5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) New testing framework
8 The testing framework has been largely rewritten and is now using
9 perl and the perl modules Test::Harness and an extended variant of
10 Test::More called OpenSSL::Test to do its work. All test scripts in
11 test/ have been rewritten into test recipes, and all direct calls to
12 executables in test/Makefile have become individual recipes using the
13 simplified testing OpenSSL::Test::Simple.
15 For documentation on our testing modules, do:
17 perldoc test/testlib/OpenSSL/Test/Simple.pm
18 perldoc test/testlib/OpenSSL/Test.pm
22 *) In DSA_generate_parameters_ex, if the provided seed is too short,
24 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
26 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
27 from RFC4279, RFC4785, RFC5487, RFC5489.
29 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
30 original RSA_PSK patch.
33 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
34 era flag was never set throughout the codebase (only read). Also removed
35 SSL3_FLAGS_POP_BUFFER which was only used if
36 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
39 *) Changed the default name options in the "ca", "crl", "req" and "x509"
40 to be "oneline" instead of "compat".
43 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
44 not aware of clients that still exhibit this bug, and the workaround
45 hasn't been working properly for a while.
48 *) The return type of BIO_number_read() and BIO_number_written() as well as
49 the corresponding num_read and num_write members in the BIO structure has
50 changed from unsigned long to uint64_t. On platforms where an unsigned
51 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
55 *) Given the pervasive nature of TLS extensions it is inadvisable to run
56 OpenSSL without support for them. It also means that maintaining
57 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
58 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
61 *) Removed support for the two export grade static DH ciphersuites
62 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
63 were newly added (along with a number of other static DH ciphersuites) to
64 1.0.2. However the two export ones have *never* worked since they were
65 introduced. It seems strange in any case to be adding new export
66 ciphersuites, and given "logjam" it also does not seem correct to fix them.
69 *) Version negotiation has been rewritten. In particular SSLv23_method(),
70 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
71 and turned into macros which simply call the new preferred function names
72 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
73 should use the new names instead. Also as part of this change the ssl23.h
74 header file has been removed.
77 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
78 code and the associated standard is no longer considered fit-for-purpose.
81 *) RT2547 was closed. When generating a private key, try to make the
82 output file readable only by the owner. This behavior change might
83 be noticeable when interacting with other software.
85 *) Added HTTP GET support to the ocsp command.
88 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
91 *) Added support for TLS extended master secret from
92 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
93 initial patch which was a great help during development.
96 *) All libssl internal structures have been removed from the public header
97 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
98 now redundant). Users should not attempt to access internal structures
99 directly. Instead they should use the provided API functions.
102 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
103 Access to deprecated functions can be re-enabled by running config with
104 "enable-deprecated". In addition applications wishing to use deprecated
105 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
106 will, by default, disable some transitive includes that previously existed
107 in the header files (e.g. ec.h will no longer, by default, include bn.h)
110 *) Added support for OCB mode. OpenSSL has been granted a patent license
111 compatible with the OpenSSL license for use of OCB. Details are available
112 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
113 for OCB can be removed by calling config with no-ocb.
116 *) SSLv2 support has been removed. It still supports receiving a SSLv2
117 compatible client hello.
120 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
121 done while fixing the error code for the key-too-small case.
122 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
124 *) CA.sh has been removmed; use CA.pl instead.
127 *) Removed old DES API.
130 *) Remove various unsupported platforms:
136 Sinix/ReliantUNIX RM400
141 16-bit platforms such as WIN16
144 *) Clean up OPENSSL_NO_xxx #define's
145 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
146 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
147 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
148 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
149 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
150 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
151 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
152 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
153 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
154 Remove MS_STATIC; it's a relic from platforms <32 bits.
157 *) Cleaned up dead code
158 Remove all but one '#ifdef undef' which is to be looked at.
161 *) Clean up calling of xxx_free routines.
162 Just like free(), fix most of the xxx_free routines to accept
163 NULL. Remove the non-null checks from callers. Save much code.
166 *) Add secure heap for storage of private keys (when possible).
167 Add BIO_s_secmem(), CBIGNUM, etc.
168 Contributed by Akamai Technologies under our Corporate CLA.
171 *) Experimental support for a new, fast, unbiased prime candidate generator,
172 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
173 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
175 *) New output format NSS in the sess_id command line tool. This allows
176 exporting the session id and the master key in NSS keylog format.
177 [Martin Kaiser <martin@kaiser.cx>]
179 *) Harmonize version and its documentation. -f flag is used to display
181 [mancha <mancha1@zoho.com>]
183 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
184 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
185 [mancha <mancha1@zoho.com>]
187 *) Fix some double frees. These are not thought to be exploitable.
188 [mancha <mancha1@zoho.com>]
190 *) A missing bounds check in the handling of the TLS heartbeat extension
191 can be used to reveal up to 64k of memory to a connected client or
194 Thanks for Neel Mehta of Google Security for discovering this bug and to
195 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
196 preparing the fix (CVE-2014-0160)
197 [Adam Langley, Bodo Moeller]
199 *) Fix for the attack described in the paper "Recovering OpenSSL
200 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
201 by Yuval Yarom and Naomi Benger. Details can be obtained from:
202 http://eprint.iacr.org/2014/140
204 Thanks to Yuval Yarom and Naomi Benger for discovering this
205 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
206 [Yuval Yarom and Naomi Benger]
208 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
209 this fixes a limitation in previous versions of OpenSSL.
212 *) Experimental encrypt-then-mac support.
214 Experimental support for encrypt then mac from
215 draft-gutmann-tls-encrypt-then-mac-02.txt
217 To enable it set the appropriate extension number (0x42 for the test
218 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
220 For non-compliant peers (i.e. just about everything) this should have no
223 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
227 *) Add EVP support for key wrapping algorithms, to avoid problems with
228 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
229 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
230 algorithms and include tests cases.
233 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
237 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
238 MGF1 digest and OAEP label.
241 *) Make openssl verify return errors.
242 [Chris Palmer <palmer@google.com> and Ben Laurie]
244 *) New function ASN1_TIME_diff to calculate the difference between two
245 ASN1_TIME structures or one structure and the current time.
248 *) Update fips_test_suite to support multiple command line options. New
249 test to induce all self test errors in sequence and check expected
253 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
254 sign or verify all in one operation.
257 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
258 test programs and fips_test_suite. Includes functionality to parse
259 the minimal script output of fipsalgest.pl directly.
262 *) Add authorisation parameter to FIPS_module_mode_set().
265 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
268 *) Use separate DRBG fields for internal and external flags. New function
269 FIPS_drbg_health_check() to perform on demand health checking. Add
270 generation tests to fips_test_suite with reduced health check interval to
271 demonstrate periodic health checking. Add "nodh" option to
272 fips_test_suite to skip very slow DH test.
275 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
279 *) More extensive health check for DRBG checking many more failure modes.
280 New function FIPS_selftest_drbg_all() to handle every possible DRBG
281 combination: call this in fips_test_suite.
284 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
285 and POST to handle Dual EC cases.
288 *) Add support for canonical generation of DSA parameter 'g'. See
291 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
292 POST to handle HMAC cases.
295 *) Add functions FIPS_module_version() and FIPS_module_version_text()
296 to return numerical and string versions of the FIPS module number.
299 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
300 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
301 outside the validated module in the FIPS capable OpenSSL.
304 *) Minor change to DRBG entropy callback semantics. In some cases
305 there is no multiple of the block length between min_len and
306 max_len. Allow the callback to return more than max_len bytes
307 of entropy but discard any extra: it is the callback's responsibility
308 to ensure that the extra data discarded does not impact the
309 requested amount of entropy.
312 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
313 information in FIPS186-3, SP800-57 and SP800-131A.
316 *) CCM support via EVP. Interface is very similar to GCM case except we
317 must supply all data in one chunk (i.e. no update, final) and the
318 message length must be supplied if AAD is used. Add algorithm test
322 *) Initial version of POST overhaul. Add POST callback to allow the status
323 of POST to be monitored and/or failures induced. Modify fips_test_suite
324 to use callback. Always run all selftests even if one fails.
327 *) XTS support including algorithm test driver in the fips_gcmtest program.
328 Note: this does increase the maximum key length from 32 to 64 bytes but
329 there should be no binary compatibility issues as existing applications
330 will never use XTS mode.
333 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
334 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
335 performs algorithm blocking for unapproved PRNG types. Also do not
336 set PRNG type in FIPS_mode_set(): leave this to the application.
337 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
338 the standard OpenSSL PRNG: set additional data to a date time vector.
341 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
342 This shouldn't present any incompatibility problems because applications
343 shouldn't be using these directly and any that are will need to rethink
344 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
347 *) Extensive self tests and health checking required by SP800-90 DRBG.
348 Remove strength parameter from FIPS_drbg_instantiate and always
349 instantiate at maximum supported strength.
352 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
355 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
358 *) New function DH_compute_key_padded() to compute a DH key and pad with
359 leading zeroes if needed: this complies with SP800-56A et al.
362 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
363 anything, incomplete, subject to change and largely untested at present.
366 *) Modify fipscanisteronly build option to only build the necessary object
367 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
370 *) Add experimental option FIPSSYMS to give all symbols in
371 fipscanister.o and FIPS or fips prefix. This will avoid
372 conflicts with future versions of OpenSSL. Add perl script
373 util/fipsas.pl to preprocess assembly language source files
374 and rename any affected symbols.
377 *) Add selftest checks and algorithm block of non-fips algorithms in
378 FIPS mode. Remove DES2 from selftests.
381 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
382 return internal method without any ENGINE dependencies. Add new
383 tiny fips sign and verify functions.
386 *) New build option no-ec2m to disable characteristic 2 code.
389 *) New build option "fipscanisteronly". This only builds fipscanister.o
390 and (currently) associated fips utilities. Uses the file Makefile.fips
391 instead of Makefile.org as the prototype.
394 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
395 Update fips_gcmtest to use IV generator.
398 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
399 setting output buffer to NULL. The *Final function must be
400 called although it will not retrieve any additional data. The tag
401 can be set or retrieved with a ctrl. The IV length is by default 12
402 bytes (96 bits) but can be set to an alternative value. If the IV
403 length exceeds the maximum IV length (currently 16 bytes) it cannot be
407 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
408 underlying do_cipher function handles all cipher semantics itself
409 including padding and finalisation. This is useful if (for example)
410 an ENGINE cipher handles block padding itself. The behaviour of
411 do_cipher is subtly changed if this flag is set: the return value
412 is the number of characters written to the output buffer (zero is
413 no longer an error code) or a negative error code. Also if the
414 input buffer is NULL and length 0 finalisation should be performed.
417 *) If a candidate issuer certificate is already part of the constructed
418 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
421 *) Improve forward-security support: add functions
423 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
424 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
426 for use by SSL/TLS servers; the callback function will be called whenever a
427 new session is created, and gets to decide whether the session may be
428 cached to make it resumable (return 0) or not (return 1). (As by the
429 SSL/TLS protocol specifications, the session_id sent by the server will be
430 empty to indicate that the session is not resumable; also, the server will
431 not generate RFC 4507 (RFC 5077) session tickets.)
433 A simple reasonable callback implementation is to return is_forward_secure.
434 This parameter will be set to 1 or 0 depending on the ciphersuite selected
435 by the SSL/TLS server library, indicating whether it can provide forward
437 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
439 *) New -verify_name option in command line utilities to set verification
443 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
444 Add CMAC pkey methods.
447 *) Experimental renegotiation in s_server -www mode. If the client
448 browses /reneg connection is renegotiated. If /renegcert it is
449 renegotiated requesting a certificate.
452 *) Add an "external" session cache for debugging purposes to s_server. This
453 should help trace issues which normally are only apparent in deployed
454 multi-process servers.
457 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
458 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
459 BIO_set_cipher() and some obscure PEM functions were changed so they
460 can now return an error. The RAND changes required a change to the
461 RAND_METHOD structure.
464 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
465 a gcc attribute to warn if the result of a function is ignored. This
466 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
467 whose return value is often ignored.
470 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
472 *) Alternate chains certificate forgery
474 During certificate verfification, OpenSSL will attempt to find an
475 alternative certificate chain if the first attempt to build such a chain
476 fails. An error in the implementation of this logic can mean that an
477 attacker could cause certain checks on untrusted certificates to be
478 bypassed, such as the CA flag, enabling them to use a valid leaf
479 certificate to act as a CA and "issue" an invalid certificate.
481 This issue was reported to OpenSSL by Adam Langley/David Benjamin
485 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
487 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
488 incompatibility in the handling of HMAC. The previous ABI has now been
492 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
494 *) Malformed ECParameters causes infinite loop
496 When processing an ECParameters structure OpenSSL enters an infinite loop
497 if the curve specified is over a specially malformed binary polynomial
500 This can be used to perform denial of service against any
501 system which processes public keys, certificate requests or
502 certificates. This includes TLS clients and TLS servers with
503 client authentication enabled.
505 This issue was reported to OpenSSL by Joseph Barr-Pixton.
509 *) Exploitable out-of-bounds read in X509_cmp_time
511 X509_cmp_time does not properly check the length of the ASN1_TIME
512 string and can read a few bytes out of bounds. In addition,
513 X509_cmp_time accepts an arbitrary number of fractional seconds in the
516 An attacker can use this to craft malformed certificates and CRLs of
517 various sizes and potentially cause a segmentation fault, resulting in
518 a DoS on applications that verify certificates or CRLs. TLS clients
519 that verify CRLs are affected. TLS clients and servers with client
520 authentication enabled may be affected if they use custom verification
523 This issue was reported to OpenSSL by Robert Swiecki (Google), and
524 independently by Hanno Böck.
528 *) PKCS7 crash with missing EnvelopedContent
530 The PKCS#7 parsing code does not handle missing inner EncryptedContent
531 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
532 with missing content and trigger a NULL pointer dereference on parsing.
534 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
535 structures from untrusted sources are affected. OpenSSL clients and
536 servers are not affected.
538 This issue was reported to OpenSSL by Michal Zalewski (Google).
542 *) CMS verify infinite loop with unknown hash function
544 When verifying a signedData message the CMS code can enter an infinite loop
545 if presented with an unknown hash function OID. This can be used to perform
546 denial of service against any system which verifies signedData messages using
548 This issue was reported to OpenSSL by Johannes Bauer.
552 *) Race condition handling NewSessionTicket
554 If a NewSessionTicket is received by a multi-threaded client when attempting to
555 reuse a previous ticket then a race condition can occur potentially leading to
556 a double free of the ticket data.
560 *) Only support 256-bit or stronger elliptic curves with the
561 'ecdh_auto' setting (server) or by default (client). Of supported
562 curves, prefer P-256 (both).
565 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
567 *) ClientHello sigalgs DoS fix
569 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
570 invalid signature algorithms extension a NULL pointer dereference will
571 occur. This can be exploited in a DoS attack against the server.
573 This issue was was reported to OpenSSL by David Ramos of Stanford
576 [Stephen Henson and Matt Caswell]
578 *) Multiblock corrupted pointer fix
580 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
581 feature only applies on 64 bit x86 architecture platforms that support AES
582 NI instructions. A defect in the implementation of "multiblock" can cause
583 OpenSSL's internal write buffer to become incorrectly set to NULL when
584 using non-blocking IO. Typically, when the user application is using a
585 socket BIO for writing, this will only result in a failed connection.
586 However if some other BIO is used then it is likely that a segmentation
587 fault will be triggered, thus enabling a potential DoS attack.
589 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
593 *) Segmentation fault in DTLSv1_listen fix
595 The DTLSv1_listen function is intended to be stateless and processes the
596 initial ClientHello from many peers. It is common for user code to loop
597 over the call to DTLSv1_listen until a valid ClientHello is received with
598 an associated cookie. A defect in the implementation of DTLSv1_listen means
599 that state is preserved in the SSL object from one invocation to the next
600 that can lead to a segmentation fault. Errors processing the initial
601 ClientHello can trigger this scenario. An example of such an error could be
602 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
605 This issue was reported to OpenSSL by Per Allansson.
609 *) Segmentation fault in ASN1_TYPE_cmp fix
611 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
612 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
613 certificate signature algorithm consistency this can be used to crash any
614 certificate verification operation and exploited in a DoS attack. Any
615 application which performs certificate verification is vulnerable including
616 OpenSSL clients and servers which enable client authentication.
620 *) Segmentation fault for invalid PSS parameters fix
622 The signature verification routines will crash with a NULL pointer
623 dereference if presented with an ASN.1 signature using the RSA PSS
624 algorithm and invalid parameters. Since these routines are used to verify
625 certificate signature algorithms this can be used to crash any
626 certificate verification operation and exploited in a DoS attack. Any
627 application which performs certificate verification is vulnerable including
628 OpenSSL clients and servers which enable client authentication.
630 This issue was was reported to OpenSSL by Brian Carpenter.
634 *) ASN.1 structure reuse memory corruption fix
636 Reusing a structure in ASN.1 parsing may allow an attacker to cause
637 memory corruption via an invalid write. Such reuse is and has been
638 strongly discouraged and is believed to be rare.
640 Applications that parse structures containing CHOICE or ANY DEFINED BY
641 components may be affected. Certificate parsing (d2i_X509 and related
642 functions) are however not affected. OpenSSL clients and servers are
647 *) PKCS7 NULL pointer dereferences fix
649 The PKCS#7 parsing code does not handle missing outer ContentInfo
650 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
651 missing content and trigger a NULL pointer dereference on parsing.
653 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
654 otherwise parse PKCS#7 structures from untrusted sources are
655 affected. OpenSSL clients and servers are not affected.
657 This issue was reported to OpenSSL by Michal Zalewski (Google).
661 *) DoS via reachable assert in SSLv2 servers fix
663 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
664 servers that both support SSLv2 and enable export cipher suites by sending
665 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
667 This issue was discovered by Sean Burford (Google) and Emilia Käsper
668 (OpenSSL development team).
672 *) Empty CKE with client auth and DHE fix
674 If client auth is used then a server can seg fault in the event of a DHE
675 ciphersuite being selected and a zero length ClientKeyExchange message
676 being sent by the client. This could be exploited in a DoS attack.
680 *) Handshake with unseeded PRNG fix
682 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
683 with an unseeded PRNG. The conditions are:
684 - The client is on a platform where the PRNG has not been seeded
685 automatically, and the user has not seeded manually
686 - A protocol specific client method version has been used (i.e. not
687 SSL_client_methodv23)
688 - A ciphersuite is used that does not require additional random data from
689 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
691 If the handshake succeeds then the client random that has been used will
692 have been generated from a PRNG with insufficient entropy and therefore the
693 output may be predictable.
695 For example using the following command with an unseeded openssl will
696 succeed on an unpatched platform:
698 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
702 *) Use After Free following d2i_ECPrivatekey error fix
704 A malformed EC private key file consumed via the d2i_ECPrivateKey function
705 could cause a use after free condition. This, in turn, could cause a double
706 free in several private key parsing functions (such as d2i_PrivateKey
707 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
708 for applications that receive EC private keys from untrusted
709 sources. This scenario is considered rare.
711 This issue was discovered by the BoringSSL project and fixed in their
716 *) X509_to_X509_REQ NULL pointer deref fix
718 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
719 the certificate key is invalid. This function is rarely used in practice.
721 This issue was discovered by Brian Carpenter.
725 *) Removed the export ciphers from the DEFAULT ciphers
728 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
730 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
731 ARMv5 through ARMv8, as opposite to "locking" it to single one.
732 So far those who have to target multiple plaforms would compromise
733 and argue that binary targeting say ARMv5 would still execute on
734 ARMv8. "Universal" build resolves this compromise by providing
735 near-optimal performance even on newer platforms.
738 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
739 (other platforms pending).
740 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
742 *) Add support for the SignedCertificateTimestampList certificate and
743 OCSP response extensions from RFC6962.
746 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
747 for corner cases. (Certain input points at infinity could lead to
748 bogus results, with non-infinity inputs mapped to infinity too.)
751 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
752 This covers AES, SHA256/512 and GHASH. "Initial" means that most
753 common cases are optimized and there still is room for further
754 improvements. Vector Permutation AES for Altivec is also added.
757 *) Add support for little-endian ppc64 Linux target.
758 [Marcelo Cerri (IBM)]
760 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
761 SHA1, SHA256 and GHASH. "Initial" means that most common cases
762 are optimized and there still is room for further improvements.
763 Both 32- and 64-bit modes are supported.
764 [Andy Polyakov, Ard Biesheuvel (Linaro)]
766 *) Improved ARMv7 NEON support.
769 *) Support for SPARC Architecture 2011 crypto extensions, first
770 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
771 SHA256/512, MD5, GHASH and modular exponentiation.
772 [Andy Polyakov, David Miller]
774 *) Accelerated modular exponentiation for Intel processors, a.k.a.
776 [Shay Gueron & Vlad Krasnov (Intel Corp)]
778 *) Support for new and upcoming Intel processors, including AVX2,
779 BMI and SHA ISA extensions. This includes additional "stitched"
780 implementations, AESNI-SHA256 and GCM, and multi-buffer support
783 This work was sponsored by Intel Corp.
786 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
787 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
788 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
791 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
792 this fixes a limiation in previous versions of OpenSSL.
795 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
796 MGF1 digest and OAEP label.
799 *) Add EVP support for key wrapping algorithms, to avoid problems with
800 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
801 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
802 algorithms and include tests cases.
805 *) Add functions to allocate and set the fields of an ECDSA_METHOD
807 [Douglas E. Engert, Steve Henson]
809 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
810 difference in days and seconds between two tm or ASN1_TIME structures.
813 *) Add -rev test option to s_server to just reverse order of characters
814 received by client and send back to server. Also prints an abbreviated
815 summary of the connection parameters.
818 *) New option -brief for s_client and s_server to print out a brief summary
819 of connection parameters.
822 *) Add callbacks for arbitrary TLS extensions.
823 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
825 *) New option -crl_download in several openssl utilities to download CRLs
826 from CRLDP extension in certificates.
829 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
832 *) New function X509_CRL_diff to generate a delta CRL from the difference
833 of two full CRLs. Add support to "crl" utility.
836 *) New functions to set lookup_crls function and to retrieve
837 X509_STORE from X509_STORE_CTX.
840 *) Print out deprecated issuer and subject unique ID fields in
844 *) Extend OCSP I/O functions so they can be used for simple general purpose
845 HTTP as well as OCSP. New wrapper function which can be used to download
846 CRLs using the OCSP API.
849 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
852 *) SSL_CONF* functions. These provide a common framework for application
853 configuration using configuration files or command lines.
856 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
857 message callback and prints the results. Needs compile time option
858 "enable-ssl-trace". New options to s_client and s_server to enable
862 *) New ctrl and macro to retrieve supported points extensions.
863 Print out extension in s_server and s_client.
866 *) New functions to retrieve certificate signature and signature
870 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
874 *) New Suite B modes for TLS code. These use and enforce the requirements
875 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
876 only use Suite B curves. The Suite B modes can be set by using the
877 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
880 *) New chain verification flags for Suite B levels of security. Check
881 algorithms are acceptable when flags are set in X509_verify_cert.
884 *) Make tls1_check_chain return a set of flags indicating checks passed
885 by a certificate chain. Add additional tests to handle client
886 certificates: checks for matching certificate type and issuer name
890 *) If an attempt is made to use a signature algorithm not in the peer
891 preference list abort the handshake. If client has no suitable
892 signature algorithms in response to a certificate request do not
896 *) If server EC tmp key is not in client preference list abort handshake.
899 *) Add support for certificate stores in CERT structure. This makes it
900 possible to have different stores per SSL structure or one store in
901 the parent SSL_CTX. Include distint stores for certificate chain
902 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
903 to build and store a certificate chain in CERT structure: returing
904 an error if the chain cannot be built: this will allow applications
905 to test if a chain is correctly configured.
907 Note: if the CERT based stores are not set then the parent SSL_CTX
908 store is used to retain compatibility with existing behaviour.
912 *) New function ssl_set_client_disabled to set a ciphersuite disabled
913 mask based on the current session, check mask when sending client
914 hello and checking the requested ciphersuite.
917 *) New ctrls to retrieve and set certificate types in a certificate
918 request message. Print out received values in s_client. If certificate
919 types is not set with custom values set sensible values based on
920 supported signature algorithms.
923 *) Support for distinct client and server supported signature algorithms.
926 *) Add certificate callback. If set this is called whenever a certificate
927 is required by client or server. An application can decide which
928 certificate chain to present based on arbitrary criteria: for example
929 supported signature algorithms. Add very simple example to s_server.
930 This fixes many of the problems and restrictions of the existing client
931 certificate callback: for example you can now clear an existing
932 certificate and specify the whole chain.
935 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
936 the certificate can be used for (if anything). Set valid_flags field
937 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
938 to have similar checks in it.
940 Add new "cert_flags" field to CERT structure and include a "strict mode".
941 This enforces some TLS certificate requirements (such as only permitting
942 certificate signature algorithms contained in the supported algorithms
943 extension) which some implementations ignore: this option should be used
944 with caution as it could cause interoperability issues.
947 *) Update and tidy signature algorithm extension processing. Work out
948 shared signature algorithms based on preferences and peer algorithms
949 and print them out in s_client and s_server. Abort handshake if no
950 shared signature algorithms.
953 *) Add new functions to allow customised supported signature algorithms
954 for SSL and SSL_CTX structures. Add options to s_client and s_server
958 *) New function SSL_certs_clear() to delete all references to certificates
959 from an SSL structure. Before this once a certificate had been added
960 it couldn't be removed.
963 *) Integrate hostname, email address and IP address checking with certificate
964 verification. New verify options supporting checking in opensl utility.
967 *) Fixes and wildcard matching support to hostname and email checking
968 functions. Add manual page.
969 [Florian Weimer (Red Hat Product Security Team)]
971 *) New functions to check a hostname email or IP address against a
972 certificate. Add options x509 utility to print results of checks against
976 *) Fix OCSP checking.
977 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
979 *) Initial experimental support for explicitly trusted non-root CAs.
980 OpenSSL still tries to build a complete chain to a root but if an
981 intermediate CA has a trust setting included that is used. The first
982 setting is used: whether to trust (e.g., -addtrust option to the x509
986 *) Add -trusted_first option which attempts to find certificates in the
987 trusted store even if an untrusted chain is also supplied.
990 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
991 platform support for Linux and Android.
994 *) Support for linux-x32, ILP32 environment in x86_64 framework.
997 *) Experimental multi-implementation support for FIPS capable OpenSSL.
998 When in FIPS mode the approved implementations are used as normal,
999 when not in FIPS mode the internal unapproved versions are used instead.
1000 This means that the FIPS capable OpenSSL isn't forced to use the
1001 (often lower performance) FIPS implementations outside FIPS mode.
1004 *) Transparently support X9.42 DH parameters when calling
1005 PEM_read_bio_DHparameters. This means existing applications can handle
1006 the new parameter format automatically.
1009 *) Initial experimental support for X9.42 DH parameter format: mainly
1010 to support use of 'q' parameter for RFC5114 parameters.
1013 *) Add DH parameters from RFC5114 including test data to dhtest.
1016 *) Support for automatic EC temporary key parameter selection. If enabled
1017 the most preferred EC parameters are automatically used instead of
1018 hardcoded fixed parameters. Now a server just has to call:
1019 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1020 support ECDH and use the most appropriate parameters.
1023 *) Enhance and tidy EC curve and point format TLS extension code. Use
1024 static structures instead of allocation if default values are used.
1025 New ctrls to set curves we wish to support and to retrieve shared curves.
1026 Print out shared curves in s_server. New options to s_server and s_client
1027 to set list of supported curves.
1030 *) New ctrls to retrieve supported signature algorithms and
1031 supported curve values as an array of NIDs. Extend openssl utility
1032 to print out received values.
1035 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1036 between NIDs and the more common NIST names such as "P-256". Enhance
1037 ecparam utility and ECC method to recognise the NIST names for curves.
1040 *) Enhance SSL/TLS certificate chain handling to support different
1041 chains for each certificate instead of one chain in the parent SSL_CTX.
1044 *) Support for fixed DH ciphersuite client authentication: where both
1045 server and client use DH certificates with common parameters.
1048 *) Support for fixed DH ciphersuites: those requiring DH server
1052 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1054 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1055 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1056 X509_CINF_get_signature were reverted post internal team review.
1058 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1060 *) Build fixes for the Windows and OpenVMS platforms
1061 [Matt Caswell and Richard Levitte]
1063 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1065 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1066 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1067 dereference. This could lead to a Denial Of Service attack. Thanks to
1068 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1072 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1073 dtls1_buffer_record function under certain conditions. In particular this
1074 could occur if an attacker sent repeated DTLS records with the same
1075 sequence number but for the next epoch. The memory leak could be exploited
1076 by an attacker in a Denial of Service attack through memory exhaustion.
1077 Thanks to Chris Mueller for reporting this issue.
1081 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1082 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1083 method would be set to NULL which could later result in a NULL pointer
1084 dereference. Thanks to Frank Schmirler for reporting this issue.
1088 *) Abort handshake if server key exchange message is omitted for ephemeral
1091 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1092 reporting this issue.
1096 *) Remove non-export ephemeral RSA code on client and server. This code
1097 violated the TLS standard by allowing the use of temporary RSA keys in
1098 non-export ciphersuites and could be used by a server to effectively
1099 downgrade the RSA key length used to a value smaller than the server
1100 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1101 INRIA or reporting this issue.
1105 *) Fixed issue where DH client certificates are accepted without verification.
1106 An OpenSSL server will accept a DH certificate for client authentication
1107 without the certificate verify message. This effectively allows a client to
1108 authenticate without the use of a private key. This only affects servers
1109 which trust a client certificate authority which issues certificates
1110 containing DH keys: these are extremely rare and hardly ever encountered.
1111 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1116 *) Ensure that the session ID context of an SSL is updated when its
1117 SSL_CTX is updated via SSL_set_SSL_CTX.
1119 The session ID context is typically set from the parent SSL_CTX,
1120 and can vary with the CTX.
1123 *) Fix various certificate fingerprint issues.
1125 By using non-DER or invalid encodings outside the signed portion of a
1126 certificate the fingerprint can be changed without breaking the signature.
1127 Although no details of the signed portion of the certificate can be changed
1128 this can cause problems with some applications: e.g. those using the
1129 certificate fingerprint for blacklists.
1131 1. Reject signatures with non zero unused bits.
1133 If the BIT STRING containing the signature has non zero unused bits reject
1134 the signature. All current signature algorithms require zero unused bits.
1136 2. Check certificate algorithm consistency.
1138 Check the AlgorithmIdentifier inside TBS matches the one in the
1139 certificate signature. NB: this will result in signature failure
1140 errors for some broken certificates.
1142 Thanks to Konrad Kraszewski from Google for reporting this issue.
1144 3. Check DSA/ECDSA signatures use DER.
1146 Reencode DSA/ECDSA signatures and compare with the original received
1147 signature. Return an error if there is a mismatch.
1149 This will reject various cases including garbage after signature
1150 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1151 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1152 (negative or with leading zeroes).
1154 Further analysis was conducted and fixes were developed by Stephen Henson
1155 of the OpenSSL core team.
1160 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1161 results on some platforms, including x86_64. This bug occurs at random
1162 with a very low probability, and is not known to be exploitable in any
1163 way, though its exact impact is difficult to determine. Thanks to Pieter
1164 Wuille (Blockstream) who reported this issue and also suggested an initial
1165 fix. Further analysis was conducted by the OpenSSL development team and
1166 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1167 the OpenSSL core team.
1171 *) Do not resume sessions on the server if the negotiated protocol
1172 version does not match the session's version. Resuming with a different
1173 version, while not strictly forbidden by the RFC, is of questionable
1174 sanity and breaks all known clients.
1175 [David Benjamin, Emilia Käsper]
1177 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1178 early CCS messages during renegotiation. (Note that because
1179 renegotiation is encrypted, this early CCS was not exploitable.)
1182 *) Tighten client-side session ticket handling during renegotiation:
1183 ensure that the client only accepts a session ticket if the server sends
1184 the extension anew in the ServerHello. Previously, a TLS client would
1185 reuse the old extension state and thus accept a session ticket if one was
1186 announced in the initial ServerHello.
1188 Similarly, ensure that the client requires a session ticket if one
1189 was advertised in the ServerHello. Previously, a TLS client would
1190 ignore a missing NewSessionTicket message.
1193 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1195 *) SRTP Memory Leak.
1197 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1198 sends a carefully crafted handshake message, to cause OpenSSL to fail
1199 to free up to 64k of memory causing a memory leak. This could be
1200 exploited in a Denial Of Service attack. This issue affects OpenSSL
1201 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1202 whether SRTP is used or configured. Implementations of OpenSSL that
1203 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1205 The fix was developed by the OpenSSL team.
1209 *) Session Ticket Memory Leak.
1211 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1212 integrity of that ticket is first verified. In the event of a session
1213 ticket integrity check failing, OpenSSL will fail to free memory
1214 causing a memory leak. By sending a large number of invalid session
1215 tickets an attacker could exploit this issue in a Denial Of Service
1220 *) Build option no-ssl3 is incomplete.
1222 When OpenSSL is configured with "no-ssl3" as a build option, servers
1223 could accept and complete a SSL 3.0 handshake, and clients could be
1224 configured to send them.
1226 [Akamai and the OpenSSL team]
1228 *) Add support for TLS_FALLBACK_SCSV.
1229 Client applications doing fallback retries should call
1230 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1232 [Adam Langley, Bodo Moeller]
1234 *) Add additional DigestInfo checks.
1236 Reencode DigestInto in DER and check against the original when
1237 verifying RSA signature: this will reject any improperly encoded
1238 DigestInfo structures.
1240 Note: this is a precautionary measure and no attacks are currently known.
1244 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1246 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1247 SRP code can be overrun an internal buffer. Add sanity check that
1248 g, A, B < N to SRP code.
1250 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1251 Group for discovering this issue.
1255 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1256 TLS 1.0 instead of higher protocol versions when the ClientHello message
1257 is badly fragmented. This allows a man-in-the-middle attacker to force a
1258 downgrade to TLS 1.0 even if both the server and the client support a
1259 higher protocol version, by modifying the client's TLS records.
1261 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1262 researching this issue.
1266 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1267 to a denial of service attack. A malicious server can crash the client
1268 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1269 ciphersuite and sending carefully crafted handshake messages.
1271 Thanks to Felix Gröbert (Google) for discovering and researching this
1276 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1277 to leak memory. This can be exploited through a Denial of Service attack.
1278 Thanks to Adam Langley for discovering and researching this issue.
1282 *) An attacker can force openssl to consume large amounts of memory whilst
1283 processing DTLS handshake messages. This can be exploited through a
1284 Denial of Service attack.
1285 Thanks to Adam Langley for discovering and researching this issue.
1289 *) An attacker can force an error condition which causes openssl to crash
1290 whilst processing DTLS packets due to memory being freed twice. This
1291 can be exploited through a Denial of Service attack.
1292 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1297 *) If a multithreaded client connects to a malicious server using a resumed
1298 session and the server sends an ec point format extension it could write
1299 up to 255 bytes to freed memory.
1301 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1306 *) A malicious server can crash an OpenSSL client with a null pointer
1307 dereference (read) by specifying an SRP ciphersuite even though it was not
1308 properly negotiated with the client. This can be exploited through a
1309 Denial of Service attack.
1311 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1312 discovering and researching this issue.
1316 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1317 X509_name_oneline, X509_name_print_ex et al. to leak some information
1318 from the stack. Applications may be affected if they echo pretty printing
1319 output to the attacker.
1321 Thanks to Ivan Fratric (Google) for discovering this issue.
1323 [Emilia Käsper, and Steve Henson]
1325 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1326 for corner cases. (Certain input points at infinity could lead to
1327 bogus results, with non-infinity inputs mapped to infinity too.)
1330 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1332 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1333 handshake can force the use of weak keying material in OpenSSL
1334 SSL/TLS clients and servers.
1336 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1337 researching this issue. (CVE-2014-0224)
1338 [KIKUCHI Masashi, Steve Henson]
1340 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1341 OpenSSL DTLS client the code can be made to recurse eventually crashing
1344 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1346 [Imre Rad, Steve Henson]
1348 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1349 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1350 client or server. This is potentially exploitable to run arbitrary
1351 code on a vulnerable client or server.
1353 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1354 [Jüri Aedla, Steve Henson]
1356 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1357 are subject to a denial of service attack.
1359 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1360 this issue. (CVE-2014-3470)
1361 [Felix Gröbert, Ivan Fratric, Steve Henson]
1363 *) Harmonize version and its documentation. -f flag is used to display
1365 [mancha <mancha1@zoho.com>]
1367 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1368 in i2d_ECPrivateKey.
1369 [mancha <mancha1@zoho.com>]
1371 *) Fix some double frees. These are not thought to be exploitable.
1372 [mancha <mancha1@zoho.com>]
1374 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1376 *) A missing bounds check in the handling of the TLS heartbeat extension
1377 can be used to reveal up to 64k of memory to a connected client or
1380 Thanks for Neel Mehta of Google Security for discovering this bug and to
1381 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1382 preparing the fix (CVE-2014-0160)
1383 [Adam Langley, Bodo Moeller]
1385 *) Fix for the attack described in the paper "Recovering OpenSSL
1386 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1387 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1388 http://eprint.iacr.org/2014/140
1390 Thanks to Yuval Yarom and Naomi Benger for discovering this
1391 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1392 [Yuval Yarom and Naomi Benger]
1394 *) TLS pad extension: draft-agl-tls-padding-03
1396 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1397 TLS client Hello record length value would otherwise be > 255 and
1398 less that 512 pad with a dummy extension containing zeroes so it
1399 is at least 512 bytes long.
1401 [Adam Langley, Steve Henson]
1403 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1405 *) Fix for TLS record tampering bug. A carefully crafted invalid
1406 handshake could crash OpenSSL with a NULL pointer exception.
1407 Thanks to Anton Johansson for reporting this issues.
1410 *) Keep original DTLS digest and encryption contexts in retransmission
1411 structures so we can use the previous session parameters if they need
1412 to be resent. (CVE-2013-6450)
1415 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1416 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1417 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1418 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1419 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1420 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1421 [Rob Stradling, Adam Langley]
1423 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1425 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1426 supporting platforms or when small records were transferred.
1427 [Andy Polyakov, Steve Henson]
1429 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1431 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1433 This addresses the flaw in CBC record processing discovered by
1434 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1435 at: http://www.isg.rhul.ac.uk/tls/
1437 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1438 Security Group at Royal Holloway, University of London
1439 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1440 Emilia Käsper for the initial patch.
1442 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1444 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1445 ciphersuites which can be exploited in a denial of service attack.
1446 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1447 and detecting this bug and to Wolfgang Ettlinger
1448 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1452 *) Return an error when checking OCSP signatures when key is NULL.
1453 This fixes a DoS attack. (CVE-2013-0166)
1456 *) Make openssl verify return errors.
1457 [Chris Palmer <palmer@google.com> and Ben Laurie]
1459 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1460 the right response is stapled. Also change SSL_get_certificate()
1461 so it returns the certificate actually sent.
1462 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1463 [Rob Stradling <rob.stradling@comodo.com>]
1465 *) Fix possible deadlock when decoding public keys.
1468 *) Don't use TLS 1.0 record version number in initial client hello
1472 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1474 *) Sanity check record length before skipping explicit IV in TLS
1475 1.2, 1.1 and DTLS to fix DoS attack.
1477 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1478 fuzzing as a service testing platform.
