5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
6 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
7 not aware of clients that still exhibit this bug, and the workaround
8 hasn't been working properly for a while.
11 *) The return type of BIO_number_read() and BIO_number_written() as well as
12 the corresponding num_read and num_write members in the BIO structure has
13 changed from unsigned long to uint64_t. On platforms where an unsigned
14 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
18 *) Given the pervasive nature of TLS extensions it is inadvisable to run
19 OpenSSL without support for them. It also means that maintaining
20 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
21 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
24 *) Removed support for the two export grade static DH ciphersuites
25 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
26 were newly added (along with a number of other static DH ciphersuites) to
27 1.0.2. However the two export ones have *never* worked since they were
28 introduced. It seems strange in any case to be adding new export
29 ciphersuites, and given "logjam" it also does not seem correct to fix them.
32 *) Version negotiation has been rewritten. In particular SSLv23_method(),
33 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
34 and turned into macros which simply call the new preferred function names
35 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
36 should use the new names instead. Also as part of this change the ssl23.h
37 header file has been removed.
40 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
41 code and the associated standard is no longer considered fit-for-purpose.
44 *) Added HTTP GET support to the ocsp command.
47 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
50 *) Added support for TLS extended master secret from
51 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
52 initial patch which was a great help during development.
55 *) All libssl internal structures have been removed from the public header
56 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
57 now redundant). Users should not attempt to access internal structures
58 directly. Instead they should use the provided API functions.
61 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
62 Access to deprecated functions can be re-enabled by running config with
63 "enable-deprecated". In addition applications wishing to use deprecated
64 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
65 will, by default, disable some transitive includes that previously existed
66 in the header files (e.g. ec.h will no longer, by default, include bn.h)
69 *) Added support for OCB mode. OpenSSL has been granted a patent license
70 compatible with the OpenSSL license for use of OCB. Details are available
71 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
72 for OCB can be removed by calling config with no-ocb.
75 *) SSLv2 support has been removed. It still supports receiving a SSLv2
76 compatible client hello.
79 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
80 done while fixing the error code for the key-too-small case.
81 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
83 *) CA.sh has been removmed; use CA.pl instead.
86 *) Removed old DES API.
89 *) Remove various unsupported platforms:
95 Sinix/ReliantUNIX RM400
100 16-bit platforms such as WIN16
103 *) Clean up OPENSSL_NO_xxx #define's
104 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
105 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
106 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
107 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
108 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
109 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
110 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
111 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
112 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
113 Remove MS_STATIC; it's a relic from platforms <32 bits.
116 *) Cleaned up dead code
117 Remove all but one '#ifdef undef' which is to be looked at.
120 *) Clean up calling of xxx_free routines.
121 Just like free(), fix most of the xxx_free routines to accept
122 NULL. Remove the non-null checks from callers. Save much code.
125 *) Experimental support for a new, fast, unbiased prime candidate generator,
126 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
127 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
129 *) New output format NSS in the sess_id command line tool. This allows
130 exporting the session id and the master key in NSS keylog format.
131 [Martin Kaiser <martin@kaiser.cx>]
133 *) Harmonize version and its documentation. -f flag is used to display
135 [mancha <mancha1@zoho.com>]
137 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
139 [mancha <mancha1@zoho.com>]
141 *) Fix some double frees. These are not thought to be exploitable.
142 [mancha <mancha1@zoho.com>]
144 *) A missing bounds check in the handling of the TLS heartbeat extension
145 can be used to reveal up to 64k of memory to a connected client or
148 Thanks for Neel Mehta of Google Security for discovering this bug and to
149 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
150 preparing the fix (CVE-2014-0160)
151 [Adam Langley, Bodo Moeller]
153 *) Fix for the attack described in the paper "Recovering OpenSSL
154 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
155 by Yuval Yarom and Naomi Benger. Details can be obtained from:
156 http://eprint.iacr.org/2014/140
158 Thanks to Yuval Yarom and Naomi Benger for discovering this
159 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
160 [Yuval Yarom and Naomi Benger]
162 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
163 this fixes a limitation in previous versions of OpenSSL.
166 *) Experimental encrypt-then-mac support.
168 Experimental support for encrypt then mac from
169 draft-gutmann-tls-encrypt-then-mac-02.txt
171 To enable it set the appropriate extension number (0x42 for the test
172 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
174 For non-compliant peers (i.e. just about everything) this should have no
177 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
181 *) Add EVP support for key wrapping algorithms, to avoid problems with
182 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
183 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
184 algorithms and include tests cases.
187 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
191 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
192 MGF1 digest and OAEP label.
195 *) Make openssl verify return errors.
196 [Chris Palmer <palmer@google.com> and Ben Laurie]
198 *) New function ASN1_TIME_diff to calculate the difference between two
199 ASN1_TIME structures or one structure and the current time.
202 *) Update fips_test_suite to support multiple command line options. New
203 test to induce all self test errors in sequence and check expected
207 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
208 sign or verify all in one operation.
211 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
212 test programs and fips_test_suite. Includes functionality to parse
213 the minimal script output of fipsalgest.pl directly.
216 *) Add authorisation parameter to FIPS_module_mode_set().
219 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
222 *) Use separate DRBG fields for internal and external flags. New function
223 FIPS_drbg_health_check() to perform on demand health checking. Add
224 generation tests to fips_test_suite with reduced health check interval to
225 demonstrate periodic health checking. Add "nodh" option to
226 fips_test_suite to skip very slow DH test.
229 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
233 *) More extensive health check for DRBG checking many more failure modes.
234 New function FIPS_selftest_drbg_all() to handle every possible DRBG
235 combination: call this in fips_test_suite.
238 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
239 and POST to handle Dual EC cases.
242 *) Add support for canonical generation of DSA parameter 'g'. See
245 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
246 POST to handle HMAC cases.
249 *) Add functions FIPS_module_version() and FIPS_module_version_text()
250 to return numerical and string versions of the FIPS module number.
253 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
254 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
255 outside the validated module in the FIPS capable OpenSSL.
258 *) Minor change to DRBG entropy callback semantics. In some cases
259 there is no multiple of the block length between min_len and
260 max_len. Allow the callback to return more than max_len bytes
261 of entropy but discard any extra: it is the callback's responsibility
262 to ensure that the extra data discarded does not impact the
263 requested amount of entropy.
266 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
267 information in FIPS186-3, SP800-57 and SP800-131A.
270 *) CCM support via EVP. Interface is very similar to GCM case except we
271 must supply all data in one chunk (i.e. no update, final) and the
272 message length must be supplied if AAD is used. Add algorithm test
276 *) Initial version of POST overhaul. Add POST callback to allow the status
277 of POST to be monitored and/or failures induced. Modify fips_test_suite
278 to use callback. Always run all selftests even if one fails.
281 *) XTS support including algorithm test driver in the fips_gcmtest program.
282 Note: this does increase the maximum key length from 32 to 64 bytes but
283 there should be no binary compatibility issues as existing applications
284 will never use XTS mode.
287 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
288 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
289 performs algorithm blocking for unapproved PRNG types. Also do not
290 set PRNG type in FIPS_mode_set(): leave this to the application.
291 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
292 the standard OpenSSL PRNG: set additional data to a date time vector.
295 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
296 This shouldn't present any incompatibility problems because applications
297 shouldn't be using these directly and any that are will need to rethink
298 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
301 *) Extensive self tests and health checking required by SP800-90 DRBG.
302 Remove strength parameter from FIPS_drbg_instantiate and always
303 instantiate at maximum supported strength.
306 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
309 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
312 *) New function DH_compute_key_padded() to compute a DH key and pad with
313 leading zeroes if needed: this complies with SP800-56A et al.
316 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
317 anything, incomplete, subject to change and largely untested at present.
320 *) Modify fipscanisteronly build option to only build the necessary object
321 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
324 *) Add experimental option FIPSSYMS to give all symbols in
325 fipscanister.o and FIPS or fips prefix. This will avoid
326 conflicts with future versions of OpenSSL. Add perl script
327 util/fipsas.pl to preprocess assembly language source files
328 and rename any affected symbols.
331 *) Add selftest checks and algorithm block of non-fips algorithms in
332 FIPS mode. Remove DES2 from selftests.
335 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
336 return internal method without any ENGINE dependencies. Add new
337 tiny fips sign and verify functions.
340 *) New build option no-ec2m to disable characteristic 2 code.
343 *) New build option "fipscanisteronly". This only builds fipscanister.o
344 and (currently) associated fips utilities. Uses the file Makefile.fips
345 instead of Makefile.org as the prototype.
348 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
349 Update fips_gcmtest to use IV generator.
352 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
353 setting output buffer to NULL. The *Final function must be
354 called although it will not retrieve any additional data. The tag
355 can be set or retrieved with a ctrl. The IV length is by default 12
356 bytes (96 bits) but can be set to an alternative value. If the IV
357 length exceeds the maximum IV length (currently 16 bytes) it cannot be
361 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
362 underlying do_cipher function handles all cipher semantics itself
363 including padding and finalisation. This is useful if (for example)
364 an ENGINE cipher handles block padding itself. The behaviour of
365 do_cipher is subtly changed if this flag is set: the return value
366 is the number of characters written to the output buffer (zero is
367 no longer an error code) or a negative error code. Also if the
368 input buffer is NULL and length 0 finalisation should be performed.
371 *) If a candidate issuer certificate is already part of the constructed
372 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
375 *) Improve forward-security support: add functions
377 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
378 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
380 for use by SSL/TLS servers; the callback function will be called whenever a
381 new session is created, and gets to decide whether the session may be
382 cached to make it resumable (return 0) or not (return 1). (As by the
383 SSL/TLS protocol specifications, the session_id sent by the server will be
384 empty to indicate that the session is not resumable; also, the server will
385 not generate RFC 4507 (RFC 5077) session tickets.)
387 A simple reasonable callback implementation is to return is_forward_secure.
388 This parameter will be set to 1 or 0 depending on the ciphersuite selected
389 by the SSL/TLS server library, indicating whether it can provide forward
391 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
393 *) New -verify_name option in command line utilities to set verification
397 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
398 Add CMAC pkey methods.
401 *) Experimental renegotiation in s_server -www mode. If the client
402 browses /reneg connection is renegotiated. If /renegcert it is
403 renegotiated requesting a certificate.
406 *) Add an "external" session cache for debugging purposes to s_server. This
407 should help trace issues which normally are only apparent in deployed
408 multi-process servers.
411 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
412 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
413 BIO_set_cipher() and some obscure PEM functions were changed so they
414 can now return an error. The RAND changes required a change to the
415 RAND_METHOD structure.
418 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
419 a gcc attribute to warn if the result of a function is ignored. This
420 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
421 whose return value is often ignored.
424 Changes between 1.0.2a and 1.0.2b [xx XXX xxxx]
426 *) Malformed ECParameters causes infinite loop
428 When processing an ECParameters structure OpenSSL enters an infinite loop
429 if the curve specified is over a specially malformed binary polynomial
432 This can be used to perform denial of service against any
433 system which processes public keys, certificate requests or
434 certificates. This includes TLS clients and TLS servers with
435 client authentication enabled.
437 This issue was reported to OpenSSL by Joseph Barr-Pixton.
441 *) Exploitable out-of-bounds read in X509_cmp_time
443 X509_cmp_time does not properly check the length of the ASN1_TIME
444 string and can read a few bytes out of bounds. In addition,
445 X509_cmp_time accepts an arbitrary number of fractional seconds in the
448 An attacker can use this to craft malformed certificates and CRLs of
449 various sizes and potentially cause a segmentation fault, resulting in
450 a DoS on applications that verify certificates or CRLs. TLS clients
451 that verify CRLs are affected. TLS clients and servers with client
452 authentication enabled may be affected if they use custom verification
455 This issue was reported to OpenSSL by Robert Swiecki (Google), and
456 independently by Hanno Böck.
460 *) PKCS7 crash with missing EnvelopedContent
462 The PKCS#7 parsing code does not handle missing inner EncryptedContent
463 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
464 with missing content and trigger a NULL pointer dereference on parsing.
466 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
467 structures from untrusted sources are affected. OpenSSL clients and
468 servers are not affected.
470 This issue was reported to OpenSSL by Michal Zalewski (Google).
474 *) CMS verify infinite loop with unknown hash function
476 When verifying a signedData message the CMS code can enter an infinite loop
477 if presented with an unknown hash function OID. This can be used to perform
478 denial of service against any system which verifies signedData messages using
480 This issue was reported to OpenSSL by Johannes Bauer.
484 *) Race condition handling NewSessionTicket
486 If a NewSessionTicket is received by a multi-threaded client when attempting to
487 reuse a previous ticket then a race condition can occur potentially leading to
488 a double free of the ticket data.
