5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) RAND_pseudo_bytes has been deprecated. Users should use RAND bytes instead.
9 *) Added support for TLS extended master secret from
10 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
11 initial patch which was a great help during development.
14 *) All libssl internal structures have been removed from the public header
15 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
16 now redundant). Users should not attempt to access internal structures
17 directly. Instead they should use the provided API functions.
20 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
21 Access to deprecated functions can be re-enabled by running config with
22 "enable-deprecated". In addition applications wishing to use deprecated
23 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
24 will, by default, disable some transitive includes that previously existed
25 in the header files (e.g. ec.h will no longer, by default, include bn.h)
28 *) Added support for OCB mode. OpenSSL has been granted a patent license
29 compatible with the OpenSSL license for use of OCB. Details are available
30 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
31 for OCB can be removed by calling config with no-ocb.
34 *) SSLv2 support has been removed. It still supports receiving a SSLv2
35 compatible client hello.
38 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
39 done while fixing the error code for the key-too-small case.
40 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
42 *) CA.sh has been removmed; use CA.pl instead.
45 *) Removed old DES API.
48 *) Remove various unsupported platforms:
54 Sinix/ReliantUNIX RM400
59 16-bit platforms such as WIN16
62 *) Clean up OPENSSL_NO_xxx #define's
63 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
64 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
65 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
66 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
67 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
68 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
69 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
70 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
71 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
72 Remove MS_STATIC; it's a relic from platforms <32 bits.
75 *) Cleaned up dead code
76 Remove all but one '#ifdef undef' which is to be looked at.
79 *) Clean up calling of xxx_free routines.
80 Just like free(), fix most of the xxx_free routines to accept
81 NULL. Remove the non-null checks from callers. Save much code.
84 *) Experimental support for a new, fast, unbiased prime candidate generator,
85 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
86 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
88 *) New output format NSS in the sess_id command line tool. This allows
89 exporting the session id and the master key in NSS keylog format.
90 [Martin Kaiser <martin@kaiser.cx>]
92 *) Harmonize version and its documentation. -f flag is used to display
94 [mancha <mancha1@zoho.com>]
96 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
98 [mancha <mancha1@zoho.com>]
100 *) Fix some double frees. These are not thought to be exploitable.
101 [mancha <mancha1@zoho.com>]
103 *) A missing bounds check in the handling of the TLS heartbeat extension
104 can be used to reveal up to 64k of memory to a connected client or
107 Thanks for Neel Mehta of Google Security for discovering this bug and to
108 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
109 preparing the fix (CVE-2014-0160)
110 [Adam Langley, Bodo Moeller]
112 *) Fix for the attack described in the paper "Recovering OpenSSL
113 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
114 by Yuval Yarom and Naomi Benger. Details can be obtained from:
115 http://eprint.iacr.org/2014/140
117 Thanks to Yuval Yarom and Naomi Benger for discovering this
118 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
119 [Yuval Yarom and Naomi Benger]
121 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
122 this fixes a limitation in previous versions of OpenSSL.
125 *) Experimental encrypt-then-mac support.
127 Experimental support for encrypt then mac from
128 draft-gutmann-tls-encrypt-then-mac-02.txt
130 To enable it set the appropriate extension number (0x42 for the test
131 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
133 For non-compliant peers (i.e. just about everything) this should have no
136 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
140 *) Add EVP support for key wrapping algorithms, to avoid problems with
141 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
142 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
143 algorithms and include tests cases.
146 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
150 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
151 MGF1 digest and OAEP label.
154 *) Make openssl verify return errors.
155 [Chris Palmer <palmer@google.com> and Ben Laurie]
157 *) New function ASN1_TIME_diff to calculate the difference between two
158 ASN1_TIME structures or one structure and the current time.
161 *) Update fips_test_suite to support multiple command line options. New
162 test to induce all self test errors in sequence and check expected
166 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
167 sign or verify all in one operation.
170 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
171 test programs and fips_test_suite. Includes functionality to parse
172 the minimal script output of fipsalgest.pl directly.
175 *) Add authorisation parameter to FIPS_module_mode_set().
178 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
181 *) Use separate DRBG fields for internal and external flags. New function
182 FIPS_drbg_health_check() to perform on demand health checking. Add
183 generation tests to fips_test_suite with reduced health check interval to
184 demonstrate periodic health checking. Add "nodh" option to
185 fips_test_suite to skip very slow DH test.
188 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
192 *) More extensive health check for DRBG checking many more failure modes.
193 New function FIPS_selftest_drbg_all() to handle every possible DRBG
194 combination: call this in fips_test_suite.
197 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
198 and POST to handle Dual EC cases.
201 *) Add support for canonical generation of DSA parameter 'g'. See
204 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
205 POST to handle HMAC cases.
208 *) Add functions FIPS_module_version() and FIPS_module_version_text()
209 to return numerical and string versions of the FIPS module number.
212 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
213 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
214 outside the validated module in the FIPS capable OpenSSL.
217 *) Minor change to DRBG entropy callback semantics. In some cases
218 there is no multiple of the block length between min_len and
219 max_len. Allow the callback to return more than max_len bytes
220 of entropy but discard any extra: it is the callback's responsibility
221 to ensure that the extra data discarded does not impact the
222 requested amount of entropy.
225 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
226 information in FIPS186-3, SP800-57 and SP800-131A.
229 *) CCM support via EVP. Interface is very similar to GCM case except we
230 must supply all data in one chunk (i.e. no update, final) and the
231 message length must be supplied if AAD is used. Add algorithm test
235 *) Initial version of POST overhaul. Add POST callback to allow the status
236 of POST to be monitored and/or failures induced. Modify fips_test_suite
237 to use callback. Always run all selftests even if one fails.
240 *) XTS support including algorithm test driver in the fips_gcmtest program.
241 Note: this does increase the maximum key length from 32 to 64 bytes but
242 there should be no binary compatibility issues as existing applications
243 will never use XTS mode.
246 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
247 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
248 performs algorithm blocking for unapproved PRNG types. Also do not
249 set PRNG type in FIPS_mode_set(): leave this to the application.
250 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
251 the standard OpenSSL PRNG: set additional data to a date time vector.
254 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
255 This shouldn't present any incompatibility problems because applications
256 shouldn't be using these directly and any that are will need to rethink
257 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
260 *) Extensive self tests and health checking required by SP800-90 DRBG.
261 Remove strength parameter from FIPS_drbg_instantiate and always
262 instantiate at maximum supported strength.
265 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
268 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
271 *) New function DH_compute_key_padded() to compute a DH key and pad with
272 leading zeroes if needed: this complies with SP800-56A et al.
275 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
276 anything, incomplete, subject to change and largely untested at present.
279 *) Modify fipscanisteronly build option to only build the necessary object
280 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
283 *) Add experimental option FIPSSYMS to give all symbols in
284 fipscanister.o and FIPS or fips prefix. This will avoid
285 conflicts with future versions of OpenSSL. Add perl script
286 util/fipsas.pl to preprocess assembly language source files
287 and rename any affected symbols.
290 *) Add selftest checks and algorithm block of non-fips algorithms in
291 FIPS mode. Remove DES2 from selftests.
294 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
295 return internal method without any ENGINE dependencies. Add new
296 tiny fips sign and verify functions.
299 *) New build option no-ec2m to disable characteristic 2 code.
302 *) New build option "fipscanisteronly". This only builds fipscanister.o
303 and (currently) associated fips utilities. Uses the file Makefile.fips
304 instead of Makefile.org as the prototype.
307 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
308 Update fips_gcmtest to use IV generator.
311 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
312 setting output buffer to NULL. The *Final function must be
313 called although it will not retrieve any additional data. The tag
314 can be set or retrieved with a ctrl. The IV length is by default 12
315 bytes (96 bits) but can be set to an alternative value. If the IV
316 length exceeds the maximum IV length (currently 16 bytes) it cannot be
320 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
321 underlying do_cipher function handles all cipher semantics itself
322 including padding and finalisation. This is useful if (for example)
323 an ENGINE cipher handles block padding itself. The behaviour of
324 do_cipher is subtly changed if this flag is set: the return value
325 is the number of characters written to the output buffer (zero is
326 no longer an error code) or a negative error code. Also if the
327 input buffer is NULL and length 0 finalisation should be performed.
330 *) If a candidate issuer certificate is already part of the constructed
331 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
334 *) Improve forward-security support: add functions
336 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
337 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
339 for use by SSL/TLS servers; the callback function will be called whenever a
340 new session is created, and gets to decide whether the session may be
341 cached to make it resumable (return 0) or not (return 1). (As by the
342 SSL/TLS protocol specifications, the session_id sent by the server will be
343 empty to indicate that the session is not resumable; also, the server will
344 not generate RFC 4507 (RFC 5077) session tickets.)
346 A simple reasonable callback implementation is to return is_forward_secure.
347 This parameter will be set to 1 or 0 depending on the ciphersuite selected
348 by the SSL/TLS server library, indicating whether it can provide forward
350 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
352 *) New -verify_name option in command line utilities to set verification
356 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
357 Add CMAC pkey methods.
360 *) Experimental renegotiation in s_server -www mode. If the client
361 browses /reneg connection is renegotiated. If /renegcert it is
362 renegotiated requesting a certificate.
365 *) Add an "external" session cache for debugging purposes to s_server. This
366 should help trace issues which normally are only apparent in deployed
367 multi-process servers.
370 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
371 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
372 BIO_set_cipher() and some obscure PEM functions were changed so they
373 can now return an error. The RAND changes required a change to the
374 RAND_METHOD structure.
377 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
378 a gcc attribute to warn if the result of a function is ignored. This
379 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
380 whose return value is often ignored.
383 Changes between 1.0.2 and 1.0.2a [xx XXX xxxx]
385 *) ClientHello sigalgs DoS fix
387 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
388 invalid signature algorithms extension a NULL pointer dereference will
389 occur. This can be exploited in a DoS attack against the server.
391 This issue was was reported to OpenSSL by David Ramos of Stanford
394 [Stephen Henson and Matt Caswell]
396 *) Multiblock corrupted pointer fix
398 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
399 feature only applies on 64 bit x86 architecture platforms that support AES
400 NI instructions. A defect in the implementation of "multiblock" can cause
401 OpenSSL's internal write buffer to become incorrectly set to NULL when
402 using non-blocking IO. Typically, when the user application is using a
403 socket BIO for writing, this will only result in a failed connection.
404 However if some other BIO is used then it is likely that a segmentation
405 fault will be triggered, thus enabling a potential DoS attack.
407 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
411 *) Segmentation fault in DTLSv1_listen fix
413 The DTLSv1_listen function is intended to be stateless and processes the
414 initial ClientHello from many peers. It is common for user code to loop
415 over the call to DTLSv1_listen until a valid ClientHello is received with
416 an associated cookie. A defect in the implementation of DTLSv1_listen means
417 that state is preserved in the SSL object from one invocation to the next
418 that can lead to a segmentation fault. Errors processing the initial
419 ClientHello can trigger this scenario. An example of such an error could be
420 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
423 This issue was reported to OpenSSL by Per Allansson.
427 *) Segmentation fault in ASN1_TYPE_cmp fix
429 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
430 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
431 certificate signature algorithm consistency this can be used to crash any
432 certificate verification operation and exploited in a DoS attack. Any
433 application which performs certificate verification is vulnerable including
434 OpenSSL clients and servers which enable client authentication.
438 *) Segmentation fault for invalid PSS parameters fix
440 The signature verification routines will crash with a NULL pointer
441 dereference if presented with an ASN.1 signature using the RSA PSS
442 algorithm and invalid parameters. Since these routines are used to verify
443 certificate signature algorithms this can be used to crash any
444 certificate verification operation and exploited in a DoS attack. Any
445 application which performs certificate verification is vulnerable including
446 OpenSSL clients and servers which enable client authentication.
448 This issue was was reported to OpenSSL by Brian Carpenter.
452 *) ASN.1 structure reuse memory corruption fix
454 Reusing a structure in ASN.1 parsing may allow an attacker to cause
455 memory corruption via an invalid write. Such reuse is and has been
456 strongly discouraged and is believed to be rare.
458 Applications that parse structures containing CHOICE or ANY DEFINED BY
459 components may be affected. Certificate parsing (d2i_X509 and related
460 functions) are however not affected. OpenSSL clients and servers are
465 *) PKCS7 NULL pointer dereferences fix
467 The PKCS#7 parsing code does not handle missing outer ContentInfo
468 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
469 missing content and trigger a NULL pointer dereference on parsing.
471 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
472 otherwise parse PKCS#7 structures from untrusted sources are
473 affected. OpenSSL clients and servers are not affected.
475 This issue was reported to OpenSSL by Michal Zalewski (Google).
479 *) DoS via reachable assert in SSLv2 servers fix
481 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
482 servers that both support SSLv2 and enable export cipher suites by sending
483 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
485 This issue was discovered by Sean Burford (Google) and Emilia Käsper
486 (OpenSSL development team).
