5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
8 This changes the decoding behaviour for some invalid messages,
9 though the change is mostly in the more lenient direction, and
10 legacy behaviour is preserved as much as possible.
13 *) New testing framework
14 The testing framework has been largely rewritten and is now using
15 perl and the perl modules Test::Harness and an extended variant of
16 Test::More called OpenSSL::Test to do its work. All test scripts in
17 test/ have been rewritten into test recipes, and all direct calls to
18 executables in test/Makefile have become individual recipes using the
19 simplified testing OpenSSL::Test::Simple.
21 For documentation on our testing modules, do:
23 perldoc test/testlib/OpenSSL/Test/Simple.pm
24 perldoc test/testlib/OpenSSL/Test.pm
28 *) In DSA_generate_parameters_ex, if the provided seed is too short,
30 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
32 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
33 from RFC4279, RFC4785, RFC5487, RFC5489.
35 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
36 original RSA_PSK patch.
39 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
40 era flag was never set throughout the codebase (only read). Also removed
41 SSL3_FLAGS_POP_BUFFER which was only used if
42 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
45 *) Changed the default name options in the "ca", "crl", "req" and "x509"
46 to be "oneline" instead of "compat".
49 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
50 not aware of clients that still exhibit this bug, and the workaround
51 hasn't been working properly for a while.
54 *) The return type of BIO_number_read() and BIO_number_written() as well as
55 the corresponding num_read and num_write members in the BIO structure has
56 changed from unsigned long to uint64_t. On platforms where an unsigned
57 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
61 *) Given the pervasive nature of TLS extensions it is inadvisable to run
62 OpenSSL without support for them. It also means that maintaining
63 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
64 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
67 *) Removed support for the two export grade static DH ciphersuites
68 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
69 were newly added (along with a number of other static DH ciphersuites) to
70 1.0.2. However the two export ones have *never* worked since they were
71 introduced. It seems strange in any case to be adding new export
72 ciphersuites, and given "logjam" it also does not seem correct to fix them.
75 *) Version negotiation has been rewritten. In particular SSLv23_method(),
76 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
77 and turned into macros which simply call the new preferred function names
78 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
79 should use the new names instead. Also as part of this change the ssl23.h
80 header file has been removed.
83 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
84 code and the associated standard is no longer considered fit-for-purpose.
87 *) RT2547 was closed. When generating a private key, try to make the
88 output file readable only by the owner. This behavior change might
89 be noticeable when interacting with other software.
91 *) Added HTTP GET support to the ocsp command.
94 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
97 *) Added support for TLS extended master secret from
98 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
99 initial patch which was a great help during development.
102 *) All libssl internal structures have been removed from the public header
103 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
104 now redundant). Users should not attempt to access internal structures
105 directly. Instead they should use the provided API functions.
108 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
109 Access to deprecated functions can be re-enabled by running config with
110 "enable-deprecated". In addition applications wishing to use deprecated
111 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
112 will, by default, disable some transitive includes that previously existed
113 in the header files (e.g. ec.h will no longer, by default, include bn.h)
116 *) Added support for OCB mode. OpenSSL has been granted a patent license
117 compatible with the OpenSSL license for use of OCB. Details are available
118 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
119 for OCB can be removed by calling config with no-ocb.
122 *) SSLv2 support has been removed. It still supports receiving a SSLv2
123 compatible client hello.
126 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
127 done while fixing the error code for the key-too-small case.
128 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
130 *) CA.sh has been removmed; use CA.pl instead.
133 *) Removed old DES API.
136 *) Remove various unsupported platforms:
142 Sinix/ReliantUNIX RM400
147 16-bit platforms such as WIN16
150 *) Clean up OPENSSL_NO_xxx #define's
151 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
152 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
153 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
154 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
155 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
156 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
157 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
158 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
159 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
160 Remove MS_STATIC; it's a relic from platforms <32 bits.
163 *) Cleaned up dead code
164 Remove all but one '#ifdef undef' which is to be looked at.
167 *) Clean up calling of xxx_free routines.
168 Just like free(), fix most of the xxx_free routines to accept
169 NULL. Remove the non-null checks from callers. Save much code.
172 *) Add secure heap for storage of private keys (when possible).
173 Add BIO_s_secmem(), CBIGNUM, etc.
174 Contributed by Akamai Technologies under our Corporate CLA.
177 *) Experimental support for a new, fast, unbiased prime candidate generator,
178 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
179 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
181 *) New output format NSS in the sess_id command line tool. This allows
182 exporting the session id and the master key in NSS keylog format.
183 [Martin Kaiser <martin@kaiser.cx>]
185 *) Harmonize version and its documentation. -f flag is used to display
187 [mancha <mancha1@zoho.com>]
189 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
190 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
191 [mancha <mancha1@zoho.com>]
193 *) Fix some double frees. These are not thought to be exploitable.
194 [mancha <mancha1@zoho.com>]
196 *) A missing bounds check in the handling of the TLS heartbeat extension
197 can be used to reveal up to 64k of memory to a connected client or
200 Thanks for Neel Mehta of Google Security for discovering this bug and to
201 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
202 preparing the fix (CVE-2014-0160)
203 [Adam Langley, Bodo Moeller]
205 *) Fix for the attack described in the paper "Recovering OpenSSL
206 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
207 by Yuval Yarom and Naomi Benger. Details can be obtained from:
208 http://eprint.iacr.org/2014/140
210 Thanks to Yuval Yarom and Naomi Benger for discovering this
211 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
212 [Yuval Yarom and Naomi Benger]
214 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
215 this fixes a limitation in previous versions of OpenSSL.
218 *) Experimental encrypt-then-mac support.
220 Experimental support for encrypt then mac from
221 draft-gutmann-tls-encrypt-then-mac-02.txt
223 To enable it set the appropriate extension number (0x42 for the test
224 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
226 For non-compliant peers (i.e. just about everything) this should have no
229 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
233 *) Add EVP support for key wrapping algorithms, to avoid problems with
234 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
235 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
236 algorithms and include tests cases.
239 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
243 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
244 MGF1 digest and OAEP label.
247 *) Make openssl verify return errors.
248 [Chris Palmer <palmer@google.com> and Ben Laurie]
250 *) New function ASN1_TIME_diff to calculate the difference between two
251 ASN1_TIME structures or one structure and the current time.
254 *) Update fips_test_suite to support multiple command line options. New
255 test to induce all self test errors in sequence and check expected
259 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
260 sign or verify all in one operation.
263 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
264 test programs and fips_test_suite. Includes functionality to parse
265 the minimal script output of fipsalgest.pl directly.
268 *) Add authorisation parameter to FIPS_module_mode_set().
271 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
274 *) Use separate DRBG fields for internal and external flags. New function
275 FIPS_drbg_health_check() to perform on demand health checking. Add
276 generation tests to fips_test_suite with reduced health check interval to
277 demonstrate periodic health checking. Add "nodh" option to
278 fips_test_suite to skip very slow DH test.
281 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
285 *) More extensive health check for DRBG checking many more failure modes.
286 New function FIPS_selftest_drbg_all() to handle every possible DRBG
287 combination: call this in fips_test_suite.
290 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
291 and POST to handle Dual EC cases.
294 *) Add support for canonical generation of DSA parameter 'g'. See
297 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
298 POST to handle HMAC cases.
301 *) Add functions FIPS_module_version() and FIPS_module_version_text()
302 to return numerical and string versions of the FIPS module number.
305 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
306 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
307 outside the validated module in the FIPS capable OpenSSL.
310 *) Minor change to DRBG entropy callback semantics. In some cases
311 there is no multiple of the block length between min_len and
312 max_len. Allow the callback to return more than max_len bytes
313 of entropy but discard any extra: it is the callback's responsibility
314 to ensure that the extra data discarded does not impact the
315 requested amount of entropy.
318 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
319 information in FIPS186-3, SP800-57 and SP800-131A.
322 *) CCM support via EVP. Interface is very similar to GCM case except we
323 must supply all data in one chunk (i.e. no update, final) and the
324 message length must be supplied if AAD is used. Add algorithm test
328 *) Initial version of POST overhaul. Add POST callback to allow the status
329 of POST to be monitored and/or failures induced. Modify fips_test_suite
330 to use callback. Always run all selftests even if one fails.
333 *) XTS support including algorithm test driver in the fips_gcmtest program.
334 Note: this does increase the maximum key length from 32 to 64 bytes but
335 there should be no binary compatibility issues as existing applications
336 will never use XTS mode.
339 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
340 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
341 performs algorithm blocking for unapproved PRNG types. Also do not
342 set PRNG type in FIPS_mode_set(): leave this to the application.
343 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
344 the standard OpenSSL PRNG: set additional data to a date time vector.
347 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
348 This shouldn't present any incompatibility problems because applications
349 shouldn't be using these directly and any that are will need to rethink
350 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
353 *) Extensive self tests and health checking required by SP800-90 DRBG.
354 Remove strength parameter from FIPS_drbg_instantiate and always
355 instantiate at maximum supported strength.
358 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
361 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
364 *) New function DH_compute_key_padded() to compute a DH key and pad with
365 leading zeroes if needed: this complies with SP800-56A et al.
368 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
369 anything, incomplete, subject to change and largely untested at present.
372 *) Modify fipscanisteronly build option to only build the necessary object
373 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
376 *) Add experimental option FIPSSYMS to give all symbols in
377 fipscanister.o and FIPS or fips prefix. This will avoid
378 conflicts with future versions of OpenSSL. Add perl script
379 util/fipsas.pl to preprocess assembly language source files
380 and rename any affected symbols.
383 *) Add selftest checks and algorithm block of non-fips algorithms in
384 FIPS mode. Remove DES2 from selftests.
387 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
388 return internal method without any ENGINE dependencies. Add new
389 tiny fips sign and verify functions.
392 *) New build option no-ec2m to disable characteristic 2 code.
395 *) New build option "fipscanisteronly". This only builds fipscanister.o
396 and (currently) associated fips utilities. Uses the file Makefile.fips
397 instead of Makefile.org as the prototype.
400 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
401 Update fips_gcmtest to use IV generator.
404 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
405 setting output buffer to NULL. The *Final function must be
406 called although it will not retrieve any additional data. The tag
407 can be set or retrieved with a ctrl. The IV length is by default 12
408 bytes (96 bits) but can be set to an alternative value. If the IV
409 length exceeds the maximum IV length (currently 16 bytes) it cannot be
413 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
414 underlying do_cipher function handles all cipher semantics itself
415 including padding and finalisation. This is useful if (for example)
416 an ENGINE cipher handles block padding itself. The behaviour of
417 do_cipher is subtly changed if this flag is set: the return value
418 is the number of characters written to the output buffer (zero is
419 no longer an error code) or a negative error code. Also if the
420 input buffer is NULL and length 0 finalisation should be performed.
423 *) If a candidate issuer certificate is already part of the constructed
424 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
427 *) Improve forward-security support: add functions
429 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
430 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
432 for use by SSL/TLS servers; the callback function will be called whenever a
433 new session is created, and gets to decide whether the session may be
434 cached to make it resumable (return 0) or not (return 1). (As by the
435 SSL/TLS protocol specifications, the session_id sent by the server will be
436 empty to indicate that the session is not resumable; also, the server will
437 not generate RFC 4507 (RFC 5077) session tickets.)
439 A simple reasonable callback implementation is to return is_forward_secure.
440 This parameter will be set to 1 or 0 depending on the ciphersuite selected
441 by the SSL/TLS server library, indicating whether it can provide forward
443 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
445 *) New -verify_name option in command line utilities to set verification
449 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
450 Add CMAC pkey methods.
453 *) Experimental renegotiation in s_server -www mode. If the client
454 browses /reneg connection is renegotiated. If /renegcert it is
455 renegotiated requesting a certificate.
458 *) Add an "external" session cache for debugging purposes to s_server. This
459 should help trace issues which normally are only apparent in deployed
460 multi-process servers.
463 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
464 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
465 BIO_set_cipher() and some obscure PEM functions were changed so they
466 can now return an error. The RAND changes required a change to the
467 RAND_METHOD structure.
470 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
471 a gcc attribute to warn if the result of a function is ignored. This
472 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
473 whose return value is often ignored.
476 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
478 *) Alternate chains certificate forgery
480 During certificate verfification, OpenSSL will attempt to find an
481 alternative certificate chain if the first attempt to build such a chain
482 fails. An error in the implementation of this logic can mean that an
483 attacker could cause certain checks on untrusted certificates to be
484 bypassed, such as the CA flag, enabling them to use a valid leaf
485 certificate to act as a CA and "issue" an invalid certificate.
487 This issue was reported to OpenSSL by Adam Langley/David Benjamin
491 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
493 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
494 incompatibility in the handling of HMAC. The previous ABI has now been
498 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
500 *) Malformed ECParameters causes infinite loop
502 When processing an ECParameters structure OpenSSL enters an infinite loop
503 if the curve specified is over a specially malformed binary polynomial
506 This can be used to perform denial of service against any
507 system which processes public keys, certificate requests or
508 certificates. This includes TLS clients and TLS servers with
509 client authentication enabled.
511 This issue was reported to OpenSSL by Joseph Barr-Pixton.
515 *) Exploitable out-of-bounds read in X509_cmp_time
517 X509_cmp_time does not properly check the length of the ASN1_TIME
518 string and can read a few bytes out of bounds. In addition,
519 X509_cmp_time accepts an arbitrary number of fractional seconds in the
522 An attacker can use this to craft malformed certificates and CRLs of
523 various sizes and potentially cause a segmentation fault, resulting in
524 a DoS on applications that verify certificates or CRLs. TLS clients
525 that verify CRLs are affected. TLS clients and servers with client
526 authentication enabled may be affected if they use custom verification
529 This issue was reported to OpenSSL by Robert Swiecki (Google), and
530 independently by Hanno Böck.
534 *) PKCS7 crash with missing EnvelopedContent
536 The PKCS#7 parsing code does not handle missing inner EncryptedContent
537 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
538 with missing content and trigger a NULL pointer dereference on parsing.
540 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
541 structures from untrusted sources are affected. OpenSSL clients and
542 servers are not affected.
544 This issue was reported to OpenSSL by Michal Zalewski (Google).
548 *) CMS verify infinite loop with unknown hash function
550 When verifying a signedData message the CMS code can enter an infinite loop
551 if presented with an unknown hash function OID. This can be used to perform
552 denial of service against any system which verifies signedData messages using
554 This issue was reported to OpenSSL by Johannes Bauer.
558 *) Race condition handling NewSessionTicket
560 If a NewSessionTicket is received by a multi-threaded client when attempting to
561 reuse a previous ticket then a race condition can occur potentially leading to
562 a double free of the ticket data.
566 *) Only support 256-bit or stronger elliptic curves with the
567 'ecdh_auto' setting (server) or by default (client). Of supported
568 curves, prefer P-256 (both).
571 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
573 *) ClientHello sigalgs DoS fix
575 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
576 invalid signature algorithms extension a NULL pointer dereference will
577 occur. This can be exploited in a DoS attack against the server.
579 This issue was was reported to OpenSSL by David Ramos of Stanford
582 [Stephen Henson and Matt Caswell]
584 *) Multiblock corrupted pointer fix
586 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
587 feature only applies on 64 bit x86 architecture platforms that support AES
588 NI instructions. A defect in the implementation of "multiblock" can cause
589 OpenSSL's internal write buffer to become incorrectly set to NULL when
590 using non-blocking IO. Typically, when the user application is using a
591 socket BIO for writing, this will only result in a failed connection.
592 However if some other BIO is used then it is likely that a segmentation
593 fault will be triggered, thus enabling a potential DoS attack.
595 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
599 *) Segmentation fault in DTLSv1_listen fix
601 The DTLSv1_listen function is intended to be stateless and processes the
602 initial ClientHello from many peers. It is common for user code to loop
603 over the call to DTLSv1_listen until a valid ClientHello is received with
604 an associated cookie. A defect in the implementation of DTLSv1_listen means
605 that state is preserved in the SSL object from one invocation to the next
606 that can lead to a segmentation fault. Errors processing the initial
607 ClientHello can trigger this scenario. An example of such an error could be
608 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
611 This issue was reported to OpenSSL by Per Allansson.
615 *) Segmentation fault in ASN1_TYPE_cmp fix
617 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
618 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
619 certificate signature algorithm consistency this can be used to crash any
620 certificate verification operation and exploited in a DoS attack. Any
621 application which performs certificate verification is vulnerable including
622 OpenSSL clients and servers which enable client authentication.
626 *) Segmentation fault for invalid PSS parameters fix
628 The signature verification routines will crash with a NULL pointer
629 dereference if presented with an ASN.1 signature using the RSA PSS
630 algorithm and invalid parameters. Since these routines are used to verify
631 certificate signature algorithms this can be used to crash any
632 certificate verification operation and exploited in a DoS attack. Any
633 application which performs certificate verification is vulnerable including
634 OpenSSL clients and servers which enable client authentication.
636 This issue was was reported to OpenSSL by Brian Carpenter.
640 *) ASN.1 structure reuse memory corruption fix
642 Reusing a structure in ASN.1 parsing may allow an attacker to cause
643 memory corruption via an invalid write. Such reuse is and has been
644 strongly discouraged and is believed to be rare.
646 Applications that parse structures containing CHOICE or ANY DEFINED BY
647 components may be affected. Certificate parsing (d2i_X509 and related
648 functions) are however not affected. OpenSSL clients and servers are
653 *) PKCS7 NULL pointer dereferences fix
655 The PKCS#7 parsing code does not handle missing outer ContentInfo
656 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
657 missing content and trigger a NULL pointer dereference on parsing.
659 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
660 otherwise parse PKCS#7 structures from untrusted sources are
661 affected. OpenSSL clients and servers are not affected.
663 This issue was reported to OpenSSL by Michal Zalewski (Google).
667 *) DoS via reachable assert in SSLv2 servers fix
669 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
670 servers that both support SSLv2 and enable export cipher suites by sending
671 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
673 This issue was discovered by Sean Burford (Google) and Emilia Käsper
674 (OpenSSL development team).
678 *) Empty CKE with client auth and DHE fix
680 If client auth is used then a server can seg fault in the event of a DHE
681 ciphersuite being selected and a zero length ClientKeyExchange message
682 being sent by the client. This could be exploited in a DoS attack.
686 *) Handshake with unseeded PRNG fix
688 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
689 with an unseeded PRNG. The conditions are:
690 - The client is on a platform where the PRNG has not been seeded
691 automatically, and the user has not seeded manually
692 - A protocol specific client method version has been used (i.e. not
693 SSL_client_methodv23)
694 - A ciphersuite is used that does not require additional random data from
695 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
697 If the handshake succeeds then the client random that has been used will
698 have been generated from a PRNG with insufficient entropy and therefore the
699 output may be predictable.
701 For example using the following command with an unseeded openssl will
702 succeed on an unpatched platform:
704 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
708 *) Use After Free following d2i_ECPrivatekey error fix
710 A malformed EC private key file consumed via the d2i_ECPrivateKey function
711 could cause a use after free condition. This, in turn, could cause a double
712 free in several private key parsing functions (such as d2i_PrivateKey
713 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
714 for applications that receive EC private keys from untrusted
715 sources. This scenario is considered rare.
717 This issue was discovered by the BoringSSL project and fixed in their
722 *) X509_to_X509_REQ NULL pointer deref fix
724 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
725 the certificate key is invalid. This function is rarely used in practice.
727 This issue was discovered by Brian Carpenter.
731 *) Removed the export ciphers from the DEFAULT ciphers
734 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
736 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
737 ARMv5 through ARMv8, as opposite to "locking" it to single one.
738 So far those who have to target multiple plaforms would compromise
739 and argue that binary targeting say ARMv5 would still execute on
740 ARMv8. "Universal" build resolves this compromise by providing
741 near-optimal performance even on newer platforms.
744 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
745 (other platforms pending).
746 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
748 *) Add support for the SignedCertificateTimestampList certificate and
749 OCSP response extensions from RFC6962.
752 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
753 for corner cases. (Certain input points at infinity could lead to
754 bogus results, with non-infinity inputs mapped to infinity too.)
757 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
758 This covers AES, SHA256/512 and GHASH. "Initial" means that most
759 common cases are optimized and there still is room for further
760 improvements. Vector Permutation AES for Altivec is also added.
763 *) Add support for little-endian ppc64 Linux target.
764 [Marcelo Cerri (IBM)]
766 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
767 SHA1, SHA256 and GHASH. "Initial" means that most common cases
768 are optimized and there still is room for further improvements.
769 Both 32- and 64-bit modes are supported.
770 [Andy Polyakov, Ard Biesheuvel (Linaro)]
772 *) Improved ARMv7 NEON support.
775 *) Support for SPARC Architecture 2011 crypto extensions, first
776 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
777 SHA256/512, MD5, GHASH and modular exponentiation.
778 [Andy Polyakov, David Miller]
780 *) Accelerated modular exponentiation for Intel processors, a.k.a.
782 [Shay Gueron & Vlad Krasnov (Intel Corp)]
784 *) Support for new and upcoming Intel processors, including AVX2,
785 BMI and SHA ISA extensions. This includes additional "stitched"
786 implementations, AESNI-SHA256 and GCM, and multi-buffer support
789 This work was sponsored by Intel Corp.
792 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
793 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
794 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
797 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
798 this fixes a limiation in previous versions of OpenSSL.
801 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
802 MGF1 digest and OAEP label.
805 *) Add EVP support for key wrapping algorithms, to avoid problems with
806 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
807 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
808 algorithms and include tests cases.
811 *) Add functions to allocate and set the fields of an ECDSA_METHOD
813 [Douglas E. Engert, Steve Henson]
815 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
816 difference in days and seconds between two tm or ASN1_TIME structures.
819 *) Add -rev test option to s_server to just reverse order of characters
820 received by client and send back to server. Also prints an abbreviated
821 summary of the connection parameters.
824 *) New option -brief for s_client and s_server to print out a brief summary
825 of connection parameters.
828 *) Add callbacks for arbitrary TLS extensions.
829 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
831 *) New option -crl_download in several openssl utilities to download CRLs
832 from CRLDP extension in certificates.
835 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
838 *) New function X509_CRL_diff to generate a delta CRL from the difference
839 of two full CRLs. Add support to "crl" utility.
842 *) New functions to set lookup_crls function and to retrieve
843 X509_STORE from X509_STORE_CTX.
846 *) Print out deprecated issuer and subject unique ID fields in
850 *) Extend OCSP I/O functions so they can be used for simple general purpose
851 HTTP as well as OCSP. New wrapper function which can be used to download
852 CRLs using the OCSP API.
855 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
858 *) SSL_CONF* functions. These provide a common framework for application
859 configuration using configuration files or command lines.
862 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
863 message callback and prints the results. Needs compile time option
864 "enable-ssl-trace". New options to s_client and s_server to enable
868 *) New ctrl and macro to retrieve supported points extensions.
869 Print out extension in s_server and s_client.
872 *) New functions to retrieve certificate signature and signature
876 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
880 *) New Suite B modes for TLS code. These use and enforce the requirements
881 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
882 only use Suite B curves. The Suite B modes can be set by using the
883 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
886 *) New chain verification flags for Suite B levels of security. Check
887 algorithms are acceptable when flags are set in X509_verify_cert.
890 *) Make tls1_check_chain return a set of flags indicating checks passed
891 by a certificate chain. Add additional tests to handle client
892 certificates: checks for matching certificate type and issuer name
896 *) If an attempt is made to use a signature algorithm not in the peer
897 preference list abort the handshake. If client has no suitable
898 signature algorithms in response to a certificate request do not
902 *) If server EC tmp key is not in client preference list abort handshake.