492 *) Only support 256-bit or stronger elliptic curves with the
493 'ecdh_auto' setting (server) or by default (client). Of supported
494 curves, prefer P-256 (both).
497 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
499 *) ClientHello sigalgs DoS fix
501 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
502 invalid signature algorithms extension a NULL pointer dereference will
503 occur. This can be exploited in a DoS attack against the server.
505 This issue was was reported to OpenSSL by David Ramos of Stanford
508 [Stephen Henson and Matt Caswell]
510 *) Multiblock corrupted pointer fix
512 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
513 feature only applies on 64 bit x86 architecture platforms that support AES
514 NI instructions. A defect in the implementation of "multiblock" can cause
515 OpenSSL's internal write buffer to become incorrectly set to NULL when
516 using non-blocking IO. Typically, when the user application is using a
517 socket BIO for writing, this will only result in a failed connection.
518 However if some other BIO is used then it is likely that a segmentation
519 fault will be triggered, thus enabling a potential DoS attack.
521 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
525 *) Segmentation fault in DTLSv1_listen fix
527 The DTLSv1_listen function is intended to be stateless and processes the
528 initial ClientHello from many peers. It is common for user code to loop
529 over the call to DTLSv1_listen until a valid ClientHello is received with
530 an associated cookie. A defect in the implementation of DTLSv1_listen means
531 that state is preserved in the SSL object from one invocation to the next
532 that can lead to a segmentation fault. Errors processing the initial
533 ClientHello can trigger this scenario. An example of such an error could be
534 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
537 This issue was reported to OpenSSL by Per Allansson.
541 *) Segmentation fault in ASN1_TYPE_cmp fix
543 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
544 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
545 certificate signature algorithm consistency this can be used to crash any
546 certificate verification operation and exploited in a DoS attack. Any
547 application which performs certificate verification is vulnerable including
548 OpenSSL clients and servers which enable client authentication.
552 *) Segmentation fault for invalid PSS parameters fix
554 The signature verification routines will crash with a NULL pointer
555 dereference if presented with an ASN.1 signature using the RSA PSS
556 algorithm and invalid parameters. Since these routines are used to verify
557 certificate signature algorithms this can be used to crash any
558 certificate verification operation and exploited in a DoS attack. Any
559 application which performs certificate verification is vulnerable including
560 OpenSSL clients and servers which enable client authentication.
562 This issue was was reported to OpenSSL by Brian Carpenter.
566 *) ASN.1 structure reuse memory corruption fix
568 Reusing a structure in ASN.1 parsing may allow an attacker to cause
569 memory corruption via an invalid write. Such reuse is and has been
570 strongly discouraged and is believed to be rare.
572 Applications that parse structures containing CHOICE or ANY DEFINED BY
573 components may be affected. Certificate parsing (d2i_X509 and related
574 functions) are however not affected. OpenSSL clients and servers are
579 *) PKCS7 NULL pointer dereferences fix
581 The PKCS#7 parsing code does not handle missing outer ContentInfo
582 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
583 missing content and trigger a NULL pointer dereference on parsing.
585 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
586 otherwise parse PKCS#7 structures from untrusted sources are
587 affected. OpenSSL clients and servers are not affected.
589 This issue was reported to OpenSSL by Michal Zalewski (Google).
593 *) DoS via reachable assert in SSLv2 servers fix
595 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
596 servers that both support SSLv2 and enable export cipher suites by sending
597 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
599 This issue was discovered by Sean Burford (Google) and Emilia Käsper
600 (OpenSSL development team).
604 *) Empty CKE with client auth and DHE fix
606 If client auth is used then a server can seg fault in the event of a DHE
607 ciphersuite being selected and a zero length ClientKeyExchange message
608 being sent by the client. This could be exploited in a DoS attack.
612 *) Handshake with unseeded PRNG fix
614 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
615 with an unseeded PRNG. The conditions are:
616 - The client is on a platform where the PRNG has not been seeded
617 automatically, and the user has not seeded manually
618 - A protocol specific client method version has been used (i.e. not
619 SSL_client_methodv23)
620 - A ciphersuite is used that does not require additional random data from
621 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
623 If the handshake succeeds then the client random that has been used will
624 have been generated from a PRNG with insufficient entropy and therefore the
625 output may be predictable.
627 For example using the following command with an unseeded openssl will
628 succeed on an unpatched platform:
630 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
634 *) Use After Free following d2i_ECPrivatekey error fix
636 A malformed EC private key file consumed via the d2i_ECPrivateKey function
637 could cause a use after free condition. This, in turn, could cause a double
638 free in several private key parsing functions (such as d2i_PrivateKey
639 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
640 for applications that receive EC private keys from untrusted
641 sources. This scenario is considered rare.
643 This issue was discovered by the BoringSSL project and fixed in their
648 *) X509_to_X509_REQ NULL pointer deref fix
650 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
651 the certificate key is invalid. This function is rarely used in practice.
653 This issue was discovered by Brian Carpenter.
657 *) Removed the export ciphers from the DEFAULT ciphers
660 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
662 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
663 ARMv5 through ARMv8, as opposite to "locking" it to single one.
664 So far those who have to target multiple plaforms would compromise
665 and argue that binary targeting say ARMv5 would still execute on
666 ARMv8. "Universal" build resolves this compromise by providing
667 near-optimal performance even on newer platforms.
670 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
671 (other platforms pending).
672 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
674 *) Add support for the SignedCertificateTimestampList certificate and
675 OCSP response extensions from RFC6962.
678 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
679 for corner cases. (Certain input points at infinity could lead to
680 bogus results, with non-infinity inputs mapped to infinity too.)
683 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
684 This covers AES, SHA256/512 and GHASH. "Initial" means that most
685 common cases are optimized and there still is room for further
686 improvements. Vector Permutation AES for Altivec is also added.
689 *) Add support for little-endian ppc64 Linux target.
690 [Marcelo Cerri (IBM)]
692 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
693 SHA1, SHA256 and GHASH. "Initial" means that most common cases
694 are optimized and there still is room for further improvements.
695 Both 32- and 64-bit modes are supported.
696 [Andy Polyakov, Ard Biesheuvel (Linaro)]
698 *) Improved ARMv7 NEON support.
701 *) Support for SPARC Architecture 2011 crypto extensions, first
702 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
703 SHA256/512, MD5, GHASH and modular exponentiation.
704 [Andy Polyakov, David Miller]
706 *) Accelerated modular exponentiation for Intel processors, a.k.a.
708 [Shay Gueron & Vlad Krasnov (Intel Corp)]
710 *) Support for new and upcoming Intel processors, including AVX2,
711 BMI and SHA ISA extensions. This includes additional "stitched"
712 implementations, AESNI-SHA256 and GCM, and multi-buffer support
715 This work was sponsored by Intel Corp.
718 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
719 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
720 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
723 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
724 this fixes a limiation in previous versions of OpenSSL.
727 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
728 MGF1 digest and OAEP label.
731 *) Add EVP support for key wrapping algorithms, to avoid problems with
732 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
733 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
734 algorithms and include tests cases.
737 *) Add functions to allocate and set the fields of an ECDSA_METHOD
739 [Douglas E. Engert, Steve Henson]
741 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
742 difference in days and seconds between two tm or ASN1_TIME structures.
745 *) Add -rev test option to s_server to just reverse order of characters
746 received by client and send back to server. Also prints an abbreviated
747 summary of the connection parameters.
750 *) New option -brief for s_client and s_server to print out a brief summary
751 of connection parameters.
754 *) Add callbacks for arbitrary TLS extensions.
755 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
757 *) New option -crl_download in several openssl utilities to download CRLs
758 from CRLDP extension in certificates.
761 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
764 *) New function X509_CRL_diff to generate a delta CRL from the difference
765 of two full CRLs. Add support to "crl" utility.
768 *) New functions to set lookup_crls function and to retrieve
769 X509_STORE from X509_STORE_CTX.
772 *) Print out deprecated issuer and subject unique ID fields in
776 *) Extend OCSP I/O functions so they can be used for simple general purpose
777 HTTP as well as OCSP. New wrapper function which can be used to download
778 CRLs using the OCSP API.
781 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
784 *) SSL_CONF* functions. These provide a common framework for application
785 configuration using configuration files or command lines.
788 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
789 message callback and prints the results. Needs compile time option
790 "enable-ssl-trace". New options to s_client and s_server to enable
794 *) New ctrl and macro to retrieve supported points extensions.
795 Print out extension in s_server and s_client.
798 *) New functions to retrieve certificate signature and signature
802 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
806 *) New Suite B modes for TLS code. These use and enforce the requirements
807 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
808 only use Suite B curves. The Suite B modes can be set by using the
809 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
812 *) New chain verification flags for Suite B levels of security. Check
813 algorithms are acceptable when flags are set in X509_verify_cert.
816 *) Make tls1_check_chain return a set of flags indicating checks passed
817 by a certificate chain. Add additional tests to handle client
818 certificates: checks for matching certificate type and issuer name
822 *) If an attempt is made to use a signature algorithm not in the peer
823 preference list abort the handshake. If client has no suitable
824 signature algorithms in response to a certificate request do not
828 *) If server EC tmp key is not in client preference list abort handshake.
831 *) Add support for certificate stores in CERT structure. This makes it
832 possible to have different stores per SSL structure or one store in
833 the parent SSL_CTX. Include distint stores for certificate chain
834 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
835 to build and store a certificate chain in CERT structure: returing
836 an error if the chain cannot be built: this will allow applications
837 to test if a chain is correctly configured.
839 Note: if the CERT based stores are not set then the parent SSL_CTX
840 store is used to retain compatibility with existing behaviour.
844 *) New function ssl_set_client_disabled to set a ciphersuite disabled
845 mask based on the current session, check mask when sending client
846 hello and checking the requested ciphersuite.
849 *) New ctrls to retrieve and set certificate types in a certificate
850 request message. Print out received values in s_client. If certificate
851 types is not set with custom values set sensible values based on
852 supported signature algorithms.
855 *) Support for distinct client and server supported signature algorithms.
858 *) Add certificate callback. If set this is called whenever a certificate
859 is required by client or server. An application can decide which
860 certificate chain to present based on arbitrary criteria: for example
861 supported signature algorithms. Add very simple example to s_server.
862 This fixes many of the problems and restrictions of the existing client
863 certificate callback: for example you can now clear an existing
864 certificate and specify the whole chain.
867 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
868 the certificate can be used for (if anything). Set valid_flags field
869 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
870 to have similar checks in it.
872 Add new "cert_flags" field to CERT structure and include a "strict mode".
873 This enforces some TLS certificate requirements (such as only permitting
874 certificate signature algorithms contained in the supported algorithms
875 extension) which some implementations ignore: this option should be used
876 with caution as it could cause interoperability issues.
879 *) Update and tidy signature algorithm extension processing. Work out
880 shared signature algorithms based on preferences and peer algorithms
881 and print them out in s_client and s_server. Abort handshake if no
882 shared signature algorithms.
885 *) Add new functions to allow customised supported signature algorithms
886 for SSL and SSL_CTX structures. Add options to s_client and s_server
890 *) New function SSL_certs_clear() to delete all references to certificates
891 from an SSL structure. Before this once a certificate had been added
892 it couldn't be removed.
895 *) Integrate hostname, email address and IP address checking with certificate
896 verification. New verify options supporting checking in opensl utility.
899 *) Fixes and wildcard matching support to hostname and email checking
900 functions. Add manual page.
901 [Florian Weimer (Red Hat Product Security Team)]
903 *) New functions to check a hostname email or IP address against a
904 certificate. Add options x509 utility to print results of checks against
908 *) Fix OCSP checking.
909 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
911 *) Initial experimental support for explicitly trusted non-root CAs.
912 OpenSSL still tries to build a complete chain to a root but if an
913 intermediate CA has a trust setting included that is used. The first
914 setting is used: whether to trust (e.g., -addtrust option to the x509
918 *) Add -trusted_first option which attempts to find certificates in the
919 trusted store even if an untrusted chain is also supplied.
922 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
923 platform support for Linux and Android.
926 *) Support for linux-x32, ILP32 environment in x86_64 framework.
929 *) Experimental multi-implementation support for FIPS capable OpenSSL.
930 When in FIPS mode the approved implementations are used as normal,
931 when not in FIPS mode the internal unapproved versions are used instead.
