the script.
804 Use of the c_rehash script is considered obsolete and should be replaced
805 by the OpenSSL rehash command line tool.
808 *Daniel Fiala, Tomáš Mráz*
810 * Case insensitive string comparison no longer uses locales. It has instead
811 been directly implemented.
815 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
817 * Case insensitive string comparison is reimplemented via new locale-agnostic
818 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
819 comparison. The previous implementation had problems when the Turkish locale
824 * Fixed a bug in the c_rehash script which was not properly sanitising shell
825 metacharacters to prevent command injection. This script is distributed by
826 some operating systems in a manner where it is automatically executed. On
827 such operating systems, an attacker could execute arbitrary commands with the
828 privileges of the script.
830 Use of the c_rehash script is considered obsolete and should be replaced
831 by the OpenSSL rehash command line tool.
836 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
837 certificate on an OCSP response. The bug caused the function in the case
838 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
839 response (meaning a successful verification) even in the case where the
840 response signing certificate fails to verify.
842 It is anticipated that most users of `OCSP_basic_verify` will not use the
843 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
844 a negative value (indicating a fatal error) in the case of a certificate
845 verification failure. The normal expected return value in this case would be
848 This issue also impacts the command line OpenSSL "ocsp" application. When
849 verifying an ocsp response with the "-no_cert_checks" option the command line
850 application will report that the verification is successful even though it
851 has in fact failed. In this case the incorrect successful response will also
852 be accompanied by error messages showing the failure and contradicting the
853 apparently successful result.
858 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
859 AAD data as the MAC key. This made the MAC key trivially predictable.
861 An attacker could exploit this issue by performing a man-in-the-middle attack
862 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
863 that the modified data would still pass the MAC integrity check.
865 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
866 endpoint will always be rejected by the recipient and the connection will
867 fail at that point. Many application protocols require data to be sent from
868 the client to the server first. Therefore, in such a case, only an OpenSSL
869 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
871 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
872 sent in both directions. In this case both clients and servers could be
873 affected, regardless of the application protocol.
875 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
876 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
877 the handshake when using this ciphersuite.
879 The confidentiality of data is not impacted by this issue, i.e. an attacker
880 cannot decrypt data that has been encrypted using this ciphersuite - they can
883 In order for this attack to work both endpoints must legitimately negotiate
884 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
885 OpenSSL 3.0, and is not available within the default provider or the default
886 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
887 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
888 following must have occurred:
890 1) OpenSSL must have been compiled with the (non-default) compile time option
891 enable-weak-ssl-ciphers
893 2) OpenSSL must have had the legacy provider explicitly loaded (either
894 through application code or via configuration)
896 3) The ciphersuite must have been explicitly added to the ciphersuite list
898 4) The libssl security level must have been set to 0 (default is 1)
900 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
902 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
903 others that both endpoints have in common
908 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
909 occupied by the removed hash table entries.
911 This function is used when decoding certificates or keys. If a long lived
912 process periodically decodes certificates or keys its memory usage will
913 expand without bounds and the process might be terminated by the operating
914 system causing a denial of service. Also traversing the empty hash table
915 entries will take increasingly more time.
917 Typically such long lived processes might be TLS clients or TLS servers
918 configured to accept client certificate authentication.
921 *Hugo Landau, Aliaksei Levin*
923 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
924 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
925 statistics are no longer supported. For compatibility, these statistics are
926 still listed in the output but are now always reported as zero.
930 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
932 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
933 for non-prime moduli.
935 Internally this function is used when parsing certificates that contain
936 elliptic curve public keys in compressed form or explicit elliptic curve
937 parameters with a base point encoded in compressed form.
939 It is possible to trigger the infinite loop by crafting a certificate that
940 has invalid explicit curve parameters.
942 Since certificate parsing happens prior to verification of the certificate
943 signature, any process that parses an externally supplied certificate may thus
944 be subject to a denial of service attack. The infinite loop can also be
945 reached when parsing crafted private keys as they can contain explicit
946 elliptic curve parameters.
948 Thus vulnerable situations include:
950 - TLS clients consuming server certificates
951 - TLS servers consuming client certificates
952 - Hosting providers taking certificates or private keys from customers
953 - Certificate authorities parsing certification requests from subscribers
954 - Anything else which parses ASN.1 elliptic curve parameters
956 Also any other applications that use the BN_mod_sqrt() where the attacker
957 can control the parameter values are vulnerable to this DoS issue.
962 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
963 to the list of ciphersuites providing Perfect Forward Secrecy as
964 required by SECLEVEL >= 3.
966 *Dmitry Belyavskiy, Nicola Tuveri*
968 * Made the AES constant time code for no-asm configurations
969 optional due to the resulting 95% performance degradation.
970 The AES constant time code can be enabled, for no assembly
971 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
975 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
980 * The negative return value handling of the certificate verification callback
981 was reverted. The replacement is to set the verification retry state with
982 the SSL_set_retry_verify() function.
986 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
988 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
989 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
990 verify a certificate supplied by a server. That function may return a
991 negative return value to indicate an internal error (for example out of
992 memory). Such a negative return value is mishandled by OpenSSL and will cause
993 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
994 success and a subsequent call to SSL_get_error() to return the value
995 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
996 returned by OpenSSL if the application has previously called
997 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
998 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
999 totally unexpected and applications may not behave correctly as a result. The
1000 exact behaviour will depend on the application but it could result in
1001 crashes, infinite loops or other similar incorrect responses.
1003 This issue is made more serious in combination with a separate bug in OpenSSL
1004 3.0 that will cause X509_verify_cert() to indicate an internal error when
1005 processing a certificate chain. This will occur where a certificate does not
1006 include the Subject Alternative Name extension but where a Certificate
1007 Authority has enforced name constraints. This issue can occur even with valid
1013 * Corrected a few file name and file reference bugs in the build,
1014 installation and setup scripts, which lead to installation verification
1015 failures. Slightly enhanced the installation verification script.
1019 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1024 * Fixed PVK encoder to properly query for the passphrase.
1028 * Multiple fixes in the OSSL_HTTP API functions.
1032 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1033 OSSL_PARAM_INTEGER data type and return error on negative numbers
1034 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1035 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1039 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1043 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1047 * Multiple threading fixes.
1051 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1055 * Allow fetching an operation from the provider that owns an unexportable key
1056 as a fallback if that is still allowed by the property query.
1060 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1062 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1067 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1068 S390X capability vector to zero. This simplifies testing of different code
1069 paths on S390X architecture.
1073 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1074 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1075 SP 800-38D". The communication will fail at this point.
1079 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1080 confidential in EC_GROUP data.
1084 * The byte order mark (BOM) character is ignored if encountered at the
1085 beginning of a PEM-formatted file.
1089 * Added CMS support for the Russian GOST algorithms.
1093 * Due to move of the implementation of cryptographic operations
1094 to the providers, validation of various operation parameters can
1095 be postponed until the actual operation is executed where previously
1096 it happened immediately when an operation parameter was set.
1098 For example when setting an unsupported curve with
1099 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1100 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1102 *OpenSSL team members and many third party contributors*
1104 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1105 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1106 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1107 instead to retrieve these algorithms from a provider.
1111 * On build targets where the multilib postfix is set in the build
1112 configuration the libdir directory was changing based on whether
1113 the lib directory with the multilib postfix exists on the system
1114 or not. This unpredictable behavior was removed and eventual
1115 multilib postfix is now always added to the default libdir. Use
1116 `--libdir=lib` to override the libdir if adding the postfix is
1121 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1122 no longer interoperable with OpenSSL 1.1.1.
1126 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1127 function codes, this function can only cause problems for calling
1132 * Add a configurable flag to output date formats as ISO 8601. Does not
1133 change the default date format.
1137 * Version of MSVC earlier than 1300 could get link warnings, which could
1138 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1139 Support for this flag has been removed.
1143 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1144 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1145 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1146 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1147 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1151 * The signatures of the functions to get and set options on SSL and
1152 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1153 Some source code changes may be required.
1157 * The public definitions of conf_method_st and conf_st have been
1158 deprecated. They will be made opaque in a future release.
1160 *Rich Salz and Tomáš Mráz*
1162 * Client-initiated renegotiation is disabled by default. To allow it, use
1163 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1164 flag, or the "ClientRenegotiation" config parameter as appropriate.
1168 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1169 or modify relative pathname inclusion.
1173 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1174 validated. Please consult the README-FIPS and
1175 README-PROVIDERS files, as well as the migration guide.
1177 *OpenSSL team members and many third party contributors*
1179 * For the key types DH and DHX the allowed settable parameters are now different.
1183 * The openssl commands that read keys, certificates, and CRLs now
1184 automatically detect the PEM or DER format of the input files.
1186 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1188 * Added enhanced PKCS#12 APIs which accept a library context.
1192 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1196 * Added support for Kernel TLS (KTLS).
1198 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1200 * Support for RFC 5746 secure renegotiation is now required by default for
1201 SSL or TLS connections to succeed.
1205 * The signature of the `copy` functional parameter of the
1206 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1207 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1208 the signature of the `pub_decode` functional parameter of the
1209 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1210 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1214 * The error return values from some control calls (ctrl) have changed.
1218 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1222 * Many functions in the EVP_ namespace that are getters of values from
1223 implementations or contexts were renamed to include get or get0 in their
1224 names. Old names are provided as macro aliases for compatibility and
1229 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1230 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1231 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1236 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1239 * The output from the command line applications may have minor
1244 * The output from numerous "printing" may have minor changes.
1248 * Windows thread synchronization uses read/write primitives (SRWLock) when
1249 supported by the OS, otherwise CriticalSection continues to be used.
1253 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1254 work on read only BIO source/sinks that do not support these functions.
1255 This allows piping or redirection of a file BIO using stdin to be buffered
1256 into memory. This is used internally in OSSL_DECODER_from_bio().
1260 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1261 this function would return one of the values OSSL_STORE_INFO_NAME,
1262 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1263 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1264 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1265 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1266 using this function should be amended to handle the changed return value.
1270 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1271 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1272 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1273 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1274 contains more than one certificate identifier: This means that all
1275 certificates referenced there MUST be part of the validation chain.
1279 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1280 RC5, DESX and DES have been moved to the legacy provider.
1284 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1285 RIPEMD-160 have been moved to the legacy provider.
1289 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1294 * The deprecated functions EVP_PKEY_get0_RSA(),
1295 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1296 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1297 well as the similarly named "get1" functions behave differently in
1302 * A number of functions handling low-level keys or engines were deprecated
1303 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1304 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1305 EVP_PKEY_get0_siphash().
1309 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1310 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1311 will need to load the legacy crypto provider. This includes these PBE
1312 algorithms which use this KDF:
1313 - NID_pbeWithMD2AndDES_CBC
1314 - NID_pbeWithMD5AndDES_CBC
1315 - NID_pbeWithSHA1AndRC2_CBC
1316 - NID_pbeWithMD2AndRC2_CBC
1317 - NID_pbeWithMD5AndRC2_CBC
1318 - NID_pbeWithSHA1AndDES_CBC
1322 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1323 BIO_debug_callback() functions.
1327 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1328 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1332 * The RAND_METHOD APIs have been deprecated.
1336 * The SRP APIs have been deprecated.
1340 * Add a compile time option to prevent the caching of provider fetched
1341 algorithms. This is enabled by including the no-cached-fetch option
1342 at configuration time.
1346 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1347 count of PKCS12_DEFAULT_ITER.
1349 *Tomáš Mráz and Sahana Prasad*
1351 * The openssl speed command does not use low-level API calls anymore.
1355 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1358 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1360 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1364 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1365 providers may supply their own group implementations (using either the "key
1366 exchange" or the "key encapsulation" methods) which will automatically be
1367 detected and used by libssl.
1369 *Matt Caswell, Nicola Tuveri*
1371 * The undocumented function X509_certificate_type() has been deprecated;
1375 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1379 * Removed RSA padding mode for SSLv23 (which was only used for
1380 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1381 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1386 * Deprecated the obsolete X9.31 RSA key generation related functions.
1388 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1389 is not allowed to return a value > 1, this is no more taken as failure.
1391 *Viktor Dukhovni and David von Oheimb*
1393 * Deprecated the obsolete X9.31 RSA key generation related functions
1394 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1395 BN_X931_generate_prime_ex().
1399 * The default key generation method for the regular 2-prime RSA keys was
1400 changed to the FIPS 186-4 B.3.6 method.
1404 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1408 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1412 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1413 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1415 *Rich Salz, Richard Levitte, and David von Oheimb*
1417 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1421 * Deprecated `OCSP_parse_url()`.
1425 * Validation of SM2 keys has been separated from the validation of regular EC
1430 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1431 switches: a validation failure triggers an early exit, returning a failure
1432 exit status to the parent process.
1436 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1437 to ignore unknown ciphers.
1441 * The `-cipher-commands` and `-digest-commands` options
1442 of the command line utility `list` have been deprecated.
1443 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1447 * Added convenience functions for generating asymmetric key pairs:
1448 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1449 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1453 * All of the low-level EC_KEY functions have been deprecated.
1455 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1457 * Deprecated all the libcrypto and libssl error string loading
1462 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1463 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1468 * The `-crypt` option to the `passwd` command line tool has been removed.
1472 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1477 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1481 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1482 EVP_PKEY_get1_tls_encodedpoint().
1486 * The security callback, which can be customised by application code, supports
1487 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1488 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1492 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1493 interface. Their functionality remains unchanged.
1497 * Added new option for 'openssl list', '-providers', which will display the
1498 list of loaded providers, their names, version and status. It optionally
1499 displays their gettable parameters.
1503 * Removed EVP_PKEY_set_alias_type().
1507 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1508 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1512 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1513 parameters are still checked for type safety at compile time via helper
1518 * Remove the RAND_DRBG API
1520 *Paul Dale and Matthias St. Pierre*
1522 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1523 as well as actual hostnames.
1527 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1528 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1529 conversely, silently ignore DTLS protocol version bounds when configuring
1530 TLS-based contexts. The commands can be repeated to set bounds of both
1531 types. The same applies with the corresponding "min_protocol" and
1532 "max_protocol" command-line switches, in case some application uses both TLS
1535 SSL_CTX instances that are created for a fixed protocol version (e.g.
1536 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1537 attempts to apply bounds to these protocol versions would result in an
1538 error. Now only the "version-flexible" SSL_CTX instances are subject to
1539 limits in configuration files in command-line options.
1543 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1548 * Reworked the recorded ERR codes to make better space for system errors.
1549 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1550 given code is a system error (true) or an OpenSSL error (false).
1554 * Reworked the test perl framework to better allow parallel testing.
1556 *Nicola Tuveri and David von Oheimb*
1558 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1559 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1563 * 'Configure' has been changed to figure out the configuration target if
1564 none is given on the command line. Consequently, the 'config' script is
1565 now only a mere wrapper. All documentation is changed to only mention
1568 *Rich Salz and Richard Levitte*
1570 * Added a library context `OSSL_LIB_CTX` that applications as well as
1571 other libraries can use to form a separate context within which
1572 libcrypto operations are performed.
1576 * Added various `_ex` functions to the OpenSSL API that support using
1577 a non-default `OSSL_LIB_CTX`.
1581 * Handshake now fails if Extended Master Secret extension is dropped
1586 * Dropped interactive mode from the `openssl` program.
1590 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1592 *David von Oheimb and Shane Lontis*
1594 * Deprecated `EC_METHOD_get_field_type()`.
1598 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1599 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1600 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1604 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1608 * Add CAdES-BES signature verification support, mostly derived
1609 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1611 *Filipe Raimundo da Silva*
1613 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1617 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1618 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1622 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1626 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1627 EC_KEY_precompute_mult().
1631 * Deprecated EC_POINTs_mul().
1635 * Removed FIPS_mode() and FIPS_mode_set().
1639 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1643 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1644 EC_POINT_get_Jprojective_coordinates_GFp().
1648 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1649 arrays to be more easily constructed via a series of utility functions.
1650 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1651 the various push functions and finally convert to a passable OSSL_PARAM
1652 array using OSSL_PARAM_BLD_to_param().
1656 * The security strength of SHA1 and MD5 based signatures in TLS has been
1661 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1662 contain a provider side internal key.
1666 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1670 * Project text documents not yet having a proper file name extension
1671 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1672 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1676 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1677 have been converted to Markdown with the goal to produce documents
1678 which not only look pretty when viewed online in the browser, but
1679 remain well readable inside a plain text editor.
1681 To achieve this goal, a 'minimalistic' Markdown style has been applied
1682 which avoids formatting elements that interfere too much with the
1683 reading flow in the text file. For example, it
1685 * avoids [ATX headings][] and uses [setext headings][] instead
1686 (which works for `<h1>` and `<h2>` headings only).
1687 * avoids [inline links][] and uses [reference links][] instead.
1688 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1690 [ATX headings]: https://github.github.com/gfm/#atx-headings
1691 [setext headings]: https://github.github.com/gfm/#setext-headings
1692 [inline links]: https://github.github.com/gfm/#inline-link
1693 [reference links]: https://github.github.com/gfm/#reference-link
1694 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1695 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1697 *Matthias St. Pierre*
1699 * The test suite is changed to preserve results of each test recipe.
1700 A new directory test-runs/ with subdirectories named like the
1701 test recipes are created in the build tree for this purpose.
1705 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1706 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1707 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1709 *David von Oheimb, Martin Peylo*
1711 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1712 It supports arbitrary request and response content types, GET redirection,
1713 TLS, connections via HTTP(S) proxies, connections and exchange via
1714 user-defined BIOs (allowing implicit connections), persistent connections,
1715 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1716 The legacy OCSP-focused (and only partly documented) API
1717 is retained for backward compatibility, while most of it is deprecated.
1721 * Added `util/check-format.pl`, a tool for checking adherence to the
1722 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1723 The checks performed are incomplete and yield some false positives.
1724 Still the tool should be useful for detecting most typical glitches.
1728 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1729 If domain name resolution yields multiple IP addresses all of them are tried
1730 after `connect()` failures.
1734 * All of the low-level RSA functions have been deprecated.
1738 * X509 certificates signed using SHA1 are no longer allowed at security
1743 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1744 modified to use PKEY APIs. These commands are now in maintenance mode
1745 and no new features will be added to them.
1749 * The command line utility rsautl has been deprecated.
1753 * The command line utilities genrsa and rsa have been modified to use PKEY
1754 APIs. They now write PKCS#8 keys by default. These commands are now in
1755 maintenance mode and no new features will be added to them.
1759 * All of the low-level DH functions have been deprecated.
1761 *Paul Dale and Matt Caswell*
1763 * All of the low-level DSA functions have been deprecated.
1767 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1768 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1772 * Deprecated low-level ECDH and ECDSA functions.
1776 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1780 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1781 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1782 a new formulation to include all the things it can be used for,
1783 as well as words of caution.
1787 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1791 * All of the low-level HMAC functions have been deprecated.
1793 *Paul Dale and David von Oheimb*
1795 * Over two thousand fixes were made to the documentation, including:
1796 - Common options (such as -rand/-writerand, TLS version control, etc)
1797 were refactored and point to newly-enhanced descriptions in openssl.pod.
1798 - Added style conformance for all options (with help from Richard Levitte),
1799 documented all reported missing options, added a CI build to check
1800 that all options are documented and that no unimplemented options
1802 - Documented some internals, such as all use of environment variables.
1803 - Addressed all internal broken L<> references.
1807 * All of the low-level CMAC functions have been deprecated.
1811 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1812 functions have been deprecated.
1814 *Paul Dale and David von Oheimb*
1816 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1817 set of functions. The documentation mentioned negative values for some
1818 errors, but this was never the case, so the mention of negative values
1821 Code that followed the documentation and thereby check with something
1822 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1826 * All of the low-level cipher functions have been deprecated.
1828 *Matt Caswell and Paul Dale*
1830 * Removed include/openssl/opensslconf.h.in and replaced it with
1831 include/openssl/configuration.h.in, which differs in not including
1832 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1833 was added to include both.
1835 This allows internal hacks where one might need to modify the set
1836 of configured macros, for example this if deprecated symbols are
1837 still supposed to be available internally:
1839 #include <openssl/configuration.h>
1841 #undef OPENSSL_NO_DEPRECATED
1842 #define OPENSSL_SUPPRESS_DEPRECATED
1844 #include <openssl/macros.h>
1846 This should not be used by applications that use the exported
1847 symbols, as that will lead to linking errors.
1851 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1852 used in exponentiation with 512-bit moduli. No EC algorithms are
1853 affected. Analysis suggests that attacks against 2-prime RSA1024,
1854 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1855 difficult to perform and are not believed likely. Attacks against DH512
1856 are considered just feasible. However, for an attack the target would
1857 have to re-use the DH512 private key, which is not recommended anyway.
1858 Also applications directly using the low-level API BN_mod_exp may be
1859 affected if they use BN_FLG_CONSTTIME.
1864 * Most memory-debug features have been deprecated, and the functionality
1865 replaced with no-ops.
1869 * Added documentation for the STACK API.
1873 * Introduced a new method type and API, OSSL_ENCODER, to represent
1874 generic encoders. These do the same sort of job that PEM writers
1875 and d2i functions do, but with support for methods supplied by
1876 providers, and the possibility for providers to support other
1881 * Introduced a new method type and API, OSSL_DECODER, to represent
1882 generic decoders. These do the same sort of job that PEM readers
1883 and i2d functions do, but with support for methods supplied by
1884 providers, and the possibility for providers to support other
1889 * Added a .pragma directive to the syntax of configuration files, to
1890 allow varying behavior in a supported and predictable manner.
1891 Currently added pragma:
1895 This allows dollar signs to be a keyword character unless it's
1896 followed by a opening brace or parenthesis. This is useful for
1897 platforms where dollar signs are commonly used in names, such as
1898 volume names and system directory names on VMS.
1902 * Added functionality to create an EVP_PKEY from user data.
1906 * Change the interpretation of the '--api' configuration option to
1907 mean that this is a desired API compatibility level with no
1908 further meaning. The previous interpretation, that this would
1909 also mean to remove all deprecated symbols up to and including
1910 the given version, no requires that 'no-deprecated' is also used
1911 in the configuration.
1913 When building applications, the desired API compatibility level
1914 can be set with the OPENSSL_API_COMPAT macro like before. For
1915 API compatibility version below 3.0, the old style numerical
1916 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1917 For version 3.0 and on, the value is expected to be the decimal
1918 value calculated from the major and minor version like this:
1920 MAJOR * 10000 + MINOR * 100
1924 -DOPENSSL_API_COMPAT=30000 For 3.0
1925 -DOPENSSL_API_COMPAT=30200 For 3.2
1927 To hide declarations that are deprecated up to and including the
1928 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1929 given when building the application as well.
1933 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1934 access to certificate and CRL stores via URIs and OSSL_STORE
1937 This adds the following functions:
1939 - X509_LOOKUP_store()
1940 - X509_STORE_load_file()
1941 - X509_STORE_load_path()
1942 - X509_STORE_load_store()
1943 - SSL_add_store_cert_subjects_to_stack()
1944 - SSL_CTX_set_default_verify_store()
1945 - SSL_CTX_load_verify_file()
1946 - SSL_CTX_load_verify_dir()
1947 - SSL_CTX_load_verify_store()
1951 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1952 The presence of this system service is determined at run-time.
1956 * Added functionality to create an EVP_PKEY context based on data
1957 for methods from providers. This takes an algorithm name and a
1958 property query string and simply stores them, with the intent
1959 that any operation that uses this context will use those strings
1960 to fetch the needed methods implicitly, thereby making the port
1961 of application written for pre-3.0 OpenSSL easier.
1965 * The undocumented function NCONF_WIN32() has been deprecated; for
1966 conversion details see the HISTORY section of doc/man5/config.pod
1970 * Introduced the new functions EVP_DigestSignInit_ex() and
1971 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1972 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1973 pages for further details.
1977 * Over two thousand fixes were made to the documentation, including:
1978 adding missing command flags, better style conformance, documentation
1981 *Rich Salz, Richard Levitte*
1983 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1984 X25519, X448, Ed25519 and Ed448.
1988 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1993 * Deprecated the public definition of `ERR_STATE` as well as the function
1994 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1999 * Added ERR functionality to give callers access to the stored function
2000 names that have replaced the older function code based functions.
2002 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2003 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2004 ERR_peek_error_all() and ERR_peek_last_error_all().
2006 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2007 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2008 ERR_func_error_string().
2012 * Extended testing to be verbose for failing tests only. The make variables
2013 VERBOSE_FAILURE or VF can be used to enable this:
2015 $ make VF=1 test # Unix
2016 $ mms /macro=(VF=1) test ! OpenVMS
2017 $ nmake VF=1 test # Windows
2021 * Added the `-copy_extensions` option to the `x509` command for use with
2022 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2023 all extensions in the request are copied to the certificate or vice versa.
2025 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2027 * Added the `-copy_extensions` option to the `req` command for use with
2028 `-x509`. When given with the `copy` or `copyall` argument,
2029 all extensions in the certification request are copied to the certificate.
2033 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2034 they generate are by default RFC 5280 compliant in the following sense:
2035 There is a subjectKeyIdentifier extension with a hash value of the public key
2036 and for not self-signed certs there is an authorityKeyIdentifier extension
2037 with a keyIdentifier field or issuer information identifying the signing key.
2038 This is done unless some configuration overrides the new default behavior,
2039 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2043 * Added several checks to `X509_verify_cert()` according to requirements in
2044 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2045 (which may be done by using the CLI option `-x509_strict`):
2046 * The basicConstraints of CA certificates must be marked critical.
2047 * CA certificates must explicitly include the keyUsage extension.
2048 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2049 * The issuer name of any certificate must not be empty.
2050 * The subject name of CA certs, certs with keyUsage crlSign,
2051 and certs without subjectAlternativeName must not be empty.
2052 * If a subjectAlternativeName extension is given it must not be empty.
2053 * The signatureAlgorithm field and the cert signature must be consistent.
2054 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2055 must not be marked critical.
2056 * The authorityKeyIdentifier must be given for X.509v3 certs
2057 unless they are self-signed.
2058 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2062 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2063 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2067 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2068 used even when parsing explicit parameters, when loading a encoded key
2069 or calling `EC_GROUP_new_from_ecpkparameters()`/
2070 `EC_GROUP_new_from_ecparameters()`.
2071 This prevents bypass of security hardening and performance gains,
2072 especially for curves with specialized EC_METHODs.
2073 By default, if a key encoded with explicit parameters is loaded and later
2074 encoded, the output is still encoded with explicit parameters, even if
2075 internally a "named" EC_GROUP is used for computation.
2079 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2080 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2081 NULL. After this change, only the cofactor parameter can be NULL. It also
2082 does some minimal sanity checks on the passed order.
2087 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2088 An attack is simple, if the first CMS_recipientInfo is valid but the
2089 second CMS_recipientInfo is chosen ciphertext. If the second
2090 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2091 encryption key will be replaced by garbage, and the message cannot be
2092 decoded, but if the RSA decryption fails, the correct encryption key is
2093 used and the recipient will not notice the attack.
2094 As a work around for this potential attack the length of the decrypted
2095 key must be equal to the cipher default key length, in case the
2096 certificate is not given and all recipientInfo are tried out.
2097 The old behaviour can be re-enabled in the CMS code by setting the
2098 CMS_DEBUG_DECRYPT flag.
2102 * Early start up entropy quality from the DEVRANDOM seed source has been
2103 improved for older Linux systems. The RAND subsystem will wait for
2104 /dev/random to be producing output before seeding from /dev/urandom.
2105 The seeded state is stored for future library initialisations using
2106 a system global shared memory segment. The shared memory identifier
2107 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2108 the desired value. The default identifier is 114.
2112 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2113 when primes for RSA keys are computed.
2114 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2115 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2116 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2117 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2118 This avoids possible fingerprinting of newly generated RSA modules.
2122 * Correct the extended master secret constant on EBCDIC systems. Without this
2123 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2124 negotiate EMS will fail. Unfortunately this also means that TLS connections
2125 between EBCDIC systems with this fix, and EBCDIC systems without this
2126 fix will fail if they negotiate EMS.
2130 * Changed the library initialisation so that the config file is now loaded
2131 by default. This was already the case for libssl. It now occurs for both
2132 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2133 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2137 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2138 where the former acts as a replacement for `ERR_put_error()`, and the
2139 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2140 `ERR_raise_data()` adds more flexibility by taking a format string and
2141 an arbitrary number of arguments following it, to be processed with
2146 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2147 to check if a named provider is loaded and available. When called, it
2148 will also activate all fallback providers if such are still present.
2152 * Enforce a minimum DH modulus size of 512 bits.
2156 * Changed DH parameters to generate the order q subgroup instead of 2q.
2157 Previously generated DH parameters are still accepted by DH_check
2158 but DH_generate_key works around that by clearing bit 0 of the
2159 private key for those. This avoids leaking bit 0 of the private key.
2163 * Significantly reduce secure memory usage by the randomness pools.
2167 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2172 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2173 algorithms. An implementation of a key exchange algorithm can be obtained
2174 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2175 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2176 the older EVP_PKEY_derive_init() function. See the man pages for the new
2177 functions for further details.
2181 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2185 * Removed the function names from error messages and deprecated the
2190 * Removed NextStep support and the macro OPENSSL_UNISTD
2194 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2195 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2196 Also removed "export var as function" capability; we do not export
2197 variables, only functions.
2201 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2202 an error and 1 indicating success. In previous versions of OpenSSL this
2203 was a void type. If a key was set longer than the maximum possible this
2208 * Support SM2 signing and verification schemes with X509 certificate.
2212 * Use SHA256 as the default digest for TS query in the `ts` app.
2216 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2220 * Default cipher lists/suites are now available via a function, the
2221 #defines are deprecated.
2225 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2226 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2227 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2231 * Join the directories crypto/x509 and crypto/x509v3
2235 * Added command 'openssl kdf' that uses the EVP_KDF API.
2239 * Added command 'openssl mac' that uses the EVP_MAC API.
2243 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2244 as default directories. Also added the command 'openssl info'
2245 for scripting purposes.
2249 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2254 * Add prediction resistance to the DRBG reseeding process.
2258 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2259 mandated by IEEE Std 1619-2018.
2263 * Added newline escaping functionality to a filename when using openssl dgst.
2264 This output format is to replicate the output format found in the `*sum`
2265 checksum programs. This aims to preserve backward compatibility.
2267 *Matt Eaton, Richard Levitte, and Paul Dale*
2269 * Removed the heartbeat message in DTLS feature, as it has very
2270 little usage and doesn't seem to fulfill a valuable purpose.
2271 The configuration option is now deprecated.
2275 * Changed the output of 'openssl {digestname} < file' to display the
2276 digest name in its output.
2280 * Added a new generic trace API which provides support for enabling
2281 instrumentation through trace output.
2283 *Richard Levitte & Matthias St. Pierre*
2285 * Added build tests for C++. These are generated files that only do one
2286 thing, to include one public OpenSSL head file each. This tests that
2287 the public header files can be usefully included in a C++ application.
2289 This test isn't enabled by default. It can be enabled with the option
2290 'enable-buildtest-c++'.
2294 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2298 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2302 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2306 * Added KMAC to EVP_MAC.
2310 * Added property based algorithm implementation selection framework to
2315 * Added SCA hardening for modular field inversion in EC_GROUP through
2316 a new dedicated field_inv() pointer in EC_METHOD.
2317 This also addresses a leakage affecting conversions from projective
2318 to affine coordinates.
2320 *Billy Bob Brumley, Nicola Tuveri*
2322 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2323 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2324 those algorithms that were already supported through the EVP_PKEY API
2325 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2326 and scrypt are now wrappers that call EVP_KDF.
2330 * Build devcrypto engine as a dynamic engine.
2332 *Eneas U de Queiroz*
2334 * Add keyed BLAKE2 to EVP_MAC.
2338 * Fix a bug in the computation of the endpoint-pair shared secret used
2339 by DTLS over SCTP. This breaks interoperability with older versions
2340 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2341 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2342 interoperability with such broken implementations. However, enabling
2343 this switch breaks interoperability with correct implementations.
2345 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2346 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2350 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2354 * Changed the license to the Apache License v2.0.
2358 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2360 - Major releases (indicated by incrementing the MAJOR release number)
2361 may introduce incompatible API/ABI changes.
2362 - Minor releases (indicated by incrementing the MINOR release number)
2363 may introduce new features but retain API/ABI compatibility.
2364 - Patch releases (indicated by incrementing the PATCH number)
2365 are intended for bug fixes and other improvements of existing
2366 features only (like improving performance or adding documentation)
2367 and retain API/ABI compatibility.
2371 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2375 * Remove the 'dist' target and add a tarball building script. The
2376 'dist' target has fallen out of use, and it shouldn't be
2377 necessary to configure just to create a source distribution.
2381 * Recreate the OS390-Unix config target. It no longer relies on a
2382 special script like it did for OpenSSL pre-1.1.0.
2386 * Instead of having the source directories listed in Configure, add
2387 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2392 * Add GMAC to EVP_MAC.
2396 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2400 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2401 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2402 to facilitate the continued use of MACs through raw private keys in
2403 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2407 * Deprecate ECDH_KDF_X9_62().
2411 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2412 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2413 are retained for backwards compatibility.
2417 * AES-XTS mode now enforces that its two keys are different to mitigate
2418 the attacked described in "Efficient Instantiations of Tweakable
2419 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2420 Details of this attack can be obtained from:
2421 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2425 * Rename the object files, i.e. give them other names than in previous
2426 versions. Their names now include the name of the final product, as
2427 well as its type mnemonic (bin, lib, shlib).
2431 * Added new option for 'openssl list', '-objects', which will display the
2432 list of built in objects, i.e. OIDs with names.
2436 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2437 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2442 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2443 improves application performance by removing data copies and providing
2444 applications with zero-copy system calls such as sendfile and splice.
2448 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2452 * `PKCS12_parse` now maintains the order of the parsed certificates
2453 when outputting them via `*ca` (rather than reversing it).
2457 * Deprecated pthread fork support methods.
2461 * Added support for FFDHE key exchange in TLS 1.3.
2465 * Added a new concept for OpenSSL plugability: providers. This
2466 functionality is designed to replace the ENGINE API and ENGINE
2467 implementations, and to be much more dynamic, allowing provider
2468 authors to introduce new algorithms among other things, as long as
2469 there's an API that supports the algorithm type.
2471 With this concept comes a new core API for interaction between
2472 libcrypto and provider implementations. Public libcrypto functions
2473 that want to use providers do so through this core API.
2475 The main documentation for this core API is found in
2476 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2477 refer to other manuals describing the API specific for supported
2478 algorithm types (also called operations).
2485 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2487 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2489 * Avoid loading of a dynamic engine twice.
2493 * Prioritise DANE TLSA issuer certs over peer certs
2497 * Fixed random API for MacOS prior to 10.12
2499 These MacOS versions don't support the CommonCrypto APIs
2503 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2505 * Fixed an SM2 Decryption Buffer Overflow.
2507 In order to decrypt SM2 encrypted data an application is expected to
2508 call the API function EVP_PKEY_decrypt(). Typically an application will
2509 call this function twice. The first time, on entry, the "out" parameter
2510 can be NULL and, on exit, the "outlen" parameter is populated with the
2511 buffer size required to hold the decrypted plaintext. The application
2512 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2513 again, but this time passing a non-NULL value for the "out" parameter.
2515 A bug in the implementation of the SM2 decryption code means that the
2516 calculation of the buffer size required to hold the plaintext returned
2517 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2518 size required by the second call. This can lead to a buffer overflow
2519 when EVP_PKEY_decrypt() is called by the application a second time with
2520 a buffer that is too small.
2522 A malicious attacker who is able present SM2 content for decryption to
2523 an application could cause attacker chosen data to overflow the buffer
2524 by up to a maximum of 62 bytes altering the contents of other data held
2525 after the buffer, possibly changing application behaviour or causing
2526 the application to crash. The location of the buffer is application
2527 dependent but is typically heap allocated.
2532 * Fixed various read buffer overruns processing ASN.1 strings
2534 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2535 structure which contains a buffer holding the string data and a field
2536 holding the buffer length. This contrasts with normal C strings which
2537 are represented as a buffer for the string data which is terminated
2538 with a NUL (0) byte.
2540 Although not a strict requirement, ASN.1 strings that are parsed using
2541 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2542 well as any string whose value has been set with the ASN1_STRING_set()
2543 function will additionally NUL terminate the byte array in the
2544 ASN1_STRING structure.
2546 However, it is possible for applications to directly construct valid
2547 ASN1_STRING structures which do not NUL terminate the byte array by
2548 directly setting the "data" and "length" fields in the ASN1_STRING
2549 array. This can also happen by using the ASN1_STRING_set0() function.
2551 Numerous OpenSSL functions that print ASN.1 data have been found to
2552 assume that the ASN1_STRING byte array will be NUL terminated, even
2553 though this is not guaranteed for strings that have been directly
2554 constructed. Where an application requests an ASN.1 structure to be
2555 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2556 been directly constructed by the application without NUL terminating
2557 the "data" field, then a read buffer overrun can occur.
2559 The same thing can also occur during name constraints processing
2560 of certificates (for example if a certificate has been directly
2561 constructed by the application instead of loading it via the OpenSSL
2562 parsing functions, and the certificate contains non NUL terminated
2563 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2564 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2566 If a malicious actor can cause an application to directly construct an
2567 ASN1_STRING and then process it through one of the affected OpenSSL
2568 functions then this issue could be hit. This might result in a crash
2569 (causing a Denial of Service attack). It could also result in the
2570 disclosure of private memory contents (such as private keys, or
2571 sensitive plaintext).
2576 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2578 * Fixed a problem with verifying a certificate chain when using the
2579 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2580 the certificates present in a certificate chain. It is not set by default.
2582 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2583 the chain that have explicitly encoded elliptic curve parameters was added
2584 as an additional strict check.
2586 An error in the implementation of this check meant that the result of a
2587 previous check to confirm that certificates in the chain are valid CA
2588 certificates was overwritten. This effectively bypasses the check
2589 that non-CA certificates must not be able to issue other certificates.
2591 If a "purpose" has been configured then there is a subsequent opportunity
2592 for checks that the certificate is a valid CA. All of the named "purpose"
2593 values implemented in libcrypto perform this check. Therefore, where
2594 a purpose is set the certificate chain will still be rejected even when the
2595 strict flag has been used. A purpose is set by default in libssl client and
2596 server certificate verification routines, but it can be overridden or
2597 removed by an application.
2599 In order to be affected, an application must explicitly set the
2600 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2601 for the certificate verification or, in the case of TLS client or server
2602 applications, override the default purpose.
2607 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2608 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2609 renegotiation ClientHello omits the signature_algorithms extension (where it
2610 was present in the initial ClientHello), but includes a
2611 signature_algorithms_cert extension then a NULL pointer dereference will
2612 result, leading to a crash and a denial of service attack.
2614 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2615 (which is the default configuration). OpenSSL TLS clients are not impacted by
2619 *Peter Kästle and Samuel Sapalski*
2621 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2623 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2624 create a unique hash value based on the issuer and serial number data
2625 contained within an X509 certificate. However, it was failing to correctly
2626 handle any errors that may occur while parsing the issuer field (which might
2627 occur if the issuer field is maliciously constructed). This may subsequently
2628 result in a NULL pointer deref and a crash leading to a potential denial of
2634 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2635 padding mode to correctly check for rollback attacks. This is considered a
2636 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2641 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2642 functions. Previously they could overflow the output length argument in some
2643 cases where the input length is close to the maximum permissible length for
2644 an integer on the platform. In such cases the return value from the function
2645 call would be 1 (indicating success), but the output length value would be
2646 negative. This could cause applications to behave incorrectly or crash.
2651 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2652 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2653 could be exploited in a side channel attack to recover the password. Since
2654 the attack is local host only this is outside of the current OpenSSL
2655 threat model and therefore no CVE is assigned.
2657 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2662 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2664 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2665 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2666 If an attacker can control both items being compared then this could lead
2667 to a possible denial of service attack. OpenSSL itself uses the
2668 GENERAL_NAME_cmp function for two purposes:
2669 1) Comparing CRL distribution point names between an available CRL and a
2670 CRL distribution point embedded in an X509 certificate
2671 2) When verifying that a timestamp response token signer matches the
2672 timestamp authority name (exposed via the API functions
2673 TS_RESP_verify_response and TS_RESP_verify_token)
2678 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2680 * Certificates with explicit curve parameters are now disallowed in
2681 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2685 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2686 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2687 conversely, silently ignore DTLS protocol version bounds when configuring
2688 TLS-based contexts. The commands can be repeated to set bounds of both
2689 types. The same applies with the corresponding "min_protocol" and
2690 "max_protocol" command-line switches, in case some application uses both TLS
2693 SSL_CTX instances that are created for a fixed protocol version (e.g.
2694 TLSv1_server_method()) also silently ignore version bounds. Previously
2695 attempts to apply bounds to these protocol versions would result in an
2696 error. Now only the "version-flexible" SSL_CTX instances are subject to
2697 limits in configuration files in command-line options.
2701 * Handshake now fails if Extended Master Secret extension is dropped
2706 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2708 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2710 * Fixed segmentation fault in SSL_check_chain()
2711 Server or client applications that call the SSL_check_chain() function
2712 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2713 dereference as a result of incorrect handling of the
2714 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2715 or unrecognised signature algorithm is received from the peer. This could
2716 be exploited by a malicious peer in a Denial of Service attack.
2721 * Added AES consttime code for no-asm configurations
2722 an optional constant time support for AES was added
2723 when building openssl for no-asm.
2724 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2725 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2726 At this time this feature is by default disabled.
2727 It will be enabled by default in 3.0.
2731 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2733 * Revert the change of EOF detection while reading in libssl to avoid
2734 regressions in applications depending on the current way of reporting
2735 the EOF. As the existing method is not fully accurate the change to
2736 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2737 branch and will be present in the 3.0 release.
2741 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2742 when primes for RSA keys are computed.
2743 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2744 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2745 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2746 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2747 This avoids possible fingerprinting of newly generated RSA modules.
2751 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2753 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2754 while reading in libssl then we would report an error back to the
2755 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2756 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2757 therefore give a hint as to what went wrong.
2761 * Check that ed25519 and ed448 are allowed by the security level. Previously
2762 signature algorithms not using an MD were not being checked that they were
2763 allowed by the security level.
2767 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2768 was not quite right. The behaviour was not consistent between resumption
2769 and normal handshakes, and also not quite consistent with historical
2770 behaviour. The behaviour in various scenarios has been clarified and
2771 it has been updated to make it match historical behaviour as closely as
2776 * *[VMS only]* The header files that the VMS compilers include automatically,
2777 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2778 that the C++ compiler doesn't understand. This is a shortcoming in the
2779 compiler, but can be worked around with `__cplusplus` guards.
2781 C++ applications that use OpenSSL libraries must be compiled using the
2782 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2783 functions. Otherwise, only functions with symbols of less than 31
2784 characters can be used, as the linker will not be able to successfully
2785 resolve symbols with longer names.
2789 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2790 The presence of this system service is determined at run-time.
2794 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2799 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2801 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2802 number generator (RNG). This was intended to include protection in the
2803 event of a fork() system call in order to ensure that the parent and child
2804 processes did not share the same RNG state. However, this protection was not
2805 being used in the default case.
2807 A partial mitigation for this issue is that the output from a high
2808 precision timer is mixed into the RNG state so the likelihood of a parent
2809 and child process sharing state is significantly reduced.
2811 If an application already calls OPENSSL_init_crypto() explicitly using
2812 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2815 *Matthias St. Pierre*
2817 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2818 used even when parsing explicit parameters, when loading a encoded key
2819 or calling `EC_GROUP_new_from_ecpkparameters()`/
2820 `EC_GROUP_new_from_ecparameters()`.
2821 This prevents bypass of security hardening and performance gains,
2822 especially for curves with specialized EC_METHODs.
2823 By default, if a key encoded with explicit parameters is loaded and later
2824 encoded, the output is still encoded with explicit parameters, even if
2825 internally a "named" EC_GROUP is used for computation.
2829 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2830 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2831 NULL. After this change, only the cofactor parameter can be NULL. It also
2832 does some minimal sanity checks on the passed order.
2837 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2838 An attack is simple, if the first CMS_recipientInfo is valid but the
2839 second CMS_recipientInfo is chosen ciphertext. If the second
2840 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2841 encryption key will be replaced by garbage, and the message cannot be
2842 decoded, but if the RSA decryption fails, the correct encryption key is
2843 used and the recipient will not notice the attack.
2844 As a work around for this potential attack the length of the decrypted
2845 key must be equal to the cipher default key length, in case the
2846 certificate is not given and all recipientInfo are tried out.
2847 The old behaviour can be re-enabled in the CMS code by setting the
2848 CMS_DEBUG_DECRYPT flag.
2853 * Early start up entropy quality from the DEVRANDOM seed source has been
2854 improved for older Linux systems. The RAND subsystem will wait for
2855 /dev/random to be producing output before seeding from /dev/urandom.
2856 The seeded state is stored for future library initialisations using
2857 a system global shared memory segment. The shared memory identifier
2858 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2859 the desired value. The default identifier is 114.
2863 * Correct the extended master secret constant on EBCDIC systems. Without this
2864 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2865 negotiate EMS will fail. Unfortunately this also means that TLS connections
2866 between EBCDIC systems with this fix, and EBCDIC systems without this
2867 fix will fail if they negotiate EMS.
2871 * Use Windows installation paths in the mingw builds
2873 Mingw isn't a POSIX environment per se, which means that Windows
2874 paths should be used for installation.
2879 * Changed DH_check to accept parameters with order q and 2q subgroups.
2880 With order 2q subgroups the bit 0 of the private key is not secret
2881 but DH_generate_key works around that by clearing bit 0 of the
2882 private key for those. This avoids leaking bit 0 of the private key.
2886 * Significantly reduce secure memory usage by the randomness pools.
2890 * Revert the DEVRANDOM_WAIT feature for Linux systems
2892 The DEVRANDOM_WAIT feature added a select() call to wait for the
2893 /dev/random device to become readable before reading from the
2894 /dev/urandom device.
2896 It turned out that this change had negative side effects on
2897 performance which were not acceptable. After some discussion it
2898 was decided to revert this feature and leave it up to the OS
2899 resp. the platform maintainer to ensure a proper initialization
2900 during early boot time.
2902 *Matthias St. Pierre*
2904 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2906 * Add build tests for C++. These are generated files that only do one
2907 thing, to include one public OpenSSL head file each. This tests that
2908 the public header files can be usefully included in a C++ application.
2910 This test isn't enabled by default. It can be enabled with the option
2911 'enable-buildtest-c++'.
2915 * Enable SHA3 pre-hashing for ECDSA and DSA.
2919 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2920 This changes the size when using the `genpkey` command when no size is given.
2921 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2922 generation commands to use 2048 bits by default.
2926 * Reorganize the manual pages to consistently have RETURN VALUES,
2927 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2928 util/fix-doc-nits accordingly.
2930 *Paul Yang, Joshua Lock*
2932 * Add the missing accessor EVP_PKEY_get0_engine()
2936 * Have commands like `s_client` and `s_server` output the signature scheme
2937 along with other cipher suite parameters when debugging.
2941 * Make OPENSSL_config() error agnostic again.
2945 * Do the error handling in RSA decryption constant time.
2949 * Prevent over long nonces in ChaCha20-Poly1305.
2951 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2952 for every encryption operation. RFC 7539 specifies that the nonce value
2953 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2954 and front pads the nonce with 0 bytes if it is less than 12
2955 bytes. However it also incorrectly allows a nonce to be set of up to 16
2956 bytes. In this case only the last 12 bytes are significant and any
2957 additional leading bytes are ignored.
2959 It is a requirement of using this cipher that nonce values are
2960 unique. Messages encrypted using a reused nonce value are susceptible to
2961 serious confidentiality and integrity attacks. If an application changes
2962 the default nonce length to be longer than 12 bytes and then makes a
2963 change to the leading bytes of the nonce expecting the new value to be a
2964 new unique nonce then such an application could inadvertently encrypt
2965 messages with a reused nonce.
2967 Additionally the ignored bytes in a long nonce are not covered by the
2968 integrity guarantee of this cipher. Any application that relies on the
2969 integrity of these ignored leading bytes of a long nonce may be further
2970 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2971 is safe because no such use sets such a long nonce value. However user
2972 applications that use this cipher directly and set a non-default nonce
2973 length to be longer than 12 bytes may be vulnerable.
2975 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2981 * Add DEVRANDOM_WAIT feature for Linux systems
2983 On older Linux systems where the getrandom() system call is not available,
2984 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2985 Contrary to getrandom(), the /dev/urandom device will not block during
2986 early boot when the kernel CSPRNG has not been seeded yet.
2988 To mitigate this known weakness, use select() to wait for /dev/random to
2989 become readable before reading from /dev/urandom.
2991 * Ensure that SM2 only uses SM3 as digest algorithm
2995 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2997 * Change the info callback signals for the start and end of a post-handshake
2998 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2999 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3000 confused by this and assume that a TLSv1.2 renegotiation has started. This
3001 can break KeyUpdate handling. Instead we no longer signal the start and end
3002 of a post handshake message exchange (although the messages themselves are
3003 still signalled). This could break some applications that were expecting
3004 the old signals. However without this KeyUpdate is not usable for many
3009 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3011 * Timing vulnerability in DSA signature generation
3013 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3014 timing side channel attack. An attacker could use variations in the signing
3015 algorithm to recover the private key.
3017 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3022 * Timing vulnerability in ECDSA signature generation
3024 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3025 timing side channel attack. An attacker could use variations in the signing
3026 algorithm to recover the private key.
3028 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3033 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3034 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3035 of two gigabytes and the error handling improved.
3037 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3038 categorized as a normal bug, not a security issue, because the DRBG reseeds
3039 automatically and is fully functional even without additional randomness
3040 provided by the application.
3042 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3044 * Add a new ClientHello callback. Provides a callback interface that gives
3045 the application the ability to adjust the nascent SSL object at the
3046 earliest stage of ClientHello processing, immediately after extensions have
3047 been collected but before they have been processed. In particular, this
3048 callback can adjust the supported TLS versions in response to the contents
3053 * Add SM2 base algorithm support.
3057 * s390x assembly pack: add (improved) hardware-support for the following
3058 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3059 aes-cfb/cfb8, aes-ecb.
3063 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3064 parameter is no longer accepted, as it leads to a corrupt table. NULL
3065 pem_str is reserved for alias entries only.
3069 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3070 step for prime curves. The new implementation is based on formulae from
3071 differential addition-and-doubling in homogeneous projective coordinates
3072 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3073 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3074 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3075 to work in projective coordinates.
3077 *Billy Bob Brumley, Nicola Tuveri*
3079 * Change generating and checking of primes so that the error rate of not
3080 being prime depends on the intended use based on the size of the input.
3081 For larger primes this will result in more rounds of Miller-Rabin.
3082 The maximal error rate for primes with more than 1080 bits is lowered
3085 *Kurt Roeckx, Annie Yousar*
3087 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3091 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3092 moving between systems, and to avoid confusion when a Windows build is
3093 done with mingw vs with MSVC. For POSIX installs, there's still a
3094 symlink or copy named 'tsget' to avoid that confusion as well.
3098 * Revert blinding in ECDSA sign and instead make problematic addition
3099 length-invariant. Switch even to fixed-length Montgomery multiplication.
3103 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3104 step for binary curves. The new implementation is based on formulae from
3105 differential addition-and-doubling in mixed Lopez-Dahab projective
3106 coordinates, modified to independently blind the operands.
3108 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3110 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3111 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3112 EC_METHODs to implement their own specialized "ladder step", to take
3113 advantage of more favorable coordinate systems or more efficient
3114 differential addition-and-doubling algorithms.
3116 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3118 * Modified the random device based seed sources to keep the relevant
3119 file descriptors open rather than reopening them on each access.
3120 This allows such sources to operate in a chroot() jail without
3121 the associated device nodes being available. This behaviour can be
3122 controlled using RAND_keep_random_devices_open().
3126 * Numerous side-channel attack mitigations have been applied. This may have
3127 performance impacts for some algorithms for the benefit of improved
3128 security. Specific changes are noted in this change log by their respective
3133 * AIX shared library support overhaul. Switch to AIX "natural" way of
3134 handling shared libraries, which means collecting shared objects of
3135 different versions and bitnesses in one common archive. This allows to
3136 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3137 doesn't affect the way 3rd party applications are linked, only how
3138 multi-version installation is managed.
3142 * Make ec_group_do_inverse_ord() more robust and available to other
3143 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3144 mitigations are applied to the fallback BN_mod_inverse().
3145 When using this function rather than BN_mod_inverse() directly, new
3146 EC cryptosystem implementations are then safer-by-default.
3150 * Add coordinate blinding for EC_POINT and implement projective
3151 coordinate blinding for generic prime curves as a countermeasure to
3152 chosen point SCA attacks.
3154 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3156 * Add blinding to ECDSA and DSA signatures to protect against side channel
3157 attacks discovered by Keegan Ryan (NCC Group).
3161 * Enforce checking in the `pkeyutl` command to ensure that the input
3162 length does not exceed the maximum supported digest length when performing
3163 a sign, verify or verifyrecover operation.
3167 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3168 I/O in combination with something like select() or poll() will hang. This
3169 can be turned off again using SSL_CTX_clear_mode().
3170 Many applications do not properly handle non-application data records, and
3171 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3172 around the problems in those applications, but can also break some.
3173 It's recommended to read the manpages about SSL_read(), SSL_write(),
3174 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3175 SSL_CTX_set_read_ahead() again.
3179 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3180 now allow empty (zero character) pass phrases.
3184 * Apply blinding to binary field modular inversion and remove patent
3185 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3189 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3190 binary and prime elliptic curves.
3194 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3195 constant time fixed point multiplication.
3199 * Revise elliptic curve scalar multiplication with timing attack
3200 defenses: ec_wNAF_mul redirects to a constant time implementation
3201 when computing fixed point and variable point multiplication (which
3202 in OpenSSL are mostly used with secret scalars in keygen, sign,
3203 ECDH derive operations).
3204 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3207 * Updated CONTRIBUTING
3211 * Updated DRBG / RAND to request nonce and additional low entropy
3212 randomness from the system.
3214 *Matthias St. Pierre*
3216 * Updated 'openssl rehash' to use OpenSSL consistent default.
3220 * Moved the load of the ssl_conf module to libcrypto, which helps
3221 loading engines that libssl uses before libssl is initialised.
3225 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3229 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3231 *Ingo Schwarze, Rich Salz*
3233 * Added output of accepting IP address and port for 'openssl s_server'
3237 * Added a new API for TLSv1.3 ciphersuites:
3238 SSL_CTX_set_ciphersuites()
3239 SSL_set_ciphersuites()
3243 * Memory allocation failures consistently add an error to the error
3248 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3249 in libcrypto when run as setuid/setgid.
3253 * Load any config file by default when libssl is used.
3257 * Added new public header file <openssl/rand_drbg.h> and documentation
3258 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3260 *Matthias St. Pierre*
3262 * QNX support removed (cannot find contributors to get their approval
3263 for the license change).
3267 * TLSv1.3 replay protection for early data has been implemented. See the
3268 SSL_read_early_data() man page for further details.
3272 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3273 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3274 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3275 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3276 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3277 configuration has been separated out. See the ciphers man page or the
3278 SSL_CTX_set_ciphersuites() man page for more information.
3282 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3283 in responder mode now supports the new "-multi" option, which
3284 spawns the specified number of child processes to handle OCSP
3285 requests. The "-timeout" option now also limits the OCSP
3286 responder's patience to wait to receive the full client request
3287 on a newly accepted connection. Child processes are respawned
3288 as needed, and the CA index file is automatically reloaded
3289 when changed. This makes it possible to run the "ocsp" responder
3290 as a long-running service, making the OpenSSL CA somewhat more
3291 feature-complete. In this mode, most diagnostic messages logged
3292 after entering the event loop are logged via syslog(3) rather than
3297 * Added support for X448 and Ed448. Heavily based on original work by
3302 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3303 objects loaded. This adds the functions OSSL_STORE_expect() and
3304 OSSL_STORE_find() as well as needed tools to construct searches and
3305 get the search data out of them.
3309 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3310 version of OpenSSL should review their configuration settings to ensure
3311 that they are still appropriate for TLSv1.3. For further information see:
3312 <https://wiki.openssl.org/index.php/TLS1.3>
3316 * Grand redesign of the OpenSSL random generator
3318 The default RAND method now utilizes an AES-CTR DRBG according to
3319 NIST standard SP 800-90Ar1. The new random generator is essentially
3320 a port of the default random generator from the OpenSSL FIPS 2.0
3321 object module. It is a hybrid deterministic random bit generator
3322 using an AES-CTR bit stream and which seeds and reseeds itself
3323 automatically using trusted system entropy sources.
3325 Some of its new features are:
3326 - Support for multiple DRBG instances with seed chaining.
3327 - The default RAND method makes use of a DRBG.
3328 - There is a public and private DRBG instance.
3329 - The DRBG instances are fork-safe.
3330 - Keep all global DRBG instances on the secure heap if it is enabled.
3331 - The public and private DRBG instance are per thread for lock free
3334 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3336 * Changed Configure so it only says what it does and doesn't dump
3337 so much data. Instead, ./configdata.pm should be used as a script
3338 to display all sorts of configuration data.
3342 * Added processing of "make variables" to Configure.
3346 * Added SHA512/224 and SHA512/256 algorithm support.
3350 * The last traces of Netware support, first removed in 1.1.0, have
3355 * Get rid of Makefile.shared, and in the process, make the processing
3356 of certain files (rc.obj, or the .def/.map/.opt files produced from
3357 the ordinal files) more visible and hopefully easier to trace and
3358 debug (or make silent).
3362 * Make it possible to have environment variable assignments as
3363 arguments to config / Configure.
3367 * Add multi-prime RSA (RFC 8017) support.
3371 * Add SM3 implemented according to GB/T 32905-2016
3372 *Jack Lloyd <jack.lloyd@ribose.com>,*
3373 *Ronald Tse <ronald.tse@ribose.com>,*
3374 *Erick Borsboom <erick.borsboom@ribose.com>*
3376 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3377 as documented in RFC6066.
3378 Based on a patch from Tomasz Moń
3380 *Filipe Raimundo da Silva*
3382 * Add SM4 implemented according to GB/T 32907-2016.
