5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) State machine rewrite. The state machine code has been significantly
8 refactored in order to remove much duplication of code and solve issues
9 with the old code (see ssl/statem/README for further details). This change
10 does have some associated API changes. Notably SSL_get_state/SSL_state now
11 returns an "enum HANDSHAKE_STATE" instead of an int. The previous handshake
12 states defined in ssl.h and ssl3.h have been redefined to be the nearest
13 equivalent HANDSHAKE_STATE value. Not all states have an equivalent value,
14 (e.g. SSL_ST_CW_FLUSH). New application code should not use the old
15 handshake state values, but should instead use HANDSHAKE_STATE.
18 *) The demo files in crypto/threads were moved to demo/threads.
21 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron and sureware.
24 *) New ASN.1 embed macro.
26 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
27 structure is not allocated: it is part of the parent. That is instead of
35 This reduces memory fragmentation and make it impossible to accidentally
36 set a mandatory field to NULL.
38 This currently only works for some fields specifically a SEQUENCE, CHOICE,
39 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
40 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
44 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
47 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
48 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
49 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
50 DES and RC4 ciphersuites.
53 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
54 This changes the decoding behaviour for some invalid messages,
55 though the change is mostly in the more lenient direction, and
56 legacy behaviour is preserved as much as possible.
58 *) Fix no-stdio build.
59 [ David Woodhouse <David.Woodhouse@intel.com> and also
60 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
62 *) New testing framework
63 The testing framework has been largely rewritten and is now using
64 perl and the perl modules Test::Harness and an extended variant of
65 Test::More called OpenSSL::Test to do its work. All test scripts in
66 test/ have been rewritten into test recipes, and all direct calls to
67 executables in test/Makefile have become individual recipes using the
68 simplified testing OpenSSL::Test::Simple.
70 For documentation on our testing modules, do:
72 perldoc test/testlib/OpenSSL/Test/Simple.pm
73 perldoc test/testlib/OpenSSL/Test.pm
77 *) In DSA_generate_parameters_ex, if the provided seed is too short,
79 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
81 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
82 from RFC4279, RFC4785, RFC5487, RFC5489.
84 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
85 original RSA_PSK patch.
88 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
89 era flag was never set throughout the codebase (only read). Also removed
90 SSL3_FLAGS_POP_BUFFER which was only used if
91 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
94 *) Changed the default name options in the "ca", "crl", "req" and "x509"
95 to be "oneline" instead of "compat".
98 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
99 not aware of clients that still exhibit this bug, and the workaround
100 hasn't been working properly for a while.
103 *) The return type of BIO_number_read() and BIO_number_written() as well as
104 the corresponding num_read and num_write members in the BIO structure has
105 changed from unsigned long to uint64_t. On platforms where an unsigned
106 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
110 *) Given the pervasive nature of TLS extensions it is inadvisable to run
111 OpenSSL without support for them. It also means that maintaining
112 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
113 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
116 *) Removed support for the two export grade static DH ciphersuites
117 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
118 were newly added (along with a number of other static DH ciphersuites) to
119 1.0.2. However the two export ones have *never* worked since they were
120 introduced. It seems strange in any case to be adding new export
121 ciphersuites, and given "logjam" it also does not seem correct to fix them.
124 *) Version negotiation has been rewritten. In particular SSLv23_method(),
125 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
126 and turned into macros which simply call the new preferred function names
127 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
128 should use the new names instead. Also as part of this change the ssl23.h
129 header file has been removed.
132 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
133 code and the associated standard is no longer considered fit-for-purpose.
136 *) RT2547 was closed. When generating a private key, try to make the
137 output file readable only by the owner. This behavior change might
138 be noticeable when interacting with other software.
140 *) Added HTTP GET support to the ocsp command.
143 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
146 *) Added support for TLS extended master secret from
147 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
148 initial patch which was a great help during development.
151 *) All libssl internal structures have been removed from the public header
152 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
153 now redundant). Users should not attempt to access internal structures
154 directly. Instead they should use the provided API functions.
157 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
158 Access to deprecated functions can be re-enabled by running config with
159 "enable-deprecated". In addition applications wishing to use deprecated
160 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
161 will, by default, disable some transitive includes that previously existed
162 in the header files (e.g. ec.h will no longer, by default, include bn.h)
165 *) Added support for OCB mode. OpenSSL has been granted a patent license
166 compatible with the OpenSSL license for use of OCB. Details are available
167 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
168 for OCB can be removed by calling config with no-ocb.
171 *) SSLv2 support has been removed. It still supports receiving a SSLv2
172 compatible client hello.
175 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
176 done while fixing the error code for the key-too-small case.
177 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
179 *) CA.sh has been removmed; use CA.pl instead.
182 *) Removed old DES API.
185 *) Remove various unsupported platforms:
191 Sinix/ReliantUNIX RM400
196 16-bit platforms such as WIN16
199 *) Clean up OPENSSL_NO_xxx #define's
200 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
201 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
202 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
203 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
204 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
205 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
206 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
207 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
208 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
209 Remove MS_STATIC; it's a relic from platforms <32 bits.
212 *) Cleaned up dead code
213 Remove all but one '#ifdef undef' which is to be looked at.
216 *) Clean up calling of xxx_free routines.
217 Just like free(), fix most of the xxx_free routines to accept
218 NULL. Remove the non-null checks from callers. Save much code.
221 *) Add secure heap for storage of private keys (when possible).
222 Add BIO_s_secmem(), CBIGNUM, etc.
223 Contributed by Akamai Technologies under our Corporate CLA.
226 *) Experimental support for a new, fast, unbiased prime candidate generator,
227 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
228 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
230 *) New output format NSS in the sess_id command line tool. This allows
231 exporting the session id and the master key in NSS keylog format.
232 [Martin Kaiser <martin@kaiser.cx>]
234 *) Harmonize version and its documentation. -f flag is used to display
236 [mancha <mancha1@zoho.com>]
238 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
239 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
240 [mancha <mancha1@zoho.com>]
242 *) Fix some double frees. These are not thought to be exploitable.
243 [mancha <mancha1@zoho.com>]
245 *) A missing bounds check in the handling of the TLS heartbeat extension
246 can be used to reveal up to 64k of memory to a connected client or
249 Thanks for Neel Mehta of Google Security for discovering this bug and to
250 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
251 preparing the fix (CVE-2014-0160)
252 [Adam Langley, Bodo Moeller]
254 *) Fix for the attack described in the paper "Recovering OpenSSL
255 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
256 by Yuval Yarom and Naomi Benger. Details can be obtained from:
257 http://eprint.iacr.org/2014/140
259 Thanks to Yuval Yarom and Naomi Benger for discovering this
260 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
261 [Yuval Yarom and Naomi Benger]
263 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
264 this fixes a limitation in previous versions of OpenSSL.
267 *) Experimental encrypt-then-mac support.
269 Experimental support for encrypt then mac from
270 draft-gutmann-tls-encrypt-then-mac-02.txt
272 To enable it set the appropriate extension number (0x42 for the test
273 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
275 For non-compliant peers (i.e. just about everything) this should have no
278 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
282 *) Add EVP support for key wrapping algorithms, to avoid problems with
283 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
284 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
285 algorithms and include tests cases.
288 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
292 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
293 MGF1 digest and OAEP label.
296 *) Make openssl verify return errors.
297 [Chris Palmer <palmer@google.com> and Ben Laurie]
299 *) New function ASN1_TIME_diff to calculate the difference between two
300 ASN1_TIME structures or one structure and the current time.
303 *) Update fips_test_suite to support multiple command line options. New
304 test to induce all self test errors in sequence and check expected
308 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
309 sign or verify all in one operation.
312 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
313 test programs and fips_test_suite. Includes functionality to parse
314 the minimal script output of fipsalgest.pl directly.
317 *) Add authorisation parameter to FIPS_module_mode_set().
320 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
323 *) Use separate DRBG fields for internal and external flags. New function
324 FIPS_drbg_health_check() to perform on demand health checking. Add
325 generation tests to fips_test_suite with reduced health check interval to
326 demonstrate periodic health checking. Add "nodh" option to
327 fips_test_suite to skip very slow DH test.
330 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
334 *) More extensive health check for DRBG checking many more failure modes.
335 New function FIPS_selftest_drbg_all() to handle every possible DRBG
336 combination: call this in fips_test_suite.
339 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
340 and POST to handle Dual EC cases.
343 *) Add support for canonical generation of DSA parameter 'g'. See
346 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
347 POST to handle HMAC cases.
350 *) Add functions FIPS_module_version() and FIPS_module_version_text()
351 to return numerical and string versions of the FIPS module number.
354 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
355 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
356 outside the validated module in the FIPS capable OpenSSL.
359 *) Minor change to DRBG entropy callback semantics. In some cases
360 there is no multiple of the block length between min_len and
361 max_len. Allow the callback to return more than max_len bytes
362 of entropy but discard any extra: it is the callback's responsibility
363 to ensure that the extra data discarded does not impact the
364 requested amount of entropy.
367 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
368 information in FIPS186-3, SP800-57 and SP800-131A.
371 *) CCM support via EVP. Interface is very similar to GCM case except we
372 must supply all data in one chunk (i.e. no update, final) and the
373 message length must be supplied if AAD is used. Add algorithm test
377 *) Initial version of POST overhaul. Add POST callback to allow the status
378 of POST to be monitored and/or failures induced. Modify fips_test_suite
379 to use callback. Always run all selftests even if one fails.
382 *) XTS support including algorithm test driver in the fips_gcmtest program.
383 Note: this does increase the maximum key length from 32 to 64 bytes but
384 there should be no binary compatibility issues as existing applications
385 will never use XTS mode.
388 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
389 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
390 performs algorithm blocking for unapproved PRNG types. Also do not
391 set PRNG type in FIPS_mode_set(): leave this to the application.
392 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
393 the standard OpenSSL PRNG: set additional data to a date time vector.
396 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
397 This shouldn't present any incompatibility problems because applications
398 shouldn't be using these directly and any that are will need to rethink
399 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
402 *) Extensive self tests and health checking required by SP800-90 DRBG.
403 Remove strength parameter from FIPS_drbg_instantiate and always
404 instantiate at maximum supported strength.
407 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
410 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
413 *) New function DH_compute_key_padded() to compute a DH key and pad with
414 leading zeroes if needed: this complies with SP800-56A et al.
417 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
418 anything, incomplete, subject to change and largely untested at present.
421 *) Modify fipscanisteronly build option to only build the necessary object
422 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
425 *) Add experimental option FIPSSYMS to give all symbols in
426 fipscanister.o and FIPS or fips prefix. This will avoid
427 conflicts with future versions of OpenSSL. Add perl script
428 util/fipsas.pl to preprocess assembly language source files
429 and rename any affected symbols.
432 *) Add selftest checks and algorithm block of non-fips algorithms in
433 FIPS mode. Remove DES2 from selftests.
436 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
437 return internal method without any ENGINE dependencies. Add new
438 tiny fips sign and verify functions.
441 *) New build option no-ec2m to disable characteristic 2 code.
444 *) New build option "fipscanisteronly". This only builds fipscanister.o
445 and (currently) associated fips utilities. Uses the file Makefile.fips
446 instead of Makefile.org as the prototype.
449 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
450 Update fips_gcmtest to use IV generator.
453 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
454 setting output buffer to NULL. The *Final function must be
455 called although it will not retrieve any additional data. The tag
456 can be set or retrieved with a ctrl. The IV length is by default 12
457 bytes (96 bits) but can be set to an alternative value. If the IV
458 length exceeds the maximum IV length (currently 16 bytes) it cannot be
462 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
463 underlying do_cipher function handles all cipher semantics itself
464 including padding and finalisation. This is useful if (for example)
465 an ENGINE cipher handles block padding itself. The behaviour of
466 do_cipher is subtly changed if this flag is set: the return value
467 is the number of characters written to the output buffer (zero is
468 no longer an error code) or a negative error code. Also if the
469 input buffer is NULL and length 0 finalisation should be performed.
472 *) If a candidate issuer certificate is already part of the constructed
473 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
476 *) Improve forward-security support: add functions
478 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
479 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
481 for use by SSL/TLS servers; the callback function will be called whenever a
482 new session is created, and gets to decide whether the session may be
483 cached to make it resumable (return 0) or not (return 1). (As by the
484 SSL/TLS protocol specifications, the session_id sent by the server will be
485 empty to indicate that the session is not resumable; also, the server will
486 not generate RFC 4507 (RFC 5077) session tickets.)
488 A simple reasonable callback implementation is to return is_forward_secure.
489 This parameter will be set to 1 or 0 depending on the ciphersuite selected
490 by the SSL/TLS server library, indicating whether it can provide forward
492 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
494 *) New -verify_name option in command line utilities to set verification
498 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
499 Add CMAC pkey methods.
502 *) Experimental renegotiation in s_server -www mode. If the client
503 browses /reneg connection is renegotiated. If /renegcert it is
504 renegotiated requesting a certificate.