905 *) Add support for certificate stores in CERT structure. This makes it
906 possible to have different stores per SSL structure or one store in
907 the parent SSL_CTX. Include distint stores for certificate chain
908 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
909 to build and store a certificate chain in CERT structure: returing
910 an error if the chain cannot be built: this will allow applications
911 to test if a chain is correctly configured.
913 Note: if the CERT based stores are not set then the parent SSL_CTX
914 store is used to retain compatibility with existing behaviour.
918 *) New function ssl_set_client_disabled to set a ciphersuite disabled
919 mask based on the current session, check mask when sending client
920 hello and checking the requested ciphersuite.
923 *) New ctrls to retrieve and set certificate types in a certificate
924 request message. Print out received values in s_client. If certificate
925 types is not set with custom values set sensible values based on
926 supported signature algorithms.
929 *) Support for distinct client and server supported signature algorithms.
932 *) Add certificate callback. If set this is called whenever a certificate
933 is required by client or server. An application can decide which
934 certificate chain to present based on arbitrary criteria: for example
935 supported signature algorithms. Add very simple example to s_server.
936 This fixes many of the problems and restrictions of the existing client
937 certificate callback: for example you can now clear an existing
938 certificate and specify the whole chain.
941 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
942 the certificate can be used for (if anything). Set valid_flags field
943 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
944 to have similar checks in it.
946 Add new "cert_flags" field to CERT structure and include a "strict mode".
947 This enforces some TLS certificate requirements (such as only permitting
948 certificate signature algorithms contained in the supported algorithms
949 extension) which some implementations ignore: this option should be used
950 with caution as it could cause interoperability issues.
953 *) Update and tidy signature algorithm extension processing. Work out
954 shared signature algorithms based on preferences and peer algorithms
955 and print them out in s_client and s_server. Abort handshake if no
956 shared signature algorithms.
959 *) Add new functions to allow customised supported signature algorithms
960 for SSL and SSL_CTX structures. Add options to s_client and s_server
964 *) New function SSL_certs_clear() to delete all references to certificates
965 from an SSL structure. Before this once a certificate had been added
966 it couldn't be removed.
969 *) Integrate hostname, email address and IP address checking with certificate
970 verification. New verify options supporting checking in opensl utility.
973 *) Fixes and wildcard matching support to hostname and email checking
974 functions. Add manual page.
975 [Florian Weimer (Red Hat Product Security Team)]
977 *) New functions to check a hostname email or IP address against a
978 certificate. Add options x509 utility to print results of checks against
982 *) Fix OCSP checking.
983 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
985 *) Initial experimental support for explicitly trusted non-root CAs.
986 OpenSSL still tries to build a complete chain to a root but if an
987 intermediate CA has a trust setting included that is used. The first
988 setting is used: whether to trust (e.g., -addtrust option to the x509
992 *) Add -trusted_first option which attempts to find certificates in the
993 trusted store even if an untrusted chain is also supplied.
996 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
997 platform support for Linux and Android.
1000 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1003 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1004 When in FIPS mode the approved implementations are used as normal,
1005 when not in FIPS mode the internal unapproved versions are used instead.
1006 This means that the FIPS capable OpenSSL isn't forced to use the
1007 (often lower performance) FIPS implementations outside FIPS mode.
1010 *) Transparently support X9.42 DH parameters when calling
1011 PEM_read_bio_DHparameters. This means existing applications can handle
1012 the new parameter format automatically.
1015 *) Initial experimental support for X9.42 DH parameter format: mainly
1016 to support use of 'q' parameter for RFC5114 parameters.
1019 *) Add DH parameters from RFC5114 including test data to dhtest.
1022 *) Support for automatic EC temporary key parameter selection. If enabled
1023 the most preferred EC parameters are automatically used instead of
1024 hardcoded fixed parameters. Now a server just has to call:
1025 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1026 support ECDH and use the most appropriate parameters.
1029 *) Enhance and tidy EC curve and point format TLS extension code. Use
1030 static structures instead of allocation if default values are used.
1031 New ctrls to set curves we wish to support and to retrieve shared curves.
1032 Print out shared curves in s_server. New options to s_server and s_client
1033 to set list of supported curves.
1036 *) New ctrls to retrieve supported signature algorithms and
1037 supported curve values as an array of NIDs. Extend openssl utility
1038 to print out received values.
1041 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1042 between NIDs and the more common NIST names such as "P-256". Enhance
1043 ecparam utility and ECC method to recognise the NIST names for curves.
1046 *) Enhance SSL/TLS certificate chain handling to support different
1047 chains for each certificate instead of one chain in the parent SSL_CTX.
1050 *) Support for fixed DH ciphersuite client authentication: where both
1051 server and client use DH certificates with common parameters.
1054 *) Support for fixed DH ciphersuites: those requiring DH server
1058 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1060 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1061 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1062 X509_CINF_get_signature were reverted post internal team review.
1064 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1066 *) Build fixes for the Windows and OpenVMS platforms
1067 [Matt Caswell and Richard Levitte]
1069 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1071 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1072 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1073 dereference. This could lead to a Denial Of Service attack. Thanks to
1074 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1078 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1079 dtls1_buffer_record function under certain conditions. In particular this
1080 could occur if an attacker sent repeated DTLS records with the same
1081 sequence number but for the next epoch. The memory leak could be exploited
1082 by an attacker in a Denial of Service attack through memory exhaustion.
1083 Thanks to Chris Mueller for reporting this issue.
1087 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1088 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1089 method would be set to NULL which could later result in a NULL pointer
1090 dereference. Thanks to Frank Schmirler for reporting this issue.
1094 *) Abort handshake if server key exchange message is omitted for ephemeral
1097 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1098 reporting this issue.
1102 *) Remove non-export ephemeral RSA code on client and server. This code
1103 violated the TLS standard by allowing the use of temporary RSA keys in
1104 non-export ciphersuites and could be used by a server to effectively
1105 downgrade the RSA key length used to a value smaller than the server
1106 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1107 INRIA or reporting this issue.
1111 *) Fixed issue where DH client certificates are accepted without verification.
1112 An OpenSSL server will accept a DH certificate for client authentication
1113 without the certificate verify message. This effectively allows a client to
1114 authenticate without the use of a private key. This only affects servers
1115 which trust a client certificate authority which issues certificates
1116 containing DH keys: these are extremely rare and hardly ever encountered.
1117 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1122 *) Ensure that the session ID context of an SSL is updated when its
1123 SSL_CTX is updated via SSL_set_SSL_CTX.
1125 The session ID context is typically set from the parent SSL_CTX,
1126 and can vary with the CTX.
1129 *) Fix various certificate fingerprint issues.
1131 By using non-DER or invalid encodings outside the signed portion of a
1132 certificate the fingerprint can be changed without breaking the signature.
1133 Although no details of the signed portion of the certificate can be changed
1134 this can cause problems with some applications: e.g. those using the
1135 certificate fingerprint for blacklists.
1137 1. Reject signatures with non zero unused bits.
1139 If the BIT STRING containing the signature has non zero unused bits reject
1140 the signature. All current signature algorithms require zero unused bits.
1142 2. Check certificate algorithm consistency.
1144 Check the AlgorithmIdentifier inside TBS matches the one in the
1145 certificate signature. NB: this will result in signature failure
1146 errors for some broken certificates.
1148 Thanks to Konrad Kraszewski from Google for reporting this issue.
1150 3. Check DSA/ECDSA signatures use DER.
1152 Reencode DSA/ECDSA signatures and compare with the original received
1153 signature. Return an error if there is a mismatch.
1155 This will reject various cases including garbage after signature
1156 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
1157 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
1158 (negative or with leading zeroes).
1160 Further analysis was conducted and fixes were developed by Stephen Henson
1161 of the OpenSSL core team.
1166 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1167 results on some platforms, including x86_64. This bug occurs at random
1168 with a very low probability, and is not known to be exploitable in any
1169 way, though its exact impact is difficult to determine. Thanks to Pieter
1170 Wuille (Blockstream) who reported this issue and also suggested an initial
1171 fix. Further analysis was conducted by the OpenSSL development team and
1172 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1173 the OpenSSL core team.
1177 *) Do not resume sessions on the server if the negotiated protocol
1178 version does not match the session's version. Resuming with a different
1179 version, while not strictly forbidden by the RFC, is of questionable
1180 sanity and breaks all known clients.
1181 [David Benjamin, Emilia Käsper]
1183 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1184 early CCS messages during renegotiation. (Note that because
1185 renegotiation is encrypted, this early CCS was not exploitable.)
1188 *) Tighten client-side session ticket handling during renegotiation:
1189 ensure that the client only accepts a session ticket if the server sends
1190 the extension anew in the ServerHello. Previously, a TLS client would
1191 reuse the old extension state and thus accept a session ticket if one was
1192 announced in the initial ServerHello.
1194 Similarly, ensure that the client requires a session ticket if one
1195 was advertised in the ServerHello. Previously, a TLS client would
1196 ignore a missing NewSessionTicket message.
1199 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1201 *) SRTP Memory Leak.
1203 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1204 sends a carefully crafted handshake message, to cause OpenSSL to fail
1205 to free up to 64k of memory causing a memory leak. This could be
1206 exploited in a Denial Of Service attack. This issue affects OpenSSL
1207 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1208 whether SRTP is used or configured. Implementations of OpenSSL that
1209 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1211 The fix was developed by the OpenSSL team.
1215 *) Session Ticket Memory Leak.
1217 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1218 integrity of that ticket is first verified. In the event of a session
1219 ticket integrity check failing, OpenSSL will fail to free memory
1220 causing a memory leak. By sending a large number of invalid session
1221 tickets an attacker could exploit this issue in a Denial Of Service
1226 *) Build option no-ssl3 is incomplete.
1228 When OpenSSL is configured with "no-ssl3" as a build option, servers
1229 could accept and complete a SSL 3.0 handshake, and clients could be
1230 configured to send them.
1232 [Akamai and the OpenSSL team]
1234 *) Add support for TLS_FALLBACK_SCSV.
1235 Client applications doing fallback retries should call
1236 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1238 [Adam Langley, Bodo Moeller]
1240 *) Add additional DigestInfo checks.
1242 Reencode DigestInto in DER and check against the original when
1243 verifying RSA signature: this will reject any improperly encoded
1244 DigestInfo structures.
1246 Note: this is a precautionary measure and no attacks are currently known.
1250 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1252 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1253 SRP code can be overrun an internal buffer. Add sanity check that
1254 g, A, B < N to SRP code.
1256 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1257 Group for discovering this issue.
1261 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1262 TLS 1.0 instead of higher protocol versions when the ClientHello message
1263 is badly fragmented. This allows a man-in-the-middle attacker to force a
1264 downgrade to TLS 1.0 even if both the server and the client support a
1265 higher protocol version, by modifying the client's TLS records.
1267 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1268 researching this issue.
1272 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1273 to a denial of service attack. A malicious server can crash the client
1274 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1275 ciphersuite and sending carefully crafted handshake messages.
1277 Thanks to Felix Gröbert (Google) for discovering and researching this
1282 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1283 to leak memory. This can be exploited through a Denial of Service attack.
1284 Thanks to Adam Langley for discovering and researching this issue.
1288 *) An attacker can force openssl to consume large amounts of memory whilst
1289 processing DTLS handshake messages. This can be exploited through a
1290 Denial of Service attack.
1291 Thanks to Adam Langley for discovering and researching this issue.
1295 *) An attacker can force an error condition which causes openssl to crash
1296 whilst processing DTLS packets due to memory being freed twice. This
1297 can be exploited through a Denial of Service attack.
1298 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1303 *) If a multithreaded client connects to a malicious server using a resumed
1304 session and the server sends an ec point format extension it could write
1305 up to 255 bytes to freed memory.
1307 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1312 *) A malicious server can crash an OpenSSL client with a null pointer
1313 dereference (read) by specifying an SRP ciphersuite even though it was not
1314 properly negotiated with the client. This can be exploited through a
1315 Denial of Service attack.
1317 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1318 discovering and researching this issue.
1322 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1323 X509_name_oneline, X509_name_print_ex et al. to leak some information
1324 from the stack. Applications may be affected if they echo pretty printing
1325 output to the attacker.
1327 Thanks to Ivan Fratric (Google) for discovering this issue.
1329 [Emilia Käsper, and Steve Henson]
1331 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1332 for corner cases. (Certain input points at infinity could lead to
1333 bogus results, with non-infinity inputs mapped to infinity too.)
1336 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1338 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1339 handshake can force the use of weak keying material in OpenSSL
1340 SSL/TLS clients and servers.
1342 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1343 researching this issue. (CVE-2014-0224)
1344 [KIKUCHI Masashi, Steve Henson]
1346 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1347 OpenSSL DTLS client the code can be made to recurse eventually crashing
1350 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1352 [Imre Rad, Steve Henson]
1354 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1355 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1356 client or server. This is potentially exploitable to run arbitrary
1357 code on a vulnerable client or server.
1359 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1360 [Jüri Aedla, Steve Henson]
1362 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1363 are subject to a denial of service attack.
1365 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1366 this issue. (CVE-2014-3470)
1367 [Felix Gröbert, Ivan Fratric, Steve Henson]
1369 *) Harmonize version and its documentation. -f flag is used to display
1371 [mancha <mancha1@zoho.com>]
1373 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1374 in i2d_ECPrivateKey.
1375 [mancha <mancha1@zoho.com>]
1377 *) Fix some double frees. These are not thought to be exploitable.
1378 [mancha <mancha1@zoho.com>]
1380 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1382 *) A missing bounds check in the handling of the TLS heartbeat extension
1383 can be used to reveal up to 64k of memory to a connected client or
1386 Thanks for Neel Mehta of Google Security for discovering this bug and to
1387 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1388 preparing the fix (CVE-2014-0160)
1389 [Adam Langley, Bodo Moeller]
1391 *) Fix for the attack described in the paper "Recovering OpenSSL
1392 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1393 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1394 http://eprint.iacr.org/2014/140
1396 Thanks to Yuval Yarom and Naomi Benger for discovering this
1397 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1398 [Yuval Yarom and Naomi Benger]
1400 *) TLS pad extension: draft-agl-tls-padding-03
1402 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1403 TLS client Hello record length value would otherwise be > 255 and
1404 less that 512 pad with a dummy extension containing zeroes so it
1405 is at least 512 bytes long.
1407 [Adam Langley, Steve Henson]
1409 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1411 *) Fix for TLS record tampering bug. A carefully crafted invalid
1412 handshake could crash OpenSSL with a NULL pointer exception.
1413 Thanks to Anton Johansson for reporting this issues.
1416 *) Keep original DTLS digest and encryption contexts in retransmission
1417 structures so we can use the previous session parameters if they need
1418 to be resent. (CVE-2013-6450)
1421 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1422 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1423 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1424 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1425 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1426 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1427 [Rob Stradling, Adam Langley]
1429 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1431 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1432 supporting platforms or when small records were transferred.
1433 [Andy Polyakov, Steve Henson]
1435 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1437 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1439 This addresses the flaw in CBC record processing discovered by
1440 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1441 at: http://www.isg.rhul.ac.uk/tls/
1443 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1444 Security Group at Royal Holloway, University of London
1445 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1446 Emilia Käsper for the initial patch.
1448 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1450 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1451 ciphersuites which can be exploited in a denial of service attack.
1452 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1453 and detecting this bug and to Wolfgang Ettlinger
1454 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1458 *) Return an error when checking OCSP signatures when key is NULL.
1459 This fixes a DoS attack. (CVE-2013-0166)
1462 *) Make openssl verify return errors.
1463 [Chris Palmer <palmer@google.com> and Ben Laurie]
1465 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1466 the right response is stapled. Also change SSL_get_certificate()
1467 so it returns the certificate actually sent.
1468 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1469 [Rob Stradling <rob.stradling@comodo.com>]
1471 *) Fix possible deadlock when decoding public keys.
1474 *) Don't use TLS 1.0 record version number in initial client hello
1478 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1480 *) Sanity check record length before skipping explicit IV in TLS
1481 1.2, 1.1 and DTLS to fix DoS attack.
1483 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1484 fuzzing as a service testing platform.
1488 *) Initialise tkeylen properly when encrypting CMS messages.
1489 Thanks to Solar Designer of Openwall for reporting this issue.
1492 *) In FIPS mode don't try to use composite ciphers as they are not
1496 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1498 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1499 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1500 mean any application compiled against OpenSSL 1.0.0 headers setting
1501 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1502 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1503 0x10000000L Any application which was previously compiled against
1504 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1505 will need to be recompiled as a result. Letting be results in
1506 inability to disable specifically TLS 1.1 and in client context,
1507 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1510 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1511 disable just protocol X, but all protocols above X *if* there are
1512 protocols *below* X still enabled. In more practical terms it means
1513 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1514 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1515 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1519 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1521 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1522 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1523 in CRYPTO_realloc_clean.
1525 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1526 issue and to Adam Langley <agl@chromium.org> for fixing it.
1528 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1530 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1533 *) Workarounds for some broken servers that "hang" if a client hello
1534 record length exceeds 255 bytes.
1536 1. Do not use record version number > TLS 1.0 in initial client
1537 hello: some (but not all) hanging servers will now work.
1538 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1539 the number of ciphers sent in the client hello. This should be
1540 set to an even number, such as 50, for example by passing:
1541 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1542 Most broken servers should now work.
1543 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1544 TLS 1.2 client support entirely.
1547 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1550 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1552 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1553 STRING form instead of a DigestInfo.
1556 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1557 and the RSA_sign/RSA_verify functions. This was made more apparent when
1558 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1559 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1560 the correct format in RSA_verify so both forms transparently work.
1563 *) Some servers which support TLS 1.0 can choke if we initially indicate
1564 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1565 encrypted premaster secret. As a workaround use the maximum permitted
1566 client version in client hello, this should keep such servers happy
1567 and still work with previous versions of OpenSSL.
1570 *) Add support for TLS/DTLS heartbeats.
1571 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1573 *) Add support for SCTP.
1574 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1576 *) Improved PRNG seeding for VOS.
1577 [Paul Green <Paul.Green@stratus.com>]
1579 *) Extensive assembler packs updates, most notably:
1581 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1582 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1583 - x86_64: bit-sliced AES implementation;
1584 - ARM: NEON support, contemporary platforms optimizations;
1585 - s390x: z196 support;
1586 - *: GHASH and GF(2^m) multiplication implementations;
1590 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1591 (removal of unnecessary code)
1592 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1594 *) Add TLS key material exporter from RFC 5705.
1597 *) Add DTLS-SRTP negotiation from RFC 5764.
1600 *) Add Next Protocol Negotiation,
1601 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1602 disabled with a no-npn flag to config or Configure. Code donated
1604 [Adam Langley <agl@google.com> and Ben Laurie]
1606 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1607 NIST-P256, NIST-P521, with constant-time single point multiplication on
1608 typical inputs. Compiler support for the nonstandard type __uint128_t is
1609 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1610 Code made available under Apache License version 2.0.
1612 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1613 line to include this in your build of OpenSSL, and run "make depend" (or
1614 "make update"). This enables the following EC_METHODs:
1616 EC_GFp_nistp224_method()
1617 EC_GFp_nistp256_method()
1618 EC_GFp_nistp521_method()
1620 EC_GROUP_new_by_curve_name() will automatically use these (while
1621 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1623 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1625 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1626 all platforms. Move ssize_t definition from e_os.h to the public
1627 header file e_os2.h as it now appears in public header file cms.h
1630 *) New -sigopt option to the ca, req and x509 utilities. Additional
1631 signature parameters can be passed using this option and in
1635 *) Add RSA PSS signing function. This will generate and set the
1636 appropriate AlgorithmIdentifiers for PSS based on those in the
1637 corresponding EVP_MD_CTX structure. No application support yet.
1640 *) Support for companion algorithm specific ASN1 signing routines.
1641 New function ASN1_item_sign_ctx() signs a pre-initialised
1642 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1643 the appropriate parameters.
1646 *) Add new algorithm specific ASN1 verification initialisation function
1647 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1648 handling will be the same no matter what EVP_PKEY_METHOD is used.
1649 Add a PSS handler to support verification of PSS signatures: checked
1650 against a number of sample certificates.
1653 *) Add signature printing for PSS. Add PSS OIDs.
1654 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1656 *) Add algorithm specific signature printing. An individual ASN1 method
1657 can now print out signatures instead of the standard hex dump.
1659 More complex signatures (e.g. PSS) can print out more meaningful
1660 information. Include DSA version that prints out the signature
1664 *) Password based recipient info support for CMS library: implementing
1668 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1669 neatly separates the code into cipher and PBE sections and is required
1670 for some algorithms that split PBES2 into separate pieces (such as
1671 password based CMS).
1674 *) Session-handling fixes:
1675 - Fix handling of connections that are resuming with a session ID,
1676 but also support Session Tickets.
1677 - Fix a bug that suppressed issuing of a new ticket if the client
1678 presented a ticket with an expired session.
1679 - Try to set the ticket lifetime hint to something reasonable.
1680 - Make tickets shorter by excluding irrelevant information.
1681 - On the client side, don't ignore renewed tickets.
1682 [Adam Langley, Bodo Moeller (Google)]
1684 *) Fix PSK session representation.
1687 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1689 This work was sponsored by Intel.
1692 *) Add GCM support to TLS library. Some custom code is needed to split
1693 the IV between the fixed (from PRF) and explicit (from TLS record)
1694 portions. This adds all GCM ciphersuites supported by RFC5288 and
1695 RFC5289. Generalise some AES* cipherstrings to include GCM and
1696 add a special AESGCM string for GCM only.
1699 *) Expand range of ctrls for AES GCM. Permit setting invocation
1700 field on decrypt and retrieval of invocation field only on encrypt.
1703 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1704 As required by RFC5289 these ciphersuites cannot be used if for
1705 versions of TLS earlier than 1.2.
1708 *) For FIPS capable OpenSSL interpret a NULL default public key method
1709 as unset and return the appropriate default but do *not* set the default.
1710 This means we can return the appropriate method in applications that
1711 switch between FIPS and non-FIPS modes.
1714 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1715 ENGINE is used then we cannot handle that in the FIPS module so we
1716 keep original code iff non-FIPS operations are allowed.
1719 *) Add -attime option to openssl utilities.
1720 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1722 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1725 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1726 FIPS EC methods unconditionally for now.
1729 *) New build option no-ec2m to disable characteristic 2 code.
1732 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1733 all cases can be covered as some introduce binary incompatibilities.
1736 *) Redirect RSA operations to FIPS module including keygen,
1737 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1740 *) Add similar low level API blocking to ciphers.
1743 *) Low level digest APIs are not approved in FIPS mode: any attempt
1744 to use these will cause a fatal error. Applications that *really* want
1745 to use them can use the private_* version instead.
1748 *) Redirect cipher operations to FIPS module for FIPS builds.
1751 *) Redirect digest operations to FIPS module for FIPS builds.
1754 *) Update build system to add "fips" flag which will link in fipscanister.o
1755 for static and shared library builds embedding a signature if needed.
1758 *) Output TLS supported curves in preference order instead of numerical
1759 order. This is currently hardcoded for the highest order curves first.
1760 This should be configurable so applications can judge speed vs strength.
1763 *) Add TLS v1.2 server support for client authentication.