3383 *Jack Lloyd <jack.lloyd@ribose.com>,*
3384 *Ronald Tse <ronald.tse@ribose.com>,*
3385 *Erick Borsboom <erick.borsboom@ribose.com>*
3387 * Reimplement -newreq-nodes and ERR_error_string_n; the
3388 original author does not agree with the license change.
3392 * Add ARIA AEAD TLS support.
3396 * Some macro definitions to support VS6 have been removed. Visual
3397 Studio 6 has not worked since 1.1.0
3401 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3402 without clearing the errors.
3406 * Add "atfork" functions. If building on a system that without
3407 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3408 requirements. The RAND facility now uses/requires this.
3416 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3417 not possible to disable entirely. However, it's still possible to
3418 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3421 To disable, configure with 'no-ui-console'. 'no-ui' is still
3422 possible to use as an alias. Check at compile time with the
3423 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3424 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3428 * Add a STORE module, which implements a uniform and URI based reader of
3429 stores that can contain keys, certificates, CRLs and numerous other
3430 objects. The main API is loosely based on a few stdio functions,
3431 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3432 OSSL_STORE_error and OSSL_STORE_close.
3433 The implementation uses backends called "loaders" to implement arbitrary
3434 URI schemes. There is one built in "loader" for the 'file' scheme.
3438 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3439 then adjusted to work on FreeBSD 8.4 as well.
3440 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3441 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3445 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3446 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3447 error code calls like this:
3449 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3451 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3452 that can be encoded in C. For the foreseeable future, this will only
3455 *Richard Levitte and Tim Hudson*
3457 * Removed BSD cryptodev engine.
3461 * Add a build target 'build_all_generated', to build all generated files
3462 and only that. This can be used to prepare everything that requires
3463 things like perl for a system that lacks perl and then move everything
3464 to that system and do the rest of the build there.
3468 * In the UI interface, make it possible to duplicate the user data. This
3469 can be used by engines that need to retain the data for a longer time
3470 than just the call where this user data is passed.
3474 * Ignore the '-named_curve auto' value for compatibility of applications
3477 *Tomáš Mráz <tmraz@fedoraproject.org>*
3479 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3480 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3481 alerts across multiple records (some of which could be empty). In practice
3482 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3483 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3484 support this at all. Supporting it adds significant complexity to the
3485 record layer, and its removal is unlikely to cause interoperability
3490 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3491 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3492 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3497 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3498 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3500 *Richard Levitte, Andy Polyakov*
3502 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3507 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3508 platform rather than 'mingw'.
3512 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3513 success if they are asked to add an object which already exists
3514 in the store. This change cascades to other functions which load
3515 certificates and CRLs.
3519 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3520 facilitate stack unwinding even from assembly subroutines.
3524 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3525 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3529 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3530 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3531 which is the minimum version we support.
3535 * Certificate time validation (X509_cmp_time) enforces stricter
3536 compliance with RFC 5280. Fractional seconds and timezone offsets
3537 are no longer allowed.
3541 * Add support for ARIA
3545 * s_client will now send the Server Name Indication (SNI) extension by
3546 default unless the new "-noservername" option is used. The server name is
3547 based on the host provided to the "-connect" option unless overridden by
3548 using "-servername".
3552 * Add support for SipHash
3556 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3557 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3558 prevent issues where no progress is being made and the peer continually
3559 sends unrecognised record types, using up resources processing them.
3563 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3564 using the algorithm defined in
3565 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3569 * Heartbeat support has been removed; the ABI is changed for now.
3571 *Richard Levitte, Rich Salz*
3573 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3577 * The RSA "null" method, which was partially supported to avoid patent
3578 issues, has been replaced to always returns NULL.
3585 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3587 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3588 used even when parsing explicit parameters, when loading a encoded key
3589 or calling `EC_GROUP_new_from_ecpkparameters()`/
3590 `EC_GROUP_new_from_ecparameters()`.
3591 This prevents bypass of security hardening and performance gains,
3592 especially for curves with specialized EC_METHODs.
3593 By default, if a key encoded with explicit parameters is loaded and later
3594 encoded, the output is still encoded with explicit parameters, even if
3595 internally a "named" EC_GROUP is used for computation.
3599 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3600 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3601 NULL. After this change, only the cofactor parameter can be NULL. It also
3602 does some minimal sanity checks on the passed order.
3607 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3608 An attack is simple, if the first CMS_recipientInfo is valid but the
3609 second CMS_recipientInfo is chosen ciphertext. If the second
3610 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3611 encryption key will be replaced by garbage, and the message cannot be
3612 decoded, but if the RSA decryption fails, the correct encryption key is
3613 used and the recipient will not notice the attack.
3614 As a work around for this potential attack the length of the decrypted
3615 key must be equal to the cipher default key length, in case the
3616 certificate is not given and all recipientInfo are tried out.
3617 The old behaviour can be re-enabled in the CMS code by setting the
3618 CMS_DEBUG_DECRYPT flag.
3623 * Use Windows installation paths in the mingw builds
3625 Mingw isn't a POSIX environment per se, which means that Windows
3626 paths should be used for installation.
3631 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3633 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3634 This changes the size when using the `genpkey` command when no size is given.
3635 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3636 generation commands to use 2048 bits by default.
3640 * Prevent over long nonces in ChaCha20-Poly1305.
3642 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3643 for every encryption operation. RFC 7539 specifies that the nonce value
3644 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3645 and front pads the nonce with 0 bytes if it is less than 12
3646 bytes. However it also incorrectly allows a nonce to be set of up to 16
3647 bytes. In this case only the last 12 bytes are significant and any
3648 additional leading bytes are ignored.
3650 It is a requirement of using this cipher that nonce values are
3651 unique. Messages encrypted using a reused nonce value are susceptible to
3652 serious confidentiality and integrity attacks. If an application changes
3653 the default nonce length to be longer than 12 bytes and then makes a
3654 change to the leading bytes of the nonce expecting the new value to be a
3655 new unique nonce then such an application could inadvertently encrypt
3656 messages with a reused nonce.
3658 Additionally the ignored bytes in a long nonce are not covered by the
3659 integrity guarantee of this cipher. Any application that relies on the
3660 integrity of these ignored leading bytes of a long nonce may be further
3661 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3662 is safe because no such use sets such a long nonce value. However user
3663 applications that use this cipher directly and set a non-default nonce
3664 length to be longer than 12 bytes may be vulnerable.
3666 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3672 * Added SCA hardening for modular field inversion in EC_GROUP through
3673 a new dedicated field_inv() pointer in EC_METHOD.
3674 This also addresses a leakage affecting conversions from projective
3675 to affine coordinates.
3677 *Billy Bob Brumley, Nicola Tuveri*
3679 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3680 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3684 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3688 * Remove the 'dist' target and add a tarball building script. The
3689 'dist' target has fallen out of use, and it shouldn't be
3690 necessary to configure just to create a source distribution.
3694 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3696 * Timing vulnerability in DSA signature generation
3698 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3699 timing side channel attack. An attacker could use variations in the signing
3700 algorithm to recover the private key.
3702 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3707 * Timing vulnerability in ECDSA signature generation
3709 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3710 timing side channel attack. An attacker could use variations in the signing
3711 algorithm to recover the private key.
3713 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3718 * Add coordinate blinding for EC_POINT and implement projective
3719 coordinate blinding for generic prime curves as a countermeasure to
3720 chosen point SCA attacks.
3722 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3724 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3726 * Client DoS due to large DH parameter
3728 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3729 malicious server can send a very large prime value to the client. This will
3730 cause the client to spend an unreasonably long period of time generating a
3731 key for this prime resulting in a hang until the client has finished. This
3732 could be exploited in a Denial Of Service attack.
3734 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3739 * Cache timing vulnerability in RSA Key Generation
3741 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3742 a cache timing side channel attack. An attacker with sufficient access to
3743 mount cache timing attacks during the RSA key generation process could
3744 recover the private key.
3746 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3747 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3752 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3753 parameter is no longer accepted, as it leads to a corrupt table. NULL
3754 pem_str is reserved for alias entries only.
3758 * Revert blinding in ECDSA sign and instead make problematic addition
3759 length-invariant. Switch even to fixed-length Montgomery multiplication.
3763 * Change generating and checking of primes so that the error rate of not
3764 being prime depends on the intended use based on the size of the input.
3765 For larger primes this will result in more rounds of Miller-Rabin.
3766 The maximal error rate for primes with more than 1080 bits is lowered
3769 *Kurt Roeckx, Annie Yousar*
3771 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3775 * Add blinding to ECDSA and DSA signatures to protect against side channel
3776 attacks discovered by Keegan Ryan (NCC Group).
3780 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3781 now allow empty (zero character) pass phrases.
3785 * Certificate time validation (X509_cmp_time) enforces stricter
3786 compliance with RFC 5280. Fractional seconds and timezone offsets
3787 are no longer allowed.
3791 * Fixed a text canonicalisation bug in CMS
3793 Where a CMS detached signature is used with text content the text goes
3794 through a canonicalisation process first prior to signing or verifying a
3795 signature. This process strips trailing space at the end of lines, converts
3796 line terminators to CRLF and removes additional trailing line terminators
3797 at the end of a file. A bug in the canonicalisation process meant that
3798 some characters, such as form-feed, were incorrectly treated as whitespace
3799 and removed. This is contrary to the specification (RFC5485). This fix
3800 could mean that detached text data signed with an earlier version of
3801 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3802 signed with a fixed OpenSSL may fail to verify with an earlier version of
3803 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3804 and use the "-binary" flag (for the "cms" command line application) or set
3805 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3809 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3811 * Constructed ASN.1 types with a recursive definition could exceed the stack
3813 Constructed ASN.1 types with a recursive definition (such as can be found
3814 in PKCS7) could eventually exceed the stack given malicious input with
3815 excessive recursion. This could result in a Denial Of Service attack. There
3816 are no such structures used within SSL/TLS that come from untrusted sources
3817 so this is considered safe.
3819 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3825 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3827 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3828 effectively reduced to only comparing the least significant bit of each
3829 byte. This allows an attacker to forge messages that would be considered as
3830 authenticated in an amount of tries lower than that guaranteed by the
3831 security claims of the scheme. The module can only be compiled by the
3832 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3834 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3840 * Add a build target 'build_all_generated', to build all generated files
3841 and only that. This can be used to prepare everything that requires
3842 things like perl for a system that lacks perl and then move everything
3843 to that system and do the rest of the build there.
3847 * Backport SSL_OP_NO_RENGOTIATION
3849 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3850 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3851 changes this is no longer possible in 1.1.0. Therefore, the new
3852 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3853 1.1.0 to provide equivalent functionality.
3855 Note that if an application built against 1.1.0h headers (or above) is run
3856 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3857 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3861 * Removed the OS390-Unix config target. It relied on a script that doesn't
3866 * rsaz_1024_mul_avx2 overflow bug on x86_64
3868 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3869 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3870 Analysis suggests that attacks against RSA and DSA as a result of this
3871 defect would be very difficult to perform and are not believed likely.
3872 Attacks against DH1024 are considered just feasible, because most of the
3873 work necessary to deduce information about a private key may be performed
3874 offline. The amount of resources required for such an attack would be
3875 significant. However, for an attack on TLS to be meaningful, the server
3876 would have to share the DH1024 private key among multiple clients, which is
3877 no longer an option since CVE-2016-0701.
3879 This only affects processors that support the AVX2 but not ADX extensions
3880 like Intel Haswell (4th generation).
3882 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3883 was originally found via the OSS-Fuzz project.
3888 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3890 * bn_sqrx8x_internal carry bug on x86_64
3892 There is a carry propagating bug in the x86_64 Montgomery squaring
3893 procedure. No EC algorithms are affected. Analysis suggests that attacks
3894 against RSA and DSA as a result of this defect would be very difficult to
3895 perform and are not believed likely. Attacks against DH are considered just
3896 feasible (although very difficult) because most of the work necessary to
3897 deduce information about a private key may be performed offline. The amount
3898 of resources required for such an attack would be very significant and
3899 likely only accessible to a limited number of attackers. An attacker would
3900 additionally need online access to an unpatched system using the target
3901 private key in a scenario with persistent DH parameters and a private
3902 key that is shared between multiple clients.
3904 This only affects processors that support the BMI1, BMI2 and ADX extensions
3905 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3907 This issue was reported to OpenSSL by the OSS-Fuzz project.
3912 * Malformed X.509 IPAddressFamily could cause OOB read
3914 If an X.509 certificate has a malformed IPAddressFamily extension,
3915 OpenSSL could do a one-byte buffer overread. The most likely result
3916 would be an erroneous display of the certificate in text format.
3918 This issue was reported to OpenSSL by the OSS-Fuzz project.
3923 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3925 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3926 platform rather than 'mingw'.
3930 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3931 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3932 which is the minimum version we support.
3936 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3938 * Encrypt-Then-Mac renegotiation crash
3940 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3941 negotiated where it was not in the original handshake (or vice-versa) then
3942 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
3943 and servers are affected.
3945 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3950 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3952 * Truncated packet could crash via OOB read
3954 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3955 cipher is being used, then a truncated packet can cause that host to
3956 perform an out-of-bounds read, usually resulting in a crash.
3958 This issue was reported to OpenSSL by Robert Święcki of Google.
3963 * Bad (EC)DHE parameters cause a client crash
3965 If a malicious server supplies bad parameters for a DHE or ECDHE key
3966 exchange then this can result in the client attempting to dereference a
3967 NULL pointer leading to a client crash. This could be exploited in a Denial
3970 This issue was reported to OpenSSL by Guido Vranken.
3975 * BN_mod_exp may produce incorrect results on x86_64
3977 There is a carry propagating bug in the x86_64 Montgomery squaring
3978 procedure. No EC algorithms are affected. Analysis suggests that attacks
3979 against RSA and DSA as a result of this defect would be very difficult to
3980 perform and are not believed likely. Attacks against DH are considered just
3981 feasible (although very difficult) because most of the work necessary to
3982 deduce information about a private key may be performed offline. The amount
3983 of resources required for such an attack would be very significant and
3984 likely only accessible to a limited number of attackers. An attacker would
3985 additionally need online access to an unpatched system using the target
3986 private key in a scenario with persistent DH parameters and a private
3987 key that is shared between multiple clients. For example this can occur by
3988 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3989 similar to CVE-2015-3193 but must be treated as a separate problem.
3991 This issue was reported to OpenSSL by the OSS-Fuzz project.
3996 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3998 * ChaCha20/Poly1305 heap-buffer-overflow
4000 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4001 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4002 crash. This issue is not considered to be exploitable beyond a DoS.
4004 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4009 * CMS Null dereference
4011 Applications parsing invalid CMS structures can crash with a NULL pointer
4012 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4013 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4014 structure callback if an attempt is made to free certain invalid encodings.
4015 Only CHOICE structures using a callback which do not handle NULL value are
4018 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4023 * Montgomery multiplication may produce incorrect results
4025 There is a carry propagating bug in the Broadwell-specific Montgomery
4026 multiplication procedure that handles input lengths divisible by, but
4027 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4028 and DH private keys are impossible. This is because the subroutine in
4029 question is not used in operations with the private key itself and an input
4030 of the attacker's direct choice. Otherwise the bug can manifest itself as
4031 transient authentication and key negotiation failures or reproducible
4032 erroneous outcome of public-key operations with specially crafted input.
4033 Among EC algorithms only Brainpool P-512 curves are affected and one
4034 presumably can attack ECDH key negotiation. Impact was not analyzed in
4035 detail, because pre-requisites for attack are considered unlikely. Namely
4036 multiple clients have to choose the curve in question and the server has to
4037 share the private key among them, neither of which is default behaviour.
4038 Even then only clients that chose the curve will be affected.
4040 This issue was publicly reported as transient failures and was not
4041 initially recognized as a security issue. Thanks to Richard Morgan for
4042 providing reproducible case.
4047 * Removed automatic addition of RPATH in shared libraries and executables,
4048 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4052 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4054 * Fix Use After Free for large message sizes
4056 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4057 message larger than approx 16k is received then the underlying buffer to
4058 store the incoming message is reallocated and moved. Unfortunately a
4059 dangling pointer to the old location is left which results in an attempt to
4060 write to the previously freed location. This is likely to result in a
4061 crash, however it could potentially lead to execution of arbitrary code.
4063 This issue only affects OpenSSL 1.1.0a.
4065 This issue was reported to OpenSSL by Robert Święcki.
4070 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4072 * OCSP Status Request extension unbounded memory growth
4074 A malicious client can send an excessively large OCSP Status Request
4075 extension. If that client continually requests renegotiation, sending a
4076 large OCSP Status Request extension each time, then there will be unbounded
4077 memory growth on the server. This will eventually lead to a Denial Of
4078 Service attack through memory exhaustion. Servers with a default
4079 configuration are vulnerable even if they do not support OCSP. Builds using
4080 the "no-ocsp" build time option are not affected.
4082 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4087 * SSL_peek() hang on empty record
4089 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4090 sends an empty record. This could be exploited by a malicious peer in a
4091 Denial Of Service attack.
4093 This issue was reported to OpenSSL by Alex Gaynor.
4098 * Excessive allocation of memory in tls_get_message_header() and
4099 dtls1_preprocess_fragment()
4101 A (D)TLS message includes 3 bytes for its length in the header for the
4102 message. This would allow for messages up to 16Mb in length. Messages of
4103 this length are excessive and OpenSSL includes a check to ensure that a
4104 peer is sending reasonably sized messages in order to avoid too much memory
4105 being consumed to service a connection. A flaw in the logic of version
4106 1.1.0 means that memory for the message is allocated too early, prior to
4107 the excessive message length check. Due to way memory is allocated in
4108 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4109 to service a connection. This could lead to a Denial of Service through
4110 memory exhaustion. However, the excessive message length check still takes
4111 place, and this would cause the connection to immediately fail. Assuming
4112 that the application calls SSL_free() on the failed connection in a timely
4113 manner then the 21Mb of allocated memory will then be immediately freed
4114 again. Therefore, the excessive memory allocation will be transitory in
4115 nature. This then means that there is only a security impact if:
4117 1) The application does not call SSL_free() in a timely manner in the event
4118 that the connection fails
4120 2) The application is working in a constrained environment where there is
4121 very little free memory
4123 3) The attacker initiates multiple connection attempts such that there are
4124 multiple connections in a state where memory has been allocated for the
4125 connection; SSL_free() has not yet been called; and there is insufficient
4126 memory to service the multiple requests.
4128 Except in the instance of (1) above any Denial Of Service is likely to be
4129 transitory because as soon as the connection fails the memory is
4130 subsequently freed again in the SSL_free() call. However there is an
4131 increased risk during this period of application crashes due to the lack of
4132 memory - which would then mean a more serious Denial of Service.
4134 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4135 (CVE-2016-6307 and CVE-2016-6308)
4139 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4140 had to be removed. Primary reason is that vendor assembler can't
4141 assemble our modules with -KPIC flag. As result it, assembly
4142 support, was not even available as option. But its lack means
4143 lack of side-channel resistant code, which is incompatible with
4144 security by todays standards. Fortunately gcc is readily available
4145 prepackaged option, which we firmly point at...
4149 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4151 * Windows command-line tool supports UTF-8 opt-in option for arguments
4152 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4153 (to any value) allows Windows user to access PKCS#12 file generated
4154 with Windows CryptoAPI and protected with non-ASCII password, as well
4155 as files generated under UTF-8 locale on Linux also protected with
4160 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4161 have been disabled by default and removed from DEFAULT, just like RC4.
4162 See the RC4 item below to re-enable both.
4166 * The method for finding the storage location for the Windows RAND seed file
4167 has changed. First we check %RANDFILE%. If that is not set then we check
4168 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4169 all else fails we fall back to C:\.
4173 * The EVP_EncryptUpdate() function has had its return type changed from void
4174 to int. A return of 0 indicates and error while a return of 1 indicates
4179 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4180 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4181 off the constant time implementation for RSA, DSA and DH have been made
4182 no-ops and deprecated.
4186 * Windows RAND implementation was simplified to only get entropy by
4187 calling CryptGenRandom(). Various other RAND-related tickets
4190 *Joseph Wylie Yandle, Rich Salz*
4192 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4193 and `OPENSSL_LH_`, respectively. The old names are available
4194 with API compatibility. They new names are now completely documented.
4198 * Unify TYPE_up_ref(obj) methods signature.
4199 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4200 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4201 int (instead of void) like all others TYPE_up_ref() methods.
4202 So now these methods also check the return value of CRYPTO_atomic_add(),
4203 and the validity of object reference counter.
4205 *fdasilvayy@gmail.com*
4207 * With Windows Visual Studio builds, the .pdb files are installed
4208 alongside the installed libraries and executables. For a static
4209 library installation, ossl_static.pdb is the associate compiler
4210 generated .pdb file to be used when linking programs.
4214 * Remove openssl.spec. Packaging files belong with the packagers.
4218 * Automatic Darwin/OSX configuration has had a refresh, it will now
4219 recognise x86_64 architectures automatically. You can still decide
4220 to build for a different bitness with the environment variable
4221 KERNEL_BITS (can be 32 or 64), for example:
4223 KERNEL_BITS=32 ./config
4227 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4228 256 bit AES and HMAC with SHA256.
4232 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4236 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4240 * To enable users to have their own config files and build file templates,
4241 Configure looks in the directory indicated by the environment variable
4242 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4243 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4244 name and is used as is.
4248 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4249 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4250 X509_CERT_FILE_CTX was removed.
4254 * "shared" builds are now the default. To create only static libraries use
4255 the "no-shared" Configure option.
4259 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4260 All of these option have not worked for some while and are fundamental
4265 * Make various cleanup routines no-ops and mark them as deprecated. Most
4266 global cleanup functions are no longer required because they are handled
4267 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4268 Explicitly de-initing can cause problems (e.g. where a library that uses
4269 OpenSSL de-inits, but an application is still using it). The affected
4270 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4271 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4272 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4273 COMP_zlib_cleanup().
4277 * --strict-warnings no longer enables runtime debugging options
4278 such as REF_DEBUG. Instead, debug options are automatically
4279 enabled with '--debug' builds.
4281 *Andy Polyakov, Emilia Käsper*
4283 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4284 have been moved out of the public header files. New functions for managing
4285 these have been added.
4289 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4290 objects have been moved out of the public header files. New
4291 functions for managing these have been added.
4295 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4296 have been moved out of the public header files. New functions for managing
4297 these have been added.
4301 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4302 moved out of the public header files. New functions for managing these
4307 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4311 * Removed the mk1mf build scripts.
4315 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4316 it is always safe to #include a header now.
4320 * Removed the aged BC-32 config and all its supporting scripts
4324 * Removed support for Ultrix, Netware, and OS/2.
4328 * Add support for HKDF.
4330 *Alessandro Ghedini*
4332 * Add support for blake2b and blake2s
4336 * Added support for "pipelining". Ciphers that have the
4337 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4338 encryptions/decryptions simultaneously. There are currently no built-in
4339 ciphers with this property but the expectation is that engines will be able
4340 to offer it to significantly improve throughput. Support has been extended
4341 into libssl so that multiple records for a single connection can be
4342 processed in one go (for >=TLS 1.1).
4346 * Added the AFALG engine. This is an async capable engine which is able to
4347 offload work to the Linux kernel. In this initial version it only supports
4348 AES128-CBC. The kernel must be version 4.1.0 or greater.
4352 * OpenSSL now uses a new threading API. It is no longer necessary to
4353 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4354 are two supported threading models: pthreads and windows threads. It is
4355 also possible to configure OpenSSL at compile time for "no-threads". The
4356 old threading API should no longer be used. The functions have been
4357 replaced with "no-op" compatibility macros.
4359 *Alessandro Ghedini, Matt Caswell*
4361 * Modify behavior of ALPN to invoke callback after SNI/servername
4362 callback, such that updates to the SSL_CTX affect ALPN.
4366 * Add SSL_CIPHER queries for authentication and key-exchange.
4370 * Changes to the DEFAULT cipherlist:
4371 - Prefer (EC)DHE handshakes over plain RSA.
4372 - Prefer AEAD ciphers over legacy ciphers.
4373 - Prefer ECDSA over RSA when both certificates are available.
4374 - Prefer TLSv1.2 ciphers/PRF.
4375 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4380 * Change the ECC default curve list to be this, in order: x25519,
4381 secp256r1, secp521r1, secp384r1.
4385 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4386 disabled by default. They can be re-enabled using the
4387 enable-weak-ssl-ciphers option to Configure.
4391 * If the server has ALPN configured, but supports no protocols that the
4392 client advertises, send a fatal "no_application_protocol" alert.
4393 This behaviour is SHALL in RFC 7301, though it isn't universally
4394 implemented by other servers.
4398 * Add X25519 support.
4399 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4400 for public and private key encoding using the format documented in
4401 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4402 key generation and key derivation.
4404 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4409 * Deprecate SRP_VBASE_get_by_user.
4410 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4411 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4412 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4413 seed, even if the seed is configured.
4415 Users should use SRP_VBASE_get1_by_user instead. Note that in
4416 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4417 also that even though configuring the SRP seed attempts to hide
4418 invalid usernames by continuing the handshake with fake
4419 credentials, this behaviour is not constant time and no strong
4420 guarantees are made that the handshake is indistinguishable from
4421 that of a valid user.
4425 * Configuration change; it's now possible to build dynamic engines
4426 without having to build shared libraries and vice versa. This
4427 only applies to the engines in `engines/`, those in `crypto/engine/`
4428 will always be built into libcrypto (i.e. "static").
4430 Building dynamic engines is enabled by default; to disable, use
4431 the configuration option "disable-dynamic-engine".
4433 The only requirements for building dynamic engines are the
4434 presence of the DSO module and building with position independent
4435 code, so they will also automatically be disabled if configuring
4436 with "disable-dso" or "disable-pic".
4438 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4439 are also taken away from openssl/opensslconf.h, as they are
4444 * Configuration change; if there is a known flag to compile
4445 position independent code, it will always be applied on the
4446 libcrypto and libssl object files, and never on the application
4447 object files. This means other libraries that use routines from
4448 libcrypto / libssl can be made into shared libraries regardless
4449 of how OpenSSL was configured.
4451 If this isn't desirable, the configuration options "disable-pic"
4452 or "no-pic" can be used to disable the use of PIC. This will
4453 also disable building shared libraries and dynamic engines.
4457 * Removed JPAKE code. It was experimental and has no wide use.
4461 * The INSTALL_PREFIX Makefile variable has been renamed to
4462 DESTDIR. That makes for less confusion on what this variable
4463 is for. Also, the configuration option --install_prefix is
4468 * Heartbeat for TLS has been removed and is disabled by default
4469 for DTLS; configure with enable-heartbeats. Code that uses the
4470 old #define's might need to be updated.
4472 *Emilia Käsper, Rich Salz*
4474 * Rename REF_CHECK to REF_DEBUG.
4478 * New "unified" build system
4480 The "unified" build system is aimed to be a common system for all
4481 platforms we support. With it comes new support for VMS.
4483 This system builds supports building in a different directory tree
4484 than the source tree. It produces one Makefile (for unix family
4485 or lookalikes), or one descrip.mms (for VMS).
4487 The source of information to make the Makefile / descrip.mms is
4488 small files called 'build.info', holding the necessary
4489 information for each directory with source to compile, and a
4490 template in Configurations, like unix-Makefile.tmpl or
4493 With this change, the library names were also renamed on Windows
4494 and on VMS. They now have names that are closer to the standard
4495 on Unix, and include the major version number, and in certain
4496 cases, the architecture they are built for. See "Notes on shared
4497 libraries" in INSTALL.
4499 We rely heavily on the perl module Text::Template.
4503 * Added support for auto-initialisation and de-initialisation of the library.
4504 OpenSSL no longer requires explicit init or deinit routines to be called,
4505 except in certain circumstances. See the OPENSSL_init_crypto() and
4506 OPENSSL_init_ssl() man pages for further information.
4510 * The arguments to the DTLSv1_listen function have changed. Specifically the
4511 "peer" argument is now expected to be a BIO_ADDR object.
4513 * Rewrite of BIO networking library. The BIO library lacked consistent
4514 support of IPv6, and adding it required some more extensive
4515 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4516 which hold all types of addresses and chains of address information.
4517 It also introduces a new API, with functions like BIO_socket,
4518 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4519 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4520 have been adapted accordingly.
4524 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4529 * CRIME protection: disable compression by default, even if OpenSSL is
4530 compiled with zlib enabled. Applications can still enable compression
4531 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4532 using the SSL_CONF library to configure compression.
4536 * The signature of the session callback configured with
4537 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4538 was explicitly marked as `const unsigned char*` instead of
4543 * Always DPURIFY. Remove the use of uninitialized memory in the
4544 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4548 * Removed many obsolete configuration items, including
4549 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4550 MD2_CHAR, MD2_INT, MD2_LONG
4552 IDEA_SHORT, IDEA_LONG
4553 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4555 *Rich Salz, with advice from Andy Polyakov*
4557 * Many BN internals have been moved to an internal header file.
4559 *Rich Salz with help from Andy Polyakov*
4561 * Configuration and writing out the results from it has changed.
4562 Files such as Makefile include/openssl/opensslconf.h and are now
4563 produced through general templates, such as Makefile.in and
4564 crypto/opensslconf.h.in and some help from the perl module
4567 Also, the center of configuration information is no longer
4568 Makefile. Instead, Configure produces a perl module in
4569 configdata.pm which holds most of the config data (in the hash
4570 table %config), the target data that comes from the target
4571 configuration in one of the `Configurations/*.conf` files (in
4576 * To clarify their intended purposes, the Configure options
4577 --prefix and --openssldir change their semantics, and become more
4578 straightforward and less interdependent.
4580 --prefix shall be used exclusively to give the location INSTALLTOP
4581 where programs, scripts, libraries, include files and manuals are
4582 going to be installed. The default is now /usr/local.
4584 --openssldir shall be used exclusively to give the default
4585 location OPENSSLDIR where certificates, private keys, CRLs are
4586 managed. This is also where the default openssl.cnf gets
4588 If the directory given with this option is a relative path, the
4589 values of both the --prefix value and the --openssldir value will
4590 be combined to become OPENSSLDIR.
4591 The default for --openssldir is INSTALLTOP/ssl.
4593 Anyone who uses --openssldir to specify where OpenSSL is to be
4594 installed MUST change to use --prefix instead.
4598 * The GOST engine was out of date and therefore it has been removed. An up
4599 to date GOST engine is now being maintained in an external repository.
4600 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4601 support for GOST ciphersuites (these are only activated if a GOST engine
4606 * EGD is no longer supported by default; use enable-egd when
4609 *Ben Kaduk and Rich Salz*
4611 * The distribution now has Makefile.in files, which are used to
4612 create Makefile's when Configure is run. *Configure must be run
4613 before trying to build now.*
4617 * The return value for SSL_CIPHER_description() for error conditions
4622 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4624 Obtaining and performing DNSSEC validation of TLSA records is
4625 the application's responsibility. The application provides
4626 the TLSA records of its choice to OpenSSL, and these are then
4627 used to authenticate the peer.
4629 The TLSA records need not even come from DNS. They can, for
4630 example, be used to implement local end-entity certificate or
4631 trust-anchor "pinning", where the "pin" data takes the form
4632 of TLSA records, which can augment or replace verification
4633 based on the usual WebPKI public certification authorities.
4637 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4638 continues to support deprecated interfaces in default builds.
4639 However, applications are strongly advised to compile their
4640 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4641 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4642 or the 1.1.0 releases.
4644 In environments in which all applications have been ported to
4645 not use any deprecated interfaces OpenSSL's Configure script
4646 should be used with the --api=1.1.0 option to entirely remove
4647 support for the deprecated features from the library and
4648 unconditionally disable them in the installed headers.
4649 Essentially the same effect can be achieved with the "no-deprecated"
4650 argument to Configure, except that this will always restrict
4651 the build to just the latest API, rather than a fixed API
4654 As applications are ported to future revisions of the API,
4655 they should update their compile-time OPENSSL_API_COMPAT define
4656 accordingly, but in most cases should be able to continue to
4657 compile with later releases.
4659 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4660 0x10000000L and 0x00908000L, respectively. However those
4661 versions did not support the OPENSSL_API_COMPAT feature, and
4662 so applications are not typically tested for explicit support
4663 of just the undeprecated features of either release.
4667 * Add support for setting the minimum and maximum supported protocol.
4668 It can bet set via the SSL_set_min_proto_version() and
4669 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4670 MaxProtocol. It's recommended to use the new APIs to disable
4671 protocols instead of disabling individual protocols using
4672 SSL_set_options() or SSL_CONF's Protocol. This change also
4673 removes support for disabling TLS 1.2 in the OpenSSL TLS
4674 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4678 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4682 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4683 and integrates ECDSA and ECDH functionality into EC. Implementations can
4684 now redirect key generation and no longer need to convert to or from
4687 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4688 include the ec.h header file instead.
4692 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4693 ciphers who are no longer supported and drops support the ephemeral RSA key
4694 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4698 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4699 opaque. For HMAC_CTX, the following constructors and destructors
4702 HMAC_CTX *HMAC_CTX_new(void);
4703 void HMAC_CTX_free(HMAC_CTX *ctx);
4705 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4706 destroy such methods has been added. See EVP_MD_meth_new(3) and
4707 EVP_CIPHER_meth_new(3) for documentation.
4710 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4711 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4712 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4713 an already created structure.
4714 2) For consistency with the majority of our object creators and
4715 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4716 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4717 for deprecated builds.
4721 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4722 cryptographic operations to be performed asynchronously as long as an
4723 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4724 further details. Libssl has also had this capability integrated with the
4725 introduction of the new mode SSL_MODE_ASYNC and associated error
4726 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4727 pages. This work was developed in partnership with Intel Corp.
4731 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4732 always enabled now. If you want to disable the support you should
4733 exclude it using the list of supported ciphers. This also means that the
4734 "-no_ecdhe" option has been removed from s_server.
4738 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4739 SSL_{CTX_}set1_curves() which can set a list.
4743 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4744 curve you want to support using SSL_{CTX_}set1_curves().
4748 * State machine rewrite. The state machine code has been significantly
4749 refactored in order to remove much duplication of code and solve issues
4750 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4751 further details). This change does have some associated API changes.
4752 Notably the SSL_state() function has been removed and replaced by
4753 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4754 SSL_set_state() has been removed altogether. The previous handshake states
4755 defined in ssl.h and ssl3.h have also been removed.
4759 * All instances of the string "ssleay" in the public API were replaced
4760 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4761 Some error codes related to internal RSA_eay API's were renamed.
4765 * The demo files in crypto/threads were moved to demo/threads.
4769 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4772 *Matt Caswell, Rich Salz*
4774 * New ASN.1 embed macro.
4776 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4777 structure is not allocated: it is part of the parent. That is instead of
4785 This reduces memory fragmentation and make it impossible to accidentally
4786 set a mandatory field to NULL.
4788 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4789 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4790 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4795 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4799 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4800 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4801 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4802 DES and RC4 ciphersuites.
4806 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4807 This changes the decoding behaviour for some invalid messages,
4808 though the change is mostly in the more lenient direction, and
4809 legacy behaviour is preserved as much as possible.
4813 * Fix no-stdio build.
4814 *David Woodhouse <David.Woodhouse@intel.com> and also*
4815 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4817 * New testing framework
4818 The testing framework has been largely rewritten and is now using
4819 perl and the perl modules Test::Harness and an extended variant of
4820 Test::More called OpenSSL::Test to do its work. All test scripts in
4821 test/ have been rewritten into test recipes, and all direct calls to
4822 executables in test/Makefile have become individual recipes using the
4823 simplified testing OpenSSL::Test::Simple.
4825 For documentation on our testing modules, do:
4827 perldoc test/testlib/OpenSSL/Test/Simple.pm
4828 perldoc test/testlib/OpenSSL/Test.pm
4832 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4833 are used; the latter aborts on memory leaks (usually checked on exit).
4834 Some undocumented "set malloc, etc., hooks" functions were removed
4835 and others were changed. All are now documented.
4839 * In DSA_generate_parameters_ex, if the provided seed is too short,
4842 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4844 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4845 from RFC4279, RFC4785, RFC5487, RFC5489.
4847 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4848 original RSA_PSK patch.
4852 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4853 era flag was never set throughout the codebase (only read). Also removed
4854 SSL3_FLAGS_POP_BUFFER which was only used if
4855 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4859 * Changed the default name options in the "ca", "crl", "req" and "x509"
4860 to be "oneline" instead of "compat".
4864 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4865 not aware of clients that still exhibit this bug, and the workaround
4866 hasn't been working properly for a while.
4870 * The return type of BIO_number_read() and BIO_number_written() as well as
4871 the corresponding num_read and num_write members in the BIO structure has
4872 changed from unsigned long to uint64_t. On platforms where an unsigned
4873 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4878 * Given the pervasive nature of TLS extensions it is inadvisable to run
4879 OpenSSL without support for them. It also means that maintaining
4880 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4881 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
4885 * Removed support for the two export grade static DH ciphersuites
4886 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4887 were newly added (along with a number of other static DH ciphersuites) to
4888 1.0.2. However the two export ones have *never* worked since they were
4889 introduced. It seems strange in any case to be adding new export
4890 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4894 * Version negotiation has been rewritten. In particular SSLv23_method(),
4895 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4896 and turned into macros which simply call the new preferred function names
4897 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4898 should use the new names instead. Also as part of this change the ssl23.h
4899 header file has been removed.
4903 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4904 code and the associated standard is no longer considered fit-for-purpose.
4908 * RT2547 was closed. When generating a private key, try to make the
4909 output file readable only by the owner. This behavior change might
4910 be noticeable when interacting with other software.
4912 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4917 * Added HTTP GET support to the ocsp command.
4921 * Changed default digest for the dgst and enc commands from MD5 to
4926 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4930 * Added support for TLS extended master secret from
4931 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4932 initial patch which was a great help during development.
4936 * All libssl internal structures have been removed from the public header
4937 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4938 now redundant). Users should not attempt to access internal structures
4939 directly. Instead they should use the provided API functions.
4943 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4944 Access to deprecated functions can be re-enabled by running config with
4945 "enable-deprecated". In addition applications wishing to use deprecated
4946 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4947 will, by default, disable some transitive includes that previously existed
4948 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4952 * Added support for OCB mode. OpenSSL has been granted a patent license
4953 compatible with the OpenSSL license for use of OCB. Details are available
4954 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4955 for OCB can be removed by calling config with no-ocb.
4959 * SSLv2 support has been removed. It still supports receiving an SSLv2
4960 compatible client hello.
4964 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4965 done while fixing the error code for the key-too-small case.
4967 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4969 * CA.sh has been removed; use CA.pl instead.
4973 * Removed old DES API.
4977 * Remove various unsupported platforms:
4983 Sinix/ReliantUNIX RM400
4988 16-bit platforms such as WIN16
4992 * Clean up OPENSSL_NO_xxx #define's
4993 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4994 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4995 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4996 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4997 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4998 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4999 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5000 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5001 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5002 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5006 * Cleaned up dead code
5007 Remove all but one '#ifdef undef' which is to be looked at.
5011 * Clean up calling of xxx_free routines.
5012 Just like free(), fix most of the xxx_free routines to accept
5013 NULL. Remove the non-null checks from callers. Save much code.
5017 * Add secure heap for storage of private keys (when possible).
5018 Add BIO_s_secmem(), CBIGNUM, etc.
5019 Contributed by Akamai Technologies under our Corporate CLA.
5023 * Experimental support for a new, fast, unbiased prime candidate generator,
5024 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5026 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5028 * New output format NSS in the sess_id command line tool. This allows
5029 exporting the session id and the master key in NSS keylog format.
5031 *Martin Kaiser <martin@kaiser.cx>*
5033 * Harmonize version and its documentation. -f flag is used to display
5036 *mancha <mancha1@zoho.com>*
5038 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5039 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5041 *mancha <mancha1@zoho.com>*
5043 * Fix some double frees. These are not thought to be exploitable.
5045 *mancha <mancha1@zoho.com>*
5047 * A missing bounds check in the handling of the TLS heartbeat extension
5048 can be used to reveal up to 64k of memory to a connected client or
5051 Thanks for Neel Mehta of Google Security for discovering this bug and to
5052 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5053 preparing the fix ([CVE-2014-0160])
5055 *Adam Langley, Bodo Moeller*
5057 * Fix for the attack described in the paper "Recovering OpenSSL
5058 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5059 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5060 <http://eprint.iacr.org/2014/140>
5062 Thanks to Yuval Yarom and Naomi Benger for discovering this
5063 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5065 *Yuval Yarom and Naomi Benger*
5067 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5068 this fixes a limitation in previous versions of OpenSSL.
5072 * Experimental encrypt-then-mac support.
5074 Experimental support for encrypt then mac from
5075 draft-gutmann-tls-encrypt-then-mac-02.txt
5077 To enable it set the appropriate extension number (0x42 for the test
5078 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5080 For non-compliant peers (i.e. just about everything) this should have no
5083 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5087 * Add EVP support for key wrapping algorithms, to avoid problems with
5088 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5089 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5090 algorithms and include tests cases.
5094 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5099 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5100 MGF1 digest and OAEP label.
5104 * Make openssl verify return errors.
5106 *Chris Palmer <palmer@google.com> and Ben Laurie*
5108 * New function ASN1_TIME_diff to calculate the difference between two
5109 ASN1_TIME structures or one structure and the current time.
5113 * Update fips_test_suite to support multiple command line options. New
5114 test to induce all self test errors in sequence and check expected
5119 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5120 sign or verify all in one operation.
5124 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5125 test programs and fips_test_suite. Includes functionality to parse
5126 the minimal script output of fipsalgest.pl directly.
5130 * Add authorisation parameter to FIPS_module_mode_set().
5134 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5138 * Use separate DRBG fields for internal and external flags. New function
5139 FIPS_drbg_health_check() to perform on demand health checking. Add
5140 generation tests to fips_test_suite with reduced health check interval to
5141 demonstrate periodic health checking. Add "nodh" option to
5142 fips_test_suite to skip very slow DH test.
5146 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5151 * More extensive health check for DRBG checking many more failure modes.
5152 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5153 combination: call this in fips_test_suite.
5157 * Add support for canonical generation of DSA parameter 'g'. See
5160 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5161 POST to handle HMAC cases.
5165 * Add functions FIPS_module_version() and FIPS_module_version_text()
5166 to return numerical and string versions of the FIPS module number.
5170 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5171 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5172 outside the validated module in the FIPS capable OpenSSL.
5176 * Minor change to DRBG entropy callback semantics. In some cases
5177 there is no multiple of the block length between min_len and
5178 max_len. Allow the callback to return more than max_len bytes
5179 of entropy but discard any extra: it is the callback's responsibility
5180 to ensure that the extra data discarded does not impact the
5181 requested amount of entropy.
5185 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5186 information in FIPS186-3, SP800-57 and SP800-131A.
5190 * CCM support via EVP. Interface is very similar to GCM case except we
5191 must supply all data in one chunk (i.e. no update, final) and the
5192 message length must be supplied if AAD is used. Add algorithm test
5197 * Initial version of POST overhaul. Add POST callback to allow the status
5198 of POST to be monitored and/or failures induced. Modify fips_test_suite
5199 to use callback. Always run all selftests even if one fails.
5203 * XTS support including algorithm test driver in the fips_gcmtest program.
5204 Note: this does increase the maximum key length from 32 to 64 bytes but
5205 there should be no binary compatibility issues as existing applications
5206 will never use XTS mode.
5210 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5211 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5212 performs algorithm blocking for unapproved PRNG types. Also do not
5213 set PRNG type in FIPS_mode_set(): leave this to the application.
5214 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5215 the standard OpenSSL PRNG: set additional data to a date time vector.
5219 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5220 This shouldn't present any incompatibility problems because applications
5221 shouldn't be using these directly and any that are will need to rethink
5222 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5226 * Extensive self tests and health checking required by SP800-90 DRBG.
5227 Remove strength parameter from FIPS_drbg_instantiate and always
5228 instantiate at maximum supported strength.
5232 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5236 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5240 * New function DH_compute_key_padded() to compute a DH key and pad with
5241 leading zeroes if needed: this complies with SP800-56A et al.
5245 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5246 anything, incomplete, subject to change and largely untested at present.
5250 * Modify fipscanisteronly build option to only build the necessary object
5251 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5255 * Add experimental option FIPSSYMS to give all symbols in
5256 fipscanister.o and FIPS or fips prefix. This will avoid
5257 conflicts with future versions of OpenSSL. Add perl script
5258 util/fipsas.pl to preprocess assembly language source files
5259 and rename any affected symbols.
5263 * Add selftest checks and algorithm block of non-fips algorithms in
5264 FIPS mode. Remove DES2 from selftests.
5268 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5269 return internal method without any ENGINE dependencies. Add new
5270 tiny fips sign and verify functions.
5274 * New build option no-ec2m to disable characteristic 2 code.
5278 * New build option "fipscanisteronly". This only builds fipscanister.o
5279 and (currently) associated fips utilities. Uses the file Makefile.fips
5280 instead of Makefile.org as the prototype.
5284 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5285 Update fips_gcmtest to use IV generator.
5289 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5290 setting output buffer to NULL. The `*Final` function must be
5291 called although it will not retrieve any additional data. The tag
5292 can be set or retrieved with a ctrl. The IV length is by default 12
5293 bytes (96 bits) but can be set to an alternative value. If the IV
5294 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5299 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5300 underlying do_cipher function handles all cipher semantics itself
5301 including padding and finalisation. This is useful if (for example)
5302 an ENGINE cipher handles block padding itself. The behaviour of
5303 do_cipher is subtly changed if this flag is set: the return value
5304 is the number of characters written to the output buffer (zero is
5305 no longer an error code) or a negative error code. Also if the
5306 input buffer is NULL and length 0 finalisation should be performed.
5310 * If a candidate issuer certificate is already part of the constructed
5311 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5315 * Improve forward-security support: add functions
5317 void SSL_CTX_set_not_resumable_session_callback(
5318 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5319 void SSL_set_not_resumable_session_callback(
5320 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5322 for use by SSL/TLS servers; the callback function will be called whenever a
5323 new session is created, and gets to decide whether the session may be
5324 cached to make it resumable (return 0) or not (return 1). (As by the
5325 SSL/TLS protocol specifications, the session_id sent by the server will be
5326 empty to indicate that the session is not resumable; also, the server will
5327 not generate RFC 4507 (RFC 5077) session tickets.)
5329 A simple reasonable callback implementation is to return is_forward_secure.
5330 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5331 by the SSL/TLS server library, indicating whether it can provide forward
5334 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5336 * New -verify_name option in command line utilities to set verification
5341 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5342 Add CMAC pkey methods.
5346 * Experimental renegotiation in s_server -www mode. If the client
5347 browses /reneg connection is renegotiated. If /renegcert it is
5348 renegotiated requesting a certificate.
5352 * Add an "external" session cache for debugging purposes to s_server. This
5353 should help trace issues which normally are only apparent in deployed
5354 multi-process servers.
5358 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5359 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5360 BIO_set_cipher() and some obscure PEM functions were changed so they
5361 can now return an error. The RAND changes required a change to the
5362 RAND_METHOD structure.
5366 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5367 a gcc attribute to warn if the result of a function is ignored. This
5368 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5369 whose return value is often ignored.
5373 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5374 These allow SCTs (signed certificate timestamps) to be requested and
5375 validated when establishing a connection.
5377 *Rob Percival <robpercival@google.com>*
5382 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5384 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5385 used even when parsing explicit parameters, when loading a encoded key
5386 or calling `EC_GROUP_new_from_ecpkparameters()`/
5387 `EC_GROUP_new_from_ecparameters()`.
5388 This prevents bypass of security hardening and performance gains,
5389 especially for curves with specialized EC_METHODs.
5390 By default, if a key encoded with explicit parameters is loaded and later
5391 encoded, the output is still encoded with explicit parameters, even if
5392 internally a "named" EC_GROUP is used for computation.
5396 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5397 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5398 NULL. After this change, only the cofactor parameter can be NULL. It also
5399 does some minimal sanity checks on the passed order.
5404 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5405 An attack is simple, if the first CMS_recipientInfo is valid but the
5406 second CMS_recipientInfo is chosen ciphertext. If the second
5407 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5408 encryption key will be replaced by garbage, and the message cannot be
5409 decoded, but if the RSA decryption fails, the correct encryption key is
5410 used and the recipient will not notice the attack.
5411 As a work around for this potential attack the length of the decrypted
5412 key must be equal to the cipher default key length, in case the
5413 certificate is not given and all recipientInfo are tried out.
5414 The old behaviour can be re-enabled in the CMS code by setting the
5415 CMS_DEBUG_DECRYPT flag.
5420 * Document issue with installation paths in diverse Windows builds
5422 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5423 binaries and run-time config file.
5428 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5430 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5431 This changes the size when using the `genpkey` command when no size is given.
5432 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5433 generation commands to use 2048 bits by default.
5437 * Add FIPS support for Android Arm 64-bit
5439 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5440 Module in Version 2.0.10. For some reason, the corresponding target
5441 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5442 built with FIPS support on Android Arm 64-bit. This omission has been
5445 *Matthias St. Pierre*
5447 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5449 * 0-byte record padding oracle
5451 If an application encounters a fatal protocol error and then calls
5452 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5453 then OpenSSL can respond differently to the calling application if a 0 byte
5454 record is received with invalid padding compared to if a 0 byte record is
5455 received with an invalid MAC. If the application then behaves differently
5456 based on that in a way that is detectable to the remote peer, then this
5457 amounts to a padding oracle that could be used to decrypt data.
5459 In order for this to be exploitable "non-stitched" ciphersuites must be in
5460 use. Stitched ciphersuites are optimised implementations of certain
5461 commonly used ciphersuites. Also the application must call SSL_shutdown()
5462 twice even if a protocol error has occurred (applications should not do
5463 this but some do anyway).
5465 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5466 Aviram, with additional investigation by Steven Collison and Andrew
5467 Hourselt. It was reported to OpenSSL on 10th December 2018.
5472 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5476 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5478 * Microarchitecture timing vulnerability in ECC scalar multiplication
5480 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5481 shown to be vulnerable to a microarchitecture timing side channel attack.
5482 An attacker with sufficient access to mount local timing attacks during
5483 ECDSA signature generation could recover the private key.
5485 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5486 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5492 * Timing vulnerability in DSA signature generation
5494 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5495 timing side channel attack. An attacker could use variations in the signing
5496 algorithm to recover the private key.
5498 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5503 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5504 Module, accidentally introduced while backporting security fixes from the
5505 development branch and hindering the use of ECC in FIPS mode.
5509 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5511 * Client DoS due to large DH parameter
5513 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5514 malicious server can send a very large prime value to the client. This will
5515 cause the client to spend an unreasonably long period of time generating a
5516 key for this prime resulting in a hang until the client has finished. This
5517 could be exploited in a Denial Of Service attack.
5519 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5524 * Cache timing vulnerability in RSA Key Generation
5526 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5527 a cache timing side channel attack. An attacker with sufficient access to
5528 mount cache timing attacks during the RSA key generation process could
5529 recover the private key.
5531 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5532 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5537 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5538 parameter is no longer accepted, as it leads to a corrupt table. NULL
5539 pem_str is reserved for alias entries only.
5543 * Revert blinding in ECDSA sign and instead make problematic addition
5544 length-invariant. Switch even to fixed-length Montgomery multiplication.
5548 * Change generating and checking of primes so that the error rate of not
5549 being prime depends on the intended use based on the size of the input.
5550 For larger primes this will result in more rounds of Miller-Rabin.
5551 The maximal error rate for primes with more than 1080 bits is lowered
5554 *Kurt Roeckx, Annie Yousar*
5556 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5560 * Add blinding to ECDSA and DSA signatures to protect against side channel
5561 attacks discovered by Keegan Ryan (NCC Group).
5565 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5566 now allow empty (zero character) pass phrases.
5570 * Certificate time validation (X509_cmp_time) enforces stricter
5571 compliance with RFC 5280. Fractional seconds and timezone offsets
5572 are no longer allowed.
5576 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5578 * Constructed ASN.1 types with a recursive definition could exceed the stack
5580 Constructed ASN.1 types with a recursive definition (such as can be found
5581 in PKCS7) could eventually exceed the stack given malicious input with
5582 excessive recursion. This could result in a Denial Of Service attack. There
5583 are no such structures used within SSL/TLS that come from untrusted sources
5584 so this is considered safe.
5586 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5592 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5594 * Read/write after SSL object in error state
5596 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5597 mechanism. The intent was that if a fatal error occurred during a handshake
5598 then OpenSSL would move into the error state and would immediately fail if
5599 you attempted to continue the handshake. This works as designed for the
5600 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5601 SSL_connect()), however due to a bug it does not work correctly if
5602 SSL_read() or SSL_write() is called directly. In that scenario, if the
5603 handshake fails then a fatal error will be returned in the initial function
5604 call. If SSL_read()/SSL_write() is subsequently called by the application
5605 for the same SSL object then it will succeed and the data is passed without
5606 being decrypted/encrypted directly from the SSL/TLS record layer.
5608 In order to exploit this issue an application bug would have to be present
5609 that resulted in a call to SSL_read()/SSL_write() being issued after having
5610 already received a fatal error.
5612 This issue was reported to OpenSSL by David Benjamin (Google).
5617 * rsaz_1024_mul_avx2 overflow bug on x86_64
5619 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5620 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5621 Analysis suggests that attacks against RSA and DSA as a result of this
5622 defect would be very difficult to perform and are not believed likely.
5623 Attacks against DH1024 are considered just feasible, because most of the
5624 work necessary to deduce information about a private key may be performed
5625 offline. The amount of resources required for such an attack would be
5626 significant. However, for an attack on TLS to be meaningful, the server
5627 would have to share the DH1024 private key among multiple clients, which is
5628 no longer an option since CVE-2016-0701.
5630 This only affects processors that support the AVX2 but not ADX extensions
5631 like Intel Haswell (4th generation).
5633 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5634 was originally found via the OSS-Fuzz project.
5639 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5641 * bn_sqrx8x_internal carry bug on x86_64
5643 There is a carry propagating bug in the x86_64 Montgomery squaring
5644 procedure. No EC algorithms are affected. Analysis suggests that attacks
5645 against RSA and DSA as a result of this defect would be very difficult to
5646 perform and are not believed likely. Attacks against DH are considered just
5647 feasible (although very difficult) because most of the work necessary to
5648 deduce information about a private key may be performed offline. The amount
5649 of resources required for such an attack would be very significant and
5650 likely only accessible to a limited number of attackers. An attacker would
5651 additionally need online access to an unpatched system using the target
5652 private key in a scenario with persistent DH parameters and a private
5653 key that is shared between multiple clients.
5655 This only affects processors that support the BMI1, BMI2 and ADX extensions
5656 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5658 This issue was reported to OpenSSL by the OSS-Fuzz project.
5663 * Malformed X.509 IPAddressFamily could cause OOB read
5665 If an X.509 certificate has a malformed IPAddressFamily extension,
5666 OpenSSL could do a one-byte buffer overread. The most likely result
5667 would be an erroneous display of the certificate in text format.
5669 This issue was reported to OpenSSL by the OSS-Fuzz project.
5673 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5675 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5676 platform rather than 'mingw'.
5680 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5682 * Truncated packet could crash via OOB read
5684 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5685 cipher is being used, then a truncated packet can cause that host to
5686 perform an out-of-bounds read, usually resulting in a crash.
5688 This issue was reported to OpenSSL by Robert Święcki of Google.
5693 * BN_mod_exp may produce incorrect results on x86_64
5695 There is a carry propagating bug in the x86_64 Montgomery squaring
5696 procedure. No EC algorithms are affected. Analysis suggests that attacks
5697 against RSA and DSA as a result of this defect would be very difficult to
5698 perform and are not believed likely. Attacks against DH are considered just
5699 feasible (although very difficult) because most of the work necessary to
5700 deduce information about a private key may be performed offline. The amount
5701 of resources required for such an attack would be very significant and
5702 likely only accessible to a limited number of attackers. An attacker would
5703 additionally need online access to an unpatched system using the target
5704 private key in a scenario with persistent DH parameters and a private
5705 key that is shared between multiple clients. For example this can occur by
5706 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5707 similar to CVE-2015-3193 but must be treated as a separate problem.
5709 This issue was reported to OpenSSL by the OSS-Fuzz project.
5714 * Montgomery multiplication may produce incorrect results
5716 There is a carry propagating bug in the Broadwell-specific Montgomery
5717 multiplication procedure that handles input lengths divisible by, but
5718 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5719 and DH private keys are impossible. This is because the subroutine in
5720 question is not used in operations with the private key itself and an input
5721 of the attacker's direct choice. Otherwise the bug can manifest itself as
5722 transient authentication and key negotiation failures or reproducible
5723 erroneous outcome of public-key operations with specially crafted input.
5724 Among EC algorithms only Brainpool P-512 curves are affected and one
5725 presumably can attack ECDH key negotiation. Impact was not analyzed in
5726 detail, because pre-requisites for attack are considered unlikely. Namely
5727 multiple clients have to choose the curve in question and the server has to
5728 share the private key among them, neither of which is default behaviour.
5729 Even then only clients that chose the curve will be affected.
5731 This issue was publicly reported as transient failures and was not
5732 initially recognized as a security issue. Thanks to Richard Morgan for
5733 providing reproducible case.
5738 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5739 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5740 prevent issues where no progress is being made and the peer continually
5741 sends unrecognised record types, using up resources processing them.
5745 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5747 * Missing CRL sanity check
5749 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5750 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5751 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5753 This issue only affects the OpenSSL 1.0.2i
5758 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5760 * OCSP Status Request extension unbounded memory growth
5762 A malicious client can send an excessively large OCSP Status Request
5763 extension. If that client continually requests renegotiation, sending a
5764 large OCSP Status Request extension each time, then there will be unbounded
5765 memory growth on the server. This will eventually lead to a Denial Of
5766 Service attack through memory exhaustion. Servers with a default
5767 configuration are vulnerable even if they do not support OCSP. Builds using
5768 the "no-ocsp" build time option are not affected.
5770 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5775 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5778 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5784 * OOB write in MDC2_Update()
5786 An overflow can occur in MDC2_Update() either if called directly or
5787 through the EVP_DigestUpdate() function using MDC2. If an attacker
5788 is able to supply very large amounts of input data after a previous
5789 call to EVP_EncryptUpdate() with a partial block then a length check
5790 can overflow resulting in a heap corruption.