507 *) Add an "external" session cache for debugging purposes to s_server. This
508 should help trace issues which normally are only apparent in deployed
509 multi-process servers.
512 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
513 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
514 BIO_set_cipher() and some obscure PEM functions were changed so they
515 can now return an error. The RAND changes required a change to the
516 RAND_METHOD structure.
519 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
520 a gcc attribute to warn if the result of a function is ignored. This
521 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
522 whose return value is often ignored.
525 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
527 *) Alternate chains certificate forgery
529 During certificate verfification, OpenSSL will attempt to find an
530 alternative certificate chain if the first attempt to build such a chain
531 fails. An error in the implementation of this logic can mean that an
532 attacker could cause certain checks on untrusted certificates to be
533 bypassed, such as the CA flag, enabling them to use a valid leaf
534 certificate to act as a CA and "issue" an invalid certificate.
536 This issue was reported to OpenSSL by Adam Langley/David Benjamin
540 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
542 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
543 incompatibility in the handling of HMAC. The previous ABI has now been
547 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
549 *) Malformed ECParameters causes infinite loop
551 When processing an ECParameters structure OpenSSL enters an infinite loop
552 if the curve specified is over a specially malformed binary polynomial
555 This can be used to perform denial of service against any
556 system which processes public keys, certificate requests or
557 certificates. This includes TLS clients and TLS servers with
558 client authentication enabled.
560 This issue was reported to OpenSSL by Joseph Barr-Pixton.
564 *) Exploitable out-of-bounds read in X509_cmp_time
566 X509_cmp_time does not properly check the length of the ASN1_TIME
567 string and can read a few bytes out of bounds. In addition,
568 X509_cmp_time accepts an arbitrary number of fractional seconds in the
571 An attacker can use this to craft malformed certificates and CRLs of
572 various sizes and potentially cause a segmentation fault, resulting in
573 a DoS on applications that verify certificates or CRLs. TLS clients
574 that verify CRLs are affected. TLS clients and servers with client
575 authentication enabled may be affected if they use custom verification
578 This issue was reported to OpenSSL by Robert Swiecki (Google), and
579 independently by Hanno Böck.
583 *) PKCS7 crash with missing EnvelopedContent
585 The PKCS#7 parsing code does not handle missing inner EncryptedContent
586 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
587 with missing content and trigger a NULL pointer dereference on parsing.
589 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
590 structures from untrusted sources are affected. OpenSSL clients and
591 servers are not affected.
593 This issue was reported to OpenSSL by Michal Zalewski (Google).
597 *) CMS verify infinite loop with unknown hash function
599 When verifying a signedData message the CMS code can enter an infinite loop
600 if presented with an unknown hash function OID. This can be used to perform
601 denial of service against any system which verifies signedData messages using
603 This issue was reported to OpenSSL by Johannes Bauer.
607 *) Race condition handling NewSessionTicket
609 If a NewSessionTicket is received by a multi-threaded client when attempting to
610 reuse a previous ticket then a race condition can occur potentially leading to
611 a double free of the ticket data.
615 *) Only support 256-bit or stronger elliptic curves with the
616 'ecdh_auto' setting (server) or by default (client). Of supported
617 curves, prefer P-256 (both).
620 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
622 *) ClientHello sigalgs DoS fix
624 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
625 invalid signature algorithms extension a NULL pointer dereference will
626 occur. This can be exploited in a DoS attack against the server.
628 This issue was was reported to OpenSSL by David Ramos of Stanford
631 [Stephen Henson and Matt Caswell]
633 *) Multiblock corrupted pointer fix
635 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
636 feature only applies on 64 bit x86 architecture platforms that support AES
637 NI instructions. A defect in the implementation of "multiblock" can cause
638 OpenSSL's internal write buffer to become incorrectly set to NULL when
639 using non-blocking IO. Typically, when the user application is using a
640 socket BIO for writing, this will only result in a failed connection.
641 However if some other BIO is used then it is likely that a segmentation
642 fault will be triggered, thus enabling a potential DoS attack.
644 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
648 *) Segmentation fault in DTLSv1_listen fix
650 The DTLSv1_listen function is intended to be stateless and processes the
651 initial ClientHello from many peers. It is common for user code to loop
652 over the call to DTLSv1_listen until a valid ClientHello is received with
653 an associated cookie. A defect in the implementation of DTLSv1_listen means
654 that state is preserved in the SSL object from one invocation to the next
655 that can lead to a segmentation fault. Errors processing the initial
656 ClientHello can trigger this scenario. An example of such an error could be
657 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
660 This issue was reported to OpenSSL by Per Allansson.
664 *) Segmentation fault in ASN1_TYPE_cmp fix
666 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
667 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
668 certificate signature algorithm consistency this can be used to crash any
669 certificate verification operation and exploited in a DoS attack. Any
670 application which performs certificate verification is vulnerable including
671 OpenSSL clients and servers which enable client authentication.
675 *) Segmentation fault for invalid PSS parameters fix
677 The signature verification routines will crash with a NULL pointer
678 dereference if presented with an ASN.1 signature using the RSA PSS
679 algorithm and invalid parameters. Since these routines are used to verify
680 certificate signature algorithms this can be used to crash any
681 certificate verification operation and exploited in a DoS attack. Any
682 application which performs certificate verification is vulnerable including
683 OpenSSL clients and servers which enable client authentication.
685 This issue was was reported to OpenSSL by Brian Carpenter.
689 *) ASN.1 structure reuse memory corruption fix
691 Reusing a structure in ASN.1 parsing may allow an attacker to cause
692 memory corruption via an invalid write. Such reuse is and has been
693 strongly discouraged and is believed to be rare.
695 Applications that parse structures containing CHOICE or ANY DEFINED BY
696 components may be affected. Certificate parsing (d2i_X509 and related
697 functions) are however not affected. OpenSSL clients and servers are
702 *) PKCS7 NULL pointer dereferences fix
704 The PKCS#7 parsing code does not handle missing outer ContentInfo
705 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
706 missing content and trigger a NULL pointer dereference on parsing.
708 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
709 otherwise parse PKCS#7 structures from untrusted sources are
710 affected. OpenSSL clients and servers are not affected.
712 This issue was reported to OpenSSL by Michal Zalewski (Google).
716 *) DoS via reachable assert in SSLv2 servers fix
718 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
719 servers that both support SSLv2 and enable export cipher suites by sending
720 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
722 This issue was discovered by Sean Burford (Google) and Emilia Käsper
723 (OpenSSL development team).
727 *) Empty CKE with client auth and DHE fix
729 If client auth is used then a server can seg fault in the event of a DHE
730 ciphersuite being selected and a zero length ClientKeyExchange message
731 being sent by the client. This could be exploited in a DoS attack.
735 *) Handshake with unseeded PRNG fix
737 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
738 with an unseeded PRNG. The conditions are:
739 - The client is on a platform where the PRNG has not been seeded
740 automatically, and the user has not seeded manually
741 - A protocol specific client method version has been used (i.e. not
742 SSL_client_methodv23)
743 - A ciphersuite is used that does not require additional random data from
744 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
746 If the handshake succeeds then the client random that has been used will
747 have been generated from a PRNG with insufficient entropy and therefore the
748 output may be predictable.
750 For example using the following command with an unseeded openssl will
751 succeed on an unpatched platform:
753 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
757 *) Use After Free following d2i_ECPrivatekey error fix
759 A malformed EC private key file consumed via the d2i_ECPrivateKey function
760 could cause a use after free condition. This, in turn, could cause a double
761 free in several private key parsing functions (such as d2i_PrivateKey
762 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
763 for applications that receive EC private keys from untrusted
764 sources. This scenario is considered rare.
766 This issue was discovered by the BoringSSL project and fixed in their
771 *) X509_to_X509_REQ NULL pointer deref fix
773 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
774 the certificate key is invalid. This function is rarely used in practice.
776 This issue was discovered by Brian Carpenter.
780 *) Removed the export ciphers from the DEFAULT ciphers
783 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
785 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
786 ARMv5 through ARMv8, as opposite to "locking" it to single one.
787 So far those who have to target multiple plaforms would compromise
788 and argue that binary targeting say ARMv5 would still execute on
789 ARMv8. "Universal" build resolves this compromise by providing
790 near-optimal performance even on newer platforms.
793 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
794 (other platforms pending).
795 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
797 *) Add support for the SignedCertificateTimestampList certificate and
798 OCSP response extensions from RFC6962.
801 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
802 for corner cases. (Certain input points at infinity could lead to
803 bogus results, with non-infinity inputs mapped to infinity too.)
806 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
807 This covers AES, SHA256/512 and GHASH. "Initial" means that most
808 common cases are optimized and there still is room for further
809 improvements. Vector Permutation AES for Altivec is also added.
812 *) Add support for little-endian ppc64 Linux target.
813 [Marcelo Cerri (IBM)]
815 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
816 SHA1, SHA256 and GHASH. "Initial" means that most common cases
817 are optimized and there still is room for further improvements.
818 Both 32- and 64-bit modes are supported.
819 [Andy Polyakov, Ard Biesheuvel (Linaro)]
821 *) Improved ARMv7 NEON support.
824 *) Support for SPARC Architecture 2011 crypto extensions, first
825 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
826 SHA256/512, MD5, GHASH and modular exponentiation.
827 [Andy Polyakov, David Miller]
829 *) Accelerated modular exponentiation for Intel processors, a.k.a.
831 [Shay Gueron & Vlad Krasnov (Intel Corp)]
833 *) Support for new and upcoming Intel processors, including AVX2,
834 BMI and SHA ISA extensions. This includes additional "stitched"
835 implementations, AESNI-SHA256 and GCM, and multi-buffer support
838 This work was sponsored by Intel Corp.
841 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
842 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
843 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
846 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
847 this fixes a limiation in previous versions of OpenSSL.
850 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
851 MGF1 digest and OAEP label.
854 *) Add EVP support for key wrapping algorithms, to avoid problems with
855 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
856 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
857 algorithms and include tests cases.
860 *) Add functions to allocate and set the fields of an ECDSA_METHOD
862 [Douglas E. Engert, Steve Henson]
864 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
865 difference in days and seconds between two tm or ASN1_TIME structures.
868 *) Add -rev test option to s_server to just reverse order of characters
869 received by client and send back to server. Also prints an abbreviated
870 summary of the connection parameters.
873 *) New option -brief for s_client and s_server to print out a brief summary
874 of connection parameters.
877 *) Add callbacks for arbitrary TLS extensions.
878 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
880 *) New option -crl_download in several openssl utilities to download CRLs
881 from CRLDP extension in certificates.
884 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
887 *) New function X509_CRL_diff to generate a delta CRL from the difference
888 of two full CRLs. Add support to "crl" utility.
891 *) New functions to set lookup_crls function and to retrieve
892 X509_STORE from X509_STORE_CTX.
895 *) Print out deprecated issuer and subject unique ID fields in
899 *) Extend OCSP I/O functions so they can be used for simple general purpose
900 HTTP as well as OCSP. New wrapper function which can be used to download
901 CRLs using the OCSP API.
904 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
907 *) SSL_CONF* functions. These provide a common framework for application
908 configuration using configuration files or command lines.
911 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
912 message callback and prints the results. Needs compile time option
913 "enable-ssl-trace". New options to s_client and s_server to enable
917 *) New ctrl and macro to retrieve supported points extensions.
918 Print out extension in s_server and s_client.
921 *) New functions to retrieve certificate signature and signature
925 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
929 *) New Suite B modes for TLS code. These use and enforce the requirements
930 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
931 only use Suite B curves. The Suite B modes can be set by using the
932 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
935 *) New chain verification flags for Suite B levels of security. Check
936 algorithms are acceptable when flags are set in X509_verify_cert.
939 *) Make tls1_check_chain return a set of flags indicating checks passed
940 by a certificate chain. Add additional tests to handle client
941 certificates: checks for matching certificate type and issuer name
945 *) If an attempt is made to use a signature algorithm not in the peer
946 preference list abort the handshake. If client has no suitable
947 signature algorithms in response to a certificate request do not
951 *) If server EC tmp key is not in client preference list abort handshake.
954 *) Add support for certificate stores in CERT structure. This makes it
955 possible to have different stores per SSL structure or one store in
956 the parent SSL_CTX. Include distint stores for certificate chain
957 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
958 to build and store a certificate chain in CERT structure: returing
959 an error if the chain cannot be built: this will allow applications
960 to test if a chain is correctly configured.
962 Note: if the CERT based stores are not set then the parent SSL_CTX
963 store is used to retain compatibility with existing behaviour.