490 *) Empty CKE with client auth and DHE fix
492 If client auth is used then a server can seg fault in the event of a DHE
493 ciphersuite being selected and a zero length ClientKeyExchange message
494 being sent by the client. This could be exploited in a DoS attack.
498 *) Handshake with unseeded PRNG fix
500 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
501 with an unseeded PRNG. The conditions are:
502 - The client is on a platform where the PRNG has not been seeded
503 automatically, and the user has not seeded manually
504 - A protocol specific client method version has been used (i.e. not
505 SSL_client_methodv23)
506 - A ciphersuite is used that does not require additional random data from
507 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
509 If the handshake succeeds then the client random that has been used will
510 have been generated from a PRNG with insufficient entropy and therefore the
511 output may be predictable.
513 For example using the following command with an unseeded openssl will
514 succeed on an unpatched platform:
516 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
520 *) Use After Free following d2i_ECPrivatekey error fix
522 A malformed EC private key file consumed via the d2i_ECPrivateKey function
523 could cause a use after free condition. This, in turn, could cause a double
524 free in several private key parsing functions (such as d2i_PrivateKey
525 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
526 for applications that receive EC private keys from untrusted
527 sources. This scenario is considered rare.
529 This issue was discovered by the BoringSSL project and fixed in their
534 *) X509_to_X509_REQ NULL pointer deref fix
536 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
537 the certificate key is invalid. This function is rarely used in practice.
539 This issue was discovered by Brian Carpenter.
543 *) Removed the export ciphers from the DEFAULT ciphers
546 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
548 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
549 ARMv5 through ARMv8, as opposite to "locking" it to single one.
550 So far those who have to target multiple plaforms would compromise
551 and argue that binary targeting say ARMv5 would still execute on
552 ARMv8. "Universal" build resolves this compromise by providing
553 near-optimal performance even on newer platforms.
556 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
557 (other platforms pending).
558 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
560 *) Add support for the SignedCertificateTimestampList certificate and
561 OCSP response extensions from RFC6962.
564 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
565 for corner cases. (Certain input points at infinity could lead to
566 bogus results, with non-infinity inputs mapped to infinity too.)
569 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
570 This covers AES, SHA256/512 and GHASH. "Initial" means that most
571 common cases are optimized and there still is room for further
572 improvements. Vector Permutation AES for Altivec is also added.
575 *) Add support for little-endian ppc64 Linux target.
576 [Marcelo Cerri (IBM)]
578 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
579 SHA1, SHA256 and GHASH. "Initial" means that most common cases
580 are optimized and there still is room for further improvements.
581 Both 32- and 64-bit modes are supported.
582 [Andy Polyakov, Ard Biesheuvel (Linaro)]
584 *) Improved ARMv7 NEON support.
587 *) Support for SPARC Architecture 2011 crypto extensions, first
588 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
589 SHA256/512, MD5, GHASH and modular exponentiation.
590 [Andy Polyakov, David Miller]
592 *) Accelerated modular exponentiation for Intel processors, a.k.a.
594 [Shay Gueron & Vlad Krasnov (Intel Corp)]
596 *) Support for new and upcoming Intel processors, including AVX2,
597 BMI and SHA ISA extensions. This includes additional "stitched"
598 implementations, AESNI-SHA256 and GCM, and multi-buffer support
601 This work was sponsored by Intel Corp.
604 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
605 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
606 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
609 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
610 this fixes a limiation in previous versions of OpenSSL.
613 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
614 MGF1 digest and OAEP label.
617 *) Add EVP support for key wrapping algorithms, to avoid problems with
618 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
619 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
620 algorithms and include tests cases.
623 *) Add functions to allocate and set the fields of an ECDSA_METHOD
625 [Douglas E. Engert, Steve Henson]
627 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
628 difference in days and seconds between two tm or ASN1_TIME structures.
631 *) Add -rev test option to s_server to just reverse order of characters
632 received by client and send back to server. Also prints an abbreviated
633 summary of the connection parameters.
636 *) New option -brief for s_client and s_server to print out a brief summary
637 of connection parameters.
640 *) Add callbacks for arbitrary TLS extensions.
641 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
643 *) New option -crl_download in several openssl utilities to download CRLs
644 from CRLDP extension in certificates.
647 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
650 *) New function X509_CRL_diff to generate a delta CRL from the difference
651 of two full CRLs. Add support to "crl" utility.
654 *) New functions to set lookup_crls function and to retrieve
655 X509_STORE from X509_STORE_CTX.
658 *) Print out deprecated issuer and subject unique ID fields in
662 *) Extend OCSP I/O functions so they can be used for simple general purpose
663 HTTP as well as OCSP. New wrapper function which can be used to download
664 CRLs using the OCSP API.
667 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
670 *) SSL_CONF* functions. These provide a common framework for application
671 configuration using configuration files or command lines.
674 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
675 message callback and prints the results. Needs compile time option
676 "enable-ssl-trace". New options to s_client and s_server to enable
680 *) New ctrl and macro to retrieve supported points extensions.
681 Print out extension in s_server and s_client.
684 *) New functions to retrieve certificate signature and signature
688 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
692 *) New Suite B modes for TLS code. These use and enforce the requirements
693 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
694 only use Suite B curves. The Suite B modes can be set by using the
695 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
698 *) New chain verification flags for Suite B levels of security. Check
699 algorithms are acceptable when flags are set in X509_verify_cert.
702 *) Make tls1_check_chain return a set of flags indicating checks passed
703 by a certificate chain. Add additional tests to handle client
704 certificates: checks for matching certificate type and issuer name
708 *) If an attempt is made to use a signature algorithm not in the peer
709 preference list abort the handshake. If client has no suitable
710 signature algorithms in response to a certificate request do not
714 *) If server EC tmp key is not in client preference list abort handshake.
717 *) Add support for certificate stores in CERT structure. This makes it
718 possible to have different stores per SSL structure or one store in
719 the parent SSL_CTX. Include distint stores for certificate chain
720 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
721 to build and store a certificate chain in CERT structure: returing
722 an error if the chain cannot be built: this will allow applications
723 to test if a chain is correctly configured.
725 Note: if the CERT based stores are not set then the parent SSL_CTX
726 store is used to retain compatibility with existing behaviour.
730 *) New function ssl_set_client_disabled to set a ciphersuite disabled
731 mask based on the current session, check mask when sending client
732 hello and checking the requested ciphersuite.
735 *) New ctrls to retrieve and set certificate types in a certificate
736 request message. Print out received values in s_client. If certificate
737 types is not set with custom values set sensible values based on
738 supported signature algorithms.
741 *) Support for distinct client and server supported signature algorithms.
744 *) Add certificate callback. If set this is called whenever a certificate
745 is required by client or server. An application can decide which
746 certificate chain to present based on arbitrary criteria: for example
747 supported signature algorithms. Add very simple example to s_server.
748 This fixes many of the problems and restrictions of the existing client
749 certificate callback: for example you can now clear an existing
750 certificate and specify the whole chain.
753 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
754 the certificate can be used for (if anything). Set valid_flags field
755 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
756 to have similar checks in it.
758 Add new "cert_flags" field to CERT structure and include a "strict mode".
759 This enforces some TLS certificate requirements (such as only permitting
760 certificate signature algorithms contained in the supported algorithms
761 extension) which some implementations ignore: this option should be used
762 with caution as it could cause interoperability issues.
765 *) Update and tidy signature algorithm extension processing. Work out
766 shared signature algorithms based on preferences and peer algorithms
767 and print them out in s_client and s_server. Abort handshake if no
768 shared signature algorithms.
771 *) Add new functions to allow customised supported signature algorithms
772 for SSL and SSL_CTX structures. Add options to s_client and s_server
776 *) New function SSL_certs_clear() to delete all references to certificates
777 from an SSL structure. Before this once a certificate had been added
778 it couldn't be removed.
781 *) Integrate hostname, email address and IP address checking with certificate
782 verification. New verify options supporting checking in opensl utility.
785 *) Fixes and wildcard matching support to hostname and email checking
786 functions. Add manual page.
787 [Florian Weimer (Red Hat Product Security Team)]
789 *) New functions to check a hostname email or IP address against a
790 certificate. Add options x509 utility to print results of checks against
794 *) Fix OCSP checking.
795 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
797 *) Initial experimental support for explicitly trusted non-root CAs.
798 OpenSSL still tries to build a complete chain to a root but if an
799 intermediate CA has a trust setting included that is used. The first
800 setting is used: whether to trust (e.g., -addtrust option to the x509
804 *) Add -trusted_first option which attempts to find certificates in the
805 trusted store even if an untrusted chain is also supplied.
808 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
809 platform support for Linux and Android.
812 *) Support for linux-x32, ILP32 environment in x86_64 framework.
815 *) Experimental multi-implementation support for FIPS capable OpenSSL.
816 When in FIPS mode the approved implementations are used as normal,
817 when not in FIPS mode the internal unapproved versions are used instead.
818 This means that the FIPS capable OpenSSL isn't forced to use the
819 (often lower performance) FIPS implementations outside FIPS mode.
822 *) Transparently support X9.42 DH parameters when calling
823 PEM_read_bio_DHparameters. This means existing applications can handle
824 the new parameter format automatically.
827 *) Initial experimental support for X9.42 DH parameter format: mainly
828 to support use of 'q' parameter for RFC5114 parameters.
831 *) Add DH parameters from RFC5114 including test data to dhtest.
834 *) Support for automatic EC temporary key parameter selection. If enabled
835 the most preferred EC parameters are automatically used instead of
836 hardcoded fixed parameters. Now a server just has to call:
837 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
838 support ECDH and use the most appropriate parameters.
841 *) Enhance and tidy EC curve and point format TLS extension code. Use
842 static structures instead of allocation if default values are used.
843 New ctrls to set curves we wish to support and to retrieve shared curves.
844 Print out shared curves in s_server. New options to s_server and s_client
845 to set list of supported curves.
848 *) New ctrls to retrieve supported signature algorithms and
849 supported curve values as an array of NIDs. Extend openssl utility
850 to print out received values.
853 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
854 between NIDs and the more common NIST names such as "P-256". Enhance
855 ecparam utility and ECC method to recognise the NIST names for curves.
858 *) Enhance SSL/TLS certificate chain handling to support different
859 chains for each certificate instead of one chain in the parent SSL_CTX.
862 *) Support for fixed DH ciphersuite client authentication: where both
863 server and client use DH certificates with common parameters.
866 *) Support for fixed DH ciphersuites: those requiring DH server
870 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
872 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
873 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
874 X509_CINF_get_signature were reverted post internal team review.
876 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
878 *) Build fixes for the Windows and OpenVMS platforms
879 [Matt Caswell and Richard Levitte]
881 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
883 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
884 message can cause a segmentation fault in OpenSSL due to a NULL pointer
885 dereference. This could lead to a Denial Of Service attack. Thanks to
886 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
890 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
891 dtls1_buffer_record function under certain conditions. In particular this
892 could occur if an attacker sent repeated DTLS records with the same
893 sequence number but for the next epoch. The memory leak could be exploited
894 by an attacker in a Denial of Service attack through memory exhaustion.
895 Thanks to Chris Mueller for reporting this issue.
899 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
900 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
901 method would be set to NULL which could later result in a NULL pointer
902 dereference. Thanks to Frank Schmirler for reporting this issue.
906 *) Abort handshake if server key exchange message is omitted for ephemeral
909 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
910 reporting this issue.
914 *) Remove non-export ephemeral RSA code on client and server. This code
915 violated the TLS standard by allowing the use of temporary RSA keys in
916 non-export ciphersuites and could be used by a server to effectively
917 downgrade the RSA key length used to a value smaller than the server
918 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
919 INRIA or reporting this issue.
923 *) Fixed issue where DH client certificates are accepted without verification.
924 An OpenSSL server will accept a DH certificate for client authentication
925 without the certificate verify message. This effectively allows a client to
926 authenticate without the use of a private key. This only affects servers
927 which trust a client certificate authority which issues certificates
928 containing DH keys: these are extremely rare and hardly ever encountered.
929 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
934 *) Ensure that the session ID context of an SSL is updated when its
935 SSL_CTX is updated via SSL_set_SSL_CTX.
937 The session ID context is typically set from the parent SSL_CTX,
938 and can vary with the CTX.
941 *) Fix various certificate fingerprint issues.
943 By using non-DER or invalid encodings outside the signed portion of a
944 certificate the fingerprint can be changed without breaking the signature.
945 Although no details of the signed portion of the certificate can be changed
946 this can cause problems with some applications: e.g. those using the
947 certificate fingerprint for blacklists.
949 1. Reject signatures with non zero unused bits.