1766 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1770 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1771 FIPS modules versions.
1774 *) Add TLS v1.2 client side support for client authentication. Keep cache
1775 of handshake records longer as we don't know the hash algorithm to use
1776 until after the certificate request message is received.
1779 *) Initial TLS v1.2 client support. Add a default signature algorithms
1780 extension including all the algorithms we support. Parse new signature
1781 format in client key exchange. Relax some ECC signing restrictions for
1782 TLS v1.2 as indicated in RFC5246.
1785 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1786 to new signature format when needed using client digest preference.
1787 All server ciphersuites should now work correctly in TLS v1.2. No client
1788 support yet and no support for client certificates.
1791 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1792 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1793 ciphersuites. At present only RSA key exchange ciphersuites work with
1794 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1795 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1796 and version checking.
1799 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1800 with this defined it will not be affected by any changes to ssl internal
1801 structures. Add several utility functions to allow openssl application
1802 to work with OPENSSL_NO_SSL_INTERN defined.
1806 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1808 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1811 *) Permit abbreviated handshakes when renegotiating using the function
1812 SSL_renegotiate_abbreviated().
1813 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1815 *) Add call to ENGINE_register_all_complete() to
1816 ENGINE_load_builtin_engines(), so some implementations get used
1817 automatically instead of needing explicit application support.
1820 *) Add support for TLS key exporter as described in RFC5705.
1821 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1823 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1824 a few changes are required:
1826 Add SSL_OP_NO_TLSv1_1 flag.
1827 Add TLSv1_1 methods.
1828 Update version checking logic to handle version 1.1.
1829 Add explicit IV handling (ported from DTLS code).
1830 Add command line options to s_client/s_server.
1833 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1835 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1836 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1837 content decryption and always return the same error. Note: this attack
1838 needs on average 2^20 messages so it only affects automated senders. The
1839 old behaviour can be reenabled in the CMS code by setting the
1840 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1841 an MMA defence is not necessary.
1842 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1843 this issue. (CVE-2012-0884)
1846 *) Fix CVE-2011-4619: make sure we really are receiving a
1847 client hello before rejecting multiple SGC restarts. Thanks to
1848 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1851 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1853 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1854 Thanks to Antonio Martin, Enterprise Secure Access Research and
1855 Development, Cisco Systems, Inc. for discovering this bug and
1856 preparing a fix. (CVE-2012-0050)
1859 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1861 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1862 of the Vaudenay padding oracle attack on CBC mode encryption
1863 which enables an efficient plaintext recovery attack against
1864 the OpenSSL implementation of DTLS. Their attack exploits timing
1865 differences arising during decryption processing. A research
1866 paper describing this attack can be found at:
1867 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1868 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1869 Security Group at Royal Holloway, University of London
1870 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1871 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1872 for preparing the fix. (CVE-2011-4108)
1873 [Robin Seggelmann, Michael Tuexen]
1875 *) Clear bytes used for block padding of SSL 3.0 records.
1877 [Adam Langley (Google)]
1879 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1880 Kadianakis <desnacked@gmail.com> for discovering this issue and
1881 Adam Langley for preparing the fix. (CVE-2011-4619)
1882 [Adam Langley (Google)]
1884 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1885 [Andrey Kulikov <amdeich@gmail.com>]
1887 *) Prevent malformed RFC3779 data triggering an assertion failure.
1888 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1889 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1890 [Rob Austein <sra@hactrn.net>]
1892 *) Improved PRNG seeding for VOS.
1893 [Paul Green <Paul.Green@stratus.com>]
1895 *) Fix ssl_ciph.c set-up race.
1896 [Adam Langley (Google)]
1898 *) Fix spurious failures in ecdsatest.c.
1899 [Emilia Käsper (Google)]
1901 *) Fix the BIO_f_buffer() implementation (which was mixing different
1902 interpretations of the '..._len' fields).
1903 [Adam Langley (Google)]
1905 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1906 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1907 threads won't reuse the same blinding coefficients.
1909 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1910 lock to call BN_BLINDING_invert_ex, and avoids one use of
1911 BN_BLINDING_update for each BN_BLINDING structure (previously,
1912 the last update always remained unused).
1913 [Emilia Käsper (Google)]
1915 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1916 [Bob Buckholz (Google)]
1918 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1920 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1921 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1922 [Kaspar Brand <ossl@velox.ch>]
1924 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1925 for multi-threaded use of ECDH. (CVE-2011-3210)
1926 [Adam Langley (Google)]
1928 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1931 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1932 signature public key algorithm by using OID xref utilities instead.
1933 Before this you could only use some ECC ciphersuites with SHA1 only.
1936 *) Add protection against ECDSA timing attacks as mentioned in the paper
1937 by Billy Bob Brumley and Nicola Tuveri, see:
1939 http://eprint.iacr.org/2011/232.pdf
1941 [Billy Bob Brumley and Nicola Tuveri]
1943 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1945 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1946 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1948 *) Fix bug in string printing code: if *any* escaping is enabled we must
1949 escape the escape character (backslash) or the resulting string is
1953 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1955 *) Disable code workaround for ancient and obsolete Netscape browsers
1956 and servers: an attacker can use it in a ciphersuite downgrade attack.
1957 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1960 *) Fixed J-PAKE implementation error, originally discovered by
1961 Sebastien Martini, further info and confirmation from Stefan
1962 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1965 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1967 *) Fix extension code to avoid race conditions which can result in a buffer
1968 overrun vulnerability: resumed sessions must not be modified as they can
1969 be shared by multiple threads. CVE-2010-3864
1972 *) Fix WIN32 build system to correctly link an ENGINE directory into
1976 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1978 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1980 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1982 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1984 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1985 context. The operation can be customised via the ctrl mechanism in
1986 case ENGINEs want to include additional functionality.
1989 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1992 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1993 output hashes compatible with older versions of OpenSSL.
1994 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1996 *) Fix compression algorithm handling: if resuming a session use the
1997 compression algorithm of the resumed session instead of determining
1998 it from client hello again. Don't allow server to change algorithm.
2001 *) Add load_crls() function to apps tidying load_certs() too. Add option
2002 to verify utility to allow additional CRLs to be included.
2005 *) Update OCSP request code to permit adding custom headers to the request:
2006 some responders need this.
2009 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2011 [Julia Lawall <julia@diku.dk>]
2013 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2014 needlessly dereferenced structures, used obsolete functions and
2015 didn't handle all updated verify codes correctly.
2018 *) Disable MD2 in the default configuration.
2021 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2022 indicate the initial BIO being pushed or popped. This makes it possible
2023 to determine whether the BIO is the one explicitly called or as a result
2024 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2025 it handles reference counts correctly and doesn't zero out the I/O bio
2026 when it is not being explicitly popped. WARNING: applications which
2027 included workarounds for the old buggy behaviour will need to be modified
2028 or they could free up already freed BIOs.
2031 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2032 renaming to all platforms (within the 0.9.8 branch, this was
2033 done conditionally on Netware platforms to avoid a name clash).
2034 [Guenter <lists@gknw.net>]
2036 *) Add ECDHE and PSK support to DTLS.
2037 [Michael Tuexen <tuexen@fh-muenster.de>]
2039 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2043 *) Add "missing" function EVP_MD_flags() (without this the only way to
2044 retrieve a digest flags is by accessing the structure directly. Update
2045 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2046 or cipher is registered as in the "from" argument. Print out all
2047 registered digests in the dgst usage message instead of manually
2048 attempting to work them out.
2051 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2052 this allows the use of compression and extensions. Change default cipher
2053 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2054 by default unless an application cipher string requests it.
2057 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2058 key ids to find matching certificates and keys but some PKCS#12 files
2059 don't follow the (somewhat unwritten) rules and this strategy fails.
2060 Now just gather all certificates together and the first private key
2061 then look for the first certificate that matches the key.
2064 *) Support use of registered digest and cipher names for dgst and cipher
2065 commands instead of having to add each one as a special case. So now
2072 openssl dgst -sha256 foo
2074 and this works for ENGINE based algorithms too.
2078 *) Update Gost ENGINE to support parameter files.
2079 [Victor B. Wagner <vitus@cryptocom.ru>]
2081 *) Support GeneralizedTime in ca utility.
2082 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2084 *) Enhance the hash format used for certificate directory links. The new
2085 form uses the canonical encoding (meaning equivalent names will work
2086 even if they aren't identical) and uses SHA1 instead of MD5. This form
2087 is incompatible with the older format and as a result c_rehash should
2088 be used to rebuild symbolic links.
2091 *) Make PKCS#8 the default write format for private keys, replacing the
2092 traditional format. This form is standardised, more secure and doesn't
2093 include an implicit MD5 dependency.
2096 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2097 committed to OpenSSL should pass this lot as a minimum.
2100 *) Add session ticket override functionality for use by EAP-FAST.
2101 [Jouni Malinen <j@w1.fi>]
2103 *) Modify HMAC functions to return a value. Since these can be implemented
2104 in an ENGINE errors can occur.
2107 *) Type-checked OBJ_bsearch_ex.
2110 *) Type-checked OBJ_bsearch. Also some constification necessitated
2111 by type-checking. Still to come: TXT_DB, bsearch(?),
2112 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2116 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2117 seconds to a tm structure directly, instead of going through OS
2118 specific date routines. This avoids any issues with OS routines such
2119 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2120 and X509_time_adj_ex() to cover the extended range. The existing
2121 X509_time_adj() is still usable and will no longer have any date issues.
2124 *) Delta CRL support. New use deltas option which will attempt to locate
2125 and search any appropriate delta CRLs available.
2127 This work was sponsored by Google.
2130 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2131 code and add additional score elements. Validate alternate CRL paths
2132 as part of the CRL checking and indicate a new error "CRL path validation
2133 error" in this case. Applications wanting additional details can use
2134 the verify callback and check the new "parent" field. If this is not
2135 NULL CRL path validation is taking place. Existing applications wont
2136 see this because it requires extended CRL support which is off by
2139 This work was sponsored by Google.
2142 *) Support for freshest CRL extension.
2144 This work was sponsored by Google.
2147 *) Initial indirect CRL support. Currently only supported in the CRLs
2148 passed directly and not via lookup. Process certificate issuer
2149 CRL entry extension and lookup CRL entries by bother issuer name
2150 and serial number. Check and process CRL issuer entry in IDP extension.
2152 This work was sponsored by Google.
2155 *) Add support for distinct certificate and CRL paths. The CRL issuer
2156 certificate is validated separately in this case. Only enabled if
2157 an extended CRL support flag is set: this flag will enable additional
2158 CRL functionality in future.
2160 This work was sponsored by Google.
2163 *) Add support for policy mappings extension.
2165 This work was sponsored by Google.
2168 *) Fixes to pathlength constraint, self issued certificate handling,
2169 policy processing to align with RFC3280 and PKITS tests.
2171 This work was sponsored by Google.
2174 *) Support for name constraints certificate extension. DN, email, DNS
2175 and URI types are currently supported.
2177 This work was sponsored by Google.
2180 *) To cater for systems that provide a pointer-based thread ID rather
2181 than numeric, deprecate the current numeric thread ID mechanism and
2182 replace it with a structure and associated callback type. This
2183 mechanism allows a numeric "hash" to be extracted from a thread ID in
2184 either case, and on platforms where pointers are larger than 'long',
2185 mixing is done to help ensure the numeric 'hash' is usable even if it
2186 can't be guaranteed unique. The default mechanism is to use "&errno"
2187 as a pointer-based thread ID to distinguish between threads.
2189 Applications that want to provide their own thread IDs should now use
2190 CRYPTO_THREADID_set_callback() to register a callback that will call
2191 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2193 Note that ERR_remove_state() is now deprecated, because it is tied
2194 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2195 to free the current thread's error state should be replaced by
2196 ERR_remove_thread_state(NULL).
2198 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2199 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2200 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2201 application was previously providing a numeric thread callback that
2202 was inappropriate for distinguishing threads, then uniqueness might
2203 have been obtained with &errno that happened immediately in the
2204 intermediate development versions of OpenSSL; this is no longer the
2205 case, the numeric thread callback will now override the automatic use
2207 [Geoff Thorpe, with help from Bodo Moeller]
2209 *) Initial support for different CRL issuing certificates. This covers a
2210 simple case where the self issued certificates in the chain exist and
2211 the real CRL issuer is higher in the existing chain.
2213 This work was sponsored by Google.
2216 *) Removed effectively defunct crypto/store from the build.
2219 *) Revamp of STACK to provide stronger type-checking. Still to come:
2220 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2221 ASN1_STRING, CONF_VALUE.
2224 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2225 RAM on SSL connections. This option can save about 34k per idle SSL.
2228 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2229 STACK, TXT_DB, bsearch, qsort.
2232 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2233 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2234 support for data, signedData, compressedData, digestedData and
2235 encryptedData, envelopedData types included. Scripts to check against
2236 RFC4134 examples draft and interop and consistency checks of many
2237 content types and variants.
2240 *) Add options to enc utility to support use of zlib compression BIO.
2243 *) Extend mk1mf to support importing of options and assembly language
2244 files from Configure script, currently only included in VC-WIN32.
2245 The assembly language rules can now optionally generate the source
2246 files from the associated perl scripts.
2249 *) Implement remaining functionality needed to support GOST ciphersuites.
2250 Interop testing has been performed using CryptoPro implementations.
2251 [Victor B. Wagner <vitus@cryptocom.ru>]
2253 *) s390x assembler pack.
2256 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2260 *) Implement Opaque PRF Input TLS extension as specified in
2261 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2262 official specification yet and no extension type assignment by
2263 IANA exists, this extension (for now) will have to be explicitly
2264 enabled when building OpenSSL by providing the extension number
2265 to use. For example, specify an option
2267 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2269 to the "config" or "Configure" script to enable the extension,
2270 assuming extension number 0x9527 (which is a completely arbitrary
2271 and unofficial assignment based on the MD5 hash of the Internet
2272 Draft). Note that by doing so, you potentially lose
2273 interoperability with other TLS implementations since these might
2274 be using the same extension number for other purposes.
2276 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2277 opaque PRF input value to use in the handshake. This will create
2278 an interal copy of the length-'len' string at 'src', and will
2279 return non-zero for success.
2281 To get more control and flexibility, provide a callback function
2284 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2285 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2289 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2292 Callback function 'cb' will be called in handshakes, and is
2293 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2294 Argument 'arg' is for application purposes (the value as given to
2295 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2296 be provided to the callback function). The callback function
2297 has to return non-zero to report success: usually 1 to use opaque
2298 PRF input just if possible, or 2 to enforce use of the opaque PRF
2299 input. In the latter case, the library will abort the handshake
2300 if opaque PRF input is not successfully negotiated.
2302 Arguments 'peerinput' and 'len' given to the callback function
2303 will always be NULL and 0 in the case of a client. A server will
2304 see the client's opaque PRF input through these variables if
2305 available (NULL and 0 otherwise). Note that if the server
2306 provides an opaque PRF input, the length must be the same as the
2307 length of the client's opaque PRF input.
2309 Note that the callback function will only be called when creating
2310 a new session (session resumption can resume whatever was
2311 previously negotiated), and will not be called in SSL 2.0
2312 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2313 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2314 for applications that need to enforce opaque PRF input.
2318 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2321 [Victor B. Wagner <vitus@cryptocom.ru>]
2323 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2324 RFC4507bis. The encrypted ticket format is an encrypted encoded
2325 SSL_SESSION structure, that way new session features are automatically
2328 If a client application caches session in an SSL_SESSION structure
2329 support is transparent because tickets are now stored in the encoded
2332 The SSL_CTX structure automatically generates keys for ticket
2333 protection in servers so again support should be possible
2334 with no application modification.
2336 If a client or server wishes to disable RFC4507 support then the option
2337 SSL_OP_NO_TICKET can be set.
2339 Add a TLS extension debugging callback to allow the contents of any client
2340 or server extensions to be examined.
2342 This work was sponsored by Google.
2345 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2346 OpenSSL should now compile cleanly on gcc 4.2
2347 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2349 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2350 support including streaming MAC support: this is required for GOST
2351 ciphersuite support.
2352 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2354 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2355 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2356 to output in BER and PEM format.
2359 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2360 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2361 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2362 ENGINE support for HMAC keys which are unextractable. New -mac and
2363 -macopt options to dgst utility.
2366 *) New option -sigopt to dgst utility. Update dgst to use
2367 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2368 alternative signing paramaters such as X9.31 or PSS in the dgst
2372 *) Change ssl_cipher_apply_rule(), the internal function that does
2373 the work each time a ciphersuite string requests enabling
2374 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2375 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2376 the order of disabled ciphersuites such that those ciphersuites
2377 that most recently went from enabled to disabled not only stay
2378 in order with respect to each other, but also have higher priority
2379 than other disabled ciphersuites the next time ciphersuites are
2382 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2383 the same ciphersuites as with "HIGH" alone, but in a specific
2384 order where the PSK ciphersuites come first (since they are the
2385 most recently disabled ciphersuites when "HIGH" is parsed).
2387 Also, change ssl_create_cipher_list() (using this new
2388 funcionality) such that between otherwise identical
2389 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2393 *) Change ssl_create_cipher_list() so that it automatically
2394 arranges the ciphersuites in reasonable order before starting
2395 to process the rule string. Thus, the definition for "DEFAULT"
2396 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2397 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2398 This makes it much easier to arrive at a reasonable default order
2399 in applications for which anonymous ciphers are OK (meaning
2400 that you can't actually use DEFAULT).
2401 [Bodo Moeller; suggested by Victor Duchovni]
2403 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2404 processing) into multiple integers instead of setting
2405 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2406 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2407 (These masks as well as the individual bit definitions are hidden
2408 away into the non-exported interface ssl/ssl_locl.h, so this
2409 change to the definition of the SSL_CIPHER structure shouldn't
2410 affect applications.) This give us more bits for each of these
2411 categories, so there is no longer a need to coagulate AES128 and
2412 AES256 into a single algorithm bit, and to coagulate Camellia128
2413 and Camellia256 into a single algorithm bit, which has led to all
2416 Thus, among other things, the kludge introduced in 0.9.7m and
2417 0.9.8e for masking out AES256 independently of AES128 or masking
2418 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2420 With the change, we also introduce new ciphersuite aliases that
2421 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2425 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2426 Use the leftmost N bytes of the signature input if the input is
2427 larger than the prime q (with N being the size in bytes of q).
2430 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2431 it yet and it is largely untested.
2434 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2437 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2438 some compilers (gcc 4.2 and later) reject their use. Safestack is
2439 reimplemented. Update ASN1 to avoid use of legacy functions.
2442 *) Win32/64 targets are linked with Winsock2.
2445 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2446 to external functions. This can be used to increase CRL handling
2447 efficiency especially when CRLs are very large by (for example) storing
2448 the CRL revoked certificates in a database.
2451 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2452 new CRLs added to a directory can be used. New command line option
2453 -verify_return_error to s_client and s_server. This causes real errors
2454 to be returned by the verify callback instead of carrying on no matter
2455 what. This reflects the way a "real world" verify callback would behave.
2458 *) GOST engine, supporting several GOST algorithms and public key formats.
2459 Kindly donated by Cryptocom.
2462 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2463 partitioned by DP are handled but no indirect CRL or reason partitioning
2464 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2465 selected via a scoring technique which handles IDP and AKID in CRLs.
2468 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2469 will ultimately be used for all verify operations: this will remove the
2470 X509_STORE dependency on certificate verification and allow alternative
2471 lookup methods. X509_STORE based implementations of these two callbacks.
2474 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2475 Modify get_crl() to find a valid (unexpired) CRL if possible.
2478 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2479 this would be called X509_CRL_cmp() but that name is already used by
2480 a function that just compares CRL issuer names. Cache several CRL
2481 extensions in X509_CRL structure and cache CRLDP in X509.
2484 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2485 this maps equivalent X509_NAME structures into a consistent structure.
2486 Name comparison can then be performed rapidly using memcmp().
2489 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2493 *) Allow digests to supply their own micalg string for S/MIME type using
2494 the ctrl EVP_MD_CTRL_MICALG.
2497 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2498 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2499 ctrl. It can then customise the structure before and/or after signing
2503 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2504 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2505 to free up any added signature OIDs.
2508 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2509 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2510 digest and cipher tables. New options added to openssl utility:
2511 list-message-digest-algorithms and list-cipher-algorithms.
2514 *) Change the array representation of binary polynomials: the list
2515 of degrees of non-zero coefficients is now terminated with -1.
2516 Previously it was terminated with 0, which was also part of the
2517 value; thus, the array representation was not applicable to
2518 polynomials where t^0 has coefficient zero. This change makes
2519 the array representation useful in a more general context.
2522 *) Various modifications and fixes to SSL/TLS cipher string
2523 handling. For ECC, the code now distinguishes between fixed ECDH
2524 with RSA certificates on the one hand and with ECDSA certificates
2525 on the other hand, since these are separate ciphersuites. The
2526 unused code for Fortezza ciphersuites has been removed.
2528 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2529 (not "ECDHE"). For consistency with the code for DH
2530 certificates, use of ECDH certificates is now considered ECDH
2531 authentication, not RSA or ECDSA authentication (the latter is
2532 merely the CA's signing algorithm and not actively used in the
2535 The temporary ciphersuite alias "ECCdraft" is no longer
2536 available, and ECC ciphersuites are no longer excluded from "ALL"
2537 and "DEFAULT". The following aliases now exist for RFC 4492
2538 ciphersuites, most of these by analogy with the DH case:
2540 kECDHr - ECDH cert, signed with RSA
2541 kECDHe - ECDH cert, signed with ECDSA
2542 kECDH - ECDH cert (signed with either RSA or ECDSA)
2543 kEECDH - ephemeral ECDH
2544 ECDH - ECDH cert or ephemeral ECDH
2550 AECDH - anonymous ECDH
2551 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2555 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2556 Use correct micalg parameters depending on digest(s) in signed message.
2559 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2560 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2563 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2564 an engine to register a method. Add ENGINE lookups for methods and
2565 functional reference processing.
2568 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2569 EVP_{Sign,Verify}* which allow an application to customise the signature
2573 *) New -resign option to smime utility. This adds one or more signers
2574 to an existing PKCS#7 signedData structure. Also -md option to use an
2575 alternative message digest algorithm for signing.
2578 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2579 create PKCS7 structures containing multiple signers. Update smime
2580 application to support multiple signers.
2583 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2587 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2588 Reorganize PBE internals to lookup from a static table using NIDs,
2589 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2590 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2591 PRF which will be automatically used with PBES2.
2594 *) Replace the algorithm specific calls to generate keys in "req" with the
2598 *) Update PKCS#7 enveloped data routines to use new API. This is now
2599 supported by any public key method supporting the encrypt operation. A
2600 ctrl is added to allow the public key algorithm to examine or modify
2601 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2605 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2606 a default digest type to use. In most cases this will be SHA1 but some
2607 algorithms (such as GOST) need to specify an alternative digest. The
2608 return value indicates how strong the preference is 1 means optional and
2609 2 is mandatory (that is it is the only supported type). Modify
2610 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2611 use the default md. Update openssl utilities to use the default digest
2612 type for signing if it is not explicitly indicated.
2615 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2616 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2617 signing method from the key type. This effectively removes the link
2618 between digests and public key types.
2621 *) Add an OID cross reference table and utility functions. Its purpose is to
2622 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2623 rsaEncryption. This will allow some of the algorithm specific hackery
2624 needed to use the correct OID to be removed.