967 *) New function ssl_set_client_disabled to set a ciphersuite disabled
968 mask based on the current session, check mask when sending client
969 hello and checking the requested ciphersuite.
972 *) New ctrls to retrieve and set certificate types in a certificate
973 request message. Print out received values in s_client. If certificate
974 types is not set with custom values set sensible values based on
975 supported signature algorithms.
978 *) Support for distinct client and server supported signature algorithms.
981 *) Add certificate callback. If set this is called whenever a certificate
982 is required by client or server. An application can decide which
983 certificate chain to present based on arbitrary criteria: for example
984 supported signature algorithms. Add very simple example to s_server.
985 This fixes many of the problems and restrictions of the existing client
986 certificate callback: for example you can now clear an existing
987 certificate and specify the whole chain.
990 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
991 the certificate can be used for (if anything). Set valid_flags field
992 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
993 to have similar checks in it.
995 Add new "cert_flags" field to CERT structure and include a "strict mode".
996 This enforces some TLS certificate requirements (such as only permitting
997 certificate signature algorithms contained in the supported algorithms
998 extension) which some implementations ignore: this option should be used
999 with caution as it could cause interoperability issues.
1002 *) Update and tidy signature algorithm extension processing. Work out
1003 shared signature algorithms based on preferences and peer algorithms
1004 and print them out in s_client and s_server. Abort handshake if no
1005 shared signature algorithms.
1008 *) Add new functions to allow customised supported signature algorithms
1009 for SSL and SSL_CTX structures. Add options to s_client and s_server
1013 *) New function SSL_certs_clear() to delete all references to certificates
1014 from an SSL structure. Before this once a certificate had been added
1015 it couldn't be removed.
1018 *) Integrate hostname, email address and IP address checking with certificate
1019 verification. New verify options supporting checking in opensl utility.
1022 *) Fixes and wildcard matching support to hostname and email checking
1023 functions. Add manual page.
1024 [Florian Weimer (Red Hat Product Security Team)]
1026 *) New functions to check a hostname email or IP address against a
1027 certificate. Add options x509 utility to print results of checks against
1031 *) Fix OCSP checking.
1032 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1034 *) Initial experimental support for explicitly trusted non-root CAs.
1035 OpenSSL still tries to build a complete chain to a root but if an
1036 intermediate CA has a trust setting included that is used. The first
1037 setting is used: whether to trust (e.g., -addtrust option to the x509
1041 *) Add -trusted_first option which attempts to find certificates in the
1042 trusted store even if an untrusted chain is also supplied.
1045 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1046 platform support for Linux and Android.
1049 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1052 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1053 When in FIPS mode the approved implementations are used as normal,
1054 when not in FIPS mode the internal unapproved versions are used instead.
1055 This means that the FIPS capable OpenSSL isn't forced to use the
1056 (often lower performance) FIPS implementations outside FIPS mode.
1059 *) Transparently support X9.42 DH parameters when calling
1060 PEM_read_bio_DHparameters. This means existing applications can handle
1061 the new parameter format automatically.
1064 *) Initial experimental support for X9.42 DH parameter format: mainly
1065 to support use of 'q' parameter for RFC5114 parameters.
1068 *) Add DH parameters from RFC5114 including test data to dhtest.
1071 *) Support for automatic EC temporary key parameter selection. If enabled
1072 the most preferred EC parameters are automatically used instead of
1073 hardcoded fixed parameters. Now a server just has to call:
1074 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1075 support ECDH and use the most appropriate parameters.
1078 *) Enhance and tidy EC curve and point format TLS extension code. Use
1079 static structures instead of allocation if default values are used.
1080 New ctrls to set curves we wish to support and to retrieve shared curves.
1081 Print out shared curves in s_server. New options to s_server and s_client
1082 to set list of supported curves.
1085 *) New ctrls to retrieve supported signature algorithms and
1086 supported curve values as an array of NIDs. Extend openssl utility
1087 to print out received values.
1090 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1091 between NIDs and the more common NIST names such as "P-256". Enhance
1092 ecparam utility and ECC method to recognise the NIST names for curves.
1095 *) Enhance SSL/TLS certificate chain handling to support different
1096 chains for each certificate instead of one chain in the parent SSL_CTX.
1099 *) Support for fixed DH ciphersuite client authentication: where both
1100 server and client use DH certificates with common parameters.
1103 *) Support for fixed DH ciphersuites: those requiring DH server
1107 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1109 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1110 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1111 X509_CINF_get_signature were reverted post internal team review.
1113 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1115 *) Build fixes for the Windows and OpenVMS platforms
1116 [Matt Caswell and Richard Levitte]
1118 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1120 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1121 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1122 dereference. This could lead to a Denial Of Service attack. Thanks to
1123 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1127 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1128 dtls1_buffer_record function under certain conditions. In particular this
1129 could occur if an attacker sent repeated DTLS records with the same
1130 sequence number but for the next epoch. The memory leak could be exploited
1131 by an attacker in a Denial of Service attack through memory exhaustion.
1132 Thanks to Chris Mueller for reporting this issue.
1136 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1137 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1138 method would be set to NULL which could later result in a NULL pointer
1139 dereference. Thanks to Frank Schmirler for reporting this issue.
1143 *) Abort handshake if server key exchange message is omitted for ephemeral
1146 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1147 reporting this issue.
1151 *) Remove non-export ephemeral RSA code on client and server. This code
1152 violated the TLS standard by allowing the use of temporary RSA keys in
1153 non-export ciphersuites and could be used by a server to effectively
1154 downgrade the RSA key length used to a value smaller than the server
1155 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1156 INRIA or reporting this issue.
1160 *) Fixed issue where DH client certificates are accepted without verification.
1161 An OpenSSL server will accept a DH certificate for client authentication
1162 without the certificate verify message. This effectively allows a client to
1163 authenticate without the use of a private key. This only affects servers
1164 which trust a client certificate authority which issues certificates
1165 containing DH keys: these are extremely rare and hardly ever encountered.
1166 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1171 *) Ensure that the session ID context of an SSL is updated when its
1172 SSL_CTX is updated via SSL_set_SSL_CTX.
1174 The session ID context is typically set from the parent SSL_CTX,
1175 and can vary with the CTX.
1178 *) Fix various certificate fingerprint issues.
1180 By using non-DER or invalid encodings outside the signed portion of a
1181 certificate the fingerprint can be changed without breaking the signature.
1182 Although no details of the signed portion of the certificate can be changed
1183 this can cause problems with some applications: e.g. those using the
1184 certificate fingerprint for blacklists.
1186 1. Reject signatures with non zero unused bits.
1188 If the BIT STRING containing the signature has non zero unused bits reject
1189 the signature. All current signature algorithms require zero unused bits.
1191 2. Check certificate algorithm consistency.
1193 Check the AlgorithmIdentifier inside TBS matches the one in the
1194 certificate signature. NB: this will result in signature failure
1195 errors for some broken certificates.
1197 Thanks to Konrad Kraszewski from Google for reporting this issue.
1199 3. Check DSA/ECDSA signatures use DER.
1201 Reencode DSA/ECDSA signatures and compare with the original received
1202 signature. Return an error if there is a mismatch.
1204 This will reject various cases including garbage after signature
1205 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1206 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1207 (negative or with leading zeroes).
1209 Further analysis was conducted and fixes were developed by Stephen Henson
1210 of the OpenSSL core team.
1215 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1216 results on some platforms, including x86_64. This bug occurs at random
1217 with a very low probability, and is not known to be exploitable in any
1218 way, though its exact impact is difficult to determine. Thanks to Pieter
1219 Wuille (Blockstream) who reported this issue and also suggested an initial
1220 fix. Further analysis was conducted by the OpenSSL development team and
1221 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1222 the OpenSSL core team.
1226 *) Do not resume sessions on the server if the negotiated protocol
1227 version does not match the session's version. Resuming with a different
1228 version, while not strictly forbidden by the RFC, is of questionable
1229 sanity and breaks all known clients.
1230 [David Benjamin, Emilia Käsper]
1232 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1233 early CCS messages during renegotiation. (Note that because
1234 renegotiation is encrypted, this early CCS was not exploitable.)
1237 *) Tighten client-side session ticket handling during renegotiation:
1238 ensure that the client only accepts a session ticket if the server sends
1239 the extension anew in the ServerHello. Previously, a TLS client would
1240 reuse the old extension state and thus accept a session ticket if one was
1241 announced in the initial ServerHello.
1243 Similarly, ensure that the client requires a session ticket if one
1244 was advertised in the ServerHello. Previously, a TLS client would
1245 ignore a missing NewSessionTicket message.
1248 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1250 *) SRTP Memory Leak.
1252 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1253 sends a carefully crafted handshake message, to cause OpenSSL to fail
1254 to free up to 64k of memory causing a memory leak. This could be
1255 exploited in a Denial Of Service attack. This issue affects OpenSSL
1256 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1257 whether SRTP is used or configured. Implementations of OpenSSL that
1258 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1260 The fix was developed by the OpenSSL team.
1264 *) Session Ticket Memory Leak.
1266 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1267 integrity of that ticket is first verified. In the event of a session
1268 ticket integrity check failing, OpenSSL will fail to free memory
1269 causing a memory leak. By sending a large number of invalid session
1270 tickets an attacker could exploit this issue in a Denial Of Service
1275 *) Build option no-ssl3 is incomplete.
1277 When OpenSSL is configured with "no-ssl3" as a build option, servers
1278 could accept and complete a SSL 3.0 handshake, and clients could be
1279 configured to send them.
1281 [Akamai and the OpenSSL team]
1283 *) Add support for TLS_FALLBACK_SCSV.
1284 Client applications doing fallback retries should call
1285 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1287 [Adam Langley, Bodo Moeller]
1289 *) Add additional DigestInfo checks.
1291 Reencode DigestInto in DER and check against the original when
1292 verifying RSA signature: this will reject any improperly encoded
1293 DigestInfo structures.
1295 Note: this is a precautionary measure and no attacks are currently known.
1299 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1301 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1302 SRP code can be overrun an internal buffer. Add sanity check that
1303 g, A, B < N to SRP code.
1305 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1306 Group for discovering this issue.
1310 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1311 TLS 1.0 instead of higher protocol versions when the ClientHello message
1312 is badly fragmented. This allows a man-in-the-middle attacker to force a
1313 downgrade to TLS 1.0 even if both the server and the client support a
1314 higher protocol version, by modifying the client's TLS records.
1316 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1317 researching this issue.
1321 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1322 to a denial of service attack. A malicious server can crash the client
1323 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1324 ciphersuite and sending carefully crafted handshake messages.
1326 Thanks to Felix Gröbert (Google) for discovering and researching this
1331 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1332 to leak memory. This can be exploited through a Denial of Service attack.
1333 Thanks to Adam Langley for discovering and researching this issue.
1337 *) An attacker can force openssl to consume large amounts of memory whilst
1338 processing DTLS handshake messages. This can be exploited through a
1339 Denial of Service attack.
1340 Thanks to Adam Langley for discovering and researching this issue.
1344 *) An attacker can force an error condition which causes openssl to crash
1345 whilst processing DTLS packets due to memory being freed twice. This
1346 can be exploited through a Denial of Service attack.
1347 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1352 *) If a multithreaded client connects to a malicious server using a resumed
1353 session and the server sends an ec point format extension it could write
1354 up to 255 bytes to freed memory.
1356 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1361 *) A malicious server can crash an OpenSSL client with a null pointer
1362 dereference (read) by specifying an SRP ciphersuite even though it was not
1363 properly negotiated with the client. This can be exploited through a
1364 Denial of Service attack.
1366 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1367 discovering and researching this issue.
1371 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1372 X509_name_oneline, X509_name_print_ex et al. to leak some information
1373 from the stack. Applications may be affected if they echo pretty printing
1374 output to the attacker.
1376 Thanks to Ivan Fratric (Google) for discovering this issue.
1378 [Emilia Käsper, and Steve Henson]
1380 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1381 for corner cases. (Certain input points at infinity could lead to
1382 bogus results, with non-infinity inputs mapped to infinity too.)
1385 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1387 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1388 handshake can force the use of weak keying material in OpenSSL
1389 SSL/TLS clients and servers.
1391 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1392 researching this issue. (CVE-2014-0224)
1393 [KIKUCHI Masashi, Steve Henson]
1395 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1396 OpenSSL DTLS client the code can be made to recurse eventually crashing
1399 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1401 [Imre Rad, Steve Henson]
1403 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1404 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1405 client or server. This is potentially exploitable to run arbitrary
1406 code on a vulnerable client or server.
1408 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1409 [Jüri Aedla, Steve Henson]
1411 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1412 are subject to a denial of service attack.