951 If the BIT STRING containing the signature has non zero unused bits reject
952 the signature. All current signature algorithms require zero unused bits.
954 2. Check certificate algorithm consistency.
956 Check the AlgorithmIdentifier inside TBS matches the one in the
957 certificate signature. NB: this will result in signature failure
958 errors for some broken certificates.
960 Thanks to Konrad Kraszewski from Google for reporting this issue.
962 3. Check DSA/ECDSA signatures use DER.
964 Reencode DSA/ECDSA signatures and compare with the original received
965 signature. Return an error if there is a mismatch.
967 This will reject various cases including garbage after signature
968 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
969 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
970 (negative or with leading zeroes).
972 Further analysis was conducted and fixes were developed by Stephen Henson
973 of the OpenSSL core team.
978 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
979 results on some platforms, including x86_64. This bug occurs at random
980 with a very low probability, and is not known to be exploitable in any
981 way, though its exact impact is difficult to determine. Thanks to Pieter
982 Wuille (Blockstream) who reported this issue and also suggested an initial
983 fix. Further analysis was conducted by the OpenSSL development team and
984 Adam Langley of Google. The final fix was developed by Andy Polyakov of
985 the OpenSSL core team.
989 *) Do not resume sessions on the server if the negotiated protocol
990 version does not match the session's version. Resuming with a different
991 version, while not strictly forbidden by the RFC, is of questionable
992 sanity and breaks all known clients.
993 [David Benjamin, Emilia Käsper]
995 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
996 early CCS messages during renegotiation. (Note that because
997 renegotiation is encrypted, this early CCS was not exploitable.)
1000 *) Tighten client-side session ticket handling during renegotiation:
1001 ensure that the client only accepts a session ticket if the server sends
1002 the extension anew in the ServerHello. Previously, a TLS client would
1003 reuse the old extension state and thus accept a session ticket if one was
1004 announced in the initial ServerHello.
1006 Similarly, ensure that the client requires a session ticket if one
1007 was advertised in the ServerHello. Previously, a TLS client would
1008 ignore a missing NewSessionTicket message.
1011 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1013 *) SRTP Memory Leak.
1015 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1016 sends a carefully crafted handshake message, to cause OpenSSL to fail
1017 to free up to 64k of memory causing a memory leak. This could be
1018 exploited in a Denial Of Service attack. This issue affects OpenSSL
1019 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1020 whether SRTP is used or configured. Implementations of OpenSSL that
1021 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1023 The fix was developed by the OpenSSL team.
1027 *) Session Ticket Memory Leak.
1029 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1030 integrity of that ticket is first verified. In the event of a session
1031 ticket integrity check failing, OpenSSL will fail to free memory
1032 causing a memory leak. By sending a large number of invalid session
1033 tickets an attacker could exploit this issue in a Denial Of Service
1038 *) Build option no-ssl3 is incomplete.
1040 When OpenSSL is configured with "no-ssl3" as a build option, servers
1041 could accept and complete a SSL 3.0 handshake, and clients could be
1042 configured to send them.
1044 [Akamai and the OpenSSL team]
1046 *) Add support for TLS_FALLBACK_SCSV.
1047 Client applications doing fallback retries should call
1048 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1050 [Adam Langley, Bodo Moeller]
1052 *) Add additional DigestInfo checks.
1054 Reencode DigestInto in DER and check against the original when
1055 verifying RSA signature: this will reject any improperly encoded
1056 DigestInfo structures.
1058 Note: this is a precautionary measure and no attacks are currently known.
1062 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1064 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1065 SRP code can be overrun an internal buffer. Add sanity check that
1066 g, A, B < N to SRP code.
1068 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1069 Group for discovering this issue.
1073 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1074 TLS 1.0 instead of higher protocol versions when the ClientHello message
1075 is badly fragmented. This allows a man-in-the-middle attacker to force a
1076 downgrade to TLS 1.0 even if both the server and the client support a
1077 higher protocol version, by modifying the client's TLS records.
1079 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1080 researching this issue.
1084 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1085 to a denial of service attack. A malicious server can crash the client
1086 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1087 ciphersuite and sending carefully crafted handshake messages.
1089 Thanks to Felix Gröbert (Google) for discovering and researching this
1094 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1095 to leak memory. This can be exploited through a Denial of Service attack.
1096 Thanks to Adam Langley for discovering and researching this issue.
1100 *) An attacker can force openssl to consume large amounts of memory whilst
1101 processing DTLS handshake messages. This can be exploited through a
1102 Denial of Service attack.
1103 Thanks to Adam Langley for discovering and researching this issue.
1107 *) An attacker can force an error condition which causes openssl to crash
1108 whilst processing DTLS packets due to memory being freed twice. This
1109 can be exploited through a Denial of Service attack.
1110 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1115 *) If a multithreaded client connects to a malicious server using a resumed
1116 session and the server sends an ec point format extension it could write
1117 up to 255 bytes to freed memory.
1119 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1124 *) A malicious server can crash an OpenSSL client with a null pointer
1125 dereference (read) by specifying an SRP ciphersuite even though it was not
1126 properly negotiated with the client. This can be exploited through a
1127 Denial of Service attack.
1129 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1130 discovering and researching this issue.
1134 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1135 X509_name_oneline, X509_name_print_ex et al. to leak some information
1136 from the stack. Applications may be affected if they echo pretty printing
1137 output to the attacker.
1139 Thanks to Ivan Fratric (Google) for discovering this issue.
1141 [Emilia Käsper, and Steve Henson]
1143 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1144 for corner cases. (Certain input points at infinity could lead to
1145 bogus results, with non-infinity inputs mapped to infinity too.)
1148 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1150 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1151 handshake can force the use of weak keying material in OpenSSL
1152 SSL/TLS clients and servers.
1154 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1155 researching this issue. (CVE-2014-0224)
1156 [KIKUCHI Masashi, Steve Henson]
1158 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1159 OpenSSL DTLS client the code can be made to recurse eventually crashing
1162 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1164 [Imre Rad, Steve Henson]
1166 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1167 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1168 client or server. This is potentially exploitable to run arbitrary
1169 code on a vulnerable client or server.
1171 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1172 [Jüri Aedla, Steve Henson]
1174 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1175 are subject to a denial of service attack.
1177 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1178 this issue. (CVE-2014-3470)
1179 [Felix Gröbert, Ivan Fratric, Steve Henson]
1181 *) Harmonize version and its documentation. -f flag is used to display
1183 [mancha <mancha1@zoho.com>]
1185 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1186 in i2d_ECPrivateKey.
1187 [mancha <mancha1@zoho.com>]
1189 *) Fix some double frees. These are not thought to be exploitable.
1190 [mancha <mancha1@zoho.com>]
1192 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1194 *) A missing bounds check in the handling of the TLS heartbeat extension
1195 can be used to reveal up to 64k of memory to a connected client or
1198 Thanks for Neel Mehta of Google Security for discovering this bug and to
1199 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1200 preparing the fix (CVE-2014-0160)
1201 [Adam Langley, Bodo Moeller]
1203 *) Fix for the attack described in the paper "Recovering OpenSSL
1204 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1205 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1206 http://eprint.iacr.org/2014/140
1208 Thanks to Yuval Yarom and Naomi Benger for discovering this
1209 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1210 [Yuval Yarom and Naomi Benger]
1212 *) TLS pad extension: draft-agl-tls-padding-03
1214 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1215 TLS client Hello record length value would otherwise be > 255 and
1216 less that 512 pad with a dummy extension containing zeroes so it
1217 is at least 512 bytes long.
1219 [Adam Langley, Steve Henson]
1221 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1223 *) Fix for TLS record tampering bug. A carefully crafted invalid
1224 handshake could crash OpenSSL with a NULL pointer exception.
1225 Thanks to Anton Johansson for reporting this issues.
1228 *) Keep original DTLS digest and encryption contexts in retransmission
1229 structures so we can use the previous session parameters if they need
1230 to be resent. (CVE-2013-6450)
1233 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1234 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1235 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1236 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1237 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1238 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1239 [Rob Stradling, Adam Langley]
1241 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1243 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1244 supporting platforms or when small records were transferred.
1245 [Andy Polyakov, Steve Henson]
1247 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1249 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1251 This addresses the flaw in CBC record processing discovered by
1252 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1253 at: http://www.isg.rhul.ac.uk/tls/
1255 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1256 Security Group at Royal Holloway, University of London
1257 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1258 Emilia Käsper for the initial patch.
1260 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1262 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1263 ciphersuites which can be exploited in a denial of service attack.
1264 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1265 and detecting this bug and to Wolfgang Ettlinger
1266 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1270 *) Return an error when checking OCSP signatures when key is NULL.
1271 This fixes a DoS attack. (CVE-2013-0166)
1274 *) Make openssl verify return errors.
1275 [Chris Palmer <palmer@google.com> and Ben Laurie]
1277 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1278 the right response is stapled. Also change SSL_get_certificate()
1279 so it returns the certificate actually sent.
1280 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1281 [Rob Stradling <rob.stradling@comodo.com>]
1283 *) Fix possible deadlock when decoding public keys.
1286 *) Don't use TLS 1.0 record version number in initial client hello
1290 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1292 *) Sanity check record length before skipping explicit IV in TLS
1293 1.2, 1.1 and DTLS to fix DoS attack.
1295 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1296 fuzzing as a service testing platform.
1300 *) Initialise tkeylen properly when encrypting CMS messages.
1301 Thanks to Solar Designer of Openwall for reporting this issue.
1304 *) In FIPS mode don't try to use composite ciphers as they are not
1308 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1310 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1311 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1312 mean any application compiled against OpenSSL 1.0.0 headers setting
1313 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1314 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1315 0x10000000L Any application which was previously compiled against
1316 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1317 will need to be recompiled as a result. Letting be results in
1318 inability to disable specifically TLS 1.1 and in client context,
1319 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1322 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1323 disable just protocol X, but all protocols above X *if* there are
1324 protocols *below* X still enabled. In more practical terms it means
1325 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1326 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1327 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1331 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1333 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1334 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1335 in CRYPTO_realloc_clean.
1337 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1338 issue and to Adam Langley <agl@chromium.org> for fixing it.
1340 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1342 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1345 *) Workarounds for some broken servers that "hang" if a client hello
1346 record length exceeds 255 bytes.
1348 1. Do not use record version number > TLS 1.0 in initial client
1349 hello: some (but not all) hanging servers will now work.
1350 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1351 the number of ciphers sent in the client hello. This should be
1352 set to an even number, such as 50, for example by passing:
1353 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1354 Most broken servers should now work.
1355 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1356 TLS 1.2 client support entirely.
1359 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1362 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1364 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1365 STRING form instead of a DigestInfo.
1368 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1369 and the RSA_sign/RSA_verify functions. This was made more apparent when
1370 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1371 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1372 the correct format in RSA_verify so both forms transparently work.
1375 *) Some servers which support TLS 1.0 can choke if we initially indicate
1376 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1377 encrypted premaster secret. As a workaround use the maximum permitted
1378 client version in client hello, this should keep such servers happy
1379 and still work with previous versions of OpenSSL.
1382 *) Add support for TLS/DTLS heartbeats.
1383 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1385 *) Add support for SCTP.
1386 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1388 *) Improved PRNG seeding for VOS.
1389 [Paul Green <Paul.Green@stratus.com>]
1391 *) Extensive assembler packs updates, most notably:
1393 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1394 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1395 - x86_64: bit-sliced AES implementation;
1396 - ARM: NEON support, contemporary platforms optimizations;
1397 - s390x: z196 support;
1398 - *: GHASH and GF(2^m) multiplication implementations;
1402 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1403 (removal of unnecessary code)
1404 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1406 *) Add TLS key material exporter from RFC 5705.
1409 *) Add DTLS-SRTP negotiation from RFC 5764.
1412 *) Add Next Protocol Negotiation,
1413 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1414 disabled with a no-npn flag to config or Configure. Code donated
1416 [Adam Langley <agl@google.com> and Ben Laurie]
1418 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1419 NIST-P256, NIST-P521, with constant-time single point multiplication on
1420 typical inputs. Compiler support for the nonstandard type __uint128_t is
1421 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1422 Code made available under Apache License version 2.0.
1424 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1425 line to include this in your build of OpenSSL, and run "make depend" (or
1426 "make update"). This enables the following EC_METHODs:
1428 EC_GFp_nistp224_method()
1429 EC_GFp_nistp256_method()
1430 EC_GFp_nistp521_method()
1432 EC_GROUP_new_by_curve_name() will automatically use these (while
1433 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1435 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1437 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1438 all platforms. Move ssize_t definition from e_os.h to the public
1439 header file e_os2.h as it now appears in public header file cms.h
1442 *) New -sigopt option to the ca, req and x509 utilities. Additional
1443 signature parameters can be passed using this option and in
1447 *) Add RSA PSS signing function. This will generate and set the
1448 appropriate AlgorithmIdentifiers for PSS based on those in the
1449 corresponding EVP_MD_CTX structure. No application support yet.