5792 The amount of data needed is comparable to SIZE_MAX which is impractical
5795 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5800 * Malformed SHA512 ticket DoS
5802 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5803 DoS attack where a malformed ticket will result in an OOB read which will
5806 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5807 a custom server callback and ticket lookup mechanism.
5809 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5814 * OOB write in BN_bn2dec()
5816 The function BN_bn2dec() does not check the return value of BN_div_word().
5817 This can cause an OOB write if an application uses this function with an
5818 overly large BIGNUM. This could be a problem if an overly large certificate
5819 or CRL is printed out from an untrusted source. TLS is not affected because
5820 record limits will reject an oversized certificate before it is parsed.
5822 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5827 * OOB read in TS_OBJ_print_bio()
5829 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5830 the total length the OID text representation would use and not the amount
5831 of data written. This will result in OOB reads when large OIDs are
5834 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5839 * Pointer arithmetic undefined behaviour
5841 Avoid some undefined pointer arithmetic
5843 A common idiom in the codebase is to check limits in the following manner:
5846 Where "p" points to some malloc'd data of SIZE bytes and
5849 "len" here could be from some externally supplied data (e.g. from a TLS
5852 The rules of C pointer arithmetic are such that "p + len" is only well
5853 defined where len <= SIZE. Therefore the above idiom is actually
5854 undefined behaviour.
5856 For example this could cause problems if some malloc implementation
5857 provides an address for "p" such that "p + len" actually overflows for
5858 values of len that are too big and therefore p + len < limit.
5860 This issue was reported to OpenSSL by Guido Vranken
5865 * Constant time flag not preserved in DSA signing
5867 Operations in the DSA signing algorithm should run in constant time in
5868 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5869 implementation means that a non-constant time codepath is followed for
5870 certain operations. This has been demonstrated through a cache-timing
5871 attack to be sufficient for an attacker to recover the private DSA key.
5873 This issue was reported by César Pereida (Aalto University), Billy Brumley
5874 (Tampere University of Technology), and Yuval Yarom (The University of
5875 Adelaide and NICTA).
5880 * DTLS buffered message DoS
5882 In a DTLS connection where handshake messages are delivered out-of-order
5883 those messages that OpenSSL is not yet ready to process will be buffered
5884 for later use. Under certain circumstances, a flaw in the logic means that
5885 those messages do not get removed from the buffer even though the handshake
5886 has been completed. An attacker could force up to approx. 15 messages to
5887 remain in the buffer when they are no longer required. These messages will
5888 be cleared when the DTLS connection is closed. The default maximum size for
5889 a message is 100k. Therefore, the attacker could force an additional 1500k
5890 to be consumed per connection. By opening many simultaneous connections an
5891 attacker could cause a DoS attack through memory exhaustion.
5893 This issue was reported to OpenSSL by Quan Luo.
5898 * DTLS replay protection DoS
5900 A flaw in the DTLS replay attack protection mechanism means that records
5901 that arrive for future epochs update the replay protection "window" before
5902 the MAC for the record has been validated. This could be exploited by an
5903 attacker by sending a record for the next epoch (which does not have to
5904 decrypt or have a valid MAC), with a very large sequence number. This means
5905 that all subsequent legitimate packets are dropped causing a denial of
5906 service for a specific DTLS connection.
5908 This issue was reported to OpenSSL by the OCAP audit team.
5913 * Certificate message OOB reads
5915 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5916 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5917 theoretical DoS risk but this has not been observed in practice on common
5920 The messages affected are client certificate, client certificate request
5921 and server certificate. As a result the attack can only be performed
5922 against a client or a server which enables client authentication.
5924 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5929 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5931 * Prevent padding oracle in AES-NI CBC MAC check
5933 A MITM attacker can use a padding oracle attack to decrypt traffic
5934 when the connection uses an AES CBC cipher and the server support
5937 This issue was introduced as part of the fix for Lucky 13 padding
5938 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5939 constant time by making sure that always the same bytes are read and
5940 compared against either the MAC or padding bytes. But it no longer
5941 checked that there was enough data to have both the MAC and padding
5944 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5948 * Fix EVP_EncodeUpdate overflow
5950 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5951 Base64 encoding of binary data. If an attacker is able to supply very large
5952 amounts of input data then a length check can overflow resulting in a heap
5955 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5956 the `PEM_write_bio*` family of functions. These are mainly used within the
5957 OpenSSL command line applications, so any application which processes data
5958 from an untrusted source and outputs it as a PEM file should be considered
5959 vulnerable to this issue. User applications that call these APIs directly
5960 with large amounts of untrusted data may also be vulnerable.
5962 This issue was reported by Guido Vranken.
5967 * Fix EVP_EncryptUpdate overflow
5969 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5970 is able to supply very large amounts of input data after a previous call to
5971 EVP_EncryptUpdate() with a partial block then a length check can overflow
5972 resulting in a heap corruption. Following an analysis of all OpenSSL
5973 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5974 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5975 the first called function after an EVP_EncryptInit(), and therefore that
5976 specific call must be safe. The second form is where the length passed to
5977 EVP_EncryptUpdate() can be seen from the code to be some small value and
5978 therefore there is no possibility of an overflow. Since all instances are
5979 one of these two forms, it is believed that there can be no overflows in
5980 internal code due to this problem. It should be noted that
5981 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5982 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5983 of these calls have also been analysed too and it is believed there are no
5984 instances in internal usage where an overflow could occur.
5986 This issue was reported by Guido Vranken.
5991 * Prevent ASN.1 BIO excessive memory allocation
5993 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5994 a short invalid encoding can cause allocation of large amounts of memory
5995 potentially consuming excessive resources or exhausting memory.
5997 Any application parsing untrusted data through d2i BIO functions is
5998 affected. The memory based functions such as d2i_X509() are *not* affected.
5999 Since the memory based functions are used by the TLS library, TLS
6000 applications are not affected.
6002 This issue was reported by Brian Carpenter.
6009 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6010 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6011 in arbitrary stack data being returned in the buffer.
6013 This issue was reported by Guido Vranken.
6018 * Modify behavior of ALPN to invoke callback after SNI/servername
6019 callback, such that updates to the SSL_CTX affect ALPN.
6023 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6028 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6029 methods are enabled and ssl2 is disabled the methods return NULL.
6033 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6035 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6036 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6037 provide any "EXPORT" or "LOW" strength ciphers.
6041 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6042 is by default disabled at build-time. Builds that are not configured with
6043 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6044 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6045 will need to explicitly call either of:
6047 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6049 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6051 as appropriate. Even if either of those is used, or the application
6052 explicitly uses the version-specific SSLv2_method() or its client and
6053 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6054 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6055 ciphers, and SSLv2 56-bit DES are no longer available.
6060 * Fix a double-free in DSA code
6062 A double free bug was discovered when OpenSSL parses malformed DSA private
6063 keys and could lead to a DoS attack or memory corruption for applications
6064 that receive DSA private keys from untrusted sources. This scenario is
6067 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6073 * Disable SRP fake user seed to address a server memory leak.
6075 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6077 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6078 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6079 was changed to ignore the "fake user" SRP seed, even if the seed
6082 Users should use SRP_VBASE_get1_by_user instead. Note that in
6083 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6084 also that even though configuring the SRP seed attempts to hide
6085 invalid usernames by continuing the handshake with fake
6086 credentials, this behaviour is not constant time and no strong
6087 guarantees are made that the handshake is indistinguishable from
6088 that of a valid user.
6093 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6095 In the BN_hex2bn function the number of hex digits is calculated using an
6096 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6097 large values of `i` this can result in `bn_expand` not allocating any
6098 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6099 field as NULL leading to a subsequent NULL ptr deref. For very large values
6100 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6101 In this case memory is allocated to the internal BIGNUM data field, but it
6102 is insufficiently sized leading to heap corruption. A similar issue exists
6103 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6104 is ever called by user applications with very large untrusted hex/dec data.
6105 This is anticipated to be a rare occurrence.
6107 All OpenSSL internal usage of these functions use data that is not expected
6108 to be untrusted, e.g. config file data or application command line
6109 arguments. If user developed applications generate config file data based
6110 on untrusted data then it is possible that this could also lead to security
6111 consequences. This is also anticipated to be rare.
6113 This issue was reported to OpenSSL by Guido Vranken.
6118 * Fix memory issues in `BIO_*printf` functions
6120 The internal `fmtstr` function used in processing a "%s" format string in
6121 the `BIO_*printf` functions could overflow while calculating the length of a
6122 string and cause an OOB read when printing very long strings.
6124 Additionally the internal `doapr_outch` function can attempt to write to an
6125 OOB memory location (at an offset from the NULL pointer) in the event of a
6126 memory allocation failure. In 1.0.2 and below this could be caused where
6127 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6128 could be in processing a very long "%s" format string. Memory leaks can
6131 The first issue may mask the second issue dependent on compiler behaviour.
6132 These problems could enable attacks where large amounts of untrusted data
6133 is passed to the `BIO_*printf` functions. If applications use these functions
6134 in this way then they could be vulnerable. OpenSSL itself uses these
6135 functions when printing out human-readable dumps of ASN.1 data. Therefore
6136 applications that print this data could be vulnerable if the data is from
6137 untrusted sources. OpenSSL command line applications could also be
6138 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6139 as command line arguments.
6141 Libssl is not considered directly vulnerable. Additionally certificates etc
6142 received via remote connections via libssl are also unlikely to be able to
6143 trigger these issues because of message size limits enforced within libssl.
6145 This issue was reported to OpenSSL Guido Vranken.
6150 * Side channel attack on modular exponentiation
6152 A side-channel attack was found which makes use of cache-bank conflicts on
6153 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6154 of RSA keys. The ability to exploit this issue is limited as it relies on
6155 an attacker who has control of code in a thread running on the same
6156 hyper-threaded core as the victim thread which is performing decryptions.
6158 This issue was reported to OpenSSL by Yuval Yarom, The University of
6159 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6160 Nadia Heninger, University of Pennsylvania with more information at
6161 <http://cachebleed.info>.
6166 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6167 if no keysize is specified with default_bits. This fixes an
6168 omission in an earlier change that changed all RSA/DSA key generation
6169 commands to use 2048 bits by default.
6173 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6175 * DH small subgroups
6177 Historically OpenSSL only ever generated DH parameters based on "safe"
6178 primes. More recently (in version 1.0.2) support was provided for
6179 generating X9.42 style parameter files such as those required for RFC 5114
6180 support. The primes used in such files may not be "safe". Where an
6181 application is using DH configured with parameters based on primes that are
6182 not "safe" then an attacker could use this fact to find a peer's private
6183 DH exponent. This attack requires that the attacker complete multiple
6184 handshakes in which the peer uses the same private DH exponent. For example
6185 this could be used to discover a TLS server's private DH exponent if it's
6186 reusing the private DH exponent or it's using a static DH ciphersuite.
6188 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6189 TLS. It is not on by default. If the option is not set then the server
6190 reuses the same private DH exponent for the life of the server process and
6191 would be vulnerable to this attack. It is believed that many popular
6192 applications do set this option and would therefore not be at risk.
6194 The fix for this issue adds an additional check where a "q" parameter is
6195 available (as is the case in X9.42 based parameters). This detects the
6196 only known attack, and is the only possible defense for static DH
6197 ciphersuites. This could have some performance impact.
6199 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6200 default and cannot be disabled. This could have some performance impact.
6202 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6207 * SSLv2 doesn't block disabled ciphers
6209 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6210 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6211 been disabled, provided that the SSLv2 protocol was not also disabled via
6214 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6215 and Sebastian Schinzel.
6220 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6222 * BN_mod_exp may produce incorrect results on x86_64
6224 There is a carry propagating bug in the x86_64 Montgomery squaring
6225 procedure. No EC algorithms are affected. Analysis suggests that attacks
6226 against RSA and DSA as a result of this defect would be very difficult to
6227 perform and are not believed likely. Attacks against DH are considered just
6228 feasible (although very difficult) because most of the work necessary to
6229 deduce information about a private key may be performed offline. The amount
6230 of resources required for such an attack would be very significant and
6231 likely only accessible to a limited number of attackers. An attacker would
6232 additionally need online access to an unpatched system using the target
6233 private key in a scenario with persistent DH parameters and a private
6234 key that is shared between multiple clients. For example this can occur by
6235 default in OpenSSL DHE based SSL/TLS ciphersuites.
6237 This issue was reported to OpenSSL by Hanno Böck.
6242 * Certificate verify crash with missing PSS parameter
6244 The signature verification routines will crash with a NULL pointer
6245 dereference if presented with an ASN.1 signature using the RSA PSS
6246 algorithm and absent mask generation function parameter. Since these
6247 routines are used to verify certificate signature algorithms this can be
6248 used to crash any certificate verification operation and exploited in a
6249 DoS attack. Any application which performs certificate verification is
6250 vulnerable including OpenSSL clients and servers which enable client
6253 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6258 * X509_ATTRIBUTE memory leak
6260 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6261 memory. This structure is used by the PKCS#7 and CMS routines so any
6262 application which reads PKCS#7 or CMS data from untrusted sources is
6263 affected. SSL/TLS is not affected.
6265 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6271 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6272 This changes the decoding behaviour for some invalid messages,
6273 though the change is mostly in the more lenient direction, and
6274 legacy behaviour is preserved as much as possible.
6278 * In DSA_generate_parameters_ex, if the provided seed is too short,
6281 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6283 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6285 * Alternate chains certificate forgery
6287 During certificate verification, OpenSSL will attempt to find an
6288 alternative certificate chain if the first attempt to build such a chain
6289 fails. An error in the implementation of this logic can mean that an
6290 attacker could cause certain checks on untrusted certificates to be
6291 bypassed, such as the CA flag, enabling them to use a valid leaf
6292 certificate to act as a CA and "issue" an invalid certificate.
6294 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6299 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6301 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6302 incompatibility in the handling of HMAC. The previous ABI has now been
6307 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6309 * Malformed ECParameters causes infinite loop
6311 When processing an ECParameters structure OpenSSL enters an infinite loop
6312 if the curve specified is over a specially malformed binary polynomial
6315 This can be used to perform denial of service against any
6316 system which processes public keys, certificate requests or
6317 certificates. This includes TLS clients and TLS servers with
6318 client authentication enabled.
6320 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6325 * Exploitable out-of-bounds read in X509_cmp_time
6327 X509_cmp_time does not properly check the length of the ASN1_TIME
6328 string and can read a few bytes out of bounds. In addition,
6329 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6332 An attacker can use this to craft malformed certificates and CRLs of
6333 various sizes and potentially cause a segmentation fault, resulting in
6334 a DoS on applications that verify certificates or CRLs. TLS clients
6335 that verify CRLs are affected. TLS clients and servers with client
6336 authentication enabled may be affected if they use custom verification
6339 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6340 independently by Hanno Böck.
6345 * PKCS7 crash with missing EnvelopedContent
6347 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6348 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6349 with missing content and trigger a NULL pointer dereference on parsing.
6351 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6352 structures from untrusted sources are affected. OpenSSL clients and
6353 servers are not affected.
6355 This issue was reported to OpenSSL by Michal Zalewski (Google).
6360 * CMS verify infinite loop with unknown hash function
6362 When verifying a signedData message the CMS code can enter an infinite loop
6363 if presented with an unknown hash function OID. This can be used to perform
6364 denial of service against any system which verifies signedData messages using
6366 This issue was reported to OpenSSL by Johannes Bauer.
6371 * Race condition handling NewSessionTicket
6373 If a NewSessionTicket is received by a multi-threaded client when attempting to
6374 reuse a previous ticket then a race condition can occur potentially leading to
6375 a double free of the ticket data.
6380 * Only support 256-bit or stronger elliptic curves with the
6381 'ecdh_auto' setting (server) or by default (client). Of supported
6382 curves, prefer P-256 (both).
6386 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6388 * ClientHello sigalgs DoS fix
6390 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6391 invalid signature algorithms extension a NULL pointer dereference will
6392 occur. This can be exploited in a DoS attack against the server.
6394 This issue was was reported to OpenSSL by David Ramos of Stanford
6398 *Stephen Henson and Matt Caswell*
6400 * Multiblock corrupted pointer fix
6402 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6403 feature only applies on 64 bit x86 architecture platforms that support AES
6404 NI instructions. A defect in the implementation of "multiblock" can cause
6405 OpenSSL's internal write buffer to become incorrectly set to NULL when
6406 using non-blocking IO. Typically, when the user application is using a
6407 socket BIO for writing, this will only result in a failed connection.
6408 However if some other BIO is used then it is likely that a segmentation
6409 fault will be triggered, thus enabling a potential DoS attack.
6411 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6416 * Segmentation fault in DTLSv1_listen fix
6418 The DTLSv1_listen function is intended to be stateless and processes the
6419 initial ClientHello from many peers. It is common for user code to loop
6420 over the call to DTLSv1_listen until a valid ClientHello is received with
6421 an associated cookie. A defect in the implementation of DTLSv1_listen means
6422 that state is preserved in the SSL object from one invocation to the next
6423 that can lead to a segmentation fault. Errors processing the initial
6424 ClientHello can trigger this scenario. An example of such an error could be
6425 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6428 This issue was reported to OpenSSL by Per Allansson.
6433 * Segmentation fault in ASN1_TYPE_cmp fix
6435 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6436 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6437 certificate signature algorithm consistency this can be used to crash any
6438 certificate verification operation and exploited in a DoS attack. Any
6439 application which performs certificate verification is vulnerable including
6440 OpenSSL clients and servers which enable client authentication.
6445 * Segmentation fault for invalid PSS parameters fix
6447 The signature verification routines will crash with a NULL pointer
6448 dereference if presented with an ASN.1 signature using the RSA PSS
6449 algorithm and invalid parameters. Since these routines are used to verify
6450 certificate signature algorithms this can be used to crash any
6451 certificate verification operation and exploited in a DoS attack. Any
6452 application which performs certificate verification is vulnerable including
6453 OpenSSL clients and servers which enable client authentication.
6455 This issue was was reported to OpenSSL by Brian Carpenter.
6460 * ASN.1 structure reuse memory corruption fix
6462 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6463 memory corruption via an invalid write. Such reuse is and has been
6464 strongly discouraged and is believed to be rare.
6466 Applications that parse structures containing CHOICE or ANY DEFINED BY
6467 components may be affected. Certificate parsing (d2i_X509 and related
6468 functions) are however not affected. OpenSSL clients and servers are
6474 * PKCS7 NULL pointer dereferences fix
6476 The PKCS#7 parsing code does not handle missing outer ContentInfo
6477 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6478 missing content and trigger a NULL pointer dereference on parsing.
6480 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6481 otherwise parse PKCS#7 structures from untrusted sources are
6482 affected. OpenSSL clients and servers are not affected.
6484 This issue was reported to OpenSSL by Michal Zalewski (Google).
6489 * DoS via reachable assert in SSLv2 servers fix
6491 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6492 servers that both support SSLv2 and enable export cipher suites by sending
6493 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6495 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6496 (OpenSSL development team).
6501 * Empty CKE with client auth and DHE fix
6503 If client auth is used then a server can seg fault in the event of a DHE
6504 ciphersuite being selected and a zero length ClientKeyExchange message
6505 being sent by the client. This could be exploited in a DoS attack.
6510 * Handshake with unseeded PRNG fix
6512 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6513 with an unseeded PRNG. The conditions are:
6514 - The client is on a platform where the PRNG has not been seeded
6515 automatically, and the user has not seeded manually
6516 - A protocol specific client method version has been used (i.e. not
6517 SSL_client_methodv23)
6518 - A ciphersuite is used that does not require additional random data from
6519 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6521 If the handshake succeeds then the client random that has been used will
6522 have been generated from a PRNG with insufficient entropy and therefore the
6523 output may be predictable.
6525 For example using the following command with an unseeded openssl will
6526 succeed on an unpatched platform:
6528 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6533 * Use After Free following d2i_ECPrivatekey error fix
6535 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6536 could cause a use after free condition. This, in turn, could cause a double
6537 free in several private key parsing functions (such as d2i_PrivateKey
6538 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6539 for applications that receive EC private keys from untrusted
6540 sources. This scenario is considered rare.
6542 This issue was discovered by the BoringSSL project and fixed in their
6548 * X509_to_X509_REQ NULL pointer deref fix
6550 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6551 the certificate key is invalid. This function is rarely used in practice.
6553 This issue was discovered by Brian Carpenter.
6558 * Removed the export ciphers from the DEFAULT ciphers
6562 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6564 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6565 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6566 So far those who have to target multiple platforms would compromise
6567 and argue that binary targeting say ARMv5 would still execute on
6568 ARMv8. "Universal" build resolves this compromise by providing
6569 near-optimal performance even on newer platforms.
6573 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6574 (other platforms pending).
6576 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6578 * Add support for the SignedCertificateTimestampList certificate and
6579 OCSP response extensions from RFC6962.
6583 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6584 for corner cases. (Certain input points at infinity could lead to
6585 bogus results, with non-infinity inputs mapped to infinity too.)
6589 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6590 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6591 common cases are optimized and there still is room for further
6592 improvements. Vector Permutation AES for Altivec is also added.
6596 * Add support for little-endian ppc64 Linux target.
6598 *Marcelo Cerri (IBM)*
6600 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6601 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6602 are optimized and there still is room for further improvements.
6603 Both 32- and 64-bit modes are supported.
6605 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6607 * Improved ARMv7 NEON support.
6611 * Support for SPARC Architecture 2011 crypto extensions, first
6612 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6613 SHA256/512, MD5, GHASH and modular exponentiation.
6615 *Andy Polyakov, David Miller*
6617 * Accelerated modular exponentiation for Intel processors, a.k.a.
6620 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6622 * Support for new and upcoming Intel processors, including AVX2,
6623 BMI and SHA ISA extensions. This includes additional "stitched"
6624 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6627 This work was sponsored by Intel Corp.
6631 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6632 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6633 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6637 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6638 this fixes a limitation in previous versions of OpenSSL.
6642 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6643 MGF1 digest and OAEP label.
6647 * Add EVP support for key wrapping algorithms, to avoid problems with
6648 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6649 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6650 algorithms and include tests cases.
6654 * Add functions to allocate and set the fields of an ECDSA_METHOD
6657 *Douglas E. Engert, Steve Henson*
6659 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6660 difference in days and seconds between two tm or ASN1_TIME structures.
6664 * Add -rev test option to s_server to just reverse order of characters
6665 received by client and send back to server. Also prints an abbreviated
6666 summary of the connection parameters.
6670 * New option -brief for s_client and s_server to print out a brief summary
6671 of connection parameters.
6675 * Add callbacks for arbitrary TLS extensions.
6677 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6679 * New option -crl_download in several openssl utilities to download CRLs
6680 from CRLDP extension in certificates.
6684 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6688 * New function X509_CRL_diff to generate a delta CRL from the difference
6689 of two full CRLs. Add support to "crl" utility.
6693 * New functions to set lookup_crls function and to retrieve
6694 X509_STORE from X509_STORE_CTX.
6698 * Print out deprecated issuer and subject unique ID fields in
6703 * Extend OCSP I/O functions so they can be used for simple general purpose
6704 HTTP as well as OCSP. New wrapper function which can be used to download
6705 CRLs using the OCSP API.
6709 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6713 * `SSL_CONF*` functions. These provide a common framework for application
6714 configuration using configuration files or command lines.
6718 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6719 message callback and prints the results. Needs compile time option
6720 "enable-ssl-trace". New options to s_client and s_server to enable
6725 * New ctrl and macro to retrieve supported points extensions.
6726 Print out extension in s_server and s_client.
6730 * New functions to retrieve certificate signature and signature
6735 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6740 * New Suite B modes for TLS code. These use and enforce the requirements
6741 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6742 only use Suite B curves. The Suite B modes can be set by using the
6743 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6747 * New chain verification flags for Suite B levels of security. Check
6748 algorithms are acceptable when flags are set in X509_verify_cert.
6752 * Make tls1_check_chain return a set of flags indicating checks passed
6753 by a certificate chain. Add additional tests to handle client
6754 certificates: checks for matching certificate type and issuer name
6759 * If an attempt is made to use a signature algorithm not in the peer
6760 preference list abort the handshake. If client has no suitable
6761 signature algorithms in response to a certificate request do not
6762 use the certificate.
6766 * If server EC tmp key is not in client preference list abort handshake.
6770 * Add support for certificate stores in CERT structure. This makes it
6771 possible to have different stores per SSL structure or one store in
6772 the parent SSL_CTX. Include distinct stores for certificate chain
6773 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6774 to build and store a certificate chain in CERT structure: returning
6775 an error if the chain cannot be built: this will allow applications
6776 to test if a chain is correctly configured.
6778 Note: if the CERT based stores are not set then the parent SSL_CTX
6779 store is used to retain compatibility with existing behaviour.
6783 * New function ssl_set_client_disabled to set a ciphersuite disabled
6784 mask based on the current session, check mask when sending client
6785 hello and checking the requested ciphersuite.
6789 * New ctrls to retrieve and set certificate types in a certificate
6790 request message. Print out received values in s_client. If certificate
6791 types is not set with custom values set sensible values based on
6792 supported signature algorithms.
6796 * Support for distinct client and server supported signature algorithms.
6800 * Add certificate callback. If set this is called whenever a certificate
6801 is required by client or server. An application can decide which
6802 certificate chain to present based on arbitrary criteria: for example
6803 supported signature algorithms. Add very simple example to s_server.
6804 This fixes many of the problems and restrictions of the existing client
6805 certificate callback: for example you can now clear an existing
6806 certificate and specify the whole chain.
6810 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6811 the certificate can be used for (if anything). Set valid_flags field
6812 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6813 to have similar checks in it.
6815 Add new "cert_flags" field to CERT structure and include a "strict mode".
6816 This enforces some TLS certificate requirements (such as only permitting
6817 certificate signature algorithms contained in the supported algorithms
6818 extension) which some implementations ignore: this option should be used
6819 with caution as it could cause interoperability issues.
6823 * Update and tidy signature algorithm extension processing. Work out
6824 shared signature algorithms based on preferences and peer algorithms
6825 and print them out in s_client and s_server. Abort handshake if no
6826 shared signature algorithms.
6830 * Add new functions to allow customised supported signature algorithms
6831 for SSL and SSL_CTX structures. Add options to s_client and s_server
6836 * New function SSL_certs_clear() to delete all references to certificates
6837 from an SSL structure. Before this once a certificate had been added
6838 it couldn't be removed.
6842 * Integrate hostname, email address and IP address checking with certificate
6843 verification. New verify options supporting checking in openssl utility.
6847 * Fixes and wildcard matching support to hostname and email checking
6848 functions. Add manual page.
6850 *Florian Weimer (Red Hat Product Security Team)*
6852 * New functions to check a hostname email or IP address against a
6853 certificate. Add options x509 utility to print results of checks against
6858 * Fix OCSP checking.
6860 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6862 * Initial experimental support for explicitly trusted non-root CAs.
6863 OpenSSL still tries to build a complete chain to a root but if an
6864 intermediate CA has a trust setting included that is used. The first
6865 setting is used: whether to trust (e.g., -addtrust option to the x509
6870 * Add -trusted_first option which attempts to find certificates in the
6871 trusted store even if an untrusted chain is also supplied.
6875 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6876 platform support for Linux and Android.
6880 * Support for linux-x32, ILP32 environment in x86_64 framework.
6884 * Experimental multi-implementation support for FIPS capable OpenSSL.
6885 When in FIPS mode the approved implementations are used as normal,
6886 when not in FIPS mode the internal unapproved versions are used instead.
6887 This means that the FIPS capable OpenSSL isn't forced to use the
6888 (often lower performance) FIPS implementations outside FIPS mode.
6892 * Transparently support X9.42 DH parameters when calling
6893 PEM_read_bio_DHparameters. This means existing applications can handle
6894 the new parameter format automatically.
6898 * Initial experimental support for X9.42 DH parameter format: mainly
6899 to support use of 'q' parameter for RFC5114 parameters.
6903 * Add DH parameters from RFC5114 including test data to dhtest.
6907 * Support for automatic EC temporary key parameter selection. If enabled
6908 the most preferred EC parameters are automatically used instead of
6909 hardcoded fixed parameters. Now a server just has to call:
6910 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6911 support ECDH and use the most appropriate parameters.
6915 * Enhance and tidy EC curve and point format TLS extension code. Use
6916 static structures instead of allocation if default values are used.
6917 New ctrls to set curves we wish to support and to retrieve shared curves.
6918 Print out shared curves in s_server. New options to s_server and s_client
6919 to set list of supported curves.
6923 * New ctrls to retrieve supported signature algorithms and
6924 supported curve values as an array of NIDs. Extend openssl utility
6925 to print out received values.
6929 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6930 between NIDs and the more common NIST names such as "P-256". Enhance
6931 ecparam utility and ECC method to recognise the NIST names for curves.
6935 * Enhance SSL/TLS certificate chain handling to support different
6936 chains for each certificate instead of one chain in the parent SSL_CTX.
6940 * Support for fixed DH ciphersuite client authentication: where both
6941 server and client use DH certificates with common parameters.
6945 * Support for fixed DH ciphersuites: those requiring DH server
6950 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6952 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6953 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6954 X509_CINF_get_signature were reverted post internal team review.
6959 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6961 * OCSP Status Request extension unbounded memory growth
6963 A malicious client can send an excessively large OCSP Status Request
6964 extension. If that client continually requests renegotiation, sending a
6965 large OCSP Status Request extension each time, then there will be unbounded
6966 memory growth on the server. This will eventually lead to a Denial Of
6967 Service attack through memory exhaustion. Servers with a default
6968 configuration are vulnerable even if they do not support OCSP. Builds using
6969 the "no-ocsp" build time option are not affected.
6971 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6976 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6979 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6985 * OOB write in MDC2_Update()
6987 An overflow can occur in MDC2_Update() either if called directly or
6988 through the EVP_DigestUpdate() function using MDC2. If an attacker
6989 is able to supply very large amounts of input data after a previous
6990 call to EVP_EncryptUpdate() with a partial block then a length check
6991 can overflow resulting in a heap corruption.
6993 The amount of data needed is comparable to SIZE_MAX which is impractical
6996 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7001 * Malformed SHA512 ticket DoS
7003 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7004 DoS attack where a malformed ticket will result in an OOB read which will
7007 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7008 a custom server callback and ticket lookup mechanism.
7010 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7015 * OOB write in BN_bn2dec()
7017 The function BN_bn2dec() does not check the return value of BN_div_word().
7018 This can cause an OOB write if an application uses this function with an
7019 overly large BIGNUM. This could be a problem if an overly large certificate
7020 or CRL is printed out from an untrusted source. TLS is not affected because
7021 record limits will reject an oversized certificate before it is parsed.
7023 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7028 * OOB read in TS_OBJ_print_bio()
7030 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7031 the total length the OID text representation would use and not the amount
7032 of data written. This will result in OOB reads when large OIDs are
7035 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7040 * Pointer arithmetic undefined behaviour
7042 Avoid some undefined pointer arithmetic
7044 A common idiom in the codebase is to check limits in the following manner:
7047 Where "p" points to some malloc'd data of SIZE bytes and
7050 "len" here could be from some externally supplied data (e.g. from a TLS
7053 The rules of C pointer arithmetic are such that "p + len" is only well
7054 defined where len <= SIZE. Therefore, the above idiom is actually
7055 undefined behaviour.
7057 For example this could cause problems if some malloc implementation
7058 provides an address for "p" such that "p + len" actually overflows for
7059 values of len that are too big and therefore p + len < limit.
7061 This issue was reported to OpenSSL by Guido Vranken
7066 * Constant time flag not preserved in DSA signing
7068 Operations in the DSA signing algorithm should run in constant time in
7069 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7070 implementation means that a non-constant time codepath is followed for
7071 certain operations. This has been demonstrated through a cache-timing
7072 attack to be sufficient for an attacker to recover the private DSA key.
7074 This issue was reported by César Pereida (Aalto University), Billy Brumley
7075 (Tampere University of Technology), and Yuval Yarom (The University of
7076 Adelaide and NICTA).
7081 * DTLS buffered message DoS
7083 In a DTLS connection where handshake messages are delivered out-of-order
7084 those messages that OpenSSL is not yet ready to process will be buffered
7085 for later use. Under certain circumstances, a flaw in the logic means that
7086 those messages do not get removed from the buffer even though the handshake
7087 has been completed. An attacker could force up to approx. 15 messages to
7088 remain in the buffer when they are no longer required. These messages will
7089 be cleared when the DTLS connection is closed. The default maximum size for
7090 a message is 100k. Therefore, the attacker could force an additional 1500k
7091 to be consumed per connection. By opening many simultaneous connections an
7092 attacker could cause a DoS attack through memory exhaustion.
7094 This issue was reported to OpenSSL by Quan Luo.
7099 * DTLS replay protection DoS
7101 A flaw in the DTLS replay attack protection mechanism means that records
7102 that arrive for future epochs update the replay protection "window" before
7103 the MAC for the record has been validated. This could be exploited by an
7104 attacker by sending a record for the next epoch (which does not have to
7105 decrypt or have a valid MAC), with a very large sequence number. This means
7106 that all subsequent legitimate packets are dropped causing a denial of
7107 service for a specific DTLS connection.
7109 This issue was reported to OpenSSL by the OCAP audit team.
7114 * Certificate message OOB reads
7116 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7117 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7118 theoretical DoS risk but this has not been observed in practice on common
7121 The messages affected are client certificate, client certificate request
7122 and server certificate. As a result the attack can only be performed
7123 against a client or a server which enables client authentication.
7125 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7130 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7132 * Prevent padding oracle in AES-NI CBC MAC check
7134 A MITM attacker can use a padding oracle attack to decrypt traffic
7135 when the connection uses an AES CBC cipher and the server support
7138 This issue was introduced as part of the fix for Lucky 13 padding
7139 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7140 constant time by making sure that always the same bytes are read and
7141 compared against either the MAC or padding bytes. But it no longer
7142 checked that there was enough data to have both the MAC and padding
7145 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7150 * Fix EVP_EncodeUpdate overflow
7152 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7153 Base64 encoding of binary data. If an attacker is able to supply very large
7154 amounts of input data then a length check can overflow resulting in a heap
7157 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7158 the `PEM_write_bio*` family of functions. These are mainly used within the
7159 OpenSSL command line applications, so any application which processes data
7160 from an untrusted source and outputs it as a PEM file should be considered
7161 vulnerable to this issue. User applications that call these APIs directly
7162 with large amounts of untrusted data may also be vulnerable.
7164 This issue was reported by Guido Vranken.
7169 * Fix EVP_EncryptUpdate overflow
7171 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7172 is able to supply very large amounts of input data after a previous call to
7173 EVP_EncryptUpdate() with a partial block then a length check can overflow
7174 resulting in a heap corruption. Following an analysis of all OpenSSL
7175 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7176 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7177 the first called function after an EVP_EncryptInit(), and therefore that
7178 specific call must be safe. The second form is where the length passed to
7179 EVP_EncryptUpdate() can be seen from the code to be some small value and
7180 therefore there is no possibility of an overflow. Since all instances are
7181 one of these two forms, it is believed that there can be no overflows in
7182 internal code due to this problem. It should be noted that
7183 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7184 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7185 of these calls have also been analysed too and it is believed there are no
7186 instances in internal usage where an overflow could occur.
7188 This issue was reported by Guido Vranken.
7193 * Prevent ASN.1 BIO excessive memory allocation
7195 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7196 a short invalid encoding can casuse allocation of large amounts of memory
7197 potentially consuming excessive resources or exhausting memory.
7199 Any application parsing untrusted data through d2i BIO functions is
7200 affected. The memory based functions such as d2i_X509() are *not* affected.
7201 Since the memory based functions are used by the TLS library, TLS
7202 applications are not affected.
7204 This issue was reported by Brian Carpenter.
7211 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7212 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7213 in arbitrary stack data being returned in the buffer.
7215 This issue was reported by Guido Vranken.
7220 * Modify behavior of ALPN to invoke callback after SNI/servername
7221 callback, such that updates to the SSL_CTX affect ALPN.
7225 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7230 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7231 methods are enabled and ssl2 is disabled the methods return NULL.
7235 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7237 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7238 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7239 provide any "EXPORT" or "LOW" strength ciphers.
7243 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7244 is by default disabled at build-time. Builds that are not configured with
7245 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7246 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7247 will need to explicitly call either of:
7249 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7251 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7253 as appropriate. Even if either of those is used, or the application
7254 explicitly uses the version-specific SSLv2_method() or its client and
7255 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7256 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7257 ciphers, and SSLv2 56-bit DES are no longer available.
7262 * Fix a double-free in DSA code
7264 A double free bug was discovered when OpenSSL parses malformed DSA private
7265 keys and could lead to a DoS attack or memory corruption for applications
7266 that receive DSA private keys from untrusted sources. This scenario is
7269 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7275 * Disable SRP fake user seed to address a server memory leak.
7277 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7279 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7280 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7281 was changed to ignore the "fake user" SRP seed, even if the seed
7284 Users should use SRP_VBASE_get1_by_user instead. Note that in
7285 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7286 also that even though configuring the SRP seed attempts to hide
7287 invalid usernames by continuing the handshake with fake
7288 credentials, this behaviour is not constant time and no strong
7289 guarantees are made that the handshake is indistinguishable from
7290 that of a valid user.
7295 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7297 In the BN_hex2bn function the number of hex digits is calculated using an
7298 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7299 large values of `i` this can result in `bn_expand` not allocating any
7300 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7301 field as NULL leading to a subsequent NULL ptr deref. For very large values
7302 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7303 In this case memory is allocated to the internal BIGNUM data field, but it
7304 is insufficiently sized leading to heap corruption. A similar issue exists
7305 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7306 is ever called by user applications with very large untrusted hex/dec data.
7307 This is anticipated to be a rare occurrence.
7309 All OpenSSL internal usage of these functions use data that is not expected
7310 to be untrusted, e.g. config file data or application command line
7311 arguments. If user developed applications generate config file data based
7312 on untrusted data then it is possible that this could also lead to security
7313 consequences. This is also anticipated to be rare.
7315 This issue was reported to OpenSSL by Guido Vranken.
7320 * Fix memory issues in `BIO_*printf` functions
7322 The internal `fmtstr` function used in processing a "%s" format string in
7323 the `BIO_*printf` functions could overflow while calculating the length of a
7324 string and cause an OOB read when printing very long strings.
7326 Additionally the internal `doapr_outch` function can attempt to write to an
7327 OOB memory location (at an offset from the NULL pointer) in the event of a
7328 memory allocation failure. In 1.0.2 and below this could be caused where
7329 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7330 could be in processing a very long "%s" format string. Memory leaks can
7333 The first issue may mask the second issue dependent on compiler behaviour.
7334 These problems could enable attacks where large amounts of untrusted data
7335 is passed to the `BIO_*printf` functions. If applications use these functions
7336 in this way then they could be vulnerable. OpenSSL itself uses these
7337 functions when printing out human-readable dumps of ASN.1 data. Therefore
7338 applications that print this data could be vulnerable if the data is from
7339 untrusted sources. OpenSSL command line applications could also be
7340 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7341 as command line arguments.
7343 Libssl is not considered directly vulnerable. Additionally certificates etc
7344 received via remote connections via libssl are also unlikely to be able to
7345 trigger these issues because of message size limits enforced within libssl.
7347 This issue was reported to OpenSSL Guido Vranken.
7352 * Side channel attack on modular exponentiation
7354 A side-channel attack was found which makes use of cache-bank conflicts on
7355 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7356 of RSA keys. The ability to exploit this issue is limited as it relies on
7357 an attacker who has control of code in a thread running on the same
7358 hyper-threaded core as the victim thread which is performing decryptions.
7360 This issue was reported to OpenSSL by Yuval Yarom, The University of
7361 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7362 Nadia Heninger, University of Pennsylvania with more information at
7363 <http://cachebleed.info>.
7368 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7369 if no keysize is specified with default_bits. This fixes an
7370 omission in an earlier change that changed all RSA/DSA key generation
7371 commands to use 2048 bits by default.
7375 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7377 * Protection for DH small subgroup attacks
7379 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7380 switched on by default and cannot be disabled. This could have some
7385 * SSLv2 doesn't block disabled ciphers
7387 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7388 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7389 been disabled, provided that the SSLv2 protocol was not also disabled via
7392 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7393 and Sebastian Schinzel.
7398 * Reject DH handshakes with parameters shorter than 1024 bits.
7402 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7404 * Certificate verify crash with missing PSS parameter
7406 The signature verification routines will crash with a NULL pointer
7407 dereference if presented with an ASN.1 signature using the RSA PSS
7408 algorithm and absent mask generation function parameter. Since these
7409 routines are used to verify certificate signature algorithms this can be
7410 used to crash any certificate verification operation and exploited in a
7411 DoS attack. Any application which performs certificate verification is
7412 vulnerable including OpenSSL clients and servers which enable client
7415 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7420 * X509_ATTRIBUTE memory leak
7422 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7423 memory. This structure is used by the PKCS#7 and CMS routines so any
7424 application which reads PKCS#7 or CMS data from untrusted sources is
7425 affected. SSL/TLS is not affected.
7427 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7433 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7434 This changes the decoding behaviour for some invalid messages,
7435 though the change is mostly in the more lenient direction, and
7436 legacy behaviour is preserved as much as possible.
7440 * In DSA_generate_parameters_ex, if the provided seed is too short,
7441 use a random seed, as already documented.
7443 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7445 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7447 * Alternate chains certificate forgery
7449 During certificate verification, OpenSSL will attempt to find an
7450 alternative certificate chain if the first attempt to build such a chain
7451 fails. An error in the implementation of this logic can mean that an
7452 attacker could cause certain checks on untrusted certificates to be
7453 bypassed, such as the CA flag, enabling them to use a valid leaf
7454 certificate to act as a CA and "issue" an invalid certificate.
7456 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7462 * Race condition handling PSK identify hint
7464 If PSK identity hints are received by a multi-threaded client then
7465 the values are wrongly updated in the parent SSL_CTX structure. This can
7466 result in a race condition potentially leading to a double free of the
7472 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7474 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7475 incompatibility in the handling of HMAC. The previous ABI has now been
7478 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7480 * Malformed ECParameters causes infinite loop
7482 When processing an ECParameters structure OpenSSL enters an infinite loop
7483 if the curve specified is over a specially malformed binary polynomial
7486 This can be used to perform denial of service against any
7487 system which processes public keys, certificate requests or
7488 certificates. This includes TLS clients and TLS servers with
7489 client authentication enabled.
7491 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7496 * Exploitable out-of-bounds read in X509_cmp_time
7498 X509_cmp_time does not properly check the length of the ASN1_TIME
7499 string and can read a few bytes out of bounds. In addition,
7500 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7503 An attacker can use this to craft malformed certificates and CRLs of
7504 various sizes and potentially cause a segmentation fault, resulting in
7505 a DoS on applications that verify certificates or CRLs. TLS clients
7506 that verify CRLs are affected. TLS clients and servers with client
7507 authentication enabled may be affected if they use custom verification
7510 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7511 independently by Hanno Böck.
7516 * PKCS7 crash with missing EnvelopedContent
7518 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7519 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7520 with missing content and trigger a NULL pointer dereference on parsing.
7522 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7523 structures from untrusted sources are affected. OpenSSL clients and
7524 servers are not affected.
7526 This issue was reported to OpenSSL by Michal Zalewski (Google).
7531 * CMS verify infinite loop with unknown hash function
7533 When verifying a signedData message the CMS code can enter an infinite loop
7534 if presented with an unknown hash function OID. This can be used to perform
7535 denial of service against any system which verifies signedData messages using
7537 This issue was reported to OpenSSL by Johannes Bauer.
7542 * Race condition handling NewSessionTicket
7544 If a NewSessionTicket is received by a multi-threaded client when attempting to
7545 reuse a previous ticket then a race condition can occur potentially leading to
7546 a double free of the ticket data.
7551 * Reject DH handshakes with parameters shorter than 768 bits.
7553 *Kurt Roeckx and Emilia Kasper*
7555 * dhparam: generate 2048-bit parameters by default.
7557 *Kurt Roeckx and Emilia Kasper*
7559 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7561 * Segmentation fault in ASN1_TYPE_cmp fix
7563 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7564 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7565 certificate signature algorithm consistency this can be used to crash any
7566 certificate verification operation and exploited in a DoS attack. Any
7567 application which performs certificate verification is vulnerable including
7568 OpenSSL clients and servers which enable client authentication.
7573 * ASN.1 structure reuse memory corruption fix
7575 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7576 memory corruption via an invalid write. Such reuse is and has been
7577 strongly discouraged and is believed to be rare.
7579 Applications that parse structures containing CHOICE or ANY DEFINED BY
7580 components may be affected. Certificate parsing (d2i_X509 and related
7581 functions) are however not affected. OpenSSL clients and servers are
7587 * PKCS7 NULL pointer dereferences fix
7589 The PKCS#7 parsing code does not handle missing outer ContentInfo
7590 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7591 missing content and trigger a NULL pointer dereference on parsing.
7593 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7594 otherwise parse PKCS#7 structures from untrusted sources are
7595 affected. OpenSSL clients and servers are not affected.
7597 This issue was reported to OpenSSL by Michal Zalewski (Google).
7602 * DoS via reachable assert in SSLv2 servers fix
7604 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7605 servers that both support SSLv2 and enable export cipher suites by sending
7606 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7608 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7609 (OpenSSL development team).
7614 * Use After Free following d2i_ECPrivatekey error fix
7616 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7617 could cause a use after free condition. This, in turn, could cause a double
7618 free in several private key parsing functions (such as d2i_PrivateKey
7619 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7620 for applications that receive EC private keys from untrusted
7621 sources. This scenario is considered rare.
7623 This issue was discovered by the BoringSSL project and fixed in their
7629 * X509_to_X509_REQ NULL pointer deref fix
7631 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7632 the certificate key is invalid. This function is rarely used in practice.
7634 This issue was discovered by Brian Carpenter.
7639 * Removed the export ciphers from the DEFAULT ciphers
7643 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7645 * Build fixes for the Windows and OpenVMS platforms
7647 *Matt Caswell and Richard Levitte*
7649 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7651 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7652 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7653 dereference. This could lead to a Denial Of Service attack. Thanks to
7654 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7659 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7660 dtls1_buffer_record function under certain conditions. In particular this
7661 could occur if an attacker sent repeated DTLS records with the same
7662 sequence number but for the next epoch. The memory leak could be exploited
7663 by an attacker in a Denial of Service attack through memory exhaustion.
7664 Thanks to Chris Mueller for reporting this issue.
7669 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7670 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7671 method would be set to NULL which could later result in a NULL pointer
7672 dereference. Thanks to Frank Schmirler for reporting this issue.
7677 * Abort handshake if server key exchange message is omitted for ephemeral
7680 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7681 reporting this issue.
7686 * Remove non-export ephemeral RSA code on client and server. This code
7687 violated the TLS standard by allowing the use of temporary RSA keys in
7688 non-export ciphersuites and could be used by a server to effectively
7689 downgrade the RSA key length used to a value smaller than the server
7690 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7691 INRIA or reporting this issue.
7696 * Fixed issue where DH client certificates are accepted without verification.
7697 An OpenSSL server will accept a DH certificate for client authentication
7698 without the certificate verify message. This effectively allows a client to
7699 authenticate without the use of a private key. This only affects servers
7700 which trust a client certificate authority which issues certificates
7701 containing DH keys: these are extremely rare and hardly ever encountered.
7702 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7708 * Ensure that the session ID context of an SSL is updated when its
7709 SSL_CTX is updated via SSL_set_SSL_CTX.
7711 The session ID context is typically set from the parent SSL_CTX,
7712 and can vary with the CTX.
7716 * Fix various certificate fingerprint issues.
7718 By using non-DER or invalid encodings outside the signed portion of a
7719 certificate the fingerprint can be changed without breaking the signature.
7720 Although no details of the signed portion of the certificate can be changed
7721 this can cause problems with some applications: e.g. those using the
7722 certificate fingerprint for blacklists.
7724 1. Reject signatures with non zero unused bits.
7726 If the BIT STRING containing the signature has non zero unused bits reject
7727 the signature. All current signature algorithms require zero unused bits.
7729 2. Check certificate algorithm consistency.
7731 Check the AlgorithmIdentifier inside TBS matches the one in the
7732 certificate signature. NB: this will result in signature failure
7733 errors for some broken certificates.
7735 Thanks to Konrad Kraszewski from Google for reporting this issue.
7737 3. Check DSA/ECDSA signatures use DER.
7739 Re-encode DSA/ECDSA signatures and compare with the original received
7740 signature. Return an error if there is a mismatch.
7742 This will reject various cases including garbage after signature
7743 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7744 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7745 (negative or with leading zeroes).
7747 Further analysis was conducted and fixes were developed by Stephen Henson
7748 of the OpenSSL core team.
7754 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7755 results on some platforms, including x86_64. This bug occurs at random
7756 with a very low probability, and is not known to be exploitable in any
7757 way, though its exact impact is difficult to determine. Thanks to Pieter
7758 Wuille (Blockstream) who reported this issue and also suggested an initial
7759 fix. Further analysis was conducted by the OpenSSL development team and
7760 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7761 the OpenSSL core team.
7766 * Do not resume sessions on the server if the negotiated protocol
7767 version does not match the session's version. Resuming with a different
7768 version, while not strictly forbidden by the RFC, is of questionable
7769 sanity and breaks all known clients.
7771 *David Benjamin, Emilia Käsper*
7773 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7774 early CCS messages during renegotiation. (Note that because
7775 renegotiation is encrypted, this early CCS was not exploitable.)
7779 * Tighten client-side session ticket handling during renegotiation:
7780 ensure that the client only accepts a session ticket if the server sends
7781 the extension anew in the ServerHello. Previously, a TLS client would
7782 reuse the old extension state and thus accept a session ticket if one was
7783 announced in the initial ServerHello.
7785 Similarly, ensure that the client requires a session ticket if one
7786 was advertised in the ServerHello. Previously, a TLS client would
7787 ignore a missing NewSessionTicket message.
7791 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7795 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7796 sends a carefully crafted handshake message, to cause OpenSSL to fail
7797 to free up to 64k of memory causing a memory leak. This could be
7798 exploited in a Denial Of Service attack. This issue affects OpenSSL
7799 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7800 whether SRTP is used or configured. Implementations of OpenSSL that
7801 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7803 The fix was developed by the OpenSSL team.
7808 * Session Ticket Memory Leak.
7810 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7811 integrity of that ticket is first verified. In the event of a session
7812 ticket integrity check failing, OpenSSL will fail to free memory
7813 causing a memory leak. By sending a large number of invalid session
7814 tickets an attacker could exploit this issue in a Denial Of Service
7820 * Build option no-ssl3 is incomplete.
7822 When OpenSSL is configured with "no-ssl3" as a build option, servers
7823 could accept and complete an SSL 3.0 handshake, and clients could be
7824 configured to send them.
7827 *Akamai and the OpenSSL team*
7829 * Add support for TLS_FALLBACK_SCSV.
7830 Client applications doing fallback retries should call
7831 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7834 *Adam Langley, Bodo Moeller*
7836 * Add additional DigestInfo checks.
7838 Re-encode DigestInto in DER and check against the original when
7839 verifying RSA signature: this will reject any improperly encoded
7840 DigestInfo structures.
7842 Note: this is a precautionary measure and no attacks are currently known.
7846 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7848 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7849 SRP code can be overrun an internal buffer. Add sanity check that
7850 g, A, B < N to SRP code.
7852 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7853 Group for discovering this issue.
7858 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7859 TLS 1.0 instead of higher protocol versions when the ClientHello message
7860 is badly fragmented. This allows a man-in-the-middle attacker to force a
7861 downgrade to TLS 1.0 even if both the server and the client support a
7862 higher protocol version, by modifying the client's TLS records.
7864 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7865 researching this issue.
7870 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7871 to a denial of service attack. A malicious server can crash the client
7872 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7873 ciphersuite and sending carefully crafted handshake messages.
7875 Thanks to Felix Gröbert (Google) for discovering and researching this
7881 * By sending carefully crafted DTLS packets an attacker could cause openssl
7882 to leak memory. This can be exploited through a Denial of Service attack.
7883 Thanks to Adam Langley for discovering and researching this issue.
7888 * An attacker can force openssl to consume large amounts of memory whilst
7889 processing DTLS handshake messages. This can be exploited through a
7890 Denial of Service attack.
7891 Thanks to Adam Langley for discovering and researching this issue.
7896 * An attacker can force an error condition which causes openssl to crash
7897 whilst processing DTLS packets due to memory being freed twice. This
7898 can be exploited through a Denial of Service attack.
7899 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7905 * If a multithreaded client connects to a malicious server using a resumed
7906 session and the server sends an ec point format extension it could write
7907 up to 255 bytes to freed memory.
7909 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7915 * A malicious server can crash an OpenSSL client with a null pointer
7916 dereference (read) by specifying an SRP ciphersuite even though it was not
7917 properly negotiated with the client. This can be exploited through a
7918 Denial of Service attack.
7920 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7921 discovering and researching this issue.
7926 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7927 X509_name_oneline, X509_name_print_ex et al. to leak some information
7928 from the stack. Applications may be affected if they echo pretty printing
7929 output to the attacker.
7931 Thanks to Ivan Fratric (Google) for discovering this issue.
7934 *Emilia Käsper, and Steve Henson*
7936 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7937 for corner cases. (Certain input points at infinity could lead to
7938 bogus results, with non-infinity inputs mapped to infinity too.)
7942 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7944 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7945 handshake can force the use of weak keying material in OpenSSL
7946 SSL/TLS clients and servers.
7948 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7949 researching this issue. ([CVE-2014-0224])
7951 *KIKUCHI Masashi, Steve Henson*
7953 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7954 OpenSSL DTLS client the code can be made to recurse eventually crashing
7957 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7960 *Imre Rad, Steve Henson*
7962 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7963 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7964 client or server. This is potentially exploitable to run arbitrary
7965 code on a vulnerable client or server.
7967 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7969 *Jüri Aedla, Steve Henson*
7971 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7972 are subject to a denial of service attack.
7974 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7975 this issue. ([CVE-2014-3470])
7977 *Felix Gröbert, Ivan Fratric, Steve Henson*
7979 * Harmonize version and its documentation. -f flag is used to display
7982 *mancha <mancha1@zoho.com>*
7984 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7985 in i2d_ECPrivateKey.
7987 *mancha <mancha1@zoho.com>*
7989 * Fix some double frees. These are not thought to be exploitable.
7991 *mancha <mancha1@zoho.com>*
7993 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7995 * A missing bounds check in the handling of the TLS heartbeat extension
7996 can be used to reveal up to 64k of memory to a connected client or
7999 Thanks for Neel Mehta of Google Security for discovering this bug and to
8000 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8001 preparing the fix ([CVE-2014-0160])
8003 *Adam Langley, Bodo Moeller*
8005 * Fix for the attack described in the paper "Recovering OpenSSL
8006 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8007 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8008 <http://eprint.iacr.org/2014/140>
8010 Thanks to Yuval Yarom and Naomi Benger for discovering this
8011 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8013 *Yuval Yarom and Naomi Benger*
8015 * TLS pad extension: draft-agl-tls-padding-03
8017 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8018 TLS client Hello record length value would otherwise be > 255 and
8019 less that 512 pad with a dummy extension containing zeroes so it
8020 is at least 512 bytes long.
8022 *Adam Langley, Steve Henson*
8024 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8026 * Fix for TLS record tampering bug. A carefully crafted invalid
8027 handshake could crash OpenSSL with a NULL pointer exception.
8028 Thanks to Anton Johansson for reporting this issues.
8031 * Keep original DTLS digest and encryption contexts in retransmission
8032 structures so we can use the previous session parameters if they need
8033 to be resent. ([CVE-2013-6450])
8037 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8038 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8039 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8040 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8041 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8042 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8044 *Rob Stradling, Adam Langley*
8046 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8048 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8049 supporting platforms or when small records were transferred.
8051 *Andy Polyakov, Steve Henson*
8053 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8055 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8057 This addresses the flaw in CBC record processing discovered by
8058 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8059 at: <http://www.isg.rhul.ac.uk/tls/>
8061 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8062 Security Group at Royal Holloway, University of London
8063 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8064 Emilia Käsper for the initial patch.
8067 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8069 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8070 ciphersuites which can be exploited in a denial of service attack.
8071 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8072 and detecting this bug and to Wolfgang Ettlinger
8073 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8078 * Return an error when checking OCSP signatures when key is NULL.
8079 This fixes a DoS attack. ([CVE-2013-0166])
8083 * Make openssl verify return errors.
8085 *Chris Palmer <palmer@google.com> and Ben Laurie*
8087 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8088 the right response is stapled. Also change SSL_get_certificate()
8089 so it returns the certificate actually sent.
8090 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8092 *Rob Stradling <rob.stradling@comodo.com>*
8094 * Fix possible deadlock when decoding public keys.
8098 * Don't use TLS 1.0 record version number in initial client hello
8103 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8105 * Sanity check record length before skipping explicit IV in TLS
8106 1.2, 1.1 and DTLS to fix DoS attack.
8108 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8109 fuzzing as a service testing platform.
8114 * Initialise tkeylen properly when encrypting CMS messages.
8115 Thanks to Solar Designer of Openwall for reporting this issue.
8119 * In FIPS mode don't try to use composite ciphers as they are not
8124 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8126 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8127 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8128 mean any application compiled against OpenSSL 1.0.0 headers setting
8129 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8130 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8131 0x10000000L Any application which was previously compiled against
8132 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8133 will need to be recompiled as a result. Letting be results in
8134 inability to disable specifically TLS 1.1 and in client context,
8135 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8139 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8140 disable just protocol X, but all protocols above X *if* there are
8141 protocols *below* X still enabled. In more practical terms it means
8142 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8143 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8144 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8149 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8151 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8152 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8153 in CRYPTO_realloc_clean.
8155 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8156 issue and to Adam Langley <agl@chromium.org> for fixing it.
8159 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8161 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8165 * Workarounds for some broken servers that "hang" if a client hello
8166 record length exceeds 255 bytes.
8168 1. Do not use record version number > TLS 1.0 in initial client
8169 hello: some (but not all) hanging servers will now work.
8170 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8171 the number of ciphers sent in the client hello. This should be
8172 set to an even number, such as 50, for example by passing:
8173 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8174 Most broken servers should now work.