1414 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1415 this issue. (CVE-2014-3470)
1416 [Felix Gröbert, Ivan Fratric, Steve Henson]
1418 *) Harmonize version and its documentation. -f flag is used to display
1420 [mancha <mancha1@zoho.com>]
1422 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1423 in i2d_ECPrivateKey.
1424 [mancha <mancha1@zoho.com>]
1426 *) Fix some double frees. These are not thought to be exploitable.
1427 [mancha <mancha1@zoho.com>]
1429 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1431 *) A missing bounds check in the handling of the TLS heartbeat extension
1432 can be used to reveal up to 64k of memory to a connected client or
1435 Thanks for Neel Mehta of Google Security for discovering this bug and to
1436 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1437 preparing the fix (CVE-2014-0160)
1438 [Adam Langley, Bodo Moeller]
1440 *) Fix for the attack described in the paper "Recovering OpenSSL
1441 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1442 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1443 http://eprint.iacr.org/2014/140
1445 Thanks to Yuval Yarom and Naomi Benger for discovering this
1446 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1447 [Yuval Yarom and Naomi Benger]
1449 *) TLS pad extension: draft-agl-tls-padding-03
1451 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1452 TLS client Hello record length value would otherwise be > 255 and
1453 less that 512 pad with a dummy extension containing zeroes so it
1454 is at least 512 bytes long.
1456 [Adam Langley, Steve Henson]
1458 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1460 *) Fix for TLS record tampering bug. A carefully crafted invalid
1461 handshake could crash OpenSSL with a NULL pointer exception.
1462 Thanks to Anton Johansson for reporting this issues.
1465 *) Keep original DTLS digest and encryption contexts in retransmission
1466 structures so we can use the previous session parameters if they need
1467 to be resent. (CVE-2013-6450)
1470 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1471 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1472 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1473 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1474 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1475 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1476 [Rob Stradling, Adam Langley]
1478 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1480 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1481 supporting platforms or when small records were transferred.
1482 [Andy Polyakov, Steve Henson]
1484 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1486 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1488 This addresses the flaw in CBC record processing discovered by
1489 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1490 at: http://www.isg.rhul.ac.uk/tls/
1492 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1493 Security Group at Royal Holloway, University of London
1494 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1495 Emilia Käsper for the initial patch.
1497 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1499 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1500 ciphersuites which can be exploited in a denial of service attack.
1501 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1502 and detecting this bug and to Wolfgang Ettlinger
1503 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1507 *) Return an error when checking OCSP signatures when key is NULL.
1508 This fixes a DoS attack. (CVE-2013-0166)
1511 *) Make openssl verify return errors.
1512 [Chris Palmer <palmer@google.com> and Ben Laurie]
1514 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1515 the right response is stapled. Also change SSL_get_certificate()
1516 so it returns the certificate actually sent.
1517 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1518 [Rob Stradling <rob.stradling@comodo.com>]
1520 *) Fix possible deadlock when decoding public keys.
1523 *) Don't use TLS 1.0 record version number in initial client hello
1527 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1529 *) Sanity check record length before skipping explicit IV in TLS
1530 1.2, 1.1 and DTLS to fix DoS attack.
1532 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1533 fuzzing as a service testing platform.
1537 *) Initialise tkeylen properly when encrypting CMS messages.
1538 Thanks to Solar Designer of Openwall for reporting this issue.
1541 *) In FIPS mode don't try to use composite ciphers as they are not
1545 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1547 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1548 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1549 mean any application compiled against OpenSSL 1.0.0 headers setting
1550 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1551 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1552 0x10000000L Any application which was previously compiled against
1553 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1554 will need to be recompiled as a result. Letting be results in
1555 inability to disable specifically TLS 1.1 and in client context,
1556 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1559 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1560 disable just protocol X, but all protocols above X *if* there are
1561 protocols *below* X still enabled. In more practical terms it means
1562 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1563 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1564 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1568 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1570 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1571 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1572 in CRYPTO_realloc_clean.
1574 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1575 issue and to Adam Langley <agl@chromium.org> for fixing it.
1577 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1579 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1582 *) Workarounds for some broken servers that "hang" if a client hello
1583 record length exceeds 255 bytes.
1585 1. Do not use record version number > TLS 1.0 in initial client
1586 hello: some (but not all) hanging servers will now work.
1587 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1588 the number of ciphers sent in the client hello. This should be
1589 set to an even number, such as 50, for example by passing:
1590 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1591 Most broken servers should now work.
1592 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1593 TLS 1.2 client support entirely.
1596 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1599 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1601 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1602 STRING form instead of a DigestInfo.
1605 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1606 and the RSA_sign/RSA_verify functions. This was made more apparent when
1607 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1608 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1609 the correct format in RSA_verify so both forms transparently work.
1612 *) Some servers which support TLS 1.0 can choke if we initially indicate
1613 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1614 encrypted premaster secret. As a workaround use the maximum permitted
1615 client version in client hello, this should keep such servers happy
1616 and still work with previous versions of OpenSSL.
1619 *) Add support for TLS/DTLS heartbeats.
1620 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1622 *) Add support for SCTP.
1623 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1625 *) Improved PRNG seeding for VOS.
1626 [Paul Green <Paul.Green@stratus.com>]
1628 *) Extensive assembler packs updates, most notably:
1630 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1631 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1632 - x86_64: bit-sliced AES implementation;
1633 - ARM: NEON support, contemporary platforms optimizations;
1634 - s390x: z196 support;
1635 - *: GHASH and GF(2^m) multiplication implementations;
1639 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1640 (removal of unnecessary code)
1641 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1643 *) Add TLS key material exporter from RFC 5705.
1646 *) Add DTLS-SRTP negotiation from RFC 5764.
1649 *) Add Next Protocol Negotiation,
1650 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1651 disabled with a no-npn flag to config or Configure. Code donated
1653 [Adam Langley <agl@google.com> and Ben Laurie]
1655 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1656 NIST-P256, NIST-P521, with constant-time single point multiplication on
1657 typical inputs. Compiler support for the nonstandard type __uint128_t is
1658 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1659 Code made available under Apache License version 2.0.
1661 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1662 line to include this in your build of OpenSSL, and run "make depend" (or
1663 "make update"). This enables the following EC_METHODs:
1665 EC_GFp_nistp224_method()
1666 EC_GFp_nistp256_method()
1667 EC_GFp_nistp521_method()
1669 EC_GROUP_new_by_curve_name() will automatically use these (while
1670 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1672 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1674 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1675 all platforms. Move ssize_t definition from e_os.h to the public
1676 header file e_os2.h as it now appears in public header file cms.h
1679 *) New -sigopt option to the ca, req and x509 utilities. Additional
1680 signature parameters can be passed using this option and in
1684 *) Add RSA PSS signing function. This will generate and set the
1685 appropriate AlgorithmIdentifiers for PSS based on those in the
1686 corresponding EVP_MD_CTX structure. No application support yet.
1689 *) Support for companion algorithm specific ASN1 signing routines.
1690 New function ASN1_item_sign_ctx() signs a pre-initialised
1691 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1692 the appropriate parameters.
1695 *) Add new algorithm specific ASN1 verification initialisation function
1696 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1697 handling will be the same no matter what EVP_PKEY_METHOD is used.
1698 Add a PSS handler to support verification of PSS signatures: checked
1699 against a number of sample certificates.
1702 *) Add signature printing for PSS. Add PSS OIDs.
1703 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1705 *) Add algorithm specific signature printing. An individual ASN1 method
1706 can now print out signatures instead of the standard hex dump.
1708 More complex signatures (e.g. PSS) can print out more meaningful
1709 information. Include DSA version that prints out the signature
1713 *) Password based recipient info support for CMS library: implementing
1717 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1718 neatly separates the code into cipher and PBE sections and is required
1719 for some algorithms that split PBES2 into separate pieces (such as
1720 password based CMS).
1723 *) Session-handling fixes:
1724 - Fix handling of connections that are resuming with a session ID,
1725 but also support Session Tickets.
1726 - Fix a bug that suppressed issuing of a new ticket if the client
1727 presented a ticket with an expired session.
1728 - Try to set the ticket lifetime hint to something reasonable.
1729 - Make tickets shorter by excluding irrelevant information.
1730 - On the client side, don't ignore renewed tickets.
1731 [Adam Langley, Bodo Moeller (Google)]
1733 *) Fix PSK session representation.
1736 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1738 This work was sponsored by Intel.
1741 *) Add GCM support to TLS library. Some custom code is needed to split
1742 the IV between the fixed (from PRF) and explicit (from TLS record)
1743 portions. This adds all GCM ciphersuites supported by RFC5288 and
1744 RFC5289. Generalise some AES* cipherstrings to include GCM and
1745 add a special AESGCM string for GCM only.
1748 *) Expand range of ctrls for AES GCM. Permit setting invocation
1749 field on decrypt and retrieval of invocation field only on encrypt.
1752 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1753 As required by RFC5289 these ciphersuites cannot be used if for
1754 versions of TLS earlier than 1.2.
1757 *) For FIPS capable OpenSSL interpret a NULL default public key method
1758 as unset and return the appropriate default but do *not* set the default.
1759 This means we can return the appropriate method in applications that
1760 switch between FIPS and non-FIPS modes.
1763 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1764 ENGINE is used then we cannot handle that in the FIPS module so we
1765 keep original code iff non-FIPS operations are allowed.
1768 *) Add -attime option to openssl utilities.
1769 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1771 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1774 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1775 FIPS EC methods unconditionally for now.
1778 *) New build option no-ec2m to disable characteristic 2 code.
1781 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1782 all cases can be covered as some introduce binary incompatibilities.
1785 *) Redirect RSA operations to FIPS module including keygen,
1786 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1789 *) Add similar low level API blocking to ciphers.
1792 *) Low level digest APIs are not approved in FIPS mode: any attempt
1793 to use these will cause a fatal error. Applications that *really* want
1794 to use them can use the private_* version instead.
1797 *) Redirect cipher operations to FIPS module for FIPS builds.
1800 *) Redirect digest operations to FIPS module for FIPS builds.
1803 *) Update build system to add "fips" flag which will link in fipscanister.o
1804 for static and shared library builds embedding a signature if needed.
1807 *) Output TLS supported curves in preference order instead of numerical
1808 order. This is currently hardcoded for the highest order curves first.
1809 This should be configurable so applications can judge speed vs strength.
1812 *) Add TLS v1.2 server support for client authentication.
1815 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1819 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1820 FIPS modules versions.
1823 *) Add TLS v1.2 client side support for client authentication. Keep cache
1824 of handshake records longer as we don't know the hash algorithm to use
1825 until after the certificate request message is received.
1828 *) Initial TLS v1.2 client support. Add a default signature algorithms
1829 extension including all the algorithms we support. Parse new signature
1830 format in client key exchange. Relax some ECC signing restrictions for
1831 TLS v1.2 as indicated in RFC5246.
1834 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1835 to new signature format when needed using client digest preference.
1836 All server ciphersuites should now work correctly in TLS v1.2. No client
1837 support yet and no support for client certificates.
1840 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1841 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1842 ciphersuites. At present only RSA key exchange ciphersuites work with
1843 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1844 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1845 and version checking.
1848 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1849 with this defined it will not be affected by any changes to ssl internal
1850 structures. Add several utility functions to allow openssl application
1851 to work with OPENSSL_NO_SSL_INTERN defined.
1855 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1857 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1860 *) Permit abbreviated handshakes when renegotiating using the function
1861 SSL_renegotiate_abbreviated().
1862 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1864 *) Add call to ENGINE_register_all_complete() to
1865 ENGINE_load_builtin_engines(), so some implementations get used
1866 automatically instead of needing explicit application support.
1869 *) Add support for TLS key exporter as described in RFC5705.
1870 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1872 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1873 a few changes are required:
1875 Add SSL_OP_NO_TLSv1_1 flag.
1876 Add TLSv1_1 methods.
1877 Update version checking logic to handle version 1.1.
1878 Add explicit IV handling (ported from DTLS code).
1879 Add command line options to s_client/s_server.
1882 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1884 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1885 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1886 content decryption and always return the same error. Note: this attack
1887 needs on average 2^20 messages so it only affects automated senders. The
1888 old behaviour can be reenabled in the CMS code by setting the
1889 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1890 an MMA defence is not necessary.
1891 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1892 this issue. (CVE-2012-0884)
1895 *) Fix CVE-2011-4619: make sure we really are receiving a
1896 client hello before rejecting multiple SGC restarts. Thanks to
1897 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1900 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1902 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1903 Thanks to Antonio Martin, Enterprise Secure Access Research and
1904 Development, Cisco Systems, Inc. for discovering this bug and
1905 preparing a fix. (CVE-2012-0050)
1908 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1910 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1911 of the Vaudenay padding oracle attack on CBC mode encryption
1912 which enables an efficient plaintext recovery attack against
1913 the OpenSSL implementation of DTLS. Their attack exploits timing
1914 differences arising during decryption processing. A research
1915 paper describing this attack can be found at:
1916 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1917 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1918 Security Group at Royal Holloway, University of London
1919 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1920 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1921 for preparing the fix. (CVE-2011-4108)
1922 [Robin Seggelmann, Michael Tuexen]
1924 *) Clear bytes used for block padding of SSL 3.0 records.