932 This means that the FIPS capable OpenSSL isn't forced to use the
933 (often lower performance) FIPS implementations outside FIPS mode.
936 *) Transparently support X9.42 DH parameters when calling
937 PEM_read_bio_DHparameters. This means existing applications can handle
938 the new parameter format automatically.
941 *) Initial experimental support for X9.42 DH parameter format: mainly
942 to support use of 'q' parameter for RFC5114 parameters.
945 *) Add DH parameters from RFC5114 including test data to dhtest.
948 *) Support for automatic EC temporary key parameter selection. If enabled
949 the most preferred EC parameters are automatically used instead of
950 hardcoded fixed parameters. Now a server just has to call:
951 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
952 support ECDH and use the most appropriate parameters.
955 *) Enhance and tidy EC curve and point format TLS extension code. Use
956 static structures instead of allocation if default values are used.
957 New ctrls to set curves we wish to support and to retrieve shared curves.
958 Print out shared curves in s_server. New options to s_server and s_client
959 to set list of supported curves.
962 *) New ctrls to retrieve supported signature algorithms and
963 supported curve values as an array of NIDs. Extend openssl utility
964 to print out received values.
967 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
968 between NIDs and the more common NIST names such as "P-256". Enhance
969 ecparam utility and ECC method to recognise the NIST names for curves.
972 *) Enhance SSL/TLS certificate chain handling to support different
973 chains for each certificate instead of one chain in the parent SSL_CTX.
976 *) Support for fixed DH ciphersuite client authentication: where both
977 server and client use DH certificates with common parameters.
980 *) Support for fixed DH ciphersuites: those requiring DH server
984 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
986 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
987 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
988 X509_CINF_get_signature were reverted post internal team review.
990 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
992 *) Build fixes for the Windows and OpenVMS platforms
993 [Matt Caswell and Richard Levitte]
995 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
997 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
998 message can cause a segmentation fault in OpenSSL due to a NULL pointer
999 dereference. This could lead to a Denial Of Service attack. Thanks to
1000 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1004 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1005 dtls1_buffer_record function under certain conditions. In particular this
1006 could occur if an attacker sent repeated DTLS records with the same
1007 sequence number but for the next epoch. The memory leak could be exploited
1008 by an attacker in a Denial of Service attack through memory exhaustion.
1009 Thanks to Chris Mueller for reporting this issue.
1013 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1014 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1015 method would be set to NULL which could later result in a NULL pointer
1016 dereference. Thanks to Frank Schmirler for reporting this issue.
1020 *) Abort handshake if server key exchange message is omitted for ephemeral
1023 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1024 reporting this issue.
1028 *) Remove non-export ephemeral RSA code on client and server. This code
1029 violated the TLS standard by allowing the use of temporary RSA keys in
1030 non-export ciphersuites and could be used by a server to effectively
1031 downgrade the RSA key length used to a value smaller than the server
1032 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1033 INRIA or reporting this issue.
1037 *) Fixed issue where DH client certificates are accepted without verification.
1038 An OpenSSL server will accept a DH certificate for client authentication
1039 without the certificate verify message. This effectively allows a client to
1040 authenticate without the use of a private key. This only affects servers
1041 which trust a client certificate authority which issues certificates
1042 containing DH keys: these are extremely rare and hardly ever encountered.
1043 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1048 *) Ensure that the session ID context of an SSL is updated when its
1049 SSL_CTX is updated via SSL_set_SSL_CTX.
1051 The session ID context is typically set from the parent SSL_CTX,
1052 and can vary with the CTX.
1055 *) Fix various certificate fingerprint issues.
1057 By using non-DER or invalid encodings outside the signed portion of a
1058 certificate the fingerprint can be changed without breaking the signature.
1059 Although no details of the signed portion of the certificate can be changed
1060 this can cause problems with some applications: e.g. those using the
1061 certificate fingerprint for blacklists.
1063 1. Reject signatures with non zero unused bits.
1065 If the BIT STRING containing the signature has non zero unused bits reject
1066 the signature. All current signature algorithms require zero unused bits.
1068 2. Check certificate algorithm consistency.
1070 Check the AlgorithmIdentifier inside TBS matches the one in the
1071 certificate signature. NB: this will result in signature failure
1072 errors for some broken certificates.
1074 Thanks to Konrad Kraszewski from Google for reporting this issue.
1076 3. Check DSA/ECDSA signatures use DER.
1078 Reencode DSA/ECDSA signatures and compare with the original received
1079 signature. Return an error if there is a mismatch.
1081 This will reject various cases including garbage after signature
1082 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1083 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1084 (negative or with leading zeroes).
1086 Further analysis was conducted and fixes were developed by Stephen Henson
1087 of the OpenSSL core team.
1092 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1093 results on some platforms, including x86_64. This bug occurs at random
1094 with a very low probability, and is not known to be exploitable in any
1095 way, though its exact impact is difficult to determine. Thanks to Pieter
1096 Wuille (Blockstream) who reported this issue and also suggested an initial
1097 fix. Further analysis was conducted by the OpenSSL development team and
1098 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1099 the OpenSSL core team.
1103 *) Do not resume sessions on the server if the negotiated protocol
1104 version does not match the session's version. Resuming with a different
1105 version, while not strictly forbidden by the RFC, is of questionable
1106 sanity and breaks all known clients.
1107 [David Benjamin, Emilia Käsper]
1109 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1110 early CCS messages during renegotiation. (Note that because
1111 renegotiation is encrypted, this early CCS was not exploitable.)
1114 *) Tighten client-side session ticket handling during renegotiation:
1115 ensure that the client only accepts a session ticket if the server sends
1116 the extension anew in the ServerHello. Previously, a TLS client would
1117 reuse the old extension state and thus accept a session ticket if one was
1118 announced in the initial ServerHello.
1120 Similarly, ensure that the client requires a session ticket if one
1121 was advertised in the ServerHello. Previously, a TLS client would
1122 ignore a missing NewSessionTicket message.
1125 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1127 *) SRTP Memory Leak.
1129 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1130 sends a carefully crafted handshake message, to cause OpenSSL to fail
1131 to free up to 64k of memory causing a memory leak. This could be
1132 exploited in a Denial Of Service attack. This issue affects OpenSSL
1133 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1134 whether SRTP is used or configured. Implementations of OpenSSL that
1135 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1137 The fix was developed by the OpenSSL team.
1141 *) Session Ticket Memory Leak.
1143 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1144 integrity of that ticket is first verified. In the event of a session
1145 ticket integrity check failing, OpenSSL will fail to free memory
1146 causing a memory leak. By sending a large number of invalid session
1147 tickets an attacker could exploit this issue in a Denial Of Service
1152 *) Build option no-ssl3 is incomplete.
1154 When OpenSSL is configured with "no-ssl3" as a build option, servers
1155 could accept and complete a SSL 3.0 handshake, and clients could be
1156 configured to send them.
1158 [Akamai and the OpenSSL team]
1160 *) Add support for TLS_FALLBACK_SCSV.
1161 Client applications doing fallback retries should call
1162 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1164 [Adam Langley, Bodo Moeller]
1166 *) Add additional DigestInfo checks.
1168 Reencode DigestInto in DER and check against the original when
1169 verifying RSA signature: this will reject any improperly encoded
1170 DigestInfo structures.
1172 Note: this is a precautionary measure and no attacks are currently known.
1176 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1178 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1179 SRP code can be overrun an internal buffer. Add sanity check that
1180 g, A, B < N to SRP code.
1182 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1183 Group for discovering this issue.
1187 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1188 TLS 1.0 instead of higher protocol versions when the ClientHello message
1189 is badly fragmented. This allows a man-in-the-middle attacker to force a
1190 downgrade to TLS 1.0 even if both the server and the client support a
1191 higher protocol version, by modifying the client's TLS records.
1193 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1194 researching this issue.
1198 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1199 to a denial of service attack. A malicious server can crash the client
1200 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1201 ciphersuite and sending carefully crafted handshake messages.
1203 Thanks to Felix Gröbert (Google) for discovering and researching this
1208 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1209 to leak memory. This can be exploited through a Denial of Service attack.
1210 Thanks to Adam Langley for discovering and researching this issue.
1214 *) An attacker can force openssl to consume large amounts of memory whilst
1215 processing DTLS handshake messages. This can be exploited through a
1216 Denial of Service attack.
1217 Thanks to Adam Langley for discovering and researching this issue.
1221 *) An attacker can force an error condition which causes openssl to crash
1222 whilst processing DTLS packets due to memory being freed twice. This
1223 can be exploited through a Denial of Service attack.
1224 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1229 *) If a multithreaded client connects to a malicious server using a resumed
1230 session and the server sends an ec point format extension it could write
1231 up to 255 bytes to freed memory.
1233 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1238 *) A malicious server can crash an OpenSSL client with a null pointer
1239 dereference (read) by specifying an SRP ciphersuite even though it was not
1240 properly negotiated with the client. This can be exploited through a
1241 Denial of Service attack.
1243 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1244 discovering and researching this issue.
1248 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1249 X509_name_oneline, X509_name_print_ex et al. to leak some information
1250 from the stack. Applications may be affected if they echo pretty printing
1251 output to the attacker.
1253 Thanks to Ivan Fratric (Google) for discovering this issue.
1255 [Emilia Käsper, and Steve Henson]
1257 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1258 for corner cases. (Certain input points at infinity could lead to
1259 bogus results, with non-infinity inputs mapped to infinity too.)
1262 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1264 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1265 handshake can force the use of weak keying material in OpenSSL
1266 SSL/TLS clients and servers.
1268 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1269 researching this issue. (CVE-2014-0224)
1270 [KIKUCHI Masashi, Steve Henson]
1272 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1273 OpenSSL DTLS client the code can be made to recurse eventually crashing
1276 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1278 [Imre Rad, Steve Henson]
1280 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1281 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1282 client or server. This is potentially exploitable to run arbitrary
1283 code on a vulnerable client or server.
1285 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1286 [Jüri Aedla, Steve Henson]
1288 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1289 are subject to a denial of service attack.
1291 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1292 this issue. (CVE-2014-3470)
1293 [Felix Gröbert, Ivan Fratric, Steve Henson]
1295 *) Harmonize version and its documentation. -f flag is used to display
1297 [mancha <mancha1@zoho.com>]
1299 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1300 in i2d_ECPrivateKey.
1301 [mancha <mancha1@zoho.com>]
1303 *) Fix some double frees. These are not thought to be exploitable.
1304 [mancha <mancha1@zoho.com>]
1306 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1308 *) A missing bounds check in the handling of the TLS heartbeat extension
1309 can be used to reveal up to 64k of memory to a connected client or
1312 Thanks for Neel Mehta of Google Security for discovering this bug and to
1313 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1314 preparing the fix (CVE-2014-0160)
1315 [Adam Langley, Bodo Moeller]
1317 *) Fix for the attack described in the paper "Recovering OpenSSL
1318 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1319 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1320 http://eprint.iacr.org/2014/140
1322 Thanks to Yuval Yarom and Naomi Benger for discovering this
1323 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1324 [Yuval Yarom and Naomi Benger]
1326 *) TLS pad extension: draft-agl-tls-padding-03
1328 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1329 TLS client Hello record length value would otherwise be > 255 and
1330 less that 512 pad with a dummy extension containing zeroes so it
1331 is at least 512 bytes long.
1333 [Adam Langley, Steve Henson]
1335 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1337 *) Fix for TLS record tampering bug. A carefully crafted invalid
1338 handshake could crash OpenSSL with a NULL pointer exception.
1339 Thanks to Anton Johansson for reporting this issues.
1342 *) Keep original DTLS digest and encryption contexts in retransmission
1343 structures so we can use the previous session parameters if they need
1344 to be resent. (CVE-2013-6450)
1347 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1348 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1349 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1350 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1351 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1352 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1353 [Rob Stradling, Adam Langley]
1355 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1357 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1358 supporting platforms or when small records were transferred.
1359 [Andy Polyakov, Steve Henson]
1361 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1363 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1365 This addresses the flaw in CBC record processing discovered by
1366 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1367 at: http://www.isg.rhul.ac.uk/tls/
1369 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1370 Security Group at Royal Holloway, University of London
1371 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1372 Emilia Käsper for the initial patch.
1374 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1376 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1377 ciphersuites which can be exploited in a denial of service attack.
1378 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1379 and detecting this bug and to Wolfgang Ettlinger
1380 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1384 *) Return an error when checking OCSP signatures when key is NULL.