1482 *) Initialise tkeylen properly when encrypting CMS messages.
1483 Thanks to Solar Designer of Openwall for reporting this issue.
1486 *) In FIPS mode don't try to use composite ciphers as they are not
1490 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1492 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1493 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1494 mean any application compiled against OpenSSL 1.0.0 headers setting
1495 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1496 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1497 0x10000000L Any application which was previously compiled against
1498 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1499 will need to be recompiled as a result. Letting be results in
1500 inability to disable specifically TLS 1.1 and in client context,
1501 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1504 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1505 disable just protocol X, but all protocols above X *if* there are
1506 protocols *below* X still enabled. In more practical terms it means
1507 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1508 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1509 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1513 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1515 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1516 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1517 in CRYPTO_realloc_clean.
1519 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1520 issue and to Adam Langley <agl@chromium.org> for fixing it.
1522 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1524 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1527 *) Workarounds for some broken servers that "hang" if a client hello
1528 record length exceeds 255 bytes.
1530 1. Do not use record version number > TLS 1.0 in initial client
1531 hello: some (but not all) hanging servers will now work.
1532 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1533 the number of ciphers sent in the client hello. This should be
1534 set to an even number, such as 50, for example by passing:
1535 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1536 Most broken servers should now work.
1537 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1538 TLS 1.2 client support entirely.
1541 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1544 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1546 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1547 STRING form instead of a DigestInfo.
1550 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1551 and the RSA_sign/RSA_verify functions. This was made more apparent when
1552 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1553 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1554 the correct format in RSA_verify so both forms transparently work.
1557 *) Some servers which support TLS 1.0 can choke if we initially indicate
1558 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1559 encrypted premaster secret. As a workaround use the maximum permitted
1560 client version in client hello, this should keep such servers happy
1561 and still work with previous versions of OpenSSL.
1564 *) Add support for TLS/DTLS heartbeats.
1565 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1567 *) Add support for SCTP.
1568 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1570 *) Improved PRNG seeding for VOS.
1571 [Paul Green <Paul.Green@stratus.com>]
1573 *) Extensive assembler packs updates, most notably:
1575 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1576 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1577 - x86_64: bit-sliced AES implementation;
1578 - ARM: NEON support, contemporary platforms optimizations;
1579 - s390x: z196 support;
1580 - *: GHASH and GF(2^m) multiplication implementations;
1584 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1585 (removal of unnecessary code)
1586 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1588 *) Add TLS key material exporter from RFC 5705.
1591 *) Add DTLS-SRTP negotiation from RFC 5764.
1594 *) Add Next Protocol Negotiation,
1595 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1596 disabled with a no-npn flag to config or Configure. Code donated
1598 [Adam Langley <agl@google.com> and Ben Laurie]
1600 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1601 NIST-P256, NIST-P521, with constant-time single point multiplication on
1602 typical inputs. Compiler support for the nonstandard type __uint128_t is
1603 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1604 Code made available under Apache License version 2.0.
1606 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1607 line to include this in your build of OpenSSL, and run "make depend" (or
1608 "make update"). This enables the following EC_METHODs:
1610 EC_GFp_nistp224_method()
1611 EC_GFp_nistp256_method()
1612 EC_GFp_nistp521_method()
1614 EC_GROUP_new_by_curve_name() will automatically use these (while
1615 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1617 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1619 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1620 all platforms. Move ssize_t definition from e_os.h to the public
1621 header file e_os2.h as it now appears in public header file cms.h
1624 *) New -sigopt option to the ca, req and x509 utilities. Additional
1625 signature parameters can be passed using this option and in
1629 *) Add RSA PSS signing function. This will generate and set the
1630 appropriate AlgorithmIdentifiers for PSS based on those in the
1631 corresponding EVP_MD_CTX structure. No application support yet.
1634 *) Support for companion algorithm specific ASN1 signing routines.
1635 New function ASN1_item_sign_ctx() signs a pre-initialised
1636 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1637 the appropriate parameters.
1640 *) Add new algorithm specific ASN1 verification initialisation function
1641 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1642 handling will be the same no matter what EVP_PKEY_METHOD is used.
1643 Add a PSS handler to support verification of PSS signatures: checked
1644 against a number of sample certificates.
1647 *) Add signature printing for PSS. Add PSS OIDs.
1648 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1650 *) Add algorithm specific signature printing. An individual ASN1 method
1651 can now print out signatures instead of the standard hex dump.
1653 More complex signatures (e.g. PSS) can print out more meaningful
1654 information. Include DSA version that prints out the signature
1658 *) Password based recipient info support for CMS library: implementing
1662 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1663 neatly separates the code into cipher and PBE sections and is required
1664 for some algorithms that split PBES2 into separate pieces (such as
1665 password based CMS).
1668 *) Session-handling fixes:
1669 - Fix handling of connections that are resuming with a session ID,
1670 but also support Session Tickets.
1671 - Fix a bug that suppressed issuing of a new ticket if the client
1672 presented a ticket with an expired session.
1673 - Try to set the ticket lifetime hint to something reasonable.
1674 - Make tickets shorter by excluding irrelevant information.
1675 - On the client side, don't ignore renewed tickets.
1676 [Adam Langley, Bodo Moeller (Google)]
1678 *) Fix PSK session representation.
1681 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1683 This work was sponsored by Intel.
1686 *) Add GCM support to TLS library. Some custom code is needed to split
1687 the IV between the fixed (from PRF) and explicit (from TLS record)
1688 portions. This adds all GCM ciphersuites supported by RFC5288 and
1689 RFC5289. Generalise some AES* cipherstrings to include GCM and
1690 add a special AESGCM string for GCM only.
1693 *) Expand range of ctrls for AES GCM. Permit setting invocation
1694 field on decrypt and retrieval of invocation field only on encrypt.
1697 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1698 As required by RFC5289 these ciphersuites cannot be used if for
1699 versions of TLS earlier than 1.2.
1702 *) For FIPS capable OpenSSL interpret a NULL default public key method
1703 as unset and return the appropriate default but do *not* set the default.
1704 This means we can return the appropriate method in applications that
1705 switch between FIPS and non-FIPS modes.
1708 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1709 ENGINE is used then we cannot handle that in the FIPS module so we
1710 keep original code iff non-FIPS operations are allowed.
1713 *) Add -attime option to openssl utilities.
1714 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1716 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1719 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1720 FIPS EC methods unconditionally for now.
1723 *) New build option no-ec2m to disable characteristic 2 code.
1726 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1727 all cases can be covered as some introduce binary incompatibilities.
1730 *) Redirect RSA operations to FIPS module including keygen,
1731 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1734 *) Add similar low level API blocking to ciphers.
1737 *) Low level digest APIs are not approved in FIPS mode: any attempt
1738 to use these will cause a fatal error. Applications that *really* want
1739 to use them can use the private_* version instead.
1742 *) Redirect cipher operations to FIPS module for FIPS builds.
1745 *) Redirect digest operations to FIPS module for FIPS builds.
1748 *) Update build system to add "fips" flag which will link in fipscanister.o
1749 for static and shared library builds embedding a signature if needed.
1752 *) Output TLS supported curves in preference order instead of numerical
1753 order. This is currently hardcoded for the highest order curves first.
1754 This should be configurable so applications can judge speed vs strength.
1757 *) Add TLS v1.2 server support for client authentication.
1760 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1764 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1765 FIPS modules versions.
1768 *) Add TLS v1.2 client side support for client authentication. Keep cache
1769 of handshake records longer as we don't know the hash algorithm to use
1770 until after the certificate request message is received.
1773 *) Initial TLS v1.2 client support. Add a default signature algorithms
1774 extension including all the algorithms we support. Parse new signature
1775 format in client key exchange. Relax some ECC signing restrictions for
1776 TLS v1.2 as indicated in RFC5246.
1779 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1780 to new signature format when needed using client digest preference.
1781 All server ciphersuites should now work correctly in TLS v1.2. No client
1782 support yet and no support for client certificates.
1785 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1786 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1787 ciphersuites. At present only RSA key exchange ciphersuites work with
1788 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1789 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1790 and version checking.
1793 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1794 with this defined it will not be affected by any changes to ssl internal
1795 structures. Add several utility functions to allow openssl application
1796 to work with OPENSSL_NO_SSL_INTERN defined.
1800 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1802 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1805 *) Permit abbreviated handshakes when renegotiating using the function
1806 SSL_renegotiate_abbreviated().
1807 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1809 *) Add call to ENGINE_register_all_complete() to
1810 ENGINE_load_builtin_engines(), so some implementations get used
1811 automatically instead of needing explicit application support.
1814 *) Add support for TLS key exporter as described in RFC5705.
1815 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1817 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1818 a few changes are required:
1820 Add SSL_OP_NO_TLSv1_1 flag.
1821 Add TLSv1_1 methods.
1822 Update version checking logic to handle version 1.1.
1823 Add explicit IV handling (ported from DTLS code).
1824 Add command line options to s_client/s_server.
1827 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1829 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1830 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1831 content decryption and always return the same error. Note: this attack
1832 needs on average 2^20 messages so it only affects automated senders. The
1833 old behaviour can be reenabled in the CMS code by setting the
1834 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1835 an MMA defence is not necessary.
1836 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1837 this issue. (CVE-2012-0884)
1840 *) Fix CVE-2011-4619: make sure we really are receiving a
1841 client hello before rejecting multiple SGC restarts. Thanks to
1842 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1845 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1847 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1848 Thanks to Antonio Martin, Enterprise Secure Access Research and
1849 Development, Cisco Systems, Inc. for discovering this bug and
1850 preparing a fix. (CVE-2012-0050)
1853 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1855 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1856 of the Vaudenay padding oracle attack on CBC mode encryption
1857 which enables an efficient plaintext recovery attack against
1858 the OpenSSL implementation of DTLS. Their attack exploits timing
1859 differences arising during decryption processing. A research
1860 paper describing this attack can be found at:
1861 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1862 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1863 Security Group at Royal Holloway, University of London
1864 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1865 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1866 for preparing the fix. (CVE-2011-4108)
1867 [Robin Seggelmann, Michael Tuexen]
1869 *) Clear bytes used for block padding of SSL 3.0 records.
1871 [Adam Langley (Google)]
1873 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1874 Kadianakis <desnacked@gmail.com> for discovering this issue and
1875 Adam Langley for preparing the fix. (CVE-2011-4619)
1876 [Adam Langley (Google)]
1878 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1879 [Andrey Kulikov <amdeich@gmail.com>]
1881 *) Prevent malformed RFC3779 data triggering an assertion failure.
1882 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1883 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1884 [Rob Austein <sra@hactrn.net>]
1886 *) Improved PRNG seeding for VOS.
1887 [Paul Green <Paul.Green@stratus.com>]
1889 *) Fix ssl_ciph.c set-up race.
1890 [Adam Langley (Google)]
1892 *) Fix spurious failures in ecdsatest.c.
1893 [Emilia Käsper (Google)]
1895 *) Fix the BIO_f_buffer() implementation (which was mixing different
1896 interpretations of the '..._len' fields).
1897 [Adam Langley (Google)]
1899 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1900 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1901 threads won't reuse the same blinding coefficients.
1903 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1904 lock to call BN_BLINDING_invert_ex, and avoids one use of
1905 BN_BLINDING_update for each BN_BLINDING structure (previously,
1906 the last update always remained unused).
1907 [Emilia Käsper (Google)]
1909 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1910 [Bob Buckholz (Google)]
1912 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1914 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1915 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1916 [Kaspar Brand <ossl@velox.ch>]
1918 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1919 for multi-threaded use of ECDH. (CVE-2011-3210)
1920 [Adam Langley (Google)]
1922 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1925 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1926 signature public key algorithm by using OID xref utilities instead.
1927 Before this you could only use some ECC ciphersuites with SHA1 only.
1930 *) Add protection against ECDSA timing attacks as mentioned in the paper
1931 by Billy Bob Brumley and Nicola Tuveri, see:
1933 http://eprint.iacr.org/2011/232.pdf
1935 [Billy Bob Brumley and Nicola Tuveri]
1937 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1939 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1940 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1942 *) Fix bug in string printing code: if *any* escaping is enabled we must
1943 escape the escape character (backslash) or the resulting string is
1947 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1949 *) Disable code workaround for ancient and obsolete Netscape browsers
1950 and servers: an attacker can use it in a ciphersuite downgrade attack.
1951 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1954 *) Fixed J-PAKE implementation error, originally discovered by
1955 Sebastien Martini, further info and confirmation from Stefan
1956 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1959 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1961 *) Fix extension code to avoid race conditions which can result in a buffer
1962 overrun vulnerability: resumed sessions must not be modified as they can
1963 be shared by multiple threads. CVE-2010-3864
1966 *) Fix WIN32 build system to correctly link an ENGINE directory into
1970 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1972 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1974 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1976 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1978 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1979 context. The operation can be customised via the ctrl mechanism in
1980 case ENGINEs want to include additional functionality.
1983 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1986 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1987 output hashes compatible with older versions of OpenSSL.
1988 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1990 *) Fix compression algorithm handling: if resuming a session use the
1991 compression algorithm of the resumed session instead of determining
1992 it from client hello again. Don't allow server to change algorithm.
1995 *) Add load_crls() function to apps tidying load_certs() too. Add option
1996 to verify utility to allow additional CRLs to be included.
1999 *) Update OCSP request code to permit adding custom headers to the request:
2000 some responders need this.
2003 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2005 [Julia Lawall <julia@diku.dk>]
2007 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2008 needlessly dereferenced structures, used obsolete functions and
2009 didn't handle all updated verify codes correctly.
2012 *) Disable MD2 in the default configuration.
2015 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2016 indicate the initial BIO being pushed or popped. This makes it possible
2017 to determine whether the BIO is the one explicitly called or as a result
2018 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2019 it handles reference counts correctly and doesn't zero out the I/O bio
2020 when it is not being explicitly popped. WARNING: applications which
2021 included workarounds for the old buggy behaviour will need to be modified
2022 or they could free up already freed BIOs.
2025 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2026 renaming to all platforms (within the 0.9.8 branch, this was
2027 done conditionally on Netware platforms to avoid a name clash).
2028 [Guenter <lists@gknw.net>]
2030 *) Add ECDHE and PSK support to DTLS.
2031 [Michael Tuexen <tuexen@fh-muenster.de>]
2033 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2037 *) Add "missing" function EVP_MD_flags() (without this the only way to
2038 retrieve a digest flags is by accessing the structure directly. Update
2039 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2040 or cipher is registered as in the "from" argument. Print out all
2041 registered digests in the dgst usage message instead of manually
2042 attempting to work them out.
2045 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2046 this allows the use of compression and extensions. Change default cipher
2047 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2048 by default unless an application cipher string requests it.
2051 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2052 key ids to find matching certificates and keys but some PKCS#12 files
2053 don't follow the (somewhat unwritten) rules and this strategy fails.
2054 Now just gather all certificates together and the first private key
2055 then look for the first certificate that matches the key.
2058 *) Support use of registered digest and cipher names for dgst and cipher
2059 commands instead of having to add each one as a special case. So now
2066 openssl dgst -sha256 foo
2068 and this works for ENGINE based algorithms too.
2072 *) Update Gost ENGINE to support parameter files.
2073 [Victor B. Wagner <vitus@cryptocom.ru>]
2075 *) Support GeneralizedTime in ca utility.
2076 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2078 *) Enhance the hash format used for certificate directory links. The new
2079 form uses the canonical encoding (meaning equivalent names will work
2080 even if they aren't identical) and uses SHA1 instead of MD5. This form
2081 is incompatible with the older format and as a result c_rehash should
2082 be used to rebuild symbolic links.
2085 *) Make PKCS#8 the default write format for private keys, replacing the
2086 traditional format. This form is standardised, more secure and doesn't
2087 include an implicit MD5 dependency.
2090 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2091 committed to OpenSSL should pass this lot as a minimum.
2094 *) Add session ticket override functionality for use by EAP-FAST.
2095 [Jouni Malinen <j@w1.fi>]
2097 *) Modify HMAC functions to return a value. Since these can be implemented
2098 in an ENGINE errors can occur.
2101 *) Type-checked OBJ_bsearch_ex.
2104 *) Type-checked OBJ_bsearch. Also some constification necessitated
2105 by type-checking. Still to come: TXT_DB, bsearch(?),
2106 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2110 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2111 seconds to a tm structure directly, instead of going through OS
2112 specific date routines. This avoids any issues with OS routines such
2113 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2114 and X509_time_adj_ex() to cover the extended range. The existing
2115 X509_time_adj() is still usable and will no longer have any date issues.
2118 *) Delta CRL support. New use deltas option which will attempt to locate
2119 and search any appropriate delta CRLs available.
2121 This work was sponsored by Google.
2124 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2125 code and add additional score elements. Validate alternate CRL paths
2126 as part of the CRL checking and indicate a new error "CRL path validation
2127 error" in this case. Applications wanting additional details can use
2128 the verify callback and check the new "parent" field. If this is not
2129 NULL CRL path validation is taking place. Existing applications wont
2130 see this because it requires extended CRL support which is off by
2133 This work was sponsored by Google.
2136 *) Support for freshest CRL extension.
2138 This work was sponsored by Google.
2141 *) Initial indirect CRL support. Currently only supported in the CRLs
2142 passed directly and not via lookup. Process certificate issuer
2143 CRL entry extension and lookup CRL entries by bother issuer name
2144 and serial number. Check and process CRL issuer entry in IDP extension.
2146 This work was sponsored by Google.
2149 *) Add support for distinct certificate and CRL paths. The CRL issuer
2150 certificate is validated separately in this case. Only enabled if
2151 an extended CRL support flag is set: this flag will enable additional
2152 CRL functionality in future.
2154 This work was sponsored by Google.
2157 *) Add support for policy mappings extension.
2159 This work was sponsored by Google.
2162 *) Fixes to pathlength constraint, self issued certificate handling,
2163 policy processing to align with RFC3280 and PKITS tests.
2165 This work was sponsored by Google.
2168 *) Support for name constraints certificate extension. DN, email, DNS
2169 and URI types are currently supported.
2171 This work was sponsored by Google.
2174 *) To cater for systems that provide a pointer-based thread ID rather
2175 than numeric, deprecate the current numeric thread ID mechanism and
2176 replace it with a structure and associated callback type. This
2177 mechanism allows a numeric "hash" to be extracted from a thread ID in
2178 either case, and on platforms where pointers are larger than 'long',
2179 mixing is done to help ensure the numeric 'hash' is usable even if it
2180 can't be guaranteed unique. The default mechanism is to use "&errno"
2181 as a pointer-based thread ID to distinguish between threads.
2183 Applications that want to provide their own thread IDs should now use
2184 CRYPTO_THREADID_set_callback() to register a callback that will call
2185 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2187 Note that ERR_remove_state() is now deprecated, because it is tied
2188 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2189 to free the current thread's error state should be replaced by
2190 ERR_remove_thread_state(NULL).
2192 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2193 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2194 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2195 application was previously providing a numeric thread callback that
2196 was inappropriate for distinguishing threads, then uniqueness might
2197 have been obtained with &errno that happened immediately in the
2198 intermediate development versions of OpenSSL; this is no longer the
2199 case, the numeric thread callback will now override the automatic use
2201 [Geoff Thorpe, with help from Bodo Moeller]
2203 *) Initial support for different CRL issuing certificates. This covers a
2204 simple case where the self issued certificates in the chain exist and
2205 the real CRL issuer is higher in the existing chain.
2207 This work was sponsored by Google.
2210 *) Removed effectively defunct crypto/store from the build.
2213 *) Revamp of STACK to provide stronger type-checking. Still to come:
2214 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2215 ASN1_STRING, CONF_VALUE.
2218 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2219 RAM on SSL connections. This option can save about 34k per idle SSL.
2222 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2223 STACK, TXT_DB, bsearch, qsort.
2226 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2227 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2228 support for data, signedData, compressedData, digestedData and
2229 encryptedData, envelopedData types included. Scripts to check against
2230 RFC4134 examples draft and interop and consistency checks of many
2231 content types and variants.
2234 *) Add options to enc utility to support use of zlib compression BIO.
2237 *) Extend mk1mf to support importing of options and assembly language
2238 files from Configure script, currently only included in VC-WIN32.
2239 The assembly language rules can now optionally generate the source
2240 files from the associated perl scripts.
2243 *) Implement remaining functionality needed to support GOST ciphersuites.
2244 Interop testing has been performed using CryptoPro implementations.
2245 [Victor B. Wagner <vitus@cryptocom.ru>]
2247 *) s390x assembler pack.
2250 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2254 *) Implement Opaque PRF Input TLS extension as specified in
2255 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2256 official specification yet and no extension type assignment by
2257 IANA exists, this extension (for now) will have to be explicitly
2258 enabled when building OpenSSL by providing the extension number
2259 to use. For example, specify an option
2261 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2263 to the "config" or "Configure" script to enable the extension,
2264 assuming extension number 0x9527 (which is a completely arbitrary
2265 and unofficial assignment based on the MD5 hash of the Internet
2266 Draft). Note that by doing so, you potentially lose
2267 interoperability with other TLS implementations since these might
2268 be using the same extension number for other purposes.
2270 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2271 opaque PRF input value to use in the handshake. This will create
2272 an interal copy of the length-'len' string at 'src', and will
2273 return non-zero for success.
2275 To get more control and flexibility, provide a callback function
2278 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2279 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2283 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2286 Callback function 'cb' will be called in handshakes, and is
2287 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2288 Argument 'arg' is for application purposes (the value as given to
2289 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2290 be provided to the callback function). The callback function
2291 has to return non-zero to report success: usually 1 to use opaque
2292 PRF input just if possible, or 2 to enforce use of the opaque PRF
2293 input. In the latter case, the library will abort the handshake
2294 if opaque PRF input is not successfully negotiated.
2296 Arguments 'peerinput' and 'len' given to the callback function
2297 will always be NULL and 0 in the case of a client. A server will
2298 see the client's opaque PRF input through these variables if
2299 available (NULL and 0 otherwise). Note that if the server
2300 provides an opaque PRF input, the length must be the same as the
2301 length of the client's opaque PRF input.
2303 Note that the callback function will only be called when creating
2304 a new session (session resumption can resume whatever was
2305 previously negotiated), and will not be called in SSL 2.0
2306 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2307 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2308 for applications that need to enforce opaque PRF input.
2312 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2315 [Victor B. Wagner <vitus@cryptocom.ru>]
2317 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2318 RFC4507bis. The encrypted ticket format is an encrypted encoded
2319 SSL_SESSION structure, that way new session features are automatically
2322 If a client application caches session in an SSL_SESSION structure
2323 support is transparent because tickets are now stored in the encoded
2326 The SSL_CTX structure automatically generates keys for ticket
2327 protection in servers so again support should be possible
2328 with no application modification.
2330 If a client or server wishes to disable RFC4507 support then the option
2331 SSL_OP_NO_TICKET can be set.
2333 Add a TLS extension debugging callback to allow the contents of any client
2334 or server extensions to be examined.
2336 This work was sponsored by Google.
2339 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2340 OpenSSL should now compile cleanly on gcc 4.2
2341 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2343 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2344 support including streaming MAC support: this is required for GOST
2345 ciphersuite support.
2346 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2348 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2349 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2350 to output in BER and PEM format.
2353 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2354 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2355 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2356 ENGINE support for HMAC keys which are unextractable. New -mac and
2357 -macopt options to dgst utility.
2360 *) New option -sigopt to dgst utility. Update dgst to use
2361 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2362 alternative signing paramaters such as X9.31 or PSS in the dgst
2366 *) Change ssl_cipher_apply_rule(), the internal function that does
2367 the work each time a ciphersuite string requests enabling
2368 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2369 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2370 the order of disabled ciphersuites such that those ciphersuites
2371 that most recently went from enabled to disabled not only stay
2372 in order with respect to each other, but also have higher priority
2373 than other disabled ciphersuites the next time ciphersuites are
2376 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2377 the same ciphersuites as with "HIGH" alone, but in a specific
2378 order where the PSK ciphersuites come first (since they are the
2379 most recently disabled ciphersuites when "HIGH" is parsed).
2381 Also, change ssl_create_cipher_list() (using this new
2382 funcionality) such that between otherwise identical
2383 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2387 *) Change ssl_create_cipher_list() so that it automatically
2388 arranges the ciphersuites in reasonable order before starting
2389 to process the rule string. Thus, the definition for "DEFAULT"
2390 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2391 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2392 This makes it much easier to arrive at a reasonable default order
2393 in applications for which anonymous ciphers are OK (meaning
2394 that you can't actually use DEFAULT).
2395 [Bodo Moeller; suggested by Victor Duchovni]
2397 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2398 processing) into multiple integers instead of setting
2399 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2400 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2401 (These masks as well as the individual bit definitions are hidden
2402 away into the non-exported interface ssl/ssl_locl.h, so this
2403 change to the definition of the SSL_CIPHER structure shouldn't
2404 affect applications.) This give us more bits for each of these
2405 categories, so there is no longer a need to coagulate AES128 and
2406 AES256 into a single algorithm bit, and to coagulate Camellia128
2407 and Camellia256 into a single algorithm bit, which has led to all
2410 Thus, among other things, the kludge introduced in 0.9.7m and
2411 0.9.8e for masking out AES256 independently of AES128 or masking
2412 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2414 With the change, we also introduce new ciphersuite aliases that
2415 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2419 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2420 Use the leftmost N bytes of the signature input if the input is
2421 larger than the prime q (with N being the size in bytes of q).
2424 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2425 it yet and it is largely untested.
2428 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2431 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2432 some compilers (gcc 4.2 and later) reject their use. Safestack is
2433 reimplemented. Update ASN1 to avoid use of legacy functions.
2436 *) Win32/64 targets are linked with Winsock2.
2439 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2440 to external functions. This can be used to increase CRL handling
2441 efficiency especially when CRLs are very large by (for example) storing
2442 the CRL revoked certificates in a database.
2445 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2446 new CRLs added to a directory can be used. New command line option
2447 -verify_return_error to s_client and s_server. This causes real errors
2448 to be returned by the verify callback instead of carrying on no matter
2449 what. This reflects the way a "real world" verify callback would behave.
2452 *) GOST engine, supporting several GOST algorithms and public key formats.
2453 Kindly donated by Cryptocom.
2456 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2457 partitioned by DP are handled but no indirect CRL or reason partitioning
2458 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2459 selected via a scoring technique which handles IDP and AKID in CRLs.
2462 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2463 will ultimately be used for all verify operations: this will remove the
2464 X509_STORE dependency on certificate verification and allow alternative
2465 lookup methods. X509_STORE based implementations of these two callbacks.
2468 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2469 Modify get_crl() to find a valid (unexpired) CRL if possible.
2472 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2473 this would be called X509_CRL_cmp() but that name is already used by
2474 a function that just compares CRL issuer names. Cache several CRL
2475 extensions in X509_CRL structure and cache CRLDP in X509.
2478 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2479 this maps equivalent X509_NAME structures into a consistent structure.
2480 Name comparison can then be performed rapidly using memcmp().
2483 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2487 *) Allow digests to supply their own micalg string for S/MIME type using
2488 the ctrl EVP_MD_CTRL_MICALG.
2491 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2492 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2493 ctrl. It can then customise the structure before and/or after signing
2497 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2498 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2499 to free up any added signature OIDs.
2502 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2503 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2504 digest and cipher tables. New options added to openssl utility:
2505 list-message-digest-algorithms and list-cipher-algorithms.
2508 *) Change the array representation of binary polynomials: the list
2509 of degrees of non-zero coefficients is now terminated with -1.
2510 Previously it was terminated with 0, which was also part of the
2511 value; thus, the array representation was not applicable to
2512 polynomials where t^0 has coefficient zero. This change makes
2513 the array representation useful in a more general context.
2516 *) Various modifications and fixes to SSL/TLS cipher string
2517 handling. For ECC, the code now distinguishes between fixed ECDH
2518 with RSA certificates on the one hand and with ECDSA certificates
2519 on the other hand, since these are separate ciphersuites. The
2520 unused code for Fortezza ciphersuites has been removed.
2522 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2523 (not "ECDHE"). For consistency with the code for DH
2524 certificates, use of ECDH certificates is now considered ECDH
2525 authentication, not RSA or ECDSA authentication (the latter is
2526 merely the CA's signing algorithm and not actively used in the
2529 The temporary ciphersuite alias "ECCdraft" is no longer
2530 available, and ECC ciphersuites are no longer excluded from "ALL"
2531 and "DEFAULT". The following aliases now exist for RFC 4492
2532 ciphersuites, most of these by analogy with the DH case:
2534 kECDHr - ECDH cert, signed with RSA
2535 kECDHe - ECDH cert, signed with ECDSA
2536 kECDH - ECDH cert (signed with either RSA or ECDSA)
2537 kEECDH - ephemeral ECDH
2538 ECDH - ECDH cert or ephemeral ECDH
2544 AECDH - anonymous ECDH
2545 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2549 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2550 Use correct micalg parameters depending on digest(s) in signed message.
2553 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2554 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2557 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2558 an engine to register a method. Add ENGINE lookups for methods and
2559 functional reference processing.
2562 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2563 EVP_{Sign,Verify}* which allow an application to customise the signature
2567 *) New -resign option to smime utility. This adds one or more signers
2568 to an existing PKCS#7 signedData structure. Also -md option to use an
2569 alternative message digest algorithm for signing.
2572 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2573 create PKCS7 structures containing multiple signers. Update smime
2574 application to support multiple signers.
2577 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2581 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2582 Reorganize PBE internals to lookup from a static table using NIDs,
2583 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2584 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2585 PRF which will be automatically used with PBES2.
2588 *) Replace the algorithm specific calls to generate keys in "req" with the
2592 *) Update PKCS#7 enveloped data routines to use new API. This is now
2593 supported by any public key method supporting the encrypt operation. A
2594 ctrl is added to allow the public key algorithm to examine or modify
2595 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2599 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2600 a default digest type to use. In most cases this will be SHA1 but some
2601 algorithms (such as GOST) need to specify an alternative digest. The
2602 return value indicates how strong the preference is 1 means optional and
2603 2 is mandatory (that is it is the only supported type). Modify
2604 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2605 use the default md. Update openssl utilities to use the default digest
2606 type for signing if it is not explicitly indicated.
2609 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2610 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2611 signing method from the key type. This effectively removes the link
2612 between digests and public key types.
2615 *) Add an OID cross reference table and utility functions. Its purpose is to
2616 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2617 rsaEncryption. This will allow some of the algorithm specific hackery
2618 needed to use the correct OID to be removed.
2621 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2622 structures for PKCS7_sign(). They are now set up by the relevant public
2626 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2629 *) Add support for key derivation (agreement) in the API, DH method and
2633 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2634 public and private key formats. As a side effect these add additional
2635 command line functionality not previously available: DSA signatures can be
2636 generated and verified using pkeyutl and DH key support and generation in
2641 [Oliver Tappe <zooey@hirschkaefer.de>]
2643 *) New make target "install_html_docs" installs HTML renditions of the
2645 [Oliver Tappe <zooey@hirschkaefer.de>]
2647 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2648 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2649 support key and parameter generation and add initial key generation
2650 functionality for RSA.
2653 *) Add functions for main EVP_PKEY_method operations. The undocumented
2654 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2655 EVP_PKEY_{encrypt,decrypt}_old.
2658 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2659 key API, doesn't do much yet.
2662 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2663 public key algorithms. New option to openssl utility:
2664 "list-public-key-algorithms" to print out info.
2667 *) Implement the Supported Elliptic Curves Extension for
2668 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2671 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2672 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2675 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2676 utilities such as rsa, dsa, dsaparam etc except they process any key
2680 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2681 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2682 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2686 *) Initial support for pluggable public key ASN1.
2687 De-spaghettify the public key ASN1 handling. Move public and private
2688 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2689 algorithm specific handling to a single module within the relevant
2690 algorithm directory. Add functions to allow (near) opaque processing
2691 of public and private key structures.
2694 *) Implement the Supported Point Formats Extension for
2695 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2698 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2699 for the psk identity [hint] and the psk callback functions to the
2700 SSL_SESSION, SSL and SSL_CTX structure.
2703 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2707 SSL_CTX_use_psk_identity_hint
2708 SSL_get_psk_identity_hint
2709 SSL_get_psk_identity
2710 SSL_use_psk_identity_hint
2712 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2714 *) Add RFC 3161 compliant time stamp request creation, response generation
2715 and response verification functionality.
2716 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2718 *) Add initial support for TLS extensions, specifically for the server_name
2719 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2720 have new members for a host name. The SSL data structure has an
2721 additional member SSL_CTX *initial_ctx so that new sessions can be
2722 stored in that context to allow for session resumption, even after the
2723 SSL has been switched to a new SSL_CTX in reaction to a client's
2724 server_name extension.
2726 New functions (subject to change):
2728 SSL_get_servername()
2729 SSL_get_servername_type()
2732 New CTRL codes and macros (subject to change):
2734 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2735 - SSL_CTX_set_tlsext_servername_callback()
2736 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2737 - SSL_CTX_set_tlsext_servername_arg()
2738 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2740 openssl s_client has a new '-servername ...' option.
2742 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2743 '-key2 ...', '-servername_fatal' (subject to change). This allows
2744 testing the HostName extension for a specific single host name ('-cert'
2745 and '-key' remain fallbacks for handshakes without HostName
2746 negotiation). If the unrecognized_name alert has to be sent, this by
2747 default is a warning; it becomes fatal with the '-servername_fatal'
2750 [Peter Sylvester, Remy Allais, Christophe Renou]
2752 *) Whirlpool hash implementation is added.
2755 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2756 bn(64,32). Because of instruction set limitations it doesn't have
2757 any negative impact on performance. This was done mostly in order
2758 to make it possible to share assembler modules, such as bn_mul_mont
2759 implementations, between 32- and 64-bit builds without hassle.
2762 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2763 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2767 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2768 dedicated Montgomery multiplication procedure, is introduced.
2769 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2770 "64-bit" performance on certain 32-bit targets.
2773 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2774 in SSL structures. New SSL ctrl to set maximum send fragment size.
2775 Save memory by seeting the I/O buffer sizes dynamically instead of
2776 using the maximum available value.
2779 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2780 in addition to the text details.
2783 *) Very, very preliminary EXPERIMENTAL support for printing of general
2784 ASN1 structures. This currently produces rather ugly output and doesn't
2785 handle several customised structures at all.
2788 *) Integrated support for PVK file format and some related formats such
2789 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2790 these in the 'rsa' and 'dsa' utilities.
2793 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2796 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2797 place for the (very old) "NETSCAPE" format certificates which are now
2798 handled using new ASN1 code equivalents.
2801 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2802 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2803 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2806 *) Modify CRL distribution points extension code to print out previously
2807 unsupported fields. Enhance extension setting code to allow setting of
2811 *) Add print and set support for Issuing Distribution Point CRL extension.
2814 *) Change 'Configure' script to enable Camellia by default.
2817 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2819 *) When rejecting SSL/TLS records due to an incorrect version number, never
2820 update s->server with a new major version number. As of
2821 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2822 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2823 the previous behavior could result in a read attempt at NULL when
2824 receiving specific incorrect SSL/TLS records once record payload
2825 protection is active. (CVE-2010-0740)
2826 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2828 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2829 could be crashed if the relevant tables were not present (e.g. chrooted).
2830 [Tomas Hoger <thoger@redhat.com>]
2832 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2834 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2835 [Martin Olsson, Neel Mehta]
2837 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2838 accommodate for stack sorting, always a write lock!).
2841 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2842 excessive delays in the RAND_poll(): over a minute. As a workaround
2843 include a time check in the inner Heap32Next loop too.
2846 *) The code that handled flushing of data in SSL/TLS originally used the
2847 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2848 the problem outlined in PR#1949. The fix suggested there however can
2849 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2850 of Apache). So instead simplify the code to flush unconditionally.
2851 This should be fine since flushing with no data to flush is a no op.
2854 *) Handle TLS versions 2.0 and later properly and correctly use the
2855 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2856 off ancient servers have a habit of sticking around for a while...
2859 *) Modify compression code so it frees up structures without using the
2860 ex_data callbacks. This works around a problem where some applications
2861 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2862 restarting) then use compression (e.g. SSL with compression) later.
2863 This results in significant per-connection memory leaks and
2864 has caused some security issues including CVE-2008-1678 and
2868 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2869 change when encrypting or decrypting.
2872 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2873 connect and renegotiate with servers which do not support RI.
2874 Until RI is more widely deployed this option is enabled by default.
2877 *) Add "missing" ssl ctrls to clear options and mode.
2880 *) If client attempts to renegotiate and doesn't support RI respond with
2881 a no_renegotiation alert as required by RFC5746. Some renegotiating
2882 TLS clients will continue a connection gracefully when they receive
2883 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2884 waiting for a server hello which it will never receive. Now we treat a
2885 received no_renegotiation alert as a fatal error. This is because
2886 applications requesting a renegotiation might well expect it to succeed
2887 and would have no code in place to handle the server denying it so the
2888 only safe thing to do is to terminate the connection.
2891 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2892 peer supports secure renegotiation and 0 otherwise. Print out peer
2893 renegotiation support in s_client/s_server.
2896 *) Replace the highly broken and deprecated SPKAC certification method with
2897 the updated NID creation version. This should correctly handle UTF8.
2900 *) Implement RFC5746. Re-enable renegotiation but require the extension
2901 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2902 turns out to be a bad idea. It has been replaced by
2903 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2904 SSL_CTX_set_options(). This is really not recommended unless you
2905 know what you are doing.
2906 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2908 *) Fixes to stateless session resumption handling. Use initial_ctx when
2909 issuing and attempting to decrypt tickets in case it has changed during
2910 servername handling. Use a non-zero length session ID when attempting
2911 stateless session resumption: this makes it possible to determine if
2912 a resumption has occurred immediately after receiving server hello
2913 (several places in OpenSSL subtly assume this) instead of later in
2917 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2918 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2919 fixes for a few places where the return code is not checked
2921 [Julia Lawall <julia@diku.dk>]
2923 *) Add --strict-warnings option to Configure script to include devteam
2924 warnings in other configurations.
2927 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2928 makes it possible to install openssl libraries in locations which
2929 have names other than "lib", for example "/usr/lib64" which some
2931 [Steve Henson, based on patch from Jeremy Utley]
2933 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2934 X690 8.9.12 and can produce some misleading textual output of OIDs.
2935 [Steve Henson, reported by Dan Kaminsky]
2937 *) Delete MD2 from algorithm tables. This follows the recommendation in
2938 several standards that it is not used in new applications due to
2939 several cryptographic weaknesses. For binary compatibility reasons
2940 the MD2 API is still compiled in by default.
2943 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2947 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2948 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2950 [Guenter <lists@gknw.net>]
2952 *) Fix the server certificate chain building code to use X509_verify_cert(),
2953 it used to have an ad-hoc builder which was unable to cope with anything
2954 other than a simple chain.
2955 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2957 *) Don't check self signed certificate signatures in X509_verify_cert()
2958 by default (a flag can override this): it just wastes time without
2959 adding any security. As a useful side effect self signed root CAs
2960 with non-FIPS digests are now usable in FIPS mode.
2963 *) In dtls1_process_out_of_seq_message() the check if the current message
2964 is already buffered was missing. For every new message was memory
2965 allocated, allowing an attacker to perform an denial of service attack
2966 with sending out of seq handshake messages until there is no memory
2967 left. Additionally every future messege was buffered, even if the
2968 sequence number made no sense and would be part of another handshake.