1452 *) Support for companion algorithm specific ASN1 signing routines.
1453 New function ASN1_item_sign_ctx() signs a pre-initialised
1454 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1455 the appropriate parameters.
1458 *) Add new algorithm specific ASN1 verification initialisation function
1459 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1460 handling will be the same no matter what EVP_PKEY_METHOD is used.
1461 Add a PSS handler to support verification of PSS signatures: checked
1462 against a number of sample certificates.
1465 *) Add signature printing for PSS. Add PSS OIDs.
1466 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1468 *) Add algorithm specific signature printing. An individual ASN1 method
1469 can now print out signatures instead of the standard hex dump.
1471 More complex signatures (e.g. PSS) can print out more meaningful
1472 information. Include DSA version that prints out the signature
1476 *) Password based recipient info support for CMS library: implementing
1480 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1481 neatly separates the code into cipher and PBE sections and is required
1482 for some algorithms that split PBES2 into separate pieces (such as
1483 password based CMS).
1486 *) Session-handling fixes:
1487 - Fix handling of connections that are resuming with a session ID,
1488 but also support Session Tickets.
1489 - Fix a bug that suppressed issuing of a new ticket if the client
1490 presented a ticket with an expired session.
1491 - Try to set the ticket lifetime hint to something reasonable.
1492 - Make tickets shorter by excluding irrelevant information.
1493 - On the client side, don't ignore renewed tickets.
1494 [Adam Langley, Bodo Moeller (Google)]
1496 *) Fix PSK session representation.
1499 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1501 This work was sponsored by Intel.
1504 *) Add GCM support to TLS library. Some custom code is needed to split
1505 the IV between the fixed (from PRF) and explicit (from TLS record)
1506 portions. This adds all GCM ciphersuites supported by RFC5288 and
1507 RFC5289. Generalise some AES* cipherstrings to include GCM and
1508 add a special AESGCM string for GCM only.
1511 *) Expand range of ctrls for AES GCM. Permit setting invocation
1512 field on decrypt and retrieval of invocation field only on encrypt.
1515 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1516 As required by RFC5289 these ciphersuites cannot be used if for
1517 versions of TLS earlier than 1.2.
1520 *) For FIPS capable OpenSSL interpret a NULL default public key method
1521 as unset and return the appropriate default but do *not* set the default.
1522 This means we can return the appropriate method in applications that
1523 switch between FIPS and non-FIPS modes.
1526 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1527 ENGINE is used then we cannot handle that in the FIPS module so we
1528 keep original code iff non-FIPS operations are allowed.
1531 *) Add -attime option to openssl utilities.
1532 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1534 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1537 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1538 FIPS EC methods unconditionally for now.
1541 *) New build option no-ec2m to disable characteristic 2 code.
1544 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1545 all cases can be covered as some introduce binary incompatibilities.
1548 *) Redirect RSA operations to FIPS module including keygen,
1549 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1552 *) Add similar low level API blocking to ciphers.
1555 *) Low level digest APIs are not approved in FIPS mode: any attempt
1556 to use these will cause a fatal error. Applications that *really* want
1557 to use them can use the private_* version instead.
1560 *) Redirect cipher operations to FIPS module for FIPS builds.
1563 *) Redirect digest operations to FIPS module for FIPS builds.
1566 *) Update build system to add "fips" flag which will link in fipscanister.o
1567 for static and shared library builds embedding a signature if needed.
1570 *) Output TLS supported curves in preference order instead of numerical
1571 order. This is currently hardcoded for the highest order curves first.
1572 This should be configurable so applications can judge speed vs strength.
1575 *) Add TLS v1.2 server support for client authentication.
1578 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1582 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1583 FIPS modules versions.
1586 *) Add TLS v1.2 client side support for client authentication. Keep cache
1587 of handshake records longer as we don't know the hash algorithm to use
1588 until after the certificate request message is received.
1591 *) Initial TLS v1.2 client support. Add a default signature algorithms
1592 extension including all the algorithms we support. Parse new signature
1593 format in client key exchange. Relax some ECC signing restrictions for
1594 TLS v1.2 as indicated in RFC5246.
1597 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1598 to new signature format when needed using client digest preference.
1599 All server ciphersuites should now work correctly in TLS v1.2. No client
1600 support yet and no support for client certificates.
1603 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1604 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1605 ciphersuites. At present only RSA key exchange ciphersuites work with
1606 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1607 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1608 and version checking.
1611 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1612 with this defined it will not be affected by any changes to ssl internal
1613 structures. Add several utility functions to allow openssl application
1614 to work with OPENSSL_NO_SSL_INTERN defined.
1618 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1620 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1623 *) Permit abbreviated handshakes when renegotiating using the function
1624 SSL_renegotiate_abbreviated().
1625 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1627 *) Add call to ENGINE_register_all_complete() to
1628 ENGINE_load_builtin_engines(), so some implementations get used
1629 automatically instead of needing explicit application support.
1632 *) Add support for TLS key exporter as described in RFC5705.
1633 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1635 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1636 a few changes are required:
1638 Add SSL_OP_NO_TLSv1_1 flag.
1639 Add TLSv1_1 methods.
1640 Update version checking logic to handle version 1.1.
1641 Add explicit IV handling (ported from DTLS code).
1642 Add command line options to s_client/s_server.
1645 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1647 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1648 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1649 content decryption and always return the same error. Note: this attack
1650 needs on average 2^20 messages so it only affects automated senders. The
1651 old behaviour can be reenabled in the CMS code by setting the
1652 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1653 an MMA defence is not necessary.
1654 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1655 this issue. (CVE-2012-0884)
1658 *) Fix CVE-2011-4619: make sure we really are receiving a
1659 client hello before rejecting multiple SGC restarts. Thanks to
1660 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1663 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1665 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1666 Thanks to Antonio Martin, Enterprise Secure Access Research and
1667 Development, Cisco Systems, Inc. for discovering this bug and
1668 preparing a fix. (CVE-2012-0050)
1671 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1673 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1674 of the Vaudenay padding oracle attack on CBC mode encryption
1675 which enables an efficient plaintext recovery attack against
1676 the OpenSSL implementation of DTLS. Their attack exploits timing
1677 differences arising during decryption processing. A research
1678 paper describing this attack can be found at:
1679 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1680 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1681 Security Group at Royal Holloway, University of London
1682 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1683 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1684 for preparing the fix. (CVE-2011-4108)
1685 [Robin Seggelmann, Michael Tuexen]
1687 *) Clear bytes used for block padding of SSL 3.0 records.
1689 [Adam Langley (Google)]
1691 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1692 Kadianakis <desnacked@gmail.com> for discovering this issue and
1693 Adam Langley for preparing the fix. (CVE-2011-4619)
1694 [Adam Langley (Google)]
1696 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1697 [Andrey Kulikov <amdeich@gmail.com>]
1699 *) Prevent malformed RFC3779 data triggering an assertion failure.
1700 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1701 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1702 [Rob Austein <sra@hactrn.net>]
1704 *) Improved PRNG seeding for VOS.
1705 [Paul Green <Paul.Green@stratus.com>]
1707 *) Fix ssl_ciph.c set-up race.
1708 [Adam Langley (Google)]
1710 *) Fix spurious failures in ecdsatest.c.
1711 [Emilia Käsper (Google)]
1713 *) Fix the BIO_f_buffer() implementation (which was mixing different
1714 interpretations of the '..._len' fields).
1715 [Adam Langley (Google)]
1717 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1718 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1719 threads won't reuse the same blinding coefficients.
1721 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1722 lock to call BN_BLINDING_invert_ex, and avoids one use of
1723 BN_BLINDING_update for each BN_BLINDING structure (previously,
1724 the last update always remained unused).
1725 [Emilia Käsper (Google)]
1727 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1728 [Bob Buckholz (Google)]
1730 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1732 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1733 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1734 [Kaspar Brand <ossl@velox.ch>]
1736 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1737 for multi-threaded use of ECDH. (CVE-2011-3210)
1738 [Adam Langley (Google)]
1740 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1743 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1744 signature public key algorithm by using OID xref utilities instead.
1745 Before this you could only use some ECC ciphersuites with SHA1 only.
1748 *) Add protection against ECDSA timing attacks as mentioned in the paper
1749 by Billy Bob Brumley and Nicola Tuveri, see:
1751 http://eprint.iacr.org/2011/232.pdf
1753 [Billy Bob Brumley and Nicola Tuveri]
1755 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1757 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1758 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1760 *) Fix bug in string printing code: if *any* escaping is enabled we must
1761 escape the escape character (backslash) or the resulting string is
1765 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1767 *) Disable code workaround for ancient and obsolete Netscape browsers
1768 and servers: an attacker can use it in a ciphersuite downgrade attack.
1769 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1772 *) Fixed J-PAKE implementation error, originally discovered by
1773 Sebastien Martini, further info and confirmation from Stefan
1774 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1777 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1779 *) Fix extension code to avoid race conditions which can result in a buffer
1780 overrun vulnerability: resumed sessions must not be modified as they can
1781 be shared by multiple threads. CVE-2010-3864
1784 *) Fix WIN32 build system to correctly link an ENGINE directory into
1788 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1790 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1792 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1794 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1796 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1797 context. The operation can be customised via the ctrl mechanism in
1798 case ENGINEs want to include additional functionality.
1801 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1804 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1805 output hashes compatible with older versions of OpenSSL.
1806 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1808 *) Fix compression algorithm handling: if resuming a session use the
1809 compression algorithm of the resumed session instead of determining
1810 it from client hello again. Don't allow server to change algorithm.
1813 *) Add load_crls() function to apps tidying load_certs() too. Add option
1814 to verify utility to allow additional CRLs to be included.
1817 *) Update OCSP request code to permit adding custom headers to the request:
1818 some responders need this.
1821 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1823 [Julia Lawall <julia@diku.dk>]
1825 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1826 needlessly dereferenced structures, used obsolete functions and
1827 didn't handle all updated verify codes correctly.
1830 *) Disable MD2 in the default configuration.
1833 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1834 indicate the initial BIO being pushed or popped. This makes it possible
1835 to determine whether the BIO is the one explicitly called or as a result
1836 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1837 it handles reference counts correctly and doesn't zero out the I/O bio
1838 when it is not being explicitly popped. WARNING: applications which
1839 included workarounds for the old buggy behaviour will need to be modified
1840 or they could free up already freed BIOs.
1843 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1844 renaming to all platforms (within the 0.9.8 branch, this was
1845 done conditionally on Netware platforms to avoid a name clash).
1846 [Guenter <lists@gknw.net>]
1848 *) Add ECDHE and PSK support to DTLS.
1849 [Michael Tuexen <tuexen@fh-muenster.de>]
1851 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1855 *) Add "missing" function EVP_MD_flags() (without this the only way to
1856 retrieve a digest flags is by accessing the structure directly. Update
1857 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1858 or cipher is registered as in the "from" argument. Print out all
1859 registered digests in the dgst usage message instead of manually
1860 attempting to work them out.
1863 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1864 this allows the use of compression and extensions. Change default cipher
1865 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1866 by default unless an application cipher string requests it.
1869 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1870 key ids to find matching certificates and keys but some PKCS#12 files
1871 don't follow the (somewhat unwritten) rules and this strategy fails.
1872 Now just gather all certificates together and the first private key
1873 then look for the first certificate that matches the key.
1876 *) Support use of registered digest and cipher names for dgst and cipher
1877 commands instead of having to add each one as a special case. So now
1884 openssl dgst -sha256 foo
1886 and this works for ENGINE based algorithms too.
1890 *) Update Gost ENGINE to support parameter files.
1891 [Victor B. Wagner <vitus@cryptocom.ru>]
1893 *) Support GeneralizedTime in ca utility.
1894 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1896 *) Enhance the hash format used for certificate directory links. The new
1897 form uses the canonical encoding (meaning equivalent names will work
1898 even if they aren't identical) and uses SHA1 instead of MD5. This form
1899 is incompatible with the older format and as a result c_rehash should
1900 be used to rebuild symbolic links.
1903 *) Make PKCS#8 the default write format for private keys, replacing the
1904 traditional format. This form is standardised, more secure and doesn't
1905 include an implicit MD5 dependency.
1908 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1909 committed to OpenSSL should pass this lot as a minimum.
1912 *) Add session ticket override functionality for use by EAP-FAST.
1913 [Jouni Malinen <j@w1.fi>]
1915 *) Modify HMAC functions to return a value. Since these can be implemented
1916 in an ENGINE errors can occur.
1919 *) Type-checked OBJ_bsearch_ex.