1385 This fixes a DoS attack. (CVE-2013-0166)
1388 *) Make openssl verify return errors.
1389 [Chris Palmer <palmer@google.com> and Ben Laurie]
1391 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1392 the right response is stapled. Also change SSL_get_certificate()
1393 so it returns the certificate actually sent.
1394 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1395 [Rob Stradling <rob.stradling@comodo.com>]
1397 *) Fix possible deadlock when decoding public keys.
1400 *) Don't use TLS 1.0 record version number in initial client hello
1404 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1406 *) Sanity check record length before skipping explicit IV in TLS
1407 1.2, 1.1 and DTLS to fix DoS attack.
1409 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1410 fuzzing as a service testing platform.
1414 *) Initialise tkeylen properly when encrypting CMS messages.
1415 Thanks to Solar Designer of Openwall for reporting this issue.
1418 *) In FIPS mode don't try to use composite ciphers as they are not
1422 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1424 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1425 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1426 mean any application compiled against OpenSSL 1.0.0 headers setting
1427 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1428 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1429 0x10000000L Any application which was previously compiled against
1430 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1431 will need to be recompiled as a result. Letting be results in
1432 inability to disable specifically TLS 1.1 and in client context,
1433 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1436 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1437 disable just protocol X, but all protocols above X *if* there are
1438 protocols *below* X still enabled. In more practical terms it means
1439 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1440 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1441 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1445 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1447 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1448 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1449 in CRYPTO_realloc_clean.
1451 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1452 issue and to Adam Langley <agl@chromium.org> for fixing it.
1454 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1456 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1459 *) Workarounds for some broken servers that "hang" if a client hello
1460 record length exceeds 255 bytes.
1462 1. Do not use record version number > TLS 1.0 in initial client
1463 hello: some (but not all) hanging servers will now work.
1464 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1465 the number of ciphers sent in the client hello. This should be
1466 set to an even number, such as 50, for example by passing:
1467 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1468 Most broken servers should now work.
1469 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1470 TLS 1.2 client support entirely.
1473 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1476 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1478 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1479 STRING form instead of a DigestInfo.
1482 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1483 and the RSA_sign/RSA_verify functions. This was made more apparent when
1484 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1485 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1486 the correct format in RSA_verify so both forms transparently work.
1489 *) Some servers which support TLS 1.0 can choke if we initially indicate
1490 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1491 encrypted premaster secret. As a workaround use the maximum permitted
1492 client version in client hello, this should keep such servers happy
1493 and still work with previous versions of OpenSSL.
1496 *) Add support for TLS/DTLS heartbeats.
1497 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1499 *) Add support for SCTP.
1500 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1502 *) Improved PRNG seeding for VOS.
1503 [Paul Green <Paul.Green@stratus.com>]
1505 *) Extensive assembler packs updates, most notably:
1507 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1508 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1509 - x86_64: bit-sliced AES implementation;
1510 - ARM: NEON support, contemporary platforms optimizations;
1511 - s390x: z196 support;
1512 - *: GHASH and GF(2^m) multiplication implementations;
1516 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1517 (removal of unnecessary code)
1518 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1520 *) Add TLS key material exporter from RFC 5705.
1523 *) Add DTLS-SRTP negotiation from RFC 5764.
1526 *) Add Next Protocol Negotiation,
1527 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1528 disabled with a no-npn flag to config or Configure. Code donated
1530 [Adam Langley <agl@google.com> and Ben Laurie]
1532 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1533 NIST-P256, NIST-P521, with constant-time single point multiplication on
1534 typical inputs. Compiler support for the nonstandard type __uint128_t is
1535 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1536 Code made available under Apache License version 2.0.
1538 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1539 line to include this in your build of OpenSSL, and run "make depend" (or
1540 "make update"). This enables the following EC_METHODs:
1542 EC_GFp_nistp224_method()
1543 EC_GFp_nistp256_method()
1544 EC_GFp_nistp521_method()
1546 EC_GROUP_new_by_curve_name() will automatically use these (while
1547 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1549 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1551 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1552 all platforms. Move ssize_t definition from e_os.h to the public
1553 header file e_os2.h as it now appears in public header file cms.h
1556 *) New -sigopt option to the ca, req and x509 utilities. Additional
1557 signature parameters can be passed using this option and in
1561 *) Add RSA PSS signing function. This will generate and set the
1562 appropriate AlgorithmIdentifiers for PSS based on those in the
1563 corresponding EVP_MD_CTX structure. No application support yet.
1566 *) Support for companion algorithm specific ASN1 signing routines.
1567 New function ASN1_item_sign_ctx() signs a pre-initialised
1568 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1569 the appropriate parameters.
1572 *) Add new algorithm specific ASN1 verification initialisation function
1573 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1574 handling will be the same no matter what EVP_PKEY_METHOD is used.
1575 Add a PSS handler to support verification of PSS signatures: checked
1576 against a number of sample certificates.
1579 *) Add signature printing for PSS. Add PSS OIDs.
1580 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1582 *) Add algorithm specific signature printing. An individual ASN1 method
1583 can now print out signatures instead of the standard hex dump.
1585 More complex signatures (e.g. PSS) can print out more meaningful
1586 information. Include DSA version that prints out the signature
1590 *) Password based recipient info support for CMS library: implementing
1594 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1595 neatly separates the code into cipher and PBE sections and is required
1596 for some algorithms that split PBES2 into separate pieces (such as
1597 password based CMS).
1600 *) Session-handling fixes:
1601 - Fix handling of connections that are resuming with a session ID,
1602 but also support Session Tickets.
1603 - Fix a bug that suppressed issuing of a new ticket if the client
1604 presented a ticket with an expired session.
1605 - Try to set the ticket lifetime hint to something reasonable.
1606 - Make tickets shorter by excluding irrelevant information.
1607 - On the client side, don't ignore renewed tickets.
1608 [Adam Langley, Bodo Moeller (Google)]
1610 *) Fix PSK session representation.
1613 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1615 This work was sponsored by Intel.
1618 *) Add GCM support to TLS library. Some custom code is needed to split
1619 the IV between the fixed (from PRF) and explicit (from TLS record)
1620 portions. This adds all GCM ciphersuites supported by RFC5288 and
1621 RFC5289. Generalise some AES* cipherstrings to include GCM and
1622 add a special AESGCM string for GCM only.
1625 *) Expand range of ctrls for AES GCM. Permit setting invocation
1626 field on decrypt and retrieval of invocation field only on encrypt.
1629 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1630 As required by RFC5289 these ciphersuites cannot be used if for
1631 versions of TLS earlier than 1.2.
1634 *) For FIPS capable OpenSSL interpret a NULL default public key method
1635 as unset and return the appropriate default but do *not* set the default.
1636 This means we can return the appropriate method in applications that
1637 switch between FIPS and non-FIPS modes.
1640 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1641 ENGINE is used then we cannot handle that in the FIPS module so we
1642 keep original code iff non-FIPS operations are allowed.
1645 *) Add -attime option to openssl utilities.
1646 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1648 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1651 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1652 FIPS EC methods unconditionally for now.
1655 *) New build option no-ec2m to disable characteristic 2 code.
1658 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1659 all cases can be covered as some introduce binary incompatibilities.
1662 *) Redirect RSA operations to FIPS module including keygen,
1663 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1666 *) Add similar low level API blocking to ciphers.
1669 *) Low level digest APIs are not approved in FIPS mode: any attempt
1670 to use these will cause a fatal error. Applications that *really* want
1671 to use them can use the private_* version instead.
1674 *) Redirect cipher operations to FIPS module for FIPS builds.
1677 *) Redirect digest operations to FIPS module for FIPS builds.
1680 *) Update build system to add "fips" flag which will link in fipscanister.o
1681 for static and shared library builds embedding a signature if needed.
1684 *) Output TLS supported curves in preference order instead of numerical
1685 order. This is currently hardcoded for the highest order curves first.
1686 This should be configurable so applications can judge speed vs strength.
1689 *) Add TLS v1.2 server support for client authentication.
1692 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1696 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1697 FIPS modules versions.
1700 *) Add TLS v1.2 client side support for client authentication. Keep cache
1701 of handshake records longer as we don't know the hash algorithm to use
1702 until after the certificate request message is received.
1705 *) Initial TLS v1.2 client support. Add a default signature algorithms
1706 extension including all the algorithms we support. Parse new signature
1707 format in client key exchange. Relax some ECC signing restrictions for
1708 TLS v1.2 as indicated in RFC5246.
1711 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1712 to new signature format when needed using client digest preference.
1713 All server ciphersuites should now work correctly in TLS v1.2. No client
1714 support yet and no support for client certificates.
1717 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1718 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1719 ciphersuites. At present only RSA key exchange ciphersuites work with
1720 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1721 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1722 and version checking.
1725 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1726 with this defined it will not be affected by any changes to ssl internal
1727 structures. Add several utility functions to allow openssl application
1728 to work with OPENSSL_NO_SSL_INTERN defined.
1732 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1734 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1737 *) Permit abbreviated handshakes when renegotiating using the function
1738 SSL_renegotiate_abbreviated().
1739 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1741 *) Add call to ENGINE_register_all_complete() to
1742 ENGINE_load_builtin_engines(), so some implementations get used
1743 automatically instead of needing explicit application support.
1746 *) Add support for TLS key exporter as described in RFC5705.
1747 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1749 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1750 a few changes are required:
1752 Add SSL_OP_NO_TLSv1_1 flag.
1753 Add TLSv1_1 methods.
1754 Update version checking logic to handle version 1.1.
1755 Add explicit IV handling (ported from DTLS code).
1756 Add command line options to s_client/s_server.
1759 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1761 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1762 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1763 content decryption and always return the same error. Note: this attack
1764 needs on average 2^20 messages so it only affects automated senders. The
1765 old behaviour can be reenabled in the CMS code by setting the
1766 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1767 an MMA defence is not necessary.
1768 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1769 this issue. (CVE-2012-0884)
1772 *) Fix CVE-2011-4619: make sure we really are receiving a
1773 client hello before rejecting multiple SGC restarts. Thanks to
1774 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1777 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1779 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1780 Thanks to Antonio Martin, Enterprise Secure Access Research and
1781 Development, Cisco Systems, Inc. for discovering this bug and
1782 preparing a fix. (CVE-2012-0050)
1785 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1787 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1788 of the Vaudenay padding oracle attack on CBC mode encryption
1789 which enables an efficient plaintext recovery attack against
1790 the OpenSSL implementation of DTLS. Their attack exploits timing
1791 differences arising during decryption processing. A research
1792 paper describing this attack can be found at:
1793 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1794 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1795 Security Group at Royal Holloway, University of London
1796 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1797 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1798 for preparing the fix. (CVE-2011-4108)
1799 [Robin Seggelmann, Michael Tuexen]
1801 *) Clear bytes used for block padding of SSL 3.0 records.
1803 [Adam Langley (Google)]
1805 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1806 Kadianakis <desnacked@gmail.com> for discovering this issue and
1807 Adam Langley for preparing the fix. (CVE-2011-4619)
1808 [Adam Langley (Google)]
1810 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1811 [Andrey Kulikov <amdeich@gmail.com>]
1813 *) Prevent malformed RFC3779 data triggering an assertion failure.
1814 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1815 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1816 [Rob Austein <sra@hactrn.net>]
1818 *) Improved PRNG seeding for VOS.
1819 [Paul Green <Paul.Green@stratus.com>]
1821 *) Fix ssl_ciph.c set-up race.
1822 [Adam Langley (Google)]
1824 *) Fix spurious failures in ecdsatest.c.
1825 [Emilia Käsper (Google)]
1827 *) Fix the BIO_f_buffer() implementation (which was mixing different
1828 interpretations of the '..._len' fields).
1829 [Adam Langley (Google)]
1831 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1832 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1833 threads won't reuse the same blinding coefficients.
1835 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1836 lock to call BN_BLINDING_invert_ex, and avoids one use of
1837 BN_BLINDING_update for each BN_BLINDING structure (previously,
1838 the last update always remained unused).
1839 [Emilia Käsper (Google)]
1841 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1842 [Bob Buckholz (Google)]
1844 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1846 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1847 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1848 [Kaspar Brand <ossl@velox.ch>]
1850 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1851 for multi-threaded use of ECDH. (CVE-2011-3210)
1852 [Adam Langley (Google)]
1854 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1857 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1858 signature public key algorithm by using OID xref utilities instead.