2969 So only messages with sequence numbers less than 10 in advance will be
2970 buffered. (CVE-2009-1378)
2971 [Robin Seggelmann, discovered by Daniel Mentz]
2973 *) Records are buffered if they arrive with a future epoch to be
2974 processed after finishing the corresponding handshake. There is
2975 currently no limitation to this buffer allowing an attacker to perform
2976 a DOS attack with sending records with future epochs until there is no
2977 memory left. This patch adds the pqueue_size() function to determine
2978 the size of a buffer and limits the record buffer to 100 entries.
2980 [Robin Seggelmann, discovered by Daniel Mentz]
2982 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2983 parent structure is freed. (CVE-2009-1379)
2986 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2987 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2990 [Ilya O. <vrghost@gmail.com>]
2992 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2994 *) Disable renegotiation completely - this fixes a severe security
2995 problem (CVE-2009-3555) at the cost of breaking all
2996 renegotiation. Renegotiation can be re-enabled by setting
2997 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2998 run-time. This is really not recommended unless you know what
3002 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3004 *) Don't set val to NULL when freeing up structures, it is freed up by
3005 underlying code. If sizeof(void *) > sizeof(long) this can result in
3006 zeroing past the valid field. (CVE-2009-0789)
3007 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3009 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3010 checked correctly. This would allow some invalid signed attributes to
3011 appear to verify correctly. (CVE-2009-0591)
3012 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3014 *) Reject UniversalString and BMPString types with invalid lengths. This
3015 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3016 a legal length. (CVE-2009-0590)
3019 *) Set S/MIME signing as the default purpose rather than setting it
3020 unconditionally. This allows applications to override it at the store
3024 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3025 to handle some structures.
3028 *) Improve efficiency of mem_gets: don't search whole buffer each time
3030 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3032 *) New -hex option for openssl rand.
3035 *) Print out UTF8String and NumericString when parsing ASN1.
3038 *) Support NumericString type for name components.
3041 *) Allow CC in the environment to override the automatically chosen
3042 compiler. Note that nothing is done to ensure flags work with the
3046 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3048 *) Properly check EVP_VerifyFinal() and similar return values
3050 [Ben Laurie, Bodo Moeller, Google Security Team]
3052 *) Enable TLS extensions by default.
3055 *) Allow the CHIL engine to be loaded, whether the application is
3056 multithreaded or not. (This does not release the developer from the
3057 obligation to set up the dynamic locking callbacks.)
3058 [Sander Temme <sander@temme.net>]
3060 *) Use correct exit code if there is an error in dgst command.
3061 [Steve Henson; problem pointed out by Roland Dirlewanger]
3063 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3064 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3067 *) Add experimental JPAKE support, including demo authentication in
3068 s_client and s_server.
3071 *) Set the comparison function in v3_addr_canonize().
3072 [Rob Austein <sra@hactrn.net>]
3074 *) Add support for XMPP STARTTLS in s_client.
3075 [Philip Paeps <philip@freebsd.org>]
3077 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3078 to ensure that even with this option, only ciphersuites in the
3079 server's preference list will be accepted. (Note that the option
3080 applies only when resuming a session, so the earlier behavior was
3081 just about the algorithm choice for symmetric cryptography.)
3084 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3086 *) Fix NULL pointer dereference if a DTLS server received
3087 ChangeCipherSpec as first record (CVE-2009-1386).
3090 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3091 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3094 *) The fix in 0.9.8c that supposedly got rid of unsafe
3095 double-checked locking was incomplete for RSA blinding,
3096 addressing just one layer of what turns out to have been
3097 doubly unsafe triple-checked locking.
3099 So now fix this for real by retiring the MONT_HELPER macro
3100 in crypto/rsa/rsa_eay.c.
3102 [Bodo Moeller; problem pointed out by Marius Schilder]
3104 *) Various precautionary measures:
3106 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3108 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3109 (NB: This would require knowledge of the secret session ticket key
3110 to exploit, in which case you'd be SOL either way.)
3112 - Change bn_nist.c so that it will properly handle input BIGNUMs
3113 outside the expected range.
3115 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3118 [Neel Mehta, Bodo Moeller]
3120 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3121 the load fails. Useful for distros.
3122 [Ben Laurie and the FreeBSD team]
3124 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3127 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3130 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3132 This work was sponsored by Logica.
3135 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3136 keystores. Support for SSL/TLS client authentication too.
3137 Not compiled unless enable-capieng specified to Configure.
3139 This work was sponsored by Logica.
3142 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3143 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3144 attribute creation routines such as certifcate requests and PKCS#12
3148 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3150 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3151 handshake which could lead to a cilent crash as found using the
3152 Codenomicon TLS test suite (CVE-2008-1672)
3153 [Steve Henson, Mark Cox]
3155 *) Fix double free in TLS server name extensions which could lead to
3156 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3159 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3161 Clear the error queue to ensure that error entries left from
3162 older function calls do not interfere with the correct operation.
3163 [Lutz Jaenicke, Erik de Castro Lopo]
3165 *) Remove root CA certificates of commercial CAs:
3167 The OpenSSL project does not recommend any specific CA and does not
3168 have any policy with respect to including or excluding any CA.
3169 Therefore it does not make any sense to ship an arbitrary selection
3170 of root CA certificates with the OpenSSL software.
3173 *) RSA OAEP patches to fix two separate invalid memory reads.
3174 The first one involves inputs when 'lzero' is greater than
3175 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3176 before the beginning of from). The second one involves inputs where
3177 the 'db' section contains nothing but zeroes (there is a one-byte
3178 invalid read after the end of 'db').
3179 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3181 *) Partial backport from 0.9.9-dev:
3183 Introduce bn_mul_mont (dedicated Montgomery multiplication
3184 procedure) as a candidate for BIGNUM assembler implementation.
3185 While 0.9.9-dev uses assembler for various architectures, only
3186 x86_64 is available by default here in the 0.9.8 branch, and
3187 32-bit x86 is available through a compile-time setting.
3189 To try the 32-bit x86 assembler implementation, use Configure
3190 option "enable-montasm" (which exists only for this backport).
3192 As "enable-montasm" for 32-bit x86 disclaims code stability
3193 anyway, in this constellation we activate additional code
3194 backported from 0.9.9-dev for further performance improvements,
3195 namely BN_from_montgomery_word. (To enable this otherwise,
3196 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3198 [Andy Polyakov (backport partially by Bodo Moeller)]
3200 *) Add TLS session ticket callback. This allows an application to set
3201 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3202 values. This is useful for key rollover for example where several key
3203 sets may exist with different names.
3206 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3207 This was broken until now in 0.9.8 releases, such that the only way
3208 a registered ENGINE could be used (assuming it initialises
3209 successfully on the host) was to explicitly set it as the default
3210 for the relevant algorithms. This is in contradiction with 0.9.7
3211 behaviour and the documentation. With this fix, when an ENGINE is
3212 registered into a given algorithm's table of implementations, the
3213 'uptodate' flag is reset so that auto-discovery will be used next
3214 time a new context for that algorithm attempts to select an
3216 [Ian Lister (tweaked by Geoff Thorpe)]
3218 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3219 implemention in the following ways:
3221 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3224 Lack of BER streaming support means one pass streaming processing is
3225 only supported if data is detached: setting the streaming flag is
3226 ignored for embedded content.
3228 CMS support is disabled by default and must be explicitly enabled
3229 with the enable-cms configuration option.
3232 *) Update the GMP engine glue to do direct copies between BIGNUM and
3233 mpz_t when openssl and GMP use the same limb size. Otherwise the
3234 existing "conversion via a text string export" trick is still used.
3235 [Paul Sheer <paulsheer@gmail.com>]
3237 *) Zlib compression BIO. This is a filter BIO which compressed and
3238 uncompresses any data passed through it.
3241 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3242 RFC3394 compatible AES key wrapping.
3245 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3246 sets string data without copying. X509_ALGOR_set0() and
3247 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3248 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3249 from an X509_ATTRIBUTE structure optionally checking it occurs only
3250 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3254 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3255 to get the expected BN_FLG_CONSTTIME behavior.
3256 [Bodo Moeller (Google)]
3260 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3261 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3262 - added some more tests to do_tests.pl
3263 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3264 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3265 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3266 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3267 - various changes to netware.pl to enable gcc-cross builds on Win32
3269 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3270 - various changes to fix missing prototype warnings
3271 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3272 - added AES, WHIRLPOOL and CPUID assembler code to build files
3273 - added missing AES assembler make rules to mk1mf.pl
3274 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3275 [Guenter Knauf <eflash@gmx.net>]
3277 *) Implement certificate status request TLS extension defined in RFC3546.
3278 A client can set the appropriate parameters and receive the encoded
3279 OCSP response via a callback. A server can query the supplied parameters
3280 and set the encoded OCSP response in the callback. Add simplified examples
3281 to s_client and s_server.
3284 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3286 *) Fix various bugs:
3287 + Binary incompatibility of ssl_ctx_st structure
3288 + DTLS interoperation with non-compliant servers
3289 + Don't call get_session_cb() without proposed session
3290 + Fix ia64 assembler code
3291 [Andy Polyakov, Steve Henson]
3293 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3295 *) DTLS Handshake overhaul. There were longstanding issues with
3296 OpenSSL DTLS implementation, which were making it impossible for
3297 RFC 4347 compliant client to communicate with OpenSSL server.
3298 Unfortunately just fixing these incompatibilities would "cut off"
3299 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3300 server keeps tolerating non RFC compliant syntax. The opposite is
3301 not true, 0.9.8f client can not communicate with earlier server.
3302 This update even addresses CVE-2007-4995.
3305 *) Changes to avoid need for function casts in OpenSSL: some compilers
3306 (gcc 4.2 and later) reject their use.
3307 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3310 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3311 RFC4507bis. The encrypted ticket format is an encrypted encoded
3312 SSL_SESSION structure, that way new session features are automatically
3315 If a client application caches session in an SSL_SESSION structure
3316 support is transparent because tickets are now stored in the encoded
3319 The SSL_CTX structure automatically generates keys for ticket
3320 protection in servers so again support should be possible
3321 with no application modification.
3323 If a client or server wishes to disable RFC4507 support then the option
3324 SSL_OP_NO_TICKET can be set.
3326 Add a TLS extension debugging callback to allow the contents of any client
3327 or server extensions to be examined.
3329 This work was sponsored by Google.
3332 *) Add initial support for TLS extensions, specifically for the server_name
3333 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3334 have new members for a host name. The SSL data structure has an
3335 additional member SSL_CTX *initial_ctx so that new sessions can be
3336 stored in that context to allow for session resumption, even after the
3337 SSL has been switched to a new SSL_CTX in reaction to a client's
3338 server_name extension.
3340 New functions (subject to change):
3342 SSL_get_servername()
3343 SSL_get_servername_type()
3346 New CTRL codes and macros (subject to change):
3348 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3349 - SSL_CTX_set_tlsext_servername_callback()
3350 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3351 - SSL_CTX_set_tlsext_servername_arg()
3352 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3354 openssl s_client has a new '-servername ...' option.
3356 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3357 '-key2 ...', '-servername_fatal' (subject to change). This allows
3358 testing the HostName extension for a specific single host name ('-cert'
3359 and '-key' remain fallbacks for handshakes without HostName
3360 negotiation). If the unrecognized_name alert has to be sent, this by
3361 default is a warning; it becomes fatal with the '-servername_fatal'
3364 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3366 *) Add AES and SSE2 assembly language support to VC++ build.
3369 *) Mitigate attack on final subtraction in Montgomery reduction.
3372 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3373 (which previously caused an internal error).
3376 *) Squeeze another 10% out of IGE mode when in != out.
3379 *) AES IGE mode speedup.
3380 [Dean Gaudet (Google)]
3382 *) Add the Korean symmetric 128-bit cipher SEED (see
3383 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3384 add SEED ciphersuites from RFC 4162:
3386 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3387 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3388 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3389 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3391 To minimize changes between patchlevels in the OpenSSL 0.9.8
3392 series, SEED remains excluded from compilation unless OpenSSL
3393 is configured with 'enable-seed'.
3394 [KISA, Bodo Moeller]
3396 *) Mitigate branch prediction attacks, which can be practical if a
3397 single processor is shared, allowing a spy process to extract
3398 information. For detailed background information, see
3399 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3400 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3401 and Necessary Software Countermeasures"). The core of the change
3402 are new versions BN_div_no_branch() and
3403 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3404 respectively, which are slower, but avoid the security-relevant
3405 conditional branches. These are automatically called by BN_div()
3406 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3407 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3408 remove a conditional branch.
3410 BN_FLG_CONSTTIME is the new name for the previous
3411 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3412 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3413 in the exponent causes BN_mod_exp_mont() to use the alternative
3414 implementation in BN_mod_exp_mont_consttime().) The old name
3415 remains as a deprecated alias.
3417 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3418 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3419 constant-time implementations for more than just exponentiation.
3420 Here too the old name is kept as a deprecated alias.
3422 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3423 the BN_BLINDING structure gets an independent copy of the
3424 modulus. This means that the previous "BIGNUM *m" argument to
3425 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3426 essentially becomes "const BIGNUM *m", although we can't actually
3427 change this in the header file before 0.9.9. It allows
3428 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3429 enable BN_FLG_CONSTTIME.
3431 [Matthew D Wood (Intel Corp)]
3433 *) In the SSL/TLS server implementation, be strict about session ID
3434 context matching (which matters if an application uses a single
3435 external cache for different purposes). Previously,
3436 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3437 set. This did ensure strict client verification, but meant that,
3438 with applications using a single external cache for quite
3439 different requirements, clients could circumvent ciphersuite
3440 restrictions for a given session ID context by starting a session
3441 in a different context.
3444 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3445 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3446 authentication-only ciphersuites.
3449 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3450 not complete and could lead to a possible single byte overflow
3451 (CVE-2007-5135) [Ben Laurie]
3453 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3455 *) Since AES128 and AES256 (and similarly Camellia128 and
3456 Camellia256) share a single mask bit in the logic of
3457 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3458 kludge to work properly if AES128 is available and AES256 isn't
3459 (or if Camellia128 is available and Camellia256 isn't).
3462 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3463 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3464 When a point or a seed is encoded in a BIT STRING, we need to
3465 prevent the removal of trailing zero bits to get the proper DER
3466 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3467 of a NamedBitList, for which trailing 0 bits need to be removed.)
3470 *) Have SSL/TLS server implementation tolerate "mismatched" record
3471 protocol version while receiving ClientHello even if the
3472 ClientHello is fragmented. (The server can't insist on the
3473 particular protocol version it has chosen before the ServerHello
3474 message has informed the client about his choice.)
3477 *) Add RFC 3779 support.
3478 [Rob Austein for ARIN, Ben Laurie]
3480 *) Load error codes if they are not already present instead of using a
3481 static variable. This allows them to be cleanly unloaded and reloaded.
3482 Improve header file function name parsing.
3485 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3486 or CAPABILITY handshake as required by RFCs.
3489 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3491 *) Introduce limits to prevent malicious keys being able to
3492 cause a denial of service. (CVE-2006-2940)
3493 [Steve Henson, Bodo Moeller]
3495 *) Fix ASN.1 parsing of certain invalid structures that can result
3496 in a denial of service. (CVE-2006-2937) [Steve Henson]
3498 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3499 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3501 *) Fix SSL client code which could crash if connecting to a
3502 malicious SSLv2 server. (CVE-2006-4343)
3503 [Tavis Ormandy and Will Drewry, Google Security Team]
3505 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3506 match only those. Before that, "AES256-SHA" would be interpreted
3507 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3508 the same strength classification in 0.9.7h) as we currently only
3509 have a single AES bit in the ciphersuite description bitmap.
3510 That change, however, also applied to ciphersuite strings such as
3511 "RC4-MD5" that intentionally matched multiple ciphersuites --
3512 namely, SSL 2.0 ciphersuites in addition to the more common ones
3513 from SSL 3.0/TLS 1.0.
3515 So we change the selection algorithm again: Naming an explicit
3516 ciphersuite selects this one ciphersuite, and any other similar
3517 ciphersuite (same bitmap) from *other* protocol versions.
3518 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3519 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3521 Since SSL 2.0 does not have any ciphersuites for which the
3522 128/256 bit distinction would be relevant, this works for now.
3523 The proper fix will be to use different bits for AES128 and
3524 AES256, which would have avoided the problems from the beginning;
3525 however, bits are scarce, so we can only do this in a new release
3526 (not just a patchlevel) when we can change the SSL_CIPHER
3527 definition to split the single 'unsigned long mask' bitmap into
3528 multiple values to extend the available space.
3532 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3534 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3535 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3537 *) Add AES IGE and biIGE modes.
3540 *) Change the Unix randomness entropy gathering to use poll() when
3541 possible instead of select(), since the latter has some
3542 undesirable limitations.
3543 [Darryl Miles via Richard Levitte and Bodo Moeller]
3545 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3546 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3547 cannot be implicitly activated as part of, e.g., the "AES" alias.
3548 However, please upgrade to OpenSSL 0.9.9[-dev] for
3549 non-experimental use of the ECC ciphersuites to get TLS extension
3550 support, which is required for curve and point format negotiation
3551 to avoid potential handshake problems.
3554 *) Disable rogue ciphersuites:
3556 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3557 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3558 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3560 The latter two were purportedly from
3561 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3564 Also deactivate the remaining ciphersuites from
3565 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3566 unofficial, and the ID has long expired.
3569 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3570 dual-core machines) and other potential thread-safety issues.
3573 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3574 versions), which is now available for royalty-free use
3575 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3576 Also, add Camellia TLS ciphersuites from RFC 4132.
3578 To minimize changes between patchlevels in the OpenSSL 0.9.8
3579 series, Camellia remains excluded from compilation unless OpenSSL
3580 is configured with 'enable-camellia'.
3583 *) Disable the padding bug check when compression is in use. The padding
3584 bug check assumes the first packet is of even length, this is not
3585 necessarily true if compresssion is enabled and can result in false
3586 positives causing handshake failure. The actual bug test is ancient
3587 code so it is hoped that implementations will either have fixed it by
3588 now or any which still have the bug do not support compression.
3591 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3593 *) When applying a cipher rule check to see if string match is an explicit
3594 cipher suite and only match that one cipher suite if it is.
3597 *) Link in manifests for VC++ if needed.
3598 [Austin Ziegler <halostatue@gmail.com>]
3600 *) Update support for ECC-based TLS ciphersuites according to
3601 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3602 TLS extensions, which are supported starting with the 0.9.9
3603 branch, not in the OpenSSL 0.9.8 branch).
3606 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3607 opaque EVP_CIPHER_CTX handling.
3610 *) Fixes and enhancements to zlib compression code. We now only use
3611 "zlib1.dll" and use the default __cdecl calling convention on Win32
3612 to conform with the standards mentioned here:
3613 http://www.zlib.net/DLL_FAQ.txt
3614 Static zlib linking now works on Windows and the new --with-zlib-include
3615 --with-zlib-lib options to Configure can be used to supply the location
3616 of the headers and library. Gracefully handle case where zlib library
3620 *) Several fixes and enhancements to the OID generation code. The old code
3621 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3622 handle numbers larger than ULONG_MAX, truncated printing and had a
3623 non standard OBJ_obj2txt() behaviour.
3626 *) Add support for building of engines under engine/ as shared libraries
3627 under VC++ build system.
3630 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3631 Hopefully, we will not see any false combination of paths any more.
3634 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3636 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3637 (part of SSL_OP_ALL). This option used to disable the
3638 countermeasure against man-in-the-middle protocol-version
3639 rollback in the SSL 2.0 server implementation, which is a bad
3640 idea. (CVE-2005-2969)
3642 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3643 for Information Security, National Institute of Advanced Industrial
3644 Science and Technology [AIST], Japan)]
3646 *) Add two function to clear and return the verify parameter flags.
3649 *) Keep cipherlists sorted in the source instead of sorting them at
3650 runtime, thus removing the need for a lock.
3653 *) Avoid some small subgroup attacks in Diffie-Hellman.
3654 [Nick Mathewson and Ben Laurie]
3656 *) Add functions for well-known primes.
3659 *) Extended Windows CE support.
3660 [Satoshi Nakamura and Andy Polyakov]
3662 *) Initialize SSL_METHOD structures at compile time instead of during
3663 runtime, thus removing the need for a lock.
3666 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3667 attempting to decrypt each encrypted key in turn. Add support to
3671 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3673 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3676 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3679 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3680 key into the same file any more.
3683 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3686 *) Add -utf8 command line and config file option to 'ca'.
3687 [Stefan <stf@udoma.org]
3689 *) Removed the macro des_crypt(), as it seems to conflict with some
3690 libraries. Use DES_crypt().
3693 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3694 involves renaming the source and generated shared-libs for
3695 both. The engines will accept the corrected or legacy ids
3696 ('ncipher' and '4758_cca' respectively) when binding. NB,
3697 this only applies when building 'shared'.
3698 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3700 *) Add attribute functions to EVP_PKEY structure. Modify
3701 PKCS12_create() to recognize a CSP name attribute and
3702 use it. Make -CSP option work again in pkcs12 utility.
3705 *) Add new functionality to the bn blinding code:
3706 - automatic re-creation of the BN_BLINDING parameters after
3707 a fixed number of uses (currently 32)
3708 - add new function for parameter creation
3709 - introduce flags to control the update behaviour of the
3710 BN_BLINDING parameters
3711 - hide BN_BLINDING structure
3712 Add a second BN_BLINDING slot to the RSA structure to improve
3713 performance when a single RSA object is shared among several
3717 *) Add support for DTLS.
3718 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3720 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3721 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3724 *) Remove buggy and incomplete DH cert support from
3725 ssl/ssl_rsa.c and ssl/s3_both.c
3728 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3729 the apps/openssl applications.
3732 *) Compile clean with "-Wall -Wmissing-prototypes
3733 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3734 DEBUG_SAFESTACK must also be set.
3737 *) Change ./Configure so that certain algorithms can be disabled by default.
3738 The new counterpiece to "no-xxx" is "enable-xxx".
3740 The patented RC5 and MDC2 algorithms will now be disabled unless
3741 "enable-rc5" and "enable-mdc2", respectively, are specified.
3743 (IDEA remains enabled despite being patented. This is because IDEA
3744 is frequently required for interoperability, and there is no license
3745 fee for non-commercial use. As before, "no-idea" can be used to
3746 avoid this algorithm.)
3750 *) Add processing of proxy certificates (see RFC 3820). This work was
3751 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3752 EGEE (Enabling Grids for E-science in Europe).
3755 *) RC4 performance overhaul on modern architectures/implementations, such
3756 as Intel P4, IA-64 and AMD64.
3759 *) New utility extract-section.pl. This can be used specify an alternative
3760 section number in a pod file instead of having to treat each file as
3761 a separate case in Makefile. This can be done by adding two lines to the
3764 =for comment openssl_section:XXX
3766 The blank line is mandatory.
3770 *) New arguments -certform, -keyform and -pass for s_client and s_server
3771 to allow alternative format key and certificate files and passphrase
3775 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3776 update associated structures and add various utility functions.
3778 Add new policy related verify parameters, include policy checking in
3779 standard verify code. Enhance 'smime' application with extra parameters
3780 to support policy checking and print out.
3783 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3784 Nehemiah processors. These extensions support AES encryption in hardware
3785 as well as RNG (though RNG support is currently disabled).
3786 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3788 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3791 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3792 [Andy Polyakov and a number of other people]
3794 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3795 implementation contributed by IBM.
3796 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3798 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3799 exponent rather than 'unsigned long'. There is a corresponding change to
3800 the new 'rsa_keygen' element of the RSA_METHOD structure.
3801 [Jelte Jansen, Geoff Thorpe]
3803 *) Functionality for creating the initial serial number file is now
3804 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3806 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3807 number file to 1, which is bound to cause problems. To avoid
3808 the problems while respecting compatibility between different 0.9.7
3809 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3810 CA.pl for serial number initialization. With the new release 0.9.8,
3811 we can fix the problem directly in the 'ca' utility.)
3814 *) Reduced header interdepencies by declaring more opaque objects in
3815 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3816 give fewer recursive includes, which could break lazy source code - so
3817 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3818 developers should define this symbol when building and using openssl to
3819 ensure they track the recommended behaviour, interfaces, [etc], but
3820 backwards-compatible behaviour prevails when this isn't defined.
3823 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3826 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3827 This will generate a random key of the appropriate length based on the
3828 cipher context. The EVP_CIPHER can provide its own random key generation
3829 routine to support keys of a specific form. This is used in the des and
3830 3des routines to generate a key of the correct parity. Update S/MIME
3831 code to use new functions and hence generate correct parity DES keys.
3832 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3833 valid (weak or incorrect parity).
3836 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3837 as looking them up. This is useful when the verified structure may contain
3838 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3839 present unless the new PKCS7_NO_CRL flag is asserted.
3842 *) Extend ASN1 oid configuration module. It now additionally accepts the
3845 shortName = some long name, 1.2.3.4
3848 *) Reimplemented the BN_CTX implementation. There is now no more static
3849 limitation on the number of variables it can handle nor the depth of the
3850 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3851 information can now expand as required, and rather than having a single
3852 static array of bignums, BN_CTX now uses a linked-list of such arrays
3853 allowing it to expand on demand whilst maintaining the usefulness of
3854 BN_CTX's "bundling".
3857 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3858 to allow all RSA operations to function using a single BN_CTX.
3861 *) Preliminary support for certificate policy evaluation and checking. This
3862 is initially intended to pass the tests outlined in "Conformance Testing
3863 of Relying Party Client Certificate Path Processing Logic" v1.07.
3866 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3867 remained unused and not that useful. A variety of other little bignum
3868 tweaks and fixes have also been made continuing on from the audit (see
3872 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3873 associated ASN1, EVP and SSL functions and old ASN1 macros.
3876 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3877 and this should never fail. So the return value from the use of
3878 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3879 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3882 *) BN_CTX_get() should return zero-valued bignums, providing the same
3883 initialised value as BN_new().
3884 [Geoff Thorpe, suggested by Ulf Möller]
3886 *) Support for inhibitAnyPolicy certificate extension.
3889 *) An audit of the BIGNUM code is underway, for which debugging code is
3890 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3891 is considered valid when processing BIGNUMs, and causes execution to
3892 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3893 further steps are taken to deliberately pollute unused data in BIGNUM
3894 structures to try and expose faulty code further on. For now, openssl will
3895 (in its default mode of operation) continue to tolerate the inconsistent
3896 forms that it has tolerated in the past, but authors and packagers should
3897 consider trying openssl and their own applications when compiled with
3898 these debugging symbols defined. It will help highlight potential bugs in
3899 their own code, and will improve the test coverage for OpenSSL itself. At
3900 some point, these tighter rules will become openssl's default to improve
3901 maintainability, though the assert()s and other overheads will remain only
3902 in debugging configurations. See bn.h for more details.
3903 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3905 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3906 that can only be obtained through BN_CTX_new() (which implicitly
3907 initialises it). The presence of this function only made it possible
3908 to overwrite an existing structure (and cause memory leaks).
3911 *) Because of the callback-based approach for implementing LHASH as a
3912 template type, lh_insert() adds opaque objects to hash-tables and
3913 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3914 to clean up those corresponding objects before destroying the hash table
3915 (and losing the object pointers). So some over-zealous constifications in
3916 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3917 objects as "const" and the lh_doall[_arg] callback wrappers are not
3918 prototyped to have "const" restrictions on the object pointers they are
3919 given (and so aren't required to cast them away any more).
3922 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3923 (speed) prefers to use its own implementation. The two implementations
3924 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3925 its object type properly exposed (MS_TM) instead of casting to/from "char
3926 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3927 aren't necessarily the greatest nomenclatures - but this is what was used
3928 internally to the implementation so I've used that for now.
3931 *) Ensure that deprecated functions do not get compiled when
3932 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3933 the self-tests were still using deprecated key-generation functions so
3934 these have been updated also.
3937 *) Reorganise PKCS#7 code to separate the digest location functionality
3938 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3939 New function PKCS7_set_digest() to set the digest type for PKCS#7
3940 digestedData type. Add additional code to correctly generate the
3941 digestedData type and add support for this type in PKCS7 initialization
3945 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3946 structure of type "other".
3949 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3950 sure the loop does correctly stop and breaking ("division by zero")
3951 modulus operations are not performed. The (pre-generated) prime
3952 table crypto/bn/bn_prime.h was already correct, but it could not be
3953 re-generated on some platforms because of the "division by zero"
3954 situation in the script.
3955 [Ralf S. Engelschall]
3957 *) Update support for ECC-based TLS ciphersuites according to
3958 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3959 SHA-1 now is only used for "small" curves (where the
3960 representation of a field element takes up to 24 bytes); for
3961 larger curves, the field element resulting from ECDH is directly
3962 used as premaster secret.
3963 [Douglas Stebila (Sun Microsystems Laboratories)]
3965 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3966 curve secp160r1 to the tests.
3967 [Douglas Stebila (Sun Microsystems Laboratories)]
3969 *) Add the possibility to load symbols globally with DSO.
3970 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3972 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3973 control of the error stack.
3976 *) Add support for STORE in ENGINE.
3979 *) Add the STORE type. The intention is to provide a common interface
3980 to certificate and key stores, be they simple file-based stores, or
3981 HSM-type store, or LDAP stores, or...
3982 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3985 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3986 pass a list of arguments to any function as well as provide a way
3987 for a function to pass data back to the caller.
3990 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3991 works like BUF_strdup() but can be used to duplicate a portion of
3992 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3996 *) Add the function sk_find_ex() which works like sk_find(), but will
3997 return an index to an element even if an exact match couldn't be
3998 found. The index is guaranteed to point at the element where the
3999 searched-for key would be inserted to preserve sorting order.
4002 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4003 takes an extra flags argument for optional functionality. Currently,
4004 the following flags are defined:
4006 OBJ_BSEARCH_VALUE_ON_NOMATCH
4007 This one gets OBJ_bsearch_ex() to return a pointer to the first
4008 element where the comparing function returns a negative or zero
4011 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4012 This one gets OBJ_bsearch_ex() to return a pointer to the first
4013 element where the comparing function returns zero. This is useful
4014 if there are more than one element where the comparing function
4018 *) Make it possible to create self-signed certificates with 'openssl ca'
4019 in such a way that the self-signed certificate becomes part of the
4020 CA database and uses the same mechanisms for serial number generation
4021 as all other certificate signing. The new flag '-selfsign' enables
4022 this functionality. Adapt CA.sh and CA.pl.in.
4025 *) Add functionality to check the public key of a certificate request
4026 against a given private. This is useful to check that a certificate
4027 request can be signed by that key (self-signing).
4030 *) Make it possible to have multiple active certificates with the same
4031 subject in the CA index file. This is done only if the keyword
4032 'unique_subject' is set to 'no' in the main CA section (default
4033 if 'CA_default') of the configuration file. The value is saved
4034 with the database itself in a separate index attribute file,
4035 named like the index file with '.attr' appended to the name.
4038 *) Generate muti valued AVAs using '+' notation in config files for
4042 *) Support for nameConstraints certificate extension.
4045 *) Support for policyConstraints certificate extension.
4048 *) Support for policyMappings certificate extension.
4051 *) Make sure the default DSA_METHOD implementation only uses its
4052 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4053 and change its own handlers to be NULL so as to remove unnecessary
4054 indirection. This lets alternative implementations fallback to the
4055 default implementation more easily.
4058 *) Support for directoryName in GeneralName related extensions
4062 *) Make it possible to link applications using Makefile.shared.
4063 Make that possible even when linking against static libraries!
4066 *) Support for single pass processing for S/MIME signing. This now
4067 means that S/MIME signing can be done from a pipe, in addition
4068 cleartext signing (multipart/signed type) is effectively streaming
4069 and the signed data does not need to be all held in memory.
4071 This is done with a new flag PKCS7_STREAM. When this flag is set
4072 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4073 is done after the data is output (and digests calculated) in
4074 SMIME_write_PKCS7().
4077 *) Add full support for -rpath/-R, both in shared libraries and
4078 applications, at least on the platforms where it's known how
4082 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4083 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4084 will now compute a table of multiples of the generator that
4085 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4086 faster (notably in the case of a single point multiplication,
4087 scalar * generator).
4088 [Nils Larsch, Bodo Moeller]
4090 *) IPv6 support for certificate extensions. The various extensions
4091 which use the IP:a.b.c.d can now take IPv6 addresses using the
4092 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4096 *) Added an ENGINE that implements RSA by performing private key
4097 exponentiations with the GMP library. The conversions to and from
4098 GMP's mpz_t format aren't optimised nor are any montgomery forms
4099 cached, and on x86 it appears OpenSSL's own performance has caught up.
4100 However there are likely to be other architectures where GMP could
4101 provide a boost. This ENGINE is not built in by default, but it can be
4102 specified at Configure time and should be accompanied by the necessary
4103 linker additions, eg;
4104 ./config -DOPENSSL_USE_GMP -lgmp
4107 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4108 testing availability of engines with "-t" - the old behaviour is
4109 produced by increasing the feature's verbosity with "-tt".
4112 *) ECDSA routines: under certain error conditions uninitialized BN objects
4113 could be freed. Solution: make sure initialization is performed early
4114 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4118 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4119 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4120 software implementations. For DSA and DH, parameter generation can
4121 also be overriden by providing the appropriate method callbacks.
4124 *) Change the "progress" mechanism used in key-generation and
4125 primality testing to functions that take a new BN_GENCB pointer in
4126 place of callback/argument pairs. The new API functions have "_ex"
4127 postfixes and the older functions are reimplemented as wrappers for
4128 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4129 declarations of the old functions to help (graceful) attempts to
4130 migrate to the new functions. Also, the new key-generation API
4131 functions operate on a caller-supplied key-structure and return
4132 success/failure rather than returning a key or NULL - this is to
4133 help make "keygen" another member function of RSA_METHOD etc.
4135 Example for using the new callback interface:
4137 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4141 BN_GENCB_set(&my_cb, my_callback, my_arg);
4143 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4144 /* For the meaning of a, b in calls to my_callback(), see the
4145 * documentation of the function that calls the callback.
4146 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4147 * my_callback should return 1 if it wants BN_is_prime_ex()
4148 * to continue, or 0 to stop.
4153 *) Change the ZLIB compression method to be stateful, and make it
4154 available to TLS with the number defined in
4155 draft-ietf-tls-compression-04.txt.
4158 *) Add the ASN.1 structures and functions for CertificatePair, which
4159 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4161 CertificatePair ::= SEQUENCE {
4162 forward [0] Certificate OPTIONAL,
4163 reverse [1] Certificate OPTIONAL,
4164 -- at least one of the pair shall be present -- }
4166 Also implement the PEM functions to read and write certificate
4167 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4169 This needed to be defined, mostly for the sake of the LDAP
4170 attribute crossCertificatePair, but may prove useful elsewhere as
4174 *) Make it possible to inhibit symlinking of shared libraries in
4175 Makefile.shared, for Cygwin's sake.
4178 *) Extend the BIGNUM API by creating a function
4179 void BN_set_negative(BIGNUM *a, int neg);
4180 and a macro that behave like
4181 int BN_is_negative(const BIGNUM *a);
4183 to avoid the need to access 'a->neg' directly in applications.
4186 *) Implement fast modular reduction for pseudo-Mersenne primes
4187 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4188 EC_GROUP_new_curve_GFp() will now automatically use this
4190 [Nils Larsch <nla@trustcenter.de>]
4192 *) Add new lock type (CRYPTO_LOCK_BN).
4195 *) Change the ENGINE framework to automatically load engines
4196 dynamically from specific directories unless they could be
4197 found to already be built in or loaded. Move all the
4198 current engines except for the cryptodev one to a new
4200 The engines in engines/ are built as shared libraries if
4201 the "shared" options was given to ./Configure or ./config.
4202 Otherwise, they are inserted in libcrypto.a.
4203 /usr/local/ssl/engines is the default directory for dynamic
4204 engines, but that can be overriden at configure time through
4205 the usual use of --prefix and/or --openssldir, and at run
4206 time with the environment variable OPENSSL_ENGINES.
4207 [Geoff Thorpe and Richard Levitte]
4209 *) Add Makefile.shared, a helper makefile to build shared
4210 libraries. Adapt Makefile.org.
4213 *) Add version info to Win32 DLLs.
4214 [Peter 'Luna' Runestig" <peter@runestig.com>]
4216 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4217 can be added using this API to created arbitrary PKCS#12
4218 files while avoiding the low level API.
4220 New options to PKCS12_create(), key or cert can be NULL and
4221 will then be omitted from the output file. The encryption
4222 algorithm NIDs can be set to -1 for no encryption, the mac
4223 iteration count can be set to 0 to omit the mac.
4225 Enhance pkcs12 utility by making the -nokeys and -nocerts
4226 options work when creating a PKCS#12 file. New option -nomac
4227 to omit the mac, NONE can be set for an encryption algorithm.
4228 New code is modified to use the enhanced PKCS12_create()
4229 instead of the low level API.
4232 *) Extend ASN1 encoder to support indefinite length constructed
4233 encoding. This can output sequences tags and octet strings in
4234 this form. Modify pk7_asn1.c to support indefinite length
4235 encoding. This is experimental and needs additional code to
4236 be useful, such as an ASN1 bio and some enhanced streaming
4239 Extend template encode functionality so that tagging is passed
4240 down to the template encoder.
4243 *) Let 'openssl req' fail if an argument to '-newkey' is not
4244 recognized instead of using RSA as a default.
4247 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4248 As these are not official, they are not included in "ALL";
4249 the "ECCdraft" ciphersuite group alias can be used to select them.
4250 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4252 *) Add ECDH engine support.
4253 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4255 *) Add ECDH in new directory crypto/ecdh/.
4256 [Douglas Stebila (Sun Microsystems Laboratories)]
4258 *) Let BN_rand_range() abort with an error after 100 iterations
4259 without success (which indicates a broken PRNG).
4262 *) Change BN_mod_sqrt() so that it verifies that the input value
4263 is really the square of the return value. (Previously,
4264 BN_mod_sqrt would show GIGO behaviour.)
4267 *) Add named elliptic curves over binary fields from X9.62, SECG,
4268 and WAP/WTLS; add OIDs that were still missing.
4270 [Sheueling Chang Shantz and Douglas Stebila
4271 (Sun Microsystems Laboratories)]
4273 *) Extend the EC library for elliptic curves over binary fields
4274 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4277 EC_GF2m_simple_method
4281 EC_GROUP_new_curve_GF2m
4282 EC_GROUP_set_curve_GF2m
4283 EC_GROUP_get_curve_GF2m
4284 EC_POINT_set_affine_coordinates_GF2m
4285 EC_POINT_get_affine_coordinates_GF2m
4286 EC_POINT_set_compressed_coordinates_GF2m
4288 Point compression for binary fields is disabled by default for
4289 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4292 As binary polynomials are represented as BIGNUMs, various members
4293 of the EC_GROUP and EC_POINT data structures can be shared
4294 between the implementations for prime fields and binary fields;
4295 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4296 are essentially identical to their ..._GFp counterparts.
4297 (For simplicity, the '..._GFp' prefix has been dropped from
4298 various internal method names.)
4300 An internal 'field_div' method (similar to 'field_mul' and
4301 'field_sqr') has been added; this is used only for binary fields.
4303 [Sheueling Chang Shantz and Douglas Stebila
4304 (Sun Microsystems Laboratories)]
4306 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4307 through methods ('mul', 'precompute_mult').