8175 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8176 TLS 1.2 client support entirely.
8180 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8184 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8186 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8187 STRING form instead of a DigestInfo.
8191 * The format used for MDC2 RSA signatures is inconsistent between EVP
8192 and the RSA_sign/RSA_verify functions. This was made more apparent when
8193 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8194 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8195 the correct format in RSA_verify so both forms transparently work.
8199 * Some servers which support TLS 1.0 can choke if we initially indicate
8200 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8201 encrypted premaster secret. As a workaround use the maximum permitted
8202 client version in client hello, this should keep such servers happy
8203 and still work with previous versions of OpenSSL.
8207 * Add support for TLS/DTLS heartbeats.
8209 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8211 * Add support for SCTP.
8213 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8215 * Improved PRNG seeding for VOS.
8217 *Paul Green <Paul.Green@stratus.com>*
8219 * Extensive assembler packs updates, most notably:
8221 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8222 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8223 - x86_64: bit-sliced AES implementation;
8224 - ARM: NEON support, contemporary platforms optimizations;
8225 - s390x: z196 support;
8226 - `*`: GHASH and GF(2^m) multiplication implementations;
8230 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8231 (removal of unnecessary code)
8233 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8235 * Add TLS key material exporter from RFC 5705.
8239 * Add DTLS-SRTP negotiation from RFC 5764.
8243 * Add Next Protocol Negotiation,
8244 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8245 disabled with a no-npn flag to config or Configure. Code donated
8248 *Adam Langley <agl@google.com> and Ben Laurie*
8250 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8251 NIST-P256, NIST-P521, with constant-time single point multiplication on
8252 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8253 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8254 Code made available under Apache License version 2.0.
8256 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8257 line to include this in your build of OpenSSL, and run "make depend" (or
8258 "make update"). This enables the following EC_METHODs:
8260 EC_GFp_nistp224_method()
8261 EC_GFp_nistp256_method()
8262 EC_GFp_nistp521_method()
8264 EC_GROUP_new_by_curve_name() will automatically use these (while
8265 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8268 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8270 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8271 all platforms. Move ssize_t definition from e_os.h to the public
8272 header file e_os2.h as it now appears in public header file cms.h
8276 * New -sigopt option to the ca, req and x509 utilities. Additional
8277 signature parameters can be passed using this option and in
8282 * Add RSA PSS signing function. This will generate and set the
8283 appropriate AlgorithmIdentifiers for PSS based on those in the
8284 corresponding EVP_MD_CTX structure. No application support yet.
8288 * Support for companion algorithm specific ASN1 signing routines.
8289 New function ASN1_item_sign_ctx() signs a pre-initialised
8290 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8291 the appropriate parameters.
8295 * Add new algorithm specific ASN1 verification initialisation function
8296 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8297 handling will be the same no matter what EVP_PKEY_METHOD is used.
8298 Add a PSS handler to support verification of PSS signatures: checked
8299 against a number of sample certificates.
8303 * Add signature printing for PSS. Add PSS OIDs.
8305 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8307 * Add algorithm specific signature printing. An individual ASN1 method
8308 can now print out signatures instead of the standard hex dump.
8310 More complex signatures (e.g. PSS) can print out more meaningful
8311 information. Include DSA version that prints out the signature
8316 * Password based recipient info support for CMS library: implementing
8321 * Split password based encryption into PBES2 and PBKDF2 functions. This
8322 neatly separates the code into cipher and PBE sections and is required
8323 for some algorithms that split PBES2 into separate pieces (such as
8324 password based CMS).
8328 * Session-handling fixes:
8329 - Fix handling of connections that are resuming with a session ID,
8330 but also support Session Tickets.
8331 - Fix a bug that suppressed issuing of a new ticket if the client
8332 presented a ticket with an expired session.
8333 - Try to set the ticket lifetime hint to something reasonable.
8334 - Make tickets shorter by excluding irrelevant information.
8335 - On the client side, don't ignore renewed tickets.
8337 *Adam Langley, Bodo Moeller (Google)*
8339 * Fix PSK session representation.
8343 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8345 This work was sponsored by Intel.
8349 * Add GCM support to TLS library. Some custom code is needed to split
8350 the IV between the fixed (from PRF) and explicit (from TLS record)
8351 portions. This adds all GCM ciphersuites supported by RFC5288 and
8352 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8353 add a special AESGCM string for GCM only.
8357 * Expand range of ctrls for AES GCM. Permit setting invocation
8358 field on decrypt and retrieval of invocation field only on encrypt.
8362 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8363 As required by RFC5289 these ciphersuites cannot be used if for
8364 versions of TLS earlier than 1.2.
8368 * For FIPS capable OpenSSL interpret a NULL default public key method
8369 as unset and return the appropriate default but do *not* set the default.
8370 This means we can return the appropriate method in applications that
8371 switch between FIPS and non-FIPS modes.
8375 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8376 ENGINE is used then we cannot handle that in the FIPS module so we
8377 keep original code iff non-FIPS operations are allowed.
8381 * Add -attime option to openssl utilities.
8383 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8385 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8389 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8390 FIPS EC methods unconditionally for now.
8394 * New build option no-ec2m to disable characteristic 2 code.
8398 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8399 all cases can be covered as some introduce binary incompatibilities.
8403 * Redirect RSA operations to FIPS module including keygen,
8404 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8408 * Add similar low-level API blocking to ciphers.
8412 * low-level digest APIs are not approved in FIPS mode: any attempt
8413 to use these will cause a fatal error. Applications that *really* want
8414 to use them can use the `private_*` version instead.
8418 * Redirect cipher operations to FIPS module for FIPS builds.
8422 * Redirect digest operations to FIPS module for FIPS builds.
8426 * Update build system to add "fips" flag which will link in fipscanister.o
8427 for static and shared library builds embedding a signature if needed.
8431 * Output TLS supported curves in preference order instead of numerical
8432 order. This is currently hardcoded for the highest order curves first.
8433 This should be configurable so applications can judge speed vs strength.
8437 * Add TLS v1.2 server support for client authentication.
8441 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8446 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8447 FIPS modules versions.
8451 * Add TLS v1.2 client side support for client authentication. Keep cache
8452 of handshake records longer as we don't know the hash algorithm to use
8453 until after the certificate request message is received.
8457 * Initial TLS v1.2 client support. Add a default signature algorithms
8458 extension including all the algorithms we support. Parse new signature
8459 format in client key exchange. Relax some ECC signing restrictions for
8460 TLS v1.2 as indicated in RFC5246.
8464 * Add server support for TLS v1.2 signature algorithms extension. Switch
8465 to new signature format when needed using client digest preference.
8466 All server ciphersuites should now work correctly in TLS v1.2. No client
8467 support yet and no support for client certificates.
8471 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8472 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8473 ciphersuites. At present only RSA key exchange ciphersuites work with
8474 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8475 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8476 and version checking.
8480 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8481 with this defined it will not be affected by any changes to ssl internal
8482 structures. Add several utility functions to allow openssl application
8483 to work with OPENSSL_NO_SSL_INTERN defined.
8487 * A long standing patch to add support for SRP from EdelWeb (Peter
8488 Sylvester and Christophe Renou) was integrated.
8489 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8490 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8493 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8497 * Permit abbreviated handshakes when renegotiating using the function
8498 SSL_renegotiate_abbreviated().
8500 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8502 * Add call to ENGINE_register_all_complete() to
8503 ENGINE_load_builtin_engines(), so some implementations get used
8504 automatically instead of needing explicit application support.
8508 * Add support for TLS key exporter as described in RFC5705.
8510 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8512 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8513 a few changes are required:
8515 Add SSL_OP_NO_TLSv1_1 flag.
8516 Add TLSv1_1 methods.
8517 Update version checking logic to handle version 1.1.
8518 Add explicit IV handling (ported from DTLS code).
8519 Add command line options to s_client/s_server.
8526 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8528 * X509_ATTRIBUTE memory leak
8530 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8531 memory. This structure is used by the PKCS#7 and CMS routines so any
8532 application which reads PKCS#7 or CMS data from untrusted sources is
8533 affected. SSL/TLS is not affected.
8535 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8541 * Race condition handling PSK identify hint
8543 If PSK identity hints are received by a multi-threaded client then
8544 the values are wrongly updated in the parent SSL_CTX structure. This can
8545 result in a race condition potentially leading to a double free of the
8551 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8553 * Malformed ECParameters causes infinite loop
8555 When processing an ECParameters structure OpenSSL enters an infinite loop
8556 if the curve specified is over a specially malformed binary polynomial
8559 This can be used to perform denial of service against any
8560 system which processes public keys, certificate requests or
8561 certificates. This includes TLS clients and TLS servers with
8562 client authentication enabled.
8564 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8569 * Exploitable out-of-bounds read in X509_cmp_time
8571 X509_cmp_time does not properly check the length of the ASN1_TIME
8572 string and can read a few bytes out of bounds. In addition,
8573 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8576 An attacker can use this to craft malformed certificates and CRLs of
8577 various sizes and potentially cause a segmentation fault, resulting in
8578 a DoS on applications that verify certificates or CRLs. TLS clients
8579 that verify CRLs are affected. TLS clients and servers with client
8580 authentication enabled may be affected if they use custom verification
8583 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8584 independently by Hanno Böck.
8589 * PKCS7 crash with missing EnvelopedContent
8591 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8592 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8593 with missing content and trigger a NULL pointer dereference on parsing.
8595 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8596 structures from untrusted sources are affected. OpenSSL clients and
8597 servers are not affected.
8599 This issue was reported to OpenSSL by Michal Zalewski (Google).
8604 * CMS verify infinite loop with unknown hash function
8606 When verifying a signedData message the CMS code can enter an infinite loop
8607 if presented with an unknown hash function OID. This can be used to perform
8608 denial of service against any system which verifies signedData messages using
8610 This issue was reported to OpenSSL by Johannes Bauer.
8615 * Race condition handling NewSessionTicket
8617 If a NewSessionTicket is received by a multi-threaded client when attempting to
8618 reuse a previous ticket then a race condition can occur potentially leading to
8619 a double free of the ticket data.
8624 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8626 * Segmentation fault in ASN1_TYPE_cmp fix
8628 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8629 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8630 certificate signature algorithm consistency this can be used to crash any
8631 certificate verification operation and exploited in a DoS attack. Any
8632 application which performs certificate verification is vulnerable including
8633 OpenSSL clients and servers which enable client authentication.
8638 * ASN.1 structure reuse memory corruption fix
8640 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8641 memory corruption via an invalid write. Such reuse is and has been
8642 strongly discouraged and is believed to be rare.
8644 Applications that parse structures containing CHOICE or ANY DEFINED BY
8645 components may be affected. Certificate parsing (d2i_X509 and related
8646 functions) are however not affected. OpenSSL clients and servers are
8652 * PKCS7 NULL pointer dereferences fix
8654 The PKCS#7 parsing code does not handle missing outer ContentInfo
8655 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8656 missing content and trigger a NULL pointer dereference on parsing.
8658 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8659 otherwise parse PKCS#7 structures from untrusted sources are
8660 affected. OpenSSL clients and servers are not affected.
8662 This issue was reported to OpenSSL by Michal Zalewski (Google).
8667 * DoS via reachable assert in SSLv2 servers fix
8669 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8670 servers that both support SSLv2 and enable export cipher suites by sending
8671 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8673 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8674 (OpenSSL development team).
8679 * Use After Free following d2i_ECPrivatekey error fix
8681 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8682 could cause a use after free condition. This, in turn, could cause a double
8683 free in several private key parsing functions (such as d2i_PrivateKey
8684 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8685 for applications that receive EC private keys from untrusted
8686 sources. This scenario is considered rare.
8688 This issue was discovered by the BoringSSL project and fixed in their
8694 * X509_to_X509_REQ NULL pointer deref fix
8696 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8697 the certificate key is invalid. This function is rarely used in practice.
8699 This issue was discovered by Brian Carpenter.
8704 * Removed the export ciphers from the DEFAULT ciphers
8708 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8710 * Build fixes for the Windows and OpenVMS platforms
8712 *Matt Caswell and Richard Levitte*
8714 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8716 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8717 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8718 dereference. This could lead to a Denial Of Service attack. Thanks to
8719 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8724 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8725 dtls1_buffer_record function under certain conditions. In particular this
8726 could occur if an attacker sent repeated DTLS records with the same
8727 sequence number but for the next epoch. The memory leak could be exploited
8728 by an attacker in a Denial of Service attack through memory exhaustion.
8729 Thanks to Chris Mueller for reporting this issue.
8734 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8735 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8736 method would be set to NULL which could later result in a NULL pointer
8737 dereference. Thanks to Frank Schmirler for reporting this issue.
8742 * Abort handshake if server key exchange message is omitted for ephemeral
8745 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8746 reporting this issue.
8751 * Remove non-export ephemeral RSA code on client and server. This code
8752 violated the TLS standard by allowing the use of temporary RSA keys in
8753 non-export ciphersuites and could be used by a server to effectively
8754 downgrade the RSA key length used to a value smaller than the server
8755 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8756 INRIA or reporting this issue.
8761 * Fixed issue where DH client certificates are accepted without verification.
8762 An OpenSSL server will accept a DH certificate for client authentication
8763 without the certificate verify message. This effectively allows a client to
8764 authenticate without the use of a private key. This only affects servers
8765 which trust a client certificate authority which issues certificates
8766 containing DH keys: these are extremely rare and hardly ever encountered.
8767 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8773 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8774 results on some platforms, including x86_64. This bug occurs at random
8775 with a very low probability, and is not known to be exploitable in any
8776 way, though its exact impact is difficult to determine. Thanks to Pieter
8777 Wuille (Blockstream) who reported this issue and also suggested an initial
8778 fix. Further analysis was conducted by the OpenSSL development team and
8779 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8780 the OpenSSL core team.
8785 * Fix various certificate fingerprint issues.
8787 By using non-DER or invalid encodings outside the signed portion of a
8788 certificate the fingerprint can be changed without breaking the signature.
8789 Although no details of the signed portion of the certificate can be changed
8790 this can cause problems with some applications: e.g. those using the
8791 certificate fingerprint for blacklists.
8793 1. Reject signatures with non zero unused bits.
8795 If the BIT STRING containing the signature has non zero unused bits reject
8796 the signature. All current signature algorithms require zero unused bits.
8798 2. Check certificate algorithm consistency.
8800 Check the AlgorithmIdentifier inside TBS matches the one in the
8801 certificate signature. NB: this will result in signature failure
8802 errors for some broken certificates.
8804 Thanks to Konrad Kraszewski from Google for reporting this issue.
8806 3. Check DSA/ECDSA signatures use DER.
8808 Re-encode DSA/ECDSA signatures and compare with the original received
8809 signature. Return an error if there is a mismatch.
8811 This will reject various cases including garbage after signature
8812 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8813 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8814 (negative or with leading zeroes).
8816 Further analysis was conducted and fixes were developed by Stephen Henson
8817 of the OpenSSL core team.
8823 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8825 * Session Ticket Memory Leak.
8827 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8828 integrity of that ticket is first verified. In the event of a session
8829 ticket integrity check failing, OpenSSL will fail to free memory
8830 causing a memory leak. By sending a large number of invalid session
8831 tickets an attacker could exploit this issue in a Denial Of Service
8837 * Build option no-ssl3 is incomplete.
8839 When OpenSSL is configured with "no-ssl3" as a build option, servers
8840 could accept and complete an SSL 3.0 handshake, and clients could be
8841 configured to send them.
8844 *Akamai and the OpenSSL team*
8846 * Add support for TLS_FALLBACK_SCSV.
8847 Client applications doing fallback retries should call
8848 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8851 *Adam Langley, Bodo Moeller*
8853 * Add additional DigestInfo checks.
8855 Re-encode DigestInto in DER and check against the original when
8856 verifying RSA signature: this will reject any improperly encoded
8857 DigestInfo structures.
8859 Note: this is a precautionary measure and no attacks are currently known.
8863 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8865 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8866 to a denial of service attack. A malicious server can crash the client
8867 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8868 ciphersuite and sending carefully crafted handshake messages.
8870 Thanks to Felix Gröbert (Google) for discovering and researching this
8876 * By sending carefully crafted DTLS packets an attacker could cause openssl
8877 to leak memory. This can be exploited through a Denial of Service attack.
8878 Thanks to Adam Langley for discovering and researching this issue.
8883 * An attacker can force openssl to consume large amounts of memory whilst
8884 processing DTLS handshake messages. This can be exploited through a
8885 Denial of Service attack.
8886 Thanks to Adam Langley for discovering and researching this issue.
8891 * An attacker can force an error condition which causes openssl to crash
8892 whilst processing DTLS packets due to memory being freed twice. This
8893 can be exploited through a Denial of Service attack.
8894 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8900 * If a multithreaded client connects to a malicious server using a resumed
8901 session and the server sends an ec point format extension it could write
8902 up to 255 bytes to freed memory.
8904 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8910 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8911 X509_name_oneline, X509_name_print_ex et al. to leak some information
8912 from the stack. Applications may be affected if they echo pretty printing
8913 output to the attacker.
8915 Thanks to Ivan Fratric (Google) for discovering this issue.
8918 *Emilia Käsper, and Steve Henson*
8920 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8921 for corner cases. (Certain input points at infinity could lead to
8922 bogus results, with non-infinity inputs mapped to infinity too.)
8926 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8928 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8929 handshake can force the use of weak keying material in OpenSSL
8930 SSL/TLS clients and servers.
8932 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8933 researching this issue. ([CVE-2014-0224])
8935 *KIKUCHI Masashi, Steve Henson*
8937 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8938 OpenSSL DTLS client the code can be made to recurse eventually crashing
8941 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8944 *Imre Rad, Steve Henson*
8946 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8947 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8948 client or server. This is potentially exploitable to run arbitrary
8949 code on a vulnerable client or server.
8951 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8953 *Jüri Aedla, Steve Henson*
8955 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8956 are subject to a denial of service attack.
8958 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8959 this issue. ([CVE-2014-3470])
8961 *Felix Gröbert, Ivan Fratric, Steve Henson*
8963 * Harmonize version and its documentation. -f flag is used to display
8966 *mancha <mancha1@zoho.com>*
8968 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8969 in i2d_ECPrivateKey.
8971 *mancha <mancha1@zoho.com>*
8973 * Fix some double frees. These are not thought to be exploitable.
8975 *mancha <mancha1@zoho.com>*
8977 * Fix for the attack described in the paper "Recovering OpenSSL
8978 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8979 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8980 <http://eprint.iacr.org/2014/140>
8982 Thanks to Yuval Yarom and Naomi Benger for discovering this
8983 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8985 *Yuval Yarom and Naomi Benger*
8987 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8989 * Keep original DTLS digest and encryption contexts in retransmission
8990 structures so we can use the previous session parameters if they need
8991 to be resent. ([CVE-2013-6450])
8995 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8996 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8997 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8998 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8999 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9000 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9002 *Rob Stradling, Adam Langley*
9004 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9006 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9008 This addresses the flaw in CBC record processing discovered by
9009 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9010 at: <http://www.isg.rhul.ac.uk/tls/>
9012 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9013 Security Group at Royal Holloway, University of London
9014 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9015 Emilia Käsper for the initial patch.
9018 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9020 * Return an error when checking OCSP signatures when key is NULL.
9021 This fixes a DoS attack. ([CVE-2013-0166])
9025 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9026 the right response is stapled. Also change SSL_get_certificate()
9027 so it returns the certificate actually sent.
9028 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9029 (This is a backport)
9031 *Rob Stradling <rob.stradling@comodo.com>*
9033 * Fix possible deadlock when decoding public keys.
9037 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9039 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9042 * Sanity check record length before skipping explicit IV in DTLS
9045 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9046 fuzzing as a service testing platform.
9051 * Initialise tkeylen properly when encrypting CMS messages.
9052 Thanks to Solar Designer of Openwall for reporting this issue.
9056 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9058 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9059 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9060 in CRYPTO_realloc_clean.
9062 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9063 issue and to Adam Langley <agl@chromium.org> for fixing it.
9066 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9068 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9070 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9071 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9072 content decryption and always return the same error. Note: this attack
9073 needs on average 2^20 messages so it only affects automated senders. The
9074 old behaviour can be re-enabled in the CMS code by setting the
9075 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9076 an MMA defence is not necessary.
9077 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9078 this issue. ([CVE-2012-0884])
9082 * Fix CVE-2011-4619: make sure we really are receiving a
9083 client hello before rejecting multiple SGC restarts. Thanks to
9084 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9088 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9090 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9091 Thanks to Antonio Martin, Enterprise Secure Access Research and
9092 Development, Cisco Systems, Inc. for discovering this bug and
9093 preparing a fix. ([CVE-2012-0050])
9097 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9099 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9100 of the Vaudenay padding oracle attack on CBC mode encryption
9101 which enables an efficient plaintext recovery attack against
9102 the OpenSSL implementation of DTLS. Their attack exploits timing
9103 differences arising during decryption processing. A research
9104 paper describing this attack can be found at:
9105 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9106 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9107 Security Group at Royal Holloway, University of London
9108 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9109 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9110 for preparing the fix. ([CVE-2011-4108])
9112 *Robin Seggelmann, Michael Tuexen*
9114 * Clear bytes used for block padding of SSL 3.0 records.
9117 *Adam Langley (Google)*
9119 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9120 Kadianakis <desnacked@gmail.com> for discovering this issue and
9121 Adam Langley for preparing the fix. ([CVE-2011-4619])
9123 *Adam Langley (Google)*
9125 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9127 *Andrey Kulikov <amdeich@gmail.com>*
9129 * Prevent malformed RFC3779 data triggering an assertion failure.
9130 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9131 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9133 *Rob Austein <sra@hactrn.net>*
9135 * Improved PRNG seeding for VOS.
9137 *Paul Green <Paul.Green@stratus.com>*
9139 * Fix ssl_ciph.c set-up race.
9141 *Adam Langley (Google)*
9143 * Fix spurious failures in ecdsatest.c.
9145 *Emilia Käsper (Google)*
9147 * Fix the BIO_f_buffer() implementation (which was mixing different
9148 interpretations of the `..._len` fields).
9150 *Adam Langley (Google)*
9152 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9153 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9154 threads won't reuse the same blinding coefficients.
9156 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9157 lock to call BN_BLINDING_invert_ex, and avoids one use of
9158 BN_BLINDING_update for each BN_BLINDING structure (previously,
9159 the last update always remained unused).
9161 *Emilia Käsper (Google)*
9163 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9165 *Bob Buckholz (Google)*
9167 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9169 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9170 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9172 *Kaspar Brand <ossl@velox.ch>*
9174 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9175 for multi-threaded use of ECDH. ([CVE-2011-3210])
9177 *Adam Langley (Google)*
9179 * Fix x509_name_ex_d2i memory leak on bad inputs.
9183 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9184 signature public key algorithm by using OID xref utilities instead.
9185 Before this you could only use some ECC ciphersuites with SHA1 only.
9189 * Add protection against ECDSA timing attacks as mentioned in the paper
9190 by Billy Bob Brumley and Nicola Tuveri, see:
9191 <http://eprint.iacr.org/2011/232.pdf>
9193 *Billy Bob Brumley and Nicola Tuveri*
9195 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9197 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9199 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9201 * Fix bug in string printing code: if *any* escaping is enabled we must
9202 escape the escape character (backslash) or the resulting string is
9207 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9209 * Disable code workaround for ancient and obsolete Netscape browsers
9210 and servers: an attacker can use it in a ciphersuite downgrade attack.
9211 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9215 * Fixed J-PAKE implementation error, originally discovered by
9216 Sebastien Martini, further info and confirmation from Stefan
9217 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9221 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9223 * Fix extension code to avoid race conditions which can result in a buffer
9224 overrun vulnerability: resumed sessions must not be modified as they can
9225 be shared by multiple threads. CVE-2010-3864
9229 * Fix WIN32 build system to correctly link an ENGINE directory into
9234 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9236 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9239 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9241 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9243 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9244 context. The operation can be customised via the ctrl mechanism in
9245 case ENGINEs want to include additional functionality.
9249 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9253 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9254 output hashes compatible with older versions of OpenSSL.
9256 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9258 * Fix compression algorithm handling: if resuming a session use the
9259 compression algorithm of the resumed session instead of determining
9260 it from client hello again. Don't allow server to change algorithm.
9264 * Add load_crls() function to commands tidying load_certs() too. Add option
9265 to verify utility to allow additional CRLs to be included.
9269 * Update OCSP request code to permit adding custom headers to the request:
9270 some responders need this.
9274 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9277 *Julia Lawall <julia@diku.dk>*
9279 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9280 needlessly dereferenced structures, used obsolete functions and
9281 didn't handle all updated verify codes correctly.
9285 * Disable MD2 in the default configuration.
9289 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9290 indicate the initial BIO being pushed or popped. This makes it possible
9291 to determine whether the BIO is the one explicitly called or as a result
9292 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9293 it handles reference counts correctly and doesn't zero out the I/O bio
9294 when it is not being explicitly popped. WARNING: applications which
9295 included workarounds for the old buggy behaviour will need to be modified
9296 or they could free up already freed BIOs.
9300 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9301 renaming to all platforms (within the 0.9.8 branch, this was
9302 done conditionally on Netware platforms to avoid a name clash).
9304 *Guenter <lists@gknw.net>*
9306 * Add ECDHE and PSK support to DTLS.
9308 *Michael Tuexen <tuexen@fh-muenster.de>*
9310 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9315 * Add "missing" function EVP_MD_flags() (without this the only way to
9316 retrieve a digest flags is by accessing the structure directly. Update
9317 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9318 or cipher is registered as in the "from" argument. Print out all
9319 registered digests in the dgst usage message instead of manually
9320 attempting to work them out.
9324 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9325 this allows the use of compression and extensions. Change default cipher
9326 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9327 by default unless an application cipher string requests it.
9331 * Alter match criteria in PKCS12_parse(). It used to try to use local
9332 key ids to find matching certificates and keys but some PKCS#12 files
9333 don't follow the (somewhat unwritten) rules and this strategy fails.
9334 Now just gather all certificates together and the first private key
9335 then look for the first certificate that matches the key.
9339 * Support use of registered digest and cipher names for dgst and cipher
9340 commands instead of having to add each one as a special case. So now
9347 openssl dgst -sha256 foo
9349 and this works for ENGINE based algorithms too.
9353 * Update Gost ENGINE to support parameter files.
9355 *Victor B. Wagner <vitus@cryptocom.ru>*
9357 * Support GeneralizedTime in ca utility.
9359 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9361 * Enhance the hash format used for certificate directory links. The new
9362 form uses the canonical encoding (meaning equivalent names will work
9363 even if they aren't identical) and uses SHA1 instead of MD5. This form
9364 is incompatible with the older format and as a result c_rehash should
9365 be used to rebuild symbolic links.
9369 * Make PKCS#8 the default write format for private keys, replacing the
9370 traditional format. This form is standardised, more secure and doesn't
9371 include an implicit MD5 dependency.
9375 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9376 committed to OpenSSL should pass this lot as a minimum.
9380 * Add session ticket override functionality for use by EAP-FAST.
9382 *Jouni Malinen <j@w1.fi>*
9384 * Modify HMAC functions to return a value. Since these can be implemented
9385 in an ENGINE errors can occur.
9389 * Type-checked OBJ_bsearch_ex.
9393 * Type-checked OBJ_bsearch. Also some constification necessitated
9394 by type-checking. Still to come: TXT_DB, bsearch(?),
9395 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9400 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9401 seconds to a tm structure directly, instead of going through OS
9402 specific date routines. This avoids any issues with OS routines such
9403 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9404 and X509_time_adj_ex() to cover the extended range. The existing
9405 X509_time_adj() is still usable and will no longer have any date issues.
9409 * Delta CRL support. New use deltas option which will attempt to locate
9410 and search any appropriate delta CRLs available.
9412 This work was sponsored by Google.
9416 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9417 code and add additional score elements. Validate alternate CRL paths
9418 as part of the CRL checking and indicate a new error "CRL path validation
9419 error" in this case. Applications wanting additional details can use
9420 the verify callback and check the new "parent" field. If this is not
9421 NULL CRL path validation is taking place. Existing applications won't
9422 see this because it requires extended CRL support which is off by
9425 This work was sponsored by Google.
9429 * Support for freshest CRL extension.
9431 This work was sponsored by Google.
9435 * Initial indirect CRL support. Currently only supported in the CRLs
9436 passed directly and not via lookup. Process certificate issuer
9437 CRL entry extension and lookup CRL entries by bother issuer name
9438 and serial number. Check and process CRL issuer entry in IDP extension.
9440 This work was sponsored by Google.
9444 * Add support for distinct certificate and CRL paths. The CRL issuer
9445 certificate is validated separately in this case. Only enabled if
9446 an extended CRL support flag is set: this flag will enable additional
9447 CRL functionality in future.
9449 This work was sponsored by Google.
9453 * Add support for policy mappings extension.
9455 This work was sponsored by Google.
9459 * Fixes to pathlength constraint, self issued certificate handling,
9460 policy processing to align with RFC3280 and PKITS tests.
9462 This work was sponsored by Google.
9466 * Support for name constraints certificate extension. DN, email, DNS
9467 and URI types are currently supported.
9469 This work was sponsored by Google.
9473 * To cater for systems that provide a pointer-based thread ID rather
9474 than numeric, deprecate the current numeric thread ID mechanism and
9475 replace it with a structure and associated callback type. This
9476 mechanism allows a numeric "hash" to be extracted from a thread ID in
9477 either case, and on platforms where pointers are larger than 'long',
9478 mixing is done to help ensure the numeric 'hash' is usable even if it
9479 can't be guaranteed unique. The default mechanism is to use "&errno"
9480 as a pointer-based thread ID to distinguish between threads.
9482 Applications that want to provide their own thread IDs should now use
9483 CRYPTO_THREADID_set_callback() to register a callback that will call
9484 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9486 Note that ERR_remove_state() is now deprecated, because it is tied
9487 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9488 to free the current thread's error state should be replaced by
9489 ERR_remove_thread_state(NULL).
9491 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9492 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9493 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9494 application was previously providing a numeric thread callback that
9495 was inappropriate for distinguishing threads, then uniqueness might
9496 have been obtained with &errno that happened immediately in the
9497 intermediate development versions of OpenSSL; this is no longer the
9498 case, the numeric thread callback will now override the automatic use
9501 *Geoff Thorpe, with help from Bodo Moeller*
9503 * Initial support for different CRL issuing certificates. This covers a
9504 simple case where the self issued certificates in the chain exist and
9505 the real CRL issuer is higher in the existing chain.
9507 This work was sponsored by Google.
9511 * Removed effectively defunct crypto/store from the build.
9515 * Revamp of STACK to provide stronger type-checking. Still to come:
9516 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9517 ASN1_STRING, CONF_VALUE.
9521 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9522 RAM on SSL connections. This option can save about 34k per idle SSL.
9526 * Revamp of LHASH to provide stronger type-checking. Still to come:
9527 STACK, TXT_DB, bsearch, qsort.
9531 * Initial support for Cryptographic Message Syntax (aka CMS) based
9532 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9533 support for data, signedData, compressedData, digestedData and
9534 encryptedData, envelopedData types included. Scripts to check against
9535 RFC4134 examples draft and interop and consistency checks of many
9536 content types and variants.
9540 * Add options to enc utility to support use of zlib compression BIO.
9544 * Extend mk1mf to support importing of options and assembly language
9545 files from Configure script, currently only included in VC-WIN32.
9546 The assembly language rules can now optionally generate the source
9547 files from the associated perl scripts.
9551 * Implement remaining functionality needed to support GOST ciphersuites.
9552 Interop testing has been performed using CryptoPro implementations.
9554 *Victor B. Wagner <vitus@cryptocom.ru>*
9556 * s390x assembler pack.
9560 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9565 * Implement Opaque PRF Input TLS extension as specified in
9566 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9567 official specification yet and no extension type assignment by
9568 IANA exists, this extension (for now) will have to be explicitly
9569 enabled when building OpenSSL by providing the extension number
9570 to use. For example, specify an option
9572 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9574 to the "config" or "Configure" script to enable the extension,
9575 assuming extension number 0x9527 (which is a completely arbitrary
9576 and unofficial assignment based on the MD5 hash of the Internet
9577 Draft). Note that by doing so, you potentially lose
9578 interoperability with other TLS implementations since these might
9579 be using the same extension number for other purposes.
9581 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9582 opaque PRF input value to use in the handshake. This will create
9583 an internal copy of the length-'len' string at 'src', and will
9584 return non-zero for success.
9586 To get more control and flexibility, provide a callback function
9589 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9590 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9594 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9597 Callback function 'cb' will be called in handshakes, and is
9598 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9599 Argument 'arg' is for application purposes (the value as given to
9600 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9601 be provided to the callback function). The callback function
9602 has to return non-zero to report success: usually 1 to use opaque
9603 PRF input just if possible, or 2 to enforce use of the opaque PRF
9604 input. In the latter case, the library will abort the handshake
9605 if opaque PRF input is not successfully negotiated.
9607 Arguments 'peerinput' and 'len' given to the callback function
9608 will always be NULL and 0 in the case of a client. A server will
9609 see the client's opaque PRF input through these variables if
9610 available (NULL and 0 otherwise). Note that if the server
9611 provides an opaque PRF input, the length must be the same as the
9612 length of the client's opaque PRF input.
9614 Note that the callback function will only be called when creating
9615 a new session (session resumption can resume whatever was
9616 previously negotiated), and will not be called in SSL 2.0
9617 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9618 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9619 for applications that need to enforce opaque PRF input.
9623 * Update ssl code to support digests other than SHA1+MD5 for handshake
9626 *Victor B. Wagner <vitus@cryptocom.ru>*
9628 * Add RFC4507 support to OpenSSL. This includes the corrections in
9629 RFC4507bis. The encrypted ticket format is an encrypted encoded
9630 SSL_SESSION structure, that way new session features are automatically
9633 If a client application caches session in an SSL_SESSION structure
9634 support is transparent because tickets are now stored in the encoded
9637 The SSL_CTX structure automatically generates keys for ticket
9638 protection in servers so again support should be possible
9639 with no application modification.
9641 If a client or server wishes to disable RFC4507 support then the option
9642 SSL_OP_NO_TICKET can be set.
9644 Add a TLS extension debugging callback to allow the contents of any client
9645 or server extensions to be examined.
9647 This work was sponsored by Google.
9651 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9652 OpenSSL should now compile cleanly on gcc 4.2
9654 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9656 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9657 support including streaming MAC support: this is required for GOST
9658 ciphersuite support.
9660 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9662 * Add option -stream to use PKCS#7 streaming in smime utility. New
9663 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9664 to output in BER and PEM format.
9668 * Experimental support for use of HMAC via EVP_PKEY interface. This
9669 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9670 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9671 ENGINE support for HMAC keys which are unextractable. New -mac and
9672 -macopt options to dgst utility.
9676 * New option -sigopt to dgst utility. Update dgst to use
9677 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9678 alternative signing parameters such as X9.31 or PSS in the dgst
9683 * Change ssl_cipher_apply_rule(), the internal function that does
9684 the work each time a ciphersuite string requests enabling
9685 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9686 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9687 the order of disabled ciphersuites such that those ciphersuites
9688 that most recently went from enabled to disabled not only stay
9689 in order with respect to each other, but also have higher priority
9690 than other disabled ciphersuites the next time ciphersuites are
9693 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9694 the same ciphersuites as with "HIGH" alone, but in a specific
9695 order where the PSK ciphersuites come first (since they are the
9696 most recently disabled ciphersuites when "HIGH" is parsed).
9698 Also, change ssl_create_cipher_list() (using this new
9699 functionality) such that between otherwise identical
9700 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9705 * Change ssl_create_cipher_list() so that it automatically
9706 arranges the ciphersuites in reasonable order before starting
9707 to process the rule string. Thus, the definition for "DEFAULT"
9708 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9709 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9710 This makes it much easier to arrive at a reasonable default order
9711 in applications for which anonymous ciphers are OK (meaning
9712 that you can't actually use DEFAULT).
9714 *Bodo Moeller; suggested by Victor Duchovni*
9716 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9717 processing) into multiple integers instead of setting
9718 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9719 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9720 (These masks as well as the individual bit definitions are hidden
9721 away into the non-exported interface ssl/ssl_locl.h, so this
9722 change to the definition of the SSL_CIPHER structure shouldn't
9723 affect applications.) This give us more bits for each of these
9724 categories, so there is no longer a need to coagulate AES128 and
9725 AES256 into a single algorithm bit, and to coagulate Camellia128
9726 and Camellia256 into a single algorithm bit, which has led to all
9729 Thus, among other things, the kludge introduced in 0.9.7m and
9730 0.9.8e for masking out AES256 independently of AES128 or masking
9731 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9733 With the change, we also introduce new ciphersuite aliases that
9734 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9739 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9740 Use the leftmost N bytes of the signature input if the input is
9741 larger than the prime q (with N being the size in bytes of q).
9745 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9746 it yet and it is largely untested.
9750 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9754 * Initial incomplete changes to avoid need for function casts in OpenSSL
9755 some compilers (gcc 4.2 and later) reject their use. Safestack is
9756 reimplemented. Update ASN1 to avoid use of legacy functions.
9760 * Win32/64 targets are linked with Winsock2.
9764 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9765 to external functions. This can be used to increase CRL handling
9766 efficiency especially when CRLs are very large by (for example) storing
9767 the CRL revoked certificates in a database.
9771 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9772 new CRLs added to a directory can be used. New command line option
9773 -verify_return_error to s_client and s_server. This causes real errors
9774 to be returned by the verify callback instead of carrying on no matter
9775 what. This reflects the way a "real world" verify callback would behave.
9779 * GOST engine, supporting several GOST algorithms and public key formats.
9780 Kindly donated by Cryptocom.
9784 * Partial support for Issuing Distribution Point CRL extension. CRLs
9785 partitioned by DP are handled but no indirect CRL or reason partitioning
9786 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9787 selected via a scoring technique which handles IDP and AKID in CRLs.
9791 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9792 will ultimately be used for all verify operations: this will remove the
9793 X509_STORE dependency on certificate verification and allow alternative
9794 lookup methods. X509_STORE based implementations of these two callbacks.
9798 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9799 Modify get_crl() to find a valid (unexpired) CRL if possible.
9803 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9804 this would be called X509_CRL_cmp() but that name is already used by
9805 a function that just compares CRL issuer names. Cache several CRL
9806 extensions in X509_CRL structure and cache CRLDP in X509.
9810 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9811 this maps equivalent X509_NAME structures into a consistent structure.
9812 Name comparison can then be performed rapidly using memcmp().
9816 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9821 * Allow digests to supply their own micalg string for S/MIME type using
9822 the ctrl EVP_MD_CTRL_MICALG.
9826 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9827 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9828 ctrl. It can then customise the structure before and/or after signing
9833 * New function OBJ_add_sigid() to allow application defined signature OIDs
9834 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9835 to free up any added signature OIDs.
9839 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9840 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9841 digest and cipher tables. New options added to openssl utility:
9842 list-message-digest-algorithms and list-cipher-algorithms.
9846 * Change the array representation of binary polynomials: the list
9847 of degrees of non-zero coefficients is now terminated with -1.
9848 Previously it was terminated with 0, which was also part of the
9849 value; thus, the array representation was not applicable to
9850 polynomials where t^0 has coefficient zero. This change makes
9851 the array representation useful in a more general context.
9855 * Various modifications and fixes to SSL/TLS cipher string
9856 handling. For ECC, the code now distinguishes between fixed ECDH
9857 with RSA certificates on the one hand and with ECDSA certificates
9858 on the other hand, since these are separate ciphersuites. The
9859 unused code for Fortezza ciphersuites has been removed.
9861 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9862 (not "ECDHE"). For consistency with the code for DH
9863 certificates, use of ECDH certificates is now considered ECDH
9864 authentication, not RSA or ECDSA authentication (the latter is
9865 merely the CA's signing algorithm and not actively used in the
9868 The temporary ciphersuite alias "ECCdraft" is no longer
9869 available, and ECC ciphersuites are no longer excluded from "ALL"
9870 and "DEFAULT". The following aliases now exist for RFC 4492
9871 ciphersuites, most of these by analogy with the DH case:
9873 kECDHr - ECDH cert, signed with RSA
9874 kECDHe - ECDH cert, signed with ECDSA
9875 kECDH - ECDH cert (signed with either RSA or ECDSA)
9876 kEECDH - ephemeral ECDH
9877 ECDH - ECDH cert or ephemeral ECDH
9883 AECDH - anonymous ECDH
9884 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9888 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9889 Use correct micalg parameters depending on digest(s) in signed message.
9893 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9894 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9898 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9899 an engine to register a method. Add ENGINE lookups for methods and
9900 functional reference processing.
9904 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9905 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9910 * New -resign option to smime utility. This adds one or more signers
9911 to an existing PKCS#7 signedData structure. Also -md option to use an
9912 alternative message digest algorithm for signing.
9916 * Tidy up PKCS#7 routines and add new functions to make it easier to
9917 create PKCS7 structures containing multiple signers. Update smime
9918 application to support multiple signers.
9922 * New -macalg option to pkcs12 utility to allow setting of an alternative
9927 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9928 Reorganize PBE internals to lookup from a static table using NIDs,
9929 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9930 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9931 PRF which will be automatically used with PBES2.
9935 * Replace the algorithm specific calls to generate keys in "req" with the
9940 * Update PKCS#7 enveloped data routines to use new API. This is now
9941 supported by any public key method supporting the encrypt operation. A
9942 ctrl is added to allow the public key algorithm to examine or modify
9943 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9948 * Add a ctrl to asn1 method to allow a public key algorithm to express
9949 a default digest type to use. In most cases this will be SHA1 but some
9950 algorithms (such as GOST) need to specify an alternative digest. The
9951 return value indicates how strong the preference is 1 means optional and
9952 2 is mandatory (that is it is the only supported type). Modify
9953 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9954 use the default md. Update openssl utilities to use the default digest
9955 type for signing if it is not explicitly indicated.
9959 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9960 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9961 signing method from the key type. This effectively removes the link
9962 between digests and public key types.
9966 * Add an OID cross reference table and utility functions. Its purpose is to
9967 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9968 rsaEncryption. This will allow some of the algorithm specific hackery
9969 needed to use the correct OID to be removed.
9973 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9974 structures for PKCS7_sign(). They are now set up by the relevant public
9979 * Add provisional EC pkey method with support for ECDSA and ECDH.
9983 * Add support for key derivation (agreement) in the API, DH method and
9988 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9989 public and private key formats. As a side effect these add additional
9990 command line functionality not previously available: DSA signatures can be
9991 generated and verified using pkeyutl and DH key support and generation in
9998 *Oliver Tappe <zooey@hirschkaefer.de>*
10000 * New make target "install_html_docs" installs HTML renditions of the
10003 *Oliver Tappe <zooey@hirschkaefer.de>*
10005 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10006 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10007 support key and parameter generation and add initial key generation
10008 functionality for RSA.
10012 * Add functions for main EVP_PKEY_method operations. The undocumented
10013 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10014 `EVP_PKEY_{encrypt,decrypt}_old`.
10018 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10019 key API, doesn't do much yet.
10023 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10024 public key algorithms. New option to openssl utility:
10025 "list-public-key-algorithms" to print out info.
10029 * Implement the Supported Elliptic Curves Extension for
10030 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10034 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10035 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10039 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10040 utilities such as rsa, dsa, dsaparam etc except they process any key
10045 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10046 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10047 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10052 * Initial support for pluggable public key ASN1.
10053 De-spaghettify the public key ASN1 handling. Move public and private
10054 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10055 algorithm specific handling to a single module within the relevant
10056 algorithm directory. Add functions to allow (near) opaque processing
10057 of public and private key structures.
10061 * Implement the Supported Point Formats Extension for
10062 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10066 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10067 for the psk identity [hint] and the psk callback functions to the
10068 SSL_SESSION, SSL and SSL_CTX structure.
10071 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10075 SSL_CTX_use_psk_identity_hint
10076 SSL_get_psk_identity_hint
10077 SSL_get_psk_identity
10078 SSL_use_psk_identity_hint
10080 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10082 * Add RFC 3161 compliant time stamp request creation, response generation
10083 and response verification functionality.
10085 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10087 * Add initial support for TLS extensions, specifically for the server_name
10088 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10089 have new members for a hostname. The SSL data structure has an
10090 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10091 stored in that context to allow for session resumption, even after the
10092 SSL has been switched to a new SSL_CTX in reaction to a client's
10093 server_name extension.
10095 New functions (subject to change):
10097 SSL_get_servername()
10098 SSL_get_servername_type()
10101 New CTRL codes and macros (subject to change):
10103 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10104 - SSL_CTX_set_tlsext_servername_callback()
10105 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10106 - SSL_CTX_set_tlsext_servername_arg()
10107 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10109 openssl s_client has a new '-servername ...' option.
10111 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10112 '-key2 ...', '-servername_fatal' (subject to change). This allows
10113 testing the HostName extension for a specific single hostname ('-cert'
10114 and '-key' remain fallbacks for handshakes without HostName
10115 negotiation). If the unrecognized_name alert has to be sent, this by
10116 default is a warning; it becomes fatal with the '-servername_fatal'
10119 *Peter Sylvester, Remy Allais, Christophe Renou*
10121 * Whirlpool hash implementation is added.
10125 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10126 bn(64,32). Because of instruction set limitations it doesn't have
10127 any negative impact on performance. This was done mostly in order
10128 to make it possible to share assembler modules, such as bn_mul_mont
10129 implementations, between 32- and 64-bit builds without hassle.
10133 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10134 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10139 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10140 dedicated Montgomery multiplication procedure, is introduced.
10141 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10142 "64-bit" performance on certain 32-bit targets.
10146 * New option SSL_OP_NO_COMP to disable use of compression selectively
10147 in SSL structures. New SSL ctrl to set maximum send fragment size.
10148 Save memory by setting the I/O buffer sizes dynamically instead of
10149 using the maximum available value.
10153 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10154 in addition to the text details.
10158 * Very, very preliminary EXPERIMENTAL support for printing of general
10159 ASN1 structures. This currently produces rather ugly output and doesn't
10160 handle several customised structures at all.
10164 * Integrated support for PVK file format and some related formats such
10165 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10166 these in the 'rsa' and 'dsa' utilities.
10170 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10174 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10175 place for the (very old) "NETSCAPE" format certificates which are now
10176 handled using new ASN1 code equivalents.
10180 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10181 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10182 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10186 * Modify CRL distribution points extension code to print out previously
10187 unsupported fields. Enhance extension setting code to allow setting of
10192 * Add print and set support for Issuing Distribution Point CRL extension.
10196 * Change 'Configure' script to enable Camellia by default.
10203 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10205 * When rejecting SSL/TLS records due to an incorrect version number, never
10206 update s->server with a new major version number. As of
10207 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10208 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10209 the previous behavior could result in a read attempt at NULL when
10210 receiving specific incorrect SSL/TLS records once record payload
10211 protection is active. ([CVE-2010-0740])
10213 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10215 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10216 could be crashed if the relevant tables were not present (e.g. chrooted).
10218 *Tomas Hoger <thoger@redhat.com>*
10220 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10222 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10224 *Martin Olsson, Neel Mehta*
10226 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10227 accommodate for stack sorting, always a write lock!).
10231 * On some versions of WIN32 Heap32Next is very slow. This can cause
10232 excessive delays in the RAND_poll(): over a minute. As a workaround
10233 include a time check in the inner Heap32Next loop too.
10237 * The code that handled flushing of data in SSL/TLS originally used the
10238 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10239 the problem outlined in PR#1949. The fix suggested there however can
10240 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10241 of Apache). So instead simplify the code to flush unconditionally.
10242 This should be fine since flushing with no data to flush is a no op.
10246 * Handle TLS versions 2.0 and later properly and correctly use the
10247 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10248 off ancient servers have a habit of sticking around for a while...
10252 * Modify compression code so it frees up structures without using the
10253 ex_data callbacks. This works around a problem where some applications
10254 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10255 restarting) then use compression (e.g. SSL with compression) later.
10256 This results in significant per-connection memory leaks and
10257 has caused some security issues including CVE-2008-1678 and
10262 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10263 change when encrypting or decrypting.
10267 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10268 connect and renegotiate with servers which do not support RI.
10269 Until RI is more widely deployed this option is enabled by default.
10273 * Add "missing" ssl ctrls to clear options and mode.
10277 * If client attempts to renegotiate and doesn't support RI respond with
10278 a no_renegotiation alert as required by RFC5746. Some renegotiating
10279 TLS clients will continue a connection gracefully when they receive
10280 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10281 waiting for a server hello which it will never receive. Now we treat a
10282 received no_renegotiation alert as a fatal error. This is because
10283 applications requesting a renegotiation might well expect it to succeed
10284 and would have no code in place to handle the server denying it so the
10285 only safe thing to do is to terminate the connection.
10289 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10290 peer supports secure renegotiation and 0 otherwise. Print out peer
10291 renegotiation support in s_client/s_server.
10295 * Replace the highly broken and deprecated SPKAC certification method with
10296 the updated NID creation version. This should correctly handle UTF8.
10300 * Implement RFC5746. Re-enable renegotiation but require the extension
10301 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10302 turns out to be a bad idea. It has been replaced by
10303 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10304 SSL_CTX_set_options(). This is really not recommended unless you
10305 know what you are doing.
10307 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10309 * Fixes to stateless session resumption handling. Use initial_ctx when
10310 issuing and attempting to decrypt tickets in case it has changed during
10311 servername handling. Use a non-zero length session ID when attempting
10312 stateless session resumption: this makes it possible to determine if
10313 a resumption has occurred immediately after receiving server hello
10314 (several places in OpenSSL subtly assume this) instead of later in
10319 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10320 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10321 fixes for a few places where the return code is not checked
10324 *Julia Lawall <julia@diku.dk>*
10326 * Add --strict-warnings option to Configure script to include devteam
10327 warnings in other configurations.
10331 * Add support for --libdir option and LIBDIR variable in makefiles. This
10332 makes it possible to install openssl libraries in locations which
10333 have names other than "lib", for example "/usr/lib64" which some
10336 *Steve Henson, based on patch from Jeremy Utley*
10338 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10339 X690 8.9.12 and can produce some misleading textual output of OIDs.
10341 *Steve Henson, reported by Dan Kaminsky*
10343 * Delete MD2 from algorithm tables. This follows the recommendation in
10344 several standards that it is not used in new applications due to
10345 several cryptographic weaknesses. For binary compatibility reasons
10346 the MD2 API is still compiled in by default.
10350 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10355 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10356 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10359 *Guenter <lists@gknw.net>*
10361 * Fix the server certificate chain building code to use X509_verify_cert(),
10362 it used to have an ad-hoc builder which was unable to cope with anything
10363 other than a simple chain.
10365 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10367 * Don't check self signed certificate signatures in X509_verify_cert()
10368 by default (a flag can override this): it just wastes time without
10369 adding any security. As a useful side effect self signed root CAs
10370 with non-FIPS digests are now usable in FIPS mode.
10374 * In dtls1_process_out_of_seq_message() the check if the current message
10375 is already buffered was missing. For every new message was memory
10376 allocated, allowing an attacker to perform an denial of service attack
10377 with sending out of seq handshake messages until there is no memory
10378 left. Additionally every future message was buffered, even if the
10379 sequence number made no sense and would be part of another handshake.
10380 So only messages with sequence numbers less than 10 in advance will be
10381 buffered. ([CVE-2009-1378])
10383 *Robin Seggelmann, discovered by Daniel Mentz*
10385 * Records are buffered if they arrive with a future epoch to be
10386 processed after finishing the corresponding handshake. There is
10387 currently no limitation to this buffer allowing an attacker to perform
10388 a DOS attack with sending records with future epochs until there is no
10389 memory left. This patch adds the pqueue_size() function to determine
10390 the size of a buffer and limits the record buffer to 100 entries.
10393 *Robin Seggelmann, discovered by Daniel Mentz*
10395 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10396 parent structure is freed. ([CVE-2009-1379])
10400 * Handle non-blocking I/O properly in SSL_shutdown() call.
10402 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10404 * Add `2.5.4.*` OIDs
10406 *Ilya O. <vrghost@gmail.com>*
10408 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10410 * Disable renegotiation completely - this fixes a severe security
10411 problem ([CVE-2009-3555]) at the cost of breaking all
10412 renegotiation. Renegotiation can be re-enabled by setting
10413 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10414 run-time. This is really not recommended unless you know what
10419 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10421 * Don't set val to NULL when freeing up structures, it is freed up by
10422 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10423 zeroing past the valid field. ([CVE-2009-0789])
10425 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10427 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10428 checked correctly. This would allow some invalid signed attributes to
10429 appear to verify correctly. ([CVE-2009-0591])
10431 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10433 * Reject UniversalString and BMPString types with invalid lengths. This
10434 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10435 a legal length. ([CVE-2009-0590])
10439 * Set S/MIME signing as the default purpose rather than setting it
10440 unconditionally. This allows applications to override it at the store
10445 * Permit restricted recursion of ASN1 strings. This is needed in practice
10446 to handle some structures.
10450 * Improve efficiency of mem_gets: don't search whole buffer each time
10453 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10455 * New -hex option for openssl rand.
10459 * Print out UTF8String and NumericString when parsing ASN1.
10463 * Support NumericString type for name components.
10467 * Allow CC in the environment to override the automatically chosen
10468 compiler. Note that nothing is done to ensure flags work with the
10473 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10475 * Properly check EVP_VerifyFinal() and similar return values
10478 *Ben Laurie, Bodo Moeller, Google Security Team*
10480 * Enable TLS extensions by default.
10484 * Allow the CHIL engine to be loaded, whether the application is
10485 multithreaded or not. (This does not release the developer from the
10486 obligation to set up the dynamic locking callbacks.)
10488 *Sander Temme <sander@temme.net>*
10490 * Use correct exit code if there is an error in dgst command.
10492 *Steve Henson; problem pointed out by Roland Dirlewanger*
10494 * Tweak Configure so that you need to say "experimental-jpake" to enable
10495 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10499 * Add experimental JPAKE support, including demo authentication in
10500 s_client and s_server.
10504 * Set the comparison function in v3_addr_canonize().
10506 *Rob Austein <sra@hactrn.net>*
10508 * Add support for XMPP STARTTLS in s_client.
10510 *Philip Paeps <philip@freebsd.org>*
10512 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10513 to ensure that even with this option, only ciphersuites in the
10514 server's preference list will be accepted. (Note that the option
10515 applies only when resuming a session, so the earlier behavior was
10516 just about the algorithm choice for symmetric cryptography.)
10520 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10522 * Fix NULL pointer dereference if a DTLS server received
10523 ChangeCipherSpec as first record ([CVE-2009-1386]).
10527 * Fix a state transition in s3_srvr.c and d1_srvr.c
10528 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10530 *Nagendra Modadugu*
10532 * The fix in 0.9.8c that supposedly got rid of unsafe
10533 double-checked locking was incomplete for RSA blinding,
10534 addressing just one layer of what turns out to have been
10535 doubly unsafe triple-checked locking.
10537 So now fix this for real by retiring the MONT_HELPER macro
10538 in crypto/rsa/rsa_eay.c.
10540 *Bodo Moeller; problem pointed out by Marius Schilder*
10542 * Various precautionary measures:
10544 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10546 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10547 (NB: This would require knowledge of the secret session ticket key
10548 to exploit, in which case you'd be SOL either way.)
10550 - Change bn_nist.c so that it will properly handle input BIGNUMs
10551 outside the expected range.
10553 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10556 *Neel Mehta, Bodo Moeller*
10558 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10559 the load fails. Useful for distros.
10561 *Ben Laurie and the FreeBSD team*
10563 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10567 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10571 * Expand ENGINE to support engine supplied SSL client certificate functions.
10573 This work was sponsored by Logica.
10577 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10578 keystores. Support for SSL/TLS client authentication too.
10579 Not compiled unless enable-capieng specified to Configure.
10581 This work was sponsored by Logica.
10585 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10586 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10587 attribute creation routines such as certificate requests and PKCS#12
10592 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10594 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10595 handshake which could lead to a client crash as found using the
10596 Codenomicon TLS test suite ([CVE-2008-1672])
10598 *Steve Henson, Mark Cox*
10600 * Fix double free in TLS server name extensions which could lead to
10601 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10605 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10607 Clear the error queue to ensure that error entries left from
10608 older function calls do not interfere with the correct operation.
10610 *Lutz Jaenicke, Erik de Castro Lopo*
10612 * Remove root CA certificates of commercial CAs:
10614 The OpenSSL project does not recommend any specific CA and does not
10615 have any policy with respect to including or excluding any CA.
10616 Therefore, it does not make any sense to ship an arbitrary selection
10617 of root CA certificates with the OpenSSL software.
10621 * RSA OAEP patches to fix two separate invalid memory reads.
10622 The first one involves inputs when 'lzero' is greater than
10623 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10624 before the beginning of from). The second one involves inputs where
10625 the 'db' section contains nothing but zeroes (there is a one-byte
10626 invalid read after the end of 'db').
10628 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10630 * Partial backport from 0.9.9-dev:
10632 Introduce bn_mul_mont (dedicated Montgomery multiplication
10633 procedure) as a candidate for BIGNUM assembler implementation.
10634 While 0.9.9-dev uses assembler for various architectures, only
10635 x86_64 is available by default here in the 0.9.8 branch, and
10636 32-bit x86 is available through a compile-time setting.
10638 To try the 32-bit x86 assembler implementation, use Configure
10639 option "enable-montasm" (which exists only for this backport).
10641 As "enable-montasm" for 32-bit x86 disclaims code stability
10642 anyway, in this constellation we activate additional code
10643 backported from 0.9.9-dev for further performance improvements,
10644 namely BN_from_montgomery_word. (To enable this otherwise,
10645 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10647 *Andy Polyakov (backport partially by Bodo Moeller)*
10649 * Add TLS session ticket callback. This allows an application to set
10650 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10651 values. This is useful for key rollover for example where several key
10652 sets may exist with different names.