1859 Before this you could only use some ECC ciphersuites with SHA1 only.
1862 *) Add protection against ECDSA timing attacks as mentioned in the paper
1863 by Billy Bob Brumley and Nicola Tuveri, see:
1865 http://eprint.iacr.org/2011/232.pdf
1867 [Billy Bob Brumley and Nicola Tuveri]
1869 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1871 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1872 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1874 *) Fix bug in string printing code: if *any* escaping is enabled we must
1875 escape the escape character (backslash) or the resulting string is
1879 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1881 *) Disable code workaround for ancient and obsolete Netscape browsers
1882 and servers: an attacker can use it in a ciphersuite downgrade attack.
1883 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1886 *) Fixed J-PAKE implementation error, originally discovered by
1887 Sebastien Martini, further info and confirmation from Stefan
1888 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1891 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1893 *) Fix extension code to avoid race conditions which can result in a buffer
1894 overrun vulnerability: resumed sessions must not be modified as they can
1895 be shared by multiple threads. CVE-2010-3864
1898 *) Fix WIN32 build system to correctly link an ENGINE directory into
1902 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1904 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1906 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1908 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1910 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1911 context. The operation can be customised via the ctrl mechanism in
1912 case ENGINEs want to include additional functionality.
1915 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1918 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1919 output hashes compatible with older versions of OpenSSL.
1920 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1922 *) Fix compression algorithm handling: if resuming a session use the
1923 compression algorithm of the resumed session instead of determining
1924 it from client hello again. Don't allow server to change algorithm.
1927 *) Add load_crls() function to apps tidying load_certs() too. Add option
1928 to verify utility to allow additional CRLs to be included.
1931 *) Update OCSP request code to permit adding custom headers to the request:
1932 some responders need this.
1935 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1937 [Julia Lawall <julia@diku.dk>]
1939 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1940 needlessly dereferenced structures, used obsolete functions and
1941 didn't handle all updated verify codes correctly.
1944 *) Disable MD2 in the default configuration.
1947 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1948 indicate the initial BIO being pushed or popped. This makes it possible
1949 to determine whether the BIO is the one explicitly called or as a result
1950 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1951 it handles reference counts correctly and doesn't zero out the I/O bio
1952 when it is not being explicitly popped. WARNING: applications which
1953 included workarounds for the old buggy behaviour will need to be modified
1954 or they could free up already freed BIOs.
1957 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1958 renaming to all platforms (within the 0.9.8 branch, this was
1959 done conditionally on Netware platforms to avoid a name clash).
1960 [Guenter <lists@gknw.net>]
1962 *) Add ECDHE and PSK support to DTLS.
1963 [Michael Tuexen <tuexen@fh-muenster.de>]
1965 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1969 *) Add "missing" function EVP_MD_flags() (without this the only way to
1970 retrieve a digest flags is by accessing the structure directly. Update
1971 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1972 or cipher is registered as in the "from" argument. Print out all
1973 registered digests in the dgst usage message instead of manually
1974 attempting to work them out.
1977 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1978 this allows the use of compression and extensions. Change default cipher
1979 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1980 by default unless an application cipher string requests it.
1983 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1984 key ids to find matching certificates and keys but some PKCS#12 files
1985 don't follow the (somewhat unwritten) rules and this strategy fails.
1986 Now just gather all certificates together and the first private key
1987 then look for the first certificate that matches the key.
1990 *) Support use of registered digest and cipher names for dgst and cipher
1991 commands instead of having to add each one as a special case. So now
1998 openssl dgst -sha256 foo
2000 and this works for ENGINE based algorithms too.
2004 *) Update Gost ENGINE to support parameter files.
2005 [Victor B. Wagner <vitus@cryptocom.ru>]
2007 *) Support GeneralizedTime in ca utility.
2008 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2010 *) Enhance the hash format used for certificate directory links. The new
2011 form uses the canonical encoding (meaning equivalent names will work
2012 even if they aren't identical) and uses SHA1 instead of MD5. This form
2013 is incompatible with the older format and as a result c_rehash should
2014 be used to rebuild symbolic links.
2017 *) Make PKCS#8 the default write format for private keys, replacing the
2018 traditional format. This form is standardised, more secure and doesn't
2019 include an implicit MD5 dependency.
2022 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2023 committed to OpenSSL should pass this lot as a minimum.
2026 *) Add session ticket override functionality for use by EAP-FAST.
2027 [Jouni Malinen <j@w1.fi>]
2029 *) Modify HMAC functions to return a value. Since these can be implemented
2030 in an ENGINE errors can occur.
2033 *) Type-checked OBJ_bsearch_ex.
2036 *) Type-checked OBJ_bsearch. Also some constification necessitated
2037 by type-checking. Still to come: TXT_DB, bsearch(?),
2038 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2042 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2043 seconds to a tm structure directly, instead of going through OS
2044 specific date routines. This avoids any issues with OS routines such
2045 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2046 and X509_time_adj_ex() to cover the extended range. The existing
2047 X509_time_adj() is still usable and will no longer have any date issues.
2050 *) Delta CRL support. New use deltas option which will attempt to locate
2051 and search any appropriate delta CRLs available.
2053 This work was sponsored by Google.
2056 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2057 code and add additional score elements. Validate alternate CRL paths
2058 as part of the CRL checking and indicate a new error "CRL path validation
2059 error" in this case. Applications wanting additional details can use
2060 the verify callback and check the new "parent" field. If this is not
2061 NULL CRL path validation is taking place. Existing applications wont
2062 see this because it requires extended CRL support which is off by
2065 This work was sponsored by Google.
2068 *) Support for freshest CRL extension.
2070 This work was sponsored by Google.
2073 *) Initial indirect CRL support. Currently only supported in the CRLs
2074 passed directly and not via lookup. Process certificate issuer
2075 CRL entry extension and lookup CRL entries by bother issuer name
2076 and serial number. Check and process CRL issuer entry in IDP extension.
2078 This work was sponsored by Google.
2081 *) Add support for distinct certificate and CRL paths. The CRL issuer
2082 certificate is validated separately in this case. Only enabled if
2083 an extended CRL support flag is set: this flag will enable additional
2084 CRL functionality in future.
2086 This work was sponsored by Google.
2089 *) Add support for policy mappings extension.
2091 This work was sponsored by Google.
2094 *) Fixes to pathlength constraint, self issued certificate handling,
2095 policy processing to align with RFC3280 and PKITS tests.
2097 This work was sponsored by Google.
2100 *) Support for name constraints certificate extension. DN, email, DNS
2101 and URI types are currently supported.
2103 This work was sponsored by Google.
2106 *) To cater for systems that provide a pointer-based thread ID rather
2107 than numeric, deprecate the current numeric thread ID mechanism and
2108 replace it with a structure and associated callback type. This
2109 mechanism allows a numeric "hash" to be extracted from a thread ID in
2110 either case, and on platforms where pointers are larger than 'long',
2111 mixing is done to help ensure the numeric 'hash' is usable even if it
2112 can't be guaranteed unique. The default mechanism is to use "&errno"
2113 as a pointer-based thread ID to distinguish between threads.
2115 Applications that want to provide their own thread IDs should now use
2116 CRYPTO_THREADID_set_callback() to register a callback that will call
2117 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2119 Note that ERR_remove_state() is now deprecated, because it is tied
2120 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2121 to free the current thread's error state should be replaced by
2122 ERR_remove_thread_state(NULL).
2124 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2125 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2126 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2127 application was previously providing a numeric thread callback that
2128 was inappropriate for distinguishing threads, then uniqueness might
2129 have been obtained with &errno that happened immediately in the
2130 intermediate development versions of OpenSSL; this is no longer the
2131 case, the numeric thread callback will now override the automatic use
2133 [Geoff Thorpe, with help from Bodo Moeller]
2135 *) Initial support for different CRL issuing certificates. This covers a
2136 simple case where the self issued certificates in the chain exist and
2137 the real CRL issuer is higher in the existing chain.
2139 This work was sponsored by Google.
2142 *) Removed effectively defunct crypto/store from the build.
2145 *) Revamp of STACK to provide stronger type-checking. Still to come:
2146 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2147 ASN1_STRING, CONF_VALUE.
2150 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2151 RAM on SSL connections. This option can save about 34k per idle SSL.
2154 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2155 STACK, TXT_DB, bsearch, qsort.
2158 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2159 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2160 support for data, signedData, compressedData, digestedData and
2161 encryptedData, envelopedData types included. Scripts to check against
2162 RFC4134 examples draft and interop and consistency checks of many
2163 content types and variants.
2166 *) Add options to enc utility to support use of zlib compression BIO.
2169 *) Extend mk1mf to support importing of options and assembly language
2170 files from Configure script, currently only included in VC-WIN32.
2171 The assembly language rules can now optionally generate the source
2172 files from the associated perl scripts.
2175 *) Implement remaining functionality needed to support GOST ciphersuites.
2176 Interop testing has been performed using CryptoPro implementations.
2177 [Victor B. Wagner <vitus@cryptocom.ru>]
2179 *) s390x assembler pack.
2182 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2186 *) Implement Opaque PRF Input TLS extension as specified in
2187 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2188 official specification yet and no extension type assignment by
2189 IANA exists, this extension (for now) will have to be explicitly
2190 enabled when building OpenSSL by providing the extension number
2191 to use. For example, specify an option
2193 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2195 to the "config" or "Configure" script to enable the extension,
2196 assuming extension number 0x9527 (which is a completely arbitrary
2197 and unofficial assignment based on the MD5 hash of the Internet
2198 Draft). Note that by doing so, you potentially lose
2199 interoperability with other TLS implementations since these might
2200 be using the same extension number for other purposes.
2202 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2203 opaque PRF input value to use in the handshake. This will create
2204 an interal copy of the length-'len' string at 'src', and will
2205 return non-zero for success.
2207 To get more control and flexibility, provide a callback function
2210 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2211 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2215 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2218 Callback function 'cb' will be called in handshakes, and is
2219 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2220 Argument 'arg' is for application purposes (the value as given to
2221 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2222 be provided to the callback function). The callback function
2223 has to return non-zero to report success: usually 1 to use opaque
2224 PRF input just if possible, or 2 to enforce use of the opaque PRF
2225 input. In the latter case, the library will abort the handshake
2226 if opaque PRF input is not successfully negotiated.
2228 Arguments 'peerinput' and 'len' given to the callback function
2229 will always be NULL and 0 in the case of a client. A server will
2230 see the client's opaque PRF input through these variables if
2231 available (NULL and 0 otherwise). Note that if the server
2232 provides an opaque PRF input, the length must be the same as the
2233 length of the client's opaque PRF input.
2235 Note that the callback function will only be called when creating
2236 a new session (session resumption can resume whatever was
2237 previously negotiated), and will not be called in SSL 2.0
2238 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2239 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2240 for applications that need to enforce opaque PRF input.
2244 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2247 [Victor B. Wagner <vitus@cryptocom.ru>]
2249 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2250 RFC4507bis. The encrypted ticket format is an encrypted encoded
2251 SSL_SESSION structure, that way new session features are automatically
2254 If a client application caches session in an SSL_SESSION structure
2255 support is transparent because tickets are now stored in the encoded
2258 The SSL_CTX structure automatically generates keys for ticket
2259 protection in servers so again support should be possible
2260 with no application modification.
2262 If a client or server wishes to disable RFC4507 support then the option
2263 SSL_OP_NO_TICKET can be set.
2265 Add a TLS extension debugging callback to allow the contents of any client
2266 or server extensions to be examined.
2268 This work was sponsored by Google.
2271 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2272 OpenSSL should now compile cleanly on gcc 4.2
2273 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2275 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2276 support including streaming MAC support: this is required for GOST
2277 ciphersuite support.
2278 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2280 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2281 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2282 to output in BER and PEM format.
2285 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2286 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2287 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2288 ENGINE support for HMAC keys which are unextractable. New -mac and
2289 -macopt options to dgst utility.
2292 *) New option -sigopt to dgst utility. Update dgst to use
2293 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2294 alternative signing paramaters such as X9.31 or PSS in the dgst
2298 *) Change ssl_cipher_apply_rule(), the internal function that does
2299 the work each time a ciphersuite string requests enabling
2300 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2301 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2302 the order of disabled ciphersuites such that those ciphersuites
2303 that most recently went from enabled to disabled not only stay
2304 in order with respect to each other, but also have higher priority
2305 than other disabled ciphersuites the next time ciphersuites are
2308 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2309 the same ciphersuites as with "HIGH" alone, but in a specific
2310 order where the PSK ciphersuites come first (since they are the
2311 most recently disabled ciphersuites when "HIGH" is parsed).