4309 The generic implementations (now internally called 'ec_wNAF_mul'
4310 and 'ec_wNAF_precomputed_mult') remain the default if these
4311 methods are undefined.
4313 [Sheueling Chang Shantz and Douglas Stebila
4314 (Sun Microsystems Laboratories)]
4316 *) New function EC_GROUP_get_degree, which is defined through
4317 EC_METHOD. For curves over prime fields, this returns the bit
4318 length of the modulus.
4320 [Sheueling Chang Shantz and Douglas Stebila
4321 (Sun Microsystems Laboratories)]
4323 *) New functions EC_GROUP_dup, EC_POINT_dup.
4324 (These simply call ..._new and ..._copy).
4326 [Sheueling Chang Shantz and Douglas Stebila
4327 (Sun Microsystems Laboratories)]
4329 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4330 Polynomials are represented as BIGNUMs (where the sign bit is not
4331 used) in the following functions [macros]:
4334 BN_GF2m_sub [= BN_GF2m_add]
4335 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4336 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4337 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4339 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4340 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4341 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4342 BN_GF2m_cmp [= BN_ucmp]
4344 (Note that only the 'mod' functions are actually for fields GF(2^m).
4345 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4347 For some functions, an the irreducible polynomial defining a
4348 field can be given as an 'unsigned int[]' with strictly
4349 decreasing elements giving the indices of those bits that are set;
4350 i.e., p[] represents the polynomial
4351 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4353 p[0] > p[1] > ... > p[k] = 0.
4354 This applies to the following functions:
4359 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4360 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4362 BN_GF2m_mod_sqrt_arr
4363 BN_GF2m_mod_solve_quad_arr
4367 Conversion can be performed by the following functions:
4372 bntest.c has additional tests for binary polynomial arithmetic.
4374 Two implementations for BN_GF2m_mod_div() are available.
4375 The default algorithm simply uses BN_GF2m_mod_inv() and
4376 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4377 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4378 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4380 [Sheueling Chang Shantz and Douglas Stebila
4381 (Sun Microsystems Laboratories)]
4383 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4384 functionality is disabled at compile-time.
4385 [Douglas Stebila <douglas.stebila@sun.com>]
4387 *) Change default behaviour of 'openssl asn1parse' so that more
4388 information is visible when viewing, e.g., a certificate:
4390 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4391 mode the content of non-printable OCTET STRINGs is output in a
4392 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4393 avoid the appearance of a printable string.
4394 [Nils Larsch <nla@trustcenter.de>]
4396 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4398 EC_GROUP_set_asn1_flag()
4399 EC_GROUP_get_asn1_flag()
4400 EC_GROUP_set_point_conversion_form()
4401 EC_GROUP_get_point_conversion_form()
4402 These control ASN1 encoding details:
4403 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4404 has been set to OPENSSL_EC_NAMED_CURVE.
4405 - Points are encoded in uncompressed form by default; options for
4406 asn1_for are as for point2oct, namely
4407 POINT_CONVERSION_COMPRESSED
4408 POINT_CONVERSION_UNCOMPRESSED
4409 POINT_CONVERSION_HYBRID
4411 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4414 EC_GROUP_get0_seed()
4415 EC_GROUP_get_seed_len()
4416 This is used only for ASN1 purposes (so far).
4417 [Nils Larsch <nla@trustcenter.de>]
4419 *) Add 'field_type' member to EC_METHOD, which holds the NID
4420 of the appropriate field type OID. The new function
4421 EC_METHOD_get_field_type() returns this value.
4422 [Nils Larsch <nla@trustcenter.de>]
4427 EC_POINT_point2hex()
4428 EC_POINT_hex2point()
4429 providing useful interfaces to EC_POINT_point2oct() and
4430 EC_POINT_oct2point().
4431 [Nils Larsch <nla@trustcenter.de>]
4433 *) Change internals of the EC library so that the functions
4434 EC_GROUP_set_generator()
4435 EC_GROUP_get_generator()
4436 EC_GROUP_get_order()
4437 EC_GROUP_get_cofactor()
4438 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4439 to methods, which would lead to unnecessary code duplication when
4440 adding different types of curves.
4441 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4443 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4444 arithmetic, and such that modified wNAFs are generated
4445 (which avoid length expansion in many cases).
4448 *) Add a function EC_GROUP_check_discriminant() (defined via
4449 EC_METHOD) that verifies that the curve discriminant is non-zero.
4451 Add a function EC_GROUP_check() that makes some sanity tests
4452 on a EC_GROUP, its generator and order. This includes
4453 EC_GROUP_check_discriminant().
4454 [Nils Larsch <nla@trustcenter.de>]
4456 *) Add ECDSA in new directory crypto/ecdsa/.
4458 Add applications 'openssl ecparam' and 'openssl ecdsa'
4459 (these are based on 'openssl dsaparam' and 'openssl dsa').
4461 ECDSA support is also included in various other files across the
4462 library. Most notably,
4463 - 'openssl req' now has a '-newkey ecdsa:file' option;
4464 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4465 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4466 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4467 them suitable for ECDSA where domain parameters must be
4468 extracted before the specific public key;
4469 - ECDSA engine support has been added.
4470 [Nils Larsch <nla@trustcenter.de>]
4472 *) Include some named elliptic curves, and add OIDs from X9.62,
4473 SECG, and WAP/WTLS. Each curve can be obtained from the new
4475 EC_GROUP_new_by_curve_name(),
4476 and the list of available named curves can be obtained with
4477 EC_get_builtin_curves().
4478 Also add a 'curve_name' member to EC_GROUP objects, which can be
4480 EC_GROUP_set_curve_name()
4481 EC_GROUP_get_curve_name()
4482 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4484 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4485 was actually never needed) and in BN_mul(). The removal in BN_mul()
4486 required a small change in bn_mul_part_recursive() and the addition
4487 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4488 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4489 bn_sub_words() and bn_add_words() except they take arrays with
4493 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4495 *) Cleanse PEM buffers before freeing them since they may contain
4497 [Benjamin Bennett <ben@psc.edu>]
4499 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4500 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4501 authentication-only ciphersuites.
4504 *) Since AES128 and AES256 share a single mask bit in the logic of
4505 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4506 kludge to work properly if AES128 is available and AES256 isn't.
4509 *) Expand security boundary to match 1.1.1 module.
4512 *) Remove redundant features: hash file source, editing of test vectors
4513 modify fipsld to use external fips_premain.c signature.
4516 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4517 run algorithm test programs.
4520 *) Make algorithm test programs more tolerant of whitespace.
4523 *) Have SSL/TLS server implementation tolerate "mismatched" record
4524 protocol version while receiving ClientHello even if the
4525 ClientHello is fragmented. (The server can't insist on the
4526 particular protocol version it has chosen before the ServerHello
4527 message has informed the client about his choice.)
4530 *) Load error codes if they are not already present instead of using a
4531 static variable. This allows them to be cleanly unloaded and reloaded.
4534 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4536 *) Introduce limits to prevent malicious keys being able to
4537 cause a denial of service. (CVE-2006-2940)
4538 [Steve Henson, Bodo Moeller]
4540 *) Fix ASN.1 parsing of certain invalid structures that can result
4541 in a denial of service. (CVE-2006-2937) [Steve Henson]
4543 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4544 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4546 *) Fix SSL client code which could crash if connecting to a
4547 malicious SSLv2 server. (CVE-2006-4343)
4548 [Tavis Ormandy and Will Drewry, Google Security Team]
4550 *) Change ciphersuite string processing so that an explicit
4551 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4552 will no longer include "AES128-SHA"), and any other similar
4553 ciphersuite (same bitmap) from *other* protocol versions (so that
4554 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4555 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4556 changes from 0.9.8b and 0.9.8d.
4559 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4561 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4562 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4564 *) Change the Unix randomness entropy gathering to use poll() when
4565 possible instead of select(), since the latter has some
4566 undesirable limitations.
4567 [Darryl Miles via Richard Levitte and Bodo Moeller]
4569 *) Disable rogue ciphersuites:
4571 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4572 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4573 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4575 The latter two were purportedly from
4576 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4579 Also deactive the remaining ciphersuites from
4580 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4581 unofficial, and the ID has long expired.
4584 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4585 dual-core machines) and other potential thread-safety issues.
4588 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4590 *) Adapt fipsld and the build system to link against the validated FIPS
4591 module in FIPS mode.
4594 *) Fixes for VC++ 2005 build under Windows.
4597 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4598 from a Windows bash shell such as MSYS. It is autodetected from the
4599 "config" script when run from a VC++ environment. Modify standard VC++
4600 build to use fipscanister.o from the GNU make build.
4603 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4605 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4606 The value now differs depending on if you build for FIPS or not.
4607 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4608 safely run with a non-FIPSed libcrypto, as it may crash because of
4609 the difference induced by this change.
4612 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4614 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4615 (part of SSL_OP_ALL). This option used to disable the
4616 countermeasure against man-in-the-middle protocol-version
4617 rollback in the SSL 2.0 server implementation, which is a bad
4618 idea. (CVE-2005-2969)
4620 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4621 for Information Security, National Institute of Advanced Industrial
4622 Science and Technology [AIST], Japan)]
4624 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4625 mainly for FIPS compliance and not fully integrated at this stage.
4628 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4629 the exponentiation using a fixed-length exponent. (Otherwise,
4630 the information leaked through timing could expose the secret key
4631 after many signatures; cf. Bleichenbacher's attack on DSA with
4635 *) Make a new fixed-window mod_exp implementation the default for
4636 RSA, DSA, and DH private-key operations so that the sequence of
4637 squares and multiplies and the memory access pattern are
4638 independent of the particular secret key. This will mitigate
4639 cache-timing and potential related attacks.
4641 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4642 and this is automatically used by BN_mod_exp_mont() if the new flag
4643 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4644 will use this BN flag for private exponents unless the flag
4645 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4646 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4648 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4650 *) Change the client implementation for SSLv23_method() and
4651 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4652 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4653 (Previously, the SSL 2.0 backwards compatible Client Hello
4654 message format would be used even with SSL_OP_NO_SSLv2.)
4657 *) Add support for smime-type MIME parameter in S/MIME messages which some
4661 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4662 a threadsafe manner. Modify rsa code to use new function and add calls
4663 to dsa and dh code (which had race conditions before).
4666 *) Include the fixed error library code in the C error file definitions
4667 instead of fixing them up at runtime. This keeps the error code
4668 structures constant.
4671 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4673 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4676 *) Fixes for newer kerberos headers. NB: the casts are needed because
4677 the 'length' field is signed on one version and unsigned on another
4678 with no (?) obvious way to tell the difference, without these VC++
4679 complains. Also the "definition" of FAR (blank) is no longer included
4680 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4681 some needed definitions.
4684 *) Undo Cygwin change.
4687 *) Added support for proxy certificates according to RFC 3820.
4688 Because they may be a security thread to unaware applications,
4689 they must be explicitely allowed in run-time. See
4690 docs/HOWTO/proxy_certificates.txt for further information.
4693 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4695 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4696 server and client random values. Previously
4697 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4698 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4700 This change has negligible security impact because:
4702 1. Server and client random values still have 24 bytes of pseudo random
4705 2. Server and client random values are sent in the clear in the initial
4708 3. The master secret is derived using the premaster secret (48 bytes in
4709 size for static RSA ciphersuites) as well as client server and random
4712 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4715 [Stephen Henson, reported by UK NISCC]
4717 *) Use Windows randomness collection on Cygwin.
4720 *) Fix hang in EGD/PRNGD query when communication socket is closed
4721 prematurely by EGD/PRNGD.
4722 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4724 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4727 *) Back-port of selected performance improvements from development
4728 branch, as well as improved support for PowerPC platforms.
4731 *) Add lots of checks for memory allocation failure, error codes to indicate
4732 failure and freeing up memory if a failure occurs.
4733 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4735 *) Add new -passin argument to dgst.
4738 *) Perform some character comparisons of different types in X509_NAME_cmp:
4739 this is needed for some certificates that reencode DNs into UTF8Strings
4740 (in violation of RFC3280) and can't or wont issue name rollover
4744 *) Make an explicit check during certificate validation to see that
4745 the CA setting in each certificate on the chain is correct. As a
4746 side effect always do the following basic checks on extensions,
4747 not just when there's an associated purpose to the check:
4749 - if there is an unhandled critical extension (unless the user
4750 has chosen to ignore this fault)
4751 - if the path length has been exceeded (if one is set at all)
4752 - that certain extensions fit the associated purpose (if one has
4756 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4758 *) Avoid a race condition when CRLs are checked in a multi threaded
4759 environment. This would happen due to the reordering of the revoked
4760 entries during signature checking and serial number lookup. Now the
4761 encoding is cached and the serial number sort performed under a lock.
4762 Add new STACK function sk_is_sorted().
4765 *) Add Delta CRL to the extension code.
4768 *) Various fixes to s3_pkt.c so alerts are sent properly.
4769 [David Holmes <d.holmes@f5.com>]
4771 *) Reduce the chances of duplicate issuer name and serial numbers (in
4772 violation of RFC3280) using the OpenSSL certificate creation utilities.
4773 This is done by creating a random 64 bit value for the initial serial
4774 number when a serial number file is created or when a self signed
4775 certificate is created using 'openssl req -x509'. The initial serial
4776 number file is created using 'openssl x509 -next_serial' in CA.pl
4777 rather than being initialized to 1.
4780 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4782 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4783 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4784 [Joe Orton, Steve Henson]
4786 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4788 [Joe Orton, Steve Henson]
4790 *) Make it possible to have multiple active certificates with the same
4791 subject in the CA index file. This is done only if the keyword
4792 'unique_subject' is set to 'no' in the main CA section (default
4793 if 'CA_default') of the configuration file. The value is saved
4794 with the database itself in a separate index attribute file,
4795 named like the index file with '.attr' appended to the name.
4798 *) X509 verify fixes. Disable broken certificate workarounds when
4799 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4800 keyUsage extension present. Don't accept CRLs with unhandled critical
4801 extensions: since verify currently doesn't process CRL extensions this
4802 rejects a CRL with *any* critical extensions. Add new verify error codes
4806 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4807 A clarification of RFC2560 will require the use of OCTET STRINGs and
4808 some implementations cannot handle the current raw format. Since OpenSSL
4809 copies and compares OCSP nonces as opaque blobs without any attempt at
4810 parsing them this should not create any compatibility issues.
4813 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4814 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4815 this HMAC (and other) operations are several times slower than OpenSSL
4819 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4820 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4822 *) Use the correct content when signing type "other".
4825 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4827 *) Fix various bugs revealed by running the NISCC test suite:
4829 Stop out of bounds reads in the ASN1 code when presented with
4830 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4832 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4834 If verify callback ignores invalid public key errors don't try to check
4835 certificate signature with the NULL public key.
4839 *) New -ignore_err option in ocsp application to stop the server
4840 exiting on the first error in a request.
4843 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4844 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4848 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4849 extra data after the compression methods not only for TLS 1.0
4850 but also for SSL 3.0 (as required by the specification).
4851 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4853 *) Change X509_certificate_type() to mark the key as exported/exportable
4854 when it's 512 *bits* long, not 512 bytes.
4857 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4858 blocks during encryption.
4861 *) Various fixes to base64 BIO and non blocking I/O. On write
4862 flushes were not handled properly if the BIO retried. On read
4863 data was not being buffered properly and had various logic bugs.
4864 This also affects blocking I/O when the data being decoded is a
4868 *) Various S/MIME bugfixes and compatibility changes:
4869 output correct application/pkcs7 MIME type if
4870 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4871 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4872 of files as .eml work). Correctly handle very long lines in MIME
4876 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4878 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4879 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4880 a protocol version number mismatch like a decryption error
4881 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4884 *) Turn on RSA blinding by default in the default implementation
4885 to avoid a timing attack. Applications that don't want it can call
4886 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4887 They would be ill-advised to do so in most cases.
4888 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4890 *) Change RSA blinding code so that it works when the PRNG is not
4891 seeded (in this case, the secret RSA exponent is abused as
4892 an unpredictable seed -- if it is not unpredictable, there
4893 is no point in blinding anyway). Make RSA blinding thread-safe
4894 by remembering the creator's thread ID in rsa->blinding and
4895 having all other threads use local one-time blinding factors
4896 (this requires more computation than sharing rsa->blinding, but
4897 avoids excessive locking; and if an RSA object is not shared
4898 between threads, blinding will still be very fast).
4901 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4902 ENGINE as defaults for all supported algorithms irrespective of
4903 the 'flags' parameter. 'flags' is now honoured, so applications
4904 should make sure they are passing it correctly.
4907 *) Target "mingw" now allows native Windows code to be generated in
4908 the Cygwin environment as well as with the MinGW compiler.
4911 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4913 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4914 via timing by performing a MAC computation even if incorrect
4915 block cipher padding has been found. This is a countermeasure
4916 against active attacks where the attacker has to distinguish
4917 between bad padding and a MAC verification error. (CVE-2003-0078)
4919 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4920 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4921 Martin Vuagnoux (EPFL, Ilion)]
4923 *) Make the no-err option work as intended. The intention with no-err
4924 is not to have the whole error stack handling routines removed from
4925 libcrypto, it's only intended to remove all the function name and
4926 reason texts, thereby removing some of the footprint that may not
4927 be interesting if those errors aren't displayed anyway.
4929 NOTE: it's still possible for any application or module to have it's
4930 own set of error texts inserted. The routines are there, just not
4931 used by default when no-err is given.
4934 *) Add support for FreeBSD on IA64.
4935 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4937 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4938 Kerberos function mit_des_cbc_cksum(). Before this change,
4939 the value returned by DES_cbc_cksum() was like the one from
4940 mit_des_cbc_cksum(), except the bytes were swapped.
4941 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4943 *) Allow an application to disable the automatic SSL chain building.
4944 Before this a rather primitive chain build was always performed in
4945 ssl3_output_cert_chain(): an application had no way to send the
4946 correct chain if the automatic operation produced an incorrect result.
4948 Now the chain builder is disabled if either:
4950 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4952 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4954 The reasoning behind this is that an application would not want the
4955 auto chain building to take place if extra chain certificates are
4956 present and it might also want a means of sending no additional
4957 certificates (for example the chain has two certificates and the
4961 *) Add the possibility to build without the ENGINE framework.
4962 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4964 *) Under Win32 gmtime() can return NULL: check return value in
4965 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4968 *) DSA routines: under certain error conditions uninitialized BN objects
4969 could be freed. Solution: make sure initialization is performed early
4970 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4971 Nils Larsch <nla@trustcenter.de> via PR#459)
4974 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4975 checked on reconnect on the client side, therefore session resumption
4976 could still fail with a "ssl session id is different" error. This
4977 behaviour is masked when SSL_OP_ALL is used due to
4978 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4979 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4980 followup to PR #377.
4983 *) IA-32 assembler support enhancements: unified ELF targets, support
4984 for SCO/Caldera platforms, fix for Cygwin shared build.
4987 *) Add support for FreeBSD on sparc64. As a consequence, support for
4988 FreeBSD on non-x86 processors is separate from x86 processors on
4989 the config script, much like the NetBSD support.
4990 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4992 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4994 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4997 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4998 code (06) was taken as the first octet of the session ID and the last
4999 octet was ignored consequently. As a result SSLv2 client side session
5000 caching could not have worked due to the session ID mismatch between
5002 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5006 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5007 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5011 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5012 seems that in spite of existing for more than a year, many application
5013 author have done nothing to provide the necessary callbacks, which
5014 means that this particular engine will not work properly anywhere.
5015 This is a very unfortunate situation which forces us, in the name
5016 of usability, to give the hw_ncipher.c a static lock, which is part
5018 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5019 appear in 0.9.8 or later. We EXPECT application authors to have
5020 dealt properly with this when 0.9.8 is released (unless we actually
5021 make such changes in the libcrypto locking code that changes will
5022 have to be made anyway).
5025 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5026 octets have been read, EOF or an error occurs. Without this change
5027 some truncated ASN1 structures will not produce an error.
5030 *) Disable Heimdal support, since it hasn't been fully implemented.
5031 Still give the possibility to force the use of Heimdal, but with
5032 warnings and a request that patches get sent to openssl-dev.
5035 *) Add the VC-CE target, introduce the WINCE sysname, and add
5036 INSTALL.WCE and appropriate conditionals to make it build.
5037 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5039 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5040 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5041 edit numbers of the version.
5042 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5044 *) Introduce safe string copy and catenation functions
5045 (BUF_strlcpy() and BUF_strlcat()).
5046 [Ben Laurie (CHATS) and Richard Levitte]
5048 *) Avoid using fixed-size buffers for one-line DNs.
5049 [Ben Laurie (CHATS)]
5051 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5052 resizing buffers containing secrets, and use where appropriate.
5053 [Ben Laurie (CHATS)]
5055 *) Avoid using fixed size buffers for configuration file location.
5056 [Ben Laurie (CHATS)]
5058 *) Avoid filename truncation for various CA files.
5059 [Ben Laurie (CHATS)]
5061 *) Use sizeof in preference to magic numbers.
5062 [Ben Laurie (CHATS)]
5064 *) Avoid filename truncation in cert requests.
5065 [Ben Laurie (CHATS)]
5067 *) Add assertions to check for (supposedly impossible) buffer
5069 [Ben Laurie (CHATS)]
5071 *) Don't cache truncated DNS entries in the local cache (this could
5072 potentially lead to a spoofing attack).
5073 [Ben Laurie (CHATS)]
5075 *) Fix various buffers to be large enough for hex/decimal
5076 representations in a platform independent manner.
5077 [Ben Laurie (CHATS)]
5079 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5080 resizing buffers containing secrets, and use where appropriate.
5081 [Ben Laurie (CHATS)]
5083 *) Add BIO_indent() to avoid much slightly worrying code to do
5085 [Ben Laurie (CHATS)]
5087 *) Convert sprintf()/BIO_puts() to BIO_printf().
5088 [Ben Laurie (CHATS)]
5090 *) buffer_gets() could terminate with the buffer only half
5092 [Ben Laurie (CHATS)]
5094 *) Add assertions to prevent user-supplied crypto functions from
5095 overflowing internal buffers by having large block sizes, etc.
5096 [Ben Laurie (CHATS)]
5098 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5100 [Ben Laurie (CHATS)]
5102 *) Eliminate unused copy of key in RC4.
5103 [Ben Laurie (CHATS)]
5105 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5106 [Ben Laurie (CHATS)]
5108 *) Fix off-by-one error in EGD path.
5109 [Ben Laurie (CHATS)]
5111 *) If RANDFILE path is too long, ignore instead of truncating.
5112 [Ben Laurie (CHATS)]
5114 *) Eliminate unused and incorrectly sized X.509 structure
5116 [Ben Laurie (CHATS)]
5118 *) Eliminate unused and dangerous function knumber().
5119 [Ben Laurie (CHATS)]
5121 *) Eliminate unused and dangerous structure, KSSL_ERR.
5122 [Ben Laurie (CHATS)]
5124 *) Protect against overlong session ID context length in an encoded
5125 session object. Since these are local, this does not appear to be
5127 [Ben Laurie (CHATS)]
5129 *) Change from security patch (see 0.9.6e below) that did not affect
5130 the 0.9.6 release series:
5132 Remote buffer overflow in SSL3 protocol - an attacker could
5133 supply an oversized master key in Kerberos-enabled versions.
5135 [Ben Laurie (CHATS)]
5137 *) Change the SSL kerb5 codes to match RFC 2712.
5140 *) Make -nameopt work fully for req and add -reqopt switch.
5141 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5143 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5144 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5146 *) Make sure tests can be performed even if the corresponding algorithms
5147 have been removed entirely. This was also the last step to make
5148 OpenSSL compilable with DJGPP under all reasonable conditions.
5149 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5151 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5152 to allow version independent disabling of normally unselected ciphers,
5153 which may be activated as a side-effect of selecting a single cipher.
5155 (E.g., cipher list string "RSA" enables ciphersuites that are left
5156 out of "ALL" because they do not provide symmetric encryption.
5157 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5158 [Lutz Jaenicke, Bodo Moeller]
5160 *) Add appropriate support for separate platform-dependent build
5161 directories. The recommended way to make a platform-dependent
5162 build directory is the following (tested on Linux), maybe with
5165 # Place yourself outside of the OpenSSL source tree. In
5166 # this example, the environment variable OPENSSL_SOURCE
5167 # is assumed to contain the absolute OpenSSL source directory.
5168 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5169 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5170 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5171 mkdir -p `dirname $F`
5172 ln -s $OPENSSL_SOURCE/$F $F
5175 To be absolutely sure not to disturb the source tree, a "make clean"
5176 is a good thing. If it isn't successful, don't worry about it,
5177 it probably means the source directory is very clean.
5180 *) Make sure any ENGINE control commands make local copies of string
5181 pointers passed to them whenever necessary. Otherwise it is possible
5182 the caller may have overwritten (or deallocated) the original string
5183 data when a later ENGINE operation tries to use the stored values.
5184 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5186 *) Improve diagnostics in file reading and command-line digests.
5187 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5189 *) Add AES modes CFB and OFB to the object database. Correct an
5190 error in AES-CFB decryption.
5193 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5194 allows existing EVP_CIPHER_CTX structures to be reused after
5195 calling EVP_*Final(). This behaviour is used by encryption
5196 BIOs and some applications. This has the side effect that
5197 applications must explicitly clean up cipher contexts with
5198 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5201 *) Check the values of dna and dnb in bn_mul_recursive before calling
5202 bn_mul_comba (a non zero value means the a or b arrays do not contain
5203 n2 elements) and fallback to bn_mul_normal if either is not zero.
5206 *) Fix escaping of non-ASCII characters when using the -subj option
5207 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5210 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5211 form for "surname", serialNumber has no short form.
5212 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5213 therefore remove "mail" short name for "internet 7".
5214 The OID for unique identifiers in X509 certificates is
5215 x500UniqueIdentifier, not uniqueIdentifier.
5216 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5219 *) Add an "init" command to the ENGINE config module and auto initialize
5220 ENGINEs. Without any "init" command the ENGINE will be initialized
5221 after all ctrl commands have been executed on it. If init=1 the
5222 ENGINE is initailized at that point (ctrls before that point are run
5223 on the uninitialized ENGINE and after on the initialized one). If
5224 init=0 then the ENGINE will not be iniatialized at all.
5227 *) Fix the 'app_verify_callback' interface so that the user-defined
5228 argument is actually passed to the callback: In the
5229 SSL_CTX_set_cert_verify_callback() prototype, the callback
5230 declaration has been changed from
5233 int (*cb)(X509_STORE_CTX *,void *);
5234 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5235 i=s->ctx->app_verify_callback(&ctx)
5236 has been changed into
5237 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5239 To update applications using SSL_CTX_set_cert_verify_callback(),
5240 a dummy argument can be added to their callback functions.
5241 [D. K. Smetters <smetters@parc.xerox.com>]
5243 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5244 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5246 *) Add and OPENSSL_LOAD_CONF define which will cause
5247 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5248 This allows older applications to transparently support certain
5249 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5250 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5251 load the config file and OPENSSL_add_all_algorithms_conf() which will
5252 always load it have also been added.
5255 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5256 Adjust NIDs and EVP layer.
5257 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5259 *) Config modules support in openssl utility.
5261 Most commands now load modules from the config file,
5262 though in a few (such as version) this isn't done
5263 because it couldn't be used for anything.
5265 In the case of ca and req the config file used is
5266 the same as the utility itself: that is the -config
5267 command line option can be used to specify an
5271 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5272 use "openssl_conf" if filename is NULL use default openssl config file.
5275 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5276 config section name. Add a new flag to tolerate a missing config file
5277 and move code to CONF_modules_load_file().
5280 *) Support for crypto accelerator cards from Accelerated Encryption
5281 Processing, www.aep.ie. (Use engine 'aep')
5282 The support was copied from 0.9.6c [engine] and adapted/corrected
5283 to work with the new engine framework.
5284 [AEP Inc. and Richard Levitte]
5286 *) Support for SureWare crypto accelerator cards from Baltimore
5287 Technologies. (Use engine 'sureware')
5288 The support was copied from 0.9.6c [engine] and adapted
5289 to work with the new engine framework.
5292 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5293 make the newer ENGINE framework commands for the CHIL engine work.
5294 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5296 *) Make it possible to produce shared libraries on ReliantUNIX.
5297 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5299 *) Add the configuration target debug-linux-ppro.
5300 Make 'openssl rsa' use the general key loading routines
5301 implemented in apps.c, and make those routines able to
5302 handle the key format FORMAT_NETSCAPE and the variant
5304 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5306 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5307 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5309 *) Add -keyform to rsautl, and document -engine.
5310 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5312 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5313 BIO_R_NO_SUCH_FILE error code rather than the generic
5314 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5317 *) Add new functions
5319 ERR_peek_last_error_line
5320 ERR_peek_last_error_line_data.
5321 These are similar to
5324 ERR_peek_error_line_data,
5325 but report on the latest error recorded rather than the first one
5326 still in the error queue.
5327 [Ben Laurie, Bodo Moeller]
5329 *) default_algorithms option in ENGINE config module. This allows things
5331 default_algorithms = ALL
5332 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5335 *) Preliminary ENGINE config module.
5338 *) New experimental application configuration code.
5341 *) Change the AES code to follow the same name structure as all other
5342 symmetric ciphers, and behave the same way. Move everything to
5343 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5344 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5346 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5347 [Ben Laurie and Theo de Raadt]
5349 *) Add option to output public keys in req command.
5350 [Massimiliano Pala madwolf@openca.org]
5352 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5353 (up to about 10% better than before for P-192 and P-224).
5356 *) New functions/macros
5358 SSL_CTX_set_msg_callback(ctx, cb)
5359 SSL_CTX_set_msg_callback_arg(ctx, arg)
5360 SSL_set_msg_callback(ssl, cb)
5361 SSL_set_msg_callback_arg(ssl, arg)
5363 to request calling a callback function
5365 void cb(int write_p, int version, int content_type,
5366 const void *buf, size_t len, SSL *ssl, void *arg)
5368 whenever a protocol message has been completely received
5369 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5370 protocol version according to which the SSL library interprets
5371 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5372 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5373 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5374 specification (change_cipher_spec(20), alert(21), handshake(22)).
5375 'buf' and 'len' point to the actual message, 'ssl' to the
5376 SSL object, and 'arg' is the application-defined value set by
5377 SSL[_CTX]_set_msg_callback_arg().
5379 'openssl s_client' and 'openssl s_server' have new '-msg' options
5380 to enable a callback that displays all protocol messages.
5383 *) Change the shared library support so shared libraries are built as
5384 soon as the corresponding static library is finished, and thereby get
5385 openssl and the test programs linked against the shared library.
5386 This still only happens when the keyword "shard" has been given to
5387 the configuration scripts.
5389 NOTE: shared library support is still an experimental thing, and
5390 backward binary compatibility is still not guaranteed.
5391 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5393 *) Add support for Subject Information Access extension.
5394 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5396 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5397 additional bytes when new memory had to be allocated, not just
5398 when reusing an existing buffer.
5401 *) New command line and configuration option 'utf8' for the req command.
5402 This allows field values to be specified as UTF8 strings.
5405 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5406 runs for the former and machine-readable output for the latter.
5409 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5410 of the e-mail address in the DN (i.e., it will go into a certificate
5411 extension only). The new configuration file option 'email_in_dn = no'
5412 has the same effect.
5413 [Massimiliano Pala madwolf@openca.org]
5415 *) Change all functions with names starting with des_ to be starting
5416 with DES_ instead. Add wrappers that are compatible with libdes,
5417 but are named _ossl_old_des_*. Finally, add macros that map the
5418 des_* symbols to the corresponding _ossl_old_des_* if libdes
5419 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5420 desired, the des_* symbols will be mapped to DES_*, with one
5423 Since we provide two compatibility mappings, the user needs to
5424 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5425 compatibility is desired. The default (i.e., when that macro
5426 isn't defined) is OpenSSL 0.9.6c compatibility.
5428 There are also macros that enable and disable the support of old
5429 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5430 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5431 are defined, the default will apply: to support the old des routines.
5433 In either case, one must include openssl/des.h to get the correct
5434 definitions. Do not try to just include openssl/des_old.h, that
5437 NOTE: This is a major break of an old API into a new one. Software
5438 authors are encouraged to switch to the DES_ style functions. Some
5439 time in the future, des_old.h and the libdes compatibility functions
5440 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5441 default), and then completely removed.
5444 *) Test for certificates which contain unsupported critical extensions.
5445 If such a certificate is found during a verify operation it is
5446 rejected by default: this behaviour can be overridden by either
5447 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5448 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5449 X509_supported_extension() has also been added which returns 1 if a
5450 particular extension is supported.
5453 *) Modify the behaviour of EVP cipher functions in similar way to digests
5454 to retain compatibility with existing code.
5457 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5458 compatibility with existing code. In particular the 'ctx' parameter does
5459 not have to be to be initialized before the call to EVP_DigestInit() and
5460 it is tidied up after a call to EVP_DigestFinal(). New function
5461 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5462 EVP_MD_CTX_copy() changed to not require the destination to be
5463 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5464 requires the destination to be valid.
5466 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5467 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5470 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5471 so that complete 'Handshake' protocol structures are kept in memory
5472 instead of overwriting 'msg_type' and 'length' with 'body' data.
5475 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5476 [Massimo Santin via Richard Levitte]
5478 *) Major restructuring to the underlying ENGINE code. This includes
5479 reduction of linker bloat, separation of pure "ENGINE" manipulation
5480 (initialisation, etc) from functionality dealing with implementations
5481 of specific crypto iterfaces. This change also introduces integrated
5482 support for symmetric ciphers and digest implementations - so ENGINEs
5483 can now accelerate these by providing EVP_CIPHER and EVP_MD
5484 implementations of their own. This is detailed in crypto/engine/README
5485 as it couldn't be adequately described here. However, there are a few
5486 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5487 were changed in the original introduction of ENGINE code have now
5488 reverted back - the hooking from this code to ENGINE is now a good
5489 deal more passive and at run-time, operations deal directly with
5490 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5491 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5492 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5493 they were not being used by the framework as there is no concept of a
5494 BIGNUM_METHOD and they could not be generalised to the new
5495 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5496 ENGINE_cpy() has been removed as it cannot be consistently defined in
5500 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5503 *) Change mkdef.pl to sort symbols that get the same entry number,
5504 and make sure the automatically generated functions ERR_load_*
5505 become part of libeay.num as well.
5508 *) New function SSL_renegotiate_pending(). This returns true once
5509 renegotiation has been requested (either SSL_renegotiate() call
5510 or HelloRequest/ClientHello received from the peer) and becomes
5511 false once a handshake has been completed.
5512 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5513 sends a HelloRequest, but does not ensure that a handshake takes
5514 place. SSL_renegotiate_pending() is useful for checking if the
5515 client has followed the request.)
5518 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5519 By default, clients may request session resumption even during
5520 renegotiation (if session ID contexts permit); with this option,
5521 session resumption is possible only in the first handshake.
5523 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5524 more bits available for options that should not be part of
5525 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5528 *) Add some demos for certificate and certificate request creation.
5531 *) Make maximum certificate chain size accepted from the peer application
5532 settable (SSL*_get/set_max_cert_list()), as proposed by
5533 "Douglas E. Engert" <deengert@anl.gov>.
5536 *) Add support for shared libraries for Unixware-7
5537 (Boyd Lynn Gerber <gerberb@zenez.com>).
5540 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5541 be done prior to destruction. Use this to unload error strings from
5542 ENGINEs that load their own error strings. NB: This adds two new API
5543 functions to "get" and "set" this destroy handler in an ENGINE.
5546 *) Alter all existing ENGINE implementations (except "openssl" and
5547 "openbsd") to dynamically instantiate their own error strings. This
5548 makes them more flexible to be built both as statically-linked ENGINEs
5549 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5550 Also, add stub code to each that makes building them as self-contained
5551 shared-libraries easier (see README.ENGINE).
5554 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5555 implementations into applications that are completely implemented in
5556 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5557 commands that can be used to configure what shared-library to load and
5558 to control aspects of the way it is handled. Also, made an update to
5559 the README.ENGINE file that brings its information up-to-date and
5560 provides some information and instructions on the "dynamic" ENGINE
5561 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5564 *) Make it possible to unload ranges of ERR strings with a new
5565 "ERR_unload_strings" function.
5568 *) Add a copy() function to EVP_MD.
5571 *) Make EVP_MD routines take a context pointer instead of just the
5572 md_data void pointer.
5575 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5576 that the digest can only process a single chunk of data
5577 (typically because it is provided by a piece of
5578 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5579 is only going to provide a single chunk of data, and hence the
5580 framework needn't accumulate the data for oneshot drivers.
5583 *) As with "ERR", make it possible to replace the underlying "ex_data"
5584 functions. This change also alters the storage and management of global
5585 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5586 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5587 index counters. The API functions that use this state have been changed
5588 to take a "class_index" rather than pointers to the class's local STACK
5589 and counter, and there is now an API function to dynamically create new
5590 classes. This centralisation allows us to (a) plug a lot of the
5591 thread-safety problems that existed, and (b) makes it possible to clean
5592 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5593 such data would previously have always leaked in application code and
5594 workarounds were in place to make the memory debugging turn a blind eye
5595 to it. Application code that doesn't use this new function will still
5596 leak as before, but their memory debugging output will announce it now
5597 rather than letting it slide.
5599 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5600 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5601 has a return value to indicate success or failure.
5604 *) Make it possible to replace the underlying "ERR" functions such that the
5605 global state (2 LHASH tables and 2 locks) is only used by the "default"
5606 implementation. This change also adds two functions to "get" and "set"
5607 the implementation prior to it being automatically set the first time
5608 any other ERR function takes place. Ie. an application can call "get",
5609 pass the return value to a module it has just loaded, and that module
5610 can call its own "set" function using that value. This means the
5611 module's "ERR" operations will use (and modify) the error state in the
5612 application and not in its own statically linked copy of OpenSSL code.
5615 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5616 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5617 the operation, and provides a more encapsulated way for external code
5618 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5619 to use these functions rather than manually incrementing the counts.
5621 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5624 *) Add EVP test program.
5627 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5630 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5631 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5632 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5633 These allow a CRL to be built without having to access X509_CRL fields
5634 directly. Modify 'ca' application to use new functions.
5637 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5638 bug workarounds. Rollback attack detection is a security feature.
5639 The problem will only arise on OpenSSL servers when TLSv1 is not
5640 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5641 Software authors not wanting to support TLSv1 will have special reasons
5642 for their choice and can explicitly enable this option.
5643 [Bodo Moeller, Lutz Jaenicke]
5645 *) Rationalise EVP so it can be extended: don't include a union of
5646 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5647 (similar to those existing for EVP_CIPHER_CTX).
5652 EVP_MD_CTX_init(&md); /* new function call */
5653 EVP_DigestInit(&md, EVP_sha1());
5654 EVP_DigestUpdate(&md, in, len);
5655 EVP_DigestFinal(&md, out, NULL);
5656 EVP_MD_CTX_cleanup(&md); /* new function call */
5660 *) Make DES key schedule conform to the usual scheme, as well as
5661 correcting its structure. This means that calls to DES functions
5662 now have to pass a pointer to a des_key_schedule instead of a
5663 plain des_key_schedule (which was actually always a pointer
5666 des_key_schedule ks;
5668 des_set_key_checked(..., &ks);
5669 des_ncbc_encrypt(..., &ks, ...);
5671 (Note that a later change renames 'des_...' into 'DES_...'.)