1922 *) Type-checked OBJ_bsearch. Also some constification necessitated
1923 by type-checking. Still to come: TXT_DB, bsearch(?),
1924 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1928 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1929 seconds to a tm structure directly, instead of going through OS
1930 specific date routines. This avoids any issues with OS routines such
1931 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1932 and X509_time_adj_ex() to cover the extended range. The existing
1933 X509_time_adj() is still usable and will no longer have any date issues.
1936 *) Delta CRL support. New use deltas option which will attempt to locate
1937 and search any appropriate delta CRLs available.
1939 This work was sponsored by Google.
1942 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1943 code and add additional score elements. Validate alternate CRL paths
1944 as part of the CRL checking and indicate a new error "CRL path validation
1945 error" in this case. Applications wanting additional details can use
1946 the verify callback and check the new "parent" field. If this is not
1947 NULL CRL path validation is taking place. Existing applications wont
1948 see this because it requires extended CRL support which is off by
1951 This work was sponsored by Google.
1954 *) Support for freshest CRL extension.
1956 This work was sponsored by Google.
1959 *) Initial indirect CRL support. Currently only supported in the CRLs
1960 passed directly and not via lookup. Process certificate issuer
1961 CRL entry extension and lookup CRL entries by bother issuer name
1962 and serial number. Check and process CRL issuer entry in IDP extension.
1964 This work was sponsored by Google.
1967 *) Add support for distinct certificate and CRL paths. The CRL issuer
1968 certificate is validated separately in this case. Only enabled if
1969 an extended CRL support flag is set: this flag will enable additional
1970 CRL functionality in future.
1972 This work was sponsored by Google.
1975 *) Add support for policy mappings extension.
1977 This work was sponsored by Google.
1980 *) Fixes to pathlength constraint, self issued certificate handling,
1981 policy processing to align with RFC3280 and PKITS tests.
1983 This work was sponsored by Google.
1986 *) Support for name constraints certificate extension. DN, email, DNS
1987 and URI types are currently supported.
1989 This work was sponsored by Google.
1992 *) To cater for systems that provide a pointer-based thread ID rather
1993 than numeric, deprecate the current numeric thread ID mechanism and
1994 replace it with a structure and associated callback type. This
1995 mechanism allows a numeric "hash" to be extracted from a thread ID in
1996 either case, and on platforms where pointers are larger than 'long',
1997 mixing is done to help ensure the numeric 'hash' is usable even if it
1998 can't be guaranteed unique. The default mechanism is to use "&errno"
1999 as a pointer-based thread ID to distinguish between threads.
2001 Applications that want to provide their own thread IDs should now use
2002 CRYPTO_THREADID_set_callback() to register a callback that will call
2003 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2005 Note that ERR_remove_state() is now deprecated, because it is tied
2006 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2007 to free the current thread's error state should be replaced by
2008 ERR_remove_thread_state(NULL).
2010 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2011 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2012 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2013 application was previously providing a numeric thread callback that
2014 was inappropriate for distinguishing threads, then uniqueness might
2015 have been obtained with &errno that happened immediately in the
2016 intermediate development versions of OpenSSL; this is no longer the
2017 case, the numeric thread callback will now override the automatic use
2019 [Geoff Thorpe, with help from Bodo Moeller]
2021 *) Initial support for different CRL issuing certificates. This covers a
2022 simple case where the self issued certificates in the chain exist and
2023 the real CRL issuer is higher in the existing chain.
2025 This work was sponsored by Google.
2028 *) Removed effectively defunct crypto/store from the build.
2031 *) Revamp of STACK to provide stronger type-checking. Still to come:
2032 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2033 ASN1_STRING, CONF_VALUE.
2036 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2037 RAM on SSL connections. This option can save about 34k per idle SSL.
2040 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2041 STACK, TXT_DB, bsearch, qsort.
2044 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2045 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2046 support for data, signedData, compressedData, digestedData and
2047 encryptedData, envelopedData types included. Scripts to check against
2048 RFC4134 examples draft and interop and consistency checks of many
2049 content types and variants.
2052 *) Add options to enc utility to support use of zlib compression BIO.
2055 *) Extend mk1mf to support importing of options and assembly language
2056 files from Configure script, currently only included in VC-WIN32.
2057 The assembly language rules can now optionally generate the source
2058 files from the associated perl scripts.
2061 *) Implement remaining functionality needed to support GOST ciphersuites.
2062 Interop testing has been performed using CryptoPro implementations.
2063 [Victor B. Wagner <vitus@cryptocom.ru>]
2065 *) s390x assembler pack.
2068 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2072 *) Implement Opaque PRF Input TLS extension as specified in
2073 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2074 official specification yet and no extension type assignment by
2075 IANA exists, this extension (for now) will have to be explicitly
2076 enabled when building OpenSSL by providing the extension number
2077 to use. For example, specify an option
2079 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2081 to the "config" or "Configure" script to enable the extension,
2082 assuming extension number 0x9527 (which is a completely arbitrary
2083 and unofficial assignment based on the MD5 hash of the Internet
2084 Draft). Note that by doing so, you potentially lose
2085 interoperability with other TLS implementations since these might
2086 be using the same extension number for other purposes.
2088 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2089 opaque PRF input value to use in the handshake. This will create
2090 an interal copy of the length-'len' string at 'src', and will
2091 return non-zero for success.
2093 To get more control and flexibility, provide a callback function
2096 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2097 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2101 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2104 Callback function 'cb' will be called in handshakes, and is
2105 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2106 Argument 'arg' is for application purposes (the value as given to
2107 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2108 be provided to the callback function). The callback function
2109 has to return non-zero to report success: usually 1 to use opaque
2110 PRF input just if possible, or 2 to enforce use of the opaque PRF
2111 input. In the latter case, the library will abort the handshake
2112 if opaque PRF input is not successfully negotiated.
2114 Arguments 'peerinput' and 'len' given to the callback function
2115 will always be NULL and 0 in the case of a client. A server will
2116 see the client's opaque PRF input through these variables if
2117 available (NULL and 0 otherwise). Note that if the server
2118 provides an opaque PRF input, the length must be the same as the
2119 length of the client's opaque PRF input.
2121 Note that the callback function will only be called when creating
2122 a new session (session resumption can resume whatever was
2123 previously negotiated), and will not be called in SSL 2.0
2124 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2125 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2126 for applications that need to enforce opaque PRF input.
2130 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2133 [Victor B. Wagner <vitus@cryptocom.ru>]
2135 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2136 RFC4507bis. The encrypted ticket format is an encrypted encoded
2137 SSL_SESSION structure, that way new session features are automatically
2140 If a client application caches session in an SSL_SESSION structure
2141 support is transparent because tickets are now stored in the encoded
2144 The SSL_CTX structure automatically generates keys for ticket
2145 protection in servers so again support should be possible
2146 with no application modification.
2148 If a client or server wishes to disable RFC4507 support then the option
2149 SSL_OP_NO_TICKET can be set.
2151 Add a TLS extension debugging callback to allow the contents of any client
2152 or server extensions to be examined.
2154 This work was sponsored by Google.
2157 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2158 OpenSSL should now compile cleanly on gcc 4.2
2159 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2161 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2162 support including streaming MAC support: this is required for GOST
2163 ciphersuite support.
2164 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2166 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2167 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2168 to output in BER and PEM format.
2171 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2172 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2173 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2174 ENGINE support for HMAC keys which are unextractable. New -mac and
2175 -macopt options to dgst utility.
2178 *) New option -sigopt to dgst utility. Update dgst to use
2179 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2180 alternative signing paramaters such as X9.31 or PSS in the dgst
2184 *) Change ssl_cipher_apply_rule(), the internal function that does
2185 the work each time a ciphersuite string requests enabling
2186 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2187 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2188 the order of disabled ciphersuites such that those ciphersuites
2189 that most recently went from enabled to disabled not only stay
2190 in order with respect to each other, but also have higher priority
2191 than other disabled ciphersuites the next time ciphersuites are
2194 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2195 the same ciphersuites as with "HIGH" alone, but in a specific
2196 order where the PSK ciphersuites come first (since they are the
2197 most recently disabled ciphersuites when "HIGH" is parsed).
2199 Also, change ssl_create_cipher_list() (using this new
2200 funcionality) such that between otherwise identical
2201 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2205 *) Change ssl_create_cipher_list() so that it automatically
2206 arranges the ciphersuites in reasonable order before starting
2207 to process the rule string. Thus, the definition for "DEFAULT"
2208 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2209 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2210 This makes it much easier to arrive at a reasonable default order
2211 in applications for which anonymous ciphers are OK (meaning
2212 that you can't actually use DEFAULT).
2213 [Bodo Moeller; suggested by Victor Duchovni]
2215 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2216 processing) into multiple integers instead of setting
2217 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2218 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2219 (These masks as well as the individual bit definitions are hidden
2220 away into the non-exported interface ssl/ssl_locl.h, so this
2221 change to the definition of the SSL_CIPHER structure shouldn't
2222 affect applications.) This give us more bits for each of these
2223 categories, so there is no longer a need to coagulate AES128 and
2224 AES256 into a single algorithm bit, and to coagulate Camellia128
2225 and Camellia256 into a single algorithm bit, which has led to all
2228 Thus, among other things, the kludge introduced in 0.9.7m and
2229 0.9.8e for masking out AES256 independently of AES128 or masking
2230 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2232 With the change, we also introduce new ciphersuite aliases that
2233 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2237 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2238 Use the leftmost N bytes of the signature input if the input is
2239 larger than the prime q (with N being the size in bytes of q).
2242 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2243 it yet and it is largely untested.
2246 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2249 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2250 some compilers (gcc 4.2 and later) reject their use. Safestack is
2251 reimplemented. Update ASN1 to avoid use of legacy functions.
2254 *) Win32/64 targets are linked with Winsock2.
2257 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2258 to external functions. This can be used to increase CRL handling
2259 efficiency especially when CRLs are very large by (for example) storing
2260 the CRL revoked certificates in a database.
2263 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2264 new CRLs added to a directory can be used. New command line option
2265 -verify_return_error to s_client and s_server. This causes real errors
2266 to be returned by the verify callback instead of carrying on no matter
2267 what. This reflects the way a "real world" verify callback would behave.
2270 *) GOST engine, supporting several GOST algorithms and public key formats.
2271 Kindly donated by Cryptocom.
2274 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2275 partitioned by DP are handled but no indirect CRL or reason partitioning
2276 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2277 selected via a scoring technique which handles IDP and AKID in CRLs.
2280 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2281 will ultimately be used for all verify operations: this will remove the
2282 X509_STORE dependency on certificate verification and allow alternative
2283 lookup methods. X509_STORE based implementations of these two callbacks.
2286 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2287 Modify get_crl() to find a valid (unexpired) CRL if possible.
2290 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2291 this would be called X509_CRL_cmp() but that name is already used by
2292 a function that just compares CRL issuer names. Cache several CRL
2293 extensions in X509_CRL structure and cache CRLDP in X509.
2296 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2297 this maps equivalent X509_NAME structures into a consistent structure.
2298 Name comparison can then be performed rapidly using memcmp().
2301 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2305 *) Allow digests to supply their own micalg string for S/MIME type using
2306 the ctrl EVP_MD_CTRL_MICALG.
2309 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2310 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2311 ctrl. It can then customise the structure before and/or after signing
2315 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2316 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2317 to free up any added signature OIDs.
2320 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2321 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2322 digest and cipher tables. New options added to openssl utility:
2323 list-message-digest-algorithms and list-cipher-algorithms.
2326 *) Change the array representation of binary polynomials: the list
2327 of degrees of non-zero coefficients is now terminated with -1.
2328 Previously it was terminated with 0, which was also part of the
2329 value; thus, the array representation was not applicable to
2330 polynomials where t^0 has coefficient zero. This change makes
2331 the array representation useful in a more general context.
2334 *) Various modifications and fixes to SSL/TLS cipher string
2335 handling. For ECC, the code now distinguishes between fixed ECDH
2336 with RSA certificates on the one hand and with ECDSA certificates
2337 on the other hand, since these are separate ciphersuites. The
2338 unused code for Fortezza ciphersuites has been removed.
2340 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2341 (not "ECDHE"). For consistency with the code for DH
2342 certificates, use of ECDH certificates is now considered ECDH
2343 authentication, not RSA or ECDSA authentication (the latter is
2344 merely the CA's signing algorithm and not actively used in the
2347 The temporary ciphersuite alias "ECCdraft" is no longer
2348 available, and ECC ciphersuites are no longer excluded from "ALL"
2349 and "DEFAULT". The following aliases now exist for RFC 4492
2350 ciphersuites, most of these by analogy with the DH case:
2352 kECDHr - ECDH cert, signed with RSA
2353 kECDHe - ECDH cert, signed with ECDSA
2354 kECDH - ECDH cert (signed with either RSA or ECDSA)
2355 kEECDH - ephemeral ECDH
2356 ECDH - ECDH cert or ephemeral ECDH
2362 AECDH - anonymous ECDH
2363 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2367 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2368 Use correct micalg parameters depending on digest(s) in signed message.