1926 [Adam Langley (Google)]
1928 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1929 Kadianakis <desnacked@gmail.com> for discovering this issue and
1930 Adam Langley for preparing the fix. (CVE-2011-4619)
1931 [Adam Langley (Google)]
1933 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1934 [Andrey Kulikov <amdeich@gmail.com>]
1936 *) Prevent malformed RFC3779 data triggering an assertion failure.
1937 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1938 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1939 [Rob Austein <sra@hactrn.net>]
1941 *) Improved PRNG seeding for VOS.
1942 [Paul Green <Paul.Green@stratus.com>]
1944 *) Fix ssl_ciph.c set-up race.
1945 [Adam Langley (Google)]
1947 *) Fix spurious failures in ecdsatest.c.
1948 [Emilia Käsper (Google)]
1950 *) Fix the BIO_f_buffer() implementation (which was mixing different
1951 interpretations of the '..._len' fields).
1952 [Adam Langley (Google)]
1954 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1955 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1956 threads won't reuse the same blinding coefficients.
1958 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1959 lock to call BN_BLINDING_invert_ex, and avoids one use of
1960 BN_BLINDING_update for each BN_BLINDING structure (previously,
1961 the last update always remained unused).
1962 [Emilia Käsper (Google)]
1964 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1965 [Bob Buckholz (Google)]
1967 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1969 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1970 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1971 [Kaspar Brand <ossl@velox.ch>]
1973 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1974 for multi-threaded use of ECDH. (CVE-2011-3210)
1975 [Adam Langley (Google)]
1977 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1980 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1981 signature public key algorithm by using OID xref utilities instead.
1982 Before this you could only use some ECC ciphersuites with SHA1 only.
1985 *) Add protection against ECDSA timing attacks as mentioned in the paper
1986 by Billy Bob Brumley and Nicola Tuveri, see:
1988 http://eprint.iacr.org/2011/232.pdf
1990 [Billy Bob Brumley and Nicola Tuveri]
1992 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1994 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1995 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1997 *) Fix bug in string printing code: if *any* escaping is enabled we must
1998 escape the escape character (backslash) or the resulting string is
2002 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2004 *) Disable code workaround for ancient and obsolete Netscape browsers
2005 and servers: an attacker can use it in a ciphersuite downgrade attack.
2006 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2009 *) Fixed J-PAKE implementation error, originally discovered by
2010 Sebastien Martini, further info and confirmation from Stefan
2011 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2014 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2016 *) Fix extension code to avoid race conditions which can result in a buffer
2017 overrun vulnerability: resumed sessions must not be modified as they can
2018 be shared by multiple threads. CVE-2010-3864
2021 *) Fix WIN32 build system to correctly link an ENGINE directory into
2025 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2027 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2029 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2031 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2033 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2034 context. The operation can be customised via the ctrl mechanism in
2035 case ENGINEs want to include additional functionality.
2038 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2041 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2042 output hashes compatible with older versions of OpenSSL.
2043 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2045 *) Fix compression algorithm handling: if resuming a session use the
2046 compression algorithm of the resumed session instead of determining
2047 it from client hello again. Don't allow server to change algorithm.
2050 *) Add load_crls() function to apps tidying load_certs() too. Add option
2051 to verify utility to allow additional CRLs to be included.
2054 *) Update OCSP request code to permit adding custom headers to the request:
2055 some responders need this.
2058 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2060 [Julia Lawall <julia@diku.dk>]
2062 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2063 needlessly dereferenced structures, used obsolete functions and
2064 didn't handle all updated verify codes correctly.
2067 *) Disable MD2 in the default configuration.
2070 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2071 indicate the initial BIO being pushed or popped. This makes it possible
2072 to determine whether the BIO is the one explicitly called or as a result
2073 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2074 it handles reference counts correctly and doesn't zero out the I/O bio
2075 when it is not being explicitly popped. WARNING: applications which
2076 included workarounds for the old buggy behaviour will need to be modified
2077 or they could free up already freed BIOs.
2080 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2081 renaming to all platforms (within the 0.9.8 branch, this was
2082 done conditionally on Netware platforms to avoid a name clash).
2083 [Guenter <lists@gknw.net>]
2085 *) Add ECDHE and PSK support to DTLS.
2086 [Michael Tuexen <tuexen@fh-muenster.de>]
2088 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2092 *) Add "missing" function EVP_MD_flags() (without this the only way to
2093 retrieve a digest flags is by accessing the structure directly. Update
2094 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2095 or cipher is registered as in the "from" argument. Print out all
2096 registered digests in the dgst usage message instead of manually
2097 attempting to work them out.
2100 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2101 this allows the use of compression and extensions. Change default cipher
2102 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2103 by default unless an application cipher string requests it.
2106 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2107 key ids to find matching certificates and keys but some PKCS#12 files
2108 don't follow the (somewhat unwritten) rules and this strategy fails.
2109 Now just gather all certificates together and the first private key
2110 then look for the first certificate that matches the key.
2113 *) Support use of registered digest and cipher names for dgst and cipher
2114 commands instead of having to add each one as a special case. So now
2121 openssl dgst -sha256 foo
2123 and this works for ENGINE based algorithms too.
2127 *) Update Gost ENGINE to support parameter files.
2128 [Victor B. Wagner <vitus@cryptocom.ru>]
2130 *) Support GeneralizedTime in ca utility.
2131 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2133 *) Enhance the hash format used for certificate directory links. The new
2134 form uses the canonical encoding (meaning equivalent names will work
2135 even if they aren't identical) and uses SHA1 instead of MD5. This form
2136 is incompatible with the older format and as a result c_rehash should
2137 be used to rebuild symbolic links.
2140 *) Make PKCS#8 the default write format for private keys, replacing the
2141 traditional format. This form is standardised, more secure and doesn't
2142 include an implicit MD5 dependency.
2145 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2146 committed to OpenSSL should pass this lot as a minimum.
2149 *) Add session ticket override functionality for use by EAP-FAST.
2150 [Jouni Malinen <j@w1.fi>]
2152 *) Modify HMAC functions to return a value. Since these can be implemented
2153 in an ENGINE errors can occur.
2156 *) Type-checked OBJ_bsearch_ex.
2159 *) Type-checked OBJ_bsearch. Also some constification necessitated
2160 by type-checking. Still to come: TXT_DB, bsearch(?),
2161 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2165 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2166 seconds to a tm structure directly, instead of going through OS
2167 specific date routines. This avoids any issues with OS routines such
2168 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2169 and X509_time_adj_ex() to cover the extended range. The existing
2170 X509_time_adj() is still usable and will no longer have any date issues.
2173 *) Delta CRL support. New use deltas option which will attempt to locate
2174 and search any appropriate delta CRLs available.
2176 This work was sponsored by Google.
2179 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2180 code and add additional score elements. Validate alternate CRL paths
2181 as part of the CRL checking and indicate a new error "CRL path validation
2182 error" in this case. Applications wanting additional details can use
2183 the verify callback and check the new "parent" field. If this is not
2184 NULL CRL path validation is taking place. Existing applications wont
2185 see this because it requires extended CRL support which is off by
2188 This work was sponsored by Google.
2191 *) Support for freshest CRL extension.
2193 This work was sponsored by Google.
2196 *) Initial indirect CRL support. Currently only supported in the CRLs
2197 passed directly and not via lookup. Process certificate issuer
2198 CRL entry extension and lookup CRL entries by bother issuer name
2199 and serial number. Check and process CRL issuer entry in IDP extension.
2201 This work was sponsored by Google.
2204 *) Add support for distinct certificate and CRL paths. The CRL issuer
2205 certificate is validated separately in this case. Only enabled if
2206 an extended CRL support flag is set: this flag will enable additional
2207 CRL functionality in future.
2209 This work was sponsored by Google.
2212 *) Add support for policy mappings extension.
2214 This work was sponsored by Google.
2217 *) Fixes to pathlength constraint, self issued certificate handling,
2218 policy processing to align with RFC3280 and PKITS tests.
2220 This work was sponsored by Google.
2223 *) Support for name constraints certificate extension. DN, email, DNS
2224 and URI types are currently supported.
2226 This work was sponsored by Google.
2229 *) To cater for systems that provide a pointer-based thread ID rather
2230 than numeric, deprecate the current numeric thread ID mechanism and
2231 replace it with a structure and associated callback type. This
2232 mechanism allows a numeric "hash" to be extracted from a thread ID in
2233 either case, and on platforms where pointers are larger than 'long',
2234 mixing is done to help ensure the numeric 'hash' is usable even if it
2235 can't be guaranteed unique. The default mechanism is to use "&errno"
2236 as a pointer-based thread ID to distinguish between threads.
2238 Applications that want to provide their own thread IDs should now use
2239 CRYPTO_THREADID_set_callback() to register a callback that will call
2240 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2242 Note that ERR_remove_state() is now deprecated, because it is tied
2243 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2244 to free the current thread's error state should be replaced by
2245 ERR_remove_thread_state(NULL).
2247 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2248 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2249 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2250 application was previously providing a numeric thread callback that
2251 was inappropriate for distinguishing threads, then uniqueness might
2252 have been obtained with &errno that happened immediately in the
2253 intermediate development versions of OpenSSL; this is no longer the
2254 case, the numeric thread callback will now override the automatic use
2256 [Geoff Thorpe, with help from Bodo Moeller]
2258 *) Initial support for different CRL issuing certificates. This covers a
2259 simple case where the self issued certificates in the chain exist and
2260 the real CRL issuer is higher in the existing chain.
2262 This work was sponsored by Google.
2265 *) Removed effectively defunct crypto/store from the build.
2268 *) Revamp of STACK to provide stronger type-checking. Still to come:
2269 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2270 ASN1_STRING, CONF_VALUE.
2273 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2274 RAM on SSL connections. This option can save about 34k per idle SSL.
2277 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2278 STACK, TXT_DB, bsearch, qsort.
2281 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2282 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2283 support for data, signedData, compressedData, digestedData and
2284 encryptedData, envelopedData types included. Scripts to check against
2285 RFC4134 examples draft and interop and consistency checks of many
2286 content types and variants.
2289 *) Add options to enc utility to support use of zlib compression BIO.
2292 *) Extend mk1mf to support importing of options and assembly language
2293 files from Configure script, currently only included in VC-WIN32.
2294 The assembly language rules can now optionally generate the source
2295 files from the associated perl scripts.
2298 *) Implement remaining functionality needed to support GOST ciphersuites.
2299 Interop testing has been performed using CryptoPro implementations.
2300 [Victor B. Wagner <vitus@cryptocom.ru>]
2302 *) s390x assembler pack.
2305 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2309 *) Implement Opaque PRF Input TLS extension as specified in
2310 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2311 official specification yet and no extension type assignment by
2312 IANA exists, this extension (for now) will have to be explicitly
2313 enabled when building OpenSSL by providing the extension number
2314 to use. For example, specify an option
2316 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2318 to the "config" or "Configure" script to enable the extension,
2319 assuming extension number 0x9527 (which is a completely arbitrary
2320 and unofficial assignment based on the MD5 hash of the Internet
2321 Draft). Note that by doing so, you potentially lose
2322 interoperability with other TLS implementations since these might
2323 be using the same extension number for other purposes.
2325 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2326 opaque PRF input value to use in the handshake. This will create
2327 an interal copy of the length-'len' string at 'src', and will
2328 return non-zero for success.
2330 To get more control and flexibility, provide a callback function
2333 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2334 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2338 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2341 Callback function 'cb' will be called in handshakes, and is
2342 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2343 Argument 'arg' is for application purposes (the value as given to
2344 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2345 be provided to the callback function). The callback function
2346 has to return non-zero to report success: usually 1 to use opaque
2347 PRF input just if possible, or 2 to enforce use of the opaque PRF
2348 input. In the latter case, the library will abort the handshake
2349 if opaque PRF input is not successfully negotiated.
2351 Arguments 'peerinput' and 'len' given to the callback function
2352 will always be NULL and 0 in the case of a client. A server will
2353 see the client's opaque PRF input through these variables if
2354 available (NULL and 0 otherwise). Note that if the server
2355 provides an opaque PRF input, the length must be the same as the
2356 length of the client's opaque PRF input.
2358 Note that the callback function will only be called when creating
2359 a new session (session resumption can resume whatever was
2360 previously negotiated), and will not be called in SSL 2.0
2361 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2362 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2363 for applications that need to enforce opaque PRF input.