10656 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10657 This was broken until now in 0.9.8 releases, such that the only way
10658 a registered ENGINE could be used (assuming it initialises
10659 successfully on the host) was to explicitly set it as the default
10660 for the relevant algorithms. This is in contradiction with 0.9.7
10661 behaviour and the documentation. With this fix, when an ENGINE is
10662 registered into a given algorithm's table of implementations, the
10663 'uptodate' flag is reset so that auto-discovery will be used next
10664 time a new context for that algorithm attempts to select an
10667 *Ian Lister (tweaked by Geoff Thorpe)*
10669 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10670 implementation in the following ways:
10672 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10675 Lack of BER streaming support means one pass streaming processing is
10676 only supported if data is detached: setting the streaming flag is
10677 ignored for embedded content.
10679 CMS support is disabled by default and must be explicitly enabled
10680 with the enable-cms configuration option.
10684 * Update the GMP engine glue to do direct copies between BIGNUM and
10685 mpz_t when openssl and GMP use the same limb size. Otherwise the
10686 existing "conversion via a text string export" trick is still used.
10688 *Paul Sheer <paulsheer@gmail.com>*
10690 * Zlib compression BIO. This is a filter BIO which compressed and
10691 uncompresses any data passed through it.
10695 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10696 RFC3394 compatible AES key wrapping.
10700 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10701 sets string data without copying. X509_ALGOR_set0() and
10702 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10703 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10704 from an X509_ATTRIBUTE structure optionally checking it occurs only
10705 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10710 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10711 to get the expected BN_FLG_CONSTTIME behavior.
10713 *Bodo Moeller (Google)*
10717 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10718 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10719 - added some more tests to do_tests.pl
10720 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10721 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10722 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10723 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10724 - various changes to netware.pl to enable gcc-cross builds on Win32
10726 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10727 - various changes to fix missing prototype warnings
10728 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10729 - added AES, WHIRLPOOL and CPUID assembler code to build files
10730 - added missing AES assembler make rules to mk1mf.pl
10731 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10733 *Guenter Knauf <eflash@gmx.net>*
10735 * Implement certificate status request TLS extension defined in RFC3546.
10736 A client can set the appropriate parameters and receive the encoded
10737 OCSP response via a callback. A server can query the supplied parameters
10738 and set the encoded OCSP response in the callback. Add simplified examples
10739 to s_client and s_server.
10743 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10745 * Fix various bugs:
10746 + Binary incompatibility of ssl_ctx_st structure
10747 + DTLS interoperation with non-compliant servers
10748 + Don't call get_session_cb() without proposed session
10749 + Fix ia64 assembler code
10751 *Andy Polyakov, Steve Henson*
10753 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10755 * DTLS Handshake overhaul. There were longstanding issues with
10756 OpenSSL DTLS implementation, which were making it impossible for
10757 RFC 4347 compliant client to communicate with OpenSSL server.
10758 Unfortunately just fixing these incompatibilities would "cut off"
10759 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10760 server keeps tolerating non RFC compliant syntax. The opposite is
10761 not true, 0.9.8f client can not communicate with earlier server.
10762 This update even addresses CVE-2007-4995.
10766 * Changes to avoid need for function casts in OpenSSL: some compilers
10767 (gcc 4.2 and later) reject their use.
10768 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10771 * Add RFC4507 support to OpenSSL. This includes the corrections in
10772 RFC4507bis. The encrypted ticket format is an encrypted encoded
10773 SSL_SESSION structure, that way new session features are automatically
10776 If a client application caches session in an SSL_SESSION structure
10777 support is transparent because tickets are now stored in the encoded
10780 The SSL_CTX structure automatically generates keys for ticket
10781 protection in servers so again support should be possible
10782 with no application modification.
10784 If a client or server wishes to disable RFC4507 support then the option
10785 SSL_OP_NO_TICKET can be set.
10787 Add a TLS extension debugging callback to allow the contents of any client
10788 or server extensions to be examined.
10790 This work was sponsored by Google.
10794 * Add initial support for TLS extensions, specifically for the server_name
10795 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10796 have new members for a hostname. The SSL data structure has an
10797 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10798 stored in that context to allow for session resumption, even after the
10799 SSL has been switched to a new SSL_CTX in reaction to a client's
10800 server_name extension.
10802 New functions (subject to change):
10804 SSL_get_servername()
10805 SSL_get_servername_type()
10808 New CTRL codes and macros (subject to change):
10810 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10811 - SSL_CTX_set_tlsext_servername_callback()
10812 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10813 - SSL_CTX_set_tlsext_servername_arg()
10814 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10816 openssl s_client has a new '-servername ...' option.
10818 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10819 '-key2 ...', '-servername_fatal' (subject to change). This allows
10820 testing the HostName extension for a specific single hostname ('-cert'
10821 and '-key' remain fallbacks for handshakes without HostName
10822 negotiation). If the unrecognized_name alert has to be sent, this by
10823 default is a warning; it becomes fatal with the '-servername_fatal'
10826 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10828 * Add AES and SSE2 assembly language support to VC++ build.
10832 * Mitigate attack on final subtraction in Montgomery reduction.
10836 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10837 (which previously caused an internal error).
10841 * Squeeze another 10% out of IGE mode when in != out.
10845 * AES IGE mode speedup.
10847 *Dean Gaudet (Google)*
10849 * Add the Korean symmetric 128-bit cipher SEED (see
10850 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10851 add SEED ciphersuites from RFC 4162:
10853 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10854 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10855 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10856 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10858 To minimize changes between patchlevels in the OpenSSL 0.9.8
10859 series, SEED remains excluded from compilation unless OpenSSL
10860 is configured with 'enable-seed'.
10862 *KISA, Bodo Moeller*
10864 * Mitigate branch prediction attacks, which can be practical if a
10865 single processor is shared, allowing a spy process to extract
10866 information. For detailed background information, see
10867 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10868 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10869 and Necessary Software Countermeasures"). The core of the change
10870 are new versions BN_div_no_branch() and
10871 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10872 respectively, which are slower, but avoid the security-relevant
10873 conditional branches. These are automatically called by BN_div()
10874 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10875 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10876 remove a conditional branch.
10878 BN_FLG_CONSTTIME is the new name for the previous
10879 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10880 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10881 in the exponent causes BN_mod_exp_mont() to use the alternative
10882 implementation in BN_mod_exp_mont_consttime().) The old name
10883 remains as a deprecated alias.
10885 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10886 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10887 constant-time implementations for more than just exponentiation.
10888 Here too the old name is kept as a deprecated alias.
10890 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10891 the BN_BLINDING structure gets an independent copy of the
10892 modulus. This means that the previous `BIGNUM *m` argument to
10893 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10894 essentially becomes `const BIGNUM *m`, although we can't actually
10895 change this in the header file before 0.9.9. It allows
10896 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10897 enable BN_FLG_CONSTTIME.
10899 *Matthew D Wood (Intel Corp)*
10901 * In the SSL/TLS server implementation, be strict about session ID
10902 context matching (which matters if an application uses a single
10903 external cache for different purposes). Previously,
10904 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10905 set. This did ensure strict client verification, but meant that,
10906 with applications using a single external cache for quite
10907 different requirements, clients could circumvent ciphersuite
10908 restrictions for a given session ID context by starting a session
10909 in a different context.
10913 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10914 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10915 authentication-only ciphersuites.
10919 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10920 not complete and could lead to a possible single byte overflow
10921 ([CVE-2007-5135]) [Ben Laurie]
10923 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10925 * Since AES128 and AES256 (and similarly Camellia128 and
10926 Camellia256) share a single mask bit in the logic of
10927 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10928 kludge to work properly if AES128 is available and AES256 isn't
10929 (or if Camellia128 is available and Camellia256 isn't).
10933 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10934 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10935 When a point or a seed is encoded in a BIT STRING, we need to
10936 prevent the removal of trailing zero bits to get the proper DER
10937 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10938 of a NamedBitList, for which trailing 0 bits need to be removed.)
10942 * Have SSL/TLS server implementation tolerate "mismatched" record
10943 protocol version while receiving ClientHello even if the
10944 ClientHello is fragmented. (The server can't insist on the
10945 particular protocol version it has chosen before the ServerHello
10946 message has informed the client about his choice.)
10950 * Add RFC 3779 support.
10952 *Rob Austein for ARIN, Ben Laurie*
10954 * Load error codes if they are not already present instead of using a
10955 static variable. This allows them to be cleanly unloaded and reloaded.
10956 Improve header file function name parsing.
10960 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10961 or CAPABILITY handshake as required by RFCs.
10963 *Goetz Babin-Ebell*
10965 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10967 * Introduce limits to prevent malicious keys being able to
10968 cause a denial of service. ([CVE-2006-2940])
10970 *Steve Henson, Bodo Moeller*
10972 * Fix ASN.1 parsing of certain invalid structures that can result
10973 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10975 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10976 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10978 * Fix SSL client code which could crash if connecting to a
10979 malicious SSLv2 server. ([CVE-2006-4343])
10981 *Tavis Ormandy and Will Drewry, Google Security Team*
10983 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10984 match only those. Before that, "AES256-SHA" would be interpreted
10985 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10986 the same strength classification in 0.9.7h) as we currently only
10987 have a single AES bit in the ciphersuite description bitmap.
10988 That change, however, also applied to ciphersuite strings such as
10989 "RC4-MD5" that intentionally matched multiple ciphersuites --
10990 namely, SSL 2.0 ciphersuites in addition to the more common ones
10991 from SSL 3.0/TLS 1.0.
10993 So we change the selection algorithm again: Naming an explicit
10994 ciphersuite selects this one ciphersuite, and any other similar
10995 ciphersuite (same bitmap) from *other* protocol versions.
10996 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10997 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10999 Since SSL 2.0 does not have any ciphersuites for which the
11000 128/256 bit distinction would be relevant, this works for now.
11001 The proper fix will be to use different bits for AES128 and
11002 AES256, which would have avoided the problems from the beginning;
11003 however, bits are scarce, so we can only do this in a new release
11004 (not just a patchlevel) when we can change the SSL_CIPHER
11005 definition to split the single 'unsigned long mask' bitmap into
11006 multiple values to extend the available space.
11010 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11012 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11013 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11015 * Add AES IGE and biIGE modes.
11019 * Change the Unix randomness entropy gathering to use poll() when
11020 possible instead of select(), since the latter has some
11021 undesirable limitations.
11023 *Darryl Miles via Richard Levitte and Bodo Moeller*
11025 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11026 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11027 cannot be implicitly activated as part of, e.g., the "AES" alias.
11028 However, please upgrade to OpenSSL 0.9.9[-dev] for
11029 non-experimental use of the ECC ciphersuites to get TLS extension
11030 support, which is required for curve and point format negotiation
11031 to avoid potential handshake problems.
11035 * Disable rogue ciphersuites:
11037 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11038 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11039 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11041 The latter two were purportedly from
11042 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11045 Also deactivate the remaining ciphersuites from
11046 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11047 unofficial, and the ID has long expired.
11051 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11052 dual-core machines) and other potential thread-safety issues.
11056 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11057 versions), which is now available for royalty-free use
11058 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11059 Also, add Camellia TLS ciphersuites from RFC 4132.
11061 To minimize changes between patchlevels in the OpenSSL 0.9.8
11062 series, Camellia remains excluded from compilation unless OpenSSL
11063 is configured with 'enable-camellia'.
11067 * Disable the padding bug check when compression is in use. The padding
11068 bug check assumes the first packet is of even length, this is not
11069 necessarily true if compression is enabled and can result in false
11070 positives causing handshake failure. The actual bug test is ancient
11071 code so it is hoped that implementations will either have fixed it by
11072 now or any which still have the bug do not support compression.
11076 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11078 * When applying a cipher rule check to see if string match is an explicit
11079 cipher suite and only match that one cipher suite if it is.
11083 * Link in manifests for VC++ if needed.
11085 *Austin Ziegler <halostatue@gmail.com>*
11087 * Update support for ECC-based TLS ciphersuites according to
11088 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11089 TLS extensions, which are supported starting with the 0.9.9
11090 branch, not in the OpenSSL 0.9.8 branch).
11094 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11095 opaque EVP_CIPHER_CTX handling.
11099 * Fixes and enhancements to zlib compression code. We now only use
11100 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11101 to conform with the standards mentioned here:
11102 <http://www.zlib.net/DLL_FAQ.txt>
11103 Static zlib linking now works on Windows and the new --with-zlib-include
11104 --with-zlib-lib options to Configure can be used to supply the location
11105 of the headers and library. Gracefully handle case where zlib library
11110 * Several fixes and enhancements to the OID generation code. The old code
11111 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11112 handle numbers larger than ULONG_MAX, truncated printing and had a
11113 non standard OBJ_obj2txt() behaviour.
11117 * Add support for building of engines under engine/ as shared libraries
11118 under VC++ build system.
11122 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11123 Hopefully, we will not see any false combination of paths any more.
11127 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11129 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11130 (part of SSL_OP_ALL). This option used to disable the
11131 countermeasure against man-in-the-middle protocol-version
11132 rollback in the SSL 2.0 server implementation, which is a bad
11133 idea. ([CVE-2005-2969])
11135 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11136 for Information Security, National Institute of Advanced Industrial
11137 Science and Technology [AIST], Japan)*
11139 * Add two function to clear and return the verify parameter flags.
11143 * Keep cipherlists sorted in the source instead of sorting them at
11144 runtime, thus removing the need for a lock.
11148 * Avoid some small subgroup attacks in Diffie-Hellman.
11150 *Nick Mathewson and Ben Laurie*
11152 * Add functions for well-known primes.
11156 * Extended Windows CE support.
11158 *Satoshi Nakamura and Andy Polyakov*
11160 * Initialize SSL_METHOD structures at compile time instead of during
11161 runtime, thus removing the need for a lock.
11165 * Make PKCS7_decrypt() work even if no certificate is supplied by
11166 attempting to decrypt each encrypted key in turn. Add support to
11171 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11173 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11176 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11180 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11181 key into the same file any more.
11185 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11189 * Add -utf8 command line and config file option to 'ca'.
11191 *Stefan <stf@udoma.org*
11193 * Removed the macro des_crypt(), as it seems to conflict with some
11194 libraries. Use DES_crypt().
11198 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11199 involves renaming the source and generated shared-libs for
11200 both. The engines will accept the corrected or legacy ids
11201 ('ncipher' and '4758_cca' respectively) when binding. NB,
11202 this only applies when building 'shared'.
11204 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11206 * Add attribute functions to EVP_PKEY structure. Modify
11207 PKCS12_create() to recognize a CSP name attribute and
11208 use it. Make -CSP option work again in pkcs12 utility.
11212 * Add new functionality to the bn blinding code:
11213 - automatic re-creation of the BN_BLINDING parameters after
11214 a fixed number of uses (currently 32)
11215 - add new function for parameter creation
11216 - introduce flags to control the update behaviour of the
11217 BN_BLINDING parameters
11218 - hide BN_BLINDING structure
11219 Add a second BN_BLINDING slot to the RSA structure to improve
11220 performance when a single RSA object is shared among several
11225 * Add support for DTLS.
11227 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11229 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11230 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11234 * Remove buggy and incomplete DH cert support from
11235 ssl/ssl_rsa.c and ssl/s3_both.c
11239 * Use SHA-1 instead of MD5 as the default digest algorithm for
11240 the `apps/openssl` commands.
11244 * Compile clean with "-Wall -Wmissing-prototypes
11245 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11246 DEBUG_SAFESTACK must also be set.
11250 * Change ./Configure so that certain algorithms can be disabled by default.
11251 The new counterpiece to "no-xxx" is "enable-xxx".
11253 The patented RC5 and MDC2 algorithms will now be disabled unless
11254 "enable-rc5" and "enable-mdc2", respectively, are specified.
11256 (IDEA remains enabled despite being patented. This is because IDEA
11257 is frequently required for interoperability, and there is no license
11258 fee for non-commercial use. As before, "no-idea" can be used to
11259 avoid this algorithm.)
11263 * Add processing of proxy certificates (see RFC 3820). This work was
11264 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11265 EGEE (Enabling Grids for E-science in Europe).
11269 * RC4 performance overhaul on modern architectures/implementations, such
11270 as Intel P4, IA-64 and AMD64.
11274 * New utility extract-section.pl. This can be used specify an alternative
11275 section number in a pod file instead of having to treat each file as
11276 a separate case in Makefile. This can be done by adding two lines to the
11279 =for comment openssl_section:XXX
11281 The blank line is mandatory.
11285 * New arguments -certform, -keyform and -pass for s_client and s_server
11286 to allow alternative format key and certificate files and passphrase
11291 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11292 update associated structures and add various utility functions.
11294 Add new policy related verify parameters, include policy checking in
11295 standard verify code. Enhance 'smime' application with extra parameters
11296 to support policy checking and print out.
11300 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11301 Nehemiah processors. These extensions support AES encryption in hardware
11302 as well as RNG (though RNG support is currently disabled).
11304 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11306 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11310 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11312 *Andy Polyakov and a number of other people*
11314 * Improved PowerPC platform support. Most notably BIGNUM assembler
11315 implementation contributed by IBM.
11317 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11319 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11320 exponent rather than 'unsigned long'. There is a corresponding change to
11321 the new 'rsa_keygen' element of the RSA_METHOD structure.
11323 *Jelte Jansen, Geoff Thorpe*
11325 * Functionality for creating the initial serial number file is now
11326 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11328 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11329 number file to 1, which is bound to cause problems. To avoid
11330 the problems while respecting compatibility between different 0.9.7
11331 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11332 CA.pl for serial number initialization. With the new release 0.9.8,
11333 we can fix the problem directly in the 'ca' utility.)
11337 * Reduced header interdependencies by declaring more opaque objects in
11338 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11339 give fewer recursive includes, which could break lazy source code - so
11340 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11341 developers should define this symbol when building and using openssl to
11342 ensure they track the recommended behaviour, interfaces, [etc], but
11343 backwards-compatible behaviour prevails when this isn't defined.
11347 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11351 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11352 This will generate a random key of the appropriate length based on the
11353 cipher context. The EVP_CIPHER can provide its own random key generation
11354 routine to support keys of a specific form. This is used in the des and
11355 3des routines to generate a key of the correct parity. Update S/MIME
11356 code to use new functions and hence generate correct parity DES keys.
11357 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11358 valid (weak or incorrect parity).
11362 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11363 as looking them up. This is useful when the verified structure may contain
11364 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11365 present unless the new PKCS7_NO_CRL flag is asserted.
11369 * Extend ASN1 oid configuration module. It now additionally accepts the
11372 shortName = some long name, 1.2.3.4
11376 * Reimplemented the BN_CTX implementation. There is now no more static
11377 limitation on the number of variables it can handle nor the depth of the
11378 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11379 information can now expand as required, and rather than having a single
11380 static array of bignums, BN_CTX now uses a linked-list of such arrays
11381 allowing it to expand on demand whilst maintaining the usefulness of
11382 BN_CTX's "bundling".
11386 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11387 to allow all RSA operations to function using a single BN_CTX.
11391 * Preliminary support for certificate policy evaluation and checking. This
11392 is initially intended to pass the tests outlined in "Conformance Testing
11393 of Relying Party Client Certificate Path Processing Logic" v1.07.
11397 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11398 remained unused and not that useful. A variety of other little bignum
11399 tweaks and fixes have also been made continuing on from the audit (see
11404 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11405 associated ASN1, EVP and SSL functions and old ASN1 macros.
11409 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11410 and this should never fail. So the return value from the use of
11411 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11412 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11416 * BN_CTX_get() should return zero-valued bignums, providing the same
11417 initialised value as BN_new().
11419 *Geoff Thorpe, suggested by Ulf Möller*
11421 * Support for inhibitAnyPolicy certificate extension.
11425 * An audit of the BIGNUM code is underway, for which debugging code is
11426 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11427 is considered valid when processing BIGNUMs, and causes execution to
11428 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11429 further steps are taken to deliberately pollute unused data in BIGNUM
11430 structures to try and expose faulty code further on. For now, openssl will
11431 (in its default mode of operation) continue to tolerate the inconsistent
11432 forms that it has tolerated in the past, but authors and packagers should
11433 consider trying openssl and their own applications when compiled with
11434 these debugging symbols defined. It will help highlight potential bugs in
11435 their own code, and will improve the test coverage for OpenSSL itself. At
11436 some point, these tighter rules will become openssl's default to improve
11437 maintainability, though the assert()s and other overheads will remain only
11438 in debugging configurations. See bn.h for more details.
11440 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11442 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11443 that can only be obtained through BN_CTX_new() (which implicitly
11444 initialises it). The presence of this function only made it possible
11445 to overwrite an existing structure (and cause memory leaks).
11449 * Because of the callback-based approach for implementing LHASH as a
11450 template type, lh_insert() adds opaque objects to hash-tables and
11451 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11452 to clean up those corresponding objects before destroying the hash table
11453 (and losing the object pointers). So some over-zealous constifications in
11454 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11455 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11456 prototyped to have "const" restrictions on the object pointers they are
11457 given (and so aren't required to cast them away any more).
11461 * The tmdiff.h API was so ugly and minimal that our own timing utility
11462 (speed) prefers to use its own implementation. The two implementations
11463 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11464 its object type properly exposed (MS_TM) instead of casting to/from
11465 `char *`. This may still change yet if someone realises MS_TM and
11467 aren't necessarily the greatest nomenclatures - but this is what was used
11468 internally to the implementation so I've used that for now.
11472 * Ensure that deprecated functions do not get compiled when
11473 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11474 the self-tests were still using deprecated key-generation functions so
11475 these have been updated also.
11479 * Reorganise PKCS#7 code to separate the digest location functionality
11480 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11481 New function PKCS7_set_digest() to set the digest type for PKCS#7
11482 digestedData type. Add additional code to correctly generate the
11483 digestedData type and add support for this type in PKCS7 initialization
11488 * New function PKCS7_set0_type_other() this initializes a PKCS7
11489 structure of type "other".
11493 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11494 sure the loop does correctly stop and breaking ("division by zero")
11495 modulus operations are not performed. The (pre-generated) prime
11496 table crypto/bn/bn_prime.h was already correct, but it could not be
11497 re-generated on some platforms because of the "division by zero"
11498 situation in the script.
11500 *Ralf S. Engelschall*
11502 * Update support for ECC-based TLS ciphersuites according to
11503 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11504 SHA-1 now is only used for "small" curves (where the
11505 representation of a field element takes up to 24 bytes); for
11506 larger curves, the field element resulting from ECDH is directly
11507 used as premaster secret.
11509 *Douglas Stebila (Sun Microsystems Laboratories)*
11511 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11512 curve secp160r1 to the tests.
11514 *Douglas Stebila (Sun Microsystems Laboratories)*
11516 * Add the possibility to load symbols globally with DSO.
11518 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11520 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11521 control of the error stack.
11525 * Add support for STORE in ENGINE.
11529 * Add the STORE type. The intention is to provide a common interface
11530 to certificate and key stores, be they simple file-based stores, or
11531 HSM-type store, or LDAP stores, or...
11532 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11536 * Add a generic structure called OPENSSL_ITEM. This can be used to
11537 pass a list of arguments to any function as well as provide a way
11538 for a function to pass data back to the caller.
11542 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11543 works like BUF_strdup() but can be used to duplicate a portion of
11544 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11549 * Add the function sk_find_ex() which works like sk_find(), but will
11550 return an index to an element even if an exact match couldn't be
11551 found. The index is guaranteed to point at the element where the
11552 searched-for key would be inserted to preserve sorting order.
11556 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11557 takes an extra flags argument for optional functionality. Currently,
11558 the following flags are defined:
11560 OBJ_BSEARCH_VALUE_ON_NOMATCH
11561 This one gets OBJ_bsearch_ex() to return a pointer to the first
11562 element where the comparing function returns a negative or zero
11565 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11566 This one gets OBJ_bsearch_ex() to return a pointer to the first
11567 element where the comparing function returns zero. This is useful
11568 if there are more than one element where the comparing function
11573 * Make it possible to create self-signed certificates with 'openssl ca'
11574 in such a way that the self-signed certificate becomes part of the
11575 CA database and uses the same mechanisms for serial number generation
11576 as all other certificate signing. The new flag '-selfsign' enables
11577 this functionality. Adapt CA.sh and CA.pl.in.
11581 * Add functionality to check the public key of a certificate request
11582 against a given private. This is useful to check that a certificate
11583 request can be signed by that key (self-signing).
11587 * Make it possible to have multiple active certificates with the same
11588 subject in the CA index file. This is done only if the keyword
11589 'unique_subject' is set to 'no' in the main CA section (default
11590 if 'CA_default') of the configuration file. The value is saved
11591 with the database itself in a separate index attribute file,
11592 named like the index file with '.attr' appended to the name.
11596 * Generate multi-valued AVAs using '+' notation in config files for
11601 * Support for nameConstraints certificate extension.
11605 * Support for policyConstraints certificate extension.
11609 * Support for policyMappings certificate extension.
11613 * Make sure the default DSA_METHOD implementation only uses its
11614 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11615 and change its own handlers to be NULL so as to remove unnecessary
11616 indirection. This lets alternative implementations fallback to the
11617 default implementation more easily.
11621 * Support for directoryName in GeneralName related extensions
11626 * Make it possible to link applications using Makefile.shared.
11627 Make that possible even when linking against static libraries!
11631 * Support for single pass processing for S/MIME signing. This now
11632 means that S/MIME signing can be done from a pipe, in addition
11633 cleartext signing (multipart/signed type) is effectively streaming
11634 and the signed data does not need to be all held in memory.
11636 This is done with a new flag PKCS7_STREAM. When this flag is set
11637 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11638 is done after the data is output (and digests calculated) in
11639 SMIME_write_PKCS7().
11643 * Add full support for -rpath/-R, both in shared libraries and
11644 applications, at least on the platforms where it's known how
11649 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11650 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11651 will now compute a table of multiples of the generator that
11652 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11653 faster (notably in the case of a single point multiplication,
11654 scalar * generator).
11656 *Nils Larsch, Bodo Moeller*
11658 * IPv6 support for certificate extensions. The various extensions
11659 which use the IP:a.b.c.d can now take IPv6 addresses using the
11660 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11665 * Added an ENGINE that implements RSA by performing private key
11666 exponentiations with the GMP library. The conversions to and from
11667 GMP's mpz_t format aren't optimised nor are any montgomery forms
11668 cached, and on x86 it appears OpenSSL's own performance has caught up.
11669 However there are likely to be other architectures where GMP could
11670 provide a boost. This ENGINE is not built in by default, but it can be
11671 specified at Configure time and should be accompanied by the necessary
11672 linker additions, eg;
11673 ./config -DOPENSSL_USE_GMP -lgmp
11677 * "openssl engine" will not display ENGINE/DSO load failure errors when
11678 testing availability of engines with "-t" - the old behaviour is
11679 produced by increasing the feature's verbosity with "-tt".
11683 * ECDSA routines: under certain error conditions uninitialized BN objects
11684 could be freed. Solution: make sure initialization is performed early
11685 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11690 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11691 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11692 software implementations. For DSA and DH, parameter generation can
11693 also be overridden by providing the appropriate method callbacks.
11697 * Change the "progress" mechanism used in key-generation and
11698 primality testing to functions that take a new BN_GENCB pointer in
11699 place of callback/argument pairs. The new API functions have `_ex`
11700 postfixes and the older functions are reimplemented as wrappers for
11701 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11702 declarations of the old functions to help (graceful) attempts to
11703 migrate to the new functions. Also, the new key-generation API
11704 functions operate on a caller-supplied key-structure and return
11705 success/failure rather than returning a key or NULL - this is to
11706 help make "keygen" another member function of RSA_METHOD etc.
11708 Example for using the new callback interface:
11710 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11711 void *my_arg = ...;
11714 BN_GENCB_set(&my_cb, my_callback, my_arg);
11716 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11717 /* For the meaning of a, b in calls to my_callback(), see the
11718 * documentation of the function that calls the callback.
11719 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11720 * my_callback should return 1 if it wants BN_is_prime_ex()
11721 * to continue, or 0 to stop.
11726 * Change the ZLIB compression method to be stateful, and make it
11727 available to TLS with the number defined in
11728 draft-ietf-tls-compression-04.txt.
11732 * Add the ASN.1 structures and functions for CertificatePair, which
11733 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11735 CertificatePair ::= SEQUENCE {
11736 forward [0] Certificate OPTIONAL,
11737 reverse [1] Certificate OPTIONAL,
11738 -- at least one of the pair shall be present -- }
11740 Also implement the PEM functions to read and write certificate
11741 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11743 This needed to be defined, mostly for the sake of the LDAP
11744 attribute crossCertificatePair, but may prove useful elsewhere as
11749 * Make it possible to inhibit symlinking of shared libraries in
11750 Makefile.shared, for Cygwin's sake.
11754 * Extend the BIGNUM API by creating a function
11755 void BN_set_negative(BIGNUM *a, int neg);
11756 and a macro that behave like
11757 int BN_is_negative(const BIGNUM *a);
11759 to avoid the need to access 'a->neg' directly in applications.
11763 * Implement fast modular reduction for pseudo-Mersenne primes
11764 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11765 EC_GROUP_new_curve_GFp() will now automatically use this
11768 *Nils Larsch <nla@trustcenter.de>*
11770 * Add new lock type (CRYPTO_LOCK_BN).
11774 * Change the ENGINE framework to automatically load engines
11775 dynamically from specific directories unless they could be
11776 found to already be built in or loaded. Move all the
11777 current engines except for the cryptodev one to a new
11778 directory engines/.
11779 The engines in engines/ are built as shared libraries if
11780 the "shared" options was given to ./Configure or ./config.
11781 Otherwise, they are inserted in libcrypto.a.
11782 /usr/local/ssl/engines is the default directory for dynamic
11783 engines, but that can be overridden at configure time through
11784 the usual use of --prefix and/or --openssldir, and at run
11785 time with the environment variable OPENSSL_ENGINES.
11787 *Geoff Thorpe and Richard Levitte*
11789 * Add Makefile.shared, a helper makefile to build shared
11790 libraries. Adapt Makefile.org.
11794 * Add version info to Win32 DLLs.
11796 *Peter 'Luna' Runestig" <peter@runestig.com>*
11798 * Add new 'medium level' PKCS#12 API. Certificates and keys
11799 can be added using this API to created arbitrary PKCS#12
11800 files while avoiding the low-level API.
11802 New options to PKCS12_create(), key or cert can be NULL and
11803 will then be omitted from the output file. The encryption
11804 algorithm NIDs can be set to -1 for no encryption, the mac
11805 iteration count can be set to 0 to omit the mac.
11807 Enhance pkcs12 utility by making the -nokeys and -nocerts
11808 options work when creating a PKCS#12 file. New option -nomac
11809 to omit the mac, NONE can be set for an encryption algorithm.
11810 New code is modified to use the enhanced PKCS12_create()
11811 instead of the low-level API.
11815 * Extend ASN1 encoder to support indefinite length constructed
11816 encoding. This can output sequences tags and octet strings in
11817 this form. Modify pk7_asn1.c to support indefinite length
11818 encoding. This is experimental and needs additional code to
11819 be useful, such as an ASN1 bio and some enhanced streaming
11822 Extend template encode functionality so that tagging is passed
11823 down to the template encoder.
11827 * Let 'openssl req' fail if an argument to '-newkey' is not
11828 recognized instead of using RSA as a default.
11832 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11833 As these are not official, they are not included in "ALL";
11834 the "ECCdraft" ciphersuite group alias can be used to select them.
11836 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11838 * Add ECDH engine support.
11840 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11842 * Add ECDH in new directory crypto/ecdh/.
11844 *Douglas Stebila (Sun Microsystems Laboratories)*
11846 * Let BN_rand_range() abort with an error after 100 iterations
11847 without success (which indicates a broken PRNG).
11851 * Change BN_mod_sqrt() so that it verifies that the input value
11852 is really the square of the return value. (Previously,
11853 BN_mod_sqrt would show GIGO behaviour.)
11857 * Add named elliptic curves over binary fields from X9.62, SECG,
11858 and WAP/WTLS; add OIDs that were still missing.
11860 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11862 * Extend the EC library for elliptic curves over binary fields
11863 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11866 EC_GF2m_simple_method
11870 EC_GROUP_new_curve_GF2m
11871 EC_GROUP_set_curve_GF2m
11872 EC_GROUP_get_curve_GF2m
11873 EC_POINT_set_affine_coordinates_GF2m
11874 EC_POINT_get_affine_coordinates_GF2m
11875 EC_POINT_set_compressed_coordinates_GF2m
11877 Point compression for binary fields is disabled by default for
11878 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11881 As binary polynomials are represented as BIGNUMs, various members
11882 of the EC_GROUP and EC_POINT data structures can be shared
11883 between the implementations for prime fields and binary fields;
11884 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11885 are essentially identical to their `..._GFp` counterparts.
11886 (For simplicity, the `..._GFp` prefix has been dropped from
11887 various internal method names.)
11889 An internal 'field_div' method (similar to 'field_mul' and
11890 'field_sqr') has been added; this is used only for binary fields.
11892 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11894 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11895 through methods ('mul', 'precompute_mult').
11897 The generic implementations (now internally called 'ec_wNAF_mul'
11898 and 'ec_wNAF_precomputed_mult') remain the default if these
11899 methods are undefined.
11901 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11903 * New function EC_GROUP_get_degree, which is defined through
11904 EC_METHOD. For curves over prime fields, this returns the bit
11905 length of the modulus.
11907 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11909 * New functions EC_GROUP_dup, EC_POINT_dup.
11910 (These simply call ..._new and ..._copy).
11912 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11914 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11915 Polynomials are represented as BIGNUMs (where the sign bit is not
11916 used) in the following functions [macros]:
11919 BN_GF2m_sub [= BN_GF2m_add]
11920 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11921 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11922 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11924 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11925 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11926 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11927 BN_GF2m_cmp [= BN_ucmp]
11929 (Note that only the 'mod' functions are actually for fields GF(2^m).
11930 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11932 For some functions, an the irreducible polynomial defining a
11933 field can be given as an 'unsigned int[]' with strictly
11934 decreasing elements giving the indices of those bits that are set;
11935 i.e., p[] represents the polynomial
11936 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11938 p[0] > p[1] > ... > p[k] = 0.
11939 This applies to the following functions:
11942 BN_GF2m_mod_mul_arr
11943 BN_GF2m_mod_sqr_arr
11944 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11945 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11946 BN_GF2m_mod_exp_arr
11947 BN_GF2m_mod_sqrt_arr
11948 BN_GF2m_mod_solve_quad_arr
11952 Conversion can be performed by the following functions:
11957 bntest.c has additional tests for binary polynomial arithmetic.
11959 Two implementations for BN_GF2m_mod_div() are available.
11960 The default algorithm simply uses BN_GF2m_mod_inv() and
11961 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11962 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11963 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11965 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11967 * Add new error code 'ERR_R_DISABLED' that can be used when some
11968 functionality is disabled at compile-time.
11970 *Douglas Stebila <douglas.stebila@sun.com>*
11972 * Change default behaviour of 'openssl asn1parse' so that more
11973 information is visible when viewing, e.g., a certificate:
11975 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11976 mode the content of non-printable OCTET STRINGs is output in a
11977 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11978 avoid the appearance of a printable string.
11980 *Nils Larsch <nla@trustcenter.de>*
11982 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11984 EC_GROUP_set_asn1_flag()
11985 EC_GROUP_get_asn1_flag()
11986 EC_GROUP_set_point_conversion_form()
11987 EC_GROUP_get_point_conversion_form()
11988 These control ASN1 encoding details:
11989 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11990 has been set to OPENSSL_EC_NAMED_CURVE.
11991 - Points are encoded in uncompressed form by default; options for
11992 asn1_for are as for point2oct, namely
11993 POINT_CONVERSION_COMPRESSED
11994 POINT_CONVERSION_UNCOMPRESSED
11995 POINT_CONVERSION_HYBRID
11997 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11999 EC_GROUP_set_seed()
12000 EC_GROUP_get0_seed()
12001 EC_GROUP_get_seed_len()
12002 This is used only for ASN1 purposes (so far).
12004 *Nils Larsch <nla@trustcenter.de>*
12006 * Add 'field_type' member to EC_METHOD, which holds the NID
12007 of the appropriate field type OID. The new function
12008 EC_METHOD_get_field_type() returns this value.
12010 *Nils Larsch <nla@trustcenter.de>*
12013 EC_POINT_point2bn()
12014 EC_POINT_bn2point()
12015 EC_POINT_point2hex()
12016 EC_POINT_hex2point()
12017 providing useful interfaces to EC_POINT_point2oct() and
12018 EC_POINT_oct2point().
12020 *Nils Larsch <nla@trustcenter.de>*
12022 * Change internals of the EC library so that the functions
12023 EC_GROUP_set_generator()
12024 EC_GROUP_get_generator()
12025 EC_GROUP_get_order()
12026 EC_GROUP_get_cofactor()
12027 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12028 to methods, which would lead to unnecessary code duplication when
12029 adding different types of curves.
12031 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12033 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12034 arithmetic, and such that modified wNAFs are generated
12035 (which avoid length expansion in many cases).
12039 * Add a function EC_GROUP_check_discriminant() (defined via
12040 EC_METHOD) that verifies that the curve discriminant is non-zero.
12042 Add a function EC_GROUP_check() that makes some sanity tests
12043 on a EC_GROUP, its generator and order. This includes
12044 EC_GROUP_check_discriminant().
12046 *Nils Larsch <nla@trustcenter.de>*
12048 * Add ECDSA in new directory crypto/ecdsa/.
12050 Add applications 'openssl ecparam' and 'openssl ecdsa'
12051 (these are based on 'openssl dsaparam' and 'openssl dsa').
12053 ECDSA support is also included in various other files across the
12054 library. Most notably,
12055 - 'openssl req' now has a '-newkey ecdsa:file' option;
12056 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12057 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12058 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12059 them suitable for ECDSA where domain parameters must be
12060 extracted before the specific public key;
12061 - ECDSA engine support has been added.
12063 *Nils Larsch <nla@trustcenter.de>*
12065 * Include some named elliptic curves, and add OIDs from X9.62,
12066 SECG, and WAP/WTLS. Each curve can be obtained from the new
12068 EC_GROUP_new_by_curve_name(),
12069 and the list of available named curves can be obtained with
12070 EC_get_builtin_curves().
12071 Also add a 'curve_name' member to EC_GROUP objects, which can be
12073 EC_GROUP_set_curve_name()
12074 EC_GROUP_get_curve_name()
12076 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12078 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12079 was actually never needed) and in BN_mul(). The removal in BN_mul()
12080 required a small change in bn_mul_part_recursive() and the addition
12081 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12082 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12083 bn_sub_words() and bn_add_words() except they take arrays with
12088 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12090 * Cleanse PEM buffers before freeing them since they may contain
12093 *Benjamin Bennett <ben@psc.edu>*
12095 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12096 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12097 authentication-only ciphersuites.
12101 * Since AES128 and AES256 share a single mask bit in the logic of
12102 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12103 kludge to work properly if AES128 is available and AES256 isn't.
12107 * Expand security boundary to match 1.1.1 module.
12111 * Remove redundant features: hash file source, editing of test vectors
12112 modify fipsld to use external fips_premain.c signature.
12116 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12117 run algorithm test programs.
12121 * Make algorithm test programs more tolerant of whitespace.
12125 * Have SSL/TLS server implementation tolerate "mismatched" record
12126 protocol version while receiving ClientHello even if the
12127 ClientHello is fragmented. (The server can't insist on the
12128 particular protocol version it has chosen before the ServerHello
12129 message has informed the client about his choice.)
12133 * Load error codes if they are not already present instead of using a
12134 static variable. This allows them to be cleanly unloaded and reloaded.
12138 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12140 * Introduce limits to prevent malicious keys being able to
12141 cause a denial of service. ([CVE-2006-2940])
12143 *Steve Henson, Bodo Moeller*
12145 * Fix ASN.1 parsing of certain invalid structures that can result
12146 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12148 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12149 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12151 * Fix SSL client code which could crash if connecting to a
12152 malicious SSLv2 server. ([CVE-2006-4343])
12154 *Tavis Ormandy and Will Drewry, Google Security Team*
12156 * Change ciphersuite string processing so that an explicit
12157 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12158 will no longer include "AES128-SHA"), and any other similar
12159 ciphersuite (same bitmap) from *other* protocol versions (so that
12160 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12161 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12162 changes from 0.9.8b and 0.9.8d.
12166 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12168 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12169 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12171 * Change the Unix randomness entropy gathering to use poll() when
12172 possible instead of select(), since the latter has some
12173 undesirable limitations.
12175 *Darryl Miles via Richard Levitte and Bodo Moeller*
12177 * Disable rogue ciphersuites:
12179 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12180 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12181 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12183 The latter two were purportedly from
12184 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12187 Also deactivate the remaining ciphersuites from
12188 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12189 unofficial, and the ID has long expired.
12193 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12194 dual-core machines) and other potential thread-safety issues.
12198 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12200 * Adapt fipsld and the build system to link against the validated FIPS
12201 module in FIPS mode.
12205 * Fixes for VC++ 2005 build under Windows.
12209 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12210 from a Windows bash shell such as MSYS. It is autodetected from the
12211 "config" script when run from a VC++ environment. Modify standard VC++
12212 build to use fipscanister.o from the GNU make build.
12216 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12218 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12219 The value now differs depending on if you build for FIPS or not.
12220 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12221 safely run with a non-FIPSed libcrypto, as it may crash because of
12222 the difference induced by this change.
12226 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12228 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12229 (part of SSL_OP_ALL). This option used to disable the
12230 countermeasure against man-in-the-middle protocol-version
12231 rollback in the SSL 2.0 server implementation, which is a bad
12232 idea. ([CVE-2005-2969])
12234 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12235 for Information Security, National Institute of Advanced Industrial
12236 Science and Technology [AIST, Japan)]*
12238 * Minimal support for X9.31 signatures and PSS padding modes. This is
12239 mainly for FIPS compliance and not fully integrated at this stage.
12243 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12244 the exponentiation using a fixed-length exponent. (Otherwise,
12245 the information leaked through timing could expose the secret key
12246 after many signatures; cf. Bleichenbacher's attack on DSA with
12251 * Make a new fixed-window mod_exp implementation the default for
12252 RSA, DSA, and DH private-key operations so that the sequence of
12253 squares and multiplies and the memory access pattern are
12254 independent of the particular secret key. This will mitigate
12255 cache-timing and potential related attacks.
12257 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12258 and this is automatically used by BN_mod_exp_mont() if the new flag
12259 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12260 will use this BN flag for private exponents unless the flag
12261 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12262 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12264 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12266 * Change the client implementation for SSLv23_method() and
12267 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12268 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12269 (Previously, the SSL 2.0 backwards compatible Client Hello
12270 message format would be used even with SSL_OP_NO_SSLv2.)
12274 * Add support for smime-type MIME parameter in S/MIME messages which some
12279 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12280 a threadsafe manner. Modify rsa code to use new function and add calls
12281 to dsa and dh code (which had race conditions before).
12285 * Include the fixed error library code in the C error file definitions
12286 instead of fixing them up at runtime. This keeps the error code
12287 structures constant.
12291 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12293 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12296 * Fixes for newer kerberos headers. NB: the casts are needed because
12297 the 'length' field is signed on one version and unsigned on another
12298 with no (?) obvious way to tell the difference, without these VC++
12299 complains. Also the "definition" of FAR (blank) is no longer included
12300 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12301 some needed definitions.
12305 * Undo Cygwin change.
12309 * Added support for proxy certificates according to RFC 3820.
12310 Because they may be a security thread to unaware applications,
12311 they must be explicitly allowed in run-time. See
12312 docs/HOWTO/proxy_certificates.txt for further information.
12316 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12318 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12319 server and client random values. Previously
12320 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12321 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12323 This change has negligible security impact because:
12325 1. Server and client random values still have 24 bytes of pseudo random
12328 2. Server and client random values are sent in the clear in the initial
12331 3. The master secret is derived using the premaster secret (48 bytes in
12332 size for static RSA ciphersuites) as well as client server and random
12335 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12338 *Stephen Henson, reported by UK NISCC*
12340 * Use Windows randomness collection on Cygwin.
12344 * Fix hang in EGD/PRNGD query when communication socket is closed
12345 prematurely by EGD/PRNGD.
12347 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12349 * Prompt for pass phrases when appropriate for PKCS12 input format.
12353 * Back-port of selected performance improvements from development
12354 branch, as well as improved support for PowerPC platforms.
12358 * Add lots of checks for memory allocation failure, error codes to indicate
12359 failure and freeing up memory if a failure occurs.
12361 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12363 * Add new -passin argument to dgst.
12367 * Perform some character comparisons of different types in X509_NAME_cmp:
12368 this is needed for some certificates that re-encode DNs into UTF8Strings
12369 (in violation of RFC3280) and can't or won't issue name rollover
12374 * Make an explicit check during certificate validation to see that
12375 the CA setting in each certificate on the chain is correct. As a
12376 side effect always do the following basic checks on extensions,
12377 not just when there's an associated purpose to the check:
12379 - if there is an unhandled critical extension (unless the user
12380 has chosen to ignore this fault)
12381 - if the path length has been exceeded (if one is set at all)
12382 - that certain extensions fit the associated purpose (if one has
12387 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12389 * Avoid a race condition when CRLs are checked in a multi threaded
12390 environment. This would happen due to the reordering of the revoked
12391 entries during signature checking and serial number lookup. Now the
12392 encoding is cached and the serial number sort performed under a lock.
12393 Add new STACK function sk_is_sorted().
12397 * Add Delta CRL to the extension code.
12401 * Various fixes to s3_pkt.c so alerts are sent properly.
12403 *David Holmes <d.holmes@f5.com>*
12405 * Reduce the chances of duplicate issuer name and serial numbers (in
12406 violation of RFC3280) using the OpenSSL certificate creation utilities.
12407 This is done by creating a random 64 bit value for the initial serial
12408 number when a serial number file is created or when a self signed
12409 certificate is created using 'openssl req -x509'. The initial serial
12410 number file is created using 'openssl x509 -next_serial' in CA.pl
12411 rather than being initialized to 1.
12415 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12417 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12418 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12420 *Joe Orton, Steve Henson*
12422 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12425 *Joe Orton, Steve Henson*
12427 * Make it possible to have multiple active certificates with the same
12428 subject in the CA index file. This is done only if the keyword
12429 'unique_subject' is set to 'no' in the main CA section (default
12430 if 'CA_default') of the configuration file. The value is saved
12431 with the database itself in a separate index attribute file,
12432 named like the index file with '.attr' appended to the name.
12436 * X509 verify fixes. Disable broken certificate workarounds when
12437 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12438 keyUsage extension present. Don't accept CRLs with unhandled critical
12439 extensions: since verify currently doesn't process CRL extensions this
12440 rejects a CRL with *any* critical extensions. Add new verify error codes
12445 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12446 A clarification of RFC2560 will require the use of OCTET STRINGs and
12447 some implementations cannot handle the current raw format. Since OpenSSL
12448 copies and compares OCSP nonces as opaque blobs without any attempt at
12449 parsing them this should not create any compatibility issues.
12453 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12454 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12455 this HMAC (and other) operations are several times slower than OpenSSL
12460 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12462 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12464 * Use the correct content when signing type "other".
12468 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12470 * Fix various bugs revealed by running the NISCC test suite:
12472 Stop out of bounds reads in the ASN1 code when presented with
12473 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12475 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12477 If verify callback ignores invalid public key errors don't try to check
12478 certificate signature with the NULL public key.
12482 * New -ignore_err option in ocsp application to stop the server
12483 exiting on the first error in a request.
12487 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12488 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12493 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12494 extra data after the compression methods not only for TLS 1.0
12495 but also for SSL 3.0 (as required by the specification).
12497 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12499 * Change X509_certificate_type() to mark the key as exported/exportable
12500 when it's 512 *bits* long, not 512 bytes.
12504 * Change AES_cbc_encrypt() so it outputs exact multiple of
12505 blocks during encryption.
12509 * Various fixes to base64 BIO and non blocking I/O. On write
12510 flushes were not handled properly if the BIO retried. On read
12511 data was not being buffered properly and had various logic bugs.
12512 This also affects blocking I/O when the data being decoded is a
12517 * Various S/MIME bugfixes and compatibility changes:
12518 output correct application/pkcs7 MIME type if
12519 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12520 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12521 of files as .eml work). Correctly handle very long lines in MIME
12526 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12528 * Countermeasure against the Klima-Pokorny-Rosa extension of
12529 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12530 a protocol version number mismatch like a decryption error
12531 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12535 * Turn on RSA blinding by default in the default implementation
12536 to avoid a timing attack. Applications that don't want it can call
12537 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12538 They would be ill-advised to do so in most cases.
12540 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12542 * Change RSA blinding code so that it works when the PRNG is not
12543 seeded (in this case, the secret RSA exponent is abused as
12544 an unpredictable seed -- if it is not unpredictable, there
12545 is no point in blinding anyway). Make RSA blinding thread-safe
12546 by remembering the creator's thread ID in rsa->blinding and
12547 having all other threads use local one-time blinding factors
12548 (this requires more computation than sharing rsa->blinding, but
12549 avoids excessive locking; and if an RSA object is not shared
12550 between threads, blinding will still be very fast).
12554 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12555 ENGINE as defaults for all supported algorithms irrespective of
12556 the 'flags' parameter. 'flags' is now honoured, so applications
12557 should make sure they are passing it correctly.
12561 * Target "mingw" now allows native Windows code to be generated in
12562 the Cygwin environment as well as with the MinGW compiler.
12566 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12568 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12569 via timing by performing a MAC computation even if incorrect
12570 block cipher padding has been found. This is a countermeasure
12571 against active attacks where the attacker has to distinguish
12572 between bad padding and a MAC verification error. ([CVE-2003-0078])
12574 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12575 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12576 Martin Vuagnoux (EPFL, Ilion)*
12578 * Make the no-err option work as intended. The intention with no-err
12579 is not to have the whole error stack handling routines removed from
12580 libcrypto, it's only intended to remove all the function name and
12581 reason texts, thereby removing some of the footprint that may not
12582 be interesting if those errors aren't displayed anyway.
12584 NOTE: it's still possible for any application or module to have its
12585 own set of error texts inserted. The routines are there, just not
12586 used by default when no-err is given.
12590 * Add support for FreeBSD on IA64.
12592 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12594 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12595 Kerberos function mit_des_cbc_cksum(). Before this change,
12596 the value returned by DES_cbc_cksum() was like the one from
12597 mit_des_cbc_cksum(), except the bytes were swapped.
12599 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12601 * Allow an application to disable the automatic SSL chain building.
12602 Before this a rather primitive chain build was always performed in
12603 ssl3_output_cert_chain(): an application had no way to send the
12604 correct chain if the automatic operation produced an incorrect result.
12606 Now the chain builder is disabled if either:
12608 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12610 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12612 The reasoning behind this is that an application would not want the
12613 auto chain building to take place if extra chain certificates are
12614 present and it might also want a means of sending no additional
12615 certificates (for example the chain has two certificates and the
12620 * Add the possibility to build without the ENGINE framework.
12622 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12624 * Under Win32 gmtime() can return NULL: check return value in
12625 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12629 * DSA routines: under certain error conditions uninitialized BN objects
12630 could be freed. Solution: make sure initialization is performed early
12631 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12632 Nils Larsch <nla@trustcenter.de> via PR#459)
12636 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12637 checked on reconnect on the client side, therefore session resumption
12638 could still fail with a "ssl session id is different" error. This
12639 behaviour is masked when SSL_OP_ALL is used due to
12640 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12641 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12642 followup to PR #377.
12646 * IA-32 assembler support enhancements: unified ELF targets, support
12647 for SCO/Caldera platforms, fix for Cygwin shared build.
12651 * Add support for FreeBSD on sparc64. As a consequence, support for
12652 FreeBSD on non-x86 processors is separate from x86 processors on
12653 the config script, much like the NetBSD support.
12655 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12657 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12659 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12662 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12663 code (06) was taken as the first octet of the session ID and the last
12664 octet was ignored consequently. As a result SSLv2 client side session
12665 caching could not have worked due to the session ID mismatch between
12667 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12672 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12673 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12678 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12679 seems that in spite of existing for more than a year, many application
12680 author have done nothing to provide the necessary callbacks, which
12681 means that this particular engine will not work properly anywhere.
12682 This is a very unfortunate situation which forces us, in the name
12683 of usability, to give the hw_ncipher.c a static lock, which is part
12685 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12686 appear in 0.9.8 or later. We EXPECT application authors to have
12687 dealt properly with this when 0.9.8 is released (unless we actually
12688 make such changes in the libcrypto locking code that changes will
12689 have to be made anyway).
12693 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12694 octets have been read, EOF or an error occurs. Without this change
12695 some truncated ASN1 structures will not produce an error.
12699 * Disable Heimdal support, since it hasn't been fully implemented.
12700 Still give the possibility to force the use of Heimdal, but with
12701 warnings and a request that patches get sent to openssl-dev.
12705 * Add the VC-CE target, introduce the WINCE sysname, and add
12706 INSTALL.WCE and appropriate conditionals to make it build.
12708 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12710 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12711 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12712 edit numbers of the version.
12714 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12716 * Introduce safe string copy and catenation functions
12717 (BUF_strlcpy() and BUF_strlcat()).
12719 *Ben Laurie (CHATS) and Richard Levitte*
12721 * Avoid using fixed-size buffers for one-line DNs.
12723 *Ben Laurie (CHATS)*
12725 * Add BUF_MEM_grow_clean() to avoid information leakage when
12726 resizing buffers containing secrets, and use where appropriate.
12728 *Ben Laurie (CHATS)*
12730 * Avoid using fixed size buffers for configuration file location.
12732 *Ben Laurie (CHATS)*
12734 * Avoid filename truncation for various CA files.
12736 *Ben Laurie (CHATS)*
12738 * Use sizeof in preference to magic numbers.
12740 *Ben Laurie (CHATS)*
12742 * Avoid filename truncation in cert requests.
12744 *Ben Laurie (CHATS)*
12746 * Add assertions to check for (supposedly impossible) buffer
12749 *Ben Laurie (CHATS)*
12751 * Don't cache truncated DNS entries in the local cache (this could
12752 potentially lead to a spoofing attack).
12754 *Ben Laurie (CHATS)*
12756 * Fix various buffers to be large enough for hex/decimal
12757 representations in a platform independent manner.
12759 *Ben Laurie (CHATS)*
12761 * Add CRYPTO_realloc_clean() to avoid information leakage when
12762 resizing buffers containing secrets, and use where appropriate.
12764 *Ben Laurie (CHATS)*
12766 * Add BIO_indent() to avoid much slightly worrying code to do
12769 *Ben Laurie (CHATS)*
12771 * Convert sprintf()/BIO_puts() to BIO_printf().
12773 *Ben Laurie (CHATS)*
12775 * buffer_gets() could terminate with the buffer only half
12778 *Ben Laurie (CHATS)*
12780 * Add assertions to prevent user-supplied crypto functions from
12781 overflowing internal buffers by having large block sizes, etc.
12783 *Ben Laurie (CHATS)*
12785 * New OPENSSL_assert() macro (similar to assert(), but enabled
12788 *Ben Laurie (CHATS)*
12790 * Eliminate unused copy of key in RC4.
12792 *Ben Laurie (CHATS)*
12794 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12796 *Ben Laurie (CHATS)*
12798 * Fix off-by-one error in EGD path.
12800 *Ben Laurie (CHATS)*
12802 * If RANDFILE path is too long, ignore instead of truncating.
12804 *Ben Laurie (CHATS)*
12806 * Eliminate unused and incorrectly sized X.509 structure
12809 *Ben Laurie (CHATS)*
12811 * Eliminate unused and dangerous function knumber().
12813 *Ben Laurie (CHATS)*
12815 * Eliminate unused and dangerous structure, KSSL_ERR.
12817 *Ben Laurie (CHATS)*
12819 * Protect against overlong session ID context length in an encoded
12820 session object. Since these are local, this does not appear to be
12823 *Ben Laurie (CHATS)*
12825 * Change from security patch (see 0.9.6e below) that did not affect
12826 the 0.9.6 release series:
12828 Remote buffer overflow in SSL3 protocol - an attacker could
12829 supply an oversized master key in Kerberos-enabled versions.
12832 *Ben Laurie (CHATS)*
12834 * Change the SSL kerb5 codes to match RFC 2712.
12838 * Make -nameopt work fully for req and add -reqopt switch.
12840 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12842 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12844 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12846 * Make sure tests can be performed even if the corresponding algorithms
12847 have been removed entirely. This was also the last step to make
12848 OpenSSL compilable with DJGPP under all reasonable conditions.
12850 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12852 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12853 to allow version independent disabling of normally unselected ciphers,
12854 which may be activated as a side-effect of selecting a single cipher.
12856 (E.g., cipher list string "RSA" enables ciphersuites that are left
12857 out of "ALL" because they do not provide symmetric encryption.
12858 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12860 *Lutz Jaenicke, Bodo Moeller*
12862 * Add appropriate support for separate platform-dependent build
12863 directories. The recommended way to make a platform-dependent
12864 build directory is the following (tested on Linux), maybe with
12867 # Place yourself outside of the OpenSSL source tree. In
12868 # this example, the environment variable OPENSSL_SOURCE
12869 # is assumed to contain the absolute OpenSSL source directory.
12870 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12871 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12872 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12873 mkdir -p `dirname $F`
12874 ln -s $OPENSSL_SOURCE/$F $F
12877 To be absolutely sure not to disturb the source tree, a "make clean"
12878 is a good thing. If it isn't successful, don't worry about it,
12879 it probably means the source directory is very clean.
12883 * Make sure any ENGINE control commands make local copies of string
12884 pointers passed to them whenever necessary. Otherwise it is possible
12885 the caller may have overwritten (or deallocated) the original string
12886 data when a later ENGINE operation tries to use the stored values.
12888 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12890 * Improve diagnostics in file reading and command-line digests.
12892 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12894 * Add AES modes CFB and OFB to the object database. Correct an
12895 error in AES-CFB decryption.
12899 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12900 allows existing EVP_CIPHER_CTX structures to be reused after
12901 calling `EVP_*Final()`. This behaviour is used by encryption
12902 BIOs and some applications. This has the side effect that
12903 applications must explicitly clean up cipher contexts with
12904 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12908 * Check the values of dna and dnb in bn_mul_recursive before calling
12909 bn_mul_comba (a non zero value means the a or b arrays do not contain
12910 n2 elements) and fallback to bn_mul_normal if either is not zero.
12914 * Fix escaping of non-ASCII characters when using the -subj option
12915 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12919 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12920 form for "surname", serialNumber has no short form.