2627 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2628 structures for PKCS7_sign(). They are now set up by the relevant public
2632 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2635 *) Add support for key derivation (agreement) in the API, DH method and
2639 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2640 public and private key formats. As a side effect these add additional
2641 command line functionality not previously available: DSA signatures can be
2642 generated and verified using pkeyutl and DH key support and generation in
2647 [Oliver Tappe <zooey@hirschkaefer.de>]
2649 *) New make target "install_html_docs" installs HTML renditions of the
2651 [Oliver Tappe <zooey@hirschkaefer.de>]
2653 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2654 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2655 support key and parameter generation and add initial key generation
2656 functionality for RSA.
2659 *) Add functions for main EVP_PKEY_method operations. The undocumented
2660 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2661 EVP_PKEY_{encrypt,decrypt}_old.
2664 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2665 key API, doesn't do much yet.
2668 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2669 public key algorithms. New option to openssl utility:
2670 "list-public-key-algorithms" to print out info.
2673 *) Implement the Supported Elliptic Curves Extension for
2674 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2677 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2678 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2681 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2682 utilities such as rsa, dsa, dsaparam etc except they process any key
2686 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2687 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2688 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2692 *) Initial support for pluggable public key ASN1.
2693 De-spaghettify the public key ASN1 handling. Move public and private
2694 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2695 algorithm specific handling to a single module within the relevant
2696 algorithm directory. Add functions to allow (near) opaque processing
2697 of public and private key structures.
2700 *) Implement the Supported Point Formats Extension for
2701 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2704 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2705 for the psk identity [hint] and the psk callback functions to the
2706 SSL_SESSION, SSL and SSL_CTX structure.
2709 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2713 SSL_CTX_use_psk_identity_hint
2714 SSL_get_psk_identity_hint
2715 SSL_get_psk_identity
2716 SSL_use_psk_identity_hint
2718 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2720 *) Add RFC 3161 compliant time stamp request creation, response generation
2721 and response verification functionality.
2722 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2724 *) Add initial support for TLS extensions, specifically for the server_name
2725 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2726 have new members for a host name. The SSL data structure has an
2727 additional member SSL_CTX *initial_ctx so that new sessions can be
2728 stored in that context to allow for session resumption, even after the
2729 SSL has been switched to a new SSL_CTX in reaction to a client's
2730 server_name extension.
2732 New functions (subject to change):
2734 SSL_get_servername()
2735 SSL_get_servername_type()
2738 New CTRL codes and macros (subject to change):
2740 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2741 - SSL_CTX_set_tlsext_servername_callback()
2742 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2743 - SSL_CTX_set_tlsext_servername_arg()
2744 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2746 openssl s_client has a new '-servername ...' option.
2748 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2749 '-key2 ...', '-servername_fatal' (subject to change). This allows
2750 testing the HostName extension for a specific single host name ('-cert'
2751 and '-key' remain fallbacks for handshakes without HostName
2752 negotiation). If the unrecognized_name alert has to be sent, this by
2753 default is a warning; it becomes fatal with the '-servername_fatal'
2756 [Peter Sylvester, Remy Allais, Christophe Renou]
2758 *) Whirlpool hash implementation is added.
2761 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2762 bn(64,32). Because of instruction set limitations it doesn't have
2763 any negative impact on performance. This was done mostly in order
2764 to make it possible to share assembler modules, such as bn_mul_mont
2765 implementations, between 32- and 64-bit builds without hassle.
2768 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2769 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2773 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2774 dedicated Montgomery multiplication procedure, is introduced.
2775 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2776 "64-bit" performance on certain 32-bit targets.
2779 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2780 in SSL structures. New SSL ctrl to set maximum send fragment size.
2781 Save memory by seeting the I/O buffer sizes dynamically instead of
2782 using the maximum available value.
2785 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2786 in addition to the text details.
2789 *) Very, very preliminary EXPERIMENTAL support for printing of general
2790 ASN1 structures. This currently produces rather ugly output and doesn't
2791 handle several customised structures at all.
2794 *) Integrated support for PVK file format and some related formats such
2795 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2796 these in the 'rsa' and 'dsa' utilities.
2799 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2802 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2803 place for the (very old) "NETSCAPE" format certificates which are now
2804 handled using new ASN1 code equivalents.
2807 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2808 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2809 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2812 *) Modify CRL distribution points extension code to print out previously
2813 unsupported fields. Enhance extension setting code to allow setting of
2817 *) Add print and set support for Issuing Distribution Point CRL extension.
2820 *) Change 'Configure' script to enable Camellia by default.
2823 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2825 *) When rejecting SSL/TLS records due to an incorrect version number, never
2826 update s->server with a new major version number. As of
2827 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2828 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2829 the previous behavior could result in a read attempt at NULL when
2830 receiving specific incorrect SSL/TLS records once record payload
2831 protection is active. (CVE-2010-0740)
2832 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2834 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2835 could be crashed if the relevant tables were not present (e.g. chrooted).
2836 [Tomas Hoger <thoger@redhat.com>]
2838 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2840 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2841 [Martin Olsson, Neel Mehta]
2843 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2844 accommodate for stack sorting, always a write lock!).
2847 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2848 excessive delays in the RAND_poll(): over a minute. As a workaround
2849 include a time check in the inner Heap32Next loop too.
2852 *) The code that handled flushing of data in SSL/TLS originally used the
2853 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2854 the problem outlined in PR#1949. The fix suggested there however can
2855 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2856 of Apache). So instead simplify the code to flush unconditionally.
2857 This should be fine since flushing with no data to flush is a no op.
2860 *) Handle TLS versions 2.0 and later properly and correctly use the
2861 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2862 off ancient servers have a habit of sticking around for a while...
2865 *) Modify compression code so it frees up structures without using the
2866 ex_data callbacks. This works around a problem where some applications
2867 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2868 restarting) then use compression (e.g. SSL with compression) later.
2869 This results in significant per-connection memory leaks and
2870 has caused some security issues including CVE-2008-1678 and
2874 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2875 change when encrypting or decrypting.
2878 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2879 connect and renegotiate with servers which do not support RI.
2880 Until RI is more widely deployed this option is enabled by default.
2883 *) Add "missing" ssl ctrls to clear options and mode.
2886 *) If client attempts to renegotiate and doesn't support RI respond with
2887 a no_renegotiation alert as required by RFC5746. Some renegotiating
2888 TLS clients will continue a connection gracefully when they receive
2889 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2890 waiting for a server hello which it will never receive. Now we treat a
2891 received no_renegotiation alert as a fatal error. This is because
2892 applications requesting a renegotiation might well expect it to succeed
2893 and would have no code in place to handle the server denying it so the
2894 only safe thing to do is to terminate the connection.
2897 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2898 peer supports secure renegotiation and 0 otherwise. Print out peer
2899 renegotiation support in s_client/s_server.
2902 *) Replace the highly broken and deprecated SPKAC certification method with
2903 the updated NID creation version. This should correctly handle UTF8.
2906 *) Implement RFC5746. Re-enable renegotiation but require the extension
2907 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2908 turns out to be a bad idea. It has been replaced by
2909 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2910 SSL_CTX_set_options(). This is really not recommended unless you
2911 know what you are doing.
2912 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2914 *) Fixes to stateless session resumption handling. Use initial_ctx when
2915 issuing and attempting to decrypt tickets in case it has changed during
2916 servername handling. Use a non-zero length session ID when attempting
2917 stateless session resumption: this makes it possible to determine if
2918 a resumption has occurred immediately after receiving server hello
2919 (several places in OpenSSL subtly assume this) instead of later in
2923 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2924 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2925 fixes for a few places where the return code is not checked
2927 [Julia Lawall <julia@diku.dk>]
2929 *) Add --strict-warnings option to Configure script to include devteam
2930 warnings in other configurations.
2933 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2934 makes it possible to install openssl libraries in locations which
2935 have names other than "lib", for example "/usr/lib64" which some
2937 [Steve Henson, based on patch from Jeremy Utley]
2939 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2940 X690 8.9.12 and can produce some misleading textual output of OIDs.
2941 [Steve Henson, reported by Dan Kaminsky]
2943 *) Delete MD2 from algorithm tables. This follows the recommendation in
2944 several standards that it is not used in new applications due to
2945 several cryptographic weaknesses. For binary compatibility reasons
2946 the MD2 API is still compiled in by default.
2949 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2953 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2954 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2956 [Guenter <lists@gknw.net>]
2958 *) Fix the server certificate chain building code to use X509_verify_cert(),
2959 it used to have an ad-hoc builder which was unable to cope with anything
2960 other than a simple chain.
2961 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2963 *) Don't check self signed certificate signatures in X509_verify_cert()
2964 by default (a flag can override this): it just wastes time without
2965 adding any security. As a useful side effect self signed root CAs
2966 with non-FIPS digests are now usable in FIPS mode.
2969 *) In dtls1_process_out_of_seq_message() the check if the current message
2970 is already buffered was missing. For every new message was memory
2971 allocated, allowing an attacker to perform an denial of service attack
2972 with sending out of seq handshake messages until there is no memory
2973 left. Additionally every future messege was buffered, even if the
2974 sequence number made no sense and would be part of another handshake.
2975 So only messages with sequence numbers less than 10 in advance will be
2976 buffered. (CVE-2009-1378)
2977 [Robin Seggelmann, discovered by Daniel Mentz]
2979 *) Records are buffered if they arrive with a future epoch to be
2980 processed after finishing the corresponding handshake. There is
2981 currently no limitation to this buffer allowing an attacker to perform
2982 a DOS attack with sending records with future epochs until there is no
2983 memory left. This patch adds the pqueue_size() function to determine
2984 the size of a buffer and limits the record buffer to 100 entries.
2986 [Robin Seggelmann, discovered by Daniel Mentz]
2988 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2989 parent structure is freed. (CVE-2009-1379)
2992 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2993 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2996 [Ilya O. <vrghost@gmail.com>]
2998 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3000 *) Disable renegotiation completely - this fixes a severe security
3001 problem (CVE-2009-3555) at the cost of breaking all
3002 renegotiation. Renegotiation can be re-enabled by setting
3003 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3004 run-time. This is really not recommended unless you know what
3008 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3010 *) Don't set val to NULL when freeing up structures, it is freed up by
3011 underlying code. If sizeof(void *) > sizeof(long) this can result in
3012 zeroing past the valid field. (CVE-2009-0789)
3013 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3015 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3016 checked correctly. This would allow some invalid signed attributes to
3017 appear to verify correctly. (CVE-2009-0591)
3018 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3020 *) Reject UniversalString and BMPString types with invalid lengths. This
3021 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3022 a legal length. (CVE-2009-0590)
3025 *) Set S/MIME signing as the default purpose rather than setting it
3026 unconditionally. This allows applications to override it at the store
3030 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3031 to handle some structures.
3034 *) Improve efficiency of mem_gets: don't search whole buffer each time
3036 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3038 *) New -hex option for openssl rand.
3041 *) Print out UTF8String and NumericString when parsing ASN1.
3044 *) Support NumericString type for name components.
3047 *) Allow CC in the environment to override the automatically chosen
3048 compiler. Note that nothing is done to ensure flags work with the
3052 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3054 *) Properly check EVP_VerifyFinal() and similar return values
3056 [Ben Laurie, Bodo Moeller, Google Security Team]
3058 *) Enable TLS extensions by default.
3061 *) Allow the CHIL engine to be loaded, whether the application is
3062 multithreaded or not. (This does not release the developer from the
3063 obligation to set up the dynamic locking callbacks.)
3064 [Sander Temme <sander@temme.net>]
3066 *) Use correct exit code if there is an error in dgst command.
3067 [Steve Henson; problem pointed out by Roland Dirlewanger]
3069 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3070 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3073 *) Add experimental JPAKE support, including demo authentication in
3074 s_client and s_server.
3077 *) Set the comparison function in v3_addr_canonize().
3078 [Rob Austein <sra@hactrn.net>]
3080 *) Add support for XMPP STARTTLS in s_client.
3081 [Philip Paeps <philip@freebsd.org>]
3083 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
3084 to ensure that even with this option, only ciphersuites in the
3085 server's preference list will be accepted. (Note that the option
3086 applies only when resuming a session, so the earlier behavior was
3087 just about the algorithm choice for symmetric cryptography.)
3090 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
3092 *) Fix NULL pointer dereference if a DTLS server received
3093 ChangeCipherSpec as first record (CVE-2009-1386).
3096 *) Fix a state transition in s3_srvr.c and d1_srvr.c
3097 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
3100 *) The fix in 0.9.8c that supposedly got rid of unsafe
3101 double-checked locking was incomplete for RSA blinding,
3102 addressing just one layer of what turns out to have been
3103 doubly unsafe triple-checked locking.
3105 So now fix this for real by retiring the MONT_HELPER macro
3106 in crypto/rsa/rsa_eay.c.
3108 [Bodo Moeller; problem pointed out by Marius Schilder]
3110 *) Various precautionary measures:
3112 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
3114 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
3115 (NB: This would require knowledge of the secret session ticket key
3116 to exploit, in which case you'd be SOL either way.)
3118 - Change bn_nist.c so that it will properly handle input BIGNUMs
3119 outside the expected range.
3121 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
3124 [Neel Mehta, Bodo Moeller]
3126 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
3127 the load fails. Useful for distros.
3128 [Ben Laurie and the FreeBSD team]
3130 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3133 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3136 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3138 This work was sponsored by Logica.
3141 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3142 keystores. Support for SSL/TLS client authentication too.
3143 Not compiled unless enable-capieng specified to Configure.
3145 This work was sponsored by Logica.
3148 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3149 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3150 attribute creation routines such as certifcate requests and PKCS#12
3154 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3156 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3157 handshake which could lead to a cilent crash as found using the
3158 Codenomicon TLS test suite (CVE-2008-1672)
3159 [Steve Henson, Mark Cox]
3161 *) Fix double free in TLS server name extensions which could lead to
3162 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3165 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3167 Clear the error queue to ensure that error entries left from
3168 older function calls do not interfere with the correct operation.
3169 [Lutz Jaenicke, Erik de Castro Lopo]
3171 *) Remove root CA certificates of commercial CAs:
3173 The OpenSSL project does not recommend any specific CA and does not
3174 have any policy with respect to including or excluding any CA.
3175 Therefore it does not make any sense to ship an arbitrary selection
3176 of root CA certificates with the OpenSSL software.
3179 *) RSA OAEP patches to fix two separate invalid memory reads.
3180 The first one involves inputs when 'lzero' is greater than
3181 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3182 before the beginning of from). The second one involves inputs where
3183 the 'db' section contains nothing but zeroes (there is a one-byte
3184 invalid read after the end of 'db').
3185 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3187 *) Partial backport from 0.9.9-dev:
3189 Introduce bn_mul_mont (dedicated Montgomery multiplication
3190 procedure) as a candidate for BIGNUM assembler implementation.
3191 While 0.9.9-dev uses assembler for various architectures, only
3192 x86_64 is available by default here in the 0.9.8 branch, and
3193 32-bit x86 is available through a compile-time setting.
3195 To try the 32-bit x86 assembler implementation, use Configure
3196 option "enable-montasm" (which exists only for this backport).
3198 As "enable-montasm" for 32-bit x86 disclaims code stability
3199 anyway, in this constellation we activate additional code
3200 backported from 0.9.9-dev for further performance improvements,
3201 namely BN_from_montgomery_word. (To enable this otherwise,
3202 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3204 [Andy Polyakov (backport partially by Bodo Moeller)]
3206 *) Add TLS session ticket callback. This allows an application to set
3207 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3208 values. This is useful for key rollover for example where several key
3209 sets may exist with different names.
3212 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3213 This was broken until now in 0.9.8 releases, such that the only way
3214 a registered ENGINE could be used (assuming it initialises
3215 successfully on the host) was to explicitly set it as the default
3216 for the relevant algorithms. This is in contradiction with 0.9.7
3217 behaviour and the documentation. With this fix, when an ENGINE is
3218 registered into a given algorithm's table of implementations, the
3219 'uptodate' flag is reset so that auto-discovery will be used next
3220 time a new context for that algorithm attempts to select an
3222 [Ian Lister (tweaked by Geoff Thorpe)]
3224 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3225 implemention in the following ways:
3227 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3230 Lack of BER streaming support means one pass streaming processing is
3231 only supported if data is detached: setting the streaming flag is
3232 ignored for embedded content.
3234 CMS support is disabled by default and must be explicitly enabled
3235 with the enable-cms configuration option.
3238 *) Update the GMP engine glue to do direct copies between BIGNUM and
3239 mpz_t when openssl and GMP use the same limb size. Otherwise the
3240 existing "conversion via a text string export" trick is still used.
3241 [Paul Sheer <paulsheer@gmail.com>]
3243 *) Zlib compression BIO. This is a filter BIO which compressed and
3244 uncompresses any data passed through it.
3247 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3248 RFC3394 compatible AES key wrapping.
3251 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3252 sets string data without copying. X509_ALGOR_set0() and
3253 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3254 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3255 from an X509_ATTRIBUTE structure optionally checking it occurs only
3256 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3260 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3261 to get the expected BN_FLG_CONSTTIME behavior.
3262 [Bodo Moeller (Google)]
3266 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3267 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3268 - added some more tests to do_tests.pl
3269 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3270 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3271 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3272 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3273 - various changes to netware.pl to enable gcc-cross builds on Win32
3275 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3276 - various changes to fix missing prototype warnings
3277 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3278 - added AES, WHIRLPOOL and CPUID assembler code to build files
3279 - added missing AES assembler make rules to mk1mf.pl
3280 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3281 [Guenter Knauf <eflash@gmx.net>]
3283 *) Implement certificate status request TLS extension defined in RFC3546.
3284 A client can set the appropriate parameters and receive the encoded
3285 OCSP response via a callback. A server can query the supplied parameters
3286 and set the encoded OCSP response in the callback. Add simplified examples
3287 to s_client and s_server.
3290 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3292 *) Fix various bugs:
3293 + Binary incompatibility of ssl_ctx_st structure
3294 + DTLS interoperation with non-compliant servers
3295 + Don't call get_session_cb() without proposed session
3296 + Fix ia64 assembler code
3297 [Andy Polyakov, Steve Henson]
3299 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3301 *) DTLS Handshake overhaul. There were longstanding issues with
3302 OpenSSL DTLS implementation, which were making it impossible for
3303 RFC 4347 compliant client to communicate with OpenSSL server.
3304 Unfortunately just fixing these incompatibilities would "cut off"
3305 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3306 server keeps tolerating non RFC compliant syntax. The opposite is
3307 not true, 0.9.8f client can not communicate with earlier server.
3308 This update even addresses CVE-2007-4995.
3311 *) Changes to avoid need for function casts in OpenSSL: some compilers
3312 (gcc 4.2 and later) reject their use.
3313 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3316 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3317 RFC4507bis. The encrypted ticket format is an encrypted encoded
3318 SSL_SESSION structure, that way new session features are automatically
3321 If a client application caches session in an SSL_SESSION structure
3322 support is transparent because tickets are now stored in the encoded
3325 The SSL_CTX structure automatically generates keys for ticket
3326 protection in servers so again support should be possible
3327 with no application modification.
3329 If a client or server wishes to disable RFC4507 support then the option
3330 SSL_OP_NO_TICKET can be set.
3332 Add a TLS extension debugging callback to allow the contents of any client
3333 or server extensions to be examined.
3335 This work was sponsored by Google.
3338 *) Add initial support for TLS extensions, specifically for the server_name
3339 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3340 have new members for a host name. The SSL data structure has an
3341 additional member SSL_CTX *initial_ctx so that new sessions can be
3342 stored in that context to allow for session resumption, even after the
3343 SSL has been switched to a new SSL_CTX in reaction to a client's
3344 server_name extension.
3346 New functions (subject to change):
3348 SSL_get_servername()
3349 SSL_get_servername_type()
3352 New CTRL codes and macros (subject to change):
3354 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3355 - SSL_CTX_set_tlsext_servername_callback()
3356 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3357 - SSL_CTX_set_tlsext_servername_arg()
3358 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3360 openssl s_client has a new '-servername ...' option.
3362 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3363 '-key2 ...', '-servername_fatal' (subject to change). This allows
3364 testing the HostName extension for a specific single host name ('-cert'
3365 and '-key' remain fallbacks for handshakes without HostName
3366 negotiation). If the unrecognized_name alert has to be sent, this by
3367 default is a warning; it becomes fatal with the '-servername_fatal'
3370 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3372 *) Add AES and SSE2 assembly language support to VC++ build.
3375 *) Mitigate attack on final subtraction in Montgomery reduction.
3378 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3379 (which previously caused an internal error).
3382 *) Squeeze another 10% out of IGE mode when in != out.
3385 *) AES IGE mode speedup.
3386 [Dean Gaudet (Google)]
3388 *) Add the Korean symmetric 128-bit cipher SEED (see
3389 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3390 add SEED ciphersuites from RFC 4162:
3392 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3393 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3394 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3395 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3397 To minimize changes between patchlevels in the OpenSSL 0.9.8
3398 series, SEED remains excluded from compilation unless OpenSSL
3399 is configured with 'enable-seed'.
3400 [KISA, Bodo Moeller]
3402 *) Mitigate branch prediction attacks, which can be practical if a
3403 single processor is shared, allowing a spy process to extract
3404 information. For detailed background information, see
3405 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3406 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3407 and Necessary Software Countermeasures"). The core of the change
3408 are new versions BN_div_no_branch() and
3409 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3410 respectively, which are slower, but avoid the security-relevant
3411 conditional branches. These are automatically called by BN_div()
3412 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3413 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3414 remove a conditional branch.
3416 BN_FLG_CONSTTIME is the new name for the previous
3417 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3418 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3419 in the exponent causes BN_mod_exp_mont() to use the alternative
3420 implementation in BN_mod_exp_mont_consttime().) The old name
3421 remains as a deprecated alias.
3423 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3424 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3425 constant-time implementations for more than just exponentiation.
3426 Here too the old name is kept as a deprecated alias.
3428 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3429 the BN_BLINDING structure gets an independent copy of the
3430 modulus. This means that the previous "BIGNUM *m" argument to
3431 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3432 essentially becomes "const BIGNUM *m", although we can't actually
3433 change this in the header file before 0.9.9. It allows
3434 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3435 enable BN_FLG_CONSTTIME.
3437 [Matthew D Wood (Intel Corp)]
3439 *) In the SSL/TLS server implementation, be strict about session ID
3440 context matching (which matters if an application uses a single
3441 external cache for different purposes). Previously,
3442 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3443 set. This did ensure strict client verification, but meant that,
3444 with applications using a single external cache for quite
3445 different requirements, clients could circumvent ciphersuite
3446 restrictions for a given session ID context by starting a session
3447 in a different context.
3450 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3451 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3452 authentication-only ciphersuites.
3455 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3456 not complete and could lead to a possible single byte overflow
3457 (CVE-2007-5135) [Ben Laurie]
3459 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3461 *) Since AES128 and AES256 (and similarly Camellia128 and
3462 Camellia256) share a single mask bit in the logic of
3463 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3464 kludge to work properly if AES128 is available and AES256 isn't
3465 (or if Camellia128 is available and Camellia256 isn't).
3468 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3469 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3470 When a point or a seed is encoded in a BIT STRING, we need to
3471 prevent the removal of trailing zero bits to get the proper DER
3472 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3473 of a NamedBitList, for which trailing 0 bits need to be removed.)