2367 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2370 [Victor B. Wagner <vitus@cryptocom.ru>]
2372 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2373 RFC4507bis. The encrypted ticket format is an encrypted encoded
2374 SSL_SESSION structure, that way new session features are automatically
2377 If a client application caches session in an SSL_SESSION structure
2378 support is transparent because tickets are now stored in the encoded
2381 The SSL_CTX structure automatically generates keys for ticket
2382 protection in servers so again support should be possible
2383 with no application modification.
2385 If a client or server wishes to disable RFC4507 support then the option
2386 SSL_OP_NO_TICKET can be set.
2388 Add a TLS extension debugging callback to allow the contents of any client
2389 or server extensions to be examined.
2391 This work was sponsored by Google.
2394 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2395 OpenSSL should now compile cleanly on gcc 4.2
2396 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2398 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2399 support including streaming MAC support: this is required for GOST
2400 ciphersuite support.
2401 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2403 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2404 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2405 to output in BER and PEM format.
2408 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2409 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2410 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2411 ENGINE support for HMAC keys which are unextractable. New -mac and
2412 -macopt options to dgst utility.
2415 *) New option -sigopt to dgst utility. Update dgst to use
2416 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2417 alternative signing paramaters such as X9.31 or PSS in the dgst
2421 *) Change ssl_cipher_apply_rule(), the internal function that does
2422 the work each time a ciphersuite string requests enabling
2423 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2424 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2425 the order of disabled ciphersuites such that those ciphersuites
2426 that most recently went from enabled to disabled not only stay
2427 in order with respect to each other, but also have higher priority
2428 than other disabled ciphersuites the next time ciphersuites are
2431 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2432 the same ciphersuites as with "HIGH" alone, but in a specific
2433 order where the PSK ciphersuites come first (since they are the
2434 most recently disabled ciphersuites when "HIGH" is parsed).
2436 Also, change ssl_create_cipher_list() (using this new
2437 funcionality) such that between otherwise identical
2438 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2442 *) Change ssl_create_cipher_list() so that it automatically
2443 arranges the ciphersuites in reasonable order before starting
2444 to process the rule string. Thus, the definition for "DEFAULT"
2445 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2446 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2447 This makes it much easier to arrive at a reasonable default order
2448 in applications for which anonymous ciphers are OK (meaning
2449 that you can't actually use DEFAULT).
2450 [Bodo Moeller; suggested by Victor Duchovni]
2452 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2453 processing) into multiple integers instead of setting
2454 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2455 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2456 (These masks as well as the individual bit definitions are hidden
2457 away into the non-exported interface ssl/ssl_locl.h, so this
2458 change to the definition of the SSL_CIPHER structure shouldn't
2459 affect applications.) This give us more bits for each of these
2460 categories, so there is no longer a need to coagulate AES128 and
2461 AES256 into a single algorithm bit, and to coagulate Camellia128
2462 and Camellia256 into a single algorithm bit, which has led to all
2465 Thus, among other things, the kludge introduced in 0.9.7m and
2466 0.9.8e for masking out AES256 independently of AES128 or masking
2467 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2469 With the change, we also introduce new ciphersuite aliases that
2470 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2474 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2475 Use the leftmost N bytes of the signature input if the input is
2476 larger than the prime q (with N being the size in bytes of q).
2479 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2480 it yet and it is largely untested.
2483 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2486 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2487 some compilers (gcc 4.2 and later) reject their use. Safestack is
2488 reimplemented. Update ASN1 to avoid use of legacy functions.
2491 *) Win32/64 targets are linked with Winsock2.
2494 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2495 to external functions. This can be used to increase CRL handling
2496 efficiency especially when CRLs are very large by (for example) storing
2497 the CRL revoked certificates in a database.
2500 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2501 new CRLs added to a directory can be used. New command line option
2502 -verify_return_error to s_client and s_server. This causes real errors
2503 to be returned by the verify callback instead of carrying on no matter
2504 what. This reflects the way a "real world" verify callback would behave.
2507 *) GOST engine, supporting several GOST algorithms and public key formats.
2508 Kindly donated by Cryptocom.
2511 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2512 partitioned by DP are handled but no indirect CRL or reason partitioning
2513 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2514 selected via a scoring technique which handles IDP and AKID in CRLs.
2517 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2518 will ultimately be used for all verify operations: this will remove the
2519 X509_STORE dependency on certificate verification and allow alternative
2520 lookup methods. X509_STORE based implementations of these two callbacks.
2523 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2524 Modify get_crl() to find a valid (unexpired) CRL if possible.
2527 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2528 this would be called X509_CRL_cmp() but that name is already used by
2529 a function that just compares CRL issuer names. Cache several CRL
2530 extensions in X509_CRL structure and cache CRLDP in X509.
2533 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2534 this maps equivalent X509_NAME structures into a consistent structure.
2535 Name comparison can then be performed rapidly using memcmp().
2538 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2542 *) Allow digests to supply their own micalg string for S/MIME type using
2543 the ctrl EVP_MD_CTRL_MICALG.
2546 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2547 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2548 ctrl. It can then customise the structure before and/or after signing
2552 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2553 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2554 to free up any added signature OIDs.
2557 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2558 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2559 digest and cipher tables. New options added to openssl utility:
2560 list-message-digest-algorithms and list-cipher-algorithms.
2563 *) Change the array representation of binary polynomials: the list
2564 of degrees of non-zero coefficients is now terminated with -1.
2565 Previously it was terminated with 0, which was also part of the
2566 value; thus, the array representation was not applicable to
2567 polynomials where t^0 has coefficient zero. This change makes
2568 the array representation useful in a more general context.
2571 *) Various modifications and fixes to SSL/TLS cipher string
2572 handling. For ECC, the code now distinguishes between fixed ECDH
2573 with RSA certificates on the one hand and with ECDSA certificates
2574 on the other hand, since these are separate ciphersuites. The
2575 unused code for Fortezza ciphersuites has been removed.
2577 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2578 (not "ECDHE"). For consistency with the code for DH
2579 certificates, use of ECDH certificates is now considered ECDH
2580 authentication, not RSA or ECDSA authentication (the latter is
2581 merely the CA's signing algorithm and not actively used in the
2584 The temporary ciphersuite alias "ECCdraft" is no longer
2585 available, and ECC ciphersuites are no longer excluded from "ALL"
2586 and "DEFAULT". The following aliases now exist for RFC 4492
2587 ciphersuites, most of these by analogy with the DH case:
2589 kECDHr - ECDH cert, signed with RSA
2590 kECDHe - ECDH cert, signed with ECDSA
2591 kECDH - ECDH cert (signed with either RSA or ECDSA)
2592 kEECDH - ephemeral ECDH
2593 ECDH - ECDH cert or ephemeral ECDH
2599 AECDH - anonymous ECDH
2600 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2604 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2605 Use correct micalg parameters depending on digest(s) in signed message.
2608 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2609 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2612 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2613 an engine to register a method. Add ENGINE lookups for methods and
2614 functional reference processing.
2617 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2618 EVP_{Sign,Verify}* which allow an application to customise the signature
2622 *) New -resign option to smime utility. This adds one or more signers
2623 to an existing PKCS#7 signedData structure. Also -md option to use an
2624 alternative message digest algorithm for signing.
2627 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2628 create PKCS7 structures containing multiple signers. Update smime
2629 application to support multiple signers.
2632 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2636 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2637 Reorganize PBE internals to lookup from a static table using NIDs,
2638 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2639 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2640 PRF which will be automatically used with PBES2.
2643 *) Replace the algorithm specific calls to generate keys in "req" with the
2647 *) Update PKCS#7 enveloped data routines to use new API. This is now
2648 supported by any public key method supporting the encrypt operation. A
2649 ctrl is added to allow the public key algorithm to examine or modify
2650 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2654 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2655 a default digest type to use. In most cases this will be SHA1 but some
2656 algorithms (such as GOST) need to specify an alternative digest. The
2657 return value indicates how strong the preference is 1 means optional and
2658 2 is mandatory (that is it is the only supported type). Modify
2659 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2660 use the default md. Update openssl utilities to use the default digest
2661 type for signing if it is not explicitly indicated.
2664 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2665 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2666 signing method from the key type. This effectively removes the link
2667 between digests and public key types.
2670 *) Add an OID cross reference table and utility functions. Its purpose is to
2671 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2672 rsaEncryption. This will allow some of the algorithm specific hackery
2673 needed to use the correct OID to be removed.
2676 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2677 structures for PKCS7_sign(). They are now set up by the relevant public
2681 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2684 *) Add support for key derivation (agreement) in the API, DH method and
2688 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2689 public and private key formats. As a side effect these add additional
2690 command line functionality not previously available: DSA signatures can be
2691 generated and verified using pkeyutl and DH key support and generation in
2696 [Oliver Tappe <zooey@hirschkaefer.de>]
2698 *) New make target "install_html_docs" installs HTML renditions of the
2700 [Oliver Tappe <zooey@hirschkaefer.de>]
2702 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2703 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2704 support key and parameter generation and add initial key generation
2705 functionality for RSA.
2708 *) Add functions for main EVP_PKEY_method operations. The undocumented
2709 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2710 EVP_PKEY_{encrypt,decrypt}_old.
2713 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2714 key API, doesn't do much yet.
2717 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2718 public key algorithms. New option to openssl utility:
2719 "list-public-key-algorithms" to print out info.
2722 *) Implement the Supported Elliptic Curves Extension for
2723 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2726 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2727 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2730 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2731 utilities such as rsa, dsa, dsaparam etc except they process any key
2735 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2736 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2737 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2741 *) Initial support for pluggable public key ASN1.
2742 De-spaghettify the public key ASN1 handling. Move public and private
2743 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2744 algorithm specific handling to a single module within the relevant
2745 algorithm directory. Add functions to allow (near) opaque processing
2746 of public and private key structures.
2749 *) Implement the Supported Point Formats Extension for
2750 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2753 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2754 for the psk identity [hint] and the psk callback functions to the
2755 SSL_SESSION, SSL and SSL_CTX structure.
2758 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2762 SSL_CTX_use_psk_identity_hint
2763 SSL_get_psk_identity_hint
2764 SSL_get_psk_identity
2765 SSL_use_psk_identity_hint
2767 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2769 *) Add RFC 3161 compliant time stamp request creation, response generation
2770 and response verification functionality.
2771 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2773 *) Add initial support for TLS extensions, specifically for the server_name
2774 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2775 have new members for a host name. The SSL data structure has an
2776 additional member SSL_CTX *initial_ctx so that new sessions can be
2777 stored in that context to allow for session resumption, even after the
2778 SSL has been switched to a new SSL_CTX in reaction to a client's
2779 server_name extension.
2781 New functions (subject to change):
2783 SSL_get_servername()
2784 SSL_get_servername_type()
2787 New CTRL codes and macros (subject to change):
2789 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2790 - SSL_CTX_set_tlsext_servername_callback()
2791 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2792 - SSL_CTX_set_tlsext_servername_arg()
2793 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2795 openssl s_client has a new '-servername ...' option.
2797 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2798 '-key2 ...', '-servername_fatal' (subject to change). This allows
2799 testing the HostName extension for a specific single host name ('-cert'
2800 and '-key' remain fallbacks for handshakes without HostName
2801 negotiation). If the unrecognized_name alert has to be sent, this by
2802 default is a warning; it becomes fatal with the '-servername_fatal'
2805 [Peter Sylvester, Remy Allais, Christophe Renou]
2807 *) Whirlpool hash implementation is added.
2810 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2811 bn(64,32). Because of instruction set limitations it doesn't have
2812 any negative impact on performance. This was done mostly in order
2813 to make it possible to share assembler modules, such as bn_mul_mont
2814 implementations, between 32- and 64-bit builds without hassle.
2817 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2818 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2822 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2823 dedicated Montgomery multiplication procedure, is introduced.
2824 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2825 "64-bit" performance on certain 32-bit targets.
2828 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2829 in SSL structures. New SSL ctrl to set maximum send fragment size.
2830 Save memory by seeting the I/O buffer sizes dynamically instead of
2831 using the maximum available value.
2834 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2835 in addition to the text details.
2838 *) Very, very preliminary EXPERIMENTAL support for printing of general
2839 ASN1 structures. This currently produces rather ugly output and doesn't
2840 handle several customised structures at all.
2843 *) Integrated support for PVK file format and some related formats such
2844 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2845 these in the 'rsa' and 'dsa' utilities.
2848 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2851 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2852 place for the (very old) "NETSCAPE" format certificates which are now
2853 handled using new ASN1 code equivalents.
2856 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2857 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2858 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2861 *) Modify CRL distribution points extension code to print out previously
2862 unsupported fields. Enhance extension setting code to allow setting of
2866 *) Add print and set support for Issuing Distribution Point CRL extension.