12921 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12922 therefore remove "mail" short name for "internet 7".
12923 The OID for unique identifiers in X509 certificates is
12924 x500UniqueIdentifier, not uniqueIdentifier.
12925 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12929 * Add an "init" command to the ENGINE config module and auto initialize
12930 ENGINEs. Without any "init" command the ENGINE will be initialized
12931 after all ctrl commands have been executed on it. If init=1 the
12932 ENGINE is initialized at that point (ctrls before that point are run
12933 on the uninitialized ENGINE and after on the initialized one). If
12934 init=0 then the ENGINE will not be initialized at all.
12938 * Fix the 'app_verify_callback' interface so that the user-defined
12939 argument is actually passed to the callback: In the
12940 SSL_CTX_set_cert_verify_callback() prototype, the callback
12941 declaration has been changed from
12944 int (*cb)(X509_STORE_CTX *,void *);
12945 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12946 i=s->ctx->app_verify_callback(&ctx)
12947 has been changed into
12948 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12950 To update applications using SSL_CTX_set_cert_verify_callback(),
12951 a dummy argument can be added to their callback functions.
12953 *D. K. Smetters <smetters@parc.xerox.com>*
12955 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12957 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12959 * Add and OPENSSL_LOAD_CONF define which will cause
12960 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12961 This allows older applications to transparently support certain
12962 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12963 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12964 load the config file and OPENSSL_add_all_algorithms_conf() which will
12965 always load it have also been added.
12969 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12970 Adjust NIDs and EVP layer.
12972 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12974 * Config modules support in openssl utility.
12976 Most commands now load modules from the config file,
12977 though in a few (such as version) this isn't done
12978 because it couldn't be used for anything.
12980 In the case of ca and req the config file used is
12981 the same as the utility itself: that is the -config
12982 command line option can be used to specify an
12987 * Move default behaviour from OPENSSL_config(). If appname is NULL
12988 use "openssl_conf" if filename is NULL use default openssl config file.
12992 * Add an argument to OPENSSL_config() to allow the use of an alternative
12993 config section name. Add a new flag to tolerate a missing config file
12994 and move code to CONF_modules_load_file().
12998 * Support for crypto accelerator cards from Accelerated Encryption
12999 Processing, www.aep.ie. (Use engine 'aep')
13000 The support was copied from 0.9.6c [engine] and adapted/corrected
13001 to work with the new engine framework.
13003 *AEP Inc. and Richard Levitte*
13005 * Support for SureWare crypto accelerator cards from Baltimore
13006 Technologies. (Use engine 'sureware')
13007 The support was copied from 0.9.6c [engine] and adapted
13008 to work with the new engine framework.
13012 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13013 make the newer ENGINE framework commands for the CHIL engine work.
13015 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13017 * Make it possible to produce shared libraries on ReliantUNIX.
13019 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13021 * Add the configuration target debug-linux-ppro.
13022 Make 'openssl rsa' use the general key loading routines
13023 implemented in `apps.c`, and make those routines able to
13024 handle the key format FORMAT_NETSCAPE and the variant
13027 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13029 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13031 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13033 * Add -keyform to rsautl, and document -engine.
13035 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13037 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13038 BIO_R_NO_SUCH_FILE error code rather than the generic
13039 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13043 * Add new functions
13044 ERR_peek_last_error
13045 ERR_peek_last_error_line
13046 ERR_peek_last_error_line_data.
13047 These are similar to
13049 ERR_peek_error_line
13050 ERR_peek_error_line_data,
13051 but report on the latest error recorded rather than the first one
13052 still in the error queue.
13054 *Ben Laurie, Bodo Moeller*
13056 * default_algorithms option in ENGINE config module. This allows things
13058 default_algorithms = ALL
13059 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13063 * Preliminary ENGINE config module.
13067 * New experimental application configuration code.
13071 * Change the AES code to follow the same name structure as all other
13072 symmetric ciphers, and behave the same way. Move everything to
13073 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13075 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13077 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13079 *Ben Laurie and Theo de Raadt*
13081 * Add option to output public keys in req command.
13083 *Massimiliano Pala madwolf@openca.org*
13085 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13086 (up to about 10% better than before for P-192 and P-224).
13090 * New functions/macros
13092 SSL_CTX_set_msg_callback(ctx, cb)
13093 SSL_CTX_set_msg_callback_arg(ctx, arg)
13094 SSL_set_msg_callback(ssl, cb)
13095 SSL_set_msg_callback_arg(ssl, arg)
13097 to request calling a callback function
13099 void cb(int write_p, int version, int content_type,
13100 const void *buf, size_t len, SSL *ssl, void *arg)
13102 whenever a protocol message has been completely received
13103 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13104 protocol version according to which the SSL library interprets
13105 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13106 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13107 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13108 specification (change_cipher_spec(20), alert(21), handshake(22)).
13109 'buf' and 'len' point to the actual message, 'ssl' to the
13110 SSL object, and 'arg' is the application-defined value set by
13111 SSL[_CTX]_set_msg_callback_arg().
13113 'openssl s_client' and 'openssl s_server' have new '-msg' options
13114 to enable a callback that displays all protocol messages.
13118 * Change the shared library support so shared libraries are built as
13119 soon as the corresponding static library is finished, and thereby get
13120 openssl and the test programs linked against the shared library.
13121 This still only happens when the keyword "shard" has been given to
13122 the configuration scripts.
13124 NOTE: shared library support is still an experimental thing, and
13125 backward binary compatibility is still not guaranteed.
13127 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13129 * Add support for Subject Information Access extension.
13131 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13133 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13134 additional bytes when new memory had to be allocated, not just
13135 when reusing an existing buffer.
13139 * New command line and configuration option 'utf8' for the req command.
13140 This allows field values to be specified as UTF8 strings.
13144 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13145 runs for the former and machine-readable output for the latter.
13149 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13150 of the e-mail address in the DN (i.e., it will go into a certificate
13151 extension only). The new configuration file option 'email_in_dn = no'
13152 has the same effect.
13154 *Massimiliano Pala madwolf@openca.org*
13156 * Change all functions with names starting with `des_` to be starting
13157 with `DES_` instead. Add wrappers that are compatible with libdes,
13158 but are named `_ossl_old_des_*`. Finally, add macros that map the
13159 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13160 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13161 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13164 Since we provide two compatibility mappings, the user needs to
13165 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13166 compatibility is desired. The default (i.e., when that macro
13167 isn't defined) is OpenSSL 0.9.6c compatibility.
13169 There are also macros that enable and disable the support of old
13170 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13171 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13172 are defined, the default will apply: to support the old des routines.
13174 In either case, one must include openssl/des.h to get the correct
13175 definitions. Do not try to just include openssl/des_old.h, that
13178 NOTE: This is a major break of an old API into a new one. Software
13179 authors are encouraged to switch to the `DES_` style functions. Some
13180 time in the future, des_old.h and the libdes compatibility functions
13181 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13182 default), and then completely removed.
13186 * Test for certificates which contain unsupported critical extensions.
13187 If such a certificate is found during a verify operation it is
13188 rejected by default: this behaviour can be overridden by either
13189 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13190 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13191 X509_supported_extension() has also been added which returns 1 if a
13192 particular extension is supported.
13196 * Modify the behaviour of EVP cipher functions in similar way to digests
13197 to retain compatibility with existing code.
13201 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13202 compatibility with existing code. In particular the 'ctx' parameter does
13203 not have to be to be initialized before the call to EVP_DigestInit() and
13204 it is tidied up after a call to EVP_DigestFinal(). New function
13205 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13206 EVP_MD_CTX_copy() changed to not require the destination to be
13207 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13208 requires the destination to be valid.
13210 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13211 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13215 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13216 so that complete 'Handshake' protocol structures are kept in memory
13217 instead of overwriting 'msg_type' and 'length' with 'body' data.
13221 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13223 *Massimo Santin via Richard Levitte*
13225 * Major restructuring to the underlying ENGINE code. This includes
13226 reduction of linker bloat, separation of pure "ENGINE" manipulation
13227 (initialisation, etc) from functionality dealing with implementations
13228 of specific crypto interfaces. This change also introduces integrated
13229 support for symmetric ciphers and digest implementations - so ENGINEs
13230 can now accelerate these by providing EVP_CIPHER and EVP_MD
13231 implementations of their own. This is detailed in
13232 [crypto/engine/README.md](crypto/engine/README.md)
13233 as it couldn't be adequately described here. However, there are a few
13234 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13235 were changed in the original introduction of ENGINE code have now
13236 reverted back - the hooking from this code to ENGINE is now a good
13237 deal more passive and at run-time, operations deal directly with
13238 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13239 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13240 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13241 they were not being used by the framework as there is no concept of a
13242 BIGNUM_METHOD and they could not be generalised to the new
13243 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13244 ENGINE_cpy() has been removed as it cannot be consistently defined in
13249 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13253 * Change mkdef.pl to sort symbols that get the same entry number,
13254 and make sure the automatically generated functions `ERR_load_*`
13255 become part of libeay.num as well.
13259 * New function SSL_renegotiate_pending(). This returns true once
13260 renegotiation has been requested (either SSL_renegotiate() call
13261 or HelloRequest/ClientHello received from the peer) and becomes
13262 false once a handshake has been completed.
13263 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13264 sends a HelloRequest, but does not ensure that a handshake takes
13265 place. SSL_renegotiate_pending() is useful for checking if the
13266 client has followed the request.)
13270 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13271 By default, clients may request session resumption even during
13272 renegotiation (if session ID contexts permit); with this option,
13273 session resumption is possible only in the first handshake.
13275 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13276 more bits available for options that should not be part of
13277 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13281 * Add some demos for certificate and certificate request creation.
13285 * Make maximum certificate chain size accepted from the peer application
13286 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13287 "Douglas E. Engert" <deengert@anl.gov>.
13291 * Add support for shared libraries for Unixware-7
13292 (Boyd Lynn Gerber <gerberb@zenez.com>).
13296 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13297 be done prior to destruction. Use this to unload error strings from
13298 ENGINEs that load their own error strings. NB: This adds two new API
13299 functions to "get" and "set" this destroy handler in an ENGINE.
13303 * Alter all existing ENGINE implementations (except "openssl" and
13304 "openbsd") to dynamically instantiate their own error strings. This
13305 makes them more flexible to be built both as statically-linked ENGINEs
13306 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13307 Also, add stub code to each that makes building them as self-contained
13308 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13312 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13313 implementations into applications that are completely implemented in
13314 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13315 commands that can be used to configure what shared-library to load and
13316 to control aspects of the way it is handled. Also, made an update to
13317 the [README-Engine.md](README-Engine.md) file
13318 that brings its information up-to-date and
13319 provides some information and instructions on the "dynamic" ENGINE
13320 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13324 * Make it possible to unload ranges of ERR strings with a new
13325 "ERR_unload_strings" function.
13329 * Add a copy() function to EVP_MD.
13333 * Make EVP_MD routines take a context pointer instead of just the
13334 md_data void pointer.
13338 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13339 that the digest can only process a single chunk of data
13340 (typically because it is provided by a piece of
13341 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13342 is only going to provide a single chunk of data, and hence the
13343 framework needn't accumulate the data for oneshot drivers.
13347 * As with "ERR", make it possible to replace the underlying "ex_data"
13348 functions. This change also alters the storage and management of global
13349 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13350 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13351 index counters. The API functions that use this state have been changed
13352 to take a "class_index" rather than pointers to the class's local STACK
13353 and counter, and there is now an API function to dynamically create new
13354 classes. This centralisation allows us to (a) plug a lot of the
13355 thread-safety problems that existed, and (b) makes it possible to clean
13356 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13357 such data would previously have always leaked in application code and
13358 workarounds were in place to make the memory debugging turn a blind eye
13359 to it. Application code that doesn't use this new function will still
13360 leak as before, but their memory debugging output will announce it now
13361 rather than letting it slide.
13363 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13364 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13365 has a return value to indicate success or failure.
13369 * Make it possible to replace the underlying "ERR" functions such that the
13370 global state (2 LHASH tables and 2 locks) is only used by the "default"
13371 implementation. This change also adds two functions to "get" and "set"
13372 the implementation prior to it being automatically set the first time
13373 any other ERR function takes place. Ie. an application can call "get",
13374 pass the return value to a module it has just loaded, and that module
13375 can call its own "set" function using that value. This means the
13376 module's "ERR" operations will use (and modify) the error state in the
13377 application and not in its own statically linked copy of OpenSSL code.
13381 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13382 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13383 the operation, and provides a more encapsulated way for external code
13384 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13385 to use these functions rather than manually incrementing the counts.
13387 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13391 * Add EVP test program.
13395 * Add symmetric cipher support to ENGINE. Expect the API to change!
13399 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13400 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13401 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13402 These allow a CRL to be built without having to access X509_CRL fields
13403 directly. Modify 'ca' application to use new functions.
13407 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13408 bug workarounds. Rollback attack detection is a security feature.
13409 The problem will only arise on OpenSSL servers when TLSv1 is not
13410 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13411 Software authors not wanting to support TLSv1 will have special reasons
13412 for their choice and can explicitly enable this option.
13414 *Bodo Moeller, Lutz Jaenicke*
13416 * Rationalise EVP so it can be extended: don't include a union of
13417 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13418 (similar to those existing for EVP_CIPHER_CTX).
13423 EVP_MD_CTX_init(&md); /* new function call */
13424 EVP_DigestInit(&md, EVP_sha1());
13425 EVP_DigestUpdate(&md, in, len);
13426 EVP_DigestFinal(&md, out, NULL);
13427 EVP_MD_CTX_cleanup(&md); /* new function call */
13431 * Make DES key schedule conform to the usual scheme, as well as
13432 correcting its structure. This means that calls to DES functions
13433 now have to pass a pointer to a des_key_schedule instead of a
13434 plain des_key_schedule (which was actually always a pointer
13437 des_key_schedule ks;
13439 des_set_key_checked(..., &ks);
13440 des_ncbc_encrypt(..., &ks, ...);
13442 (Note that a later change renames 'des_...' into 'DES_...'.)
13446 * Initial reduction of linker bloat: the use of some functions, such as
13447 PEM causes large amounts of unused functions to be linked in due to
13448 poor organisation. For example pem_all.c contains every PEM function
13449 which has a knock on effect of linking in large amounts of (unused)
13450 ASN1 code. Grouping together similar functions and splitting unrelated
13451 functions prevents this.
13455 * Cleanup of EVP macros.
13459 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13460 correct `_ecb suffix`.
13464 * Add initial OCSP responder support to ocsp application. The
13465 revocation information is handled using the text based index
13466 use by the ca application. The responder can either handle
13467 requests generated internally, supplied in files (for example
13468 via a CGI script) or using an internal minimal server.
13472 * Add configuration choices to get zlib compression for TLS.
13476 * Changes to Kerberos SSL for RFC 2712 compliance:
13477 1. Implemented real KerberosWrapper, instead of just using
13478 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13479 2. Implemented optional authenticator field of KerberosWrapper.
13481 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13482 and authenticator structs; see crypto/krb5/.
13484 Generalized Kerberos calls to support multiple Kerberos libraries.
13485 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13486 via Richard Levitte*
13488 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13489 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13490 values for each of the key sizes rather than having just
13491 parameters (and 'speed' generating keys each time).
13495 * Speed up EVP routines.
13498 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13499 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13500 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13501 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13503 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13504 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13505 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13508 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13510 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13514 * Added the OS2-EMX target.
13516 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13518 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13519 New functions to support `NCONF` routines in extension code.
13520 New function `CONF_set_nconf()`
13521 to allow functions which take an `NCONF` to also handle the old `LHASH`
13522 structure: this means that the old `CONF` compatible routines can be
13523 retained (in particular w.rt. extensions) without having to duplicate the
13524 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13528 * Enhance the general user interface with mechanisms for inner control
13529 and with possibilities to have yes/no kind of prompts.
13533 * Change all calls to low-level digest routines in the library and
13534 applications to use EVP. Add missing calls to HMAC_cleanup() and
13535 don't assume HMAC_CTX can be copied using memcpy().
13537 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13539 * Add the possibility to control engines through control names but with
13540 arbitrary arguments instead of just a string.
13541 Change the key loaders to take a UI_METHOD instead of a callback
13542 function pointer. NOTE: this breaks binary compatibility with earlier
13543 versions of OpenSSL [engine].
13544 Adapt the nCipher code for these new conditions and add a card insertion
13549 * Enhance the general user interface with mechanisms to better support
13550 dialog box interfaces, application-defined prompts, the possibility
13551 to use defaults (for example default passwords from somewhere else)
13552 and interrupts/cancellations.
13556 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13557 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13561 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13562 tidy up some unnecessarily weird code in 'sk_new()').
13564 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13566 * Change the key loading routines for ENGINEs to use the same kind
13567 callback (pem_password_cb) as all other routines that need this
13572 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13573 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13574 than this minimum value is recommended.
13578 * New random seeder for OpenVMS, using the system process statistics
13579 that are easily reachable.
13583 * Windows apparently can't transparently handle global
13584 variables defined in DLLs. Initialisations such as:
13586 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13588 won't compile. This is used by the any applications that need to
13589 declare their own ASN1 modules. This was fixed by adding the option
13590 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13591 needed for static libraries under Win32.
13595 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13596 setting of purpose and trust fields. New X509_STORE trust and
13597 purpose functions and tidy up setting in other SSL functions.
13601 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13602 structure. These are inherited by X509_STORE_CTX when it is
13603 initialised. This allows various defaults to be set in the
13604 X509_STORE structure (such as flags for CRL checking and custom
13605 purpose or trust settings) for functions which only use X509_STORE_CTX
13606 internally such as S/MIME.
13608 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13609 trust settings if they are not set in X509_STORE. This allows X509_STORE
13610 purposes and trust (in S/MIME for example) to override any set by default.
13612 Add command line options for CRL checking to smime, s_client and s_server
13617 * Initial CRL based revocation checking. If the CRL checking flag(s)
13618 are set then the CRL is looked up in the X509_STORE structure and
13619 its validity and signature checked, then if the certificate is found
13620 in the CRL the verify fails with a revoked error.
13622 Various new CRL related callbacks added to X509_STORE_CTX structure.
13624 Command line options added to 'verify' application to support this.
13626 This needs some additional work, such as being able to handle multiple
13627 CRLs with different times, extension based lookup (rather than just
13628 by subject name) and ultimately more complete V2 CRL extension
13633 * Add a general user interface API (crypto/ui/). This is designed
13634 to replace things like des_read_password and friends (backward
13635 compatibility functions using this new API are provided).
13636 The purpose is to remove prompting functions from the DES code
13637 section as well as provide for prompting through dialog boxes in
13638 a window system and the like.
13642 * Add "ex_data" support to ENGINE so implementations can add state at a
13643 per-structure level rather than having to store it globally.
13647 * Make it possible for ENGINE structures to be copied when retrieved by
13648 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13649 This causes the "original" ENGINE structure to act like a template,
13650 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13651 operational state can be localised to each ENGINE structure, despite the
13652 fact they all share the same "methods". New ENGINE structures returned in
13653 this case have no functional references and the return value is the single
13654 structural reference. This matches the single structural reference returned
13655 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13660 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13661 needs to match any other type at all we need to manually clear the
13666 * Changes to the "openssl engine" utility to include;
13667 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13668 about an ENGINE's available control commands.
13669 - executing control commands from command line arguments using the
13670 '-pre' and '-post' switches. '-post' is only used if '-t' is
13671 specified and the ENGINE is successfully initialised. The syntax for
13672 the individual commands are colon-separated, for example;
13673 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13677 * New dynamic control command support for ENGINEs. ENGINEs can now
13678 declare their own commands (numbers), names (strings), descriptions,
13679 and input types for run-time discovery by calling applications. A
13680 subset of these commands are implicitly classed as "executable"
13681 depending on their input type, and only these can be invoked through
13682 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13683 can be based on user input, config files, etc). The distinction is
13684 that "executable" commands cannot return anything other than a boolean
13685 result and can only support numeric or string input, whereas some
13686 discoverable commands may only be for direct use through
13687 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13688 pointers, or other custom uses. The "executable" commands are to
13689 support parameterisations of ENGINE behaviour that can be
13690 unambiguously defined by ENGINEs and used consistently across any
13691 OpenSSL-based application. Commands have been added to all the
13692 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13693 control over shared-library paths without source code alterations.
13697 * Changed all ENGINE implementations to dynamically allocate their
13698 ENGINEs rather than declaring them statically. Apart from this being
13699 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13700 this also allows the implementations to compile without using the
13701 internal engine_int.h header.
13705 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13706 'const' value. Any code that should be able to modify a RAND_METHOD
13707 should already have non-const pointers to it (ie. they should only
13708 modify their own ones).
13712 * Made a variety of little tweaks to the ENGINE code.
13713 - "atalla" and "ubsec" string definitions were moved from header files
13714 to C code. "nuron" string definitions were placed in variables
13715 rather than hard-coded - allowing parameterisation of these values
13716 later on via ctrl() commands.
13717 - Removed unused "#if 0"'d code.
13718 - Fixed engine list iteration code so it uses ENGINE_free() to release
13719 structural references.
13720 - Constified the RAND_METHOD element of ENGINE structures.
13721 - Constified various get/set functions as appropriate and added
13722 missing functions (including a catch-all ENGINE_cpy that duplicates
13723 all ENGINE values onto a new ENGINE except reference counts/state).
13724 - Removed NULL parameter checks in get/set functions. Setting a method
13725 or function to NULL is a way of cancelling out a previously set
13726 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13727 and doesn't justify the extra error symbols and code.
13728 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13729 flags from engine_int.h to engine.h.
13730 - Changed prototypes for ENGINE handler functions (init(), finish(),
13731 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13735 * Implement binary inversion algorithm for BN_mod_inverse in addition
13736 to the algorithm using long division. The binary algorithm can be
13737 used only if the modulus is odd. On 32-bit systems, it is faster
13738 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13739 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13740 up to 450 bits. In 64-bit environments, the binary algorithm
13741 appears to be advantageous for much longer moduli; here we use it
13742 for moduli up to 2048 bits.
13746 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13747 could not support the combine flag in choice fields.
13751 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13752 extensions from a certificate request to the certificate.
13756 * Allow multiple 'certopt' and 'nameopt' options to be separated
13757 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13758 file: this allows the display of the certificate about to be
13759 signed to be customised, to allow certain fields to be included
13760 or excluded and extension details. The old system didn't display
13761 multicharacter strings properly, omitted fields not in the policy
13762 and couldn't display additional details such as extensions.
13766 * Function EC_POINTs_mul for multiple scalar multiplication
13767 of an arbitrary number of elliptic curve points
13768 \sum scalars[i]*points[i],
13769 optionally including the generator defined for the EC_GROUP:
13770 scalar*generator + \sum scalars[i]*points[i].
13772 EC_POINT_mul is a simple wrapper function for the typical case
13773 that the point list has just one item (besides the optional
13778 * First EC_METHODs for curves over GF(p):
13780 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13781 operations and provides various method functions that can also
13782 operate with faster implementations of modular arithmetic.
13784 EC_GFp_mont_method() reuses most functions that are part of
13785 EC_GFp_simple_method, but uses Montgomery arithmetic.
13787 *Bodo Moeller; point addition and point doubling
13788 implementation directly derived from source code provided by
13789 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13791 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13792 crypto/ec/ec_lib.c):
13794 Curves are EC_GROUP objects (with an optional group generator)
13795 based on EC_METHODs that are built into the library.
13797 Points are EC_POINT objects based on EC_GROUP objects.
13799 Most of the framework would be able to handle curves over arbitrary
13800 finite fields, but as there are no obvious types for fields other
13801 than GF(p), some functions are limited to that for now.
13805 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13806 that the file contains a complete HTTP response.
13810 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13811 change the def and num file printf format specifier from "%-40sXXX"
13812 to "%-39s XXX". The latter will always guarantee a space after the
13813 field while the former will cause them to run together if the field
13814 is 40 of more characters long.
13818 * Constify the cipher and digest 'method' functions and structures
13819 and modify related functions to take constant EVP_MD and EVP_CIPHER
13824 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13825 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13829 * Modify `EVP_Digest*()` routines so they now return values. Although the
13830 internal software routines can never fail additional hardware versions
13835 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13837 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13838 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13841 ERR_R_NESTED_ASN1_ERROR
13843 ERR_R_MISSING_ASN1_EOS
13844 were 4 .. 9, conflicting with
13845 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13847 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13848 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13850 Add new error code 'ERR_R_INTERNAL_ERROR'.
13854 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13859 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13860 sets the subject name for a new request or supersedes the
13861 subject name in a given request. Formats that can be parsed are
13862 'CN=Some Name, OU=myOU, C=IT'
13864 'CN=Some Name/OU=myOU/C=IT'.
13866 Add options '-batch' and '-verbose' to 'openssl req'.
13868 *Massimiliano Pala <madwolf@hackmasters.net>*
13870 * Introduce the possibility to access global variables through
13871 functions on platform were that's the best way to handle exporting
13872 global variables in shared libraries. To enable this functionality,
13873 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13874 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13875 is normally done by Configure or something similar).
13877 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13878 in the source file (foo.c) like this:
13880 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13881 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13883 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13884 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13886 OPENSSL_DECLARE_GLOBAL(int,foo);
13887 #define foo OPENSSL_GLOBAL_REF(foo)
13888 OPENSSL_DECLARE_GLOBAL(double,bar);
13889 #define bar OPENSSL_GLOBAL_REF(bar)
13891 The #defines are very important, and therefore so is including the
13892 header file everywhere where the defined globals are used.
13894 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13895 of ASN.1 items, but that structure is a bit different.
13897 The largest change is in util/mkdef.pl which has been enhanced with
13898 better and easier to understand logic to choose which symbols should
13899 go into the Windows .def files as well as a number of fixes and code
13900 cleanup (among others, algorithm keywords are now sorted
13901 lexicographically to avoid constant rewrites).
13905 * In BN_div() keep a copy of the sign of 'num' before writing the
13906 result to 'rm' because if rm==num the value will be overwritten
13907 and produce the wrong result if 'num' is negative: this caused
13908 problems with BN_mod() and BN_nnmod().
13912 * Function OCSP_request_verify(). This checks the signature on an
13913 OCSP request and verifies the signer certificate. The signer
13914 certificate is just checked for a generic purpose and OCSP request
13919 * Add OCSP_check_validity() function to check the validity of OCSP
13920 responses. OCSP responses are prepared in real time and may only
13921 be a few seconds old. Simply checking that the current time lies
13922 between thisUpdate and nextUpdate max reject otherwise valid responses
13923 caused by either OCSP responder or client clock inaccuracy. Instead
13924 we allow thisUpdate and nextUpdate to fall within a certain period of
13925 the current time. The age of the response can also optionally be
13926 checked. Two new options -validity_period and -status_age added to
13931 * If signature or public key algorithm is unrecognized print out its
13932 OID rather that just UNKNOWN.
13936 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13937 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13938 ID to be generated from the issuer certificate alone which can then be
13939 passed to OCSP_id_issuer_cmp().
13943 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13944 ASN1 modules to export functions returning ASN1_ITEM pointers
13945 instead of the ASN1_ITEM structures themselves. This adds several
13946 new macros which allow the underlying ASN1 function/structure to
13947 be accessed transparently. As a result code should not use ASN1_ITEM
13948 references directly (such as &X509_it) but instead use the relevant
13949 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13950 use of the new ASN1 code on platforms where exporting structures
13951 is problematical (for example in shared libraries) but exporting
13952 functions returning pointers to structures is not.
13956 * Add support for overriding the generation of SSL/TLS session IDs.
13957 These callbacks can be registered either in an SSL_CTX or per SSL.
13958 The purpose of this is to allow applications to control, if they wish,
13959 the arbitrary values chosen for use as session IDs, particularly as it
13960 can be useful for session caching in multiple-server environments. A
13961 command-line switch for testing this (and any client code that wishes
13962 to use such a feature) has been added to "s_server".
13964 *Geoff Thorpe, Lutz Jaenicke*
13966 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13967 of the form `#if defined(...) || defined(...) || ...` and
13968 `#if !defined(...) && !defined(...) && ...`. This also avoids
13969 the growing number of special cases it was previously handling.
13973 * Make all configuration macros available for application by making
13974 sure they are available in opensslconf.h, by giving them names starting
13975 with `OPENSSL_` to avoid conflicts with other packages and by making
13976 sure e_os2.h will cover all platform-specific cases together with
13978 Additionally, it is now possible to define configuration/platform-
13979 specific names (called "system identities"). In the C code, these
13980 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13981 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13982 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13987 * New option -set_serial to 'req' and 'x509' this allows the serial
13988 number to use to be specified on the command line. Previously self
13989 signed certificates were hard coded with serial number 0 and the
13990 CA options of 'x509' had to use a serial number in a file which was
13995 * New options to 'ca' utility to support V2 CRL entry extensions.
13996 Currently CRL reason, invalidity date and hold instruction are
13997 supported. Add new CRL extensions to V3 code and some new objects.
14001 * New function EVP_CIPHER_CTX_set_padding() this is used to
14002 disable standard block padding (aka PKCS#5 padding) in the EVP
14003 API, which was previously mandatory. This means that the data is
14004 not padded in any way and so the total length much be a multiple
14005 of the block size, otherwise an error occurs.
14009 * Initial (incomplete) OCSP SSL support.
14013 * New function OCSP_parse_url(). This splits up a URL into its host,
14014 port and path components: primarily to parse OCSP URLs. New -url
14015 option to ocsp utility.
14019 * New nonce behavior. The return value of OCSP_check_nonce() now
14020 reflects the various checks performed. Applications can decide
14021 whether to tolerate certain situations such as an absent nonce
14022 in a response when one was present in a request: the ocsp application
14023 just prints out a warning. New function OCSP_add1_basic_nonce()
14024 this is to allow responders to include a nonce in a response even if
14025 the request is nonce-less.
14029 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14030 skipped when using openssl x509 multiple times on a single input file,
14031 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14035 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14036 set string type: to handle setting ASN1_TIME structures. Fix ca
14037 utility to correctly initialize revocation date of CRLs.
14041 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14042 the clients preferred ciphersuites and rather use its own preferences.
14043 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14044 Internet Explorer by ensuring unchanged hash method during stepup.
14045 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14049 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14050 to aes and add a new 'exist' option to print out symbols that don't
14055 * Additional options to ocsp utility to allow flags to be set and
14056 additional certificates supplied.
14060 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14061 OCSP client a number of certificate to only verify the response
14066 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14067 handle the new API. Currently only ECB, CBC modes supported. Add new
14070 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14071 Encryption Standard (AES) Ciphersuites for Transport Layer
14072 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14073 not enabled by default and were not part of the "ALL" ciphersuite
14074 alias because they were not yet official; they could be
14075 explicitly requested by specifying the "AESdraft" ciphersuite
14076 group alias. In the final release of OpenSSL 0.9.7, the group
14077 alias is called "AES" and is part of "ALL".)
14079 *Ben Laurie, Steve Henson, Bodo Moeller*
14081 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14082 request to response.
14086 * Functions for OCSP responders. OCSP_request_onereq_count(),
14087 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14088 extract information from a certificate request. OCSP_response_create()
14089 creates a response and optionally adds a basic response structure.
14090 OCSP_basic_add1_status() adds a complete single response to a basic
14091 response and returns the OCSP_SINGLERESP structure just added (to allow
14092 extensions to be included for example). OCSP_basic_add1_cert() adds a
14093 certificate to a basic response and OCSP_basic_sign() signs a basic
14094 response with various flags. New helper functions ASN1_TIME_check()
14095 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14096 (converts ASN1_TIME to GeneralizedTime).
14100 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14101 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14102 structure from a certificate. X509_pubkey_digest() digests the public_key
14103 contents: this is used in various key identifiers.
14107 * Make sk_sort() tolerate a NULL argument.
14109 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14111 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14112 passed by the function are trusted implicitly. If any of them signed the
14113 response then it is assumed to be valid and is not verified.
14117 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14118 to data. This was previously part of the PKCS7 ASN1 code. This
14119 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14120 *Steve Henson, reported by Kenneth R. Robinette
14121 <support@securenetterm.com>*
14123 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14124 routines: without these tracing memory leaks is very painful.
14125 Fix leaks in PKCS12 and PKCS7 routines.
14129 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14130 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14131 effectively meant GeneralizedTime would never be used. Now it
14132 is initialised to -1 but X509_time_adj() now has to check the value
14133 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14134 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14135 *Steve Henson, reported by Kenneth R. Robinette
14136 <support@securenetterm.com>*
14138 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14139 result in a zero length in the ASN1_INTEGER structure which was
14140 not consistent with the structure when d2i_ASN1_INTEGER() was used
14141 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14142 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14143 where it did not print out a minus for negative ASN1_INTEGER.
14147 * Add summary printout to ocsp utility. The various functions which
14148 convert status values to strings have been renamed to:
14149 OCSP_response_status_str(), OCSP_cert_status_str() and
14150 OCSP_crl_reason_str() and are no longer static. New options
14151 to verify nonce values and to disable verification. OCSP response
14152 printout format cleaned up.
14156 * Add additional OCSP certificate checks. These are those specified
14157 in RFC2560. This consists of two separate checks: the CA of the
14158 certificate being checked must either be the OCSP signer certificate
14159 or the issuer of the OCSP signer certificate. In the latter case the
14160 OCSP signer certificate must contain the OCSP signing extended key
14161 usage. This check is performed by attempting to match the OCSP
14162 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14163 in the OCSP_CERTID structures of the response.
14167 * Initial OCSP certificate verification added to OCSP_basic_verify()
14168 and related routines. This uses the standard OpenSSL certificate
14169 verify routines to perform initial checks (just CA validity) and
14170 to obtain the certificate chain. Then additional checks will be
14171 performed on the chain. Currently the root CA is checked to see
14172 if it is explicitly trusted for OCSP signing. This is used to set
14173 a root CA as a global signing root: that is any certificate that
14174 chains to that CA is an acceptable OCSP signing certificate.
14178 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14179 extensions from a separate configuration file.
14180 As when reading extensions from the main configuration file,
14181 the '-extensions ...' option may be used for specifying the
14184 *Massimiliano Pala <madwolf@comune.modena.it>*
14186 * New OCSP utility. Allows OCSP requests to be generated or
14187 read. The request can be sent to a responder and the output
14188 parsed, outputted or printed in text form. Not complete yet:
14189 still needs to check the OCSP response validity.
14193 * New subcommands for 'openssl ca':
14194 `openssl ca -status <serial>` prints the status of the cert with
14195 the given serial number (according to the index file).
14196 `openssl ca -updatedb` updates the expiry status of certificates
14199 *Massimiliano Pala <madwolf@comune.modena.it>*
14201 * New '-newreq-nodes' command option to CA.pl. This is like
14202 '-newreq', but calls 'openssl req' with the '-nodes' option
14203 so that the resulting key is not encrypted.
14205 *Damien Miller <djm@mindrot.org>*
14207 * New configuration for the GNU Hurd.
14209 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14211 * Initial code to implement OCSP basic response verify. This
14212 is currently incomplete. Currently just finds the signer's
14213 certificate and verifies the signature on the response.
14217 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14218 value of OPENSSLDIR. This is available via the new '-d' option
14219 to 'openssl version', and is also included in 'openssl version -a'.
14223 * Allowing defining memory allocation callbacks that will be given
14224 file name and line number information in additional arguments
14225 (a `const char*` and an int). The basic functionality remains, as
14226 well as the original possibility to just replace malloc(),
14227 realloc() and free() by functions that do not know about these
14228 additional arguments. To register and find out the current
14229 settings for extended allocation functions, the following
14230 functions are provided:
14232 CRYPTO_set_mem_ex_functions
14233 CRYPTO_set_locked_mem_ex_functions
14234 CRYPTO_get_mem_ex_functions
14235 CRYPTO_get_locked_mem_ex_functions
14237 These work the same way as CRYPTO_set_mem_functions and friends.
14238 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14239 extended allocation function is enabled.
14240 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14241 a conventional allocation function is enabled.
14243 *Richard Levitte, Bodo Moeller*
14245 * Finish off removing the remaining LHASH function pointer casts.
14246 There should no longer be any prototype-casting required when using
14247 the LHASH abstraction, and any casts that remain are "bugs". See
14248 the callback types and macros at the head of lhash.h for details
14249 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14253 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14254 If /dev/[u]random devices are not available or do not return enough
14255 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14257 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14258 /etc/entropy will be queried once each in this sequence, querying stops
14259 when enough entropy was collected without querying more sockets.
14263 * Change the Unix RAND_poll() variant to be able to poll several
14264 random devices, as specified by DEVRANDOM, until a sufficient amount
14265 of data has been collected. We spend at most 10 ms on each file
14266 (select timeout) and read in non-blocking mode. DEVRANDOM now
14267 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14268 (previously it was just the string "/dev/urandom"), so on typical
14269 platforms the 10 ms delay will never occur.
14270 Also separate out the Unix variant to its own file, rand_unix.c.
14271 For VMS, there's a currently-empty rand_vms.c.
14275 * Move OCSP client related routines to ocsp_cl.c. These
14276 provide utility functions which an application needing
14277 to issue a request to an OCSP responder and analyse the
14278 response will typically need: as opposed to those which an
14279 OCSP responder itself would need which will be added later.
14281 OCSP_request_sign() signs an OCSP request with an API similar
14282 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14283 response. OCSP_response_get1_basic() extracts basic response
14284 from response. OCSP_resp_find_status(): finds and extracts status
14285 information from an OCSP_CERTID structure (which will be created
14286 when the request structure is built). These are built from lower
14287 level functions which work on OCSP_SINGLERESP structures but
14288 won't normally be used unless the application wishes to examine
14289 extensions in the OCSP response for example.
14291 Replace nonce routines with a pair of functions.
14292 OCSP_request_add1_nonce() adds a nonce value and optionally
14293 generates a random value. OCSP_check_nonce() checks the
14294 validity of the nonce in an OCSP response.
14298 * Change function OCSP_request_add() to OCSP_request_add0_id().
14299 This doesn't copy the supplied OCSP_CERTID and avoids the
14300 need to free up the newly created id. Change return type
14301 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14302 This can then be used to add extensions to the request.
14303 Deleted OCSP_request_new(), since most of its functionality
14304 is now in OCSP_REQUEST_new() (and the case insensitive name
14305 clash) apart from the ability to set the request name which
14306 will be added elsewhere.
14310 * Update OCSP API. Remove obsolete extensions argument from
14311 various functions. Extensions are now handled using the new
14312 OCSP extension code. New simple OCSP HTTP function which
14313 can be used to send requests and parse the response.
14317 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14318 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14319 uses the special reorder version of SET OF to sort the attributes
14320 and reorder them to match the encoded order. This resolves a long
14321 standing problem: a verify on a PKCS7 structure just after signing
14322 it used to fail because the attribute order did not match the
14323 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14324 it uses the received order. This is necessary to tolerate some broken
14325 software that does not order SET OF. This is handled by encoding
14326 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14327 to produce the required SET OF.
14331 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14332 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14333 files to get correct declarations of the ASN.1 item variables.
14337 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14338 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14339 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14340 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14341 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14342 ASN1_ITEM and no wrapper functions.
14346 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14347 replace the old function pointer based I/O routines. Change most of
14348 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14352 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14353 lines, recognize more "algorithms" that can be deselected, and make
14354 it complain about algorithm deselection that isn't recognised.
14358 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14359 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14360 to use new functions. Add NO_ASN1_OLD which can be set to remove
14361 some old style ASN1 functions: this can be used to determine if old
14362 code will still work when these eventually go away.
14366 * New extension functions for OCSP structures, these follow the
14367 same conventions as certificates and CRLs.
14371 * New function X509V3_add1_i2d(). This automatically encodes and
14372 adds an extension. Its behaviour can be customised with various
14373 flags to append, replace or delete. Various wrappers added for
14374 certificates and CRLs.
14378 * Fix to avoid calling the underlying ASN1 print routine when
14379 an extension cannot be parsed. Correct a typo in the
14380 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14384 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14385 entries for variables.
14389 * Add functionality to `apps/openssl.c` for detecting locking
14390 problems: As the program is single-threaded, all we have
14391 to do is register a locking callback using an array for
14392 storing which locks are currently held by the program.
14396 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14397 SSL_get_ex_data_X509_STORE_idx(), which is used in
14398 ssl_verify_cert_chain() and thus can be called at any time
14399 during TLS/SSL handshakes so that thread-safety is essential.
14400 Unfortunately, the ex_data design is not at all suited
14401 for multi-threaded use, so it probably should be abolished.
14405 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14407 *Broadcom, tweaked and integrated by Geoff Thorpe*
14409 * Move common extension printing code to new function
14410 X509V3_print_extensions(). Reorganise OCSP print routines and
14411 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14415 * New function X509_signature_print() to remove duplication in some
14420 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14421 set (this was treated exactly the same as SET OF previously). This
14422 is used to reorder the STACK representing the structure to match the
14423 encoding. This will be used to get round a problem where a PKCS7
14424 structure which was signed could not be verified because the STACK
14425 order did not reflect the encoded order.
14429 * Reimplement the OCSP ASN1 module using the new code.
14433 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14434 for its ASN1 operations. The old style function pointers still exist
14435 for now but they will eventually go away.
14439 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14440 completely replaces the old ASN1 functionality with a table driven
14441 encoder and decoder which interprets an ASN1_ITEM structure describing
14442 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14443 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14444 has also been converted to the new form.
14448 * Change BN_mod_exp_recp so that negative moduli are tolerated
14449 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14450 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14451 for negative moduli.
14455 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14456 of not touching the result's sign bit.
14460 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14465 * Changed the LHASH code to use prototypes for callbacks, and created
14466 macros to declare and implement thin (optionally static) functions
14467 that provide type-safety and avoid function pointer casting for the
14468 type-specific callbacks.
14472 * Added Kerberos Cipher Suites to be used with TLS, as written in
14474 *Veers Staats <staatsvr@asc.hpc.mil>,
14475 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14477 * Reformat the FAQ so the different questions and answers can be divided
14478 in sections depending on the subject.
14482 * Have the zlib compression code load ZLIB.DLL dynamically under
14487 * New function BN_mod_sqrt for computing square roots modulo a prime
14488 (using the probabilistic Tonelli-Shanks algorithm unless
14489 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14490 be handled deterministically).
14492 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14494 * Make BN_mod_inverse faster by explicitly handling small quotients
14495 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14496 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14500 * New function BN_kronecker.
14504 * Fix BN_gcd so that it works on negative inputs; the result is
14505 positive unless both parameters are zero.
14506 Previously something reasonably close to an infinite loop was
14507 possible because numbers could be growing instead of shrinking
14508 in the implementation of Euclid's algorithm.
14512 * Fix BN_is_word() and BN_is_one() macros to take into account the
14513 sign of the number in question.
14515 Fix BN_is_word(a,w) to work correctly for w == 0.
14517 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14518 because its test if the absolute value of 'a' equals 'w'.
14519 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14520 it exists mostly for use in the implementations of BN_is_zero(),
14521 BN_is_one(), and BN_is_word().
14525 * New function BN_swap.
14529 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14530 the exponentiation functions are more likely to produce reasonable
14531 results on negative inputs.
14535 * Change BN_mod_mul so that the result is always non-negative.
14536 Previously, it could be negative if one of the factors was negative;
14537 I don't think anyone really wanted that behaviour.
14541 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14542 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14543 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14544 and add new functions:
14553 BN_mod_lshift1_quick
14555 BN_mod_lshift_quick
14557 These functions always generate non-negative results.
14559 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14560 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14562 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14563 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14564 be reduced modulo `m`.
14566 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14569 The following entry accidentally appeared in the CHANGES file
14570 distributed with OpenSSL 0.9.7. The modifications described in
14571 it do *not* apply to OpenSSL 0.9.7.
14573 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14574 was actually never needed) and in BN_mul(). The removal in BN_mul()
14575 required a small change in bn_mul_part_recursive() and the addition
14576 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14577 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14578 bn_sub_words() and bn_add_words() except they take arrays with
14584 * In 'openssl passwd', verify passwords read from the terminal
14585 unless the '-salt' option is used (which usually means that
14586 verification would just waste user's time since the resulting
14587 hash is going to be compared with some given password hash)
14588 or the new '-noverify' option is used.
14590 This is an incompatible change, but it does not affect
14591 non-interactive use of 'openssl passwd' (passwords on the command
14592 line, '-stdin' option, '-in ...' option) and thus should not
14593 cause any problems.
14597 * Remove all references to RSAref, since there's no more need for it.
14601 * Make DSO load along a path given through an environment variable
14602 (SHLIB_PATH) with shl_load().
14606 * Constify the ENGINE code as a result of BIGNUM constification.
14607 Also constify the RSA code and most things related to it. In a
14608 few places, most notable in the depth of the ASN.1 code, ugly
14609 casts back to non-const were required (to be solved at a later
14614 * Make it so the openssl application has all engines loaded by default.
14618 * Constify the BIGNUM routines a little more.
14622 * Add the following functions:
14624 ENGINE_load_cswift()
14626 ENGINE_load_atalla()
14627 ENGINE_load_nuron()
14628 ENGINE_load_builtin_engines()
14630 That way, an application can itself choose if external engines that
14631 are built-in in OpenSSL shall ever be used or not. The benefit is
14632 that applications won't have to be linked with libdl or other dso
14633 libraries unless it's really needed.
14635 Changed 'openssl engine' to load all engines on demand.
14636 Changed the engine header files to avoid the duplication of some
14637 declarations (they differed!).
14641 * 'openssl engine' can now list capabilities.
14645 * Better error reporting in 'openssl engine'.
14649 * Never call load_dh_param(NULL) in s_server.
14653 * Add engine application. It can currently list engines by name and
14654 identity, and test if they are actually available.
14658 * Improve RPM specification file by forcing symbolic linking and making
14659 sure the installed documentation is also owned by root.root.
14661 *Damien Miller <djm@mindrot.org>*
14663 * Give the OpenSSL applications more possibilities to make use of
14664 keys (public as well as private) handled by engines.
14668 * Add OCSP code that comes from CertCo.
14672 * Add VMS support for the Rijndael code.
14676 * Added untested support for Nuron crypto accelerator.
14680 * Add support for external cryptographic devices. This code was
14681 previously distributed separately as the "engine" branch.
14683 *Geoff Thorpe, Richard Levitte*
14685 * Rework the filename-translation in the DSO code. It is now possible to
14686 have far greater control over how a "name" is turned into a filename
14687 depending on the operating environment and any oddities about the
14688 different shared library filenames on each system.
14692 * Support threads on FreeBSD-elf in Configure.
14696 * Fix for SHA1 assembly problem with MASM: it produces
14697 warnings about corrupt line number information when assembling
14698 with debugging information. This is caused by the overlapping
14701 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14704 NCONF_get_number() has no error checking at all. As a replacement,
14705 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14706 promoted strongly. The old NCONF_get_number is kept around for
14707 binary backward compatibility.
14708 Make it possible for methods to load from something other than a BIO,
14709 by providing a function pointer that is given a name instead of a BIO.
14710 For example, this could be used to load configuration data from an
14715 * Fix for non blocking accept BIOs. Added new I/O special reason
14716 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14717 with non blocking I/O was not possible because no retry code was
14718 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14723 * Added the beginnings of Rijndael support.
14727 * Fix for bug in DirectoryString mask setting. Add support for
14728 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14729 to allow certificate printing to more controllable, additional
14730 'certopt' option to 'x509' to allow new printing options to be
14735 * Clean old EAY MD5 hack from e_os.h.
14739 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14741 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14742 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14744 *Joe Orton, Steve Henson*
14746 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14748 * Fix additional bug revealed by the NISCC test suite:
14750 Stop bug triggering large recursion when presented with
14751 certain ASN.1 tags ([CVE-2003-0851])
14755 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14757 * Fix various bugs revealed by running the NISCC test suite:
14759 Stop out of bounds reads in the ASN1 code when presented with
14760 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14762 If verify callback ignores invalid public key errors don't try to check
14763 certificate signature with the NULL public key.
14767 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14768 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14773 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14774 extra data after the compression methods not only for TLS 1.0
14775 but also for SSL 3.0 (as required by the specification).
14777 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14779 * Change X509_certificate_type() to mark the key as exported/exportable
14780 when it's 512 *bits* long, not 512 bytes.
14784 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14786 * Countermeasure against the Klima-Pokorny-Rosa extension of
14787 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14788 a protocol version number mismatch like a decryption error
14789 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14793 * Turn on RSA blinding by default in the default implementation
14794 to avoid a timing attack. Applications that don't want it can call
14795 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14796 They would be ill-advised to do so in most cases.
14798 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14800 * Change RSA blinding code so that it works when the PRNG is not
14801 seeded (in this case, the secret RSA exponent is abused as
14802 an unpredictable seed -- if it is not unpredictable, there
14803 is no point in blinding anyway). Make RSA blinding thread-safe
14804 by remembering the creator's thread ID in rsa->blinding and
14805 having all other threads use local one-time blinding factors
14806 (this requires more computation than sharing rsa->blinding, but
14807 avoids excessive locking; and if an RSA object is not shared
14808 between threads, blinding will still be very fast).
14812 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14814 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14815 via timing by performing a MAC computation even if incorrect
14816 block cipher padding has been found. This is a countermeasure
14817 against active attacks where the attacker has to distinguish
14818 between bad padding and a MAC verification error. ([CVE-2003-0078])
14820 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14821 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14822 Martin Vuagnoux (EPFL, Ilion)*
14824 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14826 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14827 memory from its contents. This is done with a counter that will
14828 place alternating values in each byte. This can be used to solve
14829 two issues: 1) the removal of calls to memset() by highly optimizing
14830 compilers, and 2) cleansing with other values than 0, since those can
14831 be read through on certain media, for example a swap space on disk.
14835 * Bugfix: client side session caching did not work with external caching,
14836 because the session->cipher setting was not restored when reloading
14837 from the external cache. This problem was masked, when
14838 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14839 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14843 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14844 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14846 *Zeev Lieber <zeev-l@yahoo.com>*
14848 * Undo an undocumented change introduced in 0.9.6e which caused
14849 repeated calls to OpenSSL_add_all_ciphers() and
14850 OpenSSL_add_all_digests() to be ignored, even after calling
14855 * Change the default configuration reader to deal with last line not
14856 being properly terminated.
14860 * Change X509_NAME_cmp() so it applies the special rules on handling
14861 DN values that are of type PrintableString, as well as RDNs of type
14862 emailAddress where the value has the type ia5String.
14864 *stefank@valicert.com via Richard Levitte*
14866 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14867 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14868 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14869 the bitwise-OR of the two for use by the majority of applications
14870 wanting this behaviour, and update the docs. The documented
14871 behaviour and actual behaviour were inconsistent and had been
14872 changing anyway, so this is more a bug-fix than a behavioural
14875 *Geoff Thorpe, diagnosed by Nadav Har'El*
14877 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14878 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14882 * Fix initialization code race conditions in
14883 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14884 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14885 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14886 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14887 ssl2_get_cipher_by_char(),
14888 ssl3_get_cipher_by_char().
14890 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14892 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14893 the cached sessions are flushed, as the remove_cb() might use ex_data
14894 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14895 (see [openssl.org #212]).
14897 *Geoff Thorpe, Lutz Jaenicke*
14899 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14900 length, instead of the encoding length to d2i_ASN1_OBJECT.
14904 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14906 * [In 0.9.6g-engine release:]
14907 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14909 *Lynn Gazis <lgazis@rainbow.com>*
14911 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14913 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14914 and get fix the header length calculation.
14915 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14916 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14918 * Use proper error handling instead of 'assertions' in buffer
14919 overflow checks added in 0.9.6e. This prevents DoS (the
14920 assertions could call abort()).
14922 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14924 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14926 * Add various sanity checks to asn1_get_length() to reject
14927 the ASN1 length bytes if they exceed sizeof(long), will appear
14928 negative or the content length exceeds the length of the
14931 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14933 * Fix cipher selection routines: ciphers without encryption had no flags
14934 for the cipher strength set and where therefore not handled correctly
14935 by the selection routines (PR #130).
14939 * Fix EVP_dsa_sha macro.
14944 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14945 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14946 that was added in OpenSSL 0.9.6d.
14948 As the countermeasure turned out to be incompatible with some
14949 broken SSL implementations, the new option is part of SSL_OP_ALL.
14950 SSL_OP_ALL is usually employed when compatibility with weird SSL
14951 implementations is desired (e.g. '-bugs' option to 's_client' and
14952 's_server'), so the new option is automatically set in many
14957 * Changes in security patch:
14959 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14960 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14961 Air Force Materiel Command, USAF, under agreement number
14964 * Add various sanity checks to asn1_get_length() to reject
14965 the ASN1 length bytes if they exceed sizeof(long), will appear
14966 negative or the content length exceeds the length of the
14967 supplied buffer. ([CVE-2002-0659])
14969 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14971 * Assertions for various potential buffer overflows, not known to
14972 happen in practice.
14974 *Ben Laurie (CHATS)*
14976 * Various temporary buffers to hold ASCII versions of integers were
14977 too small for 64 bit platforms. ([CVE-2002-0655])
14978 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14980 * Remote buffer overflow in SSL3 protocol - an attacker could
14981 supply an oversized session ID to a client. ([CVE-2002-0656])
14983 *Ben Laurie (CHATS)*
14985 * Remote buffer overflow in SSL2 protocol - an attacker could
14986 supply an oversized client master key. ([CVE-2002-0656])
14988 *Ben Laurie (CHATS)*
14990 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14992 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14993 encoded as NULL) with id-dsa-with-sha1.
14995 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14997 * Check various `X509_...()` return values in `apps/req.c`.
14999 *Nils Larsch <nla@trustcenter.de>*
15001 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15002 an end-of-file condition would erroneously be flagged, when the CRLF
15003 was just at the end of a processed block. The bug was discovered when
15004 processing data through a buffering memory BIO handing the data to a
15005 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15006 <ptsekov@syntrex.com> and Nedelcho Stanev.
15010 * Implement a countermeasure against a vulnerability recently found
15011 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15012 before application data chunks to avoid the use of known IVs
15013 with data potentially chosen by the attacker.
15017 * Fix length checks in ssl3_get_client_hello().
15021 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15022 to prevent ssl3_read_internal() from incorrectly assuming that
15023 ssl3_read_bytes() found application data while handshake
15024 processing was enabled when in fact s->s3->in_read_app_data was
15025 merely automatically cleared during the initial handshake.
15027 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15029 * Fix object definitions for Private and Enterprise: they were not
15030 recognized in their shortname (=lowercase) representation. Extend
15031 obj_dat.pl to issue an error when using undefined keywords instead
15032 of silently ignoring the problem (Svenning Sorensen
15033 <sss@sss.dnsalias.net>).
15037 * Fix DH_generate_parameters() so that it works for 'non-standard'
15038 generators, i.e. generators other than 2 and 5. (Previously, the
15039 code did not properly initialise the 'add' and 'rem' values to
15040 BN_generate_prime().)
15042 In the new general case, we do not insist that 'generator' is
15043 actually a primitive root: This requirement is rather pointless;
15044 a generator of the order-q subgroup is just as good, if not
15049 * Map new X509 verification errors to alerts. Discovered and submitted by
15050 Tom Wu <tom@arcot.com>.
15054 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15055 returning non-zero before the data has been completely received
15056 when using non-blocking I/O.
15058 *Bodo Moeller; problem pointed out by John Hughes*
15060 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15062 *Ben Laurie, Lutz Jaenicke*
15064 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15065 Yoram Zahavi <YoramZ@gilian.com>).
15069 * Add information about CygWin 1.3 and on, and preserve proper
15070 configuration for the versions before that.
15072 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15074 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15075 check whether we deal with a copy of a session and do not delete from
15076 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15077 <izhar@checkpoint.com>.
15081 * Do not store session data into the internal session cache, if it
15082 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15083 flag is set). Proposed by Aslam <aslam@funk.com>.
15087 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15092 * [In 0.9.6d-engine release:]
15093 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15095 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15097 * Add the configuration target linux-s390x.
15099 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15101 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15102 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15103 variable as an indication that a ClientHello message has been
15104 received. As the flag value will be lost between multiple
15105 invocations of ssl3_accept when using non-blocking I/O, the
15106 function may not be aware that a handshake has actually taken
15107 place, thus preventing a new session from being added to the
15110 To avoid this problem, we now set s->new_session to 2 instead of
15111 using a local variable.
15113 *Lutz Jaenicke, Bodo Moeller*
15115 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15116 if the SSL_R_LENGTH_MISMATCH error is detected.
15118 *Geoff Thorpe, Bodo Moeller*
15120 * New 'shared_ldflag' column in Configure platform table.
15124 * Fix EVP_CIPHER_mode macro.
15126 *"Dan S. Camper" <dan@bti.net>*
15128 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15129 type, we must throw them away by setting rr->length to 0.
15131 *D P Chang <dpc@qualys.com>*
15133 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15135 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15136 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15137 worked incorrectly for those cases where range = `10..._2` and
15138 `3*range` is two bits longer than range.)
15142 * Only add signing time to PKCS7 structures if it is not already
15147 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15148 OBJ_ld_ce should be OBJ_id_ce.
15149 Also some ip-pda OIDs in crypto/objects/objects.txt were
15150 incorrect (cf. RFC 3039).
15152 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15154 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15155 returns early because it has nothing to do.
15157 *Andy Schneider <andy.schneider@bjss.co.uk>*
15159 * [In 0.9.6c-engine release:]
15160 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15162 *Andy Schneider <andy.schneider@bjss.co.uk>*
15164 * [In 0.9.6c-engine release:]
15165 Add support for Cryptographic Appliance's keyserver technology.
15166 (Use engine 'keyclient')
15168 *Cryptographic Appliances and Geoff Thorpe*
15170 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15171 is called via tools/c89.sh because arguments have to be
15172 rearranged (all '-L' options must appear before the first object
15175 *Richard Shapiro <rshapiro@abinitio.com>*
15177 * [In 0.9.6c-engine release:]
15178 Add support for Broadcom crypto accelerator cards, backported
15181 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15183 * [In 0.9.6c-engine release:]
15184 Add support for SureWare crypto accelerator cards from
15185 Baltimore Technologies. (Use engine 'sureware')
15187 *Baltimore Technologies and Mark Cox*
15189 * [In 0.9.6c-engine release:]
15190 Add support for crypto accelerator cards from Accelerated
15191 Encryption Processing, www.aep.ie. (Use engine 'aep')
15193 *AEP Inc. and Mark Cox*
15195 * Add a configuration entry for gcc on UnixWare.