2313 Also, change ssl_create_cipher_list() (using this new
2314 funcionality) such that between otherwise identical
2315 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2319 *) Change ssl_create_cipher_list() so that it automatically
2320 arranges the ciphersuites in reasonable order before starting
2321 to process the rule string. Thus, the definition for "DEFAULT"
2322 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2323 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2324 This makes it much easier to arrive at a reasonable default order
2325 in applications for which anonymous ciphers are OK (meaning
2326 that you can't actually use DEFAULT).
2327 [Bodo Moeller; suggested by Victor Duchovni]
2329 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2330 processing) into multiple integers instead of setting
2331 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2332 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2333 (These masks as well as the individual bit definitions are hidden
2334 away into the non-exported interface ssl/ssl_locl.h, so this
2335 change to the definition of the SSL_CIPHER structure shouldn't
2336 affect applications.) This give us more bits for each of these
2337 categories, so there is no longer a need to coagulate AES128 and
2338 AES256 into a single algorithm bit, and to coagulate Camellia128
2339 and Camellia256 into a single algorithm bit, which has led to all
2342 Thus, among other things, the kludge introduced in 0.9.7m and
2343 0.9.8e for masking out AES256 independently of AES128 or masking
2344 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2346 With the change, we also introduce new ciphersuite aliases that
2347 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2351 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2352 Use the leftmost N bytes of the signature input if the input is
2353 larger than the prime q (with N being the size in bytes of q).
2356 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2357 it yet and it is largely untested.
2360 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2363 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2364 some compilers (gcc 4.2 and later) reject their use. Safestack is
2365 reimplemented. Update ASN1 to avoid use of legacy functions.
2368 *) Win32/64 targets are linked with Winsock2.
2371 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2372 to external functions. This can be used to increase CRL handling
2373 efficiency especially when CRLs are very large by (for example) storing
2374 the CRL revoked certificates in a database.
2377 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2378 new CRLs added to a directory can be used. New command line option
2379 -verify_return_error to s_client and s_server. This causes real errors
2380 to be returned by the verify callback instead of carrying on no matter
2381 what. This reflects the way a "real world" verify callback would behave.
2384 *) GOST engine, supporting several GOST algorithms and public key formats.
2385 Kindly donated by Cryptocom.
2388 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2389 partitioned by DP are handled but no indirect CRL or reason partitioning
2390 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2391 selected via a scoring technique which handles IDP and AKID in CRLs.
2394 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2395 will ultimately be used for all verify operations: this will remove the
2396 X509_STORE dependency on certificate verification and allow alternative
2397 lookup methods. X509_STORE based implementations of these two callbacks.
2400 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2401 Modify get_crl() to find a valid (unexpired) CRL if possible.
2404 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2405 this would be called X509_CRL_cmp() but that name is already used by
2406 a function that just compares CRL issuer names. Cache several CRL
2407 extensions in X509_CRL structure and cache CRLDP in X509.
2410 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2411 this maps equivalent X509_NAME structures into a consistent structure.
2412 Name comparison can then be performed rapidly using memcmp().
2415 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2419 *) Allow digests to supply their own micalg string for S/MIME type using
2420 the ctrl EVP_MD_CTRL_MICALG.
2423 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2424 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2425 ctrl. It can then customise the structure before and/or after signing
2429 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2430 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2431 to free up any added signature OIDs.
2434 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2435 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2436 digest and cipher tables. New options added to openssl utility:
2437 list-message-digest-algorithms and list-cipher-algorithms.
2440 *) Change the array representation of binary polynomials: the list
2441 of degrees of non-zero coefficients is now terminated with -1.
2442 Previously it was terminated with 0, which was also part of the
2443 value; thus, the array representation was not applicable to
2444 polynomials where t^0 has coefficient zero. This change makes
2445 the array representation useful in a more general context.
2448 *) Various modifications and fixes to SSL/TLS cipher string
2449 handling. For ECC, the code now distinguishes between fixed ECDH
2450 with RSA certificates on the one hand and with ECDSA certificates
2451 on the other hand, since these are separate ciphersuites. The
2452 unused code for Fortezza ciphersuites has been removed.
2454 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2455 (not "ECDHE"). For consistency with the code for DH
2456 certificates, use of ECDH certificates is now considered ECDH
2457 authentication, not RSA or ECDSA authentication (the latter is
2458 merely the CA's signing algorithm and not actively used in the
2461 The temporary ciphersuite alias "ECCdraft" is no longer
2462 available, and ECC ciphersuites are no longer excluded from "ALL"
2463 and "DEFAULT". The following aliases now exist for RFC 4492
2464 ciphersuites, most of these by analogy with the DH case:
2466 kECDHr - ECDH cert, signed with RSA
2467 kECDHe - ECDH cert, signed with ECDSA
2468 kECDH - ECDH cert (signed with either RSA or ECDSA)
2469 kEECDH - ephemeral ECDH
2470 ECDH - ECDH cert or ephemeral ECDH
2476 AECDH - anonymous ECDH
2477 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2481 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2482 Use correct micalg parameters depending on digest(s) in signed message.
2485 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2486 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2489 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2490 an engine to register a method. Add ENGINE lookups for methods and
2491 functional reference processing.
2494 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2495 EVP_{Sign,Verify}* which allow an application to customise the signature
2499 *) New -resign option to smime utility. This adds one or more signers
2500 to an existing PKCS#7 signedData structure. Also -md option to use an
2501 alternative message digest algorithm for signing.
2504 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2505 create PKCS7 structures containing multiple signers. Update smime
2506 application to support multiple signers.
2509 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2513 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2514 Reorganize PBE internals to lookup from a static table using NIDs,
2515 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2516 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2517 PRF which will be automatically used with PBES2.
2520 *) Replace the algorithm specific calls to generate keys in "req" with the
2524 *) Update PKCS#7 enveloped data routines to use new API. This is now
2525 supported by any public key method supporting the encrypt operation. A
2526 ctrl is added to allow the public key algorithm to examine or modify
2527 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2531 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2532 a default digest type to use. In most cases this will be SHA1 but some
2533 algorithms (such as GOST) need to specify an alternative digest. The
2534 return value indicates how strong the preference is 1 means optional and
2535 2 is mandatory (that is it is the only supported type). Modify
2536 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2537 use the default md. Update openssl utilities to use the default digest
2538 type for signing if it is not explicitly indicated.
2541 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2542 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2543 signing method from the key type. This effectively removes the link
2544 between digests and public key types.
2547 *) Add an OID cross reference table and utility functions. Its purpose is to
2548 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2549 rsaEncryption. This will allow some of the algorithm specific hackery
2550 needed to use the correct OID to be removed.
2553 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2554 structures for PKCS7_sign(). They are now set up by the relevant public
2558 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2561 *) Add support for key derivation (agreement) in the API, DH method and
2565 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2566 public and private key formats. As a side effect these add additional
2567 command line functionality not previously available: DSA signatures can be
2568 generated and verified using pkeyutl and DH key support and generation in
2573 [Oliver Tappe <zooey@hirschkaefer.de>]
2575 *) New make target "install_html_docs" installs HTML renditions of the
2577 [Oliver Tappe <zooey@hirschkaefer.de>]
2579 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2580 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2581 support key and parameter generation and add initial key generation
2582 functionality for RSA.
2585 *) Add functions for main EVP_PKEY_method operations. The undocumented
2586 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2587 EVP_PKEY_{encrypt,decrypt}_old.
2590 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2591 key API, doesn't do much yet.
2594 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2595 public key algorithms. New option to openssl utility:
2596 "list-public-key-algorithms" to print out info.
2599 *) Implement the Supported Elliptic Curves Extension for
2600 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2603 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2604 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2607 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2608 utilities such as rsa, dsa, dsaparam etc except they process any key
2612 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2613 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2614 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2618 *) Initial support for pluggable public key ASN1.
2619 De-spaghettify the public key ASN1 handling. Move public and private
2620 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2621 algorithm specific handling to a single module within the relevant
2622 algorithm directory. Add functions to allow (near) opaque processing
2623 of public and private key structures.
2626 *) Implement the Supported Point Formats Extension for
2627 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2630 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2631 for the psk identity [hint] and the psk callback functions to the
2632 SSL_SESSION, SSL and SSL_CTX structure.
2635 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2639 SSL_CTX_use_psk_identity_hint
2640 SSL_get_psk_identity_hint
2641 SSL_get_psk_identity
2642 SSL_use_psk_identity_hint
2644 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2646 *) Add RFC 3161 compliant time stamp request creation, response generation
2647 and response verification functionality.
2648 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2650 *) Add initial support for TLS extensions, specifically for the server_name
2651 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2652 have new members for a host name. The SSL data structure has an
2653 additional member SSL_CTX *initial_ctx so that new sessions can be
2654 stored in that context to allow for session resumption, even after the
2655 SSL has been switched to a new SSL_CTX in reaction to a client's
2656 server_name extension.
2658 New functions (subject to change):
2660 SSL_get_servername()
2661 SSL_get_servername_type()
2664 New CTRL codes and macros (subject to change):
2666 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2667 - SSL_CTX_set_tlsext_servername_callback()
2668 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2669 - SSL_CTX_set_tlsext_servername_arg()
2670 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2672 openssl s_client has a new '-servername ...' option.
2674 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2675 '-key2 ...', '-servername_fatal' (subject to change). This allows
2676 testing the HostName extension for a specific single host name ('-cert'
2677 and '-key' remain fallbacks for handshakes without HostName
2678 negotiation). If the unrecognized_name alert has to be sent, this by
2679 default is a warning; it becomes fatal with the '-servername_fatal'
2682 [Peter Sylvester, Remy Allais, Christophe Renou]
2684 *) Whirlpool hash implementation is added.
2687 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2688 bn(64,32). Because of instruction set limitations it doesn't have
2689 any negative impact on performance. This was done mostly in order
2690 to make it possible to share assembler modules, such as bn_mul_mont
2691 implementations, between 32- and 64-bit builds without hassle.
2694 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2695 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2699 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2700 dedicated Montgomery multiplication procedure, is introduced.
2701 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2702 "64-bit" performance on certain 32-bit targets.
2705 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2706 in SSL structures. New SSL ctrl to set maximum send fragment size.
2707 Save memory by seeting the I/O buffer sizes dynamically instead of
2708 using the maximum available value.
2711 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2712 in addition to the text details.
2715 *) Very, very preliminary EXPERIMENTAL support for printing of general
2716 ASN1 structures. This currently produces rather ugly output and doesn't
2717 handle several customised structures at all.
2720 *) Integrated support for PVK file format and some related formats such
2721 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2722 these in the 'rsa' and 'dsa' utilities.
2725 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2728 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2729 place for the (very old) "NETSCAPE" format certificates which are now
2730 handled using new ASN1 code equivalents.
2733 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2734 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2735 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2738 *) Modify CRL distribution points extension code to print out previously
2739 unsupported fields. Enhance extension setting code to allow setting of
2743 *) Add print and set support for Issuing Distribution Point CRL extension.
2746 *) Change 'Configure' script to enable Camellia by default.
2749 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2751 *) When rejecting SSL/TLS records due to an incorrect version number, never
2752 update s->server with a new major version number. As of
2753 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2754 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2755 the previous behavior could result in a read attempt at NULL when
2756 receiving specific incorrect SSL/TLS records once record payload
2757 protection is active. (CVE-2010-0740)
2758 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2760 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2761 could be crashed if the relevant tables were not present (e.g. chrooted).
2762 [Tomas Hoger <thoger@redhat.com>]
2764 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2766 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2767 [Martin Olsson, Neel Mehta]
2769 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2770 accommodate for stack sorting, always a write lock!).
2773 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2774 excessive delays in the RAND_poll(): over a minute. As a workaround
2775 include a time check in the inner Heap32Next loop too.
2778 *) The code that handled flushing of data in SSL/TLS originally used the
2779 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2780 the problem outlined in PR#1949. The fix suggested there however can
2781 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2782 of Apache). So instead simplify the code to flush unconditionally.
2783 This should be fine since flushing with no data to flush is a no op.
2786 *) Handle TLS versions 2.0 and later properly and correctly use the
2787 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2788 off ancient servers have a habit of sticking around for a while...
2791 *) Modify compression code so it frees up structures without using the
2792 ex_data callbacks. This works around a problem where some applications
2793 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2794 restarting) then use compression (e.g. SSL with compression) later.
2795 This results in significant per-connection memory leaks and
2796 has caused some security issues including CVE-2008-1678 and
2800 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2801 change when encrypting or decrypting.