2869 *) Change 'Configure' script to enable Camellia by default.
2872 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2874 *) When rejecting SSL/TLS records due to an incorrect version number, never
2875 update s->server with a new major version number. As of
2876 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2877 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2878 the previous behavior could result in a read attempt at NULL when
2879 receiving specific incorrect SSL/TLS records once record payload
2880 protection is active. (CVE-2010-0740)
2881 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2883 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2884 could be crashed if the relevant tables were not present (e.g. chrooted).
2885 [Tomas Hoger <thoger@redhat.com>]
2887 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2889 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2890 [Martin Olsson, Neel Mehta]
2892 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2893 accommodate for stack sorting, always a write lock!).
2896 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2897 excessive delays in the RAND_poll(): over a minute. As a workaround
2898 include a time check in the inner Heap32Next loop too.
2901 *) The code that handled flushing of data in SSL/TLS originally used the
2902 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2903 the problem outlined in PR#1949. The fix suggested there however can
2904 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2905 of Apache). So instead simplify the code to flush unconditionally.
2906 This should be fine since flushing with no data to flush is a no op.
2909 *) Handle TLS versions 2.0 and later properly and correctly use the
2910 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2911 off ancient servers have a habit of sticking around for a while...
2914 *) Modify compression code so it frees up structures without using the
2915 ex_data callbacks. This works around a problem where some applications
2916 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2917 restarting) then use compression (e.g. SSL with compression) later.
2918 This results in significant per-connection memory leaks and
2919 has caused some security issues including CVE-2008-1678 and
2923 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2924 change when encrypting or decrypting.
2927 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2928 connect and renegotiate with servers which do not support RI.
2929 Until RI is more widely deployed this option is enabled by default.
2932 *) Add "missing" ssl ctrls to clear options and mode.
2935 *) If client attempts to renegotiate and doesn't support RI respond with
2936 a no_renegotiation alert as required by RFC5746. Some renegotiating
2937 TLS clients will continue a connection gracefully when they receive
2938 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2939 waiting for a server hello which it will never receive. Now we treat a
2940 received no_renegotiation alert as a fatal error. This is because
2941 applications requesting a renegotiation might well expect it to succeed
2942 and would have no code in place to handle the server denying it so the
2943 only safe thing to do is to terminate the connection.
2946 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2947 peer supports secure renegotiation and 0 otherwise. Print out peer
2948 renegotiation support in s_client/s_server.
2951 *) Replace the highly broken and deprecated SPKAC certification method with
2952 the updated NID creation version. This should correctly handle UTF8.
2955 *) Implement RFC5746. Re-enable renegotiation but require the extension
2956 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2957 turns out to be a bad idea. It has been replaced by
2958 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2959 SSL_CTX_set_options(). This is really not recommended unless you
2960 know what you are doing.
2961 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2963 *) Fixes to stateless session resumption handling. Use initial_ctx when
2964 issuing and attempting to decrypt tickets in case it has changed during
2965 servername handling. Use a non-zero length session ID when attempting
2966 stateless session resumption: this makes it possible to determine if
2967 a resumption has occurred immediately after receiving server hello
2968 (several places in OpenSSL subtly assume this) instead of later in
2972 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2973 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2974 fixes for a few places where the return code is not checked
2976 [Julia Lawall <julia@diku.dk>]
2978 *) Add --strict-warnings option to Configure script to include devteam
2979 warnings in other configurations.
2982 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2983 makes it possible to install openssl libraries in locations which
2984 have names other than "lib", for example "/usr/lib64" which some
2986 [Steve Henson, based on patch from Jeremy Utley]
2988 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2989 X690 8.9.12 and can produce some misleading textual output of OIDs.
2990 [Steve Henson, reported by Dan Kaminsky]
2992 *) Delete MD2 from algorithm tables. This follows the recommendation in
2993 several standards that it is not used in new applications due to
2994 several cryptographic weaknesses. For binary compatibility reasons
2995 the MD2 API is still compiled in by default.
2998 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3002 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3003 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3005 [Guenter <lists@gknw.net>]
3007 *) Fix the server certificate chain building code to use X509_verify_cert(),
3008 it used to have an ad-hoc builder which was unable to cope with anything
3009 other than a simple chain.
3010 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3012 *) Don't check self signed certificate signatures in X509_verify_cert()
3013 by default (a flag can override this): it just wastes time without
3014 adding any security. As a useful side effect self signed root CAs
3015 with non-FIPS digests are now usable in FIPS mode.
3018 *) In dtls1_process_out_of_seq_message() the check if the current message
3019 is already buffered was missing. For every new message was memory
3020 allocated, allowing an attacker to perform an denial of service attack
3021 with sending out of seq handshake messages until there is no memory
3022 left. Additionally every future messege was buffered, even if the
3023 sequence number made no sense and would be part of another handshake.
3024 So only messages with sequence numbers less than 10 in advance will be
3025 buffered. (CVE-2009-1378)
3026 [Robin Seggelmann, discovered by Daniel Mentz]
3028 *) Records are buffered if they arrive with a future epoch to be
3029 processed after finishing the corresponding handshake. There is
3030 currently no limitation to this buffer allowing an attacker to perform
3031 a DOS attack with sending records with future epochs until there is no
3032 memory left. This patch adds the pqueue_size() function to determine
3033 the size of a buffer and limits the record buffer to 100 entries.
3035 [Robin Seggelmann, discovered by Daniel Mentz]
3037 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3038 parent structure is freed. (CVE-2009-1379)
3041 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3042 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3045 [Ilya O. <vrghost@gmail.com>]
3047 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3049 *) Disable renegotiation completely - this fixes a severe security
3050 problem (CVE-2009-3555) at the cost of breaking all
3051 renegotiation. Renegotiation can be re-enabled by setting
3052 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3053 run-time. This is really not recommended unless you know what
3057 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3059 *) Don't set val to NULL when freeing up structures, it is freed up by
3060 underlying code. If sizeof(void *) > sizeof(long) this can result in
3061 zeroing past the valid field. (CVE-2009-0789)
3062 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3064 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3065 checked correctly. This would allow some invalid signed attributes to
3066 appear to verify correctly. (CVE-2009-0591)
3067 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3069 *) Reject UniversalString and BMPString types with invalid lengths. This
3070 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3071 a legal length. (CVE-2009-0590)
3074 *) Set S/MIME signing as the default purpose rather than setting it
3075 unconditionally. This allows applications to override it at the store
3079 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3080 to handle some structures.
3083 *) Improve efficiency of mem_gets: don't search whole buffer each time
3085 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3087 *) New -hex option for openssl rand.
3090 *) Print out UTF8String and NumericString when parsing ASN1.
3093 *) Support NumericString type for name components.
3096 *) Allow CC in the environment to override the automatically chosen
3097 compiler. Note that nothing is done to ensure flags work with the
3101 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3103 *) Properly check EVP_VerifyFinal() and similar return values
3105 [Ben Laurie, Bodo Moeller, Google Security Team]
3107 *) Enable TLS extensions by default.
3110 *) Allow the CHIL engine to be loaded, whether the application is
3111 multithreaded or not. (This does not release the developer from the
3112 obligation to set up the dynamic locking callbacks.)
3113 [Sander Temme <sander@temme.net>]
3115 *) Use correct exit code if there is an error in dgst command.
3116 [Steve Henson; problem pointed out by Roland Dirlewanger]
3118 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3119 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3122 *) Add experimental JPAKE support, including demo authentication in
3123 s_client and s_server.
3126 *) Set the comparison function in v3_addr_canonize().
3127 [Rob Austein <sra@hactrn.net>]
3129 *) Add support for XMPP STARTTLS in s_client.
3130 [Philip Paeps <philip@freebsd.org>]
3132 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3133 to ensure that even with this option, only ciphersuites in the
3134 server's preference list will be accepted. (Note that the option
3135 applies only when resuming a session, so the earlier behavior was
3136 just about the algorithm choice for symmetric cryptography.)
3139 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3141 *) Fix NULL pointer dereference if a DTLS server received
3142 ChangeCipherSpec as first record (CVE-2009-1386).
3145 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3146 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3149 *) The fix in 0.9.8c that supposedly got rid of unsafe
3150 double-checked locking was incomplete for RSA blinding,
3151 addressing just one layer of what turns out to have been
3152 doubly unsafe triple-checked locking.
3154 So now fix this for real by retiring the MONT_HELPER macro
3155 in crypto/rsa/rsa_eay.c.
3157 [Bodo Moeller; problem pointed out by Marius Schilder]
3159 *) Various precautionary measures:
3161 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3163 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3164 (NB: This would require knowledge of the secret session ticket key
3165 to exploit, in which case you'd be SOL either way.)
3167 - Change bn_nist.c so that it will properly handle input BIGNUMs
3168 outside the expected range.
3170 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3173 [Neel Mehta, Bodo Moeller]
3175 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3176 the load fails. Useful for distros.
3177 [Ben Laurie and the FreeBSD team]
3179 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3182 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3185 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3187 This work was sponsored by Logica.
3190 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3191 keystores. Support for SSL/TLS client authentication too.
3192 Not compiled unless enable-capieng specified to Configure.
3194 This work was sponsored by Logica.
3197 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3198 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3199 attribute creation routines such as certifcate requests and PKCS#12
3203 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3205 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3206 handshake which could lead to a cilent crash as found using the
3207 Codenomicon TLS test suite (CVE-2008-1672)
3208 [Steve Henson, Mark Cox]
3210 *) Fix double free in TLS server name extensions which could lead to
3211 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3214 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3216 Clear the error queue to ensure that error entries left from
3217 older function calls do not interfere with the correct operation.
3218 [Lutz Jaenicke, Erik de Castro Lopo]
3220 *) Remove root CA certificates of commercial CAs:
3222 The OpenSSL project does not recommend any specific CA and does not
3223 have any policy with respect to including or excluding any CA.
3224 Therefore it does not make any sense to ship an arbitrary selection
3225 of root CA certificates with the OpenSSL software.
3228 *) RSA OAEP patches to fix two separate invalid memory reads.
3229 The first one involves inputs when 'lzero' is greater than
3230 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3231 before the beginning of from). The second one involves inputs where
3232 the 'db' section contains nothing but zeroes (there is a one-byte
3233 invalid read after the end of 'db').
3234 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3236 *) Partial backport from 0.9.9-dev:
3238 Introduce bn_mul_mont (dedicated Montgomery multiplication
3239 procedure) as a candidate for BIGNUM assembler implementation.
3240 While 0.9.9-dev uses assembler for various architectures, only
3241 x86_64 is available by default here in the 0.9.8 branch, and
3242 32-bit x86 is available through a compile-time setting.
3244 To try the 32-bit x86 assembler implementation, use Configure
3245 option "enable-montasm" (which exists only for this backport).
3247 As "enable-montasm" for 32-bit x86 disclaims code stability
3248 anyway, in this constellation we activate additional code
3249 backported from 0.9.9-dev for further performance improvements,
3250 namely BN_from_montgomery_word. (To enable this otherwise,
3251 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3253 [Andy Polyakov (backport partially by Bodo Moeller)]
3255 *) Add TLS session ticket callback. This allows an application to set
3256 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3257 values. This is useful for key rollover for example where several key
3258 sets may exist with different names.
3261 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3262 This was broken until now in 0.9.8 releases, such that the only way
3263 a registered ENGINE could be used (assuming it initialises
3264 successfully on the host) was to explicitly set it as the default
3265 for the relevant algorithms. This is in contradiction with 0.9.7
3266 behaviour and the documentation. With this fix, when an ENGINE is
3267 registered into a given algorithm's table of implementations, the
3268 'uptodate' flag is reset so that auto-discovery will be used next
3269 time a new context for that algorithm attempts to select an
3271 [Ian Lister (tweaked by Geoff Thorpe)]
3273 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3274 implemention in the following ways:
3276 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3279 Lack of BER streaming support means one pass streaming processing is
3280 only supported if data is detached: setting the streaming flag is
3281 ignored for embedded content.
3283 CMS support is disabled by default and must be explicitly enabled
3284 with the enable-cms configuration option.
3287 *) Update the GMP engine glue to do direct copies between BIGNUM and
3288 mpz_t when openssl and GMP use the same limb size. Otherwise the
3289 existing "conversion via a text string export" trick is still used.
3290 [Paul Sheer <paulsheer@gmail.com>]
3292 *) Zlib compression BIO. This is a filter BIO which compressed and
3293 uncompresses any data passed through it.
3296 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3297 RFC3394 compatible AES key wrapping.
3300 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3301 sets string data without copying. X509_ALGOR_set0() and
3302 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3303 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3304 from an X509_ATTRIBUTE structure optionally checking it occurs only
3305 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3309 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3310 to get the expected BN_FLG_CONSTTIME behavior.