2371 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2372 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2375 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2376 an engine to register a method. Add ENGINE lookups for methods and
2377 functional reference processing.
2380 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2381 EVP_{Sign,Verify}* which allow an application to customise the signature
2385 *) New -resign option to smime utility. This adds one or more signers
2386 to an existing PKCS#7 signedData structure. Also -md option to use an
2387 alternative message digest algorithm for signing.
2390 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2391 create PKCS7 structures containing multiple signers. Update smime
2392 application to support multiple signers.
2395 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2399 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2400 Reorganize PBE internals to lookup from a static table using NIDs,
2401 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2402 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2403 PRF which will be automatically used with PBES2.
2406 *) Replace the algorithm specific calls to generate keys in "req" with the
2410 *) Update PKCS#7 enveloped data routines to use new API. This is now
2411 supported by any public key method supporting the encrypt operation. A
2412 ctrl is added to allow the public key algorithm to examine or modify
2413 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2417 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2418 a default digest type to use. In most cases this will be SHA1 but some
2419 algorithms (such as GOST) need to specify an alternative digest. The
2420 return value indicates how strong the preference is 1 means optional and
2421 2 is mandatory (that is it is the only supported type). Modify
2422 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2423 use the default md. Update openssl utilities to use the default digest
2424 type for signing if it is not explicitly indicated.
2427 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2428 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2429 signing method from the key type. This effectively removes the link
2430 between digests and public key types.
2433 *) Add an OID cross reference table and utility functions. Its purpose is to
2434 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2435 rsaEncryption. This will allow some of the algorithm specific hackery
2436 needed to use the correct OID to be removed.
2439 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2440 structures for PKCS7_sign(). They are now set up by the relevant public
2444 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2447 *) Add support for key derivation (agreement) in the API, DH method and
2451 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2452 public and private key formats. As a side effect these add additional
2453 command line functionality not previously available: DSA signatures can be
2454 generated and verified using pkeyutl and DH key support and generation in
2459 [Oliver Tappe <zooey@hirschkaefer.de>]
2461 *) New make target "install_html_docs" installs HTML renditions of the
2463 [Oliver Tappe <zooey@hirschkaefer.de>]
2465 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2466 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2467 support key and parameter generation and add initial key generation
2468 functionality for RSA.
2471 *) Add functions for main EVP_PKEY_method operations. The undocumented
2472 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2473 EVP_PKEY_{encrypt,decrypt}_old.
2476 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2477 key API, doesn't do much yet.
2480 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2481 public key algorithms. New option to openssl utility:
2482 "list-public-key-algorithms" to print out info.
2485 *) Implement the Supported Elliptic Curves Extension for
2486 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2489 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2490 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2493 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2494 utilities such as rsa, dsa, dsaparam etc except they process any key
2498 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2499 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2500 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2504 *) Initial support for pluggable public key ASN1.
2505 De-spaghettify the public key ASN1 handling. Move public and private
2506 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2507 algorithm specific handling to a single module within the relevant
2508 algorithm directory. Add functions to allow (near) opaque processing
2509 of public and private key structures.
2512 *) Implement the Supported Point Formats Extension for
2513 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2516 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2517 for the psk identity [hint] and the psk callback functions to the
2518 SSL_SESSION, SSL and SSL_CTX structure.
2521 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2525 SSL_CTX_use_psk_identity_hint
2526 SSL_get_psk_identity_hint
2527 SSL_get_psk_identity
2528 SSL_use_psk_identity_hint
2530 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2532 *) Add RFC 3161 compliant time stamp request creation, response generation
2533 and response verification functionality.
2534 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2536 *) Add initial support for TLS extensions, specifically for the server_name
2537 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2538 have new members for a host name. The SSL data structure has an
2539 additional member SSL_CTX *initial_ctx so that new sessions can be
2540 stored in that context to allow for session resumption, even after the
2541 SSL has been switched to a new SSL_CTX in reaction to a client's
2542 server_name extension.
2544 New functions (subject to change):
2546 SSL_get_servername()
2547 SSL_get_servername_type()
2550 New CTRL codes and macros (subject to change):
2552 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2553 - SSL_CTX_set_tlsext_servername_callback()
2554 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2555 - SSL_CTX_set_tlsext_servername_arg()
2556 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2558 openssl s_client has a new '-servername ...' option.
2560 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2561 '-key2 ...', '-servername_fatal' (subject to change). This allows
2562 testing the HostName extension for a specific single host name ('-cert'
2563 and '-key' remain fallbacks for handshakes without HostName
2564 negotiation). If the unrecognized_name alert has to be sent, this by
2565 default is a warning; it becomes fatal with the '-servername_fatal'
2568 [Peter Sylvester, Remy Allais, Christophe Renou]
2570 *) Whirlpool hash implementation is added.
2573 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2574 bn(64,32). Because of instruction set limitations it doesn't have
2575 any negative impact on performance. This was done mostly in order
2576 to make it possible to share assembler modules, such as bn_mul_mont
2577 implementations, between 32- and 64-bit builds without hassle.
2580 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2581 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2585 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2586 dedicated Montgomery multiplication procedure, is introduced.
2587 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2588 "64-bit" performance on certain 32-bit targets.
2591 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2592 in SSL structures. New SSL ctrl to set maximum send fragment size.
2593 Save memory by seeting the I/O buffer sizes dynamically instead of
2594 using the maximum available value.
2597 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2598 in addition to the text details.
2601 *) Very, very preliminary EXPERIMENTAL support for printing of general
2602 ASN1 structures. This currently produces rather ugly output and doesn't
2603 handle several customised structures at all.
2606 *) Integrated support for PVK file format and some related formats such
2607 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2608 these in the 'rsa' and 'dsa' utilities.
2611 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2614 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2615 place for the (very old) "NETSCAPE" format certificates which are now
2616 handled using new ASN1 code equivalents.
2619 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2620 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2621 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2624 *) Modify CRL distribution points extension code to print out previously
2625 unsupported fields. Enhance extension setting code to allow setting of
2629 *) Add print and set support for Issuing Distribution Point CRL extension.
2632 *) Change 'Configure' script to enable Camellia by default.
2635 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2637 *) When rejecting SSL/TLS records due to an incorrect version number, never
2638 update s->server with a new major version number. As of
2639 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2640 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2641 the previous behavior could result in a read attempt at NULL when
2642 receiving specific incorrect SSL/TLS records once record payload
2643 protection is active. (CVE-2010-0740)
2644 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2646 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2647 could be crashed if the relevant tables were not present (e.g. chrooted).
2648 [Tomas Hoger <thoger@redhat.com>]
2650 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2652 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2653 [Martin Olsson, Neel Mehta]
2655 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2656 accommodate for stack sorting, always a write lock!).
2659 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2660 excessive delays in the RAND_poll(): over a minute. As a workaround
2661 include a time check in the inner Heap32Next loop too.
2664 *) The code that handled flushing of data in SSL/TLS originally used the
2665 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2666 the problem outlined in PR#1949. The fix suggested there however can
2667 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2668 of Apache). So instead simplify the code to flush unconditionally.
2669 This should be fine since flushing with no data to flush is a no op.
2672 *) Handle TLS versions 2.0 and later properly and correctly use the
2673 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2674 off ancient servers have a habit of sticking around for a while...
2677 *) Modify compression code so it frees up structures without using the
2678 ex_data callbacks. This works around a problem where some applications
2679 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2680 restarting) then use compression (e.g. SSL with compression) later.
2681 This results in significant per-connection memory leaks and
2682 has caused some security issues including CVE-2008-1678 and
2686 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2687 change when encrypting or decrypting.
2690 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2691 connect and renegotiate with servers which do not support RI.
2692 Until RI is more widely deployed this option is enabled by default.
2695 *) Add "missing" ssl ctrls to clear options and mode.
2698 *) If client attempts to renegotiate and doesn't support RI respond with
2699 a no_renegotiation alert as required by RFC5746. Some renegotiating
2700 TLS clients will continue a connection gracefully when they receive
2701 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2702 waiting for a server hello which it will never receive. Now we treat a
2703 received no_renegotiation alert as a fatal error. This is because
2704 applications requesting a renegotiation might well expect it to succeed
2705 and would have no code in place to handle the server denying it so the
2706 only safe thing to do is to terminate the connection.
2709 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2710 peer supports secure renegotiation and 0 otherwise. Print out peer
2711 renegotiation support in s_client/s_server.
2714 *) Replace the highly broken and deprecated SPKAC certification method with
2715 the updated NID creation version. This should correctly handle UTF8.
2718 *) Implement RFC5746. Re-enable renegotiation but require the extension
2719 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2720 turns out to be a bad idea. It has been replaced by
2721 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2722 SSL_CTX_set_options(). This is really not recommended unless you
2723 know what you are doing.
2724 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2726 *) Fixes to stateless session resumption handling. Use initial_ctx when
2727 issuing and attempting to decrypt tickets in case it has changed during
2728 servername handling. Use a non-zero length session ID when attempting
2729 stateless session resumption: this makes it possible to determine if
2730 a resumption has occurred immediately after receiving server hello
2731 (several places in OpenSSL subtly assume this) instead of later in
2735 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2736 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2737 fixes for a few places where the return code is not checked
2739 [Julia Lawall <julia@diku.dk>]
2741 *) Add --strict-warnings option to Configure script to include devteam
2742 warnings in other configurations.
2745 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2746 makes it possible to install openssl libraries in locations which
2747 have names other than "lib", for example "/usr/lib64" which some
2749 [Steve Henson, based on patch from Jeremy Utley]
2751 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2752 X690 8.9.12 and can produce some misleading textual output of OIDs.
2753 [Steve Henson, reported by Dan Kaminsky]
2755 *) Delete MD2 from algorithm tables. This follows the recommendation in
2756 several standards that it is not used in new applications due to
2757 several cryptographic weaknesses. For binary compatibility reasons
2758 the MD2 API is still compiled in by default.
2761 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2765 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2766 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2768 [Guenter <lists@gknw.net>]
2770 *) Fix the server certificate chain building code to use X509_verify_cert(),
2771 it used to have an ad-hoc builder which was unable to cope with anything
2772 other than a simple chain.
2773 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2775 *) Don't check self signed certificate signatures in X509_verify_cert()
2776 by default (a flag can override this): it just wastes time without
2777 adding any security. As a useful side effect self signed root CAs
2778 with non-FIPS digests are now usable in FIPS mode.
2781 *) In dtls1_process_out_of_seq_message() the check if the current message
2782 is already buffered was missing. For every new message was memory
2783 allocated, allowing an attacker to perform an denial of service attack
2784 with sending out of seq handshake messages until there is no memory
2785 left. Additionally every future messege was buffered, even if the
2786 sequence number made no sense and would be part of another handshake.
2787 So only messages with sequence numbers less than 10 in advance will be
2788 buffered. (CVE-2009-1378)
2789 [Robin Seggelmann, discovered by Daniel Mentz]
2791 *) Records are buffered if they arrive with a future epoch to be
2792 processed after finishing the corresponding handshake. There is
2793 currently no limitation to this buffer allowing an attacker to perform
2794 a DOS attack with sending records with future epochs until there is no
2795 memory left. This patch adds the pqueue_size() function to determine
2796 the size of a buffer and limits the record buffer to 100 entries.
2798 [Robin Seggelmann, discovered by Daniel Mentz]
2800 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2801 parent structure is freed. (CVE-2009-1379)
2804 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2805 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2808 [Ilya O. <vrghost@gmail.com>]
2810 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2812 *) Disable renegotiation completely - this fixes a severe security
2813 problem (CVE-2009-3555) at the cost of breaking all
2814 renegotiation. Renegotiation can be re-enabled by setting
2815 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2816 run-time. This is really not recommended unless you know what
2820 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2822 *) Don't set val to NULL when freeing up structures, it is freed up by
2823 underlying code. If sizeof(void *) > sizeof(long) this can result in
2824 zeroing past the valid field. (CVE-2009-0789)
2825 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2827 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2828 checked correctly. This would allow some invalid signed attributes to
2829 appear to verify correctly. (CVE-2009-0591)
2830 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2832 *) Reject UniversalString and BMPString types with invalid lengths. This
2833 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2834 a legal length. (CVE-2009-0590)
2837 *) Set S/MIME signing as the default purpose rather than setting it
2838 unconditionally. This allows applications to override it at the store
2842 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2843 to handle some structures.
2846 *) Improve efficiency of mem_gets: don't search whole buffer each time
2848 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2850 *) New -hex option for openssl rand.
2853 *) Print out UTF8String and NumericString when parsing ASN1.
2856 *) Support NumericString type for name components.