5674 *) Initial reduction of linker bloat: the use of some functions, such as
5675 PEM causes large amounts of unused functions to be linked in due to
5676 poor organisation. For example pem_all.c contains every PEM function
5677 which has a knock on effect of linking in large amounts of (unused)
5678 ASN1 code. Grouping together similar functions and splitting unrelated
5679 functions prevents this.
5682 *) Cleanup of EVP macros.
5685 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5686 correct _ecb suffix.
5689 *) Add initial OCSP responder support to ocsp application. The
5690 revocation information is handled using the text based index
5691 use by the ca application. The responder can either handle
5692 requests generated internally, supplied in files (for example
5693 via a CGI script) or using an internal minimal server.
5696 *) Add configuration choices to get zlib compression for TLS.
5699 *) Changes to Kerberos SSL for RFC 2712 compliance:
5700 1. Implemented real KerberosWrapper, instead of just using
5701 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5702 2. Implemented optional authenticator field of KerberosWrapper.
5704 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5705 and authenticator structs; see crypto/krb5/.
5707 Generalized Kerberos calls to support multiple Kerberos libraries.
5708 [Vern Staats <staatsvr@asc.hpc.mil>,
5709 Jeffrey Altman <jaltman@columbia.edu>
5710 via Richard Levitte]
5712 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5713 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5714 values for each of the key sizes rather than having just
5715 parameters (and 'speed' generating keys each time).
5718 *) Speed up EVP routines.
5721 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5722 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5723 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5724 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5726 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5727 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5728 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5731 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5733 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5736 *) Added the OS2-EMX target.
5737 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5739 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5740 to support NCONF routines in extension code. New function CONF_set_nconf()
5741 to allow functions which take an NCONF to also handle the old LHASH
5742 structure: this means that the old CONF compatible routines can be
5743 retained (in particular wrt extensions) without having to duplicate the
5744 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5747 *) Enhance the general user interface with mechanisms for inner control
5748 and with possibilities to have yes/no kind of prompts.
5751 *) Change all calls to low level digest routines in the library and
5752 applications to use EVP. Add missing calls to HMAC_cleanup() and
5753 don't assume HMAC_CTX can be copied using memcpy().
5754 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5756 *) Add the possibility to control engines through control names but with
5757 arbitrary arguments instead of just a string.
5758 Change the key loaders to take a UI_METHOD instead of a callback
5759 function pointer. NOTE: this breaks binary compatibility with earlier
5760 versions of OpenSSL [engine].
5761 Adapt the nCipher code for these new conditions and add a card insertion
5765 *) Enhance the general user interface with mechanisms to better support
5766 dialog box interfaces, application-defined prompts, the possibility
5767 to use defaults (for example default passwords from somewhere else)
5768 and interrupts/cancellations.
5771 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5772 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5775 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5776 tidy up some unnecessarily weird code in 'sk_new()').
5777 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5779 *) Change the key loading routines for ENGINEs to use the same kind
5780 callback (pem_password_cb) as all other routines that need this
5784 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5785 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5786 than this minimum value is recommended.
5789 *) New random seeder for OpenVMS, using the system process statistics
5790 that are easily reachable.
5793 *) Windows apparently can't transparently handle global
5794 variables defined in DLLs. Initialisations such as:
5796 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5798 wont compile. This is used by the any applications that need to
5799 declare their own ASN1 modules. This was fixed by adding the option
5800 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5801 needed for static libraries under Win32.
5804 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5805 setting of purpose and trust fields. New X509_STORE trust and
5806 purpose functions and tidy up setting in other SSL functions.
5809 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5810 structure. These are inherited by X509_STORE_CTX when it is
5811 initialised. This allows various defaults to be set in the
5812 X509_STORE structure (such as flags for CRL checking and custom
5813 purpose or trust settings) for functions which only use X509_STORE_CTX
5814 internally such as S/MIME.
5816 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5817 trust settings if they are not set in X509_STORE. This allows X509_STORE
5818 purposes and trust (in S/MIME for example) to override any set by default.
5820 Add command line options for CRL checking to smime, s_client and s_server
5824 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5825 are set then the CRL is looked up in the X509_STORE structure and
5826 its validity and signature checked, then if the certificate is found
5827 in the CRL the verify fails with a revoked error.
5829 Various new CRL related callbacks added to X509_STORE_CTX structure.
5831 Command line options added to 'verify' application to support this.
5833 This needs some additional work, such as being able to handle multiple
5834 CRLs with different times, extension based lookup (rather than just
5835 by subject name) and ultimately more complete V2 CRL extension
5839 *) Add a general user interface API (crypto/ui/). This is designed
5840 to replace things like des_read_password and friends (backward
5841 compatibility functions using this new API are provided).
5842 The purpose is to remove prompting functions from the DES code
5843 section as well as provide for prompting through dialog boxes in
5844 a window system and the like.
5847 *) Add "ex_data" support to ENGINE so implementations can add state at a
5848 per-structure level rather than having to store it globally.
5851 *) Make it possible for ENGINE structures to be copied when retrieved by
5852 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5853 This causes the "original" ENGINE structure to act like a template,
5854 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5855 operational state can be localised to each ENGINE structure, despite the
5856 fact they all share the same "methods". New ENGINE structures returned in
5857 this case have no functional references and the return value is the single
5858 structural reference. This matches the single structural reference returned
5859 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5863 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5864 needs to match any other type at all we need to manually clear the
5868 *) Changes to the "openssl engine" utility to include;
5869 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5870 about an ENGINE's available control commands.
5871 - executing control commands from command line arguments using the
5872 '-pre' and '-post' switches. '-post' is only used if '-t' is
5873 specified and the ENGINE is successfully initialised. The syntax for
5874 the individual commands are colon-separated, for example;
5875 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5878 *) New dynamic control command support for ENGINEs. ENGINEs can now
5879 declare their own commands (numbers), names (strings), descriptions,
5880 and input types for run-time discovery by calling applications. A
5881 subset of these commands are implicitly classed as "executable"
5882 depending on their input type, and only these can be invoked through
5883 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5884 can be based on user input, config files, etc). The distinction is
5885 that "executable" commands cannot return anything other than a boolean
5886 result and can only support numeric or string input, whereas some
5887 discoverable commands may only be for direct use through
5888 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5889 pointers, or other custom uses. The "executable" commands are to
5890 support parameterisations of ENGINE behaviour that can be
5891 unambiguously defined by ENGINEs and used consistently across any
5892 OpenSSL-based application. Commands have been added to all the
5893 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5894 control over shared-library paths without source code alterations.
5897 *) Changed all ENGINE implementations to dynamically allocate their
5898 ENGINEs rather than declaring them statically. Apart from this being
5899 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5900 this also allows the implementations to compile without using the
5901 internal engine_int.h header.
5904 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5905 'const' value. Any code that should be able to modify a RAND_METHOD
5906 should already have non-const pointers to it (ie. they should only
5907 modify their own ones).
5910 *) Made a variety of little tweaks to the ENGINE code.
5911 - "atalla" and "ubsec" string definitions were moved from header files
5912 to C code. "nuron" string definitions were placed in variables
5913 rather than hard-coded - allowing parameterisation of these values
5914 later on via ctrl() commands.
5915 - Removed unused "#if 0"'d code.
5916 - Fixed engine list iteration code so it uses ENGINE_free() to release
5917 structural references.
5918 - Constified the RAND_METHOD element of ENGINE structures.
5919 - Constified various get/set functions as appropriate and added
5920 missing functions (including a catch-all ENGINE_cpy that duplicates
5921 all ENGINE values onto a new ENGINE except reference counts/state).
5922 - Removed NULL parameter checks in get/set functions. Setting a method
5923 or function to NULL is a way of cancelling out a previously set
5924 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5925 and doesn't justify the extra error symbols and code.
5926 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5927 flags from engine_int.h to engine.h.
5928 - Changed prototypes for ENGINE handler functions (init(), finish(),
5929 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5932 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5933 to the algorithm using long division. The binary algorithm can be
5934 used only if the modulus is odd. On 32-bit systems, it is faster
5935 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5936 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5937 up to 450 bits. In 64-bit environments, the binary algorithm
5938 appears to be advantageous for much longer moduli; here we use it
5939 for moduli up to 2048 bits.
5942 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5943 could not support the combine flag in choice fields.
5946 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5947 extensions from a certificate request to the certificate.
5950 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5951 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5952 file: this allows the display of the certificate about to be
5953 signed to be customised, to allow certain fields to be included
5954 or excluded and extension details. The old system didn't display
5955 multicharacter strings properly, omitted fields not in the policy
5956 and couldn't display additional details such as extensions.
5959 *) Function EC_POINTs_mul for multiple scalar multiplication
5960 of an arbitrary number of elliptic curve points
5961 \sum scalars[i]*points[i],
5962 optionally including the generator defined for the EC_GROUP:
5963 scalar*generator + \sum scalars[i]*points[i].
5965 EC_POINT_mul is a simple wrapper function for the typical case
5966 that the point list has just one item (besides the optional
5970 *) First EC_METHODs for curves over GF(p):
5972 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5973 operations and provides various method functions that can also
5974 operate with faster implementations of modular arithmetic.
5976 EC_GFp_mont_method() reuses most functions that are part of
5977 EC_GFp_simple_method, but uses Montgomery arithmetic.
5979 [Bodo Moeller; point addition and point doubling
5980 implementation directly derived from source code provided by
5981 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5983 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5984 crypto/ec/ec_lib.c):
5986 Curves are EC_GROUP objects (with an optional group generator)
5987 based on EC_METHODs that are built into the library.
5989 Points are EC_POINT objects based on EC_GROUP objects.
5991 Most of the framework would be able to handle curves over arbitrary
5992 finite fields, but as there are no obvious types for fields other
5993 than GF(p), some functions are limited to that for now.
5996 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5997 that the file contains a complete HTTP response.
6000 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6001 change the def and num file printf format specifier from "%-40sXXX"
6002 to "%-39s XXX". The latter will always guarantee a space after the
6003 field while the former will cause them to run together if the field
6004 is 40 of more characters long.
6007 *) Constify the cipher and digest 'method' functions and structures
6008 and modify related functions to take constant EVP_MD and EVP_CIPHER
6012 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6013 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6016 *) Modify EVP_Digest*() routines so they now return values. Although the
6017 internal software routines can never fail additional hardware versions
6021 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6023 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6024 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6027 ERR_R_NESTED_ASN1_ERROR
6029 ERR_R_MISSING_ASN1_EOS
6030 were 4 .. 9, conflicting with
6031 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6033 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6034 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6036 Add new error code 'ERR_R_INTERNAL_ERROR'.
6039 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6043 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6044 sets the subject name for a new request or supersedes the
6045 subject name in a given request. Formats that can be parsed are
6046 'CN=Some Name, OU=myOU, C=IT'
6048 'CN=Some Name/OU=myOU/C=IT'.
6050 Add options '-batch' and '-verbose' to 'openssl req'.
6051 [Massimiliano Pala <madwolf@hackmasters.net>]
6053 *) Introduce the possibility to access global variables through
6054 functions on platform were that's the best way to handle exporting
6055 global variables in shared libraries. To enable this functionality,
6056 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6057 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6058 is normally done by Configure or something similar).
6060 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6061 in the source file (foo.c) like this:
6063 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6064 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6066 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6067 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6069 OPENSSL_DECLARE_GLOBAL(int,foo);
6070 #define foo OPENSSL_GLOBAL_REF(foo)
6071 OPENSSL_DECLARE_GLOBAL(double,bar);
6072 #define bar OPENSSL_GLOBAL_REF(bar)
6074 The #defines are very important, and therefore so is including the
6075 header file everywhere where the defined globals are used.
6077 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6078 of ASN.1 items, but that structure is a bit different.
6080 The largest change is in util/mkdef.pl which has been enhanced with
6081 better and easier to understand logic to choose which symbols should
6082 go into the Windows .def files as well as a number of fixes and code
6083 cleanup (among others, algorithm keywords are now sorted
6084 lexicographically to avoid constant rewrites).
6087 *) In BN_div() keep a copy of the sign of 'num' before writing the
6088 result to 'rm' because if rm==num the value will be overwritten
6089 and produce the wrong result if 'num' is negative: this caused
6090 problems with BN_mod() and BN_nnmod().
6093 *) Function OCSP_request_verify(). This checks the signature on an
6094 OCSP request and verifies the signer certificate. The signer
6095 certificate is just checked for a generic purpose and OCSP request
6099 *) Add OCSP_check_validity() function to check the validity of OCSP
6100 responses. OCSP responses are prepared in real time and may only
6101 be a few seconds old. Simply checking that the current time lies
6102 between thisUpdate and nextUpdate max reject otherwise valid responses
6103 caused by either OCSP responder or client clock inaccuracy. Instead
6104 we allow thisUpdate and nextUpdate to fall within a certain period of
6105 the current time. The age of the response can also optionally be
6106 checked. Two new options -validity_period and -status_age added to
6110 *) If signature or public key algorithm is unrecognized print out its
6111 OID rather that just UNKNOWN.
6114 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6115 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6116 ID to be generated from the issuer certificate alone which can then be
6117 passed to OCSP_id_issuer_cmp().
6120 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6121 ASN1 modules to export functions returning ASN1_ITEM pointers
6122 instead of the ASN1_ITEM structures themselves. This adds several
6123 new macros which allow the underlying ASN1 function/structure to
6124 be accessed transparently. As a result code should not use ASN1_ITEM
6125 references directly (such as &X509_it) but instead use the relevant
6126 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6127 use of the new ASN1 code on platforms where exporting structures
6128 is problematical (for example in shared libraries) but exporting
6129 functions returning pointers to structures is not.
6132 *) Add support for overriding the generation of SSL/TLS session IDs.
6133 These callbacks can be registered either in an SSL_CTX or per SSL.
6134 The purpose of this is to allow applications to control, if they wish,
6135 the arbitrary values chosen for use as session IDs, particularly as it
6136 can be useful for session caching in multiple-server environments. A
6137 command-line switch for testing this (and any client code that wishes
6138 to use such a feature) has been added to "s_server".
6139 [Geoff Thorpe, Lutz Jaenicke]
6141 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6142 of the form '#if defined(...) || defined(...) || ...' and
6143 '#if !defined(...) && !defined(...) && ...'. This also avoids
6144 the growing number of special cases it was previously handling.
6147 *) Make all configuration macros available for application by making
6148 sure they are available in opensslconf.h, by giving them names starting
6149 with "OPENSSL_" to avoid conflicts with other packages and by making
6150 sure e_os2.h will cover all platform-specific cases together with
6152 Additionally, it is now possible to define configuration/platform-
6153 specific names (called "system identities"). In the C code, these
6154 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6155 macro with the name beginning with "OPENSSL_SYS_", which is determined
6156 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6160 *) New option -set_serial to 'req' and 'x509' this allows the serial
6161 number to use to be specified on the command line. Previously self
6162 signed certificates were hard coded with serial number 0 and the
6163 CA options of 'x509' had to use a serial number in a file which was
6167 *) New options to 'ca' utility to support V2 CRL entry extensions.
6168 Currently CRL reason, invalidity date and hold instruction are
6169 supported. Add new CRL extensions to V3 code and some new objects.
6172 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6173 disable standard block padding (aka PKCS#5 padding) in the EVP
6174 API, which was previously mandatory. This means that the data is
6175 not padded in any way and so the total length much be a multiple
6176 of the block size, otherwise an error occurs.
6179 *) Initial (incomplete) OCSP SSL support.
6182 *) New function OCSP_parse_url(). This splits up a URL into its host,
6183 port and path components: primarily to parse OCSP URLs. New -url
6184 option to ocsp utility.
6187 *) New nonce behavior. The return value of OCSP_check_nonce() now
6188 reflects the various checks performed. Applications can decide
6189 whether to tolerate certain situations such as an absent nonce
6190 in a response when one was present in a request: the ocsp application
6191 just prints out a warning. New function OCSP_add1_basic_nonce()
6192 this is to allow responders to include a nonce in a response even if
6193 the request is nonce-less.
6196 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6197 skipped when using openssl x509 multiple times on a single input file,
6198 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6201 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6202 set string type: to handle setting ASN1_TIME structures. Fix ca
6203 utility to correctly initialize revocation date of CRLs.
6206 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6207 the clients preferred ciphersuites and rather use its own preferences.
6208 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6209 Internet Explorer by ensuring unchanged hash method during stepup.
6210 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6213 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6214 to aes and add a new 'exist' option to print out symbols that don't
6218 *) Additional options to ocsp utility to allow flags to be set and
6219 additional certificates supplied.
6222 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6223 OCSP client a number of certificate to only verify the response
6227 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6228 handle the new API. Currently only ECB, CBC modes supported. Add new
6231 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6232 Encryption Standard (AES) Ciphersuites for Transport Layer
6233 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6234 not enabled by default and were not part of the "ALL" ciphersuite
6235 alias because they were not yet official; they could be
6236 explicitly requested by specifying the "AESdraft" ciphersuite
6237 group alias. In the final release of OpenSSL 0.9.7, the group
6238 alias is called "AES" and is part of "ALL".)
6239 [Ben Laurie, Steve Henson, Bodo Moeller]
6241 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6242 request to response.
6245 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6246 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6247 extract information from a certificate request. OCSP_response_create()
6248 creates a response and optionally adds a basic response structure.
6249 OCSP_basic_add1_status() adds a complete single response to a basic
6250 response and returns the OCSP_SINGLERESP structure just added (to allow
6251 extensions to be included for example). OCSP_basic_add1_cert() adds a
6252 certificate to a basic response and OCSP_basic_sign() signs a basic
6253 response with various flags. New helper functions ASN1_TIME_check()
6254 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6255 (converts ASN1_TIME to GeneralizedTime).
6258 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6259 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6260 structure from a certificate. X509_pubkey_digest() digests the public_key
6261 contents: this is used in various key identifiers.
6264 *) Make sk_sort() tolerate a NULL argument.
6265 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6267 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6268 passed by the function are trusted implicitly. If any of them signed the
6269 response then it is assumed to be valid and is not verified.
6272 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6273 to data. This was previously part of the PKCS7 ASN1 code. This
6274 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6275 [Steve Henson, reported by Kenneth R. Robinette
6276 <support@securenetterm.com>]
6278 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6279 routines: without these tracing memory leaks is very painful.
6280 Fix leaks in PKCS12 and PKCS7 routines.
6283 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6284 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6285 effectively meant GeneralizedTime would never be used. Now it
6286 is initialised to -1 but X509_time_adj() now has to check the value
6287 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6288 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6289 [Steve Henson, reported by Kenneth R. Robinette
6290 <support@securenetterm.com>]
6292 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6293 result in a zero length in the ASN1_INTEGER structure which was
6294 not consistent with the structure when d2i_ASN1_INTEGER() was used
6295 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6296 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6297 where it did not print out a minus for negative ASN1_INTEGER.
6300 *) Add summary printout to ocsp utility. The various functions which
6301 convert status values to strings have been renamed to:
6302 OCSP_response_status_str(), OCSP_cert_status_str() and
6303 OCSP_crl_reason_str() and are no longer static. New options
6304 to verify nonce values and to disable verification. OCSP response
6305 printout format cleaned up.
6308 *) Add additional OCSP certificate checks. These are those specified
6309 in RFC2560. This consists of two separate checks: the CA of the
6310 certificate being checked must either be the OCSP signer certificate
6311 or the issuer of the OCSP signer certificate. In the latter case the
6312 OCSP signer certificate must contain the OCSP signing extended key
6313 usage. This check is performed by attempting to match the OCSP
6314 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6315 in the OCSP_CERTID structures of the response.
6318 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6319 and related routines. This uses the standard OpenSSL certificate
6320 verify routines to perform initial checks (just CA validity) and
6321 to obtain the certificate chain. Then additional checks will be
6322 performed on the chain. Currently the root CA is checked to see
6323 if it is explicitly trusted for OCSP signing. This is used to set
6324 a root CA as a global signing root: that is any certificate that
6325 chains to that CA is an acceptable OCSP signing certificate.
6328 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6329 extensions from a separate configuration file.
6330 As when reading extensions from the main configuration file,
6331 the '-extensions ...' option may be used for specifying the
6333 [Massimiliano Pala <madwolf@comune.modena.it>]
6335 *) New OCSP utility. Allows OCSP requests to be generated or
6336 read. The request can be sent to a responder and the output
6337 parsed, outputed or printed in text form. Not complete yet:
6338 still needs to check the OCSP response validity.
6341 *) New subcommands for 'openssl ca':
6342 'openssl ca -status <serial>' prints the status of the cert with
6343 the given serial number (according to the index file).
6344 'openssl ca -updatedb' updates the expiry status of certificates
6346 [Massimiliano Pala <madwolf@comune.modena.it>]
6348 *) New '-newreq-nodes' command option to CA.pl. This is like
6349 '-newreq', but calls 'openssl req' with the '-nodes' option
6350 so that the resulting key is not encrypted.
6351 [Damien Miller <djm@mindrot.org>]
6353 *) New configuration for the GNU Hurd.
6354 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6356 *) Initial code to implement OCSP basic response verify. This
6357 is currently incomplete. Currently just finds the signer's
6358 certificate and verifies the signature on the response.
6361 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6362 value of OPENSSLDIR. This is available via the new '-d' option
6363 to 'openssl version', and is also included in 'openssl version -a'.
6366 *) Allowing defining memory allocation callbacks that will be given
6367 file name and line number information in additional arguments
6368 (a const char* and an int). The basic functionality remains, as
6369 well as the original possibility to just replace malloc(),
6370 realloc() and free() by functions that do not know about these
6371 additional arguments. To register and find out the current
6372 settings for extended allocation functions, the following
6373 functions are provided:
6375 CRYPTO_set_mem_ex_functions
6376 CRYPTO_set_locked_mem_ex_functions
6377 CRYPTO_get_mem_ex_functions
6378 CRYPTO_get_locked_mem_ex_functions
6380 These work the same way as CRYPTO_set_mem_functions and friends.
6381 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6382 extended allocation function is enabled.
6383 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6384 a conventional allocation function is enabled.
6385 [Richard Levitte, Bodo Moeller]
6387 *) Finish off removing the remaining LHASH function pointer casts.
6388 There should no longer be any prototype-casting required when using
6389 the LHASH abstraction, and any casts that remain are "bugs". See
6390 the callback types and macros at the head of lhash.h for details
6391 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6394 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6395 If /dev/[u]random devices are not available or do not return enough
6396 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6398 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6399 /etc/entropy will be queried once each in this sequence, quering stops
6400 when enough entropy was collected without querying more sockets.
6403 *) Change the Unix RAND_poll() variant to be able to poll several
6404 random devices, as specified by DEVRANDOM, until a sufficient amount
6405 of data has been collected. We spend at most 10 ms on each file
6406 (select timeout) and read in non-blocking mode. DEVRANDOM now
6407 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6408 (previously it was just the string "/dev/urandom"), so on typical
6409 platforms the 10 ms delay will never occur.
6410 Also separate out the Unix variant to its own file, rand_unix.c.
6411 For VMS, there's a currently-empty rand_vms.c.
6414 *) Move OCSP client related routines to ocsp_cl.c. These
6415 provide utility functions which an application needing
6416 to issue a request to an OCSP responder and analyse the
6417 response will typically need: as opposed to those which an
6418 OCSP responder itself would need which will be added later.
6420 OCSP_request_sign() signs an OCSP request with an API similar
6421 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6422 response. OCSP_response_get1_basic() extracts basic response
6423 from response. OCSP_resp_find_status(): finds and extracts status
6424 information from an OCSP_CERTID structure (which will be created
6425 when the request structure is built). These are built from lower
6426 level functions which work on OCSP_SINGLERESP structures but
6427 wont normally be used unless the application wishes to examine
6428 extensions in the OCSP response for example.
6430 Replace nonce routines with a pair of functions.
6431 OCSP_request_add1_nonce() adds a nonce value and optionally
6432 generates a random value. OCSP_check_nonce() checks the
6433 validity of the nonce in an OCSP response.
6436 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6437 This doesn't copy the supplied OCSP_CERTID and avoids the
6438 need to free up the newly created id. Change return type
6439 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6440 This can then be used to add extensions to the request.
6441 Deleted OCSP_request_new(), since most of its functionality
6442 is now in OCSP_REQUEST_new() (and the case insensitive name
6443 clash) apart from the ability to set the request name which
6444 will be added elsewhere.
6447 *) Update OCSP API. Remove obsolete extensions argument from
6448 various functions. Extensions are now handled using the new
6449 OCSP extension code. New simple OCSP HTTP function which
6450 can be used to send requests and parse the response.
6453 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6454 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6455 uses the special reorder version of SET OF to sort the attributes
6456 and reorder them to match the encoded order. This resolves a long
6457 standing problem: a verify on a PKCS7 structure just after signing
6458 it used to fail because the attribute order did not match the
6459 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6460 it uses the received order. This is necessary to tolerate some broken
6461 software that does not order SET OF. This is handled by encoding
6462 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6463 to produce the required SET OF.
6466 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6467 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6468 files to get correct declarations of the ASN.1 item variables.
6471 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6472 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6473 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6474 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6475 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6476 ASN1_ITEM and no wrapper functions.
6479 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6480 replace the old function pointer based I/O routines. Change most of
6481 the *_d2i_bio() and *_d2i_fp() functions to use these.
6484 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6485 lines, recognice more "algorithms" that can be deselected, and make
6486 it complain about algorithm deselection that isn't recognised.
6489 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6490 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6491 to use new functions. Add NO_ASN1_OLD which can be set to remove
6492 some old style ASN1 functions: this can be used to determine if old
6493 code will still work when these eventually go away.
6496 *) New extension functions for OCSP structures, these follow the
6497 same conventions as certificates and CRLs.
6500 *) New function X509V3_add1_i2d(). This automatically encodes and
6501 adds an extension. Its behaviour can be customised with various
6502 flags to append, replace or delete. Various wrappers added for
6503 certifcates and CRLs.
6506 *) Fix to avoid calling the underlying ASN1 print routine when
6507 an extension cannot be parsed. Correct a typo in the
6508 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6511 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6512 entries for variables.
6515 *) Add functionality to apps/openssl.c for detecting locking
6516 problems: As the program is single-threaded, all we have
6517 to do is register a locking callback using an array for
6518 storing which locks are currently held by the program.
6521 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6522 SSL_get_ex_data_X509_STORE_idx(), which is used in
6523 ssl_verify_cert_chain() and thus can be called at any time
6524 during TLS/SSL handshakes so that thread-safety is essential.
6525 Unfortunately, the ex_data design is not at all suited
6526 for multi-threaded use, so it probably should be abolished.
6529 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6530 [Broadcom, tweaked and integrated by Geoff Thorpe]
6532 *) Move common extension printing code to new function
6533 X509V3_print_extensions(). Reorganise OCSP print routines and
6534 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6537 *) New function X509_signature_print() to remove duplication in some
6541 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6542 set (this was treated exactly the same as SET OF previously). This
6543 is used to reorder the STACK representing the structure to match the
6544 encoding. This will be used to get round a problem where a PKCS7
6545 structure which was signed could not be verified because the STACK
6546 order did not reflect the encoded order.
6549 *) Reimplement the OCSP ASN1 module using the new code.
6552 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6553 for its ASN1 operations. The old style function pointers still exist
6554 for now but they will eventually go away.
6557 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6558 completely replaces the old ASN1 functionality with a table driven
6559 encoder and decoder which interprets an ASN1_ITEM structure describing
6560 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6561 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6562 has also been converted to the new form.
6565 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6566 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6567 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6568 for negative moduli.
6571 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6572 of not touching the result's sign bit.
6575 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6579 *) Changed the LHASH code to use prototypes for callbacks, and created
6580 macros to declare and implement thin (optionally static) functions
6581 that provide type-safety and avoid function pointer casting for the
6582 type-specific callbacks.
6585 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6587 [Veers Staats <staatsvr@asc.hpc.mil>,
6588 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6590 *) Reformat the FAQ so the different questions and answers can be divided
6591 in sections depending on the subject.
6594 *) Have the zlib compression code load ZLIB.DLL dynamically under
6598 *) New function BN_mod_sqrt for computing square roots modulo a prime
6599 (using the probabilistic Tonelli-Shanks algorithm unless
6600 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6601 be handled deterministically).
6602 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6604 *) Make BN_mod_inverse faster by explicitly handling small quotients
6605 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6606 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6609 *) New function BN_kronecker.
6612 *) Fix BN_gcd so that it works on negative inputs; the result is
6613 positive unless both parameters are zero.
6614 Previously something reasonably close to an infinite loop was
6615 possible because numbers could be growing instead of shrinking
6616 in the implementation of Euclid's algorithm.
6619 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6620 sign of the number in question.
6622 Fix BN_is_word(a,w) to work correctly for w == 0.
6624 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6625 because its test if the absolute value of 'a' equals 'w'.
6626 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6627 it exists mostly for use in the implementations of BN_is_zero(),
6628 BN_is_one(), and BN_is_word().
6631 *) New function BN_swap.
6634 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6635 the exponentiation functions are more likely to produce reasonable
6636 results on negative inputs.
6639 *) Change BN_mod_mul so that the result is always non-negative.
6640 Previously, it could be negative if one of the factors was negative;
6641 I don't think anyone really wanted that behaviour.
6644 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6645 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6646 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6647 and add new functions:
6656 BN_mod_lshift1_quick
6660 These functions always generate non-negative results.
6662 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6663 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6665 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6666 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6667 be reduced modulo m.
6668 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6671 The following entry accidentally appeared in the CHANGES file
6672 distributed with OpenSSL 0.9.7. The modifications described in
6673 it do *not* apply to OpenSSL 0.9.7.
6675 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6676 was actually never needed) and in BN_mul(). The removal in BN_mul()
6677 required a small change in bn_mul_part_recursive() and the addition
6678 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6679 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6680 bn_sub_words() and bn_add_words() except they take arrays with
6685 *) In 'openssl passwd', verify passwords read from the terminal
6686 unless the '-salt' option is used (which usually means that
6687 verification would just waste user's time since the resulting
6688 hash is going to be compared with some given password hash)
6689 or the new '-noverify' option is used.
6691 This is an incompatible change, but it does not affect
6692 non-interactive use of 'openssl passwd' (passwords on the command
6693 line, '-stdin' option, '-in ...' option) and thus should not
6697 *) Remove all references to RSAref, since there's no more need for it.
6700 *) Make DSO load along a path given through an environment variable
6701 (SHLIB_PATH) with shl_load().
6704 *) Constify the ENGINE code as a result of BIGNUM constification.
6705 Also constify the RSA code and most things related to it. In a
6706 few places, most notable in the depth of the ASN.1 code, ugly
6707 casts back to non-const were required (to be solved at a later
6711 *) Make it so the openssl application has all engines loaded by default.
6714 *) Constify the BIGNUM routines a little more.
6717 *) Add the following functions:
6719 ENGINE_load_cswift()
6721 ENGINE_load_atalla()
6723 ENGINE_load_builtin_engines()
6725 That way, an application can itself choose if external engines that
6726 are built-in in OpenSSL shall ever be used or not. The benefit is
6727 that applications won't have to be linked with libdl or other dso
6728 libraries unless it's really needed.
6730 Changed 'openssl engine' to load all engines on demand.
6731 Changed the engine header files to avoid the duplication of some
6732 declarations (they differed!).
6735 *) 'openssl engine' can now list capabilities.
6738 *) Better error reporting in 'openssl engine'.
6741 *) Never call load_dh_param(NULL) in s_server.
6744 *) Add engine application. It can currently list engines by name and
6745 identity, and test if they are actually available.
6748 *) Improve RPM specification file by forcing symbolic linking and making
6749 sure the installed documentation is also owned by root.root.
6750 [Damien Miller <djm@mindrot.org>]
6752 *) Give the OpenSSL applications more possibilities to make use of
6753 keys (public as well as private) handled by engines.
6756 *) Add OCSP code that comes from CertCo.
6759 *) Add VMS support for the Rijndael code.
6762 *) Added untested support for Nuron crypto accelerator.
6765 *) Add support for external cryptographic devices. This code was
6766 previously distributed separately as the "engine" branch.
6767 [Geoff Thorpe, Richard Levitte]
6769 *) Rework the filename-translation in the DSO code. It is now possible to
6770 have far greater control over how a "name" is turned into a filename
6771 depending on the operating environment and any oddities about the
6772 different shared library filenames on each system.
6775 *) Support threads on FreeBSD-elf in Configure.
6778 *) Fix for SHA1 assembly problem with MASM: it produces
6779 warnings about corrupt line number information when assembling
6780 with debugging information. This is caused by the overlapping
6782 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6785 NCONF_get_number() has no error checking at all. As a replacement,
6786 NCONF_get_number_e() is defined (_e for "error checking") and is
6787 promoted strongly. The old NCONF_get_number is kept around for
6788 binary backward compatibility.
6789 Make it possible for methods to load from something other than a BIO,
6790 by providing a function pointer that is given a name instead of a BIO.
6791 For example, this could be used to load configuration data from an
6795 *) Fix for non blocking accept BIOs. Added new I/O special reason
6796 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6797 with non blocking I/O was not possible because no retry code was
6798 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6802 *) Added the beginnings of Rijndael support.
6805 *) Fix for bug in DirectoryString mask setting. Add support for
6806 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6807 to allow certificate printing to more controllable, additional
6808 'certopt' option to 'x509' to allow new printing options to be
6812 *) Clean old EAY MD5 hack from e_os.h.
6815 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6817 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6818 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6819 [Joe Orton, Steve Henson]
6821 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6823 *) Fix additional bug revealed by the NISCC test suite:
6825 Stop bug triggering large recursion when presented with
6826 certain ASN.1 tags (CVE-2003-0851)
6829 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6831 *) Fix various bugs revealed by running the NISCC test suite:
6833 Stop out of bounds reads in the ASN1 code when presented with
6834 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6836 If verify callback ignores invalid public key errors don't try to check
6837 certificate signature with the NULL public key.
6841 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6842 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6846 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6847 extra data after the compression methods not only for TLS 1.0
6848 but also for SSL 3.0 (as required by the specification).
6849 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6851 *) Change X509_certificate_type() to mark the key as exported/exportable
6852 when it's 512 *bits* long, not 512 bytes.
6855 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6857 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6858 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6859 a protocol version number mismatch like a decryption error
6860 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6863 *) Turn on RSA blinding by default in the default implementation
6864 to avoid a timing attack. Applications that don't want it can call
6865 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6866 They would be ill-advised to do so in most cases.
6867 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6869 *) Change RSA blinding code so that it works when the PRNG is not
6870 seeded (in this case, the secret RSA exponent is abused as
6871 an unpredictable seed -- if it is not unpredictable, there
6872 is no point in blinding anyway). Make RSA blinding thread-safe
6873 by remembering the creator's thread ID in rsa->blinding and
6874 having all other threads use local one-time blinding factors
6875 (this requires more computation than sharing rsa->blinding, but
6876 avoids excessive locking; and if an RSA object is not shared
6877 between threads, blinding will still be very fast).
6880 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6882 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6883 via timing by performing a MAC computation even if incorrrect
6884 block cipher padding has been found. This is a countermeasure
6885 against active attacks where the attacker has to distinguish
6886 between bad padding and a MAC verification error. (CVE-2003-0078)
6888 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6889 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6890 Martin Vuagnoux (EPFL, Ilion)]
6892 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6894 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6895 memory from it's contents. This is done with a counter that will
6896 place alternating values in each byte. This can be used to solve
6897 two issues: 1) the removal of calls to memset() by highly optimizing
6898 compilers, and 2) cleansing with other values than 0, since those can
6899 be read through on certain media, for example a swap space on disk.
6902 *) Bugfix: client side session caching did not work with external caching,
6903 because the session->cipher setting was not restored when reloading
6904 from the external cache. This problem was masked, when
6905 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6906 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6909 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6910 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6911 [Zeev Lieber <zeev-l@yahoo.com>]
6913 *) Undo an undocumented change introduced in 0.9.6e which caused
6914 repeated calls to OpenSSL_add_all_ciphers() and
6915 OpenSSL_add_all_digests() to be ignored, even after calling
6919 *) Change the default configuration reader to deal with last line not
6920 being properly terminated.
6923 *) Change X509_NAME_cmp() so it applies the special rules on handling
6924 DN values that are of type PrintableString, as well as RDNs of type
6925 emailAddress where the value has the type ia5String.
6926 [stefank@valicert.com via Richard Levitte]
6928 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6929 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6930 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6931 the bitwise-OR of the two for use by the majority of applications
6932 wanting this behaviour, and update the docs. The documented
6933 behaviour and actual behaviour were inconsistent and had been
6934 changing anyway, so this is more a bug-fix than a behavioural
6936 [Geoff Thorpe, diagnosed by Nadav Har'El]
6938 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6939 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6942 *) Fix initialization code race conditions in
6943 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6944 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6945 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6946 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6947 ssl2_get_cipher_by_char(),
6948 ssl3_get_cipher_by_char().
6949 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6951 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6952 the cached sessions are flushed, as the remove_cb() might use ex_data
6953 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6954 (see [openssl.org #212]).
6955 [Geoff Thorpe, Lutz Jaenicke]
6957 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6958 length, instead of the encoding length to d2i_ASN1_OBJECT.
6961 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6963 *) [In 0.9.6g-engine release:]
6964 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6965 [Lynn Gazis <lgazis@rainbow.com>]
6967 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6969 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6970 and get fix the header length calculation.
6971 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6972 Alon Kantor <alonk@checkpoint.com> (and others),
6975 *) Use proper error handling instead of 'assertions' in buffer
6976 overflow checks added in 0.9.6e. This prevents DoS (the
6977 assertions could call abort()).
6978 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6980 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6982 *) Add various sanity checks to asn1_get_length() to reject
6983 the ASN1 length bytes if they exceed sizeof(long), will appear
6984 negative or the content length exceeds the length of the
6986 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6988 *) Fix cipher selection routines: ciphers without encryption had no flags
6989 for the cipher strength set and where therefore not handled correctly
6990 by the selection routines (PR #130).
6993 *) Fix EVP_dsa_sha macro.
6997 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6998 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6999 that was added in OpenSSL 0.9.6d.
7001 As the countermeasure turned out to be incompatible with some
7002 broken SSL implementations, the new option is part of SSL_OP_ALL.
7003 SSL_OP_ALL is usually employed when compatibility with weird SSL
7004 implementations is desired (e.g. '-bugs' option to 's_client' and
7005 's_server'), so the new option is automatically set in many
7009 *) Changes in security patch:
7011 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7012 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7013 Air Force Materiel Command, USAF, under agreement number
7016 *) Add various sanity checks to asn1_get_length() to reject
7017 the ASN1 length bytes if they exceed sizeof(long), will appear
7018 negative or the content length exceeds the length of the
7019 supplied buffer. (CVE-2002-0659)
7020 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7022 *) Assertions for various potential buffer overflows, not known to
7024 [Ben Laurie (CHATS)]
7026 *) Various temporary buffers to hold ASCII versions of integers were
7027 too small for 64 bit platforms. (CVE-2002-0655)
7028 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7030 *) Remote buffer overflow in SSL3 protocol - an attacker could
7031 supply an oversized session ID to a client. (CVE-2002-0656)
7032 [Ben Laurie (CHATS)]
7034 *) Remote buffer overflow in SSL2 protocol - an attacker could
7035 supply an oversized client master key. (CVE-2002-0656)
7036 [Ben Laurie (CHATS)]
7038 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7040 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7041 encoded as NULL) with id-dsa-with-sha1.