3311 [Bodo Moeller (Google)]
3315 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3316 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3317 - added some more tests to do_tests.pl
3318 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3319 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3320 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3321 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3322 - various changes to netware.pl to enable gcc-cross builds on Win32
3324 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3325 - various changes to fix missing prototype warnings
3326 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3327 - added AES, WHIRLPOOL and CPUID assembler code to build files
3328 - added missing AES assembler make rules to mk1mf.pl
3329 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3330 [Guenter Knauf <eflash@gmx.net>]
3332 *) Implement certificate status request TLS extension defined in RFC3546.
3333 A client can set the appropriate parameters and receive the encoded
3334 OCSP response via a callback. A server can query the supplied parameters
3335 and set the encoded OCSP response in the callback. Add simplified examples
3336 to s_client and s_server.
3339 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3341 *) Fix various bugs:
3342 + Binary incompatibility of ssl_ctx_st structure
3343 + DTLS interoperation with non-compliant servers
3344 + Don't call get_session_cb() without proposed session
3345 + Fix ia64 assembler code
3346 [Andy Polyakov, Steve Henson]
3348 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3350 *) DTLS Handshake overhaul. There were longstanding issues with
3351 OpenSSL DTLS implementation, which were making it impossible for
3352 RFC 4347 compliant client to communicate with OpenSSL server.
3353 Unfortunately just fixing these incompatibilities would "cut off"
3354 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3355 server keeps tolerating non RFC compliant syntax. The opposite is
3356 not true, 0.9.8f client can not communicate with earlier server.
3357 This update even addresses CVE-2007-4995.
3360 *) Changes to avoid need for function casts in OpenSSL: some compilers
3361 (gcc 4.2 and later) reject their use.
3362 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3365 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3366 RFC4507bis. The encrypted ticket format is an encrypted encoded
3367 SSL_SESSION structure, that way new session features are automatically
3370 If a client application caches session in an SSL_SESSION structure
3371 support is transparent because tickets are now stored in the encoded
3374 The SSL_CTX structure automatically generates keys for ticket
3375 protection in servers so again support should be possible
3376 with no application modification.
3378 If a client or server wishes to disable RFC4507 support then the option
3379 SSL_OP_NO_TICKET can be set.
3381 Add a TLS extension debugging callback to allow the contents of any client
3382 or server extensions to be examined.
3384 This work was sponsored by Google.
3387 *) Add initial support for TLS extensions, specifically for the server_name
3388 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3389 have new members for a host name. The SSL data structure has an
3390 additional member SSL_CTX *initial_ctx so that new sessions can be
3391 stored in that context to allow for session resumption, even after the
3392 SSL has been switched to a new SSL_CTX in reaction to a client's
3393 server_name extension.
3395 New functions (subject to change):
3397 SSL_get_servername()
3398 SSL_get_servername_type()
3401 New CTRL codes and macros (subject to change):
3403 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3404 - SSL_CTX_set_tlsext_servername_callback()
3405 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3406 - SSL_CTX_set_tlsext_servername_arg()
3407 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3409 openssl s_client has a new '-servername ...' option.
3411 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3412 '-key2 ...', '-servername_fatal' (subject to change). This allows
3413 testing the HostName extension for a specific single host name ('-cert'
3414 and '-key' remain fallbacks for handshakes without HostName
3415 negotiation). If the unrecognized_name alert has to be sent, this by
3416 default is a warning; it becomes fatal with the '-servername_fatal'
3419 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3421 *) Add AES and SSE2 assembly language support to VC++ build.
3424 *) Mitigate attack on final subtraction in Montgomery reduction.
3427 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3428 (which previously caused an internal error).
3431 *) Squeeze another 10% out of IGE mode when in != out.
3434 *) AES IGE mode speedup.
3435 [Dean Gaudet (Google)]
3437 *) Add the Korean symmetric 128-bit cipher SEED (see
3438 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3439 add SEED ciphersuites from RFC 4162:
3441 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3442 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3443 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3444 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3446 To minimize changes between patchlevels in the OpenSSL 0.9.8
3447 series, SEED remains excluded from compilation unless OpenSSL
3448 is configured with 'enable-seed'.
3449 [KISA, Bodo Moeller]
3451 *) Mitigate branch prediction attacks, which can be practical if a
3452 single processor is shared, allowing a spy process to extract
3453 information. For detailed background information, see
3454 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3455 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3456 and Necessary Software Countermeasures"). The core of the change
3457 are new versions BN_div_no_branch() and
3458 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3459 respectively, which are slower, but avoid the security-relevant
3460 conditional branches. These are automatically called by BN_div()
3461 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3462 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3463 remove a conditional branch.
3465 BN_FLG_CONSTTIME is the new name for the previous
3466 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3467 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3468 in the exponent causes BN_mod_exp_mont() to use the alternative
3469 implementation in BN_mod_exp_mont_consttime().) The old name
3470 remains as a deprecated alias.
3472 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3473 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3474 constant-time implementations for more than just exponentiation.
3475 Here too the old name is kept as a deprecated alias.
3477 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3478 the BN_BLINDING structure gets an independent copy of the
3479 modulus. This means that the previous "BIGNUM *m" argument to
3480 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3481 essentially becomes "const BIGNUM *m", although we can't actually
3482 change this in the header file before 0.9.9. It allows
3483 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3484 enable BN_FLG_CONSTTIME.
3486 [Matthew D Wood (Intel Corp)]
3488 *) In the SSL/TLS server implementation, be strict about session ID
3489 context matching (which matters if an application uses a single
3490 external cache for different purposes). Previously,
3491 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3492 set. This did ensure strict client verification, but meant that,
3493 with applications using a single external cache for quite
3494 different requirements, clients could circumvent ciphersuite
3495 restrictions for a given session ID context by starting a session
3496 in a different context.
3499 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3500 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3501 authentication-only ciphersuites.
3504 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3505 not complete and could lead to a possible single byte overflow
3506 (CVE-2007-5135) [Ben Laurie]
3508 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3510 *) Since AES128 and AES256 (and similarly Camellia128 and
3511 Camellia256) share a single mask bit in the logic of
3512 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3513 kludge to work properly if AES128 is available and AES256 isn't
3514 (or if Camellia128 is available and Camellia256 isn't).
3517 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3518 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3519 When a point or a seed is encoded in a BIT STRING, we need to
3520 prevent the removal of trailing zero bits to get the proper DER
3521 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3522 of a NamedBitList, for which trailing 0 bits need to be removed.)
3525 *) Have SSL/TLS server implementation tolerate "mismatched" record
3526 protocol version while receiving ClientHello even if the
3527 ClientHello is fragmented. (The server can't insist on the
3528 particular protocol version it has chosen before the ServerHello
3529 message has informed the client about his choice.)
3532 *) Add RFC 3779 support.
3533 [Rob Austein for ARIN, Ben Laurie]
3535 *) Load error codes if they are not already present instead of using a
3536 static variable. This allows them to be cleanly unloaded and reloaded.
3537 Improve header file function name parsing.
3540 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3541 or CAPABILITY handshake as required by RFCs.
3544 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3546 *) Introduce limits to prevent malicious keys being able to
3547 cause a denial of service. (CVE-2006-2940)
3548 [Steve Henson, Bodo Moeller]
3550 *) Fix ASN.1 parsing of certain invalid structures that can result
3551 in a denial of service. (CVE-2006-2937) [Steve Henson]
3553 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3554 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3556 *) Fix SSL client code which could crash if connecting to a
3557 malicious SSLv2 server. (CVE-2006-4343)
3558 [Tavis Ormandy and Will Drewry, Google Security Team]
3560 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3561 match only those. Before that, "AES256-SHA" would be interpreted
3562 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3563 the same strength classification in 0.9.7h) as we currently only
3564 have a single AES bit in the ciphersuite description bitmap.
3565 That change, however, also applied to ciphersuite strings such as
3566 "RC4-MD5" that intentionally matched multiple ciphersuites --
3567 namely, SSL 2.0 ciphersuites in addition to the more common ones
3568 from SSL 3.0/TLS 1.0.
3570 So we change the selection algorithm again: Naming an explicit
3571 ciphersuite selects this one ciphersuite, and any other similar
3572 ciphersuite (same bitmap) from *other* protocol versions.
3573 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3574 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3576 Since SSL 2.0 does not have any ciphersuites for which the
3577 128/256 bit distinction would be relevant, this works for now.
3578 The proper fix will be to use different bits for AES128 and
3579 AES256, which would have avoided the problems from the beginning;
3580 however, bits are scarce, so we can only do this in a new release
3581 (not just a patchlevel) when we can change the SSL_CIPHER
3582 definition to split the single 'unsigned long mask' bitmap into
3583 multiple values to extend the available space.
3587 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3589 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3590 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3592 *) Add AES IGE and biIGE modes.
3595 *) Change the Unix randomness entropy gathering to use poll() when
3596 possible instead of select(), since the latter has some
3597 undesirable limitations.
3598 [Darryl Miles via Richard Levitte and Bodo Moeller]
3600 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3601 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3602 cannot be implicitly activated as part of, e.g., the "AES" alias.
3603 However, please upgrade to OpenSSL 0.9.9[-dev] for
3604 non-experimental use of the ECC ciphersuites to get TLS extension
3605 support, which is required for curve and point format negotiation
3606 to avoid potential handshake problems.
3609 *) Disable rogue ciphersuites:
3611 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3612 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3613 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3615 The latter two were purportedly from
3616 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3619 Also deactivate the remaining ciphersuites from
3620 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3621 unofficial, and the ID has long expired.
3624 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3625 dual-core machines) and other potential thread-safety issues.
3628 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3629 versions), which is now available for royalty-free use
3630 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3631 Also, add Camellia TLS ciphersuites from RFC 4132.
3633 To minimize changes between patchlevels in the OpenSSL 0.9.8
3634 series, Camellia remains excluded from compilation unless OpenSSL
3635 is configured with 'enable-camellia'.
3638 *) Disable the padding bug check when compression is in use. The padding
3639 bug check assumes the first packet is of even length, this is not
3640 necessarily true if compresssion is enabled and can result in false
3641 positives causing handshake failure. The actual bug test is ancient
3642 code so it is hoped that implementations will either have fixed it by
3643 now or any which still have the bug do not support compression.
3646 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3648 *) When applying a cipher rule check to see if string match is an explicit
3649 cipher suite and only match that one cipher suite if it is.
3652 *) Link in manifests for VC++ if needed.
3653 [Austin Ziegler <halostatue@gmail.com>]
3655 *) Update support for ECC-based TLS ciphersuites according to
3656 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3657 TLS extensions, which are supported starting with the 0.9.9
3658 branch, not in the OpenSSL 0.9.8 branch).
3661 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3662 opaque EVP_CIPHER_CTX handling.
3665 *) Fixes and enhancements to zlib compression code. We now only use
3666 "zlib1.dll" and use the default __cdecl calling convention on Win32
3667 to conform with the standards mentioned here:
3668 http://www.zlib.net/DLL_FAQ.txt
3669 Static zlib linking now works on Windows and the new --with-zlib-include
3670 --with-zlib-lib options to Configure can be used to supply the location
3671 of the headers and library. Gracefully handle case where zlib library
3675 *) Several fixes and enhancements to the OID generation code. The old code
3676 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3677 handle numbers larger than ULONG_MAX, truncated printing and had a
3678 non standard OBJ_obj2txt() behaviour.
3681 *) Add support for building of engines under engine/ as shared libraries
3682 under VC++ build system.
3685 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3686 Hopefully, we will not see any false combination of paths any more.
3689 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3691 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3692 (part of SSL_OP_ALL). This option used to disable the
3693 countermeasure against man-in-the-middle protocol-version
3694 rollback in the SSL 2.0 server implementation, which is a bad
3695 idea. (CVE-2005-2969)
3697 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3698 for Information Security, National Institute of Advanced Industrial
3699 Science and Technology [AIST], Japan)]
3701 *) Add two function to clear and return the verify parameter flags.
3704 *) Keep cipherlists sorted in the source instead of sorting them at
3705 runtime, thus removing the need for a lock.
3708 *) Avoid some small subgroup attacks in Diffie-Hellman.
3709 [Nick Mathewson and Ben Laurie]
3711 *) Add functions for well-known primes.
3714 *) Extended Windows CE support.
3715 [Satoshi Nakamura and Andy Polyakov]
3717 *) Initialize SSL_METHOD structures at compile time instead of during
3718 runtime, thus removing the need for a lock.
3721 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3722 attempting to decrypt each encrypted key in turn. Add support to
3726 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3728 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3731 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3734 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3735 key into the same file any more.
3738 *) Add initial support for Win64, both IA64 and AMD64/x64&nbs
projects / openssl.git / blob