2804 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2805 connect and renegotiate with servers which do not support RI.
2806 Until RI is more widely deployed this option is enabled by default.
2809 *) Add "missing" ssl ctrls to clear options and mode.
2812 *) If client attempts to renegotiate and doesn't support RI respond with
2813 a no_renegotiation alert as required by RFC5746. Some renegotiating
2814 TLS clients will continue a connection gracefully when they receive
2815 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2816 waiting for a server hello which it will never receive. Now we treat a
2817 received no_renegotiation alert as a fatal error. This is because
2818 applications requesting a renegotiation might well expect it to succeed
2819 and would have no code in place to handle the server denying it so the
2820 only safe thing to do is to terminate the connection.
2823 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2824 peer supports secure renegotiation and 0 otherwise. Print out peer
2825 renegotiation support in s_client/s_server.
2828 *) Replace the highly broken and deprecated SPKAC certification method with
2829 the updated NID creation version. This should correctly handle UTF8.
2832 *) Implement RFC5746. Re-enable renegotiation but require the extension
2833 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2834 turns out to be a bad idea. It has been replaced by
2835 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2836 SSL_CTX_set_options(). This is really not recommended unless you
2837 know what you are doing.
2838 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2840 *) Fixes to stateless session resumption handling. Use initial_ctx when
2841 issuing and attempting to decrypt tickets in case it has changed during
2842 servername handling. Use a non-zero length session ID when attempting
2843 stateless session resumption: this makes it possible to determine if
2844 a resumption has occurred immediately after receiving server hello
2845 (several places in OpenSSL subtly assume this) instead of later in
2849 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2850 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2851 fixes for a few places where the return code is not checked
2853 [Julia Lawall <julia@diku.dk>]
2855 *) Add --strict-warnings option to Configure script to include devteam
2856 warnings in other configurations.
2859 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2860 makes it possible to install openssl libraries in locations which
2861 have names other than "lib", for example "/usr/lib64" which some
2863 [Steve Henson, based on patch from Jeremy Utley]
2865 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2866 X690 8.9.12 and can produce some misleading textual output of OIDs.
2867 [Steve Henson, reported by Dan Kaminsky]
2869 *) Delete MD2 from algorithm tables. This follows the recommendation in
2870 several standards that it is not used in new applications due to
2871 several cryptographic weaknesses. For binary compatibility reasons
2872 the MD2 API is still compiled in by default.
2875 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2879 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2880 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2882 [Guenter <lists@gknw.net>]
2884 *) Fix the server certificate chain building code to use X509_verify_cert(),
2885 it used to have an ad-hoc builder which was unable to cope with anything
2886 other than a simple chain.
2887 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2889 *) Don't check self signed certificate signatures in X509_verify_cert()
2890 by default (a flag can override this): it just wastes time without
2891 adding any security. As a useful side effect self signed root CAs
2892 with non-FIPS digests are now usable in FIPS mode.
2895 *) In dtls1_process_out_of_seq_message() the check if the current message
2896 is already buffered was missing. For every new message was memory
2897 allocated, allowing an attacker to perform an denial of service attack
2898 with sending out of seq handshake messages until there is no memory
2899 left. Additionally every future messege was buffered, even if the
2900 sequence number made no sense and would be part of another handshake.
2901 So only messages with sequence numbers less than 10 in advance will be
2902 buffered. (CVE-2009-1378)
2903 [Robin Seggelmann, discovered by Daniel Mentz]
2905 *) Records are buffered if they arrive with a future epoch to be
2906 processed after finishing the corresponding handshake. There is
2907 currently no limitation to this buffer allowing an attacker to perform
2908 a DOS attack with sending records with future epochs until there is no
2909 memory left. This patch adds the pqueue_size() function to determine
2910 the size of a buffer and limits the record buffer to 100 entries.
2912 [Robin Seggelmann, discovered by Daniel Mentz]
2914 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2915 parent structure is freed. (CVE-2009-1379)
2918 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2919 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2922 [Ilya O. <vrghost@gmail.com>]
2924 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2926 *) Disable renegotiation completely - this fixes a severe security
2927 problem (CVE-2009-3555) at the cost of breaking all
2928 renegotiation. Renegotiation can be re-enabled by setting
2929 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2930 run-time. This is really not recommended unless you know what
2934 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2936 *) Don't set val to NULL when freeing up structures, it is freed up by
2937 underlying code. If sizeof(void *) > sizeof(long) this can result in
2938 zeroing past the valid field. (CVE-2009-0789)
2939 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2941 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2942 checked correctly. This would allow some invalid signed attributes to
2943 appear to verify correctly. (CVE-2009-0591)
2944 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2946 *) Reject UniversalString and BMPString types with invalid lengths. This
2947 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2948 a legal length. (CVE-2009-0590)
2951 *) Set S/MIME signing as the default purpose rather than setting it
2952 unconditionally. This allows applications to override it at the store
2956 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2957 to handle some structures.
2960 *) Improve efficiency of mem_gets: don't search whole buffer each time
2962 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2964 *) New -hex option for openssl rand.
2967 *) Print out UTF8String and NumericString when parsing ASN1.
2970 *) Support NumericString type for name components.
2973 *) Allow CC in the environment to override the automatically chosen
2974 compiler. Note that nothing is done to ensure flags work with the
2978 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2980 *) Properly check EVP_VerifyFinal() and similar return values
2982 [Ben Laurie, Bodo Moeller, Google Security Team]
2984 *) Enable TLS extensions by default.
2987 *) Allow the CHIL engine to be loaded, whether the application is
2988 multithreaded or not. (This does not release the developer from the
2989 obligation to set up the dynamic locking callbacks.)
2990 [Sander Temme <sander@temme.net>]
2992 *) Use correct exit code if there is an error in dgst command.
2993 [Steve Henson; problem pointed out by Roland Dirlewanger]
2995 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2996 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2999 *) Add experimental JPAKE support, including demo authentication in
3000 s_client and s_server.
3003 *) Set the comparison function in v3_addr_canonize().
3004 [Rob Austein <sra@hactrn.net>]
3006 *) Add support for XMPP STARTTLS in s_client.
3007 [Philip Paeps <philip@freebsd.org>]
3009 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3010 to ensure that even with this option, only ciphersuites in the
3011 server's preference list will be accepted. (Note that the option
3012 applies only when resuming a session, so the earlier behavior was
3013 just about the algorithm choice for symmetric cryptography.)
3016 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3018 *) Fix NULL pointer dereference if a DTLS server received
3019 ChangeCipherSpec as first record (CVE-2009-1386).
3022 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3023 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3026 *) The fix in 0.9.8c that supposedly got rid of unsafe
3027 double-checked locking was incomplete for RSA blinding,
3028 addressing just one layer of what turns out to have been
3029 doubly unsafe triple-checked locking.
3031 So now fix this for real by retiring the MONT_HELPER macro
3032 in crypto/rsa/rsa_eay.c.
3034 [Bodo Moeller; problem pointed out by Marius Schilder]
3036 *) Various precautionary measures:
3038 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3040 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3041 (NB: This would require knowledge of the secret session ticket key
3042 to exploit, in which case you'd be SOL either way.)
3044 - Change bn_nist.c so that it will properly handle input BIGNUMs
3045 outside the expected range.
3047 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3050 [Neel Mehta, Bodo Moeller]
3052 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3053 the load fails. Useful for distros.
3054 [Ben Laurie and the FreeBSD team]
3056 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3059 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3062 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3064 This work was sponsored by Logica.
3067 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3068 keystores. Support for SSL/TLS client authentication too.
3069 Not compiled unless enable-capieng specified to Configure.
3071 This work was sponsored by Logica.
3074 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3075 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3076 attribute creation routines such as certifcate requests and PKCS#12
3080 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3082 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3083 handshake which could lead to a cilent crash as found using the
3084 Codenomicon TLS test suite (CVE-2008-1672)
3085 [Steve Henson, Mark Cox]
3087 *) Fix double free in TLS server name extensions which could lead to
3088 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3091 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3093 Clear the error queue to ensure that error entries left from
3094 older function calls do not interfere with the correct operation.
3095 [Lutz Jaenicke, Erik de Castro Lopo]
3097 *) Remove root CA certificates of commercial CAs:
3099 The OpenSSL project does not recommend any specific CA and does not
3100 have any policy with respect to including or excluding any CA.
3101 Therefore it does not make any sense to ship an arbitrary selection
3102 of root CA certificates with the OpenSSL software.
3105 *) RSA OAEP patches to fix two separate invalid memory reads.
3106 The first one involves inputs when 'lzero' is greater than
3107 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3108 before the beginning of from). The second one involves inputs where
3109 the 'db' section contains nothing but zeroes (there is a one-byte
3110 invalid read after the end of 'db').
3111 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3113 *) Partial backport from 0.9.9-dev:
3115 Introduce bn_mul_mont (dedicated Montgomery multiplication
3116 procedure) as a candidate for BIGNUM assembler implementation.
3117 While 0.9.9-dev uses assembler for various architectures, only
3118 x86_64 is available by default here in the 0.9.8 branch, and
3119 32-bit x86 is available through a compile-time setting.
3121 To try the 32-bit x86 assembler implementation, use Configure
3122 option "enable-montasm" (which exists only for this backport).
3124 As "enable-montasm" for 32-bit x86 disclaims code stability
3125 anyway, in this constellation we activate additional code
3126 backported from 0.9.9-dev for further performance improvements,
3127 namely BN_from_montgomery_word. (To enable this otherwise,
3128 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3130 [Andy Polyakov (backport partially by Bodo Moeller)]
3132 *) Add TLS session ticket callback. This allows an application to set
3133 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3134 values. This is useful for key rollover for example where several key
3135 sets may exist with different names.
3138 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3139 This was broken until now in 0.9.8 releases, such that the only way
3140 a registered ENGINE could be used (assuming it initialises
3141 successfully on the host) was to explicitly set it as the default
3142 for the relevant algorithms. This is in contradiction with 0.9.7
3143 behaviour and the documentation. With this fix, when an ENGINE is
3144 registered into a given algorithm's table of implementations, the
3145 'uptodate' flag is reset so that auto-discovery will be used next
3146 time a new context for that algorithm attempts to select an
3148 [Ian Lister (tweaked by Geoff Thorpe)]
3150 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3151 implemention in the following ways:
3153 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3156 Lack of BER streaming support means one pass streaming processing is
3157 only supported if data is detached: setting the streaming flag is
3158 ignored for embedded content.
3160 CMS support is disabled by default and must be explicitly enabled
3161 with the enable-cms configuration option.
3164 *) Update the GMP engine glue to do direct copies between BIGNUM and
3165 mpz_t when openssl and GMP use the same limb size. Otherwise the
3166 existing "conversion via a text string export" trick is still used.
3167 [Paul Sheer <paulsheer@gmail.com>]
3169 *) Zlib compression BIO. This is a filter BIO which compressed and
3170 uncompresses any data passed through it.
3173 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3174 RFC3394 compatible AES key wrapping.
3177 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3178 sets string data without copying. X509_ALGOR_set0() and
3179 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3180 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3181 from an X509_ATTRIBUTE structure optionally checking it occurs only
3182 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3186 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3187 to get the expected BN_FLG_CONSTTIME behavior.
3188 [Bodo Moeller (Google)]
3192 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3193 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3194 - added some more tests to do_tests.pl
3195 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3196 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3197 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3198 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3199 - various changes to netware.pl to enable gcc-cross builds on Win32
3201 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3202 - various changes to fix missing prototype warnings
3203 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3204 - added AES, WHIRLPOOL and CPUID assembler code to build files
3205 - added missing AES assembler make rules to mk1mf.pl
3206 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3207 [Guenter Knauf <eflash@gmx.net>]
3209 *) Implement certificate status request TLS extension defined in RFC3546.
3210 A client can set the appropriate parameters and receive the encoded
3211 OCSP response via a callback. A server can query the supplied parameters
3212 and set the encoded OCSP response in the callback. Add simplified examples
3213 to s_client and s_server.
3216 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3218 *) Fix various bugs:
3219 + Binary incompatibility of ssl_ctx_st structure
3220 + DTLS interoperation with non-compliant servers
3221 + Don't call get_session_cb() without proposed session
3222 + Fix ia64 assembler code
3223 [Andy Polyakov, Steve Henson]
3225 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3227 *) DTLS Handshake overhaul. There were longstanding issues with
3228 OpenSSL DTLS implementation, which were making it impossible for
3229 RFC 4347 compliant client to communicate with OpenSSL server.
3230 Unfortunately just fixing these incompatibilities would "cut off"
3231 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3232 server keeps tolerating non RFC compliant syntax. The opposite is
3233 not true, 0.9.8f client can not communicate with earlier server.