3476 *) Have SSL/TLS server implementation tolerate "mismatched" record
3477 protocol version while receiving ClientHello even if the
3478 ClientHello is fragmented. (The server can't insist on the
3479 particular protocol version it has chosen before the ServerHello
3480 message has informed the client about his choice.)
3483 *) Add RFC 3779 support.
3484 [Rob Austein for ARIN, Ben Laurie]
3486 *) Load error codes if they are not already present instead of using a
3487 static variable. This allows them to be cleanly unloaded and reloaded.
3488 Improve header file function name parsing.
3491 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3492 or CAPABILITY handshake as required by RFCs.
3495 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3497 *) Introduce limits to prevent malicious keys being able to
3498 cause a denial of service. (CVE-2006-2940)
3499 [Steve Henson, Bodo Moeller]
3501 *) Fix ASN.1 parsing of certain invalid structures that can result
3502 in a denial of service. (CVE-2006-2937) [Steve Henson]
3504 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3505 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3507 *) Fix SSL client code which could crash if connecting to a
3508 malicious SSLv2 server. (CVE-2006-4343)
3509 [Tavis Ormandy and Will Drewry, Google Security Team]
3511 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3512 match only those. Before that, "AES256-SHA" would be interpreted
3513 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3514 the same strength classification in 0.9.7h) as we currently only
3515 have a single AES bit in the ciphersuite description bitmap.
3516 That change, however, also applied to ciphersuite strings such as
3517 "RC4-MD5" that intentionally matched multiple ciphersuites --
3518 namely, SSL 2.0 ciphersuites in addition to the more common ones
3519 from SSL 3.0/TLS 1.0.
3521 So we change the selection algorithm again: Naming an explicit
3522 ciphersuite selects this one ciphersuite, and any other similar
3523 ciphersuite (same bitmap) from *other* protocol versions.
3524 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3525 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3527 Since SSL 2.0 does not have any ciphersuites for which the
3528 128/256 bit distinction would be relevant, this works for now.
3529 The proper fix will be to use different bits for AES128 and
3530 AES256, which would have avoided the problems from the beginning;
3531 however, bits are scarce, so we can only do this in a new release
3532 (not just a patchlevel) when we can change the SSL_CIPHER
3533 definition to split the single 'unsigned long mask' bitmap into
3534 multiple values to extend the available space.
3538 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3540 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3541 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3543 *) Add AES IGE and biIGE modes.
3546 *) Change the Unix randomness entropy gathering to use poll() when
3547 possible instead of select(), since the latter has some
3548 undesirable limitations.
3549 [Darryl Miles via Richard Levitte and Bodo Moeller]
3551 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3552 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3553 cannot be implicitly activated as part of, e.g., the "AES" alias.
3554 However, please upgrade to OpenSSL 0.9.9[-dev] for
3555 non-experimental use of the ECC ciphersuites to get TLS extension
3556 support, which is required for curve and point format negotiation
3557 to avoid potential handshake problems.
3560 *) Disable rogue ciphersuites:
3562 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3563 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3564 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3566 The latter two were purportedly from
3567 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3570 Also deactivate the remaining ciphersuites from
3571 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3572 unofficial, and the ID has long expired.
3575 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3576 dual-core machines) and other potential thread-safety issues.
3579 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3580 versions), which is now available for royalty-free use
3581 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3582 Also, add Camellia TLS ciphersuites from RFC 4132.
3584 To minimize changes between patchlevels in the OpenSSL 0.9.8
3585 series, Camellia remains excluded from compilation unless OpenSSL
3586 is configured with 'enable-camellia'.
3589 *) Disable the padding bug check when compression is in use. The padding
3590 bug check assumes the first packet is of even length, this is not
3591 necessarily true if compresssion is enabled and can result in false
3592 positives causing handshake failure. The actual bug test is ancient
3593 code so it is hoped that implementations will either have fixed it by
3594 now or any which still have the bug do not support compression.
3597 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3599 *) When applying a cipher rule check to see if string match is an explicit
3600 cipher suite and only match that one cipher suite if it is.
3603 *) Link in manifests for VC++ if needed.
3604 [Austin Ziegler <halostatue@gmail.com>]
3606 *) Update support for ECC-based TLS ciphersuites according to
3607 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3608 TLS extensions, which are supported starting with the 0.9.9
3609 branch, not in the OpenSSL 0.9.8 branch).
3612 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3613 opaque EVP_CIPHER_CTX handling.
3616 *) Fixes and enhancements to zlib compression code. We now only use
3617 "zlib1.dll" and use the default __cdecl calling convention on Win32
3618 to conform with the standards mentioned here:
3619 http://www.zlib.net/DLL_FAQ.txt
3620 Static zlib linking now works on Windows and the new --with-zlib-include
3621 --with-zlib-lib options to Configure can be used to supply the location
3622 of the headers and library. Gracefully handle case where zlib library
3626 *) Several fixes and enhancements to the OID generation code. The old code
3627 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3628 handle numbers larger than ULONG_MAX, truncated printing and had a
3629 non standard OBJ_obj2txt() behaviour.
3632 *) Add support for building of engines under engine/ as shared libraries
3633 under VC++ build system.
3636 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3637 Hopefully, we will not see any false combination of paths any more.
3640 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3642 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3643 (part of SSL_OP_ALL). This option used to disable the
3644 countermeasure against man-in-the-middle protocol-version
3645 rollback in the SSL 2.0 server implementation, which is a bad
3646 idea. (CVE-2005-2969)
3648 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3649 for Information Security, National Institute of Advanced Industrial
3650 Science and Technology [AIST], Japan)]
3652 *) Add two function to clear and return the verify parameter flags.
3655 *) Keep cipherlists sorted in the source instead of sorting them at
3656 runtime, thus removing the need for a lock.
3659 *) Avoid some small subgroup attacks in Diffie-Hellman.
3660 [Nick Mathewson and Ben Laurie]
3662 *) Add functions for well-known primes.
3665 *) Extended Windows CE support.
3666 [Satoshi Nakamura and Andy Polyakov]
3668 *) Initialize SSL_METHOD structures at compile time instead of during
3669 runtime, thus removing the need for a lock.
3672 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3673 attempting to decrypt each encrypted key in turn. Add support to
3677 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3679 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3682 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3685 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3686 key into the same file any more.
3689 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3692 *) Add -utf8 command line and config file option to 'ca'.
3693 [Stefan <stf@udoma.org]
3695 *) Removed the macro des_crypt(), as it seems to conflict with some
3696 libraries. Use DES_crypt().
3699 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3700 involves renaming the source and generated shared-libs for
3701 both. The engines will accept the corrected or legacy ids
3702 ('ncipher' and '4758_cca' respectively) when binding. NB,
3703 this only applies when building 'shared'.
3704 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3706 *) Add attribute functions to EVP_PKEY structure. Modify
3707 PKCS12_create() to recognize a CSP name attribute and
3708 use it. Make -CSP option work again in pkcs12 utility.
3711 *) Add new functionality to the bn blinding code:
3712 - automatic re-creation of the BN_BLINDING parameters after
3713 a fixed number of uses (currently 32)
3714 - add new function for parameter creation
3715 - introduce flags to control the update behaviour of the
3716 BN_BLINDING parameters
3717 - hide BN_BLINDING structure
3718 Add a second BN_BLINDING slot to the RSA structure to improve
3719 performance when a single RSA object is shared among several
3723 *) Add support for DTLS.
3724 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3726 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3727 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3730 *) Remove buggy and incomplete DH cert support from
3731 ssl/ssl_rsa.c and ssl/s3_both.c
3734 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3735 the apps/openssl applications.
3738 *) Compile clean with "-Wall -Wmissing-prototypes
3739 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3740 DEBUG_SAFESTACK must also be set.
3743 *) Change ./Configure so that certain algorithms can be disabled by default.
3744 The new counterpiece to "no-xxx" is "enable-xxx".
3746 The patented RC5 and MDC2 algorithms will now be disabled unless
3747 "enable-rc5" and "enable-mdc2", respectively, are specified.
3749 (IDEA remains enabled despite being patented. This is because IDEA
3750 is frequently required for interoperability, and there is no license
3751 fee for non-commercial use. As before, "no-idea" can be used to
3752 avoid this algorithm.)
3756 *) Add processing of proxy certificates (see RFC 3820). This work was
3757 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3758 EGEE (Enabling Grids for E-science in Europe).
3761 *) RC4 performance overhaul on modern architectures/implementations, such
3762 as Intel P4, IA-64 and AMD64.
3765 *) New utility extract-section.pl. This can be used specify an alternative
3766 section number in a pod file instead of having to treat each file as
3767 a separate case in Makefile. This can be done by adding two lines to the
3770 =for comment openssl_section:XXX
3772 The blank line is mandatory.
3776 *) New arguments -certform, -keyform and -pass for s_client and s_server
3777 to allow alternative format key and certificate files and passphrase
3781 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3782 update associated structures and add various utility functions.
3784 Add new policy related verify parameters, include policy checking in
3785 standard verify code. Enhance 'smime' application with extra parameters
3786 to support policy checking and print out.
3789 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3790 Nehemiah processors. These extensions support AES encryption in hardware
3791 as well as RNG (though RNG support is currently disabled).
3792 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3794 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3797 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3798 [Andy Polyakov and a number of other people]
3800 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3801 implementation contributed by IBM.
3802 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3804 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3805 exponent rather than 'unsigned long'. There is a corresponding change to
3806 the new 'rsa_keygen' element of the RSA_METHOD structure.
3807 [Jelte Jansen, Geoff Thorpe]
3809 *) Functionality for creating the initial serial number file is now
3810 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3812 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3813 number file to 1, which is bound to cause problems. To avoid
3814 the problems while respecting compatibility between different 0.9.7
3815 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3816 CA.pl for serial number initialization. With the new release 0.9.8,
3817 we can fix the problem directly in the 'ca' utility.)
3820 *) Reduced header interdepencies by declaring more opaque objects in
3821 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3822 give fewer recursive includes, which could break lazy source code - so
3823 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3824 developers should define this symbol when building and using openssl to
3825 ensure they track the recommended behaviour, interfaces, [etc], but
3826 backwards-compatible behaviour prevails when this isn't defined.
3829 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3832 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3833 This will generate a random key of the appropriate length based on the
3834 cipher context. The EVP_CIPHER can provide its own random key generation
3835 routine to support keys of a specific form. This is used in the des and
3836 3des routines to generate a key of the correct parity. Update S/MIME
3837 code to use new functions and hence generate correct parity DES keys.
3838 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3839 valid (weak or incorrect parity).
3842 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3843 as looking them up. This is useful when the verified structure may contain
3844 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3845 present unless the new PKCS7_NO_CRL flag is asserted.
3848 *) Extend ASN1 oid configuration module. It now additionally accepts the
3851 shortName = some long name, 1.2.3.4
3854 *) Reimplemented the BN_CTX implementation. There is now no more static
3855 limitation on the number of variables it can handle nor the depth of the
3856 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3857 information can now expand as required, and rather than having a single
3858 static array of bignums, BN_CTX now uses a linked-list of such arrays
3859 allowing it to expand on demand whilst maintaining the usefulness of
3860 BN_CTX's "bundling".
3863 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3864 to allow all RSA operations to function using a single BN_CTX.
3867 *) Preliminary support for certificate policy evaluation and checking. This
3868 is initially intended to pass the tests outlined in "Conformance Testing
3869 of Relying Party Client Certificate Path Processing Logic" v1.07.
3872 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3873 remained unused and not that useful. A variety of other little bignum
3874 tweaks and fixes have also been made continuing on from the audit (see
3878 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3879 associated ASN1, EVP and SSL functions and old ASN1 macros.
3882 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3883 and this should never fail. So the return value from the use of
3884 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3885 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3888 *) BN_CTX_get() should return zero-valued bignums, providing the same
3889 initialised value as BN_new().
3890 [Geoff Thorpe, suggested by Ulf Möller]
3892 *) Support for inhibitAnyPolicy certificate extension.
3895 *) An audit of the BIGNUM code is underway, for which debugging code is
3896 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3897 is considered valid when processing BIGNUMs, and causes execution to
3898 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3899 further steps are taken to deliberately pollute unused data in BIGNUM
3900 structures to try and expose faulty code further on. For now, openssl will
3901 (in its default mode of operation) continue to tolerate the inconsistent
3902 forms that it has tolerated in the past, but authors and packagers should
3903 consider trying openssl and their own applications when compiled with
3904 these debugging symbols defined. It will help highlight potential bugs in
3905 their own code, and will improve the test coverage for OpenSSL itself. At
3906 some point, these tighter rules will become openssl's default to improve
3907 maintainability, though the assert()s and other overheads will remain only
3908 in debugging configurations. See bn.h for more details.
3909 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3911 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3912 that can only be obtained through BN_CTX_new() (which implicitly
3913 initialises it). The presence of this function only made it possible
3914 to overwrite an existing structure (and cause memory leaks).
3917 *) Because of the callback-based approach for implementing LHASH as a
3918 template type, lh_insert() adds opaque objects to hash-tables and
3919 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3920 to clean up those corresponding objects before destroying the hash table
3921 (and losing the object pointers). So some over-zealous constifications in
3922 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3923 objects as "const" and the lh_doall[_arg] callback wrappers are not
3924 prototyped to have "const" restrictions on the object pointers they are
3925 given (and so aren't required to cast them away any more).
3928 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3929 (speed) prefers to use its own implementation. The two implementations
3930 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3931 its object type properly exposed (MS_TM) instead of casting to/from "char
3932 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3933 aren't necessarily the greatest nomenclatures - but this is what was used
3934 internally to the implementation so I've used that for now.
3937 *) Ensure that deprecated functions do not get compiled when
3938 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3939 the self-tests were still using deprecated key-generation functions so
3940 these have been updated also.
3943 *) Reorganise PKCS#7 code to separate the digest location functionality
3944 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3945 New function PKCS7_set_digest() to set the digest type for PKCS#7
3946 digestedData type. Add additional code to correctly generate the
3947 digestedData type and add support for this type in PKCS7 initialization
3951 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3952 structure of type "other".
3955 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3956 sure the loop does correctly stop and breaking ("division by zero")
3957 modulus operations are not performed. The (pre-generated) prime
3958 table crypto/bn/bn_prime.h was already correct, but it could not be
3959 re-generated on some platforms because of the "division by zero"
3960 situation in the script.
3961 [Ralf S. Engelschall]
3963 *) Update support for ECC-based TLS ciphersuites according to
3964 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3965 SHA-1 now is only used for "small" curves (where the
3966 representation of a field element takes up to 24 bytes); for
3967 larger curves, the field element resulting from ECDH is directly
3968 used as premaster secret.
3969 [Douglas Stebila (Sun Microsystems Laboratories)]
3971 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3972 curve secp160r1 to the tests.
3973 [Douglas Stebila (Sun Microsystems Laboratories)]
3975 *) Add the possibility to load symbols globally with DSO.
3976 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3978 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3979 control of the error stack.
3982 *) Add support for STORE in ENGINE.
3985 *) Add the STORE type. The intention is to provide a common interface
3986 to certificate and key stores, be they simple file-based stores, or
3987 HSM-type store, or LDAP stores, or...
3988 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3991 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3992 pass a list of arguments to any function as well as provide a way
3993 for a function to pass data back to the caller.
3996 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3997 works like BUF_strdup() but can be used to duplicate a portion of
3998 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
4002 *) Add the function sk_find_ex() which works like sk_find(), but will
4003 return an index to an element even if an exact match couldn't be
4004 found. The index is guaranteed to point at the element where the
4005 searched-for key would be inserted to preserve sorting order.
4008 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
4009 takes an extra flags argument for optional functionality. Currently,
4010 the following flags are defined:
4012 OBJ_BSEARCH_VALUE_ON_NOMATCH
4013 This one gets OBJ_bsearch_ex() to return a pointer to the first
4014 element where the comparing function returns a negative or zero
4017 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
4018 This one gets OBJ_bsearch_ex() to return a pointer to the first
4019 element where the comparing function returns zero. This is useful
4020 if there are more than one element where the comparing function
4024 *) Make it possible to create self-signed certificates with 'openssl ca'
4025 in such a way that the self-signed certificate becomes part of the
4026 CA database and uses the same mechanisms for serial number generation
4027 as all other certificate signing. The new flag '-selfsign' enables
4028 this functionality. Adapt CA.sh and CA.pl.in.
4031 *) Add functionality to check the public key of a certificate request
4032 against a given private. This is useful to check that a certificate
4033 request can be signed by that key (self-signing).
4036 *) Make it possible to have multiple active certificates with the same
4037 subject in the CA index file. This is done only if the keyword
4038 'unique_subject' is set to 'no' in the main CA section (default
4039 if 'CA_default') of the configuration file. The value is saved
4040 with the database itself in a separate index attribute file,
4041 named like the index file with '.attr' appended to the name.
4044 *) Generate muti valued AVAs using '+' notation in config files for
4048 *) Support for nameConstraints certificate extension.
4051 *) Support for policyConstraints certificate extension.
4054 *) Support for policyMappings certificate extension.
4057 *) Make sure the default DSA_METHOD implementation only uses its
4058 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
4059 and change its own handlers to be NULL so as to remove unnecessary
4060 indirection. This lets alternative implementations fallback to the
4061 default implementation more easily.
4064 *) Support for directoryName in GeneralName related extensions
4068 *) Make it possible to link applications using Makefile.shared.
4069 Make that possible even when linking against static libraries!
4072 *) Support for single pass processing for S/MIME signing. This now
4073 means that S/MIME signing can be done from a pipe, in addition
4074 cleartext signing (multipart/signed type) is effectively streaming
4075 and the signed data does not need to be all held in memory.
4077 This is done with a new flag PKCS7_STREAM. When this flag is set
4078 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
4079 is done after the data is output (and digests calculated) in
4080 SMIME_write_PKCS7().
4083 *) Add full support for -rpath/-R, both in shared libraries and
4084 applications, at least on the platforms where it's known how
4088 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
4089 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
4090 will now compute a table of multiples of the generator that
4091 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
4092 faster (notably in the case of a single point multiplication,
4093 scalar * generator).
4094 [Nils Larsch, Bodo Moeller]
4096 *) IPv6 support for certificate extensions. The various extensions
4097 which use the IP:a.b.c.d can now take IPv6 addresses using the
4098 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
4102 *) Added an ENGINE that implements RSA by performing private key
4103 exponentiations with the GMP library. The conversions to and from
4104 GMP's mpz_t format aren't optimised nor are any montgomery forms
4105 cached, and on x86 it appears OpenSSL's own performance has caught up.
4106 However there are likely to be other architectures where GMP could
4107 provide a boost. This ENGINE is not built in by default, but it can be
4108 specified at Configure time and should be accompanied by the necessary
4109 linker additions, eg;
4110 ./config -DOPENSSL_USE_GMP -lgmp
4113 *) "openssl engine" will not display ENGINE/DSO load failure errors when
4114 testing availability of engines with "-t" - the old behaviour is
4115 produced by increasing the feature's verbosity with "-tt".
4118 *) ECDSA routines: under certain error conditions uninitialized BN objects
4119 could be freed. Solution: make sure initialization is performed early
4120 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
4124 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
4125 and DH_METHOD (eg. by ENGINE implementations) to override the normal
4126 software implementations. For DSA and DH, parameter generation can
4127 also be overriden by providing the appropriate method callbacks.
4130 *) Change the "progress" mechanism used in key-generation and
4131 primality testing to functions that take a new BN_GENCB pointer in
4132 place of callback/argument pairs. The new API functions have "_ex"
4133 postfixes and the older functions are reimplemented as wrappers for
4134 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4135 declarations of the old functions to help (graceful) attempts to
4136 migrate to the new functions. Also, the new key-generation API
4137 functions operate on a caller-supplied key-structure and return
4138 success/failure rather than returning a key or NULL - this is to
4139 help make "keygen" another member function of RSA_METHOD etc.
4141 Example for using the new callback interface:
4143 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4147 BN_GENCB_set(&my_cb, my_callback, my_arg);
4149 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4150 /* For the meaning of a, b in calls to my_callback(), see the
4151 * documentation of the function that calls the callback.
4152 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4153 * my_callback should return 1 if it wants BN_is_prime_ex()
4154 * to continue, or 0 to stop.
4159 *) Change the ZLIB compression method to be stateful, and make it
4160 available to TLS with the number defined in
4161 draft-ietf-tls-compression-04.txt.
4164 *) Add the ASN.1 structures and functions for CertificatePair, which
4165 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4167 CertificatePair ::= SEQUENCE {
4168 forward [0] Certificate OPTIONAL,
4169 reverse [1] Certificate OPTIONAL,
4170 -- at least one of the pair shall be present -- }
4172 Also implement the PEM functions to read and write certificate
4173 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4175 This needed to be defined, mostly for the sake of the LDAP
4176 attribute crossCertificatePair, but may prove useful elsewhere as
4180 *) Make it possible to inhibit symlinking of shared libraries in
4181 Makefile.shared, for Cygwin's sake.
4184 *) Extend the BIGNUM API by creating a function
4185 void BN_set_negative(BIGNUM *a, int neg);
4186 and a macro that behave like
4187 int BN_is_negative(const BIGNUM *a);
4189 to avoid the need to access 'a->neg' directly in applications.
4192 *) Implement fast modular reduction for pseudo-Mersenne primes
4193 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4194 EC_GROUP_new_curve_GFp() will now automatically use this
4196 [Nils Larsch <nla@trustcenter.de>]
4198 *) Add new lock type (CRYPTO_LOCK_BN).
4201 *) Change the ENGINE framework to automatically load engines
4202 dynamically from specific directories unless they could be
4203 found to already be built in or loaded. Move all the
4204 current engines except for the cryptodev one to a new
4206 The engines in engines/ are built as shared libraries if
4207 the "shared" options was given to ./Configure or ./config.
4208 Otherwise, they are inserted in libcrypto.a.
4209 /usr/local/ssl/engines is the default directory for dynamic
4210 engines, but that can be overriden at configure time through
4211 the usual use of --prefix and/or --openssldir, and at run
4212 time with the environment variable OPENSSL_ENGINES.
4213 [Geoff Thorpe and Richard Levitte]
4215 *) Add Makefile.shared, a helper makefile to build shared
4216 libraries. Adapt Makefile.org.
4219 *) Add version info to Win32 DLLs.
4220 [Peter 'Luna' Runestig" <peter@runestig.com>]
4222 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4223 can be added using this API to created arbitrary PKCS#12
4224 files while avoiding the low level API.
4226 New options to PKCS12_create(), key or cert can be NULL and
4227 will then be omitted from the output file. The encryption
4228 algorithm NIDs can be set to -1 for no encryption, the mac
4229 iteration count can be set to 0 to omit the mac.
4231 Enhance pkcs12 utility by making the -nokeys and -nocerts
4232 options work when creating a PKCS#12 file. New option -nomac
4233 to omit the mac, NONE can be set for an encryption algorithm.
4234 New code is modified to use the enhanced PKCS12_create()
4235 instead of the low level API.
4238 *) Extend ASN1 encoder to support indefinite length constructed
4239 encoding. This can output sequences tags and octet strings in
4240 this form. Modify pk7_asn1.c to support indefinite length
4241 encoding. This is experimental and needs additional code to
4242 be useful, such as an ASN1 bio and some enhanced streaming
4245 Extend template encode functionality so that tagging is passed
4246 down to the template encoder.
4249 *) Let 'openssl req' fail if an argument to '-newkey' is not
4250 recognized instead of using RSA as a default.