7042 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7044 *) Check various X509_...() return values in apps/req.c.
7045 [Nils Larsch <nla@trustcenter.de>]
7047 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7048 an end-of-file condition would erronously be flagged, when the CRLF
7049 was just at the end of a processed block. The bug was discovered when
7050 processing data through a buffering memory BIO handing the data to a
7051 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7052 <ptsekov@syntrex.com> and Nedelcho Stanev.
7055 *) Implement a countermeasure against a vulnerability recently found
7056 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7057 before application data chunks to avoid the use of known IVs
7058 with data potentially chosen by the attacker.
7061 *) Fix length checks in ssl3_get_client_hello().
7064 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7065 to prevent ssl3_read_internal() from incorrectly assuming that
7066 ssl3_read_bytes() found application data while handshake
7067 processing was enabled when in fact s->s3->in_read_app_data was
7068 merely automatically cleared during the initial handshake.
7069 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7071 *) Fix object definitions for Private and Enterprise: they were not
7072 recognized in their shortname (=lowercase) representation. Extend
7073 obj_dat.pl to issue an error when using undefined keywords instead
7074 of silently ignoring the problem (Svenning Sorensen
7075 <sss@sss.dnsalias.net>).
7078 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7079 generators, i.e. generators other than 2 and 5. (Previously, the
7080 code did not properly initialise the 'add' and 'rem' values to
7081 BN_generate_prime().)
7083 In the new general case, we do not insist that 'generator' is
7084 actually a primitive root: This requirement is rather pointless;
7085 a generator of the order-q subgroup is just as good, if not
7089 *) Map new X509 verification errors to alerts. Discovered and submitted by
7090 Tom Wu <tom@arcot.com>.
7093 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7094 returning non-zero before the data has been completely received
7095 when using non-blocking I/O.
7096 [Bodo Moeller; problem pointed out by John Hughes]
7098 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7099 [Ben Laurie, Lutz Jaenicke]
7101 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7102 Yoram Zahavi <YoramZ@gilian.com>).
7105 *) Add information about CygWin 1.3 and on, and preserve proper
7106 configuration for the versions before that.
7107 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7109 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7110 check whether we deal with a copy of a session and do not delete from
7111 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7112 <izhar@checkpoint.com>.
7115 *) Do not store session data into the internal session cache, if it
7116 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7117 flag is set). Proposed by Aslam <aslam@funk.com>.
7120 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7124 *) [In 0.9.6d-engine release:]
7125 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7126 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7128 *) Add the configuration target linux-s390x.
7129 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7131 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7132 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7133 variable as an indication that a ClientHello message has been
7134 received. As the flag value will be lost between multiple
7135 invocations of ssl3_accept when using non-blocking I/O, the
7136 function may not be aware that a handshake has actually taken
7137 place, thus preventing a new session from being added to the
7140 To avoid this problem, we now set s->new_session to 2 instead of
7141 using a local variable.
7142 [Lutz Jaenicke, Bodo Moeller]
7144 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7145 if the SSL_R_LENGTH_MISMATCH error is detected.
7146 [Geoff Thorpe, Bodo Moeller]
7148 *) New 'shared_ldflag' column in Configure platform table.
7151 *) Fix EVP_CIPHER_mode macro.
7152 ["Dan S. Camper" <dan@bti.net>]
7154 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7155 type, we must throw them away by setting rr->length to 0.
7156 [D P Chang <dpc@qualys.com>]
7158 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7160 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7161 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7162 worked incorrectly for those cases where range = 10..._2 and
7163 3*range is two bits longer than range.)
7166 *) Only add signing time to PKCS7 structures if it is not already
7170 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7171 OBJ_ld_ce should be OBJ_id_ce.
7172 Also some ip-pda OIDs in crypto/objects/objects.txt were
7173 incorrect (cf. RFC 3039).
7174 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7176 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7177 returns early because it has nothing to do.
7178 [Andy Schneider <andy.schneider@bjss.co.uk>]
7180 *) [In 0.9.6c-engine release:]
7181 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7182 [Andy Schneider <andy.schneider@bjss.co.uk>]
7184 *) [In 0.9.6c-engine release:]
7185 Add support for Cryptographic Appliance's keyserver technology.
7186 (Use engine 'keyclient')
7187 [Cryptographic Appliances and Geoff Thorpe]
7189 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7190 is called via tools/c89.sh because arguments have to be
7191 rearranged (all '-L' options must appear before the first object
7193 [Richard Shapiro <rshapiro@abinitio.com>]
7195 *) [In 0.9.6c-engine release:]
7196 Add support for Broadcom crypto accelerator cards, backported
7198 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7200 *) [In 0.9.6c-engine release:]
7201 Add support for SureWare crypto accelerator cards from
7202 Baltimore Technologies. (Use engine 'sureware')
7203 [Baltimore Technologies and Mark Cox]
7205 *) [In 0.9.6c-engine release:]
7206 Add support for crypto accelerator cards from Accelerated
7207 Encryption Processing, www.aep.ie. (Use engine 'aep')
7208 [AEP Inc. and Mark Cox]
7210 *) Add a configuration entry for gcc on UnixWare.
7211 [Gary Benson <gbenson@redhat.com>]
7213 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7214 messages are stored in a single piece (fixed-length part and
7215 variable-length part combined) and fix various bugs found on the way.
7218 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7219 instead. BIO_gethostbyname() does not know what timeouts are
7220 appropriate, so entries would stay in cache even when they have
7222 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7224 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7225 faced with a pathologically small ClientHello fragment that does
7226 not contain client_version: Instead of aborting with an error,
7227 simply choose the highest available protocol version (i.e.,
7228 TLS 1.0 unless it is disabled). In practice, ClientHello
7229 messages are never sent like this, but this change gives us
7230 strictly correct behaviour at least for TLS.
7233 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7234 never resets s->method to s->ctx->method when called from within
7235 one of the SSL handshake functions.
7236 [Bodo Moeller; problem pointed out by Niko Baric]
7238 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7239 (sent using the client's version number) if client_version is
7240 smaller than the protocol version in use. Also change
7241 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7242 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7243 the client will at least see that alert.
7246 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7250 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7251 client receives HelloRequest while in a handshake.
7252 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7254 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7255 should end in 'break', not 'goto end' which circumvents various
7256 cleanups done in state SSL_ST_OK. But session related stuff
7257 must be disabled for SSL_ST_OK in the case that we just sent a
7260 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7261 before just sending a HelloRequest.
7262 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7264 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7265 reveal whether illegal block cipher padding was found or a MAC
7266 verification error occured. (Neither SSLerr() codes nor alerts
7267 are directly visible to potential attackers, but the information
7268 may leak via logfiles.)
7270 Similar changes are not required for the SSL 2.0 implementation
7271 because the number of padding bytes is sent in clear for SSL 2.0,
7272 and the extra bytes are just ignored. However ssl/s2_pkt.c
7273 failed to verify that the purported number of padding bytes is in
7277 *) Add OpenUNIX-8 support including shared libraries
7278 (Boyd Lynn Gerber <gerberb@zenez.com>).
7281 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7282 'wristwatch attack' using huge encoding parameters (cf.
7283 James H. Manger's CRYPTO 2001 paper). Note that the
7284 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7285 encoding parameters and hence was not vulnerable.
7288 *) BN_sqr() bug fix.
7289 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7291 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7292 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7293 followed by modular reduction.
7294 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7296 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7297 equivalent based on BN_pseudo_rand() instead of BN_rand().
7300 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7301 This function was broken, as the check for a new client hello message
7302 to handle SGC did not allow these large messages.
7303 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7306 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7309 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7310 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7313 *) Rework the configuration and shared library support for Tru64 Unix.
7314 The configuration part makes use of modern compiler features and
7315 still retains old compiler behavior for those that run older versions
7316 of the OS. The shared library support part includes a variant that
7317 uses the RPATH feature, and is available through the special
7318 configuration target "alpha-cc-rpath", which will never be selected
7320 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7322 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7323 with the same message size as in ssl3_get_certificate_request().
7324 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7325 messages might inadvertently be reject as too long.
7326 [Petr Lampa <lampa@fee.vutbr.cz>]
7328 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7331 *) Modified SSL library such that the verify_callback that has been set
7332 specificly for an SSL object with SSL_set_verify() is actually being
7333 used. Before the change, a verify_callback set with this function was
7334 ignored and the verify_callback() set in the SSL_CTX at the time of
7335 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7336 to allow the necessary settings.
7339 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7340 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7341 done automatically (in contradiction to the requirements of the C
7342 standard). This made problems when used from OpenSSH.
7345 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7346 dh->length and always used
7348 BN_rand_range(priv_key, dh->p).
7350 BN_rand_range() is not necessary for Diffie-Hellman, and this
7351 specific range makes Diffie-Hellman unnecessarily inefficient if
7352 dh->length (recommended exponent length) is much smaller than the
7353 length of dh->p. We could use BN_rand_range() if the order of
7354 the subgroup was stored in the DH structure, but we only have
7359 BN_rand(priv_key, l, ...)
7361 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7367 RSA_eay_public_encrypt
7368 RSA_eay_private_decrypt
7369 RSA_eay_private_encrypt (signing)
7370 RSA_eay_public_decrypt (signature verification)
7372 (default implementations for RSA_public_encrypt,
7373 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7374 always reject numbers >= n.
7377 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7378 to synchronize access to 'locking_thread'. This is necessary on
7379 systems where access to 'locking_thread' (an 'unsigned long'
7380 variable) is not atomic.
7383 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7384 *before* setting the 'crypto_lock_rand' flag. The previous code had
7385 a race condition if 0 is a valid thread ID.
7386 [Travis Vitek <vitek@roguewave.com>]
7388 *) Add support for shared libraries under Irix.
7389 [Albert Chin-A-Young <china@thewrittenword.com>]
7391 *) Add configuration option to build on Linux on both big-endian and
7393 [Ralf Baechle <ralf@uni-koblenz.de>]
7395 *) Add the possibility to create shared libraries on HP-UX.
7398 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7400 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7401 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7402 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7403 PRNG state recovery was possible based on the output of
7404 one PRNG request appropriately sized to gain knowledge on
7405 'md' followed by enough consecutive 1-byte PRNG requests
7406 to traverse all of 'state'.
7408 1. When updating 'md_local' (the current thread's copy of 'md')
7409 during PRNG output generation, hash all of the previous
7410 'md_local' value, not just the half used for PRNG output.
7412 2. Make the number of bytes from 'state' included into the hash
7413 independent from the number of PRNG bytes requested.
7415 The first measure alone would be sufficient to avoid
7416 Markku-Juhani's attack. (Actually it had never occurred
7417 to me that the half of 'md_local' used for chaining was the
7418 half from which PRNG output bytes were taken -- I had always
7419 assumed that the secret half would be used.) The second
7420 measure makes sure that additional data from 'state' is never
7421 mixed into 'md_local' in small portions; this heuristically
7422 further strengthens the PRNG.
7425 *) Fix crypto/bn/asm/mips3.s.
7428 *) When only the key is given to "enc", the IV is undefined. Print out
7429 an error message in this case.
7432 *) Handle special case when X509_NAME is empty in X509 printing routines.
7435 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7436 positive and less than q.
7439 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7440 used: it isn't thread safe and the add_lock_callback should handle
7442 [Paul Rose <Paul.Rose@bridge.com>]
7444 *) Verify that incoming data obeys the block size in
7445 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7449 [Ulf Möller, Bodo Möller]
7451 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7452 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7453 when fixing the server behaviour for backwards-compatible 'client
7454 hello' messages. (Note that the attack is impractical against
7455 SSL 3.0 and TLS 1.0 anyway because length and version checking
7456 means that the probability of guessing a valid ciphertext is
7457 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7460 Before 0.9.5, the countermeasure (hide the error by generating a
7461 random 'decryption result') did not work properly because
7462 ERR_clear_error() was missing, meaning that SSL_get_error() would
7463 detect the supposedly ignored error.
7465 Both problems are now fixed.
7468 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7469 (previously it was 1024).
7472 *) Fix for compatibility mode trust settings: ignore trust settings
7473 unless some valid trust or reject settings are present.
7476 *) Fix for blowfish EVP: its a variable length cipher.
7479 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7480 parameters in DSA public key structures and return an error in the
7481 DSA routines if parameters are absent.
7484 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7485 in the current directory if neither $RANDFILE nor $HOME was set.
7486 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7487 caused some confusion to Windows users who haven't defined $HOME.
7488 Thus RAND_file_name() is changed again: e_os.h can define a
7489 DEFAULT_HOME, which will be used if $HOME is not set.
7490 For Windows, we use "C:"; on other platforms, we still require
7491 environment variables.
7493 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7494 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7495 having multiple threads call RAND_poll() concurrently.
7498 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7499 combination of a flag and a thread ID variable.
7500 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7501 flag), *other* threads can enter ssleay_add_bytes without obeying
7502 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7503 that they do not hold after the first thread unsets add_do_not_lock).
7506 *) Change bctest again: '-x' expressions are not available in all
7510 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7512 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7513 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7515 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7516 the default extension for executables, if any. Also, make the perl
7517 scripts that use symlink() to test if it really exists and use "cp"
7518 if it doesn't. All this made OpenSSL compilable and installable in
7522 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7523 If SEQUENCE is length is indefinite just set c->slen to the total
7524 amount of data available.
7525 [Steve Henson, reported by shige@FreeBSD.org]
7526 [This change does not apply to 0.9.7.]
7528 *) Change bctest to avoid here-documents inside command substitution
7529 (workaround for FreeBSD /bin/sh bug).
7530 For compatibility with Ultrix, avoid shell functions (introduced
7531 in the bctest version that searches along $PATH).
7534 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7535 with des_encrypt() defined on some operating systems, like Solaris
7539 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7540 On the Importance of Eliminating Errors in Cryptographic
7541 Computations, J. Cryptology 14 (2001) 2, 101-119,
7542 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7545 *) MIPS assembler BIGNUM division bug fix.
7548 *) Disabled incorrect Alpha assembler code.
7551 *) Fix PKCS#7 decode routines so they correctly update the length
7552 after reading an EOC for the EXPLICIT tag.
7554 [This change does not apply to 0.9.7.]
7556 *) Fix bug in PKCS#12 key generation routines. This was triggered
7557 if a 3DES key was generated with a 0 initial byte. Include
7558 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7559 (but broken) behaviour.
7562 *) Enhance bctest to search for a working bc along $PATH and print
7564 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7566 *) Fix memory leaks in err.c: free err_data string if necessary;
7567 don't write to the wrong index in ERR_set_error_data.
7570 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7574 *) Replace rdtsc with _emit statements for VC++ version 5.
7575 [Jeremy Cooper <jeremy@baymoo.org>]
7577 *) Make it possible to reuse SSLv2 sessions.
7580 *) In copy_email() check for >= 0 as a return value for
7581 X509_NAME_get_index_by_NID() since 0 is a valid index.
7582 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7584 *) Avoid coredump with unsupported or invalid public keys by checking if
7585 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7586 PKCS7_verify() fails with non detached data.
7589 *) Don't use getenv in library functions when run as setuid/setgid.
7590 New function OPENSSL_issetugid().
7593 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7594 due to incorrect handling of multi-threading:
7596 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7598 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7600 3. Count how many times MemCheck_off() has been called so that
7601 nested use can be treated correctly. This also avoids
7602 inband-signalling in the previous code (which relied on the
7603 assumption that thread ID 0 is impossible).
7606 *) Add "-rand" option also to s_client and s_server.
7609 *) Fix CPU detection on Irix 6.x.
7610 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7611 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7613 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7616 [This change does not apply to 0.9.7.]
7618 *) Use the cached encoding of an X509_NAME structure rather than
7619 copying it. This is apparently the reason for the libsafe "errors"
7620 but the code is actually correct.
7623 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7624 Bleichenbacher's DSA attack.
7625 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7626 to be set and top=0 forces the highest bit to be set; top=-1 is new
7627 and leaves the highest bit random.
7628 [Ulf Moeller, Bodo Moeller]
7630 *) In the NCONF_...-based implementations for CONF_... queries
7631 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7632 a temporary CONF structure with the data component set to NULL
7633 (which gives segmentation faults in lh_retrieve).
7634 Instead, use NULL for the CONF pointer in CONF_get_string and
7635 CONF_get_number (which may use environment variables) and directly
7636 return NULL from CONF_get_section.
7639 *) Fix potential buffer overrun for EBCDIC.
7642 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7643 keyUsage if basicConstraints absent for a CA.
7646 *) Make SMIME_write_PKCS7() write mail header values with a format that
7647 is more generally accepted (no spaces before the semicolon), since
7648 some programs can't parse those values properly otherwise. Also make
7649 sure BIO's that break lines after each write do not create invalid
7653 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7654 macros previously used would not encode an empty SEQUENCE OF
7655 and break the signature.
7657 [This change does not apply to 0.9.7.]
7659 *) Zero the premaster secret after deriving the master secret in
7663 *) Add some EVP_add_digest_alias registrations (as found in
7664 OpenSSL_add_all_digests()) to SSL_library_init()
7665 aka OpenSSL_add_ssl_algorithms(). This provides improved
7666 compatibility with peers using X.509 certificates
7667 with unconventional AlgorithmIdentifier OIDs.
7670 *) Fix for Irix with NO_ASM.
7671 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7673 *) ./config script fixes.
7674 [Ulf Moeller, Richard Levitte]
7676 *) Fix 'openssl passwd -1'.
7679 *) Change PKCS12_key_gen_asc() so it can cope with non null
7680 terminated strings whose length is passed in the passlen
7681 parameter, for example from PEM callbacks. This was done
7682 by adding an extra length parameter to asc2uni().
7683 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7685 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7686 call failed, free the DSA structure.
7689 *) Fix to uni2asc() to cope with zero length Unicode strings.
7690 These are present in some PKCS#12 files.
7693 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7694 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7695 when writing a 32767 byte record.
7696 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7698 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7699 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7701 (RSA objects have a reference count access to which is protected
7702 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7703 so they are meant to be shared between threads.)
7704 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7705 "Reddie, Steven" <Steven.Reddie@ca.com>]
7707 *) Fix a deadlock in CRYPTO_mem_leaks().
7710 *) Use better test patterns in bntest.
7713 *) rand_win.c fix for Borland C.
7716 *) BN_rshift bugfix for n == 0.
7719 *) Add a 'bctest' script that checks for some known 'bc' bugs
7720 so that 'make test' does not abort just because 'bc' is broken.
7723 *) Store verify_result within SSL_SESSION also for client side to
7724 avoid potential security hole. (Re-used sessions on the client side
7725 always resulted in verify_result==X509_V_OK, not using the original
7726 result of the server certificate verification.)
7729 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7730 SSL3_RT_APPLICATION_DATA, return 0.
7731 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7735 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7736 releases, have been re-implemented by renaming the previous
7737 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7738 and ssl3_read_internal, respectively, and adding 'peek' parameters
7739 to them. The new ssl[23]_{read,peek} functions are calls to
7740 ssl[23]_read_internal with the 'peek' flag set appropriately.
7741 A 'peek' parameter has also been added to ssl3_read_bytes, which
7742 does the actual work for ssl3_read_internal.
7745 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7746 the method-specific "init()" handler. Also clean up ex_data after
7747 calling the method-specific "finish()" handler. Previously, this was
7748 happening the other way round.
7751 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7752 The previous value, 12, was not always sufficient for BN_mod_exp().
7755 *) Make sure that shared libraries get the internal name engine with
7756 the full version number and not just 0. This should mark the
7757 shared libraries as not backward compatible. Of course, this should
7758 be changed again when we can guarantee backward binary compatibility.
7761 *) Fix typo in get_cert_by_subject() in by_dir.c
7762 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7764 *) Rework the system to generate shared libraries:
7766 - Make note of the expected extension for the shared libraries and
7767 if there is a need for symbolic links from for example libcrypto.so.0
7768 to libcrypto.so.0.9.7. There is extended info in Configure for
7771 - Make as few rebuilds of the shared libraries as possible.
7773 - Still avoid linking the OpenSSL programs with the shared libraries.
7775 - When installing, install the shared libraries separately from the
7779 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7781 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7782 and not in SSL_clear because the latter is also used by the
7783 accept/connect functions; previously, the settings made by
7784 SSL_set_read_ahead would be lost during the handshake.
7785 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7787 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7788 Previously, it would create entries for disabled algorithms no
7792 *) Added several new manual pages for SSL_* function.
7795 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7797 *) In ssl23_get_client_hello, generate an error message when faced
7798 with an initial SSL 3.0/TLS record that is too small to contain the
7799 first two bytes of the ClientHello message, i.e. client_version.
7800 (Note that this is a pathologic case that probably has never happened
7801 in real life.) The previous approach was to use the version number
7802 from the record header as a substitute; but our protocol choice
7803 should not depend on that one because it is not authenticated
7804 by the Finished messages.
7807 *) More robust randomness gathering functions for Windows.
7808 [Jeffrey Altman <jaltman@columbia.edu>]
7810 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7811 not set then we don't setup the error code for issuer check errors
7812 to avoid possibly overwriting other errors which the callback does
7813 handle. If an application does set the flag then we assume it knows
7814 what it is doing and can handle the new informational codes
7818 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7819 a general "ANY" type, as such it should be able to decode anything
7820 including tagged types. However it didn't check the class so it would
7821 wrongly interpret tagged types in the same way as their universal
7822 counterpart and unknown types were just rejected. Changed so that the
7823 tagged and unknown types are handled in the same way as a SEQUENCE:
7824 that is the encoding is stored intact. There is also a new type
7825 "V_ASN1_OTHER" which is used when the class is not universal, in this
7826 case we have no idea what the actual type is so we just lump them all
7830 *) On VMS, stdout may very well lead to a file that is written to
7831 in a record-oriented fashion. That means that every write() will
7832 write a separate record, which will be read separately by the
7833 programs trying to read from it. This can be very confusing.
7835 The solution is to put a BIO filter in the way that will buffer
7836 text until a linefeed is reached, and then write everything a
7837 line at a time, so every record written will be an actual line,
7838 not chunks of lines and not (usually doesn't happen, but I've
7839 seen it once) several lines in one record. BIO_f_linebuffer() is
7842 Currently, it's a VMS-only method, because that's where it has
7843 been tested well enough.
7846 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7847 it can return incorrect results.
7848 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7849 but it was in 0.9.6-beta[12].)
7852 *) Disable the check for content being present when verifying detached
7853 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7854 include zero length content when signing messages.
7857 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7858 BIO_ctrl (for BIO pairs).
7861 *) Add DSO method for VMS.
7864 *) Bug fix: Montgomery multiplication could produce results with the
7868 *) Add RPM specification openssl.spec and modify it to build three
7869 packages. The default package contains applications, application
7870 documentation and run-time libraries. The devel package contains
7871 include files, static libraries and function documentation. The
7872 doc package contains the contents of the doc directory. The original
7873 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7876 *) Add a large number of documentation files for many SSL routines.
7877 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7879 *) Add a configuration entry for Sony News 4.
7880 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7882 *) Don't set the two most significant bits to one when generating a
7883 random number < q in the DSA library.
7886 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7887 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7888 the underlying transport is blocking) if a handshake took place.
7889 (The default behaviour is needed by applications such as s_client
7890 and s_server that use select() to determine when to use SSL_read;
7891 but for applications that know in advance when to expect data, it
7892 just makes things more complicated.)
7895 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7899 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7900 work better on such systems.
7901 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7903 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7904 Update PKCS12_parse() so it copies the friendlyName and the
7905 keyid to the certificates aux info.
7908 *) Fix bug in PKCS7_verify() which caused an infinite loop
7909 if there was more than one signature.
7910 [Sven Uszpelkat <su@celocom.de>]
7912 *) Major change in util/mkdef.pl to include extra information
7913 about each symbol, as well as presenting variables as well
7914 as functions. This change means that there's n more need
7915 to rebuild the .num files when some algorithms are excluded.
7918 *) Allow the verify time to be set by an application,
7919 rather than always using the current time.
7922 *) Phase 2 verify code reorganisation. The certificate
7923 verify code now looks up an issuer certificate by a
7924 number of criteria: subject name, authority key id
7925 and key usage. It also verifies self signed certificates
7926 by the same criteria. The main comparison function is
7927 X509_check_issued() which performs these checks.
7929 Lot of changes were necessary in order to support this
7930 without completely rewriting the lookup code.
7932 Authority and subject key identifier are now cached.
7934 The LHASH 'certs' is X509_STORE has now been replaced
7935 by a STACK_OF(X509_OBJECT). This is mainly because an
7936 LHASH can't store or retrieve multiple objects with
7937 the same hash value.
7939 As a result various functions (which were all internal
7940 use only) have changed to handle the new X509_STORE
7941 structure. This will break anything that messed round
7942 with X509_STORE internally.
7944 The functions X509_STORE_add_cert() now checks for an
7945 exact match, rather than just subject name.
7947 The X509_STORE API doesn't directly support the retrieval
7948 of multiple certificates matching a given criteria, however
7949 this can be worked round by performing a lookup first
7950 (which will fill the cache with candidate certificates)
7951 and then examining the cache for matches. This is probably
7952 the best we can do without throwing out X509_LOOKUP
7953 entirely (maybe later...).
7955 The X509_VERIFY_CTX structure has been enhanced considerably.
7957 All certificate lookup operations now go via a get_issuer()
7958 callback. Although this currently uses an X509_STORE it
7959 can be replaced by custom lookups. This is a simple way
7960 to bypass the X509_STORE hackery necessary to make this
7961 work and makes it possible to use more efficient techniques
7962 in future. A very simple version which uses a simple
7963 STACK for its trusted certificate store is also provided
7964 using X509_STORE_CTX_trusted_stack().
7966 The verify_cb() and verify() callbacks now have equivalents
7967 in the X509_STORE_CTX structure.
7969 X509_STORE_CTX also has a 'flags' field which can be used
7970 to customise the verify behaviour.
7973 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7974 excludes S/MIME capabilities.
7977 *) When a certificate request is read in keep a copy of the
7978 original encoding of the signed data and use it when outputing
7979 again. Signatures then use the original encoding rather than
7980 a decoded, encoded version which may cause problems if the
7981 request is improperly encoded.
7984 *) For consistency with other BIO_puts implementations, call
7985 buffer_write(b, ...) directly in buffer_puts instead of calling
7988 In BIO_puts, increment b->num_write as in BIO_write.
7989 [Peter.Sylvester@EdelWeb.fr]
7991 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7992 BN_zero, we may not return a BIGNUM with an array consisting of
7996 *) Avoid calling abort() from within the library when problems are
7997 detected, except if preprocessor symbols have been defined
7998 (such as REF_CHECK, BN_DEBUG etc.).
8001 *) New openssl application 'rsautl'. This utility can be
8002 used for low level RSA operations. DER public key
8003 BIO/fp routines also added.
8006 *) New Configure entry and patches for compiling on QNX 4.
8007 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8009 *) A demo state-machine implementation was sponsored by
8010 Nuron (http://www.nuron.com/) and is now available in
8011 demos/state_machine.
8014 *) New options added to the 'dgst' utility for signature
8015 generation and verification.
8018 *) Unrecognized PKCS#7 content types are now handled via a
8019 catch all ASN1_TYPE structure. This allows unsupported
8020 types to be stored as a "blob" and an application can
8021 encode and decode it manually.
8024 *) Fix various signed/unsigned issues to make a_strex.c
8026 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8028 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8029 length if passed a buffer. ASN1_INTEGER_to_BN failed
8030 if passed a NULL BN and its argument was negative.
8031 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8033 *) Modification to PKCS#7 encoding routines to output definite
8034 length encoding. Since currently the whole structures are in
8035 memory there's not real point in using indefinite length
8036 constructed encoding. However if OpenSSL is compiled with
8037 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8040 *) Added BIO_vprintf() and BIO_vsnprintf().
8043 *) Added more prefixes to parse for in the the strings written
8044 through a logging bio, to cover all the levels that are available
8045 through syslog. The prefixes are now:
8047 PANIC, EMERG, EMR => LOG_EMERG
8048 ALERT, ALR => LOG_ALERT
8049 CRIT, CRI => LOG_CRIT
8050 ERROR, ERR => LOG_ERR
8051 WARNING, WARN, WAR => LOG_WARNING
8052 NOTICE, NOTE, NOT => LOG_NOTICE
8053 INFO, INF => LOG_INFO
8054 DEBUG, DBG => LOG_DEBUG
8056 and as before, if none of those prefixes are present at the
8057 beginning of the string, LOG_ERR is chosen.
8059 On Win32, the LOG_* levels are mapped according to this:
8061 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8062 LOG_WARNING => EVENTLOG_WARNING_TYPE
8063 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8067 *) Made it possible to reconfigure with just the configuration
8068 argument "reconf" or "reconfigure". The command line arguments
8069 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8070 and are retrieved from there when reconfiguring.
8074 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8076 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8079 *) The obj_dat.pl script was messing up the sorting of object
8080 names. The reason was that it compared the quoted version
8081 of strings as a result "OCSP" > "OCSP Signing" because
8082 " > SPACE. Changed script to store unquoted versions of
8083 names and add quotes on output. It was also omitting some
8084 names from the lookup table if they were given a default
8085 value (that is if SN is missing it is given the same
8086 value as LN and vice versa), these are now added on the
8087 grounds that if an object has a name we should be able to
8088 look it up. Finally added warning output when duplicate
8089 short or long names are found.
8092 *) Changes needed for Tandem NSK.
8093 [Scott Uroff <scott@xypro.com>]
8095 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8096 RSA_padding_check_SSLv23(), special padding was never detected
8097 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8098 version rollback attacks was not effective.
8100 In s23_clnt.c, don't use special rollback-attack detection padding
8101 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8102 client; similarly, in s23_srvr.c, don't do the rollback check if
8103 SSL 2.0 is the only protocol enabled in the server.
8106 *) Make it possible to get hexdumps of unprintable data with 'openssl
8107 asn1parse'. By implication, the functions ASN1_parse_dump() and
8108 BIO_dump_indent() are added.
8111 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8112 these print out strings and name structures based on various
8113 flags including RFC2253 support and proper handling of
8114 multibyte characters. Added options to the 'x509' utility
8115 to allow the various flags to be set.
8118 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8119 Also change the functions X509_cmp_current_time() and
8120 X509_gmtime_adj() work with an ASN1_TIME structure,
8121 this will enable certificates using GeneralizedTime in validity
8122 dates to be checked.
8125 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8126 negative public key encodings) on by default,
8127 NO_NEG_PUBKEY_BUG can be set to disable it.
8130 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8131 content octets. An i2c_ASN1_OBJECT is unnecessary because
8132 the encoding can be trivially obtained from the structure.
8135 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8136 not read locks (CRYPTO_r_[un]lock).
8139 *) A first attempt at creating official support for shared
8140 libraries through configuration. I've kept it so the
8141 default is static libraries only, and the OpenSSL programs
8142 are always statically linked for now, but there are
8143 preparations for dynamic linking in place.
8144 This has been tested on Linux and Tru64.
8147 *) Randomness polling function for Win9x, as described in:
8148 Peter Gutmann, Software Generation of Practically Strong
8152 *) Fix so PRNG is seeded in req if using an already existing
8156 *) New options to smime application. -inform and -outform
8157 allow alternative formats for the S/MIME message including
8158 PEM and DER. The -content option allows the content to be
8159 specified separately. This should allow things like Netscape
8160 form signing output easier to verify.
8163 *) Fix the ASN1 encoding of tags using the 'long form'.
8166 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8167 STRING types. These convert content octets to and from the
8168 underlying type. The actual tag and length octets are
8169 already assumed to have been read in and checked. These
8170 are needed because all other string types have virtually
8171 identical handling apart from the tag. By having versions
8172 of the ASN1 functions that just operate on content octets
8173 IMPLICIT tagging can be handled properly. It also allows
8174 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8175 and ASN1_INTEGER are identical apart from the tag.
8178 *) Change the handling of OID objects as follows:
8180 - New object identifiers are inserted in objects.txt, following
8181 the syntax given in objects.README.
8182 - objects.pl is used to process obj_mac.num and create a new
8184 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8187 This is currently kind of a hack, and the perl code in objects.pl
8188 isn't very elegant, but it works as I intended. The simplest way
8189 to check that it worked correctly is to look in obj_dat.h and
8190 check the array nid_objs and make sure the objects haven't moved
8191 around (this is important!). Additions are OK, as well as
8192 consistent name changes.
8195 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8198 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8199 The given file adds to whatever has already been seeded into the
8200 random pool through the RANDFILE configuration file option or
8201 environment variable, or the default random state file.
8204 *) mkstack.pl now sorts each macro group into lexical order.
8205 Previously the output order depended on the order the files
8206 appeared in the directory, resulting in needless rewriting
8210 *) Patches to make OpenSSL compile under Win32 again. Mostly
8211 work arounds for the VC++ problem that it treats func() as
8212 func(void). Also stripped out the parts of mkdef.pl that
8213 added extra typesafe functions: these no longer exist.
8216 *) Reorganisation of the stack code. The macros are now all
8217 collected in safestack.h . Each macro is defined in terms of
8218 a "stack macro" of the form SKM_<name>(type, a, b). The
8219 DEBUG_SAFESTACK is now handled in terms of function casts,
8220 this has the advantage of retaining type safety without the
8221 use of additional functions. If DEBUG_SAFESTACK is not defined
8222 then the non typesafe macros are used instead. Also modified the
8223 mkstack.pl script to handle the new form. Needs testing to see
8224 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8225 the default if no major problems. Similar behaviour for ASN1_SET_OF
8226 and PKCS12_STACK_OF.
8229 *) When some versions of IIS use the 'NET' form of private key the
8230 key derivation algorithm is different. Normally MD5(password) is
8231 used as a 128 bit RC4 key. In the modified case
8232 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8233 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8234 as the old Netscape_RSA functions except they have an additional
8235 'sgckey' parameter which uses the modified algorithm. Also added
8236 an -sgckey command line option to the rsa utility. Thanks to
8237 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8238 algorithm to openssl-dev.
8241 *) The evp_local.h macros were using 'c.##kname' which resulted in
8242 invalid expansion on some systems (SCO 5.0.5 for example).
8243 Corrected to 'c.kname'.
8244 [Phillip Porch <root@theporch.com>]
8246 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8247 a STACK of email addresses from a certificate or request, these look
8248 in the subject name and the subject alternative name extensions and
8249 omit any duplicate addresses.
8252 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8253 This makes DSA verification about 2 % faster.
8256 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8257 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8258 plus overhead for 1024 bit moduli).
8259 This makes exponentiations about 0.5 % faster for 1024 bit
8260 exponents (as measured by "openssl speed rsa2048").
8263 *) Rename memory handling macros to avoid conflicts with other
8265 Malloc => OPENSSL_malloc
8266 Malloc_locked => OPENSSL_malloc_locked
8267 Realloc => OPENSSL_realloc
8268 Free => OPENSSL_free
8271 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8272 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8275 *) CygWin32 support.
8276 [John Jarvie <jjarvie@newsguy.com>]
8278 *) The type-safe stack code has been rejigged. It is now only compiled
8279 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8280 by default all type-specific stack functions are "#define"d back to
8281 standard stack functions. This results in more streamlined output
8282 but retains the type-safety checking possibilities of the original
8286 *) The STACK code has been cleaned up, and certain type declarations
8287 that didn't make a lot of sense have been brought in line. This has
8288 also involved a cleanup of sorts in safestack.h to more correctly
8289 map type-safe stack functions onto their plain stack counterparts.
8290 This work has also resulted in a variety of "const"ifications of
8291 lots of the code, especially "_cmp" operations which should normally
8292 be prototyped with "const" parameters anyway.
8295 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8296 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8297 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8298 where all of 'md' is used each time the PRNG is used, but 'state'
8299 is used only indexed by a cyclic counter. As entropy may not be
8300 well distributed from the beginning, 'md' is important as a
8301 chaining variable. However, the output function chains only half
8302 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8303 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8304 in all of 'state' being rewritten, with the new values depending
8305 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8308 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8309 the handshake is continued after ssl_verify_cert_chain();
8310 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8311 can lead to 'unexplainable' connection aborts later.
8312 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8314 *) Major EVP API cipher revision.
8315 Add hooks for extra EVP features. This allows various cipher
8316 parameters to be set in the EVP interface. Support added for variable
8317 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8318 setting of RC2 and RC5 parameters.
8320 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8323 Remove lots of duplicated code from the EVP library. For example *every*
8324 cipher init() function handles the 'iv' in the same way according to the
8325 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8326 for CFB and OFB modes they zero ctx->num.
8328 New functionality allows removal of S/MIME code RC2 hack.
8330 Most of the routines have the same form and so can be declared in terms
8333 By shifting this to the top level EVP_CipherInit() it can be removed from
8334 all individual ciphers. If the cipher wants to handle IVs or keys
8335 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8338 Change lots of functions like EVP_EncryptUpdate() to now return a
8339 value: although software versions of the algorithms cannot fail
8340 any installed hardware versions can.
8343 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8344 this option is set, tolerate broken clients that send the negotiated
8345 protocol version number instead of the requested protocol version
8349 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8350 i.e. non-zero for export ciphersuites, zero otherwise.
8351 Previous versions had this flag inverted, inconsistent with
8352 rsa_tmp_cb (..._TMP_RSA_CB).
8353 [Bodo Moeller; problem reported by Amit Chopra]
8355 *) Add missing DSA library text string. Work around for some IIS
8356 key files with invalid SEQUENCE encoding.
8359 *) Add a document (doc/standards.txt) that list all kinds of standards
8360 and so on that are implemented in OpenSSL.
8363 *) Enhance c_rehash script. Old version would mishandle certificates
8364 with the same subject name hash and wouldn't handle CRLs at all.
8365 Added -fingerprint option to crl utility, to support new c_rehash
8369 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8372 *) Fix for SSL server purpose checking. Server checking was
8373 rejecting certificates which had extended key usage present
8374 but no ssl client purpose.
8375 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8377 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8378 is a little unclear about how a blank password is handled.
8379 Since the password in encoded as a BMPString with terminating
8380 double NULL a zero length password would end up as just the
8381 double NULL. However no password at all is different and is
8382 handled differently in the PKCS#12 key generation code. NS
8383 treats a blank password as zero length. MSIE treats it as no
8384 password on export: but it will try both on import. We now do
8385 the same: PKCS12_parse() tries zero length and no password if
8386 the password is set to "" or NULL (NULL is now a valid password:
8387 it wasn't before) as does the pkcs12 application.
8390 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8391 perror when PEM_read_bio_X509_REQ fails, the error message must
8392 be obtained from the error queue.
8395 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8396 it in ERR_remove_state if appropriate, and change ERR_get_state
8397 accordingly to avoid race conditions (this is necessary because
8398 thread_hash is no longer constant once set).
8401 *) Bugfix for linux-elf makefile.one.
8404 *) RSA_get_default_method() will now cause a default
8405 RSA_METHOD to be chosen if one doesn't exist already.
8406 Previously this was only set during a call to RSA_new()
8407 or RSA_new_method(NULL) meaning it was possible for
8408 RSA_get_default_method() to return NULL.
8411 *) Added native name translation to the existing DSO code
8412 that will convert (if the flag to do so is set) filenames
8413 that are sufficiently small and have no path information
8414 into a canonical native form. Eg. "blah" converted to
8415 "libblah.so" or "blah.dll" etc.