3234 This update even addresses CVE-2007-4995.
3237 *) Changes to avoid need for function casts in OpenSSL: some compilers
3238 (gcc 4.2 and later) reject their use.
3239 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3242 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3243 RFC4507bis. The encrypted ticket format is an encrypted encoded
3244 SSL_SESSION structure, that way new session features are automatically
3247 If a client application caches session in an SSL_SESSION structure
3248 support is transparent because tickets are now stored in the encoded
3251 The SSL_CTX structure automatically generates keys for ticket
3252 protection in servers so again support should be possible
3253 with no application modification.
3255 If a client or server wishes to disable RFC4507 support then the option
3256 SSL_OP_NO_TICKET can be set.
3258 Add a TLS extension debugging callback to allow the contents of any client
3259 or server extensions to be examined.
3261 This work was sponsored by Google.
3264 *) Add initial support for TLS extensions, specifically for the server_name
3265 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3266 have new members for a host name. The SSL data structure has an
3267 additional member SSL_CTX *initial_ctx so that new sessions can be
3268 stored in that context to allow for session resumption, even after the
3269 SSL has been switched to a new SSL_CTX in reaction to a client's
3270 server_name extension.
3272 New functions (subject to change):
3274 SSL_get_servername()
3275 SSL_get_servername_type()
3278 New CTRL codes and macros (subject to change):
3280 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3281 - SSL_CTX_set_tlsext_servername_callback()
3282 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3283 - SSL_CTX_set_tlsext_servername_arg()
3284 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3286 openssl s_client has a new '-servername ...' option.
3288 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3289 '-key2 ...', '-servername_fatal' (subject to change). This allows
3290 testing the HostName extension for a specific single host name ('-cert'
3291 and '-key' remain fallbacks for handshakes without HostName
3292 negotiation). If the unrecognized_name alert has to be sent, this by
3293 default is a warning; it becomes fatal with the '-servername_fatal'
3296 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3298 *) Add AES and SSE2 assembly language support to VC++ build.
3301 *) Mitigate attack on final subtraction in Montgomery reduction.
3304 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3305 (which previously caused an internal error).
3308 *) Squeeze another 10% out of IGE mode when in != out.
3311 *) AES IGE mode speedup.
3312 [Dean Gaudet (Google)]
3314 *) Add the Korean symmetric 128-bit cipher SEED (see
3315 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3316 add SEED ciphersuites from RFC 4162:
3318 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3319 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3320 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3321 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3323 To minimize changes between patchlevels in the OpenSSL 0.9.8
3324 series, SEED remains excluded from compilation unless OpenSSL
3325 is configured with 'enable-seed'.
3326 [KISA, Bodo Moeller]
3328 *) Mitigate branch prediction attacks, which can be practical if a
3329 single processor is shared, allowing a spy process to extract
3330 information. For detailed background information, see
3331 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3332 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3333 and Necessary Software Countermeasures"). The core of the change
3334 are new versions BN_div_no_branch() and
3335 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3336 respectively, which are slower, but avoid the security-relevant
3337 conditional branches. These are automatically called by BN_div()
3338 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3339 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3340 remove a conditional branch.
3342 BN_FLG_CONSTTIME is the new name for the previous
3343 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3344 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3345 in the exponent causes BN_mod_exp_mont() to use the alternative
3346 implementation in BN_mod_exp_mont_consttime().) The old name
3347 remains as a deprecated alias.
3349 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3350 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3351 constant-time implementations for more than just exponentiation.
3352 Here too the old name is kept as a deprecated alias.
3354 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3355 the BN_BLINDING structure gets an independent copy of the
3356 modulus. This means that the previous "BIGNUM *m" argument to
3357 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3358 essentially becomes "const BIGNUM *m", although we can't actually
3359 change this in the header file before 0.9.9. It allows
3360 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3361 enable BN_FLG_CONSTTIME.
3363 [Matthew D Wood (Intel Corp)]
3365 *) In the SSL/TLS server implementation, be strict about session ID
3366 context matching (which matters if an application uses a single
3367 external cache for different purposes). Previously,
3368 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3369 set. This did ensure strict client verification, but meant that,
3370 with applications using a single external cache for quite
3371 different requirements, clients could circumvent ciphersuite
3372 restrictions for a given session ID context by starting a session
3373 in a different context.
3376 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3377 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3378 authentication-only ciphersuites.
3381 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3382 not complete and could lead to a possible single byte overflow
3383 (CVE-2007-5135) [Ben Laurie]
3385 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3387 *) Since AES128 and AES256 (and similarly Camellia128 and
3388 Camellia256) share a single mask bit in the logic of
3389 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3390 kludge to work properly if AES128 is available and AES256 isn't
3391 (or if Camellia128 is available and Camellia256 isn't).
3394 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3395 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3396 When a point or a seed is encoded in a BIT STRING, we need to
3397 prevent the removal of trailing zero bits to get the proper DER
3398 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3399 of a NamedBitList, for which trailing 0 bits need to be removed.)
3402 *) Have SSL/TLS server implementation tolerate "mismatched" record
3403 protocol version while receiving ClientHello even if the
3404 ClientHello is fragmented. (The server can't insist on the
3405 particular protocol version it has chosen before the ServerHello
3406 message has informed the client about his choice.)
3409 *) Add RFC 3779 support.
3410 [Rob Austein for ARIN, Ben Laurie]
3412 *) Load error codes if they are not already present instead of using a
3413 static variable. This allows them to be cleanly unloaded and reloaded.
3414 Improve header file function name parsing.
3417 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3418 or CAPABILITY handshake as required by RFCs.
3421 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3423 *) Introduce limits to prevent malicious keys being able to
3424 cause a denial of service. (CVE-2006-2940)
3425 [Steve Henson, Bodo Moeller]
3427 *) Fix ASN.1 parsing of certain invalid structures that can result
3428 in a denial of service. (CVE-2006-2937) [Steve Henson]
3430 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3431 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3433 *) Fix SSL client code which could crash if connecting to a
3434 malicious SSLv2 server. (CVE-2006-4343)
3435 [Tavis Ormandy and Will Drewry, Google Security Team]
3437 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3438 match only those. Before that, "AES256-SHA" would be interpreted
3439 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3440 the same strength classification in 0.9.7h) as we currently only
3441 have a single AES bit in the ciphersuite description bitmap.
3442 That change, however, also applied to ciphersuite strings such as
3443 "RC4-MD5" that intentionally matched multiple ciphersuites --
3444 namely, SSL 2.0 ciphersuites in addition to the more common ones
3445 from SSL 3.0/TLS 1.0.
3447 So we change the selection algorithm again: Naming an explicit
3448 ciphersuite selects this one ciphersuite, and any other similar
3449 ciphersuite (same bitmap) from *other* protocol versions.
3450 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3451 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3453 Since SSL 2.0 does not have any ciphersuites for which the
3454 128/256 bit distinction would be relevant, this works for now.
3455 The proper fix will be to use different bits for AES128 and
3456 AES256, which would have avoided the problems from the beginning;
3457 however, bits are scarce, so we can only do this in a new release
3458 (not just a patchlevel) when we can change the SSL_CIPHER
3459 definition to split the single 'unsigned long mask' bitmap into
3460 multiple values to extend the available space.
3464 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3466 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3467 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3469 *) Add AES IGE and biIGE modes.
3472 *) Change the Unix randomness entropy gathering to use poll() when
3473 possible instead of select(), since the latter has some
3474 undesirable limitations.
3475 [Darryl Miles via Richard Levitte and Bodo Moeller]
3477 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3478 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3479 cannot be implicitly activated as part of, e.g., the "AES" alias.
3480 However, please upgrade to OpenSSL 0.9.9[-dev] for
3481 non-experimental use of the ECC ciphersuites to get TLS extension
3482 support, which is required for curve and point format negotiation
3483 to avoid potential handshake problems.
3486 *) Disable rogue ciphersuites:
3488 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3489 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3490 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3492 The latter two were purportedly from
3493 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3496 Also deactivate the remaining ciphersuites from
3497 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3498 unofficial, and the ID has long expired.
3501 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3502 dual-core machines) and other potential thread-safety issues.
3505 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3506 versions), which is now available for royalty-free use
3507 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3508 Also, add Camellia TLS ciphersuites from RFC 4132.
3510 To minimize changes between patchlevels in the OpenSSL 0.9.8
3511 series, Camellia remains excluded from compilation unless OpenSSL
3512 is configured with 'enable-camellia'.
3515 *) Disable the padding bug check when compression is in use. The padding
3516 bug check assumes the first packet is of even length, this is not
3517 necessarily true if compresssion is enabled and can result in false
3518 positives causing handshake failure. The actual bug test is ancient
3519 code so it is hoped that implementations will either have fixed it by
3520 now or any which still have the bug do not support compression.
3523 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3525 *) When applying a cipher rule check to see if string match is an explicit
3526 cipher suite and only match that one cipher suite if it is.
3529 *) Link in manifests for VC++ if needed.
3530 [Austin Ziegler <halostatue@gmail.com>]
3532 *) Update support for ECC-based TLS ciphersuites according to
3533 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3534 TLS extensions, which are supported starting with the 0.9.9
3535 branch, not in the OpenSSL 0.9.8 branch).
3538 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3539 opaque EVP_CIPHER_CTX handling.
3542 *) Fixes and enhancements to zlib compression code. We now only use
3543 "zlib1.dll" and use the default __cdecl calling convention on Win32
3544 to conform with the standards mentioned here:
3545 http://www.zlib.net/DLL_FAQ.txt
3546 Static zlib linking now works on Windows and the new --with-zlib-include
3547 --with-zlib-lib options to Configure can be used to supply the location
3548 of the headers and library. Gracefully handle case where zlib library
3552 *) Several fixes and enhancements to the OID generation code. The old code
3553 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3554 handle numbers larger than ULONG_MAX, truncated printing and had a
3555 non standard OBJ_obj2txt() behaviour.
3558 *) Add support for building of engines under engine/ as shared libraries
3559 under VC++ build system.
3562 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3563 Hopefully, we will not see any false combination of paths any more.
3566 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3568 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3569 (part of SSL_OP_ALL). This option used to disable the
3570 countermeasure against man-in-the-middle protocol-version
3571 rollback in the SSL 2.0 server implementation, which is a bad
3572 idea. (CVE-2005-2969)
3574 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3575 for Information Security, National Institute of Advanced Industrial
3576 Science and Technology [AIST], Japan)]
3578 *) Add two function to clear and return the verify parameter flags.
3581 *) Keep cipherlists sorted in the source instead of sorting them at
3582 runtime, thus removing the need for a lock.
3585 *) Avoid some small subgroup attacks in Diffie-Hellman.
3586 [Nick Mathewson and Ben Laurie]
3588 *) Add functions for well-known primes.
3591 *) Extended Windows CE support.
3592 [Satoshi Nakamura and Andy Polyakov]
3594 *) Initialize SSL_METHOD structures at compile time instead of during
3595 runtime, thus removing the need for a lock.
3598 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3599 attempting to decrypt each encrypted key in turn. Add support to
3603 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3605 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3608 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3611 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3612 key into the same file any more.
3615 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3618 *) Add -utf8 command line and config file option to 'ca'.
3619 [Stefan <stf@udoma.org]
3621 *) Removed the macro des_crypt(), as it seems to conflict with some
3622 libraries. Use DES_crypt().
3625 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3626 involves renaming the source and generated shared-libs for
3627 both. The engines will accept the corrected or legacy ids
3628 ('ncipher' and '4758_cca' respectively) when binding. NB,
3629 this only applies when building 'shared'.
3630 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3632 *) Add attribute functions to EVP_PKEY structure. Modify
3633 PKCS12_create() to recognize a CSP name attribute and
3634 use it. Make -CSP option work again in pkcs12 utility.
3637 *) Add new functionality to the bn blinding code:
3638 - automatic re-creation of the BN_BLINDING parameters after
3639 a fixed number of uses (currently 32)
3640 - add new function for parameter creation
3641 - introduce flags to control the update behaviour of the
3642 BN_BLINDING parameters
3643 - hide BN_BLINDING structure
3644 Add a second BN_BLINDING slot to the RSA structure to improve
3645 performance when a sing
projects / openssl.git / blob