4253 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4254 As these are not official, they are not included in "ALL";
4255 the "ECCdraft" ciphersuite group alias can be used to select them.
4256 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4258 *) Add ECDH engine support.
4259 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4261 *) Add ECDH in new directory crypto/ecdh/.
4262 [Douglas Stebila (Sun Microsystems Laboratories)]
4264 *) Let BN_rand_range() abort with an error after 100 iterations
4265 without success (which indicates a broken PRNG).
4268 *) Change BN_mod_sqrt() so that it verifies that the input value
4269 is really the square of the return value. (Previously,
4270 BN_mod_sqrt would show GIGO behaviour.)
4273 *) Add named elliptic curves over binary fields from X9.62, SECG,
4274 and WAP/WTLS; add OIDs that were still missing.
4276 [Sheueling Chang Shantz and Douglas Stebila
4277 (Sun Microsystems Laboratories)]
4279 *) Extend the EC library for elliptic curves over binary fields
4280 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4283 EC_GF2m_simple_method
4287 EC_GROUP_new_curve_GF2m
4288 EC_GROUP_set_curve_GF2m
4289 EC_GROUP_get_curve_GF2m
4290 EC_POINT_set_affine_coordinates_GF2m
4291 EC_POINT_get_affine_coordinates_GF2m
4292 EC_POINT_set_compressed_coordinates_GF2m
4294 Point compression for binary fields is disabled by default for
4295 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4298 As binary polynomials are represented as BIGNUMs, various members
4299 of the EC_GROUP and EC_POINT data structures can be shared
4300 between the implementations for prime fields and binary fields;
4301 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4302 are essentially identical to their ..._GFp counterparts.
4303 (For simplicity, the '..._GFp' prefix has been dropped from
4304 various internal method names.)
4306 An internal 'field_div' method (similar to 'field_mul' and
4307 'field_sqr') has been added; this is used only for binary fields.
4309 [Sheueling Chang Shantz and Douglas Stebila
4310 (Sun Microsystems Laboratories)]
4312 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4313 through methods ('mul', 'precompute_mult').
4315 The generic implementations (now internally called 'ec_wNAF_mul'
4316 and 'ec_wNAF_precomputed_mult') remain the default if these
4317 methods are undefined.
4319 [Sheueling Chang Shantz and Douglas Stebila
4320 (Sun Microsystems Laboratories)]
4322 *) New function EC_GROUP_get_degree, which is defined through
4323 EC_METHOD. For curves over prime fields, this returns the bit
4324 length of the modulus.
4326 [Sheueling Chang Shantz and Douglas Stebila
4327 (Sun Microsystems Laboratories)]
4329 *) New functions EC_GROUP_dup, EC_POINT_dup.
4330 (These simply call ..._new and ..._copy).
4332 [Sheueling Chang Shantz and Douglas Stebila
4333 (Sun Microsystems Laboratories)]
4335 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4336 Polynomials are represented as BIGNUMs (where the sign bit is not
4337 used) in the following functions [macros]:
4340 BN_GF2m_sub [= BN_GF2m_add]
4341 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4342 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4343 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4345 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4346 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4347 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4348 BN_GF2m_cmp [= BN_ucmp]
4350 (Note that only the 'mod' functions are actually for fields GF(2^m).
4351 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4353 For some functions, an the irreducible polynomial defining a
4354 field can be given as an 'unsigned int[]' with strictly
4355 decreasing elements giving the indices of those bits that are set;
4356 i.e., p[] represents the polynomial
4357 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4359 p[0] > p[1] > ... > p[k] = 0.
4360 This applies to the following functions:
4365 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4366 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4368 BN_GF2m_mod_sqrt_arr
4369 BN_GF2m_mod_solve_quad_arr
4373 Conversion can be performed by the following functions:
4378 bntest.c has additional tests for binary polynomial arithmetic.
4380 Two implementations for BN_GF2m_mod_div() are available.
4381 The default algorithm simply uses BN_GF2m_mod_inv() and
4382 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4383 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4384 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4386 [Sheueling Chang Shantz and Douglas Stebila
4387 (Sun Microsystems Laboratories)]
4389 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4390 functionality is disabled at compile-time.
4391 [Douglas Stebila <douglas.stebila@sun.com>]
4393 *) Change default behaviour of 'openssl asn1parse' so that more
4394 information is visible when viewing, e.g., a certificate:
4396 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4397 mode the content of non-printable OCTET STRINGs is output in a
4398 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4399 avoid the appearance of a printable string.
4400 [Nils Larsch <nla@trustcenter.de>]
4402 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4404 EC_GROUP_set_asn1_flag()
4405 EC_GROUP_get_asn1_flag()
4406 EC_GROUP_set_point_conversion_form()
4407 EC_GROUP_get_point_conversion_form()
4408 These control ASN1 encoding details:
4409 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4410 has been set to OPENSSL_EC_NAMED_CURVE.
4411 - Points are encoded in uncompressed form by default; options for
4412 asn1_for are as for point2oct, namely
4413 POINT_CONVERSION_COMPRESSED
4414 POINT_CONVERSION_UNCOMPRESSED
4415 POINT_CONVERSION_HYBRID
4417 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4420 EC_GROUP_get0_seed()
4421 EC_GROUP_get_seed_len()
4422 This is used only for ASN1 purposes (so far).
4423 [Nils Larsch <nla@trustcenter.de>]
4425 *) Add 'field_type' member to EC_METHOD, which holds the NID
4426 of the appropriate field type OID. The new function
4427 EC_METHOD_get_field_type() returns this value.
4428 [Nils Larsch <nla@trustcenter.de>]
4433 EC_POINT_point2hex()
4434 EC_POINT_hex2point()
4435 providing useful interfaces to EC_POINT_point2oct() and
4436 EC_POINT_oct2point().
4437 [Nils Larsch <nla@trustcenter.de>]
4439 *) Change internals of the EC library so that the functions
4440 EC_GROUP_set_generator()
4441 EC_GROUP_get_generator()
4442 EC_GROUP_get_order()
4443 EC_GROUP_get_cofactor()
4444 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4445 to methods, which would lead to unnecessary code duplication when
4446 adding different types of curves.
4447 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4449 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4450 arithmetic, and such that modified wNAFs are generated
4451 (which avoid length expansion in many cases).
4454 *) Add a function EC_GROUP_check_discriminant() (defined via
4455 EC_METHOD) that verifies that the curve discriminant is non-zero.
4457 Add a function EC_GROUP_check() that makes some sanity tests
4458 on a EC_GROUP, its generator and order. This includes
4459 EC_GROUP_check_discriminant().
4460 [Nils Larsch <nla@trustcenter.de>]
4462 *) Add ECDSA in new directory crypto/ecdsa/.
4464 Add applications 'openssl ecparam' and 'openssl ecdsa'
4465 (these are based on 'openssl dsaparam' and 'openssl dsa').
4467 ECDSA support is also included in various other files across the
4468 library. Most notably,
4469 - 'openssl req' now has a '-newkey ecdsa:file' option;
4470 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4471 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4472 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4473 them suitable for ECDSA where domain parameters must be
4474 extracted before the specific public key;
4475 - ECDSA engine support has been added.
4476 [Nils Larsch <nla@trustcenter.de>]
4478 *) Include some named elliptic curves, and add OIDs from X9.62,
4479 SECG, and WAP/WTLS. Each curve can be obtained from the new
4481 EC_GROUP_new_by_curve_name(),
4482 and the list of available named curves can be obtained with
4483 EC_get_builtin_curves().
4484 Also add a 'curve_name' member to EC_GROUP objects, which can be
4486 EC_GROUP_set_curve_name()
4487 EC_GROUP_get_curve_name()
4488 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4490 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4491 was actually never needed) and in BN_mul(). The removal in BN_mul()
4492 required a small change in bn_mul_part_recursive() and the addition
4493 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4494 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4495 bn_sub_words() and bn_add_words() except they take arrays with
4499 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4501 *) Cleanse PEM buffers before freeing them since they may contain
4503 [Benjamin Bennett <ben@psc.edu>]
4505 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4506 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4507 authentication-only ciphersuites.
4510 *) Since AES128 and AES256 share a single mask bit in the logic of
4511 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4512 kludge to work properly if AES128 is available and AES256 isn't.
4515 *) Expand security boundary to match 1.1.1 module.
4518 *) Remove redundant features: hash file source, editing of test vectors
4519 modify fipsld to use external fips_premain.c signature.
4522 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4523 run algorithm test programs.
4526 *) Make algorithm test programs more tolerant of whitespace.
4529 *) Have SSL/TLS server implementation tolerate "mismatched" record
4530 protocol version while receiving ClientHello even if the
4531 ClientHello is fragmented. (The server can't insist on the
4532 particular protocol version it has chosen before the ServerHello
4533 message has informed the client about his choice.)
4536 *) Load error codes if they are not already present instead of using a
4537 static variable. This allows them to be cleanly unloaded and reloaded.
4540 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4542 *) Introduce limits to prevent malicious keys being able to
4543 cause a denial of service. (CVE-2006-2940)
4544 [Steve Henson, Bodo Moeller]
4546 *) Fix ASN.1 parsing of certain invalid structures that can result
4547 in a denial of service. (CVE-2006-2937) [Steve Henson]
4549 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4550 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4552 *) Fix SSL client code which could crash if connecting to a
4553 malicious SSLv2 server. (CVE-2006-4343)
4554 [Tavis Ormandy and Will Drewry, Google Security Team]
4556 *) Change ciphersuite string processing so that an explicit
4557 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4558 will no longer include "AES128-SHA"), and any other similar
4559 ciphersuite (same bitmap) from *other* protocol versions (so that
4560 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4561 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4562 changes from 0.9.8b and 0.9.8d.
4565 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4567 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4568 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4570 *) Change the Unix randomness entropy gathering to use poll() when
4571 possible instead of select(), since the latter has some
4572 undesirable limitations.
4573 [Darryl Miles via Richard Levitte and Bodo Moeller]
4575 *) Disable rogue ciphersuites:
4577 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4578 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4579 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4581 The latter two were purportedly from
4582 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4585 Also deactive the remaining ciphersuites from
4586 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4587 unofficial, and the ID has long expired.
4590 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4591 dual-core machines) and other potential thread-safety issues.
4594 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4596 *) Adapt fipsld and the build system to link against the validated FIPS
4597 module in FIPS mode.
4600 *) Fixes for VC++ 2005 build under Windows.
4603 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4604 from a Windows bash shell such as MSYS. It is autodetected from the
4605 "config" script when run from a VC++ environment. Modify standard VC++
4606 build to use fipscanister.o from the GNU make build.
4609 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4611 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4612 The value now differs depending on if you build for FIPS or not.
4613 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4614 safely run with a non-FIPSed libcrypto, as it may crash because of
4615 the difference induced by this change.
4618 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4620 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4621 (part of SSL_OP_ALL). This option used to disable the
4622 countermeasure against man-in-the-middle protocol-version
4623 rollback in the SSL 2.0 server implementation, which is a bad
4624 idea. (CVE-2005-2969)
4626 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4627 for Information Security, National Institute of Advanced Industrial
4628 Science and Technology [AIST], Japan)]
4630 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4631 mainly for FIPS compliance and not fully integrated at this stage.
4634 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4635 the exponentiation using a fixed-length exponent. (Otherwise,
4636 the information leaked through timing could expose the secret key
4637 after many signatures; cf. Bleichenbacher's attack on DSA with
4641 *) Make a new fixed-window mod_exp implementation the default for
4642 RSA, DSA, and DH private-key operations so that the sequence of
4643 squares and multiplies and the memory access pattern are
4644 independent of the particular secret key. This will mitigate
4645 cache-timing and potential related attacks.
4647 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4648 and this is automatically used by BN_mod_exp_mont() if the new flag
4649 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4650 will use this BN flag for private exponents unless the flag
4651 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4652 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4654 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4656 *) Change the client implementation for SSLv23_method() and
4657 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4658 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4659 (Previously, the SSL 2.0 backwards compatible Client Hello
4660 message format would be used even with SSL_OP_NO_SSLv2.)
4663 *) Add support for smime-type MIME parameter in S/MIME messages which some
4667 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4668 a threadsafe manner. Modify rsa code to use new function and add calls
4669 to dsa and dh code (which had race conditions before).
4672 *) Include the fixed error library code in the C error file definitions
4673 instead of fixing them up at runtime. This keeps the error code
4674 structures constant.
4677 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4679 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4682 *) Fixes for newer kerberos headers. NB: the casts are needed because
4683 the 'length' field is signed on one version and unsigned on another
4684 with no (?) obvious way to tell the difference, without these VC++
4685 complains. Also the "definition" of FAR (blank) is no longer included
4686 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4687 some needed definitions.
4690 *) Undo Cygwin change.
4693 *) Added support for proxy certificates according to RFC 3820.
4694 Because they may be a security thread to unaware applications,
4695 they must be explicitely allowed in run-time. See
4696 docs/HOWTO/proxy_certificates.txt for further information.
4699 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4701 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4702 server and client random values. Previously
4703 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4704 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4706 This change has negligible security impact because:
4708 1. Server and client random values still have 24 bytes of pseudo random
4711 2. Server and client random values are sent in the clear in the initial
4714 3. The master secret is derived using the premaster secret (48 bytes in
4715 size for static RSA ciphersuites) as well as client server and random
4718 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4721 [Stephen Henson, reported by UK NISCC]
4723 *) Use Windows randomness collection on Cygwin.
4726 *) Fix hang in EGD/PRNGD query when communication socket is closed
4727 prematurely by EGD/PRNGD.
4728 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4730 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4733 *) Back-port of selected performance improvements from development
4734 branch, as well as improved support for PowerPC platforms.
4737 *) Add lots of checks for memory allocation failure, error codes to indicate
4738 failure and freeing up memory if a failure occurs.
4739 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4741 *) Add new -passin argument to dgst.
4744 *) Perform some character comparisons of different types in X509_NAME_cmp:
4745 this is needed for some certificates that reencode DNs into UTF8Strings
4746 (in violation of RFC3280) and can't or wont issue name rollover
4750 *) Make an explicit check during certificate validation to see that
4751 the CA setting in each certificate on the chain is correct. As a
4752 side effect always do the following basic checks on extensions,
4753 not just when there's an associated purpose to the check:
4755 - if there is an unhandled critical extension (unless the user
4756 has chosen to ignore this fault)
4757 - if the path length has been exceeded (if one is set at all)
4758 - that certain extensions fit the associated purpose (if one has
4762 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4764 *) Avoid a race condition when CRLs are checked in a multi threaded
4765 environment. This would happen due to the reordering of the revoked
4766 entries during signature checking and serial number lookup. Now the
4767 encoding is cached and the serial number sort performed under a lock.
4768 Add new STACK function sk_is_sorted().
4771 *) Add Delta CRL to the extension code.
4774 *) Various fixes to s3_pkt.c so alerts are sent properly.
4775 [David Holmes <d.holmes@f5.com>]
4777 *) Reduce the chances of duplicate issuer name and serial numbers (in
4778 violation of RFC3280) using the OpenSSL certificate creation utilities.
4779 This is done by creating a random 64 bit value for the initial serial
4780 number when a serial number file is created or when a self signed
4781 certificate is created using 'openssl req -x509'. The initial serial
4782 number file is created using 'openssl x509 -next_serial' in CA.pl
4783 rather than being initialized to 1.
4786 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4788 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4789 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4790 [Joe Orton, Steve Henson]
4792 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4794 [Joe Orton, Steve Henson]
4796 *) Make it possible to have multiple active certificates with the same
4797 subject in the CA index file. This is done only if the keyword
4798 'unique_subject' is set to 'no' in the main CA section (default
4799 if 'CA_default') of the configuration file. The value is saved
4800 with the database itself in a separate index attribute file,
4801 named like the index file with '.attr' appended to the name.
4804 *) X509 verify fixes. Disable broken certificate workarounds when
4805 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4806 keyUsage extension present. Don't accept CRLs with unhandled critical
4807 extensions: since verify currently doesn't process CRL extensions this
4808 rejects a CRL with *any* critical extensions. Add new verify error codes
4812 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4813 A clarification of RFC2560 will require the use of OCTET STRINGs and
4814 some implementations cannot handle the current raw format. Since OpenSSL
4815 copies and compares OCSP nonces as opaque blobs without any attempt at
4816 parsing them this should not create any compatibility issues.
4819 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4820 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4821 this HMAC (and other) operations are several times slower than OpenSSL
4825 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4826 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4828 *) Use the correct content when signing type "other".
4831 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4833 *) Fix various bugs revealed by running the NISCC test suite:
4835 Stop out of bounds reads in the ASN1 code when presented with
4836 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4838 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4840 If verify callback ignores invalid public key errors don't try to check
4841 certificate signature with the NULL public key.
4845 *) New -ignore_err option in ocsp application to stop the server
4846 exiting on the first error in a request.
4849 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4850 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4854 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4855 extra data after the compression methods not only for TLS 1.0
4856 but also for SSL 3.0 (as required by the specification).
4857 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4859 *) Change X509_certificate_type() to mark the key as exported/exportable
4860 when it's 512 *bits* long, not 512 bytes.
4863 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4864 blocks during encryption.
4867 *) Various fixes to base64 BIO and non blocking I/O. On write
4868 flushes were not handled properly if the BIO retried. On read
4869 data was not being buffered properly and had various logic bugs.
4870 This also affects blocking I/O when the data being decoded is a
4874 *) Various S/MIME bugfixes and compatibility changes:
4875 output correct application/pkcs7 MIME type if
4876 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4877 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4878 of files as .eml work). Correctly handle very long lines in MIME
4882 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4884 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4885 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4886 a protocol version number mismatch like a decryption error
4887 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4890 *) Turn on RSA blinding by default in the default implementation
4891 to avoid a timing attack. Applications that don't want it can call
4892 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4893 They would be ill-advised to do so in most cases.
4894 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4896 *) Change RSA blinding code so that it works when the PRNG is not
4897 seeded (in this case, the secret RSA exponent is abused as
4898 an unpredictable seed -- if it is not unpredictable, there
4899 is no point in blinding anyway). Make RSA blinding thread-safe
4900 by remembering the creator's thread ID in rsa->blinding and
4901 having all other threads use local one-time blinding factors
4902 (this requires more computation than sharing rsa->blinding, but
4903 avoids excessive locking; and if an RSA object is not shared
4904 between threads, blinding will still be very fast).
4907 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4908 ENGINE as defaults for all supported algorithms irrespective of
4909 the 'flags' parameter. 'flags' is now honoured, so applications
4910 should make sure they are passing it correctly.
4913 *) Target "mingw" now allows native Windows code to be generated in
4914 the Cygwin environment as well as with the MinGW compiler.
4917 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4919 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4920 via timing by performing a MAC computation even if incorrect
4921 block cipher padding has been found. This is a countermeasure
4922 against active attacks where the attacker has to distinguish
4923 between bad padding and a MAC verification error. (CVE-2003-0078)
4925 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4926 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4927 Martin Vuagnoux (EPFL, Ilion)]
4929 *) Make the no-err option work as intended. The intention with no-err
4930 is not to have the whole error stack handling routines removed from
4931 libcrypto, it's only intended to remove all the function name and
4932 reason texts, thereby removing some of the footprint that may not
4933 be interesting if those errors aren't displayed anyway.
4935 NOTE: it's still possible for any application or module to have it's
4936 own set of error texts inserted. The routines are there, just not
4937 used by default when no-err is given.
4940 *) Add support for FreeBSD on IA64.
4941 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4943 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4944 Kerberos function mit_des_cbc_cksum(). Before this change,
4945 the value returned by DES_cbc_cksum() was like the one from
4946 mit_des_cbc_cksum(), except the bytes were swapped.
4947 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4949 *) Allow an application to disable the automatic SSL chain building.
4950 Before this a rather primitive chain build was always performed in
4951 ssl3_output_cert_chain(): an application had no way to send the
4952 correct chain if the automatic operation produced an incorrect result.
4954 Now the chain builder is disabled if either:
4956 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4958 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4960 The reasoning behind this is that an application would not want the
4961 auto chain building to take place if extra chain certificates are
4962 present and it might also want a means of sending no additional
4963 certificates (for example the chain has two certificates and the
4967 *) Add the possibility to build without the ENGINE framework.
4968 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4970 *) Under Win32 gmtime() can return NULL: check return value in
4971 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4974 *) DSA routines: under certain error conditions uninitialized BN objects
4975 could be freed. Solution: make sure initialization is performed early
4976 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4977 Nils Larsch <nla@trustcenter.de> via PR#459)
4980 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4981 checked on reconnect on the client side, therefore session resumption
4982 could still fail with a "ssl session id is different" error. This
4983 behaviour is masked when SSL_OP_ALL is used due to
4984 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4985 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4986 followup to PR #377.
4989 *) IA-32 assembler support enhancements: unified ELF targets, support
4990 for SCO/Caldera platforms, fix for Cygwin shared build.
4993 *) Add support for FreeBSD on sparc64. As a consequence, support for
4994 FreeBSD on non-x86 processors is separate from x86 processors on
4995 the config script, much like the NetBSD support.
4996 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4998 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
5000 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
5003 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
5004 code (06) was taken as the first octet of the session ID and the last
5005 octet was ignored consequently. As a result SSLv2 client side session
5006 caching could not have worked due to the session ID mismatch between
5008 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5012 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
5013 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
5017 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
5018 seems that in spite of existing for more than a year, many application
5019 author have done nothing to provide the necessary callbacks, which
5020 means that this particular engine will not work properly anywhere.
5021 This is a very unfortunate situation which forces us, in the name
5022 of usability, to give the hw_ncipher.c a static lock, which is part
5024 NOTE: This is for the 0.9.7 series ONLY. This hack will never
5025 appear in 0.9.8 or later. We EXPECT application authors to have
5026 dealt properly with this when 0.9.8 is released (unless we actually
5027 make such changes in the libcrypto locking code that changes will
5028 have to be made anyway).
5031 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
5032 octets have been read, EOF or an error occurs. Without this change
5033 some truncated ASN1 structures will not produce an error.
5036 *) Disable Heimdal support, since it hasn't been fully implemented.
5037 Still give the possibility to force the use of Heimdal, but with
5038 warnings and a request that patches get sent to openssl-dev.
5041 *) Add the VC-CE target, introduce the WINCE sysname, and add
5042 INSTALL.WCE and appropriate conditionals to make it build.
5043 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5045 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
5046 cygssl-x.y.z.dll, where x, y and z are the major, minor and
5047 edit numbers of the version.
5048 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5050 *) Introduce safe string copy and catenation functions
5051 (BUF_strlcpy() and BUF_strlcat()).
5052 [Ben Laurie (CHATS) and Richard Levitte]
5054 *) Avoid using fixed-size buffers for one-line DNs.
5055 [Ben Laurie (CHATS)]
5057 *) Add BUF_MEM_grow_clean() to avoid information leakage when
5058 resizing buffers containing secrets, and use where appropriate.
5059 [Ben Laurie (CHATS)]
5061 *) Avoid using fixed size buffers for configuration file location.
5062 [Ben Laurie (CHATS)]
5064 *) Avoid filename truncation for various CA files.
5065 [Ben Laurie (CHATS)]
5067 *) Use sizeof in preference to magic numbers.
5068 [Ben Laurie (CHATS)]
5070 *) Avoid filename truncation in cert requests.