15197 *Gary Benson <gbenson@redhat.com>*
15199 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15200 messages are stored in a single piece (fixed-length part and
15201 variable-length part combined) and fix various bugs found on the way.
15205 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15206 instead. BIO_gethostbyname() does not know what timeouts are
15207 appropriate, so entries would stay in cache even when they have
15209 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15211 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15212 faced with a pathologically small ClientHello fragment that does
15213 not contain client_version: Instead of aborting with an error,
15214 simply choose the highest available protocol version (i.e.,
15215 TLS 1.0 unless it is disabled). In practice, ClientHello
15216 messages are never sent like this, but this change gives us
15217 strictly correct behaviour at least for TLS.
15221 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15222 never resets s->method to s->ctx->method when called from within
15223 one of the SSL handshake functions.
15225 *Bodo Moeller; problem pointed out by Niko Baric*
15227 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15228 (sent using the client's version number) if client_version is
15229 smaller than the protocol version in use. Also change
15230 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15231 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15232 the client will at least see that alert.
15236 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15241 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15242 client receives HelloRequest while in a handshake.
15244 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15246 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15247 should end in 'break', not 'goto end' which circumvents various
15248 cleanups done in state SSL_ST_OK. But session related stuff
15249 must be disabled for SSL_ST_OK in the case that we just sent a
15252 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15253 before just sending a HelloRequest.
15255 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15257 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15258 reveal whether illegal block cipher padding was found or a MAC
15259 verification error occurred. (Neither SSLerr() codes nor alerts
15260 are directly visible to potential attackers, but the information
15261 may leak via logfiles.)
15263 Similar changes are not required for the SSL 2.0 implementation
15264 because the number of padding bytes is sent in clear for SSL 2.0,
15265 and the extra bytes are just ignored. However ssl/s2_pkt.c
15266 failed to verify that the purported number of padding bytes is in
15271 * Add OpenUNIX-8 support including shared libraries
15272 (Boyd Lynn Gerber <gerberb@zenez.com>).
15276 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15277 'wristwatch attack' using huge encoding parameters (cf.
15278 James H. Manger's CRYPTO 2001 paper). Note that the
15279 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15280 encoding parameters and hence was not vulnerable.
15284 * BN_sqr() bug fix.
15286 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15288 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15289 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15290 followed by modular reduction.
15292 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15294 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15295 equivalent based on BN_pseudo_rand() instead of BN_rand().
15299 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15300 This function was broken, as the check for a new client hello message
15301 to handle SGC did not allow these large messages.
15302 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15306 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15310 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15311 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15315 * Rework the configuration and shared library support for Tru64 Unix.
15316 The configuration part makes use of modern compiler features and
15317 still retains old compiler behavior for those that run older versions
15318 of the OS. The shared library support part includes a variant that
15319 uses the RPATH feature, and is available through the special
15320 configuration target "alpha-cc-rpath", which will never be selected
15323 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15325 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15326 with the same message size as in ssl3_get_certificate_request().
15327 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15328 messages might inadvertently be reject as too long.
15330 *Petr Lampa <lampa@fee.vutbr.cz>*
15332 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15336 * Modified SSL library such that the verify_callback that has been set
15337 specifically for an SSL object with SSL_set_verify() is actually being
15338 used. Before the change, a verify_callback set with this function was
15339 ignored and the verify_callback() set in the SSL_CTX at the time of
15340 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15341 to allow the necessary settings.
15345 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15346 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15347 done automatically (in contradiction to the requirements of the C
15348 standard). This made problems when used from OpenSSH.
15352 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15353 dh->length and always used
15355 BN_rand_range(priv_key, dh->p).
15357 BN_rand_range() is not necessary for Diffie-Hellman, and this
15358 specific range makes Diffie-Hellman unnecessarily inefficient if
15359 dh->length (recommended exponent length) is much smaller than the
15360 length of dh->p. We could use BN_rand_range() if the order of
15361 the subgroup was stored in the DH structure, but we only have
15366 BN_rand(priv_key, l, ...)
15368 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15375 RSA_eay_public_encrypt
15376 RSA_eay_private_decrypt
15377 RSA_eay_private_encrypt (signing)
15378 RSA_eay_public_decrypt (signature verification)
15380 (default implementations for RSA_public_encrypt,
15381 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15382 always reject numbers >= n.
15386 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15387 to synchronize access to 'locking_thread'. This is necessary on
15388 systems where access to 'locking_thread' (an 'unsigned long'
15389 variable) is not atomic.
15393 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15394 *before* setting the 'crypto_lock_rand' flag. The previous code had
15395 a race condition if 0 is a valid thread ID.
15397 *Travis Vitek <vitek@roguewave.com>*
15399 * Add support for shared libraries under Irix.
15401 *Albert Chin-A-Young <china@thewrittenword.com>*
15403 * Add configuration option to build on Linux on both big-endian and
15404 little-endian MIPS.
15406 *Ralf Baechle <ralf@uni-koblenz.de>*
15408 * Add the possibility to create shared libraries on HP-UX.
15412 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15414 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15415 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15416 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15417 PRNG state recovery was possible based on the output of
15418 one PRNG request appropriately sized to gain knowledge on
15419 'md' followed by enough consecutive 1-byte PRNG requests
15420 to traverse all of 'state'.
15422 1. When updating 'md_local' (the current thread's copy of 'md')
15423 during PRNG output generation, hash all of the previous
15424 'md_local' value, not just the half used for PRNG output.
15426 2. Make the number of bytes from 'state' included into the hash
15427 independent from the number of PRNG bytes requested.
15429 The first measure alone would be sufficient to avoid
15430 Markku-Juhani's attack. (Actually it had never occurred
15431 to me that the half of 'md_local' used for chaining was the
15432 half from which PRNG output bytes were taken -- I had always
15433 assumed that the secret half would be used.) The second
15434 measure makes sure that additional data from 'state' is never
15435 mixed into 'md_local' in small portions; this heuristically
15436 further strengthens the PRNG.
15440 * Fix crypto/bn/asm/mips3.s.
15444 * When only the key is given to "enc", the IV is undefined. Print out
15445 an error message in this case.
15449 * Handle special case when X509_NAME is empty in X509 printing routines.
15453 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15454 positive and less than q.
15458 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15459 used: it isn't thread safe and the add_lock_callback should handle
15462 *Paul Rose <Paul.Rose@bridge.com>*
15464 * Verify that incoming data obeys the block size in
15465 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15471 *Ulf Möller, Bodo Möller*
15473 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15474 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15475 when fixing the server behaviour for backwards-compatible 'client
15476 hello' messages. (Note that the attack is impractical against
15477 SSL 3.0 and TLS 1.0 anyway because length and version checking
15478 means that the probability of guessing a valid ciphertext is
15479 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15482 Before 0.9.5, the countermeasure (hide the error by generating a
15483 random 'decryption result') did not work properly because
15484 ERR_clear_error() was missing, meaning that SSL_get_error() would
15485 detect the supposedly ignored error.
15487 Both problems are now fixed.
15491 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15492 (previously it was 1024).
15496 * Fix for compatibility mode trust settings: ignore trust settings
15497 unless some valid trust or reject settings are present.
15501 * Fix for blowfish EVP: its a variable length cipher.
15505 * Fix various bugs related to DSA S/MIME verification. Handle missing
15506 parameters in DSA public key structures and return an error in the
15507 DSA routines if parameters are absent.
15511 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15512 in the current directory if neither $RANDFILE nor $HOME was set.
15513 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15514 caused some confusion to Windows users who haven't defined $HOME.
15515 Thus RAND_file_name() is changed again: e_os.h can define a
15516 DEFAULT_HOME, which will be used if $HOME is not set.
15517 For Windows, we use "C:"; on other platforms, we still require
15518 environment variables.
15520 * Move 'if (!initialized) RAND_poll()' into regions protected by
15521 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15522 having multiple threads call RAND_poll() concurrently.
15526 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15527 combination of a flag and a thread ID variable.
15528 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15529 flag), *other* threads can enter ssleay_add_bytes without obeying
15530 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15531 that they do not hold after the first thread unsets add_do_not_lock).
15535 * Change bctest again: '-x' expressions are not available in all
15536 versions of 'test'.
15540 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15542 * Fix a couple of memory leaks in PKCS7_dataDecode()
15544 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15546 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15547 the default extension for executables, if any. Also, make the perl
15548 scripts that use symlink() to test if it really exists and use "cp"
15549 if it doesn't. All this made OpenSSL compilable and installable in
15554 * Fix for asn1_GetSequence() for indefinite length constructed data.
15555 If SEQUENCE is length is indefinite just set c->slen to the total
15556 amount of data available.
15558 *Steve Henson, reported by shige@FreeBSD.org*
15560 *This change does not apply to 0.9.7.*
15562 * Change bctest to avoid here-documents inside command substitution
15563 (workaround for FreeBSD /bin/sh bug).
15564 For compatibility with Ultrix, avoid shell functions (introduced
15565 in the bctest version that searches along $PATH).
15569 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15570 with des_encrypt() defined on some operating systems, like Solaris
15575 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15576 On the Importance of Eliminating Errors in Cryptographic
15577 Computations, J. Cryptology 14 (2001) 2, 101-119,
15578 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15582 * MIPS assembler BIGNUM division bug fix.
15586 * Disabled incorrect Alpha assembler code.
15590 * Fix PKCS#7 decode routines so they correctly update the length
15591 after reading an EOC for the EXPLICIT tag.
15595 *This change does not apply to 0.9.7.*
15597 * Fix bug in PKCS#12 key generation routines. This was triggered
15598 if a 3DES key was generated with a 0 initial byte. Include
15599 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15600 (but broken) behaviour.
15604 * Enhance bctest to search for a working bc along $PATH and print
15607 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15609 * Fix memory leaks in err.c: free err_data string if necessary;
15610 don't write to the wrong index in ERR_set_error_data.
15614 * Implement ssl23_peek (analogous to ssl23_read), which previously
15619 * Replace rdtsc with `_emit` statements for VC++ version 5.
15621 *Jeremy Cooper <jeremy@baymoo.org>*
15623 * Make it possible to reuse SSLv2 sessions.
15627 * In copy_email() check for >= 0 as a return value for
15628 X509_NAME_get_index_by_NID() since 0 is a valid index.
15630 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15632 * Avoid coredump with unsupported or invalid public keys by checking if
15633 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15634 PKCS7_verify() fails with non detached data.
15638 * Don't use getenv in library functions when run as setuid/setgid.
15639 New function OPENSSL_issetugid().
15643 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15644 due to incorrect handling of multi-threading:
15646 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15648 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15650 3. Count how many times MemCheck_off() has been called so that
15651 nested use can be treated correctly. This also avoids
15652 inband-signalling in the previous code (which relied on the
15653 assumption that thread ID 0 is impossible).
15657 * Add "-rand" option also to s_client and s_server.
15661 * Fix CPU detection on Irix 6.x.
15662 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15663 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15665 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15670 *This change does not apply to 0.9.7.*
15672 * Use the cached encoding of an X509_NAME structure rather than
15673 copying it. This is apparently the reason for the libsafe "errors"
15674 but the code is actually correct.
15678 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15679 Bleichenbacher's DSA attack.
15680 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15681 to be set and top=0 forces the highest bit to be set; top=-1 is new
15682 and leaves the highest bit random.
15684 *Ulf Moeller, Bodo Moeller*
15686 * In the `NCONF_...`-based implementations for `CONF_...` queries
15687 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15688 a temporary CONF structure with the data component set to NULL
15689 (which gives segmentation faults in lh_retrieve).
15690 Instead, use NULL for the CONF pointer in CONF_get_string and
15691 CONF_get_number (which may use environment variables) and directly
15692 return NULL from CONF_get_section.
15696 * Fix potential buffer overrun for EBCDIC.
15700 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15701 keyUsage if basicConstraints absent for a CA.
15705 * Make SMIME_write_PKCS7() write mail header values with a format that
15706 is more generally accepted (no spaces before the semicolon), since
15707 some programs can't parse those values properly otherwise. Also make
15708 sure BIO's that break lines after each write do not create invalid
15713 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15714 macros previously used would not encode an empty SEQUENCE OF
15715 and break the signature.
15719 *This change does not apply to 0.9.7.*
15721 * Zero the premaster secret after deriving the master secret in
15726 * Add some EVP_add_digest_alias registrations (as found in
15727 OpenSSL_add_all_digests()) to SSL_library_init()
15728 aka OpenSSL_add_ssl_algorithms(). This provides improved
15729 compatibility with peers using X.509 certificates
15730 with unconventional AlgorithmIdentifier OIDs.
15734 * Fix for Irix with NO_ASM.
15736 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15738 * ./config script fixes.
15740 *Ulf Moeller, Richard Levitte*
15742 * Fix 'openssl passwd -1'.
15746 * Change PKCS12_key_gen_asc() so it can cope with non null
15747 terminated strings whose length is passed in the passlen
15748 parameter, for example from PEM callbacks. This was done
15749 by adding an extra length parameter to asc2uni().
15751 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15753 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15754 call failed, free the DSA structure.
15758 * Fix to uni2asc() to cope with zero length Unicode strings.
15759 These are present in some PKCS#12 files.
15763 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15764 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15765 when writing a 32767 byte record.
15767 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15769 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15770 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15772 (RSA objects have a reference count access to which is protected
15773 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15774 so they are meant to be shared between threads.)
15775 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15776 "Reddie, Steven" <Steven.Reddie@ca.com>*
15778 * Fix a deadlock in CRYPTO_mem_leaks().
15782 * Use better test patterns in bntest.
15786 * rand_win.c fix for Borland C.
15790 * BN_rshift bugfix for n == 0.
15794 * Add a 'bctest' script that checks for some known 'bc' bugs
15795 so that 'make test' does not abort just because 'bc' is broken.
15799 * Store verify_result within SSL_SESSION also for client side to
15800 avoid potential security hole. (Re-used sessions on the client side
15801 always resulted in verify_result==X509_V_OK, not using the original
15802 result of the server certificate verification.)
15806 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15807 SSL3_RT_APPLICATION_DATA, return 0.
15808 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15813 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15814 releases, have been re-implemented by renaming the previous
15815 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15816 and ssl3_read_internal, respectively, and adding 'peek' parameters
15817 to them. The new ssl[23]_{read,peek} functions are calls to
15818 ssl[23]_read_internal with the 'peek' flag set appropriately.
15819 A 'peek' parameter has also been added to ssl3_read_bytes, which
15820 does the actual work for ssl3_read_internal.
15824 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15825 the method-specific "init()" handler. Also clean up ex_data after
15826 calling the method-specific "finish()" handler. Previously, this was
15827 happening the other way round.
15831 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15832 The previous value, 12, was not always sufficient for BN_mod_exp().
15836 * Make sure that shared libraries get the internal name engine with
15837 the full version number and not just 0. This should mark the
15838 shared libraries as not backward compatible. Of course, this should
15839 be changed again when we can guarantee backward binary compatibility.
15843 * Fix typo in get_cert_by_subject() in by_dir.c
15845 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15847 * Rework the system to generate shared libraries:
15849 - Make note of the expected extension for the shared libraries and
15850 if there is a need for symbolic links from for example libcrypto.so.0
15851 to libcrypto.so.0.9.7. There is extended info in Configure for
15854 - Make as few rebuilds of the shared libraries as possible.
15856 - Still avoid linking the OpenSSL programs with the shared libraries.
15858 - When installing, install the shared libraries separately from the
15863 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15865 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15866 and not in SSL_clear because the latter is also used by the
15867 accept/connect functions; previously, the settings made by
15868 SSL_set_read_ahead would be lost during the handshake.
15870 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15872 * Correct util/mkdef.pl to be selective about disabled algorithms.
15873 Previously, it would create entries for disabled algorithms no
15878 * Added several new manual pages for SSL_* function.
15882 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15884 * In ssl23_get_client_hello, generate an error message when faced
15885 with an initial SSL 3.0/TLS record that is too small to contain the
15886 first two bytes of the ClientHello message, i.e. client_version.
15887 (Note that this is a pathologic case that probably has never happened
15888 in real life.) The previous approach was to use the version number
15889 from the record header as a substitute; but our protocol choice
15890 should not depend on that one because it is not authenticated
15891 by the Finished messages.
15895 * More robust randomness gathering functions for Windows.
15897 *Jeffrey Altman <jaltman@columbia.edu>*
15899 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15900 not set then we don't setup the error code for issuer check errors
15901 to avoid possibly overwriting other errors which the callback does
15902 handle. If an application does set the flag then we assume it knows
15903 what it is doing and can handle the new informational codes
15908 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15909 a general "ANY" type, as such it should be able to decode anything
15910 including tagged types. However it didn't check the class so it would
15911 wrongly interpret tagged types in the same way as their universal
15912 counterpart and unknown types were just rejected. Changed so that the
15913 tagged and unknown types are handled in the same way as a SEQUENCE:
15914 that is the encoding is stored intact. There is also a new type
15915 "V_ASN1_OTHER" which is used when the class is not universal, in this
15916 case we have no idea what the actual type is so we just lump them all
15921 * On VMS, stdout may very well lead to a file that is written to
15922 in a record-oriented fashion. That means that every write() will
15923 write a separate record, which will be read separately by the
15924 programs trying to read from it. This can be very confusing.
15926 The solution is to put a BIO filter in the way that will buffer
15927 text until a linefeed is reached, and then write everything a
15928 line at a time, so every record written will be an actual line,
15929 not chunks of lines and not (usually doesn't happen, but I've
15930 seen it once) several lines in one record. BIO_f_linebuffer() is
15933 Currently, it's a VMS-only method, because that's where it has
15934 been tested well enough.
15938 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15939 it can return incorrect results.
15940 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15941 but it was in 0.9.6-beta[12].)
15945 * Disable the check for content being present when verifying detached
15946 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15947 include zero length content when signing messages.
15951 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15952 BIO_ctrl (for BIO pairs).
15956 * Add DSO method for VMS.
15960 * Bug fix: Montgomery multiplication could produce results with the
15965 * Add RPM specification openssl.spec and modify it to build three
15966 packages. The default package contains applications, application
15967 documentation and run-time libraries. The devel package contains
15968 include files, static libraries and function documentation. The
15969 doc package contains the contents of the doc directory. The original
15970 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15974 * Add a large number of documentation files for many SSL routines.
15976 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15978 * Add a configuration entry for Sony News 4.
15980 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15982 * Don't set the two most significant bits to one when generating a
15983 random number < q in the DSA library.
15987 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15988 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15989 the underlying transport is blocking) if a handshake took place.
15990 (The default behaviour is needed by applications such as s_client
15991 and s_server that use select() to determine when to use SSL_read;
15992 but for applications that know in advance when to expect data, it
15993 just makes things more complicated.)
15997 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16002 * Add a few more EBCDIC conditionals that make `req` and `x509`
16003 work better on such systems.
16005 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16007 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16008 Update PKCS12_parse() so it copies the friendlyName and the
16009 keyid to the certificates aux info.
16013 * Fix bug in PKCS7_verify() which caused an infinite loop
16014 if there was more than one signature.
16016 *Sven Uszpelkat <su@celocom.de>*
16018 * Major change in util/mkdef.pl to include extra information
16019 about each symbol, as well as presenting variables as well
16020 as functions. This change means that there's n more need
16021 to rebuild the .num files when some algorithms are excluded.
16025 * Allow the verify time to be set by an application,
16026 rather than always using the current time.
16030 * Phase 2 verify code reorganisation. The certificate
16031 verify code now looks up an issuer certificate by a
16032 number of criteria: subject name, authority key id
16033 and key usage. It also verifies self signed certificates
16034 by the same criteria. The main comparison function is
16035 X509_check_issued() which performs these checks.
16037 Lot of changes were necessary in order to support this
16038 without completely rewriting the lookup code.
16040 Authority and subject key identifier are now cached.
16042 The LHASH 'certs' is X509_STORE has now been replaced
16043 by a STACK_OF(X509_OBJECT). This is mainly because an
16044 LHASH can't store or retrieve multiple objects with
16045 the same hash value.
16047 As a result various functions (which were all internal
16048 use only) have changed to handle the new X509_STORE
16049 structure. This will break anything that messed round
16050 with X509_STORE internally.
16052 The functions X509_STORE_add_cert() now checks for an
16053 exact match, rather than just subject name.
16055 The X509_STORE API doesn't directly support the retrieval
16056 of multiple certificates matching a given criteria, however
16057 this can be worked round by performing a lookup first
16058 (which will fill the cache with candidate certificates)
16059 and then examining the cache for matches. This is probably
16060 the best we can do without throwing out X509_LOOKUP
16061 entirely (maybe later...).
16063 The X509_VERIFY_CTX structure has been enhanced considerably.
16065 All certificate lookup operations now go via a get_issuer()
16066 callback. Although this currently uses an X509_STORE it
16067 can be replaced by custom lookups. This is a simple way
16068 to bypass the X509_STORE hackery necessary to make this
16069 work and makes it possible to use more efficient techniques
16070 in future. A very simple version which uses a simple
16071 STACK for its trusted certificate store is also provided
16072 using X509_STORE_CTX_trusted_stack().
16074 The verify_cb() and verify() callbacks now have equivalents
16075 in the X509_STORE_CTX structure.
16077 X509_STORE_CTX also has a 'flags' field which can be used
16078 to customise the verify behaviour.
16082 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16083 excludes S/MIME capabilities.
16087 * When a certificate request is read in keep a copy of the
16088 original encoding of the signed data and use it when outputting
16089 again. Signatures then use the original encoding rather than
16090 a decoded, encoded version which may cause problems if the
16091 request is improperly encoded.
16095 * For consistency with other BIO_puts implementations, call
16096 buffer_write(b, ...) directly in buffer_puts instead of calling
16099 In BIO_puts, increment b->num_write as in BIO_write.
16101 *Peter.Sylvester@EdelWeb.fr*
16103 * Fix BN_mul_word for the case where the word is 0. (We have to use
16104 BN_zero, we may not return a BIGNUM with an array consisting of
16105 words set to zero.)
16109 * Avoid calling abort() from within the library when problems are
16110 detected, except if preprocessor symbols have been defined
16111 (such as REF_CHECK, BN_DEBUG etc.).
16115 * New openssl application 'rsautl'. This utility can be
16116 used for low-level RSA operations. DER public key
16117 BIO/fp routines also added.
16121 * New Configure entry and patches for compiling on QNX 4.
16123 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16125 * A demo state-machine implementation was sponsored by
16126 Nuron (<http://www.nuron.com/>) and is now available in
16127 demos/state_machine.
16131 * New options added to the 'dgst' utility for signature
16132 generation and verification.
16136 * Unrecognized PKCS#7 content types are now handled via a
16137 catch all ASN1_TYPE structure. This allows unsupported
16138 types to be stored as a "blob" and an application can
16139 encode and decode it manually.
16143 * Fix various signed/unsigned issues to make a_strex.c
16144 compile under VC++.
16146 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16148 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16149 length if passed a buffer. ASN1_INTEGER_to_BN failed
16150 if passed a NULL BN and its argument was negative.
16152 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16154 * Modification to PKCS#7 encoding routines to output definite
16155 length encoding. Since currently the whole structures are in
16156 memory there's not real point in using indefinite length
16157 constructed encoding. However if OpenSSL is compiled with
16158 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16162 * Added BIO_vprintf() and BIO_vsnprintf().
16166 * Added more prefixes to parse for in the strings written
16167 through a logging bio, to cover all the levels that are available
16168 through syslog. The prefixes are now:
16170 PANIC, EMERG, EMR => LOG_EMERG
16171 ALERT, ALR => LOG_ALERT
16172 CRIT, CRI => LOG_CRIT
16173 ERROR, ERR => LOG_ERR
16174 WARNING, WARN, WAR => LOG_WARNING
16175 NOTICE, NOTE, NOT => LOG_NOTICE
16176 INFO, INF => LOG_INFO
16177 DEBUG, DBG => LOG_DEBUG
16179 and as before, if none of those prefixes are present at the
16180 beginning of the string, LOG_ERR is chosen.
16182 On Win32, the `LOG_*` levels are mapped according to this:
16184 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16185 LOG_WARNING => EVENTLOG_WARNING_TYPE
16186 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16190 * Made it possible to reconfigure with just the configuration
16191 argument "reconf" or "reconfigure". The command line arguments
16192 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16193 and are retrieved from there when reconfiguring.
16199 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16201 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16205 * The obj_dat.pl script was messing up the sorting of object
16206 names. The reason was that it compared the quoted version
16207 of strings as a result "OCSP" > "OCSP Signing" because
16208 " > SPACE. Changed script to store unquoted versions of
16209 names and add quotes on output. It was also omitting some
16210 names from the lookup table if they were given a default
16211 value (that is if SN is missing it is given the same
16212 value as LN and vice versa), these are now added on the
16213 grounds that if an object has a name we should be able to
16214 look it up. Finally added warning output when duplicate
16215 short or long names are found.
16219 * Changes needed for Tandem NSK.
16221 *Scott Uroff <scott@xypro.com>*
16223 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16224 RSA_padding_check_SSLv23(), special padding was never detected
16225 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16226 version rollback attacks was not effective.
16228 In s23_clnt.c, don't use special rollback-attack detection padding
16229 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16230 client; similarly, in s23_srvr.c, don't do the rollback check if
16231 SSL 2.0 is the only protocol enabled in the server.
16235 * Make it possible to get hexdumps of unprintable data with 'openssl
16236 asn1parse'. By implication, the functions ASN1_parse_dump() and
16237 BIO_dump_indent() are added.
16241 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16242 these print out strings and name structures based on various
16243 flags including RFC2253 support and proper handling of
16244 multibyte characters. Added options to the 'x509' utility
16245 to allow the various flags to be set.
16249 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16250 Also change the functions X509_cmp_current_time() and
16251 X509_gmtime_adj() work with an ASN1_TIME structure,
16252 this will enable certificates using GeneralizedTime in validity
16253 dates to be checked.
16257 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16258 negative public key encodings) on by default,
16259 NO_NEG_PUBKEY_BUG can be set to disable it.
16263 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16264 content octets. An i2c_ASN1_OBJECT is unnecessary because
16265 the encoding can be trivially obtained from the structure.
16269 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16270 not read locks (`CRYPTO_r_[un]lock`).
16274 * A first attempt at creating official support for shared
16275 libraries through configuration. I've kept it so the
16276 default is static libraries only, and the OpenSSL programs
16277 are always statically linked for now, but there are
16278 preparations for dynamic linking in place.
16279 This has been tested on Linux and Tru64.
16283 * Randomness polling function for Win9x, as described in:
16284 Peter Gutmann, Software Generation of Practically Strong
16289 * Fix so PRNG is seeded in req if using an already existing
16294 * New options to smime application. -inform and -outform
16295 allow alternative formats for the S/MIME message including
16296 PEM and DER. The -content option allows the content to be
16297 specified separately. This should allow things like Netscape
16298 form signing output easier to verify.
16302 * Fix the ASN1 encoding of tags using the 'long form'.
16306 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16307 STRING types. These convert content octets to and from the
16308 underlying type. The actual tag and length octets are
16309 already assumed to have been read in and checked. These
16310 are needed because all other string types have virtually
16311 identical handling apart from the tag. By having versions
16312 of the ASN1 functions that just operate on content octets
16313 IMPLICIT tagging can be handled properly. It also allows
16314 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16315 and ASN1_INTEGER are identical apart from the tag.
16319 * Change the handling of OID objects as follows:
16321 - New object identifiers are inserted in objects.txt, following
16322 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16323 - objects.pl is used to process obj_mac.num and create a new
16325 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16328 This is currently kind of a hack, and the perl code in objects.pl
16329 isn't very elegant, but it works as I intended. The simplest way
16330 to check that it worked correctly is to look in obj_dat.h and
16331 check the array nid_objs and make sure the objects haven't moved
16332 around (this is important!). Additions are OK, as well as
16333 consistent name changes.
16337 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16341 * Addition of the command line parameter '-rand file' to 'openssl req'.
16342 The given file adds to whatever has already been seeded into the
16343 random pool through the RANDFILE configuration file option or
16344 environment variable, or the default random state file.
16348 * mkstack.pl now sorts each macro group into lexical order.
16349 Previously the output order depended on the order the files
16350 appeared in the directory, resulting in needless rewriting
16355 * Patches to make OpenSSL compile under Win32 again. Mostly
16356 work arounds for the VC++ problem that it treats func() as
16357 func(void). Also stripped out the parts of mkdef.pl that
16358 added extra typesafe functions: these no longer exist.
16362 * Reorganisation of the stack code. The macros are now all
16363 collected in safestack.h . Each macro is defined in terms of
16364 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16365 DEBUG_SAFESTACK is now handled in terms of function casts,
16366 this has the advantage of retaining type safety without the
16367 use of additional functions. If DEBUG_SAFESTACK is not defined
16368 then the non typesafe macros are used instead. Also modified the
16369 mkstack.pl script to handle the new form. Needs testing to see
16370 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16371 the default if no major problems. Similar behaviour for ASN1_SET_OF
16372 and PKCS12_STACK_OF.
16376 * When some versions of IIS use the 'NET' form of private key the
16377 key derivation algorithm is different. Normally MD5(password) is
16378 used as a 128 bit RC4 key. In the modified case
16379 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16380 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16381 as the old Netscape_RSA functions except they have an additional
16382 'sgckey' parameter which uses the modified algorithm. Also added
16383 an -sgckey command line option to the rsa utility. Thanks to
16384 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16385 algorithm to openssl-dev.
16389 * The evp_local.h macros were using 'c.##kname' which resulted in
16390 invalid expansion on some systems (SCO 5.0.5 for example).
16391 Corrected to 'c.kname'.
16393 *Phillip Porch <root@theporch.com>*
16395 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16396 a STACK of email addresses from a certificate or request, these look
16397 in the subject name and the subject alternative name extensions and
16398 omit any duplicate addresses.
16402 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16403 This makes DSA verification about 2 % faster.
16407 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16408 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16409 plus overhead for 1024 bit moduli).
16410 This makes exponentiations about 0.5 % faster for 1024 bit
16411 exponents (as measured by "openssl speed rsa2048").
16415 * Rename memory handling macros to avoid conflicts with other
16417 Malloc => OPENSSL_malloc
16418 Malloc_locked => OPENSSL_malloc_locked
16419 Realloc => OPENSSL_realloc
16420 Free => OPENSSL_free
16424 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16425 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16429 * CygWin32 support.
16431 *John Jarvie <jjarvie@newsguy.com>*
16433 * The type-safe stack code has been rejigged. It is now only compiled
16434 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16435 by default all type-specific stack functions are "#define"d back to
16436 standard stack functions. This results in more streamlined output
16437 but retains the type-safety checking possibilities of the original
16442 * The STACK code has been cleaned up, and certain type declarations
16443 that didn't make a lot of sense have been brought in line. This has
16444 also involved a cleanup of sorts in safestack.h to more correctly
16445 map type-safe stack functions onto their plain stack counterparts.
16446 This work has also resulted in a variety of "const"ifications of
16447 lots of the code, especially `_cmp` operations which should normally
16448 be prototyped with "const" parameters anyway.
16452 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16453 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16454 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16455 where all of 'md' is used each time the PRNG is used, but 'state'
16456 is used only indexed by a cyclic counter. As entropy may not be
16457 well distributed from the beginning, 'md' is important as a
16458 chaining variable. However, the output function chains only half
16459 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16460 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16461 in all of 'state' being rewritten, with the new values depending
16462 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16466 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16467 the handshake is continued after ssl_verify_cert_chain();
16468 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16469 can lead to 'unexplainable' connection aborts later.
16471 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16473 * Major EVP API cipher revision.
16474 Add hooks for extra EVP features. This allows various cipher
16475 parameters to be set in the EVP interface. Support added for variable
16476 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16477 setting of RC2 and RC5 parameters.
16479 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16482 Remove lots of duplicated code from the EVP library. For example *every*
16483 cipher init() function handles the 'iv' in the same way according to the
16484 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16485 for CFB and OFB modes they zero ctx->num.
16487 New functionality allows removal of S/MIME code RC2 hack.
16489 Most of the routines have the same form and so can be declared in terms
16492 By shifting this to the top level EVP_CipherInit() it can be removed from
16493 all individual ciphers. If the cipher wants to handle IVs or keys
16494 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16497 Change lots of functions like EVP_EncryptUpdate() to now return a
16498 value: although software versions of the algorithms cannot fail
16499 any installed hardware versions can.
16503 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16504 this option is set, tolerate broken clients that send the negotiated
16505 protocol version number instead of the requested protocol version
16510 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16511 i.e. non-zero for export ciphersuites, zero otherwise.
16512 Previous versions had this flag inverted, inconsistent with
16513 rsa_tmp_cb (..._TMP_RSA_CB).
16515 *Bodo Moeller; problem reported by Amit Chopra*
16517 * Add missing DSA library text string. Work around for some IIS
16518 key files with invalid SEQUENCE encoding.
16522 * Add a document (doc/standards.txt) that list all kinds of standards
16523 and so on that are implemented in OpenSSL.
16527 * Enhance c_rehash script. Old version would mishandle certificates
16528 with the same subject name hash and wouldn't handle CRLs at all.
16529 Added -fingerprint option to crl utility, to support new c_rehash
16534 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16538 * Fix for SSL server purpose checking. Server checking was
16539 rejecting certificates which had extended key usage present
16540 but no ssl client purpose.
16542 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16544 * Make PKCS#12 code work with no password. The PKCS#12 spec
16545 is a little unclear about how a blank password is handled.
16546 Since the password in encoded as a BMPString with terminating
16547 double NULL a zero length password would end up as just the
16548 double NULL. However no password at all is different and is
16549 handled differently in the PKCS#12 key generation code. NS
16550 treats a blank password as zero length. MSIE treats it as no
16551 password on export: but it will try both on import. We now do
16552 the same: PKCS12_parse() tries zero length and no password if
16553 the password is set to "" or NULL (NULL is now a valid password:
16554 it wasn't before) as does the pkcs12 application.
16558 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16559 perror when PEM_read_bio_X509_REQ fails, the error message must
16560 be obtained from the error queue.
16564 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16565 it in ERR_remove_state if appropriate, and change ERR_get_state
16566 accordingly to avoid race conditions (this is necessary because
16567 thread_hash is no longer constant once set).
16571 * Bugfix for linux-elf makefile.one.
16575 * RSA_get_default_method() will now cause a default
16576 RSA_METHOD to be chosen if one doesn't exist already.
16577 Previously this was only set during a call to RSA_new()
16578 or RSA_new_method(NULL) meaning it was possible for
16579 RSA_get_default_method() to return NULL.
16583 * Added native name translation to the existing DSO code
16584 that will convert (if the flag to do so is set) filenames
16585 that are sufficiently small and have no path information
16586 into a canonical native form. Eg. "blah" converted to
16587 "libblah.so" or "blah.dll" etc.
16591 * New function ERR_error_string_n(e, buf, len) which is like
16592 ERR_error_string(e, buf), but writes at most 'len' bytes
16593 including the 0 terminator. For ERR_error_string_n, 'buf'
16596 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16598 * CONF library reworked to become more general. A new CONF
16599 configuration file reader "class" is implemented as well as a
16600 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16601 old `CONF_*` functions are still there, but are reimplemented to
16602 work in terms of the new functions. Also, a set of functions
16603 to handle the internal storage of the configuration data is
16604 provided to make it easier to write new configuration file
16605 reader "classes" (I can definitely see something reading a
16606 configuration file in XML format, for example), called `_CONF_*`,
16607 or "the configuration storage API"...
16609 The new configuration file reading functions are:
16611 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16612 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16614 NCONF_default, NCONF_WIN32
16616 NCONF_dump_fp, NCONF_dump_bio
16618 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16619 NCONF_new creates a new CONF object. This works in the same way
16620 as other interfaces in OpenSSL, like the BIO interface.
16621 `NCONF_dump_*` dump the internal storage of the configuration file,
16622 which is useful for debugging. All other functions take the same
16623 arguments as the old `CONF_*` functions with the exception of the
16624 first that must be a `CONF *` instead of a `LHASH *`.
16626 To make it easier to use the new classes with the old `CONF_*` functions,
16627 the function CONF_set_default_method is provided.
16631 * Add '-tls1' option to 'openssl ciphers', which was already
16632 mentioned in the documentation but had not been implemented.
16633 (This option is not yet really useful because even the additional
16634 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16638 * Initial DSO code added into libcrypto for letting OpenSSL (and
16639 OpenSSL-based applications) load shared libraries and bind to
16640 them in a portable way.
16642 *Geoff Thorpe, with contributions from Richard Levitte*
16644 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16646 * Make sure _lrotl and _lrotr are only used with MSVC.
16648 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16649 (the default implementation of RAND_status).
16651 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16652 to '-clrext' (= clear extensions), as intended and documented.
16653 *Bodo Moeller; inconsistency pointed out by Michael Attili
16654 <attili@amaxo.com>*
16656 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16657 was larger than the MD block size.
16659 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16661 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16662 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16663 using the passed key: if the passed key was a private key the result
16664 of X509_print(), for example, would be to print out all the private key
16669 * des_quad_cksum() byte order bug fix.
16670 *Ulf Möller, using the problem description in krb4-0.9.7, where
16671 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16673 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16676 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16678 * For easily testing in shell scripts whether some command
16679 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16680 returns with exit code 0 iff no command of the given name is available.
16681 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16682 the output goes to stdout and nothing is printed to stderr.
16683 Additional arguments are always ignored.
16685 Since for each cipher there is a command of the same name,
16686 the 'no-cipher' compilation switches can be tested this way.
16688 ('openssl no-XXX' is not able to detect pseudo-commands such
16689 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16693 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16697 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16698 is set; it will be thrown away anyway because each handshake creates
16700 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16701 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16702 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16703 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16707 * New s_client option -ign_eof: EOF at stdin is ignored, and
16708 'Q' and 'R' lose their special meanings (quit/renegotiate).
16709 This is part of what -quiet does; unlike -quiet, -ign_eof
16710 does not suppress any output.
16714 * Add compatibility options to the purpose and trust code. The
16715 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16716 accepts a certificate or CA, this was the previous behaviour,
16717 with all the associated security issues.
16719 X509_TRUST_COMPAT is the old trust behaviour: only and
16720 automatically trust self signed roots in certificate store. A
16721 new trust setting X509_TRUST_DEFAULT is used to specify that
16722 a purpose has no associated trust setting and it should instead
16723 use the value in the default purpose.
16727 * Fix the PKCS#8 DSA private key code so it decodes keys again
16728 and fix a memory leak.
16732 * In util/mkerr.pl (which implements 'make errors'), preserve
16733 reason strings from the previous version of the .c file, as
16734 the default to have only downcase letters (and digits) in
16735 automatically generated reasons codes is not always appropriate.
16739 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16740 using strerror. Previously, ERR_reason_error_string() returned
16741 library names as reason strings for SYSerr; but SYSerr is a special
16742 case where small numbers are errno values, not library numbers.
16746 * Add '-dsaparam' option to 'openssl dhparam' application. This
16747 converts DSA parameters into DH parameters. (When creating parameters,
16748 DSA_generate_parameters is used.)
16752 * Include 'length' (recommended exponent length) in C code generated
16753 by 'openssl dhparam -C'.
16757 * The second argument to set_label in perlasm was already being used
16758 so couldn't be used as a "file scope" flag. Moved to third argument
16763 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16764 instead of RAND_bytes for encryption IVs and salts.
16768 * Include RAND_status() into RAND_METHOD instead of implementing
16769 it only for md_rand.c Otherwise replacing the PRNG by calling
16770 RAND_set_rand_method would be impossible.
16774 * Don't let DSA_generate_key() enter an infinite loop if the random
16775 number generation fails.
16779 * New 'rand' application for creating pseudo-random output.
16783 * Added configuration support for Linux/IA64
16785 *Rolf Haberrecker <rolf@suse.de>*
16787 * Assembler module support for Mingw32.
16791 * Shared library support for HPUX (in shlib/).
16793 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16795 * Shared library support for Solaris gcc.
16797 *Lutz Behnke <behnke@trustcenter.de>*
16799 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16801 * PKCS7_encrypt() was adding text MIME headers twice because they
16802 were added manually and by SMIME_crlf_copy().
16806 * In bntest.c don't call BN_rand with zero bits argument.
16808 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16810 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16811 case was implemented. This caused BN_div_recp() to fail occasionally.
16815 * Add an optional second argument to the set_label() in the perl
16816 assembly language builder. If this argument exists and is set
16817 to 1 it signals that the assembler should use a symbol whose
16818 scope is the entire file, not just the current function. This
16819 is needed with MASM which uses the format label:: for this scope.
16821 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16823 * Change the ASN1 types so they are typedefs by default. Before
16824 almost all types were #define'd to ASN1_STRING which was causing
16825 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16830 * Change names of new functions to the new get1/get0 naming
16831 convention: After 'get1', the caller owns a reference count
16832 and has to call `..._free`; 'get0' returns a pointer to some
16833 data structure without incrementing reference counters.
16834 (Some of the existing 'get' functions increment a reference
16835 counter, some don't.)
16836 Similarly, 'set1' and 'add1' functions increase reference
16837 counters or duplicate objects.
16841 * Allow for the possibility of temp RSA key generation failure:
16842 the code used to assume it always worked and crashed on failure.
16846 * Fix potential buffer overrun problem in BIO_printf().
16847 *Ulf Möller, using public domain code by Patrick Powell; problem
16848 pointed out by David Sacerdote <das33@cornell.edu>*
16850 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16851 RAND_egd() and RAND_status(). In the command line application,
16852 the EGD socket can be specified like a seed file using RANDFILE
16857 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16858 Some CAs (e.g. Verisign) distribute certificates in this form.
16862 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16863 list to exclude them. This means that no special compilation option
16864 is needed to use anonymous DH: it just needs to be included in the
16869 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16870 EVP_MD_type. The old functionality is available in a new macro called
16871 EVP_MD_md(). Change code that uses it and update docs.
16875 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16876 where the `void *` argument is replaced by a function pointer argument.
16877 Previously `void *` was abused to point to functions, which works on
16878 many platforms, but is not correct. As these functions are usually
16879 called by macros defined in OpenSSL header files, most source code
16880 should work without changes.
16884 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16885 sections with information on -D... compiler switches used for
16886 compiling the library so that applications can see them. To enable
16887 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16888 must be defined. E.g.,
16889 #define OPENSSL_ALGORITHM_DEFINES
16890 #include <openssl/opensslconf.h>
16891 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16893 *Richard Levitte, Ulf and Bodo Möller*
16895 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16900 * Change the 'other' type in certificate aux info to a STACK_OF
16901 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16902 the required ASN1 format: arbitrary types determined by an OID.
16906 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16907 argument to 'req'. This is not because the function is newer or
16908 better than others it just uses the work 'NEW' in the certificate
16909 request header lines. Some software needs this.
16913 * Reorganise password command line arguments: now passwords can be
16914 obtained from various sources. Delete the PEM_cb function and make
16915 it the default behaviour: i.e. if the callback is NULL and the
16916 usrdata argument is not NULL interpret it as a null terminated pass
16917 phrase. If usrdata and the callback are NULL then the pass phrase
16918 is prompted for as usual.
16922 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16923 the support is automatically enabled. The resulting binaries will
16924 autodetect the card and use it if present.
16926 *Ben Laurie and Compaq Inc.*
16928 * Work around for Netscape hang bug. This sends certificate request
16929 and server done in one record. Since this is perfectly legal in the
16930 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16931 the bugs/SSLv3 entry for more info.
16935 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16939 * Add -rand argument to smime and pkcs12 applications and read/write
16944 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16948 * Add command line password options to the remaining applications.
16952 * Bug fix for BN_div_recp() for numerators with an even number of
16957 * More tests in bntest.c, and changed test_bn output.
16961 * ./config recognizes MacOS X now.
16965 * Bug fix for BN_div() when the first words of num and divisor are
16966 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16970 * Add support for various broken PKCS#8 formats, and command line
16971 options to produce them.
16975 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16976 get temporary BIGNUMs from a BN_CTX.
16980 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16985 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16986 include a #define from the old name to the new. The original intent
16987 was that statically linked binaries could for example just call
16988 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16989 link with digests. This never worked because SSLeay_add_all_digests()
16990 and SSLeay_add_all_ciphers() were in the same source file so calling
16991 one would link with the other. They are now in separate source files.
16995 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16999 * Use a less unusual form of the Miller-Rabin primality test (it used
17000 a binary algorithm for exponentiation integrated into the Miller-Rabin
17001 loop, our standard modexp algorithms are faster).
17005 * Support for the EBCDIC character set completed.
17007 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17009 * Source code cleanups: use const where appropriate, eliminate casts,
17010 use `void *` instead of `char *` in lhash.
17014 * Bugfix: ssl3_send_server_key_exchange was not restartable
17015 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17016 this the server could overwrite ephemeral keys that the client
17021 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17022 using 50 iterations of the Rabin-Miller test.
17024 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17025 iterations of the Rabin-Miller test as required by the appendix
17026 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17027 As BN_is_prime_fasttest includes trial division, DSA parameter
17028 generation becomes much faster.
17030 This implies a change for the callback functions in DSA_is_prime
17031 and DSA_generate_parameters: The callback function is called once
17032 for each positive witness in the Rabin-Miller test, not just
17033 occasionally in the inner loop; and the parameters to the
17034 callback function now provide an iteration count for the outer
17035 loop rather than for the current invocation of the inner loop.
17036 DSA_generate_parameters additionally can call the callback
17037 function with an 'iteration count' of -1, meaning that a
17038 candidate has passed the trial division test (when q is generated
17039 from an application-provided seed, trial division is skipped).
17043 * New function BN_is_prime_fasttest that optionally does trial
17044 division before starting the Rabin-Miller test and has
17045 an additional BN_CTX * argument (whereas BN_is_prime always
17046 has to allocate at least one BN_CTX).
17047 'callback(1, -1, cb_arg)' is called when a number has passed the
17048 trial division stage.
17052 * Fix for bug in CRL encoding. The validity dates weren't being handled
17057 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17061 * New function BN_pseudo_rand().
17065 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17066 bignum version of BN_from_montgomery() with the working code from
17067 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17072 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17073 made it impossible to use the same SSL_SESSION data structure in
17074 SSL2 clients in multiple threads.
17078 * The return value of RAND_load_file() no longer counts bytes obtained
17079 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17080 to seed the PRNG (previously an explicit byte count was required).
17082 *Ulf Möller, Bodo Möller*
17084 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17085 used `char *` instead of `void *` and had casts all over the place.
17089 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17093 * Retain source code compatibility for BN_prime_checks macro:
17094 BN_is_prime(..., BN_prime_checks, ...) now uses
17095 BN_prime_checks_for_size to determine the appropriate number of
17096 Rabin-Miller iterations.
17100 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17101 DH_CHECK_P_NOT_SAFE_PRIME.
17102 (Check if this is true? OpenPGP calls them "strong".)
17106 * Merge the functionality of "dh" and "gendh" programs into a new program
17107 "dhparam". The old programs are retained for now but will handle DH keys
17108 (instead of parameters) in future.
17112 * Make the ciphers, s_server and s_client programs check the return values
17113 when a new cipher list is set.
17117 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17118 ciphers. Before when the 56bit ciphers were enabled the sorting was
17121 The syntax for the cipher sorting has been extended to support sorting by
17122 cipher-strength (using the strength_bits hard coded in the tables).
17123 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17125 Fix a bug in the cipher-command parser: when supplying a cipher command
17126 string with an "undefined" symbol (neither command nor alphanumeric
17127 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17128 an error is flagged.
17130 Due to the strength-sorting extension, the code of the
17131 ssl_create_cipher_list() function was completely rearranged. I hope that
17132 the readability was also increased :-)
17134 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17136 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17137 for the first serial number and places 2 in the serial number file. This
17138 avoids problems when the root CA is created with serial number zero and
17139 the first user certificate has the same issuer name and serial number
17144 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17145 the new code. Add documentation for this stuff.
17149 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17150 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17151 structures and behave in an analogous way to the X509v3 functions:
17152 they shouldn't be called directly but wrapper functions should be used
17155 So we also now have some wrapper functions that call the X509at functions
17156 when passed certificate requests. (TO DO: similar things can be done with
17157 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17158 things. Some of these need some d2i or i2d and print functionality
17159 because they handle more complex structures.)
17163 * Add missing #ifndefs that caused missing symbols when building libssl
17164 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17165 NO_RSA in `ssl/s2*.c`.
17167 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17169 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17170 has a return value which indicates the quality of the random data
17171 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17172 error queue. New function RAND_pseudo_bytes() generates output that is
17173 guaranteed to be unique but not unpredictable. RAND_add is like
17174 RAND_seed, but takes an extra argument for an entropy estimate
17175 (RAND_seed always assumes full entropy).
17179 * Do more iterations of Rabin-Miller probable prime test (specifically,
17180 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17181 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17182 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17183 false-positive rate of at most 2^-80 for random input.
17187 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17191 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17192 in the 0.9.5 release), this returns the chain
17193 from an X509_CTX structure with a dup of the stack and all
17194 the X509 reference counts upped: so the stack will exist
17195 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17198 Also make SSL_SESSION_print() print out the verify return
17203 * Add manpage for the pkcs12 command. Also change the default
17204 behaviour so MAC iteration counts are used unless the new
17205 -nomaciter option is used. This improves file security and
17206 only older versions of MSIE (4.0 for example) need it.
17210 * Honor the no-xxx Configure options when creating .DEF files.
17214 * Add PKCS#10 attributes to field table: challengePassword,
17215 unstructuredName and unstructuredAddress. These are taken from
17216 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17217 international characters are used.
17219 More changes to X509_ATTRIBUTE code: allow the setting of types
17220 based on strings. Remove the 'loc' parameter when adding
17221 attributes because these will be a SET OF encoding which is sorted
17226 * Initial changes to the 'req' utility to allow request generation
17227 automation. This will allow an application to just generate a template
17228 file containing all the field values and have req construct the
17231 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17232 used all over the place including certificate requests and PKCS#7
17233 structures. They are currently handled manually where necessary with
17234 some primitive wrappers for PKCS#7. The new functions behave in a
17235 manner analogous to the X509 extension functions: they allow
17236 attributes to be looked up by NID and added.
17238 Later something similar to the X509V3 code would be desirable to
17239 automatically handle the encoding, decoding and printing of the
17240 more complex types. The string types like challengePassword can
17241 be handled by the string table functions.
17243 Also modified the multi byte string table handling. Now there is
17244 a 'global mask' which masks out certain types. The table itself
17245 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17246 is useful when for example there is only one permissible type
17247 (as in countryName) and using the mask might result in no valid
17252 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17253 SSL_get_peer_finished to allow applications to obtain the latest
17254 Finished messages sent to the peer or expected from the peer,
17255 respectively. (SSL_get_peer_finished is usually the Finished message
17256 actually received from the peer, otherwise the protocol will be aborted.)
17258 As the Finished message are message digests of the complete handshake
17259 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17260 be used for external authentication procedures when the authentication
17261 provided by SSL/TLS is not desired or is not enough.
17265 * Enhanced support for Alpha Linux is added. Now ./config checks if
17266 the host supports BWX extension and if Compaq C is present on the
17267 $PATH. Just exploiting of the BWX extension results in 20-30%
17268 performance kick for some algorithms, e.g. DES and RC4 to mention
17269 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17274 * Add support for MS "fast SGC". This is arguably a violation of the
17275 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17276 weak crypto and after checking the certificate is SGC a second one
17277 with strong crypto. MS SGC stops the first handshake after receiving
17278 the server certificate message and sends a second client hello. Since
17279 a server will typically do all the time consuming operations before
17280 expecting any further messages from the client (server key exchange
17281 is the most expensive) there is little difference between the two.
17283 To get OpenSSL to support MS SGC we have to permit a second client
17284 hello message after we have sent server done. In addition we have to
17285 reset the MAC if we do get this second client hello.
17289 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17290 if a DER encoded private key is RSA or DSA traditional format. Changed
17291 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17292 format DER encoded private key. Newer code should use PKCS#8 format which
17293 has the key type encoded in the ASN1 structure. Added DER private key
17294 support to pkcs8 application.
17298 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17299 ciphersuites has been selected (as required by the SSL 3/TLS 1
17300 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17301 is set, we interpret this as a request to violate the specification
17302 (the worst that can happen is a handshake failure, and 'correct'
17303 behaviour would result in a handshake failure anyway).
17307 * In SSL_CTX_add_session, take into account that there might be multiple
17308 SSL_SESSION structures with the same session ID (e.g. when two threads
17309 concurrently obtain them from an external cache).
17310 The internal cache can handle only one SSL_SESSION with a given ID,
17311 so if there's a conflict, we now throw out the old one to achieve
17316 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17317 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17318 some routines that use cipher OIDs: some ciphers do not have OIDs
17319 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17324 * Simplify the trust setting structure and code. Now we just have
17325 two sequences of OIDs for trusted and rejected settings. These will
17326 typically have values the same as the extended key usage extension
17327 and any application specific purposes.
17329 The trust checking code now has a default behaviour: it will just
17330 check for an object with the same NID as the passed id. Functions can
17331 be provided to override either the default behaviour or the behaviour
17332 for a given id. SSL client, server and email already have functions
17333 in place for compatibility: they check the NID and also return "trusted"
17334 if the certificate is self signed.
17338 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17339 traditional format into an EVP_PKEY structure.
17343 * Add a password callback function PEM_cb() which either prompts for
17344 a password if usr_data is NULL or otherwise assumes it is a null
17345 terminated password. Allow passwords to be passed on command line
17346 environment or config files in a few more utilities.
17350 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17351 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17352 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17353 Update documentation.
17357 * Support for ASN1 "NULL" type. This could be handled before by using
17358 ASN1_TYPE but there wasn't any function that would try to read a NULL
17359 and produce an error if it couldn't. For compatibility we also have
17360 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17361 don't allocate anything because they don't need to.
17365 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17368 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17370 * Rebuild of the memory allocation routines used by OpenSSL code and
17371 possibly others as well. The purpose is to make an interface that
17372 provide hooks so anyone can build a separate set of allocation and
17373 deallocation routines to be used by OpenSSL, for example memory
17374 pool implementations, or something else, which was previously hard
17375 since Malloc(), Realloc() and Free() were defined as macros having
17376 the values malloc, realloc and free, respectively (except for Win32
17377 compilations). The same is provided for memory debugging code.
17378 OpenSSL already comes with functionality to find memory leaks, but
17379 this gives people a chance to debug other memory problems.
17381 With these changes, a new set of functions and macros have appeared:
17383 CRYPTO_set_mem_debug_functions() [F]
17384 CRYPTO_get_mem_debug_functions() [F]
17385 CRYPTO_dbg_set_options() [F]
17386 CRYPTO_dbg_get_options() [F]
17387 CRYPTO_malloc_debug_init() [M]
17389 The memory debug functions are NULL by default, unless the library
17390 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17391 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17392 gives the standard debugging functions that come with OpenSSL) or
17393 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17394 provided by the library user) must be used. When the standard
17395 debugging functions are used, CRYPTO_dbg_set_options can be used to
17396 request additional information:
17397 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17398 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17400 Also, things like CRYPTO_set_mem_functions will always give the
17401 expected result (the new set of functions is used for allocation
17402 and deallocation) at all times, regardless of platform and compiler
17405 To finish it up, some functions that were never use in any other
17406 way than through macros have a new API and new semantic:
17408 CRYPTO_dbg_malloc()
17409 CRYPTO_dbg_realloc()
17412 All macros of value have retained their old syntax.
17414 *Richard Levitte and Bodo Moeller*
17416 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17417 ordering of SMIMECapabilities wasn't in "strength order" and there
17418 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17423 * Some ASN1 types with illegal zero length encoding (INTEGER,
17424 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17426 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17428 * Merge in my S/MIME library for OpenSSL. This provides a simple
17429 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17430 functionality to handle multipart/signed properly) and a utility
17431 called 'smime' to call all this stuff. This is based on code I
17432 originally wrote for Celo who have kindly allowed it to be
17433 included in OpenSSL.
17437 * Add variants des_set_key_checked and des_set_key_unchecked of
17438 des_set_key (aka des_key_sched). Global variable des_check_key
17439 decides which of these is called by des_set_key; this way
17440 des_check_key behaves as it always did, but applications and
17441 the library itself, which was buggy for des_check_key == 1,
17442 have a cleaner way to pick the version they need.
17446 * New function PKCS12_newpass() which changes the password of a
17451 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17452 dynamic mix. In both cases the ids can be used as an index into the
17453 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17454 functions so they accept a list of the field values and the
17455 application doesn't need to directly manipulate the X509_TRUST
17460 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17465 * Modify the way the V3 extension code looks up extensions. This now
17466 works in a similar way to the object code: we have some "standard"
17467 extensions in a static table which is searched with OBJ_bsearch()
17468 and the application can add dynamic ones if needed. The file
17469 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17470 updated whenever a new extension is added to the core code and kept
17471 in ext_nid order. There is a simple program 'tabtest.c' which checks
17472 this. New extensions are not added too often so this file can readily
17473 be maintained manually.
17475 There are two big advantages in doing things this way. The extensions
17476 can be looked up immediately and no longer need to be "added" using
17477 X509V3_add_standard_extensions(): this function now does nothing.
17478 Side note: I get *lots* of email saying the extension code doesn't
17479 work because people forget to call this function.
17480 Also no dynamic allocation is done unless new extensions are added:
17481 so if we don't add custom extensions there is no need to call
17482 X509V3_EXT_cleanup().
17486 * Modify enc utility's salting as follows: make salting the default. Add a
17487 magic header, so unsalted files fail gracefully instead of just decrypting
17488 to garbage. This is because not salting is a big security hole, so people
17489 should be discouraged from doing it.
17493 * Fixes and enhancements to the 'x509' utility. It allowed a message
17494 digest to be passed on the command line but it only used this
17495 parameter when signing a certificate. Modified so all relevant
17496 operations are affected by the digest parameter including the
17497 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17498 DSA key was used because it didn't fix the digest.
17502 * Initial certificate chain verify code. Currently tests the untrusted
17503 certificates for consistency with the verify purpose (which is set
17504 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17506 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17507 this is because it will reject chains with invalid extensions whereas
17508 every previous version of OpenSSL and SSLeay made no checks at all.