8418 *) New function ERR_error_string_n(e, buf, len) which is like
8419 ERR_error_string(e, buf), but writes at most 'len' bytes
8420 including the 0 terminator. For ERR_error_string_n, 'buf'
8422 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8424 *) CONF library reworked to become more general. A new CONF
8425 configuration file reader "class" is implemented as well as a
8426 new functions (NCONF_*, for "New CONF") to handle it. The now
8427 old CONF_* functions are still there, but are reimplemented to
8428 work in terms of the new functions. Also, a set of functions
8429 to handle the internal storage of the configuration data is
8430 provided to make it easier to write new configuration file
8431 reader "classes" (I can definitely see something reading a
8432 configuration file in XML format, for example), called _CONF_*,
8433 or "the configuration storage API"...
8435 The new configuration file reading functions are:
8437 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8438 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8440 NCONF_default, NCONF_WIN32
8442 NCONF_dump_fp, NCONF_dump_bio
8444 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8445 NCONF_new creates a new CONF object. This works in the same way
8446 as other interfaces in OpenSSL, like the BIO interface.
8447 NCONF_dump_* dump the internal storage of the configuration file,
8448 which is useful for debugging. All other functions take the same
8449 arguments as the old CONF_* functions wth the exception of the
8450 first that must be a `CONF *' instead of a `LHASH *'.
8452 To make it easer to use the new classes with the old CONF_* functions,
8453 the function CONF_set_default_method is provided.
8456 *) Add '-tls1' option to 'openssl ciphers', which was already
8457 mentioned in the documentation but had not been implemented.
8458 (This option is not yet really useful because even the additional
8459 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8462 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8463 OpenSSL-based applications) load shared libraries and bind to
8464 them in a portable way.
8465 [Geoff Thorpe, with contributions from Richard Levitte]
8467 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8469 *) Make sure _lrotl and _lrotr are only used with MSVC.
8471 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8472 (the default implementation of RAND_status).
8474 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8475 to '-clrext' (= clear extensions), as intended and documented.
8476 [Bodo Moeller; inconsistency pointed out by Michael Attili
8479 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8480 was larger than the MD block size.
8481 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8483 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8484 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8485 using the passed key: if the passed key was a private key the result
8486 of X509_print(), for example, would be to print out all the private key
8490 *) des_quad_cksum() byte order bug fix.
8491 [Ulf Möller, using the problem description in krb4-0.9.7, where
8492 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8494 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8496 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8498 *) For easily testing in shell scripts whether some command
8499 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8500 returns with exit code 0 iff no command of the given name is available.
8501 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8502 the output goes to stdout and nothing is printed to stderr.
8503 Additional arguments are always ignored.
8505 Since for each cipher there is a command of the same name,
8506 the 'no-cipher' compilation switches can be tested this way.
8508 ('openssl no-XXX' is not able to detect pseudo-commands such
8509 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8512 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8515 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8516 is set; it will be thrown away anyway because each handshake creates
8518 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8519 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8520 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8521 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8524 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8525 'Q' and 'R' lose their special meanings (quit/renegotiate).
8526 This is part of what -quiet does; unlike -quiet, -ign_eof
8527 does not suppress any output.
8530 *) Add compatibility options to the purpose and trust code. The
8531 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8532 accepts a certificate or CA, this was the previous behaviour,
8533 with all the associated security issues.
8535 X509_TRUST_COMPAT is the old trust behaviour: only and
8536 automatically trust self signed roots in certificate store. A
8537 new trust setting X509_TRUST_DEFAULT is used to specify that
8538 a purpose has no associated trust setting and it should instead
8539 use the value in the default purpose.
8542 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8543 and fix a memory leak.
8546 *) In util/mkerr.pl (which implements 'make errors'), preserve
8547 reason strings from the previous version of the .c file, as
8548 the default to have only downcase letters (and digits) in
8549 automatically generated reasons codes is not always appropriate.
8552 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8553 using strerror. Previously, ERR_reason_error_string() returned
8554 library names as reason strings for SYSerr; but SYSerr is a special
8555 case where small numbers are errno values, not library numbers.
8558 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8559 converts DSA parameters into DH parameters. (When creating parameters,
8560 DSA_generate_parameters is used.)
8563 *) Include 'length' (recommended exponent length) in C code generated
8564 by 'openssl dhparam -C'.
8567 *) The second argument to set_label in perlasm was already being used
8568 so couldn't be used as a "file scope" flag. Moved to third argument
8572 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8573 instead of RAND_bytes for encryption IVs and salts.
8576 *) Include RAND_status() into RAND_METHOD instead of implementing
8577 it only for md_rand.c Otherwise replacing the PRNG by calling
8578 RAND_set_rand_method would be impossible.
8581 *) Don't let DSA_generate_key() enter an infinite loop if the random
8582 number generation fails.
8585 *) New 'rand' application for creating pseudo-random output.
8588 *) Added configuration support for Linux/IA64
8589 [Rolf Haberrecker <rolf@suse.de>]
8591 *) Assembler module support for Mingw32.
8594 *) Shared library support for HPUX (in shlib/).
8595 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8597 *) Shared library support for Solaris gcc.
8598 [Lutz Behnke <behnke@trustcenter.de>]
8600 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8602 *) PKCS7_encrypt() was adding text MIME headers twice because they
8603 were added manually and by SMIME_crlf_copy().
8606 *) In bntest.c don't call BN_rand with zero bits argument.
8607 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8609 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8610 case was implemented. This caused BN_div_recp() to fail occasionally.
8613 *) Add an optional second argument to the set_label() in the perl
8614 assembly language builder. If this argument exists and is set
8615 to 1 it signals that the assembler should use a symbol whose
8616 scope is the entire file, not just the current function. This
8617 is needed with MASM which uses the format label:: for this scope.
8618 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8620 *) Change the ASN1 types so they are typedefs by default. Before
8621 almost all types were #define'd to ASN1_STRING which was causing
8622 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8626 *) Change names of new functions to the new get1/get0 naming
8627 convention: After 'get1', the caller owns a reference count
8628 and has to call ..._free; 'get0' returns a pointer to some
8629 data structure without incrementing reference counters.
8630 (Some of the existing 'get' functions increment a reference
8631 counter, some don't.)
8632 Similarly, 'set1' and 'add1' functions increase reference
8633 counters or duplicate objects.
8636 *) Allow for the possibility of temp RSA key generation failure:
8637 the code used to assume it always worked and crashed on failure.
8640 *) Fix potential buffer overrun problem in BIO_printf().
8641 [Ulf Möller, using public domain code by Patrick Powell; problem
8642 pointed out by David Sacerdote <das33@cornell.edu>]
8644 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8645 RAND_egd() and RAND_status(). In the command line application,
8646 the EGD socket can be specified like a seed file using RANDFILE
8650 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8651 Some CAs (e.g. Verisign) distribute certificates in this form.
8654 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8655 list to exclude them. This means that no special compilation option
8656 is needed to use anonymous DH: it just needs to be included in the
8660 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8661 EVP_MD_type. The old functionality is available in a new macro called
8662 EVP_MD_md(). Change code that uses it and update docs.
8665 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8666 where the 'void *' argument is replaced by a function pointer argument.
8667 Previously 'void *' was abused to point to functions, which works on
8668 many platforms, but is not correct. As these functions are usually
8669 called by macros defined in OpenSSL header files, most source code
8670 should work without changes.
8673 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8674 sections with information on -D... compiler switches used for
8675 compiling the library so that applications can see them. To enable
8676 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8677 must be defined. E.g.,
8678 #define OPENSSL_ALGORITHM_DEFINES
8679 #include <openssl/opensslconf.h>
8680 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8681 [Richard Levitte, Ulf and Bodo Möller]
8683 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8687 *) Change the 'other' type in certificate aux info to a STACK_OF
8688 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8689 the required ASN1 format: arbitrary types determined by an OID.
8692 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8693 argument to 'req'. This is not because the function is newer or
8694 better than others it just uses the work 'NEW' in the certificate
8695 request header lines. Some software needs this.
8698 *) Reorganise password command line arguments: now passwords can be
8699 obtained from various sources. Delete the PEM_cb function and make
8700 it the default behaviour: i.e. if the callback is NULL and the
8701 usrdata argument is not NULL interpret it as a null terminated pass
8702 phrase. If usrdata and the callback are NULL then the pass phrase
8703 is prompted for as usual.
8706 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8707 the support is automatically enabled. The resulting binaries will
8708 autodetect the card and use it if present.
8709 [Ben Laurie and Compaq Inc.]
8711 *) Work around for Netscape hang bug. This sends certificate request
8712 and server done in one record. Since this is perfectly legal in the
8713 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8714 the bugs/SSLv3 entry for more info.
8717 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8720 *) Add -rand argument to smime and pkcs12 applications and read/write
8724 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8727 *) Add command line password options to the remaining applications.
8730 *) Bug fix for BN_div_recp() for numerators with an even number of
8734 *) More tests in bntest.c, and changed test_bn output.
8737 *) ./config recognizes MacOS X now.
8740 *) Bug fix for BN_div() when the first words of num and divsor are
8741 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8744 *) Add support for various broken PKCS#8 formats, and command line
8745 options to produce them.
8748 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8749 get temporary BIGNUMs from a BN_CTX.
8752 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8756 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8757 include a #define from the old name to the new. The original intent
8758 was that statically linked binaries could for example just call
8759 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8760 link with digests. This never worked because SSLeay_add_all_digests()
8761 and SSLeay_add_all_ciphers() were in the same source file so calling
8762 one would link with the other. They are now in separate source files.
8765 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8768 *) Use a less unusual form of the Miller-Rabin primality test (it used
8769 a binary algorithm for exponentiation integrated into the Miller-Rabin
8770 loop, our standard modexp algorithms are faster).
8773 *) Support for the EBCDIC character set completed.
8774 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8776 *) Source code cleanups: use const where appropriate, eliminate casts,
8777 use void * instead of char * in lhash.
8780 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8781 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8782 this the server could overwrite ephemeral keys that the client
8786 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8787 using 50 iterations of the Rabin-Miller test.
8789 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8790 iterations of the Rabin-Miller test as required by the appendix
8791 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8792 As BN_is_prime_fasttest includes trial division, DSA parameter
8793 generation becomes much faster.
8795 This implies a change for the callback functions in DSA_is_prime
8796 and DSA_generate_parameters: The callback function is called once
8797 for each positive witness in the Rabin-Miller test, not just
8798 occasionally in the inner loop; and the parameters to the
8799 callback function now provide an iteration count for the outer
8800 loop rather than for the current invocation of the inner loop.
8801 DSA_generate_parameters additionally can call the callback
8802 function with an 'iteration count' of -1, meaning that a
8803 candidate has passed the trial division test (when q is generated
8804 from an application-provided seed, trial division is skipped).
8807 *) New function BN_is_prime_fasttest that optionally does trial
8808 division before starting the Rabin-Miller test and has
8809 an additional BN_CTX * argument (whereas BN_is_prime always
8810 has to allocate at least one BN_CTX).
8811 'callback(1, -1, cb_arg)' is called when a number has passed the
8812 trial division stage.
8815 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8819 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8822 *) New function BN_pseudo_rand().
8825 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8826 bignum version of BN_from_montgomery() with the working code from
8827 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8831 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8832 made it impossible to use the same SSL_SESSION data structure in
8833 SSL2 clients in multiple threads.
8836 *) The return value of RAND_load_file() no longer counts bytes obtained
8837 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8838 to seed the PRNG (previously an explicit byte count was required).
8839 [Ulf Möller, Bodo Möller]
8841 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8842 used (char *) instead of (void *) and had casts all over the place.
8845 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8848 *) Retain source code compatibility for BN_prime_checks macro:
8849 BN_is_prime(..., BN_prime_checks, ...) now uses
8850 BN_prime_checks_for_size to determine the appropriate number of
8851 Rabin-Miller iterations.
8854 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8855 DH_CHECK_P_NOT_SAFE_PRIME.
8856 (Check if this is true? OpenPGP calls them "strong".)
8859 *) Merge the functionality of "dh" and "gendh" programs into a new program
8860 "dhparam". The old programs are retained for now but will handle DH keys
8861 (instead of parameters) in future.
8864 *) Make the ciphers, s_server and s_client programs check the return values
8865 when a new cipher list is set.
8868 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8869 ciphers. Before when the 56bit ciphers were enabled the sorting was
8872 The syntax for the cipher sorting has been extended to support sorting by
8873 cipher-strength (using the strength_bits hard coded in the tables).
8874 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8876 Fix a bug in the cipher-command parser: when supplying a cipher command
8877 string with an "undefined" symbol (neither command nor alphanumeric
8878 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8879 an error is flagged.
8881 Due to the strength-sorting extension, the code of the
8882 ssl_create_cipher_list() function was completely rearranged. I hope that
8883 the readability was also increased :-)
8884 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8886 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8887 for the first serial number and places 2 in the serial number file. This
8888 avoids problems when the root CA is created with serial number zero and
8889 the first user certificate has the same issuer name and serial number
8893 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8894 the new code. Add documentation for this stuff.
8897 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8898 X509_*() to X509at_*() on the grounds that they don't handle X509
8899 structures and behave in an analogous way to the X509v3 functions:
8900 they shouldn't be called directly but wrapper functions should be used
8903 So we also now have some wrapper functions that call the X509at functions
8904 when passed certificate requests. (TO DO: similar things can be done with
8905 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8906 things. Some of these need some d2i or i2d and print functionality
8907 because they handle more complex structures.)
8910 *) Add missing #ifndefs that caused missing symbols when building libssl
8911 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8912 NO_RSA in ssl/s2*.c.
8913 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8915 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8916 has a return value which indicates the quality of the random data
8917 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8918 error queue. New function RAND_pseudo_bytes() generates output that is
8919 guaranteed to be unique but not unpredictable. RAND_add is like
8920 RAND_seed, but takes an extra argument for an entropy estimate
8921 (RAND_seed always assumes full entropy).
8924 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8925 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8926 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8927 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8928 false-positive rate of at most 2^-80 for random input.
8931 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8934 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8935 in the 0.9.5 release), this returns the chain
8936 from an X509_CTX structure with a dup of the stack and all
8937 the X509 reference counts upped: so the stack will exist
8938 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8941 Also make SSL_SESSION_print() print out the verify return
8945 *) Add manpage for the pkcs12 command. Also change the default
8946 behaviour so MAC iteration counts are used unless the new
8947 -nomaciter option is used. This improves file security and
8948 only older versions of MSIE (4.0 for example) need it.
8951 *) Honor the no-xxx Configure options when creating .DEF files.
8954 *) Add PKCS#10 attributes to field table: challengePassword,
8955 unstructuredName and unstructuredAddress. These are taken from
8956 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8957 international characters are used.
8959 More changes to X509_ATTRIBUTE code: allow the setting of types
8960 based on strings. Remove the 'loc' parameter when adding
8961 attributes because these will be a SET OF encoding which is sorted
8965 *) Initial changes to the 'req' utility to allow request generation
8966 automation. This will allow an application to just generate a template
8967 file containing all the field values and have req construct the
8970 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8971 used all over the place including certificate requests and PKCS#7
8972 structures. They are currently handled manually where necessary with
8973 some primitive wrappers for PKCS#7. The new functions behave in a
8974 manner analogous to the X509 extension functions: they allow
8975 attributes to be looked up by NID and added.
8977 Later something similar to the X509V3 code would be desirable to
8978 automatically handle the encoding, decoding and printing of the
8979 more complex types. The string types like challengePassword can
8980 be handled by the string table functions.
8982 Also modified the multi byte string table handling. Now there is
8983 a 'global mask' which masks out certain types. The table itself
8984 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8985 is useful when for example there is only one permissible type
8986 (as in countryName) and using the mask might result in no valid
8990 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8991 SSL_get_peer_finished to allow applications to obtain the latest
8992 Finished messages sent to the peer or expected from the peer,
8993 respectively. (SSL_get_peer_finished is usually the Finished message
8994 actually received from the peer, otherwise the protocol will be aborted.)
8996 As the Finished message are message digests of the complete handshake
8997 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8998 be used for external authentication procedures when the authentication
8999 provided by SSL/TLS is not desired or is not enough.
9002 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9003 the host supports BWX extension and if Compaq C is present on the
9004 $PATH. Just exploiting of the BWX extension results in 20-30%
9005 performance kick for some algorithms, e.g. DES and RC4 to mention
9006 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9010 *) Add support for MS "fast SGC". This is arguably a violation of the
9011 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9012 weak crypto and after checking the certificate is SGC a second one
9013 with strong crypto. MS SGC stops the first handshake after receiving
9014 the server certificate message and sends a second client hello. Since
9015 a server will typically do all the time consuming operations before
9016 expecting any further messages from the client (server key exchange
9017 is the most expensive) there is little difference between the two.
9019 To get OpenSSL to support MS SGC we have to permit a second client
9020 hello message after we have sent server done. In addition we have to
9021 reset the MAC if we do get this second client hello.
9024 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9025 if a DER encoded private key is RSA or DSA traditional format. Changed
9026 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9027 format DER encoded private key. Newer code should use PKCS#8 format which
9028 has the key type encoded in the ASN1 structure. Added DER private key
9029 support to pkcs8 application.
9032 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9033 ciphersuites has been selected (as required by the SSL 3/TLS 1
9034 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9035 is set, we interpret this as a request to violate the specification
9036 (the worst that can happen is a handshake failure, and 'correct'
9037 behaviour would result in a handshake failure anyway).
9040 *) In SSL_CTX_add_session, take into account that there might be multiple
9041 SSL_SESSION structures with the same session ID (e.g. when two threads
9042 concurrently obtain them from an external cache).
9043 The internal cache can handle only one SSL_SESSION with a given ID,
9044 so if there's a conflict, we now throw out the old one to achieve
9048 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9049 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9050 some routines that use cipher OIDs: some ciphers do not have OIDs
9051 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9055 *) Simplify the trust setting structure and code. Now we just have
9056 two sequences of OIDs for trusted and rejected settings. These will
9057 typically have values the same as the extended key usage extension
9058 and any application specific purposes.
9060 The trust checking code now has a default behaviour: it will just
9061 check for an object with the same NID as the passed id. Functions can
9062 be provided to override either the default behaviour or the behaviour
9063 for a given id. SSL client, server and email already have functions
9064 in place for compatibility: they check the NID and also return "trusted"
9065 if the certificate is self signed.
9068 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9069 traditional format into an EVP_PKEY structure.
9072 *) Add a password callback function PEM_cb() which either prompts for
9073 a password if usr_data is NULL or otherwise assumes it is a null
9074 terminated password. Allow passwords to be passed on command line
9075 environment or config files in a few more utilities.
9078 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9079 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9080 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9081 Update documentation.
9084 *) Support for ASN1 "NULL" type. This could be handled before by using
9085 ASN1_TYPE but there wasn't any function that would try to read a NULL
9086 and produce an error if it couldn't. For compatibility we also have
9087 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9088 don't allocate anything because they don't need to.
9091 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9093 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9095 *) Rebuild of the memory allocation routines used by OpenSSL code and
9096 possibly others as well. The purpose is to make an interface that
9097 provide hooks so anyone can build a separate set of allocation and
9098 deallocation routines to be used by OpenSSL, for example memory
9099 pool implementations, or something else, which was previously hard
9100 since Malloc(), Realloc() and Free() were defined as macros having
9101 the values malloc, realloc and free, respectively (except for Win32
9102 compilations). The same is provided for memory debugging code.
9103 OpenSSL already comes with functionality to find memory leaks, but
9104 this gives people a chance to debug other memory problems.
9106 With these changes, a new set of functions and macros have appeared:
9108 CRYPTO_set_mem_debug_functions() [F]
9109 CRYPTO_get_mem_debug_functions() [F]
9110 CRYPTO_dbg_set_options() [F]
9111 CRYPTO_dbg_get_options() [F]
9112 CRYPTO_malloc_debug_init() [M]
9114 The memory debug functions are NULL by default, unless the library
9115 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9116 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9117 gives the standard debugging functions that come with OpenSSL) or
9118 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9119 provided by the library user) must be used. When the standard
9120 debugging functions are used, CRYPTO_dbg_set_options can be used to
9121 request additional information:
9122 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9123 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9125 Also, things like CRYPTO_set_mem_functions will always give the
9126 expected result (the new set of functions is used for allocation
9127 and deallocation) at all times, regardless of platform and compiler
9130 To finish it up, some functions that were never use in any other
9131 way than through macros have a new API and new semantic:
9134 CRYPTO_dbg_realloc()
9137 All macros of value have retained their old syntax.
9138 [Richard Levitte and Bodo Moeller]
9140 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9141 ordering of SMIMECapabilities wasn't in "strength order" and there
9142 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9146 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9147 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9148 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9150 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9151 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9152 functionality to handle multipart/signed properly) and a utility
9153 called 'smime' to call all this stuff. This is based on code I
9154 originally wrote for Celo who have kindly allowed it to be
9155 included in OpenSSL.
9158 *) Add variants des_set_key_checked and des_set_key_unchecked of
9159 des_set_key (aka des_key_sched). Global variable des_check_key
9160 decides which of these is called by des_set_key; this way
9161 des_check_key behaves as it always did, but applications and
9162 the library itself, which was buggy for des_check_key == 1,
9163 have a cleaner way to pick the version they need.
9166 *) New function PKCS12_newpass() which changes the password of a
9170 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9171 dynamic mix. In both cases the ids can be used as an index into the
9172 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9173 functions so they accept a list of the field values and the
9174 application doesn't need to directly manipulate the X509_TRUST
9178 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9182 *) Modify the way the V3 extension code looks up extensions. This now
9183 works in a similar way to the object code: we have some "standard"
9184 extensions in a static table which is searched with OBJ_bsearch()
9185 and the application can add dynamic ones if needed. The file
9186 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9187 updated whenever a new extension is added to the core code and kept
9188 in ext_nid order. There is a simple program 'tabtest.c' which checks
9189 this. New extensions are not added too often so this file can readily
9190 be maintained manually.
9192 There are two big advantages in doing things this way. The extensions
9193 can be looked up immediately and no longer need to be "added" using
9194 X509V3_add_standard_extensions(): this function now does nothing.
9195 [Side note: I get *lots* of email saying the extension code doesn't
9196 work because people forget to call this function]
9197 Also no dynamic allocation is done unless new extensions are added:
9198 so if we don't add custom extensions there is no need to call
9199 X509V3_EXT_cleanup().
9202 *) Modify enc utility's salting as follows: make salting the default. Add a
9203 magic header, so unsalted files fail gracefully instead of just decrypting
9204 to garbage. This is because not salting is a big security hole, so people
9205 should be discouraged from doing it.
9208 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9209 digest to be passed on the command line but it only used this
9210 parameter when signing a certificate. Modified so all relevant
9211 operations are affected by the digest parameter including the
9212 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9213 DSA key was used because it didn't fix the digest.
9216 *) Initial certificate chain verify code. Currently tests the untrusted
9217 certificates for consistency with the verify purpose (which is set
9218 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9220 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9221 this is because it will reject chains with invalid extensions whereas
9222 every previous version of OpenSSL and SSLeay made no checks at all.
9224 Trust code: checks the root CA for the relevant trust settings. Trust
9225 settings have an initial value consistent with the verify purpose: e.g.
9226 if the verify purpose is for SSL client use it expects the CA to be
9227 trusted for SSL client use. However the default value can be changed to
9228 permit custom trust settings: one example of this would be to only trust
9229 certificates from a specific "secure" set of CAs.
9231 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9232 which should be used for version portability: especially since the
9233 verify structure is likely to change more often now.
9235 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9236 to set them. If not set then assume SSL clients will verify SSL servers
9239 Two new options to the verify program: -untrusted allows a set of
9240 untrusted certificates to be passed in and -purpose which sets the
9241 intended purpose of the certificate. If a purpose is set then the
9242 new chain verify code is used to check extension consistency.
9245 *) Support for the authority information access extension.
9248 *) Modify RSA and DSA PEM read routines to transparently handle
9249 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9250 public keys in a format compatible with certificate
9251 SubjectPublicKeyInfo structures. Unfortunately there were already
9252 functions called *_PublicKey_* which used various odd formats so
9253 these are retained for compatibility: however the DSA variants were
9254 never in a public release so they have been deleted. Changed dsa/rsa
9255 utilities to handle the new format: note no releases ever handled public
9256 keys so we should be OK.
9258 The primary motivation for this change is to avoid the same fiasco
9259 that dogs private keys: there are several incompatible private key
9260 formats some of which are standard and some OpenSSL specific and
9261 require various evil hacks to allow partial transparent handling and
9262 even then it doesn't work with DER formats. Given the option anything
9263 other than PKCS#8 should be dumped: but the other formats have to
9264 stay in the name of compatibility.
9266 With public keys and the benefit of hindsight one standard format
9267 is used which works with EVP_PKEY, RSA or DSA structures: though
9268 it clearly returns an error if you try to read the wrong kind of key.
9270 Added a -pubkey option to the 'x509' utility to output the public key.
9271 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9272 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9273 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9274 that do the same as the EVP_PKEY_assign_*() except they up the
9275 reference count of the added key (they don't "swallow" the
9279 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9280 CRLs would fail if the file contained no certificates or no CRLs:
9281 added a new function to read in both types and return the number
9282 read: this means that if none are read it will be an error. The
9283 DER versions of the certificate and CRL reader would always fail
9284 because it isn't possible to mix certificates and CRLs in DER format
9285 without choking one or the other routine. Changed this to just read
9286 a certificate: this is the best we can do. Also modified the code
9287 in apps/verify.c to take notice of return codes: it was previously
9288 attempting to read in certificates from NULL pointers and ignoring
9289 any errors: this is one reason why the cert and CRL reader seemed
9290 to work. It doesn't check return codes from the default certificate
9291 routines: these may well fail if the certificates aren't installed.
9294 *) Code to support otherName option in GeneralName.
9297 *) First update to verify code. Change the verify utility
9298 so it warns if it is passed a self signed certificate:
9299 for consistency with the normal behaviour. X509_verify
9300 has been modified to it will now verify a self signed
9301 certificate if *exactly* the same certificate appears
9302 in the store: it was previously impossible to trust a
9303 single self signed certificate. This means that:
9304 openssl verify ss.pem
9305 now gives a warning about a self signed certificate but
9306 openssl verify -CAfile ss.pem ss.pem
9310 *) For servers, store verify_result in SSL_SESSION data structure
9311 (and add it to external session representation).
9312 This is needed when client certificate verifications fails,
9313 but an application-provided verification callback (set by
9314 SSL_CTX_set_cert_verify_callback) allows accepting the session
9315 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9316 but returns 1): When the session is reused, we have to set
9317 ssl->verify_result to the appropriate error code to avoid
9319 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9321 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9322 case in PKCS7_dataInit() where the signed PKCS7 structure
9323 didn't contain any existing data because it was being created.
9324 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9326 *) Add a salt to the key derivation routines in enc.c. This
9327 forms the first 8 bytes of the encrypted file. Also add a
9328 -S option to allow a salt to be input on the command line.
9331 *) New function X509_cmp(). Oddly enough there wasn't a function
9332 to compare two certificates. We do this by working out the SHA1
9333 hash and comparing that. X509_cmp() will be needed by the trust
9337 *) SSL_get1_session() is like SSL_get_session(), but increments
9338 the reference count in the SSL_SESSION returned.
9339 [Geoff Thorpe <geoff@eu.c2.net>]
9341 *) Fix for 'req': it was adding a null to request attributes.
9342 Also change the X509_LOOKUP and X509_INFO code to handle
9343 certificate auxiliary information.
9346 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9350 *) Add the possibility to add extra information to the memory leak
9351 detecting output, to form tracebacks, showing from where each
9352 allocation was originated: CRYPTO_push_info("constant string") adds
9353 the string plus current file name and line number to a per-thread
9354 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9355 is like calling CYRPTO_pop_info() until the stack is empty.
9356 Also updated memory leak detection code to be multi-thread-safe.
9359 *) Add options -text and -noout to pkcs7 utility and delete the
9360 encryption options which never did anything. Update docs.
9363 *) Add options to some of the utilities to allow the pass phrase
9364 to be included on either the command line (not recommended on
9365 OSes like Unix) or read from the environment. Update the
9366 manpages and fix a few bugs.
9369 *) Add a few manpages for some of the openssl commands.
9372 *) Fix the -revoke option in ca. It was freeing up memory twice,
9373 leaking and not finding already revoked certificates.
9376 *) Extensive changes to support certificate auxiliary information.
9377 This involves the use of X509_CERT_AUX structure and X509_AUX
9378 functions. An X509_AUX function such as PEM_read_X509_AUX()
9379 can still read in a certificate file in the usual way but it
9380 will also read in any additional "auxiliary information". By
9381 doing things this way a fair degree of compatibility can be
9382 retained: existing certificates can have this information added
9383 using the new 'x509' options.
9385 Current auxiliary information includes an "alias" and some trust
9386 settings. The trust settings will ultimately be used in enhanced
9387 certificate chain verification routines: currently a certificate
9388 can only be trusted if it is self signed and then it is trusted
9392 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9393 The problem was that one of the replacement routines had not been working
9394 since SSLeay releases. For now the offending routine has been replaced
9395 with non-optimised assembler. Even so, this now gives around 95%
9396 performance improvement for 1024 bit RSA signs.
9399 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9400 handling. Most clients have the effective key size in bits equal to
9401 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9402 A few however don't do this and instead use the size of the decrypted key
9403 to determine the RC2 key length and the AlgorithmIdentifier to determine
9404 the effective key length. In this case the effective key length can still
9405 be 40 bits but the key length can be 168 bits for example. This is fixed
9406 by manually forcing an RC2 key into the EVP_PKEY structure because the
9407 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9408 the key length and effective key length are equal.
9411 *) Add a bunch of functions that should simplify the creation of
9412 X509_NAME structures. Now you should be able to do:
9413 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9414 and have it automatically work out the correct field type and fill in
9415 the structures. The more adventurous can try:
9416 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9417 and it will (hopefully) work out the correct multibyte encoding.
9420 *) Change the 'req' utility to use the new field handling and multibyte
9421 copy routines. Before the DN field creation was handled in an ad hoc
9422 way in req, ca, and x509 which was rather broken and didn't support
9423 BMPStrings or UTF8Strings. Since some software doesn't implement
9424 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9425 using the dirstring_type option. See the new comment in the default
9426 openssl.cnf for more info.
9429 *) Make crypto/rand/md_rand.c more robust:
9430 - Assure unique random numbers after fork().
9431 - Make sure that concurrent threads access the global counter and
9432 md serializably so that we never lose entropy in them
9433 or use exactly the same state in multiple threads.
9434 Access to the large state is not always serializable because
9435 the additional locking could be a performance killer, and
9436 md should be large enough anyway.
9439 *) New file apps/app_rand.c with commonly needed functionality
9440 for handling the random seed file.
9442 Use the random seed file in some applications that previously did not:
9444 dsaparam -genkey (which also ignored its '-rand' option),
9447 x509 (when signing).
9448 Except on systems with /dev/urandom, it is crucial to have a random
9449 seed file at least for key creation, DSA signing, and for DH exchanges;
9450 for RSA signatures we could do without one.
9452 gendh and gendsa (unlike genrsa) used to read only the first byte
9453 of each file listed in the '-rand' option. The function as previously
9454 found in genrsa is now in app_rand.c and is used by all programs
9455 that support '-rand'.
9458 *) In RAND_write_file, use mode 0600 for creating files;
9459 don't just chmod when it may be too late.
9462 *) Report an error from X509_STORE_load_locations
9463 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9466 *) New function ASN1_mbstring_copy() this copies a string in either
9467 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9468 into an ASN1_STRING type. A mask of permissible types is passed
9469 and it chooses the "minimal" type to use or an error if not type
9473 *) Add function equivalents to the various macros in asn1.h. The old
9474 macros are retained with an M_ prefix. Code inside the library can
9475 use the M_ macros. External code (including the openssl utility)
9476 should *NOT* in order to be "shared library friendly".
9479 *) Add various functions that can check a certificate's extensions
9480 to see if it usable for various purposes such as SSL client,
9481 server or S/MIME and CAs of these types. This is currently
9482 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9483 verification. Also added a -purpose flag to x509 utility to
9484 print out all the purposes.
9487 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9491 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9492 for, obtain and decode and extension and obtain its critical flag.
9493 This allows all the necessary extension code to be handled in a
9494 single function call.
9497 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9498 platforms. See crypto/rc4/rc4_enc.c for further details.
9501 *) New -noout option to asn1parse. This causes no output to be produced
9502 its main use is when combined with -strparse and -out to extract data
9503 from a file (which may not be in ASN.1 format).
9506 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9507 when producing the local key id.
9508 [Richard Levitte <levitte@stacken.kth.se>]
9510 *) New option -dhparam in s_server. This allows a DH parameter file to be
9511 stated explicitly. If it is not stated then it tries the first server
9512 certificate file. The previous behaviour hard coded the filename
9516 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9517 a public key to be input or output. For example:
9518 openssl rsa -in key.pem -pubout -out pubkey.pem
9519 Also added necessary DSA public key functions to handle this.
9522 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9523 in the message. This was handled by allowing
9524 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9525 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9527 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9528 to the end of the strings whereas this didn't. This would cause problems
9529 if strings read with d2i_ASN1_bytes() were later modified.
9530 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9532 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9533 data and it contains EOF it will end up returning an error. This is
9534 caused by input 46 bytes long. The cause is due to the way base64
9535 BIOs find the start of base64 encoded data. They do this by trying a
9536 trial decode on each line until they find one that works. When they
9537 do a flag is set and it starts again knowing it can pass all the
9538 data directly through the decoder. Unfortunately it doesn't reset
9539 the context it uses. This means that if EOF is reached an attempt
9540 is made to pass two EOFs through the context and this causes the
9541 resulting error. This can also cause other problems as well. As is
9542 usual with these problems it takes *ages* to find and the fix is
9543 trivial: move one line.
9544 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9546 *) Ugly workaround to get s_client and s_server working under Windows. The
9547 old code wouldn't work because it needed to select() on sockets and the
9548 tty (for keypresses and to see if data could be written). Win32 only
9549 supports select() on sockets so we select() with a 1s timeout on the
9550 sockets and then see if any characters are waiting to be read, if none
9551 are present then we retry, we also assume we can always write data to
9552 the tty. This isn't nice because the code then blocks until we've
9553 received a complete line of data and it is effectively polling the
9554 keyboard at 1s intervals: however it's quite a bit better than not
9555 working at all :-) A dedicated Windows application might handle this
9556 with an event loop for example.
9559 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9560 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9561 will be called when RSA_sign() and RSA_verify() are used. This is useful
9562 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9563 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9564 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9565 This necessitated the support of an extra signature type NID_md5_sha1
9566 for SSL signatures and modifications to the SSL library to use it instead
9567 of calling RSA_public_decrypt() and RSA_private_encrypt().
9570 *) Add new -verify -CAfile and -CApath options to the crl program, these
9571 will lookup a CRL issuers certificate and verify the signature in a
9572 similar way to the verify program. Tidy up the crl program so it
9573 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9574 less strict. It will now permit CRL extensions even if it is not
9575 a V2 CRL: this will allow it to tolerate some broken CRLs.
9578 *) Initialize all non-automatic variables each time one of the openssl
9579 sub-programs is started (this is necessary as they may be started
9580 multiple times from the "OpenSSL>" prompt).
9581 [Lennart Bang, Bodo Moeller]
9583 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9584 removing all other RSA functionality (this is what NO_RSA does). This
9585 is so (for example) those in the US can disable those operations covered
9586 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9590 *) Non-copying interface to BIO pairs.
9591 (still largely untested)
9594 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9595 ASCII string. This was handled independently in various places before.
9598 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9599 UTF8 strings a character at a time.
9602 *) Use client_version from client hello to select the protocol
9603 (s23_srvr.c) and for RSA client key exchange verification
9604 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9607 *) Add various utility functions to handle SPKACs, these were previously
9608 handled by poking round in the structure internals. Added new function
9609 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9610 print, verify and generate SPKACs. Based on an original idea from
9611 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9614 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9617 *) Allow the config file extension section to be overwritten on the
9618 command line. Based on an original idea from Massimiliano Pala
9619 <madwolf@comune.modena.it>. The new option is called -extensions
9620 and can be applied to ca, req and x509. Also -reqexts to override
9621 the request extensions in req and -crlexts to override the crl extensions
9625 *) Add new feature to the SPKAC handling in ca. Now you can include
9626 the same field multiple times by preceding it by "XXXX." for example:
9629 this is the same syntax as used in the req config file.
9632 *) Allow certificate extensions to be added to certificate requests. These
9633 are specified in a 'req_extensions' option of the req section of the
9634 config file. They can be printed out with the -text option to req but
9635 are otherwise ignored at present.
9638 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9639 data read consists of only the final block it would not decrypted because
9640 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9641 A misplaced 'break' also meant the decrypted final block might not be
9642 copied until the next read.
9645 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9646 a few extra parameters to the DH structure: these will be useful if
9647 for example we want the value of 'q' or implement X9.42 DH.
9650 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9651 provides hooks that allow the default DSA functions or functions on a
9652 "per key" basis to be replaced. This allows hardware acceleration and
9653 hardware key storage to be handled without major modification to the
9654 library. Also added low level modexp hooks and CRYPTO_EX structure and
9655 associated functions.
9658 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9659 as "read only": it can't be written to and the buffer it points to will
9660 not be freed. Reading from a read only BIO is much more efficient than
9661 a normal memory BIO. This was added because there are several times when
9662 an area of memory needs to be read from a BIO. The previous method was
9663 to create a memory BIO and write the data to it, this results in two
9664 copies of the data and an O(n^2) reading algorithm. There is a new
9665 function BIO_new_mem_buf() which creates a read only memory BIO from
9666 an area of memory. Also modified the PKCS#7 routines to use read only
9670 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9671 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9672 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9673 but a retry condition occured while trying to read the rest.
9676 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9677 NID_pkcs7_encrypted by default: this was wrong since this should almost
9678 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9679 the encrypted data type: this is a more sensible place to put it and it
9680 allows the PKCS#12 code to be tidied up that duplicated this
9684 *) Changed obj_dat.pl script so it takes its input and output files on
9685 the command line. This should avoid shell escape redirection problems
9689 *) Initial support for certificate extension requests, these are included
9690 in things like Xenroll certificate requests. Included functions to allow
9691 extensions to be obtained and added.
9694 *) -crlf option to s_client and s_server for sending newlines as
9695 CRLF (as required by many protocols).
9698 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9700 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9701 [Ralf S. Engelschall]
9703 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9704 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9706 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9710 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9711 DH parameters/keys (q is lost during that conversion, but the resulting
9712 DH parameters contain its length).
9714 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9715 much faster than DH_generate_parameters (which creates parameters
9716 where p = 2*q + 1), and also the smaller q makes DH computations
9717 much more efficient (160-bit exponentiation instead of 1024-bit
9718 exponentiation); so this provides a convenient way to support DHE
9719 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9720 utter importance to use
9721 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9723 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9724 when such DH parameters are used, because otherwise small subgroup
9725 attacks may become possible!
9728 *) Avoid memory leak in i2d_DHparams.
9731 *) Allow the -k option to be used more than once in the enc program:
9732 this allows the same encrypted message to be read by multiple recipients.
9735 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9736 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9737 it will always use the numerical form of the OID, even if it has a short
9741 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9742 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9743 otherwise bn_mod_exp was called. In the case of hardware keys for example
9744 no private key components need be present and it might store extra data
9745 in the RSA structure, which cannot be accessed from bn_mod_exp.
9746 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9747 private key operations.