2859 *) Allow CC in the environment to override the automatically chosen
2860 compiler. Note that nothing is done to ensure flags work with the
2864 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2866 *) Properly check EVP_VerifyFinal() and similar return values
2868 [Ben Laurie, Bodo Moeller, Google Security Team]
2870 *) Enable TLS extensions by default.
2873 *) Allow the CHIL engine to be loaded, whether the application is
2874 multithreaded or not. (This does not release the developer from the
2875 obligation to set up the dynamic locking callbacks.)
2876 [Sander Temme <sander@temme.net>]
2878 *) Use correct exit code if there is an error in dgst command.
2879 [Steve Henson; problem pointed out by Roland Dirlewanger]
2881 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2882 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2885 *) Add experimental JPAKE support, including demo authentication in
2886 s_client and s_server.
2889 *) Set the comparison function in v3_addr_canonize().
2890 [Rob Austein <sra@hactrn.net>]
2892 *) Add support for XMPP STARTTLS in s_client.
2893 [Philip Paeps <philip@freebsd.org>]
2895 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2896 to ensure that even with this option, only ciphersuites in the
2897 server's preference list will be accepted. (Note that the option
2898 applies only when resuming a session, so the earlier behavior was
2899 just about the algorithm choice for symmetric cryptography.)
2902 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2904 *) Fix NULL pointer dereference if a DTLS server received
2905 ChangeCipherSpec as first record (CVE-2009-1386).
2908 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2909 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2912 *) The fix in 0.9.8c that supposedly got rid of unsafe
2913 double-checked locking was incomplete for RSA blinding,
2914 addressing just one layer of what turns out to have been
2915 doubly unsafe triple-checked locking.
2917 So now fix this for real by retiring the MONT_HELPER macro
2918 in crypto/rsa/rsa_eay.c.
2920 [Bodo Moeller; problem pointed out by Marius Schilder]
2922 *) Various precautionary measures:
2924 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2926 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2927 (NB: This would require knowledge of the secret session ticket key
2928 to exploit, in which case you'd be SOL either way.)
2930 - Change bn_nist.c so that it will properly handle input BIGNUMs
2931 outside the expected range.
2933 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2936 [Neel Mehta, Bodo Moeller]
2938 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2939 the load fails. Useful for distros.
2940 [Ben Laurie and the FreeBSD team]
2942 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2945 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2948 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2950 This work was sponsored by Logica.
2953 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2954 keystores. Support for SSL/TLS client authentication too.
2955 Not compiled unless enable-capieng specified to Configure.
2957 This work was sponsored by Logica.
2960 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2961 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2962 attribute creation routines such as certifcate requests and PKCS#12
2966 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2968 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2969 handshake which could lead to a cilent crash as found using the
2970 Codenomicon TLS test suite (CVE-2008-1672)
2971 [Steve Henson, Mark Cox]
2973 *) Fix double free in TLS server name extensions which could lead to
2974 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2977 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2979 Clear the error queue to ensure that error entries left from
2980 older function calls do not interfere with the correct operation.
2981 [Lutz Jaenicke, Erik de Castro Lopo]
2983 *) Remove root CA certificates of commercial CAs:
2985 The OpenSSL project does not recommend any specific CA and does not
2986 have any policy with respect to including or excluding any CA.
2987 Therefore it does not make any sense to ship an arbitrary selection
2988 of root CA certificates with the OpenSSL software.
2991 *) RSA OAEP patches to fix two separate invalid memory reads.
2992 The first one involves inputs when 'lzero' is greater than
2993 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2994 before the beginning of from). The second one involves inputs where
2995 the 'db' section contains nothing but zeroes (there is a one-byte
2996 invalid read after the end of 'db').
2997 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2999 *) Partial backport from 0.9.9-dev:
3001 Introduce bn_mul_mont (dedicated Montgomery multiplication
3002 procedure) as a candidate for BIGNUM assembler implementation.
3003 While 0.9.9-dev uses assembler for various architectures, only
3004 x86_64 is available by default here in the 0.9.8 branch, and
3005 32-bit x86 is available through a compile-time setting.
3007 To try the 32-bit x86 assembler implementation, use Configure
3008 option "enable-montasm" (which exists only for this backport).
3010 As "enable-montasm" for 32-bit x86 disclaims code stability
3011 anyway, in this constellation we activate additional code
3012 backported from 0.9.9-dev for further performance improvements,
3013 namely BN_from_montgomery_word. (To enable this otherwise,
3014 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3016 [Andy Polyakov (backport partially by Bodo Moeller)]
3018 *) Add TLS session ticket callback. This allows an application to set
3019 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3020 values. This is useful for key rollover for example where several key
3021 sets may exist with different names.
3024 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3025 This was broken until now in 0.9.8 releases, such that the only way
3026 a registered ENGINE could be used (assuming it initialises
3027 successfully on the host) was to explicitly set it as the default
3028 for the relevant algorithms. This is in contradiction with 0.9.7
3029 behaviour and the documentation. With this fix, when an ENGINE is
3030 registered into a given algorithm's table of implementations, the
3031 'uptodate' flag is reset so that auto-discovery will be used next
3032 time a new context for that algorithm attempts to select an
3034 [Ian Lister (tweaked by Geoff Thorpe)]
3036 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3037 implemention in the following ways:
3039 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3042 Lack of BER streaming support means one pass streaming processing is
3043 only supported if data is detached: setting the streaming flag is
3044 ignored for embedded content.
3046 CMS support is disabled by default and must be explicitly enabled
3047 with the enable-cms configuration option.
3050 *) Update the GMP engine glue to do direct copies between BIGNUM and
3051 mpz_t when openssl and GMP use the same limb size. Otherwise the
3052 existing "conversion via a text string export" trick is still used.
3053 [Paul Sheer <paulsheer@gmail.com>]
3055 *) Zlib compression BIO. This is a filter BIO which compressed and
3056 uncompresses any data passed through it.
3059 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3060 RFC3394 compatible AES key wrapping.
3063 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3064 sets string data without copying. X509_ALGOR_set0() and
3065 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3066 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3067 from an X509_ATTRIBUTE structure optionally checking it occurs only
3068 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3072 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3073 to get the expected BN_FLG_CONSTTIME behavior.
3074 [Bodo Moeller (Google)]
3078 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3079 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3080 - added some more tests to do_tests.pl
3081 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3082 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3083 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3084 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3085 - various changes to netware.pl to enable gcc-cross builds on Win32
3087 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3088 - various changes to fix missing prototype warnings
3089 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3090 - added AES, WHIRLPOOL and CPUID assembler code to build files
3091 - added missing AES assembler make rules to mk1mf.pl
3092 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3093 [Guenter Knauf <eflash@gmx.net>]
3095 *) Implement certificate status request TLS extension defined in RFC3546.
3096 A client can set the appropriate parameters and receive the encoded
3097 OCSP response via a callback. A server can query the supplied parameters
3098 and set the encoded OCSP response in the callback. Add simplified examples
3099 to s_client and s_server.
3102 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3104 *) Fix various bugs:
3105 + Binary incompatibility of ssl_ctx_st structure
3106 + DTLS interoperation with non-compliant servers
3107 + Don't call get_session_cb() without proposed session
3108 + Fix ia64 assembler code
3109 [Andy Polyakov, Steve Henson]
3111 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3113 *) DTLS Handshake overhaul. There were longstanding issues with
3114 OpenSSL DTLS implementation, which were making it impossible for
3115 RFC 4347 compliant client to communicate with OpenSSL server.
3116 Unfortunately just fixing these incompatibilities would "cut off"
3117 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3118 server keeps tolerating non RFC compliant syntax. The opposite is
3119 not true, 0.9.8f client can not communicate with earlier server.
3120 This update even addresses CVE-2007-4995.
3123 *) Changes to avoid need for function casts in OpenSSL: some compilers
3124 (gcc 4.2 and later) reject their use.
3125 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3128 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3129 RFC4507bis. The encrypted ticket format is an encrypted encoded
3130 SSL_SESSION structure, that way new session features are automatically
3133 If a client application caches session in an SSL_SESSION structure
3134 support is transparent because tickets are now stored in the encoded
3137 The SSL_CTX structure automatically generates keys for ticket
3138 protection in servers so again support should be possible
3139 with no application modification.
3141 If a client or server wishes to disable RFC4507 support then the option
3142 SSL_OP_NO_TICKET can be set.
3144 Add a TLS extension debugging callback to allow the contents of any client
3145 or server extensions to be examined.
3147 This work was sponsored by Google.
3150 *) Add initial support for TLS extensions, specifically for the server_name
3151 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3152 have new members for a host name. The SSL data structure has an
3153 additional member SSL_CTX *initial_ctx so that new sessions can be
3154 stored in that context to allow for session resumption, even after the
3155 SSL has been switched to a new SSL_CTX in reaction to a client's
3156 server_name extension.
3158 New functions (subject to change):
3160 SSL_get_servername()
3161 SSL_get_servername_type()
3164 New CTRL codes and macros (subject to change):
3166 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3167 - SSL_CTX_set_tlsext_servername_callback()
3168 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3169 - SSL_CTX_set_tlsext_servername_arg()
3170 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3172 openssl s_client has a new '-servername ...' option.
3174 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3175 '-key2 ...', '-servername_fatal' (subject to change). This allows
3176 testing the HostName extension for a specific single host name ('-cert'
3177 and '-key' remain fallbacks for handshakes without HostName
3178 negotiation). If the unrecognized_name alert has to be sent, this by
3179 default is a warning; it becomes fatal with the '-servername_fatal'
3182 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3184 *) Add AES and SSE2 assembly language support to VC++ build.
3187 *) Mitigate attack on final subtraction in Montgomery reduction.
3190 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3191 (which previously caused an internal error).
3194 *) Squeeze another 10% out of IGE mode when in != out.
3197 *) AES IGE mode speedup.
3198 [Dean Gaudet (Google)]
3200 *) Add the Korean symmetric 128-bit cipher SEED (see
3201 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3202 add SEED ciphersuites from RFC 4162:
3204 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3205 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3206 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3207 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3209 To minimize changes between patchlevels in the OpenSSL 0.9.8
3210 series, SEED remains excluded from compilation unless OpenSSL
3211 is configured with 'enable-seed'.
3212 [KISA, Bodo Moeller]
3214 *) Mitigate branch prediction attacks, which can be practical if a
3215 single processor is shared, allowing a spy process to extract
3216 information. For detailed background information, see
3217 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3218 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3219 and Necessary Software Countermeasures"). The core of the change
3220 are new versions BN_div_no_branch() and
3221 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3222 respectively, which are slower, but avoid the security-relevant
3223 conditional branches. These are automatically called by BN_div()
3224 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3225 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3226 remove a conditional branch.
3228 BN_FLG_CONSTTIME is the new name for the previous
3229 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3230 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3231 in the exponent causes BN_mod_exp_mont() to use the alternative
3232 implementation in BN_mod_exp_mont_consttime().) The old name
3233 remains as a deprecated alias.
3235 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3236 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3237 constant-time implementations for more than just exponentiation.
3238 Here too the old name is kept as a deprecated alias.
3240 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3241 the BN_BLINDING structure gets an independent copy of the
3242 modulus. This means that the previous "BIGNUM *m" argument to
3243 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3244 essentially becomes "const BIGNUM *m", although we can't actually
3245 change this in the header file before 0.9.9. It allows
3246 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3247 enable BN_FLG_CONSTTIME.
3249 [Matthew D Wood (Intel Corp)]
3251 *) In the SSL/TLS server implementation, be strict about session ID
3252 context matching (which matters if an application uses a single
3253 external cache for different purposes). Previously,
3254 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3255 set. This did ensure strict client verification, but meant that,
3256 with applications using a single external cache for quite
3257 different requirements, clients could circumvent ciphersuite
3258 restrictions for a given session ID context by starting a session
3259 in a different context.
3262 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3263 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3264 authentication-only ciphersuites.
3267 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3268 not complete and could lead to a possible single byte overflow
3269 (CVE-2007-5135) [Ben Laurie]
3271 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3273 *) Since AES128 and AES256 (and similarly Camellia128 and
3274 Camellia256) share a single mask bit in the logic of
3275 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3276 kludge to work properly if AES128 is available and AES256 isn't
3277 (or if Camellia128 is available and Camellia256 isn't).
3280 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3281 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3282 When a point or a seed is encoded in a BIT STRING, we need to
3283 prevent the removal of trailing zero bits to get the proper DER
3284 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3285 of a NamedBitList, for which trailing 0 bits need to be removed.)
3288 *) Have SSL/TLS server implementation tolerate "mismatched" record
3289 protocol version while receiving ClientHello even if the
3290 ClientHello is fragmented. (The server can't insist on the
3291 particular protocol version it has chosen before the ServerHello
3292 message has informed the client about his choice.)
3295 *) Add RFC 3779 support.