17510 Trust code: checks the root CA for the relevant trust settings. Trust
17511 settings have an initial value consistent with the verify purpose: e.g.
17512 if the verify purpose is for SSL client use it expects the CA to be
17513 trusted for SSL client use. However the default value can be changed to
17514 permit custom trust settings: one example of this would be to only trust
17515 certificates from a specific "secure" set of CAs.
17517 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17518 which should be used for version portability: especially since the
17519 verify structure is likely to change more often now.
17521 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17522 to set them. If not set then assume SSL clients will verify SSL servers
17525 Two new options to the verify program: -untrusted allows a set of
17526 untrusted certificates to be passed in and -purpose which sets the
17527 intended purpose of the certificate. If a purpose is set then the
17528 new chain verify code is used to check extension consistency.
17532 * Support for the authority information access extension.
17536 * Modify RSA and DSA PEM read routines to transparently handle
17537 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17538 public keys in a format compatible with certificate
17539 SubjectPublicKeyInfo structures. Unfortunately there were already
17540 functions called *_PublicKey_* which used various odd formats so
17541 these are retained for compatibility: however the DSA variants were
17542 never in a public release so they have been deleted. Changed dsa/rsa
17543 utilities to handle the new format: note no releases ever handled public
17544 keys so we should be OK.
17546 The primary motivation for this change is to avoid the same fiasco
17547 that dogs private keys: there are several incompatible private key
17548 formats some of which are standard and some OpenSSL specific and
17549 require various evil hacks to allow partial transparent handling and
17550 even then it doesn't work with DER formats. Given the option anything
17551 other than PKCS#8 should be dumped: but the other formats have to
17552 stay in the name of compatibility.
17554 With public keys and the benefit of hindsight one standard format
17555 is used which works with EVP_PKEY, RSA or DSA structures: though
17556 it clearly returns an error if you try to read the wrong kind of key.
17558 Added a -pubkey option to the 'x509' utility to output the public key.
17559 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17560 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17561 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17562 that do the same as the `EVP_PKEY_assign_*()` except they up the
17563 reference count of the added key (they don't "swallow" the
17568 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17569 CRLs would fail if the file contained no certificates or no CRLs:
17570 added a new function to read in both types and return the number
17571 read: this means that if none are read it will be an error. The
17572 DER versions of the certificate and CRL reader would always fail
17573 because it isn't possible to mix certificates and CRLs in DER format
17574 without choking one or the other routine. Changed this to just read
17575 a certificate: this is the best we can do. Also modified the code
17576 in `apps/verify.c` to take notice of return codes: it was previously
17577 attempting to read in certificates from NULL pointers and ignoring
17578 any errors: this is one reason why the cert and CRL reader seemed
17579 to work. It doesn't check return codes from the default certificate
17580 routines: these may well fail if the certificates aren't installed.
17584 * Code to support otherName option in GeneralName.
17588 * First update to verify code. Change the verify utility
17589 so it warns if it is passed a self signed certificate:
17590 for consistency with the normal behaviour. X509_verify
17591 has been modified to it will now verify a self signed
17592 certificate if *exactly* the same certificate appears
17593 in the store: it was previously impossible to trust a
17594 single self signed certificate. This means that:
17595 openssl verify ss.pem
17596 now gives a warning about a self signed certificate but
17597 openssl verify -CAfile ss.pem ss.pem
17602 * For servers, store verify_result in SSL_SESSION data structure
17603 (and add it to external session representation).
17604 This is needed when client certificate verifications fails,
17605 but an application-provided verification callback (set by
17606 SSL_CTX_set_cert_verify_callback) allows accepting the session
17607 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17608 but returns 1): When the session is reused, we have to set
17609 ssl->verify_result to the appropriate error code to avoid
17612 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17614 * Fix a bug in the new PKCS#7 code: it didn't consider the
17615 case in PKCS7_dataInit() where the signed PKCS7 structure
17616 didn't contain any existing data because it was being created.
17618 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17620 * Add a salt to the key derivation routines in enc.c. This
17621 forms the first 8 bytes of the encrypted file. Also add a
17622 -S option to allow a salt to be input on the command line.
17626 * New function X509_cmp(). Oddly enough there wasn't a function
17627 to compare two certificates. We do this by working out the SHA1
17628 hash and comparing that. X509_cmp() will be needed by the trust
17633 * SSL_get1_session() is like SSL_get_session(), but increments
17634 the reference count in the SSL_SESSION returned.
17636 *Geoff Thorpe <geoff@eu.c2.net>*
17638 * Fix for 'req': it was adding a null to request attributes.
17639 Also change the X509_LOOKUP and X509_INFO code to handle
17640 certificate auxiliary information.
17644 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17649 * Add the possibility to add extra information to the memory leak
17650 detecting output, to form tracebacks, showing from where each
17651 allocation was originated: CRYPTO_push_info("constant string") adds
17652 the string plus current file name and line number to a per-thread
17653 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17654 is like calling CYRPTO_pop_info() until the stack is empty.
17655 Also updated memory leak detection code to be multi-thread-safe.
17659 * Add options -text and -noout to pkcs7 utility and delete the
17660 encryption options which never did anything. Update docs.
17664 * Add options to some of the utilities to allow the pass phrase
17665 to be included on either the command line (not recommended on
17666 OSes like Unix) or read from the environment. Update the
17667 manpages and fix a few bugs.
17671 * Add a few manpages for some of the openssl commands.
17675 * Fix the -revoke option in ca. It was freeing up memory twice,
17676 leaking and not finding already revoked certificates.
17680 * Extensive changes to support certificate auxiliary information.
17681 This involves the use of X509_CERT_AUX structure and X509_AUX
17682 functions. An X509_AUX function such as PEM_read_X509_AUX()
17683 can still read in a certificate file in the usual way but it
17684 will also read in any additional "auxiliary information". By
17685 doing things this way a fair degree of compatibility can be
17686 retained: existing certificates can have this information added
17687 using the new 'x509' options.
17689 Current auxiliary information includes an "alias" and some trust
17690 settings. The trust settings will ultimately be used in enhanced
17691 certificate chain verification routines: currently a certificate
17692 can only be trusted if it is self signed and then it is trusted
17697 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17698 The problem was that one of the replacement routines had not been working
17699 since SSLeay releases. For now the offending routine has been replaced
17700 with non-optimised assembler. Even so, this now gives around 95%
17701 performance improvement for 1024 bit RSA signs.
17705 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17706 handling. Most clients have the effective key size in bits equal to
17707 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17708 A few however don't do this and instead use the size of the decrypted key
17709 to determine the RC2 key length and the AlgorithmIdentifier to determine
17710 the effective key length. In this case the effective key length can still
17711 be 40 bits but the key length can be 168 bits for example. This is fixed
17712 by manually forcing an RC2 key into the EVP_PKEY structure because the
17713 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17714 the key length and effective key length are equal.
17718 * Add a bunch of functions that should simplify the creation of
17719 X509_NAME structures. Now you should be able to do:
17720 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17721 and have it automatically work out the correct field type and fill in
17722 the structures. The more adventurous can try:
17723 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17724 and it will (hopefully) work out the correct multibyte encoding.
17728 * Change the 'req' utility to use the new field handling and multibyte
17729 copy routines. Before the DN field creation was handled in an ad hoc
17730 way in req, ca, and x509 which was rather broken and didn't support
17731 BMPStrings or UTF8Strings. Since some software doesn't implement
17732 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17733 using the dirstring_type option. See the new comment in the default
17734 openssl.cnf for more info.
17738 * Make crypto/rand/md_rand.c more robust:
17739 - Assure unique random numbers after fork().
17740 - Make sure that concurrent threads access the global counter and
17741 md serializably so that we never lose entropy in them
17742 or use exactly the same state in multiple threads.
17743 Access to the large state is not always serializable because
17744 the additional locking could be a performance killer, and
17745 md should be large enough anyway.
17749 * New file `apps/app_rand.c` with commonly needed functionality
17750 for handling the random seed file.
17752 Use the random seed file in some applications that previously did not:
17754 dsaparam -genkey (which also ignored its '-rand' option),
17757 x509 (when signing).
17758 Except on systems with /dev/urandom, it is crucial to have a random
17759 seed file at least for key creation, DSA signing, and for DH exchanges;
17760 for RSA signatures we could do without one.
17762 gendh and gendsa (unlike genrsa) used to read only the first byte
17763 of each file listed in the '-rand' option. The function as previously
17764 found in genrsa is now in app_rand.c and is used by all programs
17765 that support '-rand'.
17769 * In RAND_write_file, use mode 0600 for creating files;
17770 don't just chmod when it may be too late.
17774 * Report an error from X509_STORE_load_locations
17775 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17779 * New function ASN1_mbstring_copy() this copies a string in either
17780 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17781 into an ASN1_STRING type. A mask of permissible types is passed
17782 and it chooses the "minimal" type to use or an error if not type
17787 * Add function equivalents to the various macros in asn1.h. The old
17788 macros are retained with an `M_` prefix. Code inside the library can
17789 use the `M_` macros. External code (including the openssl utility)
17790 should *NOT* in order to be "shared library friendly".
17794 * Add various functions that can check a certificate's extensions
17795 to see if it usable for various purposes such as SSL client,
17796 server or S/MIME and CAs of these types. This is currently
17797 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17798 verification. Also added a -purpose flag to x509 utility to
17799 print out all the purposes.
17803 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17808 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17809 for, obtain and decode and extension and obtain its critical flag.
17810 This allows all the necessary extension code to be handled in a
17811 single function call.
17815 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17816 platforms. See crypto/rc4/rc4_enc.c for further details.
17820 * New -noout option to asn1parse. This causes no output to be produced
17821 its main use is when combined with -strparse and -out to extract data
17822 from a file (which may not be in ASN.1 format).
17826 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17827 when producing the local key id.
17829 *Richard Levitte <levitte@stacken.kth.se>*
17831 * New option -dhparam in s_server. This allows a DH parameter file to be
17832 stated explicitly. If it is not stated then it tries the first server
17833 certificate file. The previous behaviour hard coded the filename
17838 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17839 a public key to be input or output. For example:
17840 openssl rsa -in key.pem -pubout -out pubkey.pem
17841 Also added necessary DSA public key functions to handle this.
17845 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17846 in the message. This was handled by allowing
17847 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17849 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17851 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17852 to the end of the strings whereas this didn't. This would cause problems
17853 if strings read with d2i_ASN1_bytes() were later modified.
17855 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17857 * Fix for base64 decode bug. When a base64 bio reads only one line of
17858 data and it contains EOF it will end up returning an error. This is
17859 caused by input 46 bytes long. The cause is due to the way base64
17860 BIOs find the start of base64 encoded data. They do this by trying a
17861 trial decode on each line until they find one that works. When they
17862 do a flag is set and it starts again knowing it can pass all the
17863 data directly through the decoder. Unfortunately it doesn't reset
17864 the context it uses. This means that if EOF is reached an attempt
17865 is made to pass two EOFs through the context and this causes the
17866 resulting error. This can also cause other problems as well. As is
17867 usual with these problems it takes *ages* to find and the fix is
17868 trivial: move one line.
17870 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17872 * Ugly workaround to get s_client and s_server working under Windows. The
17873 old code wouldn't work because it needed to select() on sockets and the
17874 tty (for keypresses and to see if data could be written). Win32 only
17875 supports select() on sockets so we select() with a 1s timeout on the
17876 sockets and then see if any characters are waiting to be read, if none
17877 are present then we retry, we also assume we can always write data to
17878 the tty. This isn't nice because the code then blocks until we've
17879 received a complete line of data and it is effectively polling the
17880 keyboard at 1s intervals: however it's quite a bit better than not
17881 working at all :-) A dedicated Windows application might handle this
17882 with an event loop for example.
17886 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17887 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17888 will be called when RSA_sign() and RSA_verify() are used. This is useful
17889 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17890 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17891 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17892 This necessitated the support of an extra signature type NID_md5_sha1
17893 for SSL signatures and modifications to the SSL library to use it instead
17894 of calling RSA_public_decrypt() and RSA_private_encrypt().
17898 * Add new -verify -CAfile and -CApath options to the crl program, these
17899 will lookup a CRL issuers certificate and verify the signature in a
17900 similar way to the verify program. Tidy up the crl program so it
17901 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17902 less strict. It will now permit CRL extensions even if it is not
17903 a V2 CRL: this will allow it to tolerate some broken CRLs.
17907 * Initialize all non-automatic variables each time one of the openssl
17908 sub-programs is started (this is necessary as they may be started
17909 multiple times from the "OpenSSL>" prompt).
17911 *Lennart Bang, Bodo Moeller*
17913 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17914 removing all other RSA functionality (this is what NO_RSA does). This
17915 is so (for example) those in the US can disable those operations covered
17916 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17921 * Non-copying interface to BIO pairs.
17922 (still largely untested)
17926 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17927 ASCII string. This was handled independently in various places before.
17931 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17932 UTF8 strings a character at a time.
17936 * Use client_version from client hello to select the protocol
17937 (s23_srvr.c) and for RSA client key exchange verification
17938 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17942 * Add various utility functions to handle SPKACs, these were previously
17943 handled by poking round in the structure internals. Added new function
17944 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17945 print, verify and generate SPKACs. Based on an original idea from
17946 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17950 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17954 * Allow the config file extension section to be overwritten on the
17955 command line. Based on an original idea from Massimiliano Pala
17956 <madwolf@comune.modena.it>. The new option is called -extensions
17957 and can be applied to ca, req and x509. Also -reqexts to override
17958 the request extensions in req and -crlexts to override the crl extensions
17963 * Add new feature to the SPKAC handling in ca. Now you can include
17964 the same field multiple times by preceding it by "XXXX." for example:
17967 this is the same syntax as used in the req config file.
17971 * Allow certificate extensions to be added to certificate requests. These
17972 are specified in a 'req_extensions' option of the req section of the
17973 config file. They can be printed out with the -text option to req but
17974 are otherwise ignored at present.
17978 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17979 data read consists of only the final block it would not decrypted because
17980 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17981 A misplaced 'break' also meant the decrypted final block might not be
17982 copied until the next read.
17986 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17987 a few extra parameters to the DH structure: these will be useful if
17988 for example we want the value of 'q' or implement X9.42 DH.
17992 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17993 provides hooks that allow the default DSA functions or functions on a
17994 "per key" basis to be replaced. This allows hardware acceleration and
17995 hardware key storage to be handled without major modification to the
17996 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17997 associated functions.
18001 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18002 as "read only": it can't be written to and the buffer it points to will
18003 not be freed. Reading from a read only BIO is much more efficient than
18004 a normal memory BIO. This was added because there are several times when
18005 an area of memory needs to be read from a BIO. The previous method was
18006 to create a memory BIO and write the data to it, this results in two
18007 copies of the data and an O(n^2) reading algorithm. There is a new
18008 function BIO_new_mem_buf() which creates a read only memory BIO from
18009 an area of memory. Also modified the PKCS#7 routines to use read only
18014 * Bugfix: ssl23_get_client_hello did not work properly when called in
18015 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18016 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18017 but a retry condition occurred while trying to read the rest.
18021 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18022 NID_pkcs7_encrypted by default: this was wrong since this should almost
18023 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18024 the encrypted data type: this is a more sensible place to put it and it
18025 allows the PKCS#12 code to be tidied up that duplicated this
18030 * Changed obj_dat.pl script so it takes its input and output files on
18031 the command line. This should avoid shell escape redirection problems
18036 * Initial support for certificate extension requests, these are included
18037 in things like Xenroll certificate requests. Included functions to allow
18038 extensions to be obtained and added.
18042 * -crlf option to s_client and s_server for sending newlines as
18043 CRLF (as required by many protocols).
18047 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18049 * Install libRSAglue.a when OpenSSL is built with RSAref.
18051 *Ralf S. Engelschall*
18053 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18055 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18057 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18062 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18063 DH parameters/keys (q is lost during that conversion, but the resulting
18064 DH parameters contain its length).
18066 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18067 much faster than DH_generate_parameters (which creates parameters
18068 where `p = 2*q + 1`), and also the smaller q makes DH computations
18069 much more efficient (160-bit exponentiation instead of 1024-bit
18070 exponentiation); so this provides a convenient way to support DHE
18071 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18072 utter importance to use
18073 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18075 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18076 when such DH parameters are used, because otherwise small subgroup
18077 attacks may become possible!
18081 * Avoid memory leak in i2d_DHparams.
18085 * Allow the -k option to be used more than once in the enc program:
18086 this allows the same encrypted message to be read by multiple recipients.
18090 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18091 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18092 it will always use the numerical form of the OID, even if it has a short
18097 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18098 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18099 otherwise bn_mod_exp was called. In the case of hardware keys for example
18100 no private key components need be present and it might store extra data
18101 in the RSA structure, which cannot be accessed from bn_mod_exp.
18102 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18103 private key operations.
18107 * Added support for SPARC Linux.
18111 * pem_password_cb function type incompatibly changed from
18112 typedef int pem_password_cb(char *buf, int size, int rwflag);
18114 ....(char *buf, int size, int rwflag, void *userdata);
18115 so that applications can pass data to their callbacks:
18116 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18117 additional void * argument, which is just handed through whenever
18118 the password callback is called.
18120 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18122 New function SSL_CTX_set_default_passwd_cb_userdata.
18124 Compatibility note: As many C implementations push function arguments
18125 onto the stack in reverse order, the new library version is likely to
18126 interoperate with programs that have been compiled with the old
18127 pem_password_cb definition (PEM_whatever takes some data that
18128 happens to be on the stack as its last argument, and the callback
18129 just ignores this garbage); but there is no guarantee whatsoever that
18132 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18133 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18134 problems not only on Windows, but also on some Unix platforms.
18135 To avoid problematic command lines, these definitions are now in an
18136 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18137 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18141 * MIPS III/IV assembler module is reimplemented.
18145 * More DES library cleanups: remove references to srand/rand and
18146 delete an unused file.
18150 * Add support for the free Netwide assembler (NASM) under Win32,
18151 since not many people have MASM (ml) and it can be hard to obtain.
18152 This is currently experimental but it seems to work OK and pass all
18153 the tests. Check out INSTALL.W32 for info.
18157 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18158 without temporary keys kept an extra copy of the server key,
18159 and connections with temporary keys did not free everything in case
18164 * New function RSA_check_key and new openssl rsa option -check
18165 for verifying the consistency of RSA keys.
18167 *Ulf Moeller, Bodo Moeller*
18169 * Various changes to make Win32 compile work:
18170 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18171 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18172 comparison" warnings.
18173 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18177 * Add a debugging option to PKCS#5 v2 key generation function: when
18178 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18179 derived keys are printed to stderr.
18183 * Copy the flags in ASN1_STRING_dup().
18185 *Roman E. Pavlov <pre@mo.msk.ru>*
18187 * The x509 application mishandled signing requests containing DSA
18188 keys when the signing key was also DSA and the parameters didn't match.
18190 It was supposed to omit the parameters when they matched the signing key:
18191 the verifying software was then supposed to automatically use the CA's
18192 parameters if they were absent from the end user certificate.
18194 Omitting parameters is no longer recommended. The test was also
18195 the wrong way round! This was probably due to unusual behaviour in
18196 EVP_cmp_parameters() which returns 1 if the parameters match.
18197 This meant that parameters were omitted when they *didn't* match and
18198 the certificate was useless. Certificates signed with 'ca' didn't have
18201 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18203 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18204 The interface is as follows:
18205 Applications can use
18206 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18207 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18208 "off" is now the default.
18209 The library internally uses
18210 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18211 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18212 to disable memory-checking temporarily.
18214 Some inconsistent states that previously were possible (and were
18215 even the default) are now avoided.
18217 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18218 with each memory chunk allocated; this is occasionally more helpful
18219 than just having a counter.
18221 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18223 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18228 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18229 which largely parallels "options", but is for changing API behaviour,
18230 whereas "options" are about protocol behaviour.
18231 Initial "mode" flags are:
18233 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18234 a single record has been written.
18235 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18236 retries use the same buffer location.
18237 (But all of the contents must be
18242 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18245 * Fix problems with no-hmac etc.
18247 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18249 * New functions RSA_get_default_method(), RSA_set_method() and
18250 RSA_get_method(). These allows replacement of RSA_METHODs without having
18251 to mess around with the internals of an RSA structure.
18255 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18256 Also really enable memory leak checks in openssl.c and in some
18259 *Chad C. Mulligan, Bodo Moeller*
18261 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18262 up the length of negative integers. This has now been simplified to just
18263 store the length when it is first determined and use it later, rather
18264 than trying to keep track of where data is copied and updating it to
18266 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18268 * Add a new function PKCS7_signatureVerify. This allows the verification
18269 of a PKCS#7 signature but with the signing certificate passed to the
18270 function itself. This contrasts with PKCS7_dataVerify which assumes the
18271 certificate is present in the PKCS#7 structure. This isn't always the
18272 case: certificates can be omitted from a PKCS#7 structure and be
18273 distributed by "out of band" means (such as a certificate database).
18277 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18278 function prototypes in pem.h, also change util/mkdef.pl to add the
18279 necessary function names.
18283 * mk1mf.pl (used by Windows builds) did not properly read the
18284 options set by Configure in the top level Makefile, and Configure
18285 was not even able to write more than one option correctly.
18286 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18290 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18291 file to be loaded from a BIO or FILE pointer. The BIO version will
18292 for example allow memory BIOs to contain config info.
18296 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18297 Whoever hopes to achieve shared-library compatibility across versions
18298 must use this, not the compile-time macro.
18299 (Exercise 0.9.4: Which is the minimum library version required by
18301 Note: All this applies only to multi-threaded programs, others don't
18306 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18307 through a BIO pair triggered the default case, i.e.
18308 SSLerr(...,SSL_R_UNKNOWN_STATE).
18312 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18313 can use the SSL library even if none of the specific BIOs is
18318 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18319 for the encoded length.
18321 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18323 * Add initial documentation of the X509V3 functions.
18327 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18328 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18329 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18330 secure PKCS#8 private key format with a high iteration count.
18334 * Fix determination of Perl interpreter: A perl or perl5
18335 *directory* in $PATH was also accepted as the interpreter.
18337 *Ralf S. Engelschall*
18339 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18340 wrong with it but it was very old and did things like calling
18341 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18342 unusual formatting.
18346 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18347 to use the new extension code.
18351 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18352 with macros. This should make it easier to change their form, add extra
18353 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18358 * Add to configuration table a new entry that can specify an alternative
18359 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18360 according to Mark Crispin <MRC@Panda.COM>.
18364 * DES CBC did not update the IV. Weird.
18368 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18369 Changing the behaviour of the former might break existing programs --
18370 where IV updating is needed, des_ncbc_encrypt can be used.
18373 * When bntest is run from "make test" it drives bc to check its
18374 calculations, as well as internally checking them. If an internal check
18375 fails, it needs to cause bc to give a non-zero result or make test carries
18376 on without noticing the failure. Fixed.
18380 * DES library cleanups.
18384 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18385 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18386 ciphers. NOTE: although the key derivation function has been verified
18387 against some published test vectors it has not been extensively tested
18388 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18393 * Instead of "mkdir -p", which is not fully portable, use new
18394 Perl script "util/mkdir-p.pl".
18398 * Rewrite the way password based encryption (PBE) is handled. It used to
18399 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18400 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18401 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18402 the 'parameter' field of the AlgorithmIdentifier is passed to the
18403 underlying key generation function so it must do its own ASN1 parsing.
18404 This has also changed the EVP_PBE_CipherInit() function which now has a
18405 'parameter' argument instead of literal salt and iteration count values
18406 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18410 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18411 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18412 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18413 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18414 value was just used as a "magic string" and not used directly its
18415 value doesn't matter.
18419 * Introduce some semblance of const correctness to BN. Shame C doesn't
18424 * "linux-sparc64" configuration (ultrapenguin).
18426 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18427 "linux-sparc" configuration.
18429 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18431 * config now generates no-xxx options for missing ciphers.
18435 * Support the EBCDIC character set (work in progress).
18436 File ebcdic.c not yet included because it has a different license.
18438 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18440 * Support BS2000/OSD-POSIX.
18442 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18444 * Make callbacks for key generation use `void *` instead of `char *`.
18448 * Make S/MIME samples compile (not yet tested).
18452 * Additional typesafe stacks.
18456 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18460 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18462 * New configuration variant "sco5-gcc".
18464 * Updated some demos.
18466 *Sean O Riordain, Wade Scholine*
18468 * Add missing BIO_free at exit of pkcs12 application.
18472 * Fix memory leak in conf.c.
18476 * Updates for Win32 to assembler version of MD5.
18480 * Set #! path to perl in `apps/der_chop` to where we found it
18481 instead of using a fixed path.
18485 * SHA library changes for irix64-mips4-cc.
18489 * Improvements for VMS support.
18493 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18495 * Bignum library bug fix. IRIX 6 passes "make test" now!
18496 This also avoids the problems with SC4.2 and unpatched SC5.
18498 *Andy Polyakov <appro@fy.chalmers.se>*
18500 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18501 These are required because of the typesafe stack would otherwise break
18502 existing code. If old code used a structure member which used to be STACK
18503 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18504 sk_num or sk_value it would produce an error because the num, data members
18505 are not present in STACK_OF. Now it just produces a warning. sk_set
18506 replaces the old method of assigning a value to sk_value
18507 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18508 that does this will no longer work (and should use sk_set instead) but
18509 this could be regarded as a "questionable" behaviour anyway.
18513 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18514 correctly handle encrypted S/MIME data.
18518 * Change type of various DES function arguments from des_cblock
18519 (which means, in function argument declarations, pointer to char)
18520 to des_cblock * (meaning pointer to array with 8 char elements),
18521 which allows the compiler to do more typechecking; it was like
18522 that back in SSLeay, but with lots of ugly casts.
18524 Introduce new type const_des_cblock.
18528 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18529 problems: find RecipientInfo structure that matches recipient certificate
18530 and initialise the ASN1 structures properly based on passed cipher.
18534 * Belatedly make the BN tests actually check the results.
18538 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18539 to and from BNs: it was completely broken. New compilation option
18540 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18541 key elements as negative integers.
18545 * Reorganize and speed up MD5.
18547 *Andy Polyakov <appro@fy.chalmers.se>*
18551 *Richard Levitte <richard@levitte.org>*
18553 * New option -out to asn1parse to allow the parsed structure to be
18554 output to a file. This is most useful when combined with the -strparse
18555 option to examine the output of things like OCTET STRINGS.
18559 * Make SSL library a little more fool-proof by not requiring any longer
18560 that `SSL_set_{accept,connect}_state` be called before
18561 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18562 in many applications because usually everything *appeared* to work as
18563 intended anyway -- now it really works as intended).
18567 * Move openssl.cnf out of lib/.
18571 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18572 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18573 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18575 *Ralf S. Engelschall*
18577 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18578 handle PKCS#7 enveloped data properly.
18580 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18582 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18583 copying pointers. The cert_st handling is changed by this in
18584 various ways (and thus what used to be known as ctx->default_cert
18585 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18586 any longer when s->cert does not give us what we need).
18587 ssl_cert_instantiate becomes obsolete by this change.
18588 As soon as we've got the new code right (possibly it already is?),
18589 we have solved a couple of bugs of the earlier code where s->cert
18590 was used as if it could not have been shared with other SSL structures.
18592 Note that using the SSL API in certain dirty ways now will result
18593 in different behaviour than observed with earlier library versions:
18594 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18595 does not influence s as it used to.
18597 In order to clean up things more thoroughly, inside SSL_SESSION
18598 we don't use CERT any longer, but a new structure SESS_CERT
18599 that holds per-session data (if available); currently, this is
18600 the peer's certificate chain and, for clients, the server's certificate
18601 and temporary key. CERT holds only those values that can have
18602 meaningful defaults in an SSL_CTX.
18606 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18607 from the internal representation. Various PKCS#7 fixes: remove some
18608 evil casts and set the enc_dig_alg field properly based on the signing
18613 * Allow PKCS#12 password to be set from the command line or the
18614 environment. Let 'ca' get its config file name from the environment
18615 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18620 * Allow certificate policies extension to use an IA5STRING for the
18621 organization field. This is contrary to the PKIX definition but
18622 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18627 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18628 without disallowing inline assembler and the like for non-pedantic builds.
18632 * Support Borland C++ builder.
18634 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18640 * SHA-1 cleanups and performance enhancements.
18642 *Andy Polyakov <appro@fy.chalmers.se>*
18644 * Sparc v8plus assembler for the bignum library.
18646 *Andy Polyakov <appro@fy.chalmers.se>*
18648 * Accept any -xxx and +xxx compiler options in Configure.
18652 * Update HPUX configuration.
18656 * Add missing `sk_<type>_unshift()` function to safestack.h
18658 *Ralf S. Engelschall*
18660 * New function SSL_CTX_use_certificate_chain_file that sets the
18661 "extra_cert"s in addition to the certificate. (This makes sense
18662 only for "PEM" format files, as chains as a whole are not
18667 * Support verify_depth from the SSL API.
18668 x509_vfy.c had what can be considered an off-by-one-error:
18669 Its depth (which was not part of the external interface)
18670 was actually counting the number of certificates in a chain;
18671 now it really counts the depth.
18675 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18676 instead of X509err, which often resulted in confusing error
18677 messages since the error codes are not globally unique
18678 (e.g. an alleged error in ssl3_accept when a certificate
18679 didn't match the private key).
18681 * New function SSL_CTX_set_session_id_context that allows to set a default
18682 value (so that you don't need SSL_set_session_id_context for each
18683 connection using the SSL_CTX).
18687 * OAEP decoding bug fix.
18691 * Support INSTALL_PREFIX for package builders, as proposed by
18696 * New Configure options "threads" and "no-threads". For systems
18697 where the proper compiler options are known (currently Solaris
18698 and Linux), "threads" is the default.
18702 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18706 * Install various scripts to $(OPENSSLDIR)/misc, not to
18707 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18708 such as /usr/local/bin.
18712 * "make linux-shared" to build shared libraries.
18714 *Niels Poppe <niels@netbox.org>*
18716 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18720 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18721 extension adding in x509 utility.
18725 * Remove NOPROTO sections and error code comments.
18729 * Partial rewrite of the DEF file generator to now parse the ANSI
18734 * New Configure options --prefix=DIR and --openssldir=DIR.
18738 * Complete rewrite of the error code script(s). It is all now handled
18739 by one script at the top level which handles error code gathering,
18740 header rewriting and C source file generation. It should be much better
18741 than the old method: it now uses a modified version of Ulf's parser to
18742 read the ANSI prototypes in all header files (thus the old K&R definitions
18743 aren't needed for error creation any more) and do a better job of
18744 translating function codes into names. The old 'ASN1 error code embedded
18745 in a comment' is no longer necessary and it doesn't use .err files which
18746 have now been deleted. Also the error code call doesn't have to appear all
18747 on one line (which resulted in some large lines...).
18751 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18755 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18756 0 (which usually indicates a closed connection), but continue reading.
18760 * Fix some race conditions.
18764 * Add support for CRL distribution points extension. Add Certificate
18765 Policies and CRL distribution points documentation.
18769 * Move the autogenerated header file parts to crypto/opensslconf.h.
18773 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18774 8 of keying material. Merlin has also confirmed interop with this fix
18775 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18777 *Merlin Hughes <merlin@baltimore.ie>*
18779 * Fix lots of warnings.
18781 *Richard Levitte <levitte@stacken.kth.se>*
18783 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18784 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18786 *Richard Levitte <levitte@stacken.kth.se>*
18788 * Fix problems with sizeof(long) == 8.
18790 *Andy Polyakov <appro@fy.chalmers.se>*
18792 * Change functions to ANSI C.
18796 * Fix typos in error codes.
18798 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18800 * Remove defunct assembler files from Configure.
18804 * SPARC v8 assembler BIGNUM implementation.
18806 *Andy Polyakov <appro@fy.chalmers.se>*
18808 * Support for Certificate Policies extension: both print and set.
18809 Various additions to support the r2i method this uses.
18813 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18814 return a const string when you are expecting an allocated buffer.
18818 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18819 types DirectoryString and DisplayText.
18823 * Add code to allow r2i extensions to access the configuration database,
18824 add an LHASH database driver and add several ctx helper functions.
18828 * Fix an evil bug in bn_expand2() which caused various BN functions to
18829 fail when they extended the size of a BIGNUM.
18833 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18834 support typesafe stack.
18838 * Fix typo in SSL_[gs]et_options().
18840 *Nils Frostberg <nils@medcom.se>*
18842 * Delete various functions and files that belonged to the (now obsolete)
18843 old X509V3 handling code.
18847 * New Configure option "rsaref".
18851 * Don't auto-generate pem.h.
18855 * Introduce type-safe ASN.1 SETs.
18859 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18861 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18863 * Introduce type-safe STACKs. This will almost certainly break lots of code
18864 that links with OpenSSL (well at least cause lots of warnings), but fear
18865 not: the conversion is trivial, and it eliminates loads of evil casts. A
18866 few STACKed things have been converted already. Feel free to convert more.
18867 In the fullness of time, I'll do away with the STACK type altogether.
18871 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18872 specified in `<certfile>` by updating the entry in the index.txt file.
18873 This way one no longer has to edit the index.txt file manually for
18874 revoking a certificate. The -revoke option does the gory details now.
18876 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18878 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18879 `-text` option at all and this way the `-noout -text` combination was
18880 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18882 *Ralf S. Engelschall*
18884 * Make sure a corresponding plain text error message exists for the
18885 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18886 verify callback function determined that a certificate was revoked.
18888 *Ralf S. Engelschall*
18890 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18891 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18892 all available ciphers including rc5, which was forgotten until now.
18893 In order to let the testing shell script know which algorithms
18894 are available, a new (up to now undocumented) command
18895 `openssl list-cipher-commands` is used.
18899 * Bugfix: s_client occasionally would sleep in select() when
18900 it should have checked SSL_pending() first.
18904 * New functions DSA_do_sign and DSA_do_verify to provide access to
18905 the raw DSA values prior to ASN.1 encoding.
18909 * Tweaks to Configure
18911 *Niels Poppe <niels@netbox.org>*
18913 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18918 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18922 * New config option to avoid instructions that are illegal on the 80386.
18923 The default code is faster, but requires at least a 486.
18927 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18928 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18929 same as SSL2_VERSION anyway.
18933 * New "-showcerts" option for s_client.
18937 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18938 application. Various cleanups and fixes.
18942 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18943 modify error routines to work internally. Add error codes and PBE init
18944 to library startup routines.
18948 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18949 packing functions to asn1 and evp. Changed function names and error
18950 codes along the way.
18954 * PKCS12 integration: and so it begins... First of several patches to
18955 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18956 objects to objects.h
18960 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18961 and display support for Thawte strong extranet extension.
18965 * Add LinuxPPC support.
18967 *Jeff Dubrule <igor@pobox.org>*
18969 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18970 bn_div_words in alpha.s.
18972 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18974 * Make sure the RSA OAEP test is skipped under -DRSAref because
18975 OAEP isn't supported when OpenSSL is built with RSAref.
18977 *Ulf Moeller <ulf@fitug.de>*
18979 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18980 so they no longer are missing under -DNOPROTO.
18982 *Soren S. Jorvang <soren@t.dk>*
18984 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18986 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18987 doesn't work when the session is reused. Coming soon!
18991 * Fix a security hole, that allows sessions to be reused in the wrong
18992 context thus bypassing client cert protection! All software that uses
18993 client certs and session caches in multiple contexts NEEDS PATCHING to
18994 allow session reuse! A fuller solution is in the works.
18996 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18998 * Some more source tree cleanups (removed obsolete files
18999 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19000 permission on "config" script to be executable) and a fix for the INSTALL
19003 *Ulf Moeller <ulf@fitug.de>*
19005 * Remove some legacy and erroneous uses of malloc, free instead of
19008 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19010 * Make rsa_oaep_test return non-zero on error.
19012 *Ulf Moeller <ulf@fitug.de>*
19014 * Add support for native Solaris shared libraries. Configure
19015 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19016 if someone would make that last step automatic.
19018 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19020 * ctx_size was not built with the right compiler during "make links". Fixed.
19024 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19025 except NULL ciphers". This means the default cipher list will no longer
19026 enable NULL ciphers. They need to be specifically enabled e.g. with
19027 the string "DEFAULT:eNULL".
19031 * Fix to RSA private encryption routines: if p < q then it would
19032 occasionally produce an invalid result. This will only happen with
19033 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19037 * Be less restrictive and allow also `perl util/perlpath.pl
19038 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19039 because this way one can also use an interpreter named `perl5` (which is
19040 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19041 installed as `perl`).
19043 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19045 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19047 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19049 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19050 advapi32.lib to Win32 build and change the pem test comparison
19051 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19052 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19053 and crypto/des/ede_cbcm_enc.c.
19057 * DES quad checksum was broken on big-endian architectures. Fixed.
19061 * Comment out two functions in bio.h that aren't implemented. Fix up the
19062 Win32 test batch file so it (might) work again. The Win32 test batch file
19063 is horrible: I feel ill....
19067 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19068 in e_os.h. Audit of header files to check ANSI and non ANSI
19069 sections: 10 functions were absent from non ANSI section and not exported
19070 from Windows DLLs. Fixed up libeay.num for new functions.
19074 * Make `openssl version` output lines consistent.
19076 *Ralf S. Engelschall*
19078 * Fix Win32 symbol export lists for BIO functions: Added
19079 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19080 to ms/libeay{16,32}.def.
19082 *Ralf S. Engelschall*
19084 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19085 fine under Unix and passes some trivial tests I've now added. But the
19086 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19087 added to make sure no one expects that this stuff really works in the
19088 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19089 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19092 *Ralf S. Engelschall*
19094 * Fix the generation of two part addresses in perl.
19096 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19098 * Add config entry for Linux on MIPS.
19100 *John Tobey <jtobey@channel1.com>*
19102 * Make links whenever Configure is run, unless we are on Windoze.
19106 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19107 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19112 * Add a useful kludge to allow package maintainers to specify compiler and
19113 other platforms details on the command line without having to patch the
19114 Configure script every time: One now can use
19115 `perl Configure <id>:<details>`,
19116 i.e. platform ids are allowed to have details appended
19117 to them (separated by colons). This is treated as there would be a static
19118 pre-configured entry in Configure's %table under key `<id>` with value
19119 `<details>` and `perl Configure <id>` is called. So, when you want to
19120 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19121 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19122 now, which overrides the FreeBSD-elf entry on-the-fly.
19124 *Ralf S. Engelschall*
19126 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19130 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19131 on the `perl Configure ...` command line. This way one can compile
19132 OpenSSL libraries with Position Independent Code (PIC) which is needed
19133 for linking it into DSOs.
19135 *Ralf S. Engelschall*
19137 * Remarkably, export ciphers were totally broken and no-one had noticed!
19142 * Cleaned up the LICENSE document: The official contact for any license
19143 questions now is the OpenSSL core team under openssl-core@openssl.org.
19144 And add a paragraph about the dual-license situation to make sure people
19145 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19146 to the OpenSSL toolkit.
19148 *Ralf S. Engelschall*
19150 * General source tree makefile cleanups: Made `making xxx in yyy...`
19151 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19152 Additionally cleaned up the `make links` target: Remove unnecessary
19153 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19154 to speed processing and no longer clutter the display with confusing
19155 stuff. Instead only the actually done links are displayed.
19157 *Ralf S. Engelschall*
19159 * Permit null encryption ciphersuites, used for authentication only. It used
19160 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19161 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19166 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19167 signed attributes when verifying signatures (this would break them),
19168 the detached data encoding was wrong and public keys obtained using
19169 X509_get_pubkey() weren't freed.
19173 * Add text documentation for the BUFFER functions. Also added a work around
19174 to a Win95 console bug. This was triggered by the password read stuff: the
19175 last character typed gets carried over to the next fread(). If you were
19176 generating a new cert request using 'req' for example then the last
19177 character of the passphrase would be CR which would then enter the first
19182 * Added the new 'Includes OpenSSL Cryptography Software' button as
19183 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19184 button and can be used by applications based on OpenSSL to show the
19185 relationship to the OpenSSL project.
19187 *Ralf S. Engelschall*
19189 * Remove confusing variables in function signatures in files
19190 ssl/ssl_lib.c and ssl/ssl.h.
19192 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19194 * Don't install bss_file.c under PREFIX/include/
19196 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19198 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19199 functions that return function pointers and has support for NT specific
19200 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19201 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19202 unsigned to signed types: this was killing the Win32 compile.
19206 * Add new certificate file to stack functions,
19207 SSL_add_dir_cert_subjects_to_stack() and
19208 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19209 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19210 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19211 This means that Apache-SSL and similar packages don't have to mess around
19212 to add as many CAs as they want to the preferred list.
19216 * Experiment with doxygen documentation. Currently only partially applied to
19218 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19219 openssl.doxy as the configuration file.
19223 * Get rid of remaining C++-style comments which strict C compilers hate.
19225 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19227 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19228 compiled in by default: it has problems with large keys.
19232 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19233 DH private keys and/or callback functions which directly correspond to
19234 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19235 is needed for applications which have to configure certificates on a
19236 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19238 For the RSA certificate situation is makes no difference, but
19239 for the DSA certificate situation this fixes the "no shared cipher"
19240 problem where the OpenSSL cipher selection procedure failed because the
19241 temporary keys were not overtaken from the context and the API provided
19242 no way to reconfigure them.
19243 The new functions now let applications reconfigure the stuff and they
19244 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19245 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19246 non-public-API function ssl_cert_instantiate() is used as a helper
19247 function and also to reduce code redundancy inside ssl_rsa.c.
19249 *Ralf S. Engelschall*
19251 * Move s_server -dcert and -dkey options out of the undocumented feature
19252 area because they are useful for the DSA situation and should be
19253 recognized by the users.
19255 *Ralf S. Engelschall*
19257 * Fix the cipher decision scheme for export ciphers: the export bits are
19258 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19259 SSL_EXP_MASK. So, the original variable has to be used instead of the
19260 already masked variable.
19262 *Richard Levitte <levitte@stacken.kth.se>*
19264 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19266 *Richard Levitte <levitte@stacken.kth.se>*
19268 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19269 from `int` to `unsigned int` because it is a length and initialized by
19270 EVP_DigestFinal() which expects an `unsigned int *`.
19272 *Richard Levitte <levitte@stacken.kth.se>*
19274 * Don't hard-code path to Perl interpreter on shebang line of Configure
19275 script. Instead use the usual Shell->Perl transition trick.
19277 *Ralf S. Engelschall*
19279 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19280 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19281 -noout -modulus` as it's already the case for `openssl rsa -noout
19282 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19283 currently the public key is printed (a decision which was already done by
19284 `openssl dsa -modulus` in the past) which serves a similar purpose.
19285 Additionally the NO_RSA no longer completely removes the whole -modulus
19286 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19289 *Ralf S. Engelschall*
19291 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19292 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19294 *Arne Ansper <arne@ats.cyber.ee>*
19296 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19297 to be added. Now both 'req' and 'ca' can use new objects defined in the
19302 * Add cool BIO that does syslog (or event log on NT).
19304 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19306 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19307 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19308 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19309 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19313 * Add preliminary config info for new extension code.
19317 * Make RSA_NO_PADDING really use no padding.
19319 *Ulf Moeller <ulf@fitug.de>*
19321 * Generate errors when private/public key check is done.
19325 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19326 for some CRL extensions and new objects added.
19330 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19331 key usage extension and fuller support for authority key id.
19335 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19336 padding method for RSA, which is recommended for new applications in PKCS
19337 #1 v2.0 (RFC 2437, October 1998).
19338 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19339 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19340 against Bleichbacher's attack on RSA.
19341 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19344 * Updates to the new SSL compression code
19346 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19348 * Fix so that the version number in the master secret, when passed
19349 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19350 (because the server will not accept higher), that the version number
19351 is 0x03,0x01, not 0x03,0x00
19353 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19355 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19356 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19357 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19361 * Support for RAW extensions where an arbitrary extension can be
19362 created by including its DER encoding. See `apps/openssl.cnf` for
19367 * Make sure latest Perl versions don't interpret some generated C array
19368 code as Perl array code in the crypto/err/err_genc.pl script.
19370 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19372 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19373 not many people have the assembler. Various Win32 compilation fixes and
19374 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19375 build instructions.
19379 * Modify configure script 'Configure' to automatically create crypto/date.h
19380 file under Win32 and also build pem.h from pem.org. New script
19381 util/mkfiles.pl to create the MINFO file on environments that can't do a
19382 'make files': perl util/mkfiles.pl >MINFO should work.
19386 * Major rework of DES function declarations, in the pursuit of correctness
19387 and purity. As a result, many evil casts evaporated, and some weirdness,
19388 too. You may find this causes warnings in your code. Zapping your evil
19389 casts will probably fix them. Mostly.
19393 * Fix for a typo in asn1.h. Bug fix to object creation script
19394 obj_dat.pl. It considered a zero in an object definition to mean
19395 "end of object": none of the objects in objects.h have any zeros
19396 so it wasn't spotted.
19398 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19400 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19401 Masking (CBCM). In the absence of test vectors, the best I have been able
19402 to do is check that the decrypt undoes the encrypt, so far. Send me test
19403 vectors if you have them.
19407 * Correct calculation of key length for export ciphers (too much space was
19408 allocated for null ciphers). This has not been tested!
19412 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19413 message is now correct (it understands "crypto" and "ssl" on its
19414 command line). There is also now an "update" option. This will update
19415 the util/ssleay.num and util/libeay.num files with any new functions.
19417 perl util/mkdef.pl crypto ssl update
19418 it will update them.
19422 * Overhauled the Perl interface:
19423 - ported BN stuff to OpenSSL's different BN library
19424 - made the perl/ source tree CVS-aware
19425 - renamed the package from SSLeay to OpenSSL (the files still contain
19426 their history because I've copied them in the repository)
19427 - removed obsolete files (the test scripts will be replaced
19428 by better Test::Harness variants in the future)
19430 *Ralf S. Engelschall*
19432 * First cut for a very conservative source tree cleanup:
19433 1. merge various obsolete readme texts into doc/ssleay.txt
19434 where we collect the old documents and readme texts.
19435 2. remove the first part of files where I'm already sure that we no
19436 longer need them because of three reasons: either they are just temporary
19437 files which were left by Eric or they are preserved original files where
19438 I've verified that the diff is also available in the CVS via "cvs diff
19439 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19440 the crypto/md/ stuff).
19442 *Ralf S. Engelschall*
19444 * More extension code. Incomplete support for subject and issuer alt
19445 name, issuer and authority key id. Change the i2v function parameters
19446 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19447 what that's for :-) Fix to ASN1 macro which messed up
19448 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19452 * Preliminary support for ENUMERATED type. This is largely copied from the
19457 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19459 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19461 * Make sure `make rehash` target really finds the `openssl` program.
19463 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19465 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19466 like to hear about it if this slows down other processors.
19470 * Add CygWin32 platform information to Configure script.
19472 *Alan Batie <batie@aahz.jf.intel.com>*
19474 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19476 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19478 * New program nseq to manipulate netscape certificate sequences
19482 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19487 * Fixes to BN code. Previously the default was to define BN_RECURSION
19488 but the BN code had some problems that would cause failures when
19489 doing certificate verification and some other functions.
19491 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19493 * Add ASN1 and PEM code to support netscape certificate sequences.
19497 * Add ASN1 and PEM code to support netscape certificate sequences.
19501 * Add several PKIX and private extended key usage OIDs.
19505 * Modify the 'ca' program to handle the new extension code. Modify
19506 openssl.cnf for new extension format, add comments.
19510 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19511 and add a sample to openssl.cnf so req -x509 now adds appropriate
19516 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19517 error code, add initial support to X509_print() and x509 application.
19521 * Takes a deep breath and start adding X509 V3 extension support code. Add
19522 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19523 stuff is currently isolated and isn't even compiled yet.
19527 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19528 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19529 Removed the versions check from X509 routines when loading extensions:
19530 this allows certain broken certificates that don't set the version
19531 properly to be processed.
19535 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19536 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19537 can still be regenerated with "make depend".
19541 * Spelling mistake in C version of CAST-128.
19543 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19545 * Changes to the error generation code. The perl script err-code.pl
19546 now reads in the old error codes and retains the old numbers, only
19547 adding new ones if necessary. It also only changes the .err files if new
19548 codes are added. The makefiles have been modified to only insert errors
19549 when needed (to avoid needlessly modifying header files). This is done
19550 by only inserting errors if the .err file is newer than the auto generated
19551 C file. To rebuild all the error codes from scratch (the old behaviour)
19552 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19553 or delete all the .err files.
19557 * CAST-128 was incorrectly implemented for short keys. The C version has
19558 been fixed, but is untested. The assembler versions are also fixed, but
19559 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19560 to regenerate it if needed.
19561 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19562 Hagino <itojun@kame.net>*
19564 * File was opened incorrectly in randfile.c.
19566 *Ulf Möller <ulf@fitug.de>*
19568 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19569 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19570 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19571 al: it's just almost always a UTCTime. Note this patch adds new error
19572 codes so do a "make errors" if there are problems.
19576 * Correct Linux 1 recognition in config.
19578 *Ulf Möller <ulf@fitug.de>*
19580 * Remove pointless MD5 hash when using DSA keys in ca.
19582 *Anonymous <nobody@replay.com>*
19584 * Generate an error if given an empty string as a cert directory. Also
19585 generate an error if handed NULL (previously returned 0 to indicate an
19586 error, but didn't set one).
19588 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19590 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19594 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19595 parameters. This was causing a warning which killed off the Win32 compile.
19599 * Remove C++ style comments from crypto/bn/bn_local.h.
19601 *Neil Costigan <neil.costigan@celocom.com>*
19603 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19604 based on a text string, looking up short and long names and finally
19605 "dot" format. The "dot" format stuff didn't work. Added new function
19606 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19607 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19608 OID is not part of the table.
19612 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19613 X509_LOOKUP_by_alias().
19617 * Sort openssl functions by name.
19621 * Get the `gendsa` command working and add it to the `list` command. Remove
19622 encryption from sample DSA keys (in case anyone is interested the password
19627 * Make *all* `*_free` functions accept a NULL pointer.
19629 *Frans Heymans <fheymans@isaserver.be>*
19631 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19634 *Anonymous <nobody@replay.com>*
19636 * s_server should send the CAfile as acceptable CAs, not its own cert.
19638 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19640 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19642 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19644 * Temp key "for export" tests were wrong in s3_srvr.c.
19646 *Anonymous <nobody@replay.com>*
19648 * Add prototype for temp key callback functions
19649 SSL_CTX_set_tmp_{rsa,dh}_callback().
19653 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19654 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19658 * X509_name_add_entry() freed the wrong thing after an error.
19660 *Arne Ansper <arne@ats.cyber.ee>*
19662 * rsa_eay.c would attempt to free a NULL context.
19664 *Arne Ansper <arne@ats.cyber.ee>*
19666 * BIO_s_socket() had a broken should_retry() on Windoze.
19668 *Arne Ansper <arne@ats.cyber.ee>*
19670 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19672 *Arne Ansper <arne@ats.cyber.ee>*
19674 * Make sure the already existing X509_STORE->depth variable is initialized
19675 in X509_STORE_new(), but document the fact that this variable is still
19676 unused in the certificate verification process.
19678 *Ralf S. Engelschall*
19680 * Fix the various library and `apps/` files to free up pkeys obtained from
19681 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19685 * Fix reference counting in X509_PUBKEY_get(). This makes
19686 demos/maurice/example2.c work, amongst others, probably.
19688 *Steve Henson and Ben Laurie*
19690 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19691 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19692 are no longer created. This way we have a single and consistent command
19693 line interface `openssl <command>`, similar to `cvs <command>`.
19695 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19697 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19698 BIT STRING wrapper always have zero unused bits.
19702 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19706 * Make the top-level INSTALL documentation easier to understand.
19710 * Makefiles updated to exit if an error occurs in a sub-directory
19711 make (including if user presses ^C) [Paul Sutton]
19713 * Make Montgomery context stuff explicit in RSA data structure.
19717 * Fix build order of pem and err to allow for generated pem.h.
19721 * Fix renumbering bug in X509_NAME_delete_entry().
19725 * Enhanced the err-ins.pl script so it makes the error library number
19726 global and can add a library name. This is needed for external ASN1 and
19727 other error libraries.
19731 * Fixed sk_insert which never worked properly.
19735 * Fix ASN1 macros so they can handle indefinite length constructed
19736 EXPLICIT tags. Some non standard certificates use these: they can now
19741 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19742 into a single doc/ssleay.txt bundle. This way the information is still
19743 preserved but no longer messes up this directory. Now it's new room for
19744 the new set of documentation files.
19746 *Ralf S. Engelschall*
19748 * SETs were incorrectly DER encoded. This was a major pain, because they
19749 shared code with SEQUENCEs, which aren't coded the same. This means that
19750 almost everything to do with SETs or SEQUENCEs has either changed name or
19751 number of arguments.
19753 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19755 * Fix test data to work with the above.
19759 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19760 was already fixed by Eric for 0.9.1 it seems.
19762 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19764 * Autodetect FreeBSD3.
19768 * Fix various bugs in Configure. This affects the following platforms:
19772 unixware-2.0-pentium
19777 * Eliminate generated files from CVS. Reorder tests to regenerate files
19778 before they are needed.
19782 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19786 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19788 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19789 changed SSLeay to OpenSSL in version strings.
19791 *Ralf S. Engelschall*
19793 * Some fixups to the top-level documents.
19797 * Fixed the nasty bug where rsaref.h was not found under compile-time
19798 because the symlink to include/ was missing.
19800 *Ralf S. Engelschall*
19802 * Incorporated the popular no-RSA/DSA-only patches
19803 which allow to compile an RSA-free SSLeay.
19805 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19807 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19808 when "ssleay" is still not found.
19810 *Ralf S. Engelschall*
19812 * Added more platforms to Configure: Cray T3E, HPUX 11,
19814 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19816 * Updated the README file.
19818 *Ralf S. Engelschall*
19820 * Added various .cvsignore files in the CVS repository subdirs
19821 to make a "cvs update" really silent.
19823 *Ralf S. Engelschall*
19825 * Recompiled the error-definition header files and added
19826 missing symbols to the Win32 linker tables.
19828 *Ralf S. Engelschall*
19830 * Cleaned up the top-level documents;
19831 o new files: CHANGES and LICENSE
19832 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19833 o merged COPYRIGHT into LICENSE
19834 o removed obsolete TODO file
19835 o renamed MICROSOFT to INSTALL.W32
19837 *Ralf S. Engelschall*
19839 * Removed dummy files from the 0.9.1b source tree:
19840 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19841 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19842 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19843 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19844 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19846 *Ralf S. Engelschall*
19848 * Added various platform portability fixes.
19852 * The Genesis of the OpenSSL rpject:
19853 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19854 Young and Tim J. Hudson created while they were working for C2Net until
19857 *The OpenSSL Project*
19859 ### Changes between 0.9.0b and 0.9.1b [not released]
19861 * Updated a few CA certificates under certs/
19865 * Changed some BIGNUM api stuff.
19869 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19870 DGUX x86, Linux Alpha, etc.
19874 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19875 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19880 * Add -strparse option to asn1pars program which parses nested
19883 *Dr Stephen Henson <shenson@bigfoot.com>*
19885 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19889 * DSA fix for "ca" program.
19893 * Added "-genkey" option to "dsaparam" program.
19897 * Added RIPE MD160 (rmd160) message digest.
19901 * Added -a (all) option to "ssleay version" command.
19905 * Added PLATFORM define which is the id given to Configure.
19909 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19913 * Extended the ASN.1 parser routines.
19917 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19921 * Added a BN_CTX to the BN library.
19925 * Fixed the weak key values in DES library
19929 * Changed API in EVP library for cipher aliases.
19933 * Added support for RC2/64bit cipher.
19937 * Converted the lhash library to the crypto/mem.c functions.
19941 * Added more recognized ASN.1 object ids.
19945 * Added more RSA padding checks for SSL/TLS.
19949 * Added BIO proxy/filter functionality.
19953 * Added extra_certs to SSL_CTX which can be used
19954 send extra CA certificates to the client in the CA cert chain sending
19955 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19959 * Now Fortezza is denied in the authentication phase because
19960 this is key exchange mechanism is not supported by SSLeay at all.
19964 * Additional PKCS1 checks.
19968 * Support the string "TLSv1" for all TLS v1 ciphers.
19972 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19973 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19977 * Fixed a few memory leaks.
19981 * Fixed various code and comment typos.
19985 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19986 bytes sent in the client random.
19988 *Edward Bishop <ebishop@spyglass.com>*
19992 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
19993 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
19994 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
19995 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
19996 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
19997 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
19998 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
19999 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20000 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20001 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20002 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20003 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20004 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20005 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20006 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20007 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20008 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20009 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20010 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20011 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20012 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20013 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20014 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20015 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20016 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20017 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20018 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20019 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20020 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20021 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20022 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20023 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20024 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20025 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20026 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20027 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20028 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20029 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20030 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20031 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20032 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20033 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20034 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20035 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20036 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20037 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20038 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20039 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20040 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20041 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20042 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20043 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20044 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20045 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20046 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20047 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20048 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20049 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20050 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20051 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20052 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20053 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20054 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20055 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20056 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20057 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20058 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20059 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20060 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20061 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20062 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20063 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20064 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20065 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20066 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20067 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20068 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20069 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20070 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20071 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20072 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20073 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20074 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20075 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20076 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20077 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20078 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20079 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20080 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20081 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20082 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20083 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20084 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20085 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20086 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20087 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20088 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20089 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20090 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20091 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20092 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20093 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20094 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20095 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20096 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20097 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20098 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20099 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20100 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20101 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20102 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20103 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20104 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20105 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20106 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20107 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20108 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20109 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20110 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20111 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20112 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20113 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20114 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20115 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20116 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20117 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20118 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20119 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20120 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20121 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20122 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20123 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20124 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20125 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20126 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20127 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20128 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20129 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20130 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20131 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20132 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20133 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20134 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20135 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20136 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20137 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20138 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20139 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20140 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20141 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20142 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20143 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20144 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20145 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20146 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20147 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20148 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20149 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20150 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20151 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20152 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20153 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20154 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20155 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20156 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20157 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20158 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20159 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20160 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20161 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20162 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20163 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20164 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20165 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20166 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20167 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20168 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20169 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20170 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655