9750 *) Added support for SPARC Linux.
9753 *) pem_password_cb function type incompatibly changed from
9754 typedef int pem_password_cb(char *buf, int size, int rwflag);
9756 ....(char *buf, int size, int rwflag, void *userdata);
9757 so that applications can pass data to their callbacks:
9758 The PEM[_ASN1]_{read,write}... functions and macros now take an
9759 additional void * argument, which is just handed through whenever
9760 the password callback is called.
9761 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9763 New function SSL_CTX_set_default_passwd_cb_userdata.
9765 Compatibility note: As many C implementations push function arguments
9766 onto the stack in reverse order, the new library version is likely to
9767 interoperate with programs that have been compiled with the old
9768 pem_password_cb definition (PEM_whatever takes some data that
9769 happens to be on the stack as its last argument, and the callback
9770 just ignores this garbage); but there is no guarantee whatsoever that
9773 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9774 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9775 problems not only on Windows, but also on some Unix platforms.
9776 To avoid problematic command lines, these definitions are now in an
9777 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9778 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9781 *) MIPS III/IV assembler module is reimplemented.
9784 *) More DES library cleanups: remove references to srand/rand and
9785 delete an unused file.
9788 *) Add support for the the free Netwide assembler (NASM) under Win32,
9789 since not many people have MASM (ml) and it can be hard to obtain.
9790 This is currently experimental but it seems to work OK and pass all
9791 the tests. Check out INSTALL.W32 for info.
9794 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9795 without temporary keys kept an extra copy of the server key,
9796 and connections with temporary keys did not free everything in case
9800 *) New function RSA_check_key and new openssl rsa option -check
9801 for verifying the consistency of RSA keys.
9802 [Ulf Moeller, Bodo Moeller]
9804 *) Various changes to make Win32 compile work:
9805 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9806 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9807 comparison" warnings.
9808 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9811 *) Add a debugging option to PKCS#5 v2 key generation function: when
9812 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9813 derived keys are printed to stderr.
9816 *) Copy the flags in ASN1_STRING_dup().
9817 [Roman E. Pavlov <pre@mo.msk.ru>]
9819 *) The x509 application mishandled signing requests containing DSA
9820 keys when the signing key was also DSA and the parameters didn't match.
9822 It was supposed to omit the parameters when they matched the signing key:
9823 the verifying software was then supposed to automatically use the CA's
9824 parameters if they were absent from the end user certificate.
9826 Omitting parameters is no longer recommended. The test was also
9827 the wrong way round! This was probably due to unusual behaviour in
9828 EVP_cmp_parameters() which returns 1 if the parameters match.
9829 This meant that parameters were omitted when they *didn't* match and
9830 the certificate was useless. Certificates signed with 'ca' didn't have
9832 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9834 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9835 The interface is as follows:
9836 Applications can use
9837 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9838 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9839 "off" is now the default.
9840 The library internally uses
9841 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9842 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9843 to disable memory-checking temporarily.
9845 Some inconsistent states that previously were possible (and were
9846 even the default) are now avoided.
9848 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9849 with each memory chunk allocated; this is occasionally more helpful
9850 than just having a counter.
9852 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9854 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9858 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9859 which largely parallels "options", but is for changing API behaviour,
9860 whereas "options" are about protocol behaviour.
9861 Initial "mode" flags are:
9863 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9864 a single record has been written.
9865 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9866 retries use the same buffer location.
9867 (But all of the contents must be
9871 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9874 *) Fix problems with no-hmac etc.
9875 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9877 *) New functions RSA_get_default_method(), RSA_set_method() and
9878 RSA_get_method(). These allows replacement of RSA_METHODs without having
9879 to mess around with the internals of an RSA structure.
9882 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9883 Also really enable memory leak checks in openssl.c and in some
9885 [Chad C. Mulligan, Bodo Moeller]
9887 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9888 up the length of negative integers. This has now been simplified to just
9889 store the length when it is first determined and use it later, rather
9890 than trying to keep track of where data is copied and updating it to
9892 [Steve Henson, reported by Brien Wheeler
9893 <bwheeler@authentica-security.com>]
9895 *) Add a new function PKCS7_signatureVerify. This allows the verification
9896 of a PKCS#7 signature but with the signing certificate passed to the
9897 function itself. This contrasts with PKCS7_dataVerify which assumes the
9898 certificate is present in the PKCS#7 structure. This isn't always the
9899 case: certificates can be omitted from a PKCS#7 structure and be
9900 distributed by "out of band" means (such as a certificate database).
9903 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9904 function prototypes in pem.h, also change util/mkdef.pl to add the
9905 necessary function names.
9908 *) mk1mf.pl (used by Windows builds) did not properly read the
9909 options set by Configure in the top level Makefile, and Configure
9910 was not even able to write more than one option correctly.
9911 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9914 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9915 file to be loaded from a BIO or FILE pointer. The BIO version will
9916 for example allow memory BIOs to contain config info.
9919 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9920 Whoever hopes to achieve shared-library compatibility across versions
9921 must use this, not the compile-time macro.
9922 (Exercise 0.9.4: Which is the minimum library version required by
9924 Note: All this applies only to multi-threaded programs, others don't
9928 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9929 through a BIO pair triggered the default case, i.e.
9930 SSLerr(...,SSL_R_UNKNOWN_STATE).
9933 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9934 can use the SSL library even if none of the specific BIOs is
9938 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9939 for the encoded length.
9940 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9942 *) Add initial documentation of the X509V3 functions.
9945 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9946 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9947 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9948 secure PKCS#8 private key format with a high iteration count.
9951 *) Fix determination of Perl interpreter: A perl or perl5
9952 _directory_ in $PATH was also accepted as the interpreter.
9953 [Ralf S. Engelschall]
9955 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9956 wrong with it but it was very old and did things like calling
9957 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9961 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9962 to use the new extension code.
9965 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9966 with macros. This should make it easier to change their form, add extra
9967 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9971 *) Add to configuration table a new entry that can specify an alternative
9972 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9973 according to Mark Crispin <MRC@Panda.COM>.
9977 *) DES CBC did not update the IV. Weird.
9980 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9981 Changing the behaviour of the former might break existing programs --
9982 where IV updating is needed, des_ncbc_encrypt can be used.
9985 *) When bntest is run from "make test" it drives bc to check its
9986 calculations, as well as internally checking them. If an internal check
9987 fails, it needs to cause bc to give a non-zero result or make test carries
9988 on without noticing the failure. Fixed.
9991 *) DES library cleanups.
9994 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9995 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9996 ciphers. NOTE: although the key derivation function has been verified
9997 against some published test vectors it has not been extensively tested
9998 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10002 *) Instead of "mkdir -p", which is not fully portable, use new
10003 Perl script "util/mkdir-p.pl".
10006 *) Rewrite the way password based encryption (PBE) is handled. It used to
10007 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10008 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10009 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10010 the 'parameter' field of the AlgorithmIdentifier is passed to the
10011 underlying key generation function so it must do its own ASN1 parsing.
10012 This has also changed the EVP_PBE_CipherInit() function which now has a
10013 'parameter' argument instead of literal salt and iteration count values
10014 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10017 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10018 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10019 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10020 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10021 value was just used as a "magic string" and not used directly its
10022 value doesn't matter.
10025 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10029 *) "linux-sparc64" configuration (ultrapenguin).
10030 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10031 "linux-sparc" configuration.
10032 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10034 *) config now generates no-xxx options for missing ciphers.
10037 *) Support the EBCDIC character set (work in progress).
10038 File ebcdic.c not yet included because it has a different license.
10039 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10041 *) Support BS2000/OSD-POSIX.
10042 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10044 *) Make callbacks for key generation use void * instead of char *.
10047 *) Make S/MIME samples compile (not yet tested).
10050 *) Additional typesafe stacks.
10053 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10057 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10059 *) New configuration variant "sco5-gcc".
10061 *) Updated some demos.
10062 [Sean O Riordain, Wade Scholine]
10064 *) Add missing BIO_free at exit of pkcs12 application.
10067 *) Fix memory leak in conf.c.
10070 *) Updates for Win32 to assembler version of MD5.
10073 *) Set #! path to perl in apps/der_chop to where we found it
10074 instead of using a fixed path.
10077 *) SHA library changes for irix64-mips4-cc.
10080 *) Improvements for VMS support.
10084 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10086 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10087 This also avoids the problems with SC4.2 and unpatched SC5.
10088 [Andy Polyakov <appro@fy.chalmers.se>]
10090 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10091 These are required because of the typesafe stack would otherwise break
10092 existing code. If old code used a structure member which used to be STACK
10093 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10094 sk_num or sk_value it would produce an error because the num, data members
10095 are not present in STACK_OF. Now it just produces a warning. sk_set
10096 replaces the old method of assigning a value to sk_value
10097 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10098 that does this will no longer work (and should use sk_set instead) but
10099 this could be regarded as a "questionable" behaviour anyway.
10102 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10103 correctly handle encrypted S/MIME data.
10106 *) Change type of various DES function arguments from des_cblock
10107 (which means, in function argument declarations, pointer to char)
10108 to des_cblock * (meaning pointer to array with 8 char elements),
10109 which allows the compiler to do more typechecking; it was like
10110 that back in SSLeay, but with lots of ugly casts.
10112 Introduce new type const_des_cblock.
10115 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10116 problems: find RecipientInfo structure that matches recipient certificate
10117 and initialise the ASN1 structures properly based on passed cipher.
10120 *) Belatedly make the BN tests actually check the results.
10123 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10124 to and from BNs: it was completely broken. New compilation option
10125 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10126 key elements as negative integers.
10129 *) Reorganize and speed up MD5.
10130 [Andy Polyakov <appro@fy.chalmers.se>]
10133 [Richard Levitte <richard@levitte.org>]
10135 *) New option -out to asn1parse to allow the parsed structure to be
10136 output to a file. This is most useful when combined with the -strparse
10137 option to examine the output of things like OCTET STRINGS.
10140 *) Make SSL library a little more fool-proof by not requiring any longer
10141 that SSL_set_{accept,connect}_state be called before
10142 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10143 in many applications because usually everything *appeared* to work as
10144 intended anyway -- now it really works as intended).
10147 *) Move openssl.cnf out of lib/.
10150 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10151 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10152 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10153 [Ralf S. Engelschall]
10155 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10156 handle PKCS#7 enveloped data properly.
10157 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10159 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10160 copying pointers. The cert_st handling is changed by this in
10161 various ways (and thus what used to be known as ctx->default_cert
10162 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10163 any longer when s->cert does not give us what we need).
10164 ssl_cert_instantiate becomes obsolete by this change.
10165 As soon as we've got the new code right (possibly it already is?),
10166 we have solved a couple of bugs of the earlier code where s->cert
10167 was used as if it could not have been shared with other SSL structures.
10169 Note that using the SSL API in certain dirty ways now will result
10170 in different behaviour than observed with earlier library versions:
10171 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10172 does not influence s as it used to.
10174 In order to clean up things more thoroughly, inside SSL_SESSION
10175 we don't use CERT any longer, but a new structure SESS_CERT
10176 that holds per-session data (if available); currently, this is
10177 the peer's certificate chain and, for clients, the server's certificate
10178 and temporary key. CERT holds only those values that can have
10179 meaningful defaults in an SSL_CTX.
10182 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10183 from the internal representation. Various PKCS#7 fixes: remove some
10184 evil casts and set the enc_dig_alg field properly based on the signing
10188 *) Allow PKCS#12 password to be set from the command line or the
10189 environment. Let 'ca' get its config file name from the environment
10190 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10194 *) Allow certificate policies extension to use an IA5STRING for the
10195 organization field. This is contrary to the PKIX definition but
10196 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10200 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10201 without disallowing inline assembler and the like for non-pedantic builds.
10204 *) Support Borland C++ builder.
10205 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10207 *) Support Mingw32.
10210 *) SHA-1 cleanups and performance enhancements.
10211 [Andy Polyakov <appro@fy.chalmers.se>]
10213 *) Sparc v8plus assembler for the bignum library.
10214 [Andy Polyakov <appro@fy.chalmers.se>]
10216 *) Accept any -xxx and +xxx compiler options in Configure.
10219 *) Update HPUX configuration.
10222 *) Add missing sk_<type>_unshift() function to safestack.h
10223 [Ralf S. Engelschall]
10225 *) New function SSL_CTX_use_certificate_chain_file that sets the
10226 "extra_cert"s in addition to the certificate. (This makes sense
10227 only for "PEM" format files, as chains as a whole are not
10231 *) Support verify_depth from the SSL API.
10232 x509_vfy.c had what can be considered an off-by-one-error:
10233 Its depth (which was not part of the external interface)
10234 was actually counting the number of certificates in a chain;
10235 now it really counts the depth.
10238 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10239 instead of X509err, which often resulted in confusing error
10240 messages since the error codes are not globally unique
10241 (e.g. an alleged error in ssl3_accept when a certificate
10242 didn't match the private key).
10244 *) New function SSL_CTX_set_session_id_context that allows to set a default
10245 value (so that you don't need SSL_set_session_id_context for each
10246 connection using the SSL_CTX).
10249 *) OAEP decoding bug fix.
10252 *) Support INSTALL_PREFIX for package builders, as proposed by
10256 *) New Configure options "threads" and "no-threads". For systems
10257 where the proper compiler options are known (currently Solaris
10258 and Linux), "threads" is the default.
10261 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10264 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10265 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10266 such as /usr/local/bin.
10269 *) "make linux-shared" to build shared libraries.
10270 [Niels Poppe <niels@netbox.org>]
10272 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10275 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10276 extension adding in x509 utility.
10279 *) Remove NOPROTO sections and error code comments.
10282 *) Partial rewrite of the DEF file generator to now parse the ANSI
10286 *) New Configure options --prefix=DIR and --openssldir=DIR.
10289 *) Complete rewrite of the error code script(s). It is all now handled
10290 by one script at the top level which handles error code gathering,
10291 header rewriting and C source file generation. It should be much better
10292 than the old method: it now uses a modified version of Ulf's parser to
10293 read the ANSI prototypes in all header files (thus the old K&R definitions
10294 aren't needed for error creation any more) and do a better job of
10295 translating function codes into names. The old 'ASN1 error code imbedded
10296 in a comment' is no longer necessary and it doesn't use .err files which
10297 have now been deleted. Also the error code call doesn't have to appear all
10298 on one line (which resulted in some large lines...).
10301 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10304 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10305 0 (which usually indicates a closed connection), but continue reading.
10308 *) Fix some race conditions.
10311 *) Add support for CRL distribution points extension. Add Certificate
10312 Policies and CRL distribution points documentation.
10315 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10318 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10319 8 of keying material. Merlin has also confirmed interop with this fix
10320 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10321 [Merlin Hughes <merlin@baltimore.ie>]
10323 *) Fix lots of warnings.
10324 [Richard Levitte <levitte@stacken.kth.se>]
10326 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10327 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10328 [Richard Levitte <levitte@stacken.kth.se>]
10330 *) Fix problems with sizeof(long) == 8.
10331 [Andy Polyakov <appro@fy.chalmers.se>]
10333 *) Change functions to ANSI C.
10336 *) Fix typos in error codes.
10337 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10339 *) Remove defunct assembler files from Configure.
10342 *) SPARC v8 assembler BIGNUM implementation.
10343 [Andy Polyakov <appro@fy.chalmers.se>]
10345 *) Support for Certificate Policies extension: both print and set.
10346 Various additions to support the r2i method this uses.
10349 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10350 return a const string when you are expecting an allocated buffer.
10353 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10354 types DirectoryString and DisplayText.
10357 *) Add code to allow r2i extensions to access the configuration database,
10358 add an LHASH database driver and add several ctx helper functions.
10361 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10362 fail when they extended the size of a BIGNUM.
10365 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10366 support typesafe stack.
10369 *) Fix typo in SSL_[gs]et_options().
10370 [Nils Frostberg <nils@medcom.se>]
10372 *) Delete various functions and files that belonged to the (now obsolete)
10373 old X509V3 handling code.
10376 *) New Configure option "rsaref".
10379 *) Don't auto-generate pem.h.
10382 *) Introduce type-safe ASN.1 SETs.
10385 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10386 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10388 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10389 that links with OpenSSL (well at least cause lots of warnings), but fear
10390 not: the conversion is trivial, and it eliminates loads of evil casts. A
10391 few STACKed things have been converted already. Feel free to convert more.
10392 In the fullness of time, I'll do away with the STACK type altogether.
10395 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10396 specified in <certfile> by updating the entry in the index.txt file.
10397 This way one no longer has to edit the index.txt file manually for
10398 revoking a certificate. The -revoke option does the gory details now.
10399 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10401 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10402 `-text' option at all and this way the `-noout -text' combination was
10403 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10404 [Ralf S. Engelschall]
10406 *) Make sure a corresponding plain text error message exists for the
10407 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10408 verify callback function determined that a certificate was revoked.
10409 [Ralf S. Engelschall]
10411 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10412 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10413 all available cipers including rc5, which was forgotten until now.
10414 In order to let the testing shell script know which algorithms
10415 are available, a new (up to now undocumented) command
10416 "openssl list-cipher-commands" is used.
10419 *) Bugfix: s_client occasionally would sleep in select() when
10420 it should have checked SSL_pending() first.
10423 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10424 the raw DSA values prior to ASN.1 encoding.
10427 *) Tweaks to Configure
10428 [Niels Poppe <niels@netbox.org>]
10430 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10434 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10437 *) New config option to avoid instructions that are illegal on the 80386.
10438 The default code is faster, but requires at least a 486.
10441 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10442 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10443 same as SSL2_VERSION anyway.
10446 *) New "-showcerts" option for s_client.
10449 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10450 application. Various cleanups and fixes.
10453 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10454 modify error routines to work internally. Add error codes and PBE init
10455 to library startup routines.
10458 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10459 packing functions to asn1 and evp. Changed function names and error
10460 codes along the way.
10463 *) PKCS12 integration: and so it begins... First of several patches to
10464 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10465 objects to objects.h
10468 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10469 and display support for Thawte strong extranet extension.
10472 *) Add LinuxPPC support.
10473 [Jeff Dubrule <igor@pobox.org>]
10475 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10476 bn_div_words in alpha.s.
10477 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10479 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10480 OAEP isn't supported when OpenSSL is built with RSAref.
10481 [Ulf Moeller <ulf@fitug.de>]
10483 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10484 so they no longer are missing under -DNOPROTO.
10485 [Soren S. Jorvang <soren@t.dk>]
10488 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10490 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10491 doesn't work when the session is reused. Coming soon!
10494 *) Fix a security hole, that allows sessions to be reused in the wrong
10495 context thus bypassing client cert protection! All software that uses
10496 client certs and session caches in multiple contexts NEEDS PATCHING to
10497 allow session reuse! A fuller solution is in the works.
10498 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10500 *) Some more source tree cleanups (removed obsolete files
10501 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10502 permission on "config" script to be executable) and a fix for the INSTALL
10504 [Ulf Moeller <ulf@fitug.de>]
10506 *) Remove some legacy and erroneous uses of malloc, free instead of
10508 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10510 *) Make rsa_oaep_test return non-zero on error.
10511 [Ulf Moeller <ulf@fitug.de>]
10513 *) Add support for native Solaris shared libraries. Configure
10514 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10515 if someone would make that last step automatic.
10516 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10518 *) ctx_size was not built with the right compiler during "make links". Fixed.
10521 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10522 except NULL ciphers". This means the default cipher list will no longer
10523 enable NULL ciphers. They need to be specifically enabled e.g. with
10524 the string "DEFAULT:eNULL".
10527 *) Fix to RSA private encryption routines: if p < q then it would
10528 occasionally produce an invalid result. This will only happen with
10529 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10532 *) Be less restrictive and allow also `perl util/perlpath.pl
10533 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10534 because this way one can also use an interpreter named `perl5' (which is
10535 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10536 installed as `perl').
10537 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10539 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10540 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10542 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10543 advapi32.lib to Win32 build and change the pem test comparison
10544 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10545 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10546 and crypto/des/ede_cbcm_enc.c.
10549 *) DES quad checksum was broken on big-endian architectures. Fixed.
10552 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10553 Win32 test batch file so it (might) work again. The Win32 test batch file
10554 is horrible: I feel ill....
10557 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10558 in e_os.h. Audit of header files to check ANSI and non ANSI
10559 sections: 10 functions were absent from non ANSI section and not exported
10560 from Windows DLLs. Fixed up libeay.num for new functions.
10563 *) Make `openssl version' output lines consistent.
10564 [Ralf S. Engelschall]
10566 *) Fix Win32 symbol export lists for BIO functions: Added
10567 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10568 to ms/libeay{16,32}.def.
10569 [Ralf S. Engelschall]
10571 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10572 fine under Unix and passes some trivial tests I've now added. But the
10573 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10574 added to make sure no one expects that this stuff really works in the
10575 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10576 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10578 [Ralf S. Engelschall]
10580 *) Fix the generation of two part addresses in perl.
10581 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10583 *) Add config entry for Linux on MIPS.
10584 [John Tobey <jtobey@channel1.com>]
10586 *) Make links whenever Configure is run, unless we are on Windoze.
10589 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10590 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10594 *) Add a useful kludge to allow package maintainers to specify compiler and
10595 other platforms details on the command line without having to patch the
10596 Configure script everytime: One now can use ``perl Configure
10597 <id>:<details>'', i.e. platform ids are allowed to have details appended
10598 to them (separated by colons). This is treated as there would be a static
10599 pre-configured entry in Configure's %table under key <id> with value
10600 <details> and ``perl Configure <id>'' is called. So, when you want to
10601 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10602 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10603 now, which overrides the FreeBSD-elf entry on-the-fly.
10604 [Ralf S. Engelschall]
10606 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10609 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10610 on the `perl Configure ...' command line. This way one can compile
10611 OpenSSL libraries with Position Independent Code (PIC) which is needed
10612 for linking it into DSOs.
10613 [Ralf S. Engelschall]
10615 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10619 *) Cleaned up the LICENSE document: The official contact for any license
10620 questions now is the OpenSSL core team under openssl-core@openssl.org.
10621 And add a paragraph about the dual-license situation to make sure people
10622 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10623 to the OpenSSL toolkit.
10624 [Ralf S. Engelschall]
10626 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10627 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10628 Additionally cleaned up the `make links' target: Remove unnecessary
10629 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10630 to speed processing and no longer clutter the display with confusing
10631 stuff. Instead only the actually done links are displayed.
10632 [Ralf S. Engelschall]
10634 *) Permit null encryption ciphersuites, used for authentication only. It used
10635 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10636 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10640 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10641 signed attributes when verifying signatures (this would break them),
10642 the detached data encoding was wrong and public keys obtained using
10643 X509_get_pubkey() weren't freed.
10646 *) Add text documentation for the BUFFER functions. Also added a work around
10647 to a Win95 console bug. This was triggered by the password read stuff: the
10648 last character typed gets carried over to the next fread(). If you were
10649 generating a new cert request using 'req' for example then the last
10650 character of the passphrase would be CR which would then enter the first
10654 *) Added the new `Includes OpenSSL Cryptography Software' button as
10655 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10656 button and can be used by applications based on OpenSSL to show the
10657 relationship to the OpenSSL project.
10658 [Ralf S. Engelschall]
10660 *) Remove confusing variables in function signatures in files
10661 ssl/ssl_lib.c and ssl/ssl.h.
10662 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10664 *) Don't install bss_file.c under PREFIX/include/
10665 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10667 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10668 functions that return function pointers and has support for NT specific
10669 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10670 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10671 unsigned to signed types: this was killing the Win32 compile.
10674 *) Add new certificate file to stack functions,
10675 SSL_add_dir_cert_subjects_to_stack() and
10676 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10677 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10678 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10679 This means that Apache-SSL and similar packages don't have to mess around
10680 to add as many CAs as they want to the preferred list.
10683 *) Experiment with doxygen documentation. Currently only partially applied to
10685 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10686 openssl.doxy as the configuration file.
10689 *) Get rid of remaining C++-style comments which strict C compilers hate.
10690 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10692 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10693 compiled in by default: it has problems with large keys.
10696 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10697 DH private keys and/or callback functions which directly correspond to
10698 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10699 is needed for applications which have to configure certificates on a
10700 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10702 For the RSA certificate situation is makes no difference, but
10703 for the DSA certificate situation this fixes the "no shared cipher"
10704 problem where the OpenSSL cipher selection procedure failed because the
10705 temporary keys were not overtaken from the context and the API provided
10706 no way to reconfigure them.
10707 The new functions now let applications reconfigure the stuff and they
10708 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10709 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10710 non-public-API function ssl_cert_instantiate() is used as a helper
10711 function and also to reduce code redundancy inside ssl_rsa.c.
10712 [Ralf S. Engelschall]
10714 *) Move s_server -dcert and -dkey options out of the undocumented feature
10715 area because they are useful for the DSA situation and should be
10716 recognized by the users.
10717 [Ralf S. Engelschall]
10719 *) Fix the cipher decision scheme for export ciphers: the export bits are
10720 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10721 SSL_EXP_MASK. So, the original variable has to be used instead of the
10722 already masked variable.
10723 [Richard Levitte <levitte@stacken.kth.se>]
10725 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10726 [Richard Levitte <levitte@stacken.kth.se>]
10728 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10729 from `int' to `unsigned int' because it's a length and initialized by
10730 EVP_DigestFinal() which expects an `unsigned int *'.
10731 [Richard Levitte <levitte@stacken.kth.se>]
10733 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10734 script. Instead use the usual Shell->Perl transition trick.
10735 [Ralf S. Engelschall]
10737 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10738 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10739 -noout -modulus' as it's already the case for `openssl rsa -noout
10740 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10741 currently the public key is printed (a decision which was already done by
10742 `openssl dsa -modulus' in the past) which serves a similar purpose.
10743 Additionally the NO_RSA no longer completely removes the whole -modulus
10744 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10746 [Ralf S. Engelschall]
10748 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10749 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10750 [Arne Ansper <arne@ats.cyber.ee>]
10752 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10753 to be added. Now both 'req' and 'ca' can use new objects defined in the
10757 *) Add cool BIO that does syslog (or event log on NT).
10758 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10760 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10761 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10762 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10763 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10766 *) Add preliminary config info for new extension code.
10769 *) Make RSA_NO_PADDING really use no padding.
10770 [Ulf Moeller <ulf@fitug.de>]
10772 *) Generate errors when private/public key check is done.
10775 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10776 for some CRL extensions and new objects added.
10779 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10780 key usage extension and fuller support for authority key id.
10783 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10784 padding method for RSA, which is recommended for new applications in PKCS
10785 #1 v2.0 (RFC 2437, October 1998).
10786 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10787 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10788 against Bleichbacher's attack on RSA.
10789 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10792 *) Updates to the new SSL compression code
10793 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10795 *) Fix so that the version number in the master secret, when passed
10796 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10797 (because the server will not accept higher), that the version number
10798 is 0x03,0x01, not 0x03,0x00
10799 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10801 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10802 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10803 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10806 *) Support for RAW extensions where an arbitrary extension can be
10807 created by including its DER encoding. See apps/openssl.cnf for
10811 *) Make sure latest Perl versions don't interpret some generated C array
10812 code as Perl array code in the crypto/err/err_genc.pl script.
10813 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10815 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10816 not many people have the assembler. Various Win32 compilation fixes and
10817 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10818 build instructions.
10821 *) Modify configure script 'Configure' to automatically create crypto/date.h
10822 file under Win32 and also build pem.h from pem.org. New script
10823 util/mkfiles.pl to create the MINFO file on environments that can't do a
10824 'make files': perl util/mkfiles.pl >MINFO should work.
10827 *) Major rework of DES function declarations, in the pursuit of correctness
10828 and purity. As a result, many evil casts evaporated, and some weirdness,
10829 too. You may find this causes warnings in your code. Zapping your evil
10830 casts will probably fix them. Mostly.
10833 *) Fix for a typo in asn1.h. Bug fix to object creation script
10834 obj_dat.pl. It considered a zero in an object definition to mean
10835 "end of object": none of the objects in objects.h have any zeros
10836 so it wasn't spotted.
10837 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10839 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10840 Masking (CBCM). In the absence of test vectors, the best I have been able
10841 to do is check that the decrypt undoes the encrypt, so far. Send me test
10842 vectors if you have them.
10845 *) Correct calculation of key length for export ciphers (too much space was
10846 allocated for null ciphers). This has not been tested!
10849 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10850 message is now correct (it understands "crypto" and "ssl" on its
10851 command line). There is also now an "update" option. This will update
10852 the util/ssleay.num and util/libeay.num files with any new functions.
10854 perl util/mkdef.pl crypto ssl update
10855 it will update them.
10858 *) Overhauled the Perl interface (perl/*):
10859 - ported BN stuff to OpenSSL's different BN library
10860 - made the perl/ source tree CVS-aware
10861 - renamed the package from SSLeay to OpenSSL (the files still contain
10862 their history because I've copied them in the repository)
10863 - removed obsolete files (the test scripts will be replaced
10864 by better Test::Harness variants in the future)
10865 [Ralf S. Engelschall]
10867 *) First cut for a very conservative source tree cleanup:
10868 1. merge various obsolete readme texts into doc/ssleay.txt
10869 where we collect the old documents and readme texts.
10870 2. remove the first part of files where I'm already sure that we no
10871 longer need them because of three reasons: either they are just temporary
10872 files which were left by Eric or they are preserved original files where
10873 I've verified that the diff is also available in the CVS via "cvs diff
10874 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10875 the crypto/md/ stuff).
10876 [Ralf S. Engelschall]
10878 *) More extension code. Incomplete support for subject and issuer alt
10879 name, issuer and authority key id. Change the i2v function parameters
10880 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10881 what that's for :-) Fix to ASN1 macro which messed up
10882 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10885 *) Preliminary support for ENUMERATED type. This is largely copied from the
10889 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10890 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10892 *) Make sure `make rehash' target really finds the `openssl' program.
10893 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10895 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10896 like to hear about it if this slows down other processors.
10899 *) Add CygWin32 platform information to Configure script.
10900 [Alan Batie <batie@aahz.jf.intel.com>]
10902 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10903 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10905 *) New program nseq to manipulate netscape certificate sequences
10908 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10912 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10913 but the BN code had some problems that would cause failures when
10914 doing certificate verification and some other functions.
10915 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10917 *) Add ASN1 and PEM code to support netscape certificate sequences.
10920 *) Add ASN1 and PEM code to support netscape certificate sequences.
10923 *) Add several PKIX and private extended key usage OIDs.
10926 *) Modify the 'ca' program to handle the new extension code. Modify
10927 openssl.cnf for new extension format, add comments.
10930 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10931 and add a sample to openssl.cnf so req -x509 now adds appropriate
10935 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10936 error code, add initial support to X509_print() and x509 application.
10939 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10940 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10941 stuff is currently isolated and isn't even compiled yet.
10944 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10945 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10946 Removed the versions check from X509 routines when loading extensions:
10947 this allows certain broken certificates that don't set the version
10948 properly to be processed.
10951 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10952 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10953 can still be regenerated with "make depend".
10956 *) Spelling mistake in C version of CAST-128.
10957 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10959 *) Changes to the error generation code. The perl script err-code.pl
10960 now reads in the old error codes and retains the old numbers, only
10961 adding new ones if necessary. It also only changes the .err files if new
10962 codes are added. The makefiles have been modified to only insert errors
10963 when needed (to avoid needlessly modifying header files). This is done
10964 by only inserting errors if the .err file is newer than the auto generated
10965 C file. To rebuild all the error codes from scratch (the old behaviour)
10966 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10967 or delete all the .err files.
10970 *) CAST-128 was incorrectly implemented for short keys. The C version has
10971 been fixed, but is untested. The assembler versions are also fixed, but
10972 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10973 to regenerate it if needed.
10974 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10975 Hagino <itojun@kame.net>]
10977 *) File was opened incorrectly in randfile.c.
10978 [Ulf Möller <ulf@fitug.de>]
10980 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10981 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10982 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10983 al: it's just almost always a UTCTime. Note this patch adds new error
10984 codes so do a "make errors" if there are problems.
10987 *) Correct Linux 1 recognition in config.
10988 [Ulf Möller <ulf@fitug.de>]
10990 *) Remove pointless MD5 hash when using DSA keys in ca.
10991 [Anonymous <nobody@replay.com>]
10993 *) Generate an error if given an empty string as a cert directory. Also
10994 generate an error if handed NULL (previously returned 0 to indicate an
10995 error, but didn't set one).
10996 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10998 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11001 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11002 parameters. This was causing a warning which killed off the Win32 compile.
11005 *) Remove C++ style comments from crypto/bn/bn_local.h.
11006 [Neil Costigan <neil.costigan@celocom.com>]
11008 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11009 based on a text string, looking up short and long names and finally
11010 "dot" format. The "dot" format stuff didn't work. Added new function
11011 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11012 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11013 OID is not part of the table.
11016 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11017 X509_LOOKUP_by_alias().
11020 *) Sort openssl functions by name.
11023 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11024 encryption from sample DSA keys (in case anyone is interested the password
11028 *) Make _all_ *_free functions accept a NULL pointer.
11029 [Frans Heymans <fheymans@isaserver.be>]
11031 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11033 [Anonymous <nobody@replay.com>]
11035 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11036 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11038 *) Don't blow it for numeric -newkey arguments to apps/req.
11039 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11041 *) Temp key "for export" tests were wrong in s3_srvr.c.
11042 [Anonymous <nobody@replay.com>]
11044 *) Add prototype for temp key callback functions
11045 SSL_CTX_set_tmp_{rsa,dh}_callback().
11048 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11049 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11052 *) X509_name_add_entry() freed the wrong thing after an error.
11053 [Arne Ansper <arne@ats.cyber.ee>]
11055 *) rsa_eay.c would attempt to free a NULL context.
11056 [Arne Ansper <arne@ats.cyber.ee>]
11058 *) BIO_s_socket() had a broken should_retry() on Windoze.
11059 [Arne Ansper <arne@ats.cyber.ee>]
11061 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11062 [Arne Ansper <arne@ats.cyber.ee>]
11064 *) Make sure the already existing X509_STORE->depth variable is initialized
11065 in X509_STORE_new(), but document the fact that this variable is still
11066 unused in the certificate verification process.
11067 [Ralf S. Engelschall]
11069 *) Fix the various library and apps files to free up pkeys obtained from
11070 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11073 *) Fix reference counting in X509_PUBKEY_get(). This makes
11074 demos/maurice/example2.c work, amongst others, probably.
11075 [Steve Henson and Ben Laurie]
11077 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11078 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11079 are no longer created. This way we have a single and consistent command
11080 line interface `openssl <command>', similar to `cvs <command>'.
11081 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11083 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11084 BIT STRING wrapper always have zero unused bits.
11087 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11090 *) Make the top-level INSTALL documentation easier to understand.
11093 *) Makefiles updated to exit if an error occurs in a sub-directory
11094 make (including if user presses ^C) [Paul Sutton]
11096 *) Make Montgomery context stuff explicit in RSA data structure.
11099 *) Fix build order of pem and err to allow for generated pem.h.
11102 *) Fix renumbering bug in X509_NAME_delete_entry().
11105 *) Enhanced the err-ins.pl script so it makes the error library number
11106 global and can add a library name. This is needed for external ASN1 and
11107 other error libraries.
11110 *) Fixed sk_insert which never worked properly.
11113 *) Fix ASN1 macros so they can handle indefinite length constructed
11114 EXPLICIT tags. Some non standard certificates use these: they can now
11118 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11119 into a single doc/ssleay.txt bundle. This way the information is still
11120 preserved but no longer messes up this directory. Now it's new room for
11121 the new set of documentation files.
11122 [Ralf S. Engelschall]
11124 *) SETs were incorrectly DER encoded. This was a major pain, because they
11125 shared code with SEQUENCEs, which aren't coded the same. This means that
11126 almost everything to do with SETs or SEQUENCEs has either changed name or
11127 number of arguments.
11128 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11130 *) Fix test data to work with the above.
11133 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11134 was already fixed by Eric for 0.9.1 it seems.
11135 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11137 *) Autodetect FreeBSD3.
11140 *) Fix various bugs in Configure. This affects the following platforms:
11144 unixware-2.0-pentium
11148 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11149 before they are needed.
11152 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11156 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11158 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11159 changed SSLeay to OpenSSL in version strings.
11160 [Ralf S. Engelschall]
11162 *) Some fixups to the top-level documents.
11165 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11166 because the symlink to include/ was missing.
11167 [Ralf S. Engelschall]
11169 *) Incorporated the popular no-RSA/DSA-only patches
11170 which allow to compile a RSA-free SSLeay.
11171 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11173 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11174 when "ssleay" is still not found.
11175 [Ralf S. Engelschall]
11177 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11178 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11180 *) Updated the README file.
11181 [Ralf S. Engelschall]
11183 *) Added various .cvsignore files in the CVS repository subdirs
11184 to make a "cvs update" really silent.
11185 [Ralf S. Engelschall]
11187 *) Recompiled the error-definition header files and added
11188 missing symbols to the Win32 linker tables.
11189 [Ralf S. Engelschall]
11191 *) Cleaned up the top-level documents;
11192 o new files: CHANGES and LICENSE
11193 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11194 o merged COPYRIGHT into LICENSE
11195 o removed obsolete TODO file
11196 o renamed MICROSOFT to INSTALL.W32
11197 [Ralf S. Engelschall]
11199 *) Removed dummy files from the 0.9.1b source tree:
11200 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11201 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11202 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11203 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11204 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11205 [Ralf S. Engelschall]
11207 *) Added various platform portability fixes.
11210 *) The Genesis of the OpenSSL rpject:
11211 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11212 Young and Tim J. Hudson created while they were working for C2Net until
11214 [The OpenSSL Project]
11217 Changes between 0.9.0b and 0.9.1b [not released]
11219 *) Updated a few CA certificates under certs/
11222 *) Changed some BIGNUM api stuff.
11225 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11226 DGUX x86, Linux Alpha, etc.
11229 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11230 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11234 *) Add -strparse option to asn1pars program which parses nested
11236 [Dr Stephen Henson <shenson@bigfoot.com>]
11238 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11241 *) DSA fix for "ca" program.
11244 *) Added "-genkey" option to "dsaparam" program.
11247 *) Added RIPE MD160 (rmd160) message digest.
11250 *) Added -a (all) option to "ssleay version" command.
11253 *) Added PLATFORM define which is the id given to Configure.
11256 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11259 *) Extended the ASN.1 parser routines.
11262 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11265 *) Added a BN_CTX to the BN library.
11268 *) Fixed the weak key values in DES library
11271 *) Changed API in EVP library for cipher aliases.
11274 *) Added support for RC2/64bit cipher.
11277 *) Converted the lhash library to the crypto/mem.c functions.
11280 *) Added more recognized ASN.1 object ids.
11283 *) Added more RSA padding checks for SSL/TLS.
11286 *) Added BIO proxy/filter functionality.
11289 *) Added extra_certs to SSL_CTX which can be used
11290 send extra CA certificates to the client in the CA cert chain sending
11291 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11294 *) Now Fortezza is denied in the authentication phase because
11295 this is key exchange mechanism is not supported by SSLeay at all.
11298 *) Additional PKCS1 checks.
11301 *) Support the string "TLSv1" for all TLS v1 ciphers.
11304 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11305 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11308 *) Fixed a few memory leaks.
11311 *) Fixed various code and comment typos.
11314 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11315 bytes sent in the client random.
11316 [Edward Bishop <ebishop@spyglass.com>]