3296 [Rob Austein for ARIN, Ben Laurie]
3298 *) Load error codes if they are not already present instead of using a
3299 static variable. This allows them to be cleanly unloaded and reloaded.
3300 Improve header file function name parsing.
3303 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3304 or CAPABILITY handshake as required by RFCs.
3307 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3309 *) Introduce limits to prevent malicious keys being able to
3310 cause a denial of service. (CVE-2006-2940)
3311 [Steve Henson, Bodo Moeller]
3313 *) Fix ASN.1 parsing of certain invalid structures that can result
3314 in a denial of service. (CVE-2006-2937) [Steve Henson]
3316 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3317 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3319 *) Fix SSL client code which could crash if connecting to a
3320 malicious SSLv2 server. (CVE-2006-4343)
3321 [Tavis Ormandy and Will Drewry, Google Security Team]
3323 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3324 match only those. Before that, "AES256-SHA" would be interpreted
3325 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3326 the same strength classification in 0.9.7h) as we currently only
3327 have a single AES bit in the ciphersuite description bitmap.
3328 That change, however, also applied to ciphersuite strings such as
3329 "RC4-MD5" that intentionally matched multiple ciphersuites --
3330 namely, SSL 2.0 ciphersuites in addition to the more common ones
3331 from SSL 3.0/TLS 1.0.
3333 So we change the selection algorithm again: Naming an explicit
3334 ciphersuite selects this one ciphersuite, and any other similar
3335 ciphersuite (same bitmap) from *other* protocol versions.
3336 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3337 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3339 Since SSL 2.0 does not have any ciphersuites for which the
3340 128/256 bit distinction would be relevant, this works for now.
3341 The proper fix will be to use different bits for AES128 and
3342 AES256, which would have avoided the problems from the beginning;
3343 however, bits are scarce, so we can only do this in a new release
3344 (not just a patchlevel) when we can change the SSL_CIPHER
3345 definition to split the single 'unsigned long mask' bitmap into
3346 multiple values to extend the available space.
3350 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3352 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3353 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3355 *) Add AES IGE and biIGE modes.
3358 *) Change the Unix randomness entropy gathering to use poll() when
3359 possible instead of select(), since the latter has some
3360 undesirable limitations.
3361 [Darryl Miles via Richard Levitte and Bodo Moeller]
3363 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3364 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3365 cannot be implicitly activated as part of, e.g., the "AES" alias.
3366 However, please upgrade to OpenSSL 0.9.9[-dev] for
3367 non-experimental use of the ECC ciphersuites to get TLS extension
3368 support, which is required for curve and point format negotiation
3369 to avoid potential handshake problems.
3372 *) Disable rogue ciphersuites:
3374 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3375 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3376 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3378 The latter two were purportedly from
3379 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3382 Also deactivate the remaining ciphersuites from
3383 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3384 unofficial, and the ID has long expired.
3387 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3388 dual-core machines) and other potential thread-safety issues.
3391 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3392 versions), which is now available for royalty-free use
3393 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3394 Also, add Camellia TLS ciphersuites from RFC 4132.
3396 To minimize changes between patchlevels in the OpenSSL 0.9.8
3397 series, Camellia remains excluded from compilation unless OpenSSL
3398 is configured with 'enable-camellia'.
3401 *) Disable the padding bug check when compression is in use. The padding
3402 bug check assumes the first packet is of even length, this is not
3403 necessarily true if compresssion is enabled and can result in false
3404 positives causing handshake failure. The actual bug test is ancient
3405 code so it is hoped that implementations will either have fixed it by
3406 now or any which still have the bug do not support compression.
3409 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3411 *) When applying a cipher rule check to see if string match is an explicit
3412 cipher suite and only match that one cipher suite if it is.
3415 *) Link in manifests for VC++ if needed.
3416 [Austin Ziegler <halostatue@gmail.com>]
3418 *) Update support for ECC-based TLS ciphersuites according to
3419 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3420 TLS extensions, which are supported starting with the 0.9.9
3421 branch, not in the OpenSSL 0.9.8 branch).
3424 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3425 opaque EVP_CIPHER_CTX handling.
3428 *) Fixes and enhancements to zlib compression code. We now only use
3429 "zlib1.dll" and use the default __cdecl calling convention on Win32
3430 to conform with the standards mentioned here:
3431 http://www.zlib.net/DLL_FAQ.txt
3432 Static zlib linking now works on Windows and the new --with-zlib-include
3433 --with-zlib-lib options to Configure can be used to supply the location
3434 of the headers and library. Gracefully handle case where zlib library
3438 *) Several fixes and enhancements to the OID generation code. The old code
3439 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3440 handle numbers larger than ULONG_MAX, truncated printing and had a
3441 non standard OBJ_obj2txt() behaviour.
3444 *) Add support for building of engines under engine/ as shared libraries
3445 under VC++ build system.
3448 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3449 Hopefully, we will not see any false combination of paths any more.
3452 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3454 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3455 (part of SSL_OP_ALL). This option used to disable the
3456 countermeasure against man-in-the-middle protocol-version
3457 rollback in the SSL 2.0 server implementation, which is a bad
3458 idea. (CVE-2005-2969)
3460 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3461 for Information Security, National Institute of Advanced Industrial
3462 Science and Technology [AIST], Japan)]
3464 *) Add two function to clear and return the verify parameter flags.
3467 *) Keep cipherlists sorted in the source instead of sorting them at
3468 runtime, thus removing the need for a lock.
3471 *) Avoid some small subgroup attacks in Diffie-Hellman.
3472 [Nick Mathewson and Ben Laurie]
3474 *) Add functions for well-known primes.
3477 *) Extended Windows CE support.
3478 [Satoshi Nakamura and Andy Polyakov]
3480 *) Initialize SSL_METHOD structures at compile time instead of during
3481 runtime, thus removing the need for a lock.
3484 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3485 attempting to decrypt each encrypted key in turn. Add support to
3489 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3491 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3494 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3497 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3498 key into the same file any more.
3501 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3504 *) Add -utf8 command line and config file option to 'ca'.
3505 [Stefan <stf@udoma.org]
3507 *) Removed the macro des_crypt(), as it seems to conflict with some
3508 libraries. Use DES_crypt().
3511 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3512 involves renaming the source and generated shared-libs for
3513 both. The engines will accept the corrected or legacy ids
3514 ('ncipher' and '4758_cca' respectively) when binding. NB,
3515 this only applies when building 'shared'.
3516 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3518 *) Add attribute functions to EVP_PKEY structure. Modify
3519 PKCS12_create() to recognize a CSP name attribute and
3520 use it. Make -CSP option work again in pkcs12 utility.
3523 *) Add new functionality to the bn blinding code:
3524 - automatic re-creation of the BN_BLINDING parameters after
3525 a fixed number of uses (currently 32)
3526 - add new function for parameter creation
3527 - introduce flags to control the update behaviour of the
3528 BN_BLINDING parameters
3529 - hide BN_BLINDING structure
3530 Add a second BN_BLINDING slot to the RSA structure to improve
3531 performance when a single RSA object is shared among several
3535 *) Add support for DTLS.
3536 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3538 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3539 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3542 *) Remove buggy and incomplete DH cert support from
3543 ssl/ssl_rsa.c and ssl/s3_both.c
3546 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3547 the apps/openssl applications.
3550 *) Compile clean with "-Wall -Wmissing-prototypes
3551 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3552 DEBUG_SAFESTACK must also be set.
3555 *) Change ./Configure so that certain algorithms can be disabled by default.
3556 The new counterpiece to "no-xxx" is "enable-xxx".
3558 The patented RC5 and MDC2 algorithms will now be disabled unless
3559 "enable-rc5" and "enable-mdc2", respectively, are specified.
3561 (IDEA remains enabled despite being patented. This is because IDEA
3562 is frequently required for interoperability, and there is no license
3563 fee for non-commercial use. As before, "no-idea" can be used to
3564 avoid this algorithm.)
3568 *) Add processing of proxy certificates (see RFC 3820). This work was
3569 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3570 EGEE (Enabling Grids for E-science in Europe).
3573 *) RC4 performance overhaul on modern architectures/implementations, such
3574 as Intel P4, IA-64 and AMD64.
3577 *) New utility extract-section.pl. This can be used specify an alternative
3578 section number in a pod file instead of having to treat each file as
3579 a separate case in Makefile. This can be done by adding two lines to the
3582 =for comment openssl_section:XXX
3584 The blank line is mandatory.
3588 *) New arguments -certform, -keyform and -pass for s_client and s_server
3589 to allow alternative format key and certificate files and passphrase
3593 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3594 update associated structures and add various utility functions.
3596 Add new policy related verify parameters, include policy checking in
3597 standard verify code. Enhance 'smime' application with extra parameters
3598 to support policy checking and print out.
3601 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3602 Nehemiah processors. These extensions support AES encryption in hardware
3603 as well as RNG (though RNG support is currently disabled).
3604 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3606 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3609 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3610 [Andy Polyakov and a number of other people]
3612 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3613 implementation contributed by IBM.
3614 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3616 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3617 exponent rather than 'unsigned long'. There is a corresponding change to
3618 the new 'rsa_keygen' element of the RSA_METHOD structure.
3619 [Jelte Jansen, Geoff Thorpe]
3621 *) Functionality for creating the initial serial number file is now
3622 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3624 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3625 number file to 1, which is bound to cause problems. To avoid
3626 the problems while respecting compatibility between different 0.9.7
3627 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3628 CA.pl for serial number initialization. With the new release 0.9.8,
3629 we can fix the problem directly in the 'ca' utility.)
3632 *) Reduced header interdepencies by declaring more opaque objects in
3633 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3634 give fewer recursive includes, which could break lazy source code - so
3635 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3636 developers should define this symbol when building and using openssl to
3637 ensure they track the recommended behaviour, interfaces, [etc], but
3638 backwards-compatible behaviour prevails when this isn't defined.
3641 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3644 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3645 This will generate a random key of the appropriate length based on the
3646 cipher context. The EVP_CIPHER can provide its own random key generation
3647 routine to support keys of a specific form. This is used in the des and
3648 3des routines to generate a key of the correct parity. Update S/MIME
3649 code to use new functions and hence generate correct parity DES keys.
3650 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3651 valid (weak or incorrect parity).
3654 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3655 as looking them up. This is useful when the verified structure may contain
3656 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3657 present unless the new PKCS7_NO_CRL flag is asserted.
3660 *) Extend ASN1 oid configuration module. It now additionally accepts the
3663 shortName = some long name, 1.2.3.4
3666 *) Reimplemented the BN_CTX implementation. There is now no more static
3667 limitation on the number of variables it can handle nor the depth of the
3668 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3669 information can now expand as required, and rather than having a single
3670 static array of bignums, BN_CTX now uses a linked-list of such arrays
3671 allowing it to expand on demand whilst maintaining the usefulness of
3672 BN_CTX's "bundling".
3675 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3676 to allow all RSA operations to function using a single BN_CTX.
3679 *) Preliminary support for certificate policy evaluation and checking. This
3680 is initially intended to pass the tests outlined in "Conformance Testing
3681 of Relying Party Client Certificate Path Processing Logic" v1.07.
3684 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3685 remained unused and not that useful. A variety of other little bignum
3686 tweaks and fixes have also been made continuing on from the audit (see
3690 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3691 associated ASN1, EVP and SSL functions and old ASN1 macros.
3694 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3695 and this should never fail. So the return value from the use of
3696 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3697 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3700 *) BN_CTX_get() should return zero-valued bignums, providing the same
3701 initialised value as BN_new().
3702 [Geoff Thorpe, suggested by Ulf Möller]
3704 *) Support for inhibitAnyPolicy certificate extension.
3707 *) An audit of the BIGNUM code is underway, for which debugging code is
3708 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3709 is considered valid when processing BIGNUMs, and causes execution to
3710 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3711 further steps are taken to deliberately pollute unused data in BIGNUM
3712 structures to try and expose faulty code further on. For now, openssl will
3713 (in its default mode of operation) continue to tolerate the inconsistent
3714 forms that it has tolerated in the past, but authors and packagers should
3715 consider trying openssl and their own applications when compiled with
3716 these debugging symbols defined. It will help highlight potential bugs in
3717 their own code, and will improve the test coverage for OpenSSL itself. At
3718 some point, these tighter rules will become openssl's default to improve
3719 maintainability, though the assert()s and other overheads will remain only
3720 in debugging configurations. See bn.h for more details.
3721 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3723 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3724 that can only be obtained through BN_CTX_new() (which implicitly
3725 initialises it). The presence of this function only made it possible
3726 to overwrite an existing structure (and cause memory leaks).
3729 *) Because of the callback-based approach for implementing LHASH as a
3730 template type, lh_insert() adds opaque objects to hash-tables and
3731 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3732 to clean up those corresponding objects before destroying the hash table
projects / openssl.git / blob