5071 [Ben Laurie (CHATS)]
5073 *) Add assertions to check for (supposedly impossible) buffer
5075 [Ben Laurie (CHATS)]
5077 *) Don't cache truncated DNS entries in the local cache (this could
5078 potentially lead to a spoofing attack).
5079 [Ben Laurie (CHATS)]
5081 *) Fix various buffers to be large enough for hex/decimal
5082 representations in a platform independent manner.
5083 [Ben Laurie (CHATS)]
5085 *) Add CRYPTO_realloc_clean() to avoid information leakage when
5086 resizing buffers containing secrets, and use where appropriate.
5087 [Ben Laurie (CHATS)]
5089 *) Add BIO_indent() to avoid much slightly worrying code to do
5091 [Ben Laurie (CHATS)]
5093 *) Convert sprintf()/BIO_puts() to BIO_printf().
5094 [Ben Laurie (CHATS)]
5096 *) buffer_gets() could terminate with the buffer only half
5098 [Ben Laurie (CHATS)]
5100 *) Add assertions to prevent user-supplied crypto functions from
5101 overflowing internal buffers by having large block sizes, etc.
5102 [Ben Laurie (CHATS)]
5104 *) New OPENSSL_assert() macro (similar to assert(), but enabled
5106 [Ben Laurie (CHATS)]
5108 *) Eliminate unused copy of key in RC4.
5109 [Ben Laurie (CHATS)]
5111 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
5112 [Ben Laurie (CHATS)]
5114 *) Fix off-by-one error in EGD path.
5115 [Ben Laurie (CHATS)]
5117 *) If RANDFILE path is too long, ignore instead of truncating.
5118 [Ben Laurie (CHATS)]
5120 *) Eliminate unused and incorrectly sized X.509 structure
5122 [Ben Laurie (CHATS)]
5124 *) Eliminate unused and dangerous function knumber().
5125 [Ben Laurie (CHATS)]
5127 *) Eliminate unused and dangerous structure, KSSL_ERR.
5128 [Ben Laurie (CHATS)]
5130 *) Protect against overlong session ID context length in an encoded
5131 session object. Since these are local, this does not appear to be
5133 [Ben Laurie (CHATS)]
5135 *) Change from security patch (see 0.9.6e below) that did not affect
5136 the 0.9.6 release series:
5138 Remote buffer overflow in SSL3 protocol - an attacker could
5139 supply an oversized master key in Kerberos-enabled versions.
5141 [Ben Laurie (CHATS)]
5143 *) Change the SSL kerb5 codes to match RFC 2712.
5146 *) Make -nameopt work fully for req and add -reqopt switch.
5147 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5149 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5150 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5152 *) Make sure tests can be performed even if the corresponding algorithms
5153 have been removed entirely. This was also the last step to make
5154 OpenSSL compilable with DJGPP under all reasonable conditions.
5155 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5157 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5158 to allow version independent disabling of normally unselected ciphers,
5159 which may be activated as a side-effect of selecting a single cipher.
5161 (E.g., cipher list string "RSA" enables ciphersuites that are left
5162 out of "ALL" because they do not provide symmetric encryption.
5163 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5164 [Lutz Jaenicke, Bodo Moeller]
5166 *) Add appropriate support for separate platform-dependent build
5167 directories. The recommended way to make a platform-dependent
5168 build directory is the following (tested on Linux), maybe with
5171 # Place yourself outside of the OpenSSL source tree. In
5172 # this example, the environment variable OPENSSL_SOURCE
5173 # is assumed to contain the absolute OpenSSL source directory.
5174 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5175 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5176 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5177 mkdir -p `dirname $F`
5178 ln -s $OPENSSL_SOURCE/$F $F
5181 To be absolutely sure not to disturb the source tree, a "make clean"
5182 is a good thing. If it isn't successful, don't worry about it,
5183 it probably means the source directory is very clean.
5186 *) Make sure any ENGINE control commands make local copies of string
5187 pointers passed to them whenever necessary. Otherwise it is possible
5188 the caller may have overwritten (or deallocated) the original string
5189 data when a later ENGINE operation tries to use the stored values.
5190 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5192 *) Improve diagnostics in file reading and command-line digests.
5193 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5195 *) Add AES modes CFB and OFB to the object database. Correct an
5196 error in AES-CFB decryption.
5199 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5200 allows existing EVP_CIPHER_CTX structures to be reused after
5201 calling EVP_*Final(). This behaviour is used by encryption
5202 BIOs and some applications. This has the side effect that
5203 applications must explicitly clean up cipher contexts with
5204 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5207 *) Check the values of dna and dnb in bn_mul_recursive before calling
5208 bn_mul_comba (a non zero value means the a or b arrays do not contain
5209 n2 elements) and fallback to bn_mul_normal if either is not zero.
5212 *) Fix escaping of non-ASCII characters when using the -subj option
5213 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5216 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5217 form for "surname", serialNumber has no short form.
5218 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5219 therefore remove "mail" short name for "internet 7".
5220 The OID for unique identifiers in X509 certificates is
5221 x500UniqueIdentifier, not uniqueIdentifier.
5222 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5225 *) Add an "init" command to the ENGINE config module and auto initialize
5226 ENGINEs. Without any "init" command the ENGINE will be initialized
5227 after all ctrl commands have been executed on it. If init=1 the
5228 ENGINE is initailized at that point (ctrls before that point are run
5229 on the uninitialized ENGINE and after on the initialized one). If
5230 init=0 then the ENGINE will not be iniatialized at all.
5233 *) Fix the 'app_verify_callback' interface so that the user-defined
5234 argument is actually passed to the callback: In the
5235 SSL_CTX_set_cert_verify_callback() prototype, the callback
5236 declaration has been changed from
5239 int (*cb)(X509_STORE_CTX *,void *);
5240 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5241 i=s->ctx->app_verify_callback(&ctx)
5242 has been changed into
5243 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5245 To update applications using SSL_CTX_set_cert_verify_callback(),
5246 a dummy argument can be added to their callback functions.
5247 [D. K. Smetters <smetters@parc.xerox.com>]
5249 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5250 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5252 *) Add and OPENSSL_LOAD_CONF define which will cause
5253 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5254 This allows older applications to transparently support certain
5255 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5256 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5257 load the config file and OPENSSL_add_all_algorithms_conf() which will
5258 always load it have also been added.
5261 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5262 Adjust NIDs and EVP layer.
5263 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5265 *) Config modules support in openssl utility.
5267 Most commands now load modules from the config file,
5268 though in a few (such as version) this isn't done
5269 because it couldn't be used for anything.
5271 In the case of ca and req the config file used is
5272 the same as the utility itself: that is the -config
5273 command line option can be used to specify an
5277 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5278 use "openssl_conf" if filename is NULL use default openssl config file.
5281 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5282 config section name. Add a new flag to tolerate a missing config file
5283 and move code to CONF_modules_load_file().
5286 *) Support for crypto accelerator cards from Accelerated Encryption
5287 Processing, www.aep.ie. (Use engine 'aep')
5288 The support was copied from 0.9.6c [engine] and adapted/corrected
5289 to work with the new engine framework.
5290 [AEP Inc. and Richard Levitte]
5292 *) Support for SureWare crypto accelerator cards from Baltimore
5293 Technologies. (Use engine 'sureware')
5294 The support was copied from 0.9.6c [engine] and adapted
5295 to work with the new engine framework.
5298 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5299 make the newer ENGINE framework commands for the CHIL engine work.
5300 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5302 *) Make it possible to produce shared libraries on ReliantUNIX.
5303 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5305 *) Add the configuration target debug-linux-ppro.
5306 Make 'openssl rsa' use the general key loading routines
5307 implemented in apps.c, and make those routines able to
5308 handle the key format FORMAT_NETSCAPE and the variant
5310 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5312 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5313 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5315 *) Add -keyform to rsautl, and document -engine.
5316 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5318 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5319 BIO_R_NO_SUCH_FILE error code rather than the generic
5320 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5323 *) Add new functions
5325 ERR_peek_last_error_line
5326 ERR_peek_last_error_line_data.
5327 These are similar to
5330 ERR_peek_error_line_data,
5331 but report on the latest error recorded rather than the first one
5332 still in the error queue.
5333 [Ben Laurie, Bodo Moeller]
5335 *) default_algorithms option in ENGINE config module. This allows things
5337 default_algorithms = ALL
5338 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5341 *) Preliminary ENGINE config module.
5344 *) New experimental application configuration code.
5347 *) Change the AES code to follow the same name structure as all other
5348 symmetric ciphers, and behave the same way. Move everything to
5349 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5350 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5352 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5353 [Ben Laurie and Theo de Raadt]
5355 *) Add option to output public keys in req command.
5356 [Massimiliano Pala madwolf@openca.org]
5358 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5359 (up to about 10% better than before for P-192 and P-224).
5362 *) New functions/macros
5364 SSL_CTX_set_msg_callback(ctx, cb)
5365 SSL_CTX_set_msg_callback_arg(ctx, arg)
5366 SSL_set_msg_callback(ssl, cb)
5367 SSL_set_msg_callback_arg(ssl, arg)
5369 to request calling a callback function
5371 void cb(int write_p, int version, int content_type,
5372 const void *buf, size_t len, SSL *ssl, void *arg)
5374 whenever a protocol message has been completely received
5375 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5376 protocol version according to which the SSL library interprets
5377 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5378 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5379 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5380 specification (change_cipher_spec(20), alert(21), handshake(22)).
5381 'buf' and 'len' point to the actual message, 'ssl' to the
5382 SSL object, and 'arg' is the application-defined value set by
5383 SSL[_CTX]_set_msg_callback_arg().
5385 'openssl s_client' and 'openssl s_server' have new '-msg' options
5386 to enable a callback that displays all protocol messages.
5389 *) Change the shared library support so shared libraries are built as
5390 soon as the corresponding static library is finished, and thereby get
5391 openssl and the test programs linked against the shared library.
5392 This still only happens when the keyword "shard" has been given to
5393 the configuration scripts.
5395 NOTE: shared library support is still an experimental thing, and
5396 backward binary compatibility is still not guaranteed.
5397 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5399 *) Add support for Subject Information Access extension.
5400 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5402 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5403 additional bytes when new memory had to be allocated, not just
5404 when reusing an existing buffer.
5407 *) New command line and configuration option 'utf8' for the req command.
5408 This allows field values to be specified as UTF8 strings.
5411 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5412 runs for the former and machine-readable output for the latter.
5415 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5416 of the e-mail address in the DN (i.e., it will go into a certificate
5417 extension only). The new configuration file option 'email_in_dn = no'
5418 has the same effect.
5419 [Massimiliano Pala madwolf@openca.org]
5421 *) Change all functions with names starting with des_ to be starting
5422 with DES_ instead. Add wrappers that are compatible with libdes,
5423 but are named _ossl_old_des_*. Finally, add macros that map the
5424 des_* symbols to the corresponding _ossl_old_des_* if libdes
5425 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5426 desired, the des_* symbols will be mapped to DES_*, with one
5429 Since we provide two compatibility mappings, the user needs to
5430 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5431 compatibility is desired. The default (i.e., when that macro
5432 isn't defined) is OpenSSL 0.9.6c compatibility.
5434 There are also macros that enable and disable the support of old
5435 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5436 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5437 are defined, the default will apply: to support the old des routines.
5439 In either case, one must include openssl/des.h to get the correct
5440 definitions. Do not try to just include openssl/des_old.h, that
5443 NOTE: This is a major break of an old API into a new one. Software
5444 authors are encouraged to switch to the DES_ style functions. Some
5445 time in the future, des_old.h and the libdes compatibility functions
5446 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5447 default), and then completely removed.
5450 *) Test for certificates which contain unsupported critical extensions.
5451 If such a certificate is found during a verify operation it is
5452 rejected by default: this behaviour can be overridden by either
5453 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5454 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5455 X509_supported_extension() has also been added which returns 1 if a
5456 particular extension is supported.
5459 *) Modify the behaviour of EVP cipher functions in similar way to digests
5460 to retain compatibility with existing code.
5463 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5464 compatibility with existing code. In particular the 'ctx' parameter does
5465 not have to be to be initialized before the call to EVP_DigestInit() and
5466 it is tidied up after a call to EVP_DigestFinal(). New function
5467 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5468 EVP_MD_CTX_copy() changed to not require the destination to be
5469 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5470 requires the destination to be valid.
5472 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5473 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5476 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5477 so that complete 'Handshake' protocol structures are kept in memory
5478 instead of overwriting 'msg_type' and 'length' with 'body' data.
5481 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5482 [Massimo Santin via Richard Levitte]
5484 *) Major restructuring to the underlying ENGINE code. This includes
5485 reduction of linker bloat, separation of pure "ENGINE" manipulation
5486 (initialisation, etc) from functionality dealing with implementations
5487 of specific crypto iterfaces. This change also introduces integrated
5488 support for symmetric ciphers and digest implementations - so ENGINEs
5489 can now accelerate these by providing EVP_CIPHER and EVP_MD
5490 implementations of their own. This is detailed in crypto/engine/README
5491 as it couldn't be adequately described here. However, there are a few
5492 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5493 were changed in the original introduction of ENGINE code have now
5494 reverted back - the hooking from this code to ENGINE is now a good
5495 deal more passive and at run-time, operations deal directly with
5496 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5497 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5498 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5499 they were not being used by the framework as there is no concept of a
5500 BIGNUM_METHOD and they could not be generalised to the new
5501 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5502 ENGINE_cpy() has been removed as it cannot be consistently defined in
5506 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5509 *) Change mkdef.pl to sort symbols that get the same entry number,
5510 and make sure the automatically generated functions ERR_load_*
5511 become part of libeay.num as well.
5514 *) New function SSL_renegotiate_pending(). This returns true once
5515 renegotiation has been requested (either SSL_renegotiate() call
5516 or HelloRequest/ClientHello received from the peer) and becomes
5517 false once a handshake has been completed.
5518 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5519 sends a HelloRequest, but does not ensure that a handshake takes
5520 place. SSL_renegotiate_pending() is useful for checking if the
5521 client has followed the request.)
5524 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5525 By default, clients may request session resumption even during
5526 renegotiation (if session ID contexts permit); with this option,
5527 session resumption is possible only in the first handshake.
5529 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5530 more bits available for options that should not be part of
5531 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5534 *) Add some demos for certificate and certificate request creation.
5537 *) Make maximum certificate chain size accepted from the peer application
5538 settable (SSL*_get/set_max_cert_list()), as proposed by
5539 "Douglas E. Engert" <deengert@anl.gov>.
5542 *) Add support for shared libraries for Unixware-7
5543 (Boyd Lynn Gerber <gerberb@zenez.com>).
5546 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5547 be done prior to destruction. Use this to unload error strings from
5548 ENGINEs that load their own error strings. NB: This adds two new API
5549 functions to "get" and "set" this destroy handler in an ENGINE.
5552 *) Alter all existing ENGINE implementations (except "openssl" and
5553 "openbsd") to dynamically instantiate their own error strings. This
5554 makes them more flexible to be built both as statically-linked ENGINEs
5555 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5556 Also, add stub code to each that makes building them as self-contained
5557 shared-libraries easier (see README.ENGINE).
5560 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5561 implementations into applications that are completely implemented in
5562 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5563 commands that can be used to configure what shared-library to load and
5564 to control aspects of the way it is handled. Also, made an update to
5565 the README.ENGINE file that brings its information up-to-date and
5566 provides some information and instructions on the "dynamic" ENGINE
5567 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5570 *) Make it possible to unload ranges of ERR strings with a new
5571 "ERR_unload_strings" function.
5574 *) Add a copy() function to EVP_MD.
5577 *) Make EVP_MD routines take a context pointer instead of just the
5578 md_data void pointer.
5581 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5582 that the digest can only process a single chunk of data
5583 (typically because it is provided by a piece of
5584 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5585 is only going to provide a single chunk of data, and hence the
5586 framework needn't accumulate the data for oneshot drivers.
5589 *) As with "ERR", make it possible to replace the underlying "ex_data"
5590 functions. This change also alters the storage and management of global
5591 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5592 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5593 index counters. The API functions that use this state have been changed
5594 to take a "class_index" rather than pointers to the class's local STACK
5595 and counter, and there is now an API function to dynamically create new
5596 classes. This centralisation allows us to (a) plug a lot of the
5597 thread-safety problems that existed, and (b) makes it possible to clean
5598 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5599 such data would previously have always leaked in application code and
5600 workarounds were in place to make the memory debugging turn a blind eye
5601 to it. Application code that doesn't use this new function will still
5602 leak as before, but their memory debugging output will announce it now
5603 rather than letting it slide.
5605 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5606 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5607 has a return value to indicate success or failure.
5610 *) Make it possible to replace the underlying "ERR" functions such that the
5611 global state (2 LHASH tables and 2 locks) is only used by the "default"
5612 implementation. This change also adds two functions to "get" and "set"
5613 the implementation prior to it being automatically set the first time
5614 any other ERR function takes place. Ie. an application can call "get",
5615 pass the return value to a module it has just loaded, and that module
5616 can call its own "set" function using that value. This means the
5617 module's "ERR" operations will use (and modify) the error state in the
5618 application and not in its own statically linked copy of OpenSSL code.
5621 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5622 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5623 the operation, and provides a more encapsulated way for external code
5624 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5625 to use these functions rather than manually incrementing the counts.
5627 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5630 *) Add EVP test program.
5633 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5636 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5637 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5638 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5639 These allow a CRL to be built without having to access X509_CRL fields
5640 directly. Modify 'ca' application to use new functions.
5643 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5644 bug workarounds. Rollback attack detection is a security feature.
5645 The problem will only arise on OpenSSL servers when TLSv1 is not
5646 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5647 Software authors not wanting to support TLSv1 will have special reasons
5648 for their choice and can explicitly enable this option.
5649 [Bodo Moeller, Lutz Jaenicke]
5651 *) Rationalise EVP so it can be extended: don't include a union of
5652 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5653 (similar to those existing for EVP_CIPHER_CTX).
5658 EVP_MD_CTX_init(&md); /* new function call */
5659 EVP_DigestInit(&md, EVP_sha1());
5660 EVP_DigestUpdate(&md, in, len);
5661 EVP_DigestFinal(&md, out, NULL);
5662 EVP_MD_CTX_cleanup(&md); /* new function call */
5666 *) Make DES key schedule conform to the usual scheme, as well as
5667 correcting its structure. This means that calls to DES functions
5668 now have to pass a pointer to a des_key_schedule instead of a
5669 plain des_key_schedule (which was actually always a pointer
5672 des_key_schedule ks;
5674 des_set_key_checked(..., &ks);
5675 des_ncbc_encrypt(..., &ks, ...);
5677 (Note that a later change renames 'des_...' into 'DES_...'.)
5680 *) Initial reduction of linker bloat: the use of some functions, such as
5681 PEM causes large amounts of unused functions to be linked in due to
5682 poor organisation. For example pem_all.c contains every PEM function
5683 which has a knock on effect of linking in large amounts of (unused)
5684 ASN1 code. Grouping together similar functions and splitting unrelated
5685 functions prevents this.
5688 *) Cleanup of EVP macros.
5691 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5692 correct _ecb suffix.
5695 *) Add initial OCSP responder support to ocsp application. The
5696 revocation information is handled using the text based index
5697 use by the ca application. The responder can either handle
5698 requests generated internally, supplied in files (for example
5699 via a CGI script) or using an internal minimal server.
5702 *) Add configuration choices to get zlib compression for TLS.
5705 *) Changes to Kerberos SSL for RFC 2712 compliance:
5706 1. Implemented real KerberosWrapper, instead of just using
5707 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5708 2. Implemented optional authenticator field of KerberosWrapper.
5710 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5711 and authenticator structs; see crypto/krb5/.
5713 Generalized Kerberos calls to support multiple Kerberos libraries.
5714 [Vern Staats <staatsvr@asc.hpc.mil>,
5715 Jeffrey Altman <jaltman@columbia.edu>
5716 via Richard Levitte]
5718 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5719 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5720 values for each of the key sizes rather than having just
5721 parameters (and 'speed' generating keys each time).
5724 *) Speed up EVP routines.
5727 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5728 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5729 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5730 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5732 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5733 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5734 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5737 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5739 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5742 *) Added the OS2-EMX target.
5743 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5745 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5746 to support NCONF routines in extension code. New function CONF_set_nconf()
5747 to allow functions which take an NCONF to also handle the old LHASH
5748 structure: this means that the old CONF compatible routines can be
5749 retained (in particular wrt extensions) without having to duplicate the
5750 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5753 *) Enhance the general user interface with mechanisms for inner control
5754 and with possibilities to have yes/no kind of prompts.
5757 *) Change all calls to low level digest routines in the library and
5758 applications to use EVP. Add missing calls to HMAC_cleanup() and
5759 don't assume HMAC_CTX can be copied using memcpy().
5760 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5762 *) Add the possibility to control engines through control names but with
5763 arbitrary arguments instead of just a string.
5764 Change the key loaders to take a UI_METHOD instead of a callback
5765 function pointer. NOTE: this breaks binary compatibility with earlier
5766 versions of OpenSSL [engine].
5767 Adapt the nCipher code for these new conditions and add a card insertion
5771 *) Enhance the general user interface with mechanisms to better support
5772 dialog box interfaces, application-defined prompts, the possibility
5773 to use defaults (for example default passwords from somewhere else)
5774 and interrupts/cancellations.
5777 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5778 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5781 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5782 tidy up some unnecessarily weird code in 'sk_new()').
5783 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5785 *) Change the key loading routines for ENGINEs to use the same kind
5786 callback (pem_password_cb) as all other routines that need this
5790 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5791 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5792 than this minimum value is recommended.
5795 *) New random seeder for OpenVMS, using the system process statistics
5796 that are easily reachable.
5799 *) Windows apparently can't transparently handle global
5800 variables defined in DLLs. Initialisations such as:
5802 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5804 wont compile. This is used by the any applications that need to
5805 declare their own ASN1 modules. This was fixed by adding the option
5806 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5807 needed for static libraries under Win32.
5810 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5811 setting of purpose and trust fields. New X509_STORE trust and
5812 purpose functions and tidy up setting in other SSL functions.
5815 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5816 structure. These are inherited by X509_STORE_CTX when it is
5817 initialised. This allows various defaults to be set in the
5818 X509_STORE structure (such as flags for CRL checking and custom
5819 purpose or trust settings) for functions which only use X509_STORE_CTX
5820 internally such as S/MIME.
5822 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5823 trust settings if they are not set in X509_STORE. This allows X509_STORE
5824 purposes and trust (in S/MIME for example) to override any set by default.
5826 Add command line options for CRL checking to smime, s_client and s_server
5830 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5831 are set then the CRL is looked up in the X509_STORE structure and
5832 its validity and signature checked, then if the certificate is found
5833 in the CRL the verify fails with a revoked error.
5835 Various new CRL related callbacks added to X509_STORE_CTX structure.
5837 Command line options added to 'verify' application to support this.
5839 This needs some additional work, such as being able to handle multiple
5840 CRLs with different times, extension based lookup (rather than just
5841 by subject name) and ultimately more complete V2 CRL extension
5845 *) Add a general user interface API (crypto/ui/). This is designed
5846 to replace things like des_read_password and friends (backward
5847 compatibility functions using this new API are provided).
5848 The purpose is to remove prompting functions from the DES code
5849 section as well as provide for prompting through dialog boxes in
5850 a window system and the like.
5853 *) Add "ex_data" support to ENGINE so implementations can add state at a
5854 per-structure level rather than having to store it globally.
5857 *) Make it possible for ENGINE structures to be copied when retrieved by
5858 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5859 This causes the "original" ENGINE structure to act like a template,
5860 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5861 operational state can be localised to each ENGINE structure, despite the
5862 fact they all share the same "methods". New ENGINE structures returned in
5863 this case have no functional references and the return value is the single
5864 structural reference. This matches the single structural reference returned
5865 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5869 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5870 needs to match any other type at all we need to manually clear the
5874 *) Changes to the "openssl engine" utility to include;
5875 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5876 about an ENGINE's available control commands.
5877 - executing control commands from command line arguments using the
5878 '-pre' and '-post' switches. '-post' is only used if '-t' is
5879 specified and the ENGINE is successfully initialised. The syntax for
5880 the individual commands are colon-separated, for example;
5881 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5884 *) New dynamic control command support for ENGINEs. ENGINEs can now
5885 declare their own commands (numbers), names (strings), descriptions,
5886 and input types for run-time discovery by calling applications. A
5887 subset of these commands are implicitly classed as "executable"
5888 depending on their input type, and only these can be invoked through
5889 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5890 can be based on user input, config files, etc). The distinction is
5891 that "executable" commands cannot return anything other than a boolean
5892 result and can only support numeric or string input, whereas some
5893 discoverable commands may only be for direct use through
5894 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5895 pointers, or other custom uses. The "executable" commands are to
5896 support parameterisations of ENGINE behaviour that can be
5897 unambiguously defined by ENGINEs and used consistently across any
5898 OpenSSL-based application. Commands have been added to all the
5899 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5900 control over shared-library paths without source code alterations.
5903 *) Changed all ENGINE implementations to dynamically allocate their
5904 ENGINEs rather than declaring them statically. Apart from this being
5905 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5906 this also allows the implementations to compile without using the
5907 internal engine_int.h header.
5910 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5911 'const' value. Any code that should be able to modify a RAND_METHOD
5912 should already have non-const pointers to it (ie. they should only
5913 modify their own ones).
5916 *) Made a variety of little tweaks to the ENGINE code.
5917 - "atalla" and "ubsec" string definitions were moved from header files
5918 to C code. "nuron" string definitions were placed in variables
5919 rather than hard-coded - allowing parameterisation of these values
5920 later on via ctrl() commands.
5921 - Removed unused "#if 0"'d code.
5922 - Fixed engine list iteration code so it uses ENGINE_free() to release
5923 structural references.
5924 - Constified the RAND_METHOD element of ENGINE structures.
5925 - Constified various get/set functions as appropriate and added
5926 missing functions (including a catch-all ENGINE_cpy that duplicates
5927 all ENGINE values onto a new ENGINE except reference counts/state).
5928 - Removed NULL parameter checks in get/set functions. Setting a method
5929 or function to NULL is a way of cancelling out a previously set
5930 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5931 and doesn't justify the extra error symbols and code.
5932 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5933 flags from engine_int.h to engine.h.
5934 - Changed prototypes for ENGINE handler functions (init(), finish(),
5935 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5938 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5939 to the algorithm using long division. The binary algorithm can be
5940 used only if the modulus is odd. On 32-bit systems, it is faster
5941 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5942 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5943 up to 450 bits. In 64-bit environments, the binary algorithm
5944 appears to be advantageous for much longer moduli; here we use it
5945 for moduli up to 2048 bits.
5948 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5949 could not support the combine flag in choice fields.
5952 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5953 extensions from a certificate request to the certificate.
5956 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5957 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5958 file: this allows the display of the certificate about to be
5959 signed to be customised, to allow certain fields to be included
5960 or excluded and extension details. The old system didn't display
5961 multicharacter strings properly, omitted fields not in the policy
5962 and couldn't display additional details such as extensions.
5965 *) Function EC_POINTs_mul for multiple scalar multiplication
5966 of an arbitrary number of elliptic curve points
5967 \sum scalars[i]*points[i],
5968 optionally including the generator defined for the EC_GROUP:
5969 scalar*generator + \sum scalars[i]*points[i].
5971 EC_POINT_mul is a simple wrapper function for the typical case
5972 that the point list has just one item (besides the optional
5976 *) First EC_METHODs for curves over GF(p):
5978 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5979 operations and provides various method functions that can also
5980 operate with faster implementations of modular arithmetic.
5982 EC_GFp_mont_method() reuses most functions that are part of
5983 EC_GFp_simple_method, but uses Montgomery arithmetic.
5985 [Bodo Moeller; point addition and point doubling
5986 implementation directly derived from source code provided by
5987 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5989 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5990 crypto/ec/ec_lib.c):
5992 Curves are EC_GROUP objects (with an optional group generator)
5993 based on EC_METHODs that are built into the library.
5995 Points are EC_POINT objects based on EC_GROUP objects.
5997 Most of the framework would be able to handle curves over arbitrary
5998 finite fields, but as there are no obvious types for fields other
5999 than GF(p), some functions are limited to that for now.
6002 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
6003 that the file contains a complete HTTP response.
6006 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
6007 change the def and num file printf format specifier from "%-40sXXX"
6008 to "%-39s XXX". The latter will always guarantee a space after the
6009 field while the former will cause them to run together if the field
6010 is 40 of more characters long.
6013 *) Constify the cipher and digest 'method' functions and structures
6014 and modify related functions to take constant EVP_MD and EVP_CIPHER
6018 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
6019 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
6022 *) Modify EVP_Digest*() routines so they now return values. Although the
6023 internal software routines can never fail additional hardware versions
6027 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
6029 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
6030 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
6033 ERR_R_NESTED_ASN1_ERROR
6035 ERR_R_MISSING_ASN1_EOS
6036 were 4 .. 9, conflicting with
6037 ERR_LIB_RSA (= ERR_R_RSA_LIB)
6039 ERR_LIB_PEM (= ERR_R_PEM_LIB).
6040 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
6042 Add new error code 'ERR_R_INTERNAL_ERROR'.
6045 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
6049 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
6050 sets the subject name for a new request or supersedes the
6051 subject name in a given request. Formats that can be parsed are
6052 'CN=Some Name, OU=myOU, C=IT'
6054 'CN=Some Name/OU=myOU/C=IT'.
6056 Add options '-batch' and '-verbose' to 'openssl req'.
6057 [Massimiliano Pala <madwolf@hackmasters.net>]
6059 *) Introduce the possibility to access global variables through
6060 functions on platform were that's the best way to handle exporting
6061 global variables in shared libraries. To enable this functionality,
6062 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
6063 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
6064 is normally done by Configure or something similar).
6066 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
6067 in the source file (foo.c) like this:
6069 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
6070 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
6072 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
6073 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
6075 OPENSSL_DECLARE_GLOBAL(int,foo);
6076 #define foo OPENSSL_GLOBAL_REF(foo)
6077 OPENSSL_DECLARE_GLOBAL(double,bar);
6078 #define bar OPENSSL_GLOBAL_REF(bar)
6080 The #defines are very important, and therefore so is including the
6081 header file everywhere where the defined globals are used.
6083 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
6084 of ASN.1 items, but that structure is a bit different.
6086 The largest change is in util/mkdef.pl which has been enhanced with
6087 better and easier to understand logic to choose which symbols should
6088 go into the Windows .def files as well as a number of fixes and code
6089 cleanup (among others, algorithm keywords are now sorted
6090 lexicographically to avoid constant rewrites).
6093 *) In BN_div() keep a copy of the sign of 'num' before writing the
6094 result to 'rm' because if rm==num the value will be overwritten
6095 and produce the wrong result if 'num' is negative: this caused
6096 problems with BN_mod() and BN_nnmod().
6099 *) Function OCSP_request_verify(). This checks the signature on an
6100 OCSP request and verifies the signer certificate. The signer
6101 certificate is just checked for a generic purpose and OCSP request
6105 *) Add OCSP_check_validity() function to check the validity of OCSP
6106 responses. OCSP responses are prepared in real time and may only
6107 be a few seconds old. Simply checking that the current time lies
6108 between thisUpdate and nextUpdate max reject otherwise valid responses
6109 caused by either OCSP responder or client clock inaccuracy. Instead
6110 we allow thisUpdate and nextUpdate to fall within a certain period of
6111 the current time. The age of the response can also optionally be
6112 checked. Two new options -validity_period and -status_age added to
6116 *) If signature or public key algorithm is unrecognized print out its
6117 OID rather that just UNKNOWN.
6120 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
6121 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
6122 ID to be generated from the issuer certificate alone which can then be
6123 passed to OCSP_id_issuer_cmp().
6126 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
6127 ASN1 modules to export functions returning ASN1_ITEM pointers
6128 instead of the ASN1_ITEM structures themselves. This adds several
6129 new macros which allow the underlying ASN1 function/structure to
6130 be accessed transparently. As a result code should not use ASN1_ITEM
6131 references directly (such as &X509_it) but instead use the relevant
6132 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6133 use of the new ASN1 code on platforms where exporting structures
6134 is problematical (for example in shared libraries) but exporting
6135 functions returning pointers to structures is not.
6138 *) Add support for overriding the generation of SSL/TLS session IDs.
6139 These callbacks can be registered either in an SSL_CTX or per SSL.
6140 The purpose of this is to allow applications to control, if they wish,
6141 the arbitrary values chosen for use as session IDs, particularly as it
6142 can be useful for session caching in multiple-server environments. A
6143 command-line switch for testing this (and any client code that wishes
6144 to use such a feature) has been added to "s_server".
6145 [Geoff Thorpe, Lutz Jaenicke]
6147 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6148 of the form '#if defined(...) || defined(...) || ...' and
6149 '#if !defined(...) && !defined(...) && ...'. This also avoids
6150 the growing number of special cases it was previously handling.
6153 *) Make all configuration macros available for application by making
6154 sure they are available in opensslconf.h, by giving them names starting
6155 with "OPENSSL_" to avoid conflicts with other packages and by making
6156 sure e_os2.h will cover all platform-specific cases together with
6158 Additionally, it is now possible to define configuration/platform-
6159 specific names (called "system identities"). In the C code, these
6160 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6161 macro with the name beginning with "OPENSSL_SYS_", which is determined
6162 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6166 *) New option -set_serial to 'req' and 'x509' this allows the serial
6167 number to use to be specified on the command line. Previously self
6168 signed certificates were hard coded with serial number 0 and the
6169 CA options of 'x509' had to use a serial number in a file which was
6173 *) New options to 'ca' utility to support V2 CRL entry extensions.
6174 Currently CRL reason, invalidity date and hold instruction are
6175 supported. Add new CRL extensions to V3 code and some new objects.
6178 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6179 disable standard block padding (aka PKCS#5 padding) in the EVP
6180 API, which was previously mandatory. This means that the data is
6181 not padded in any way and so the total length much be a multiple
6182 of the block size, otherwise an error occurs.
6185 *) Initial (incomplete) OCSP SSL support.
6188 *) New function OCSP_parse_url(). This splits up a URL into its host,
6189 port and path components: primarily to parse OCSP URLs. New -url
6190 option to ocsp utility.
6193 *) New nonce behavior. The return value of OCSP_check_nonce() now
6194 reflects the various checks performed. Applications can decide
6195 whether to tolerate certain situations such as an absent nonce
6196 in a response when one was present in a request: the ocsp application
6197 just prints out a warning. New function OCSP_add1_basic_nonce()
6198 this is to allow responders to include a nonce in a response even if
6199 the request is nonce-less.
6202 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6203 skipped when using openssl x509 multiple times on a single input file,
6204 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6207 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6208 set string type: to handle setting ASN1_TIME structures. Fix ca
6209 utility to correctly initialize revocation date of CRLs.
6212 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6213 the clients preferred ciphersuites and rather use its own preferences.
6214 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6215 Internet Explorer by ensuring unchanged hash method during stepup.
6216 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6219 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6220 to aes and add a new 'exist' option to print out symbols that don't
6224 *) Additional options to ocsp utility to allow flags to be set and
6225 additional certificates supplied.
6228 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6229 OCSP client a number of certificate to only verify the response
6233 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6234 handle the new API. Currently only ECB, CBC modes supported. Add new
6237 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6238 Encryption Standard (AES) Ciphersuites for Transport Layer
6239 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6240 not enabled by default and were not part of the "ALL" ciphersuite
6241 alias because they were not yet official; they could be
6242 explicitly requested by specifying the "AESdraft" ciphersuite
6243 group alias. In the final release of OpenSSL 0.9.7, the group
6244 alias is called "AES" and is part of "ALL".)
6245 [Ben Laurie, Steve Henson, Bodo Moeller]
6247 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6248 request to response.
6251 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6252 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6253 extract information from a certificate request. OCSP_response_create()
6254 creates a response and optionally adds a basic response structure.
6255 OCSP_basic_add1_status() adds a complete single response to a basic
6256 response and returns the OCSP_SINGLERESP structure just added (to allow
6257 extensions to be included for example). OCSP_basic_add1_cert() adds a
6258 certificate to a basic response and OCSP_basic_sign() signs a basic
6259 response with various flags. New helper functions ASN1_TIME_check()
6260 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6261 (converts ASN1_TIME to GeneralizedTime).
6264 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6265 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6266 structure from a certificate. X509_pubkey_digest() digests the public_key
6267 contents: this is used in various key identifiers.
6270 *) Make sk_sort() tolerate a NULL argument.
6271 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6273 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6274 passed by the function are trusted implicitly. If any of them signed the
6275 response then it is assumed to be valid and is not verified.
6278 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6279 to data. This was previously part of the PKCS7 ASN1 code. This
6280 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6281 [Steve Henson, reported by Kenneth R. Robinette
6282 <support@securenetterm.com>]
6284 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6285 routines: without these tracing memory leaks is very painful.
6286 Fix leaks in PKCS12 and PKCS7 routines.
6289 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6290 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6291 effectively meant GeneralizedTime would never be used. Now it
6292 is initialised to -1 but X509_time_adj() now has to check the value
6293 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6294 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6295 [Steve Henson, reported by Kenneth R. Robinette
6296 <support@securenetterm.com>]
6298 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6299 result in a zero length in the ASN1_INTEGER structure which was
6300 not consistent with the structure when d2i_ASN1_INTEGER() was used
6301 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6302 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6303 where it did not print out a minus for negative ASN1_INTEGER.
6306 *) Add summary printout to ocsp utility. The various functions which
6307 convert status values to strings have been renamed to:
6308 OCSP_response_status_str(), OCSP_cert_status_str() and
6309 OCSP_crl_reason_str() and are no longer static. New options
6310 to verify nonce values and to disable verification. OCSP response
6311 printout format cleaned up.
6314 *) Add additional OCSP certificate checks. These are those specified
6315 in RFC2560. This consists of two separate checks: the CA of the
6316 certificate being checked must either be the OCSP signer certificate
6317 or the issuer of the OCSP signer certificate. In the latter case the
6318 OCSP signer certificate must contain the OCSP signing extended key
6319 usage. This check is performed by attempting to match the OCSP
6320 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6321 in the OCSP_CERTID structures of the response.
6324 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6325 and related routines. This uses the standard OpenSSL certificate
6326 verify routines to perform initial checks (just CA validity) and
6327 to obtain the certificate chain. Then additional checks will be
6328 performed on the chain. Currently the root CA is checked to see
6329 if it is explicitly trusted for OCSP signing. This is used to set
6330 a root CA as a global signing root: that is any certificate that
6331 chains to that CA is an acceptable OCSP signing certificate.
6334 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6335 extensions from a separate configuration file.
6336 As when reading extensions from the main configuration file,
6337 the '-extensions ...' option may be used for specifying the
6339 [Massimiliano Pala <madwolf@comune.modena.it>]
6341 *) New OCSP utility. Allows OCSP requests to be generated or
6342 read. The request can be sent to a responder and the output
6343 parsed, outputed or printed in text form. Not complete yet:
6344 still needs to check the OCSP response validity.
6347 *) New subcommands for 'openssl ca':
6348 'openssl ca -status <serial>' prints the status of the cert with
6349 the given serial number (according to the index file).
6350 'openssl ca -updatedb' updates the expiry status of certificates
6352 [Massimiliano Pala <madwolf@comune.modena.it>]
6354 *) New '-newreq-nodes' command option to CA.pl. This is like
6355 '-newreq', but calls 'openssl req' with the '-nodes' option
6356 so that the resulting key is not encrypted.
6357 [Damien Miller <djm@mindrot.org>]
6359 *) New configuration for the GNU Hurd.
6360 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6362 *) Initial code to implement OCSP basic response verify. This
6363 is currently incomplete. Currently just finds the signer's
6364 certificate and verifies the signature on the response.
6367 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6368 value of OPENSSLDIR. This is available via the new '-d' option
6369 to 'openssl version', and is also included in 'openssl version -a'.
6372 *) Allowing defining memory allocation callbacks that will be given
6373 file name and line number information in additional arguments
6374 (a const char* and an int). The basic functionality remains, as
6375 well as the original possibility to just replace malloc(),
6376 realloc() and free() by functions that do not know about these
6377 additional arguments. To register and find out the current
6378 settings for extended allocation functions, the following
6379 functions are provided:
6381 CRYPTO_set_mem_ex_functions
6382 CRYPTO_set_locked_mem_ex_functions
6383 CRYPTO_get_mem_ex_functions
6384 CRYPTO_get_locked_mem_ex_functions
6386 These work the same way as CRYPTO_set_mem_functions and friends.
6387 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6388 extended allocation function is enabled.
6389 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6390 a conventional allocation function is enabled.
6391 [Richard Levitte, Bodo Moeller]
6393 *) Finish off removing the remaining LHASH function pointer casts.
6394 There should no longer be any prototype-casting required when using
6395 the LHASH abstraction, and any casts that remain are "bugs". See
6396 the callback types and macros at the head of lhash.h for details
6397 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6400 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6401 If /dev/[u]random devices are not available or do not return enough
6402 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6404 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6405 /etc/entropy will be queried once each in this sequence, quering stops
6406 when enough entropy was collected without querying more sockets.
6409 *) Change the Unix RAND_poll() variant to be able to poll several
6410 random devices, as specified by DEVRANDOM, until a sufficient amount
6411 of data has been collected. We spend at most 10 ms on each file
6412 (select timeout) and read in non-blocking mode. DEVRANDOM now
6413 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6414 (previously it was just the string "/dev/urandom"), so on typical
6415 platforms the 10 ms delay will never occur.
6416 Also separate out the Unix variant to its own file, rand_unix.c.
6417 For VMS, there's a currently-empty rand_vms.c.
6420 *) Move OCSP client related routines to ocsp_cl.c. These
6421 provide utility functions which an application needing
6422 to issue a request to an OCSP responder and analyse the
6423 response will typically need: as opposed to those which an
6424 OCSP responder itself would need which will be added later.
6426 OCSP_request_sign() signs an OCSP request with an API similar
6427 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6428 response. OCSP_response_get1_basic() extracts basic response
6429 from response. OCSP_resp_find_status(): finds and extracts status
6430 information from an OCSP_CERTID structure (which will be created
6431 when the request structure is built). These are built from lower
6432 level functions which work on OCSP_SINGLERESP structures but
6433 wont normally be used unless the application wishes to examine
6434 extensions in the OCSP response for example.
6436 Replace nonce routines with a pair of functions.
6437 OCSP_request_add1_nonce() adds a nonce value and optionally
6438 generates a random value. OCSP_check_nonce() checks the
6439 validity of the nonce in an OCSP response.
6442 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6443 This doesn't copy the supplied OCSP_CERTID and avoids the
6444 need to free up the newly created id. Change return type
6445 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6446 This can then be used to add extensions to the request.
6447 Deleted OCSP_request_new(), since most of its functionality
6448 is now in OCSP_REQUEST_new() (and the case insensitive name
6449 clash) apart from the ability to set the request name which
6450 will be added elsewhere.
6453 *) Update OCSP API. Remove obsolete extensions argument from
6454 various functions. Extensions are now handled using the new
6455 OCSP extension code. New simple OCSP HTTP function which
6456 can be used to send requests and parse the response.
6459 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6460 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6461 uses the special reorder version of SET OF to sort the attributes
6462 and reorder them to match the encoded order. This resolves a long
6463 standing problem: a verify on a PKCS7 structure just after signing
6464 it used to fail because the attribute order did not match the
6465 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6466 it uses the received order. This is necessary to tolerate some broken
6467 software that does not order SET OF. This is handled by encoding
6468 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6469 to produce the required SET OF.
6472 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6473 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6474 files to get correct declarations of the ASN.1 item variables.
6477 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6478 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6479 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6480 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6481 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6482 ASN1_ITEM and no wrapper functions.
6485 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6486 replace the old function pointer based I/O routines. Change most of
6487 the *_d2i_bio() and *_d2i_fp() functions to use these.
6490 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6491 lines, recognice more "algorithms" that can be deselected, and make
6492 it complain about algorithm deselection that isn't recognised.
6495 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6496 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6497 to use new functions. Add NO_ASN1_OLD which can be set to remove
6498 some old style ASN1 functions: this can be used to determine if old
6499 code will still work when these eventually go away.
6502 *) New extension functions for OCSP structures, these follow the
6503 same conventions as certificates and CRLs.
6506 *) New function X509V3_add1_i2d(). This automatically encodes and
6507 adds an extension. Its behaviour can be customised with various
6508 flags to append, replace or delete. Various wrappers added for
6509 certifcates and CRLs.
6512 *) Fix to avoid calling the underlying ASN1 print routine when
6513 an extension cannot be parsed. Correct a typo in the
6514 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6517 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6518 entries for variables.
6521 *) Add functionality to apps/openssl.c for detecting locking
6522 problems: As the program is single-threaded, all we have
6523 to do is register a locking callback using an array for
6524 storing which locks are currently held by the program.
6527 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6528 SSL_get_ex_data_X509_STORE_idx(), which is used in
6529 ssl_verify_cert_chain() and thus can be called at any time
6530 during TLS/SSL handshakes so that thread-safety is essential.
6531 Unfortunately, the ex_data design is not at all suited
6532 for multi-threaded use, so it probably should be abolished.
6535 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6536 [Broadcom, tweaked and integrated by Geoff Thorpe]
6538 *) Move common extension printing code to new function
6539 X509V3_print_extensions(). Reorganise OCSP print routines and
6540 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6543 *) New function X509_signature_print() to remove duplication in some
6547 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6548 set (this was treated exactly the same as SET OF previously). This
6549 is used to reorder the STACK representing the structure to match the
6550 encoding. This will be used to get round a problem where a PKCS7
6551 structure which was signed could not be verified because the STACK
6552 order did not reflect the encoded order.
6555 *) Reimplement the OCSP ASN1 module using the new code.
6558 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6559 for its ASN1 operations. The old style function pointers still exist
6560 for now but they will eventually go away.
6563 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6564 completely replaces the old ASN1 functionality with a table driven
6565 encoder and decoder which interprets an ASN1_ITEM structure describing
6566 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6567 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6568 has also been converted to the new form.
6571 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6572 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6573 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6574 for negative moduli.
6577 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6578 of not touching the result's sign bit.
6581 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6585 *) Changed the LHASH code to use prototypes for callbacks, and created
6586 macros to declare and implement thin (optionally static) functions
6587 that provide type-safety and avoid function pointer casting for the
6588 type-specific callbacks.
6591 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6593 [Veers Staats <staatsvr@asc.hpc.mil>,
6594 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6596 *) Reformat the FAQ so the different questions and answers can be divided
6597 in sections depending on the subject.
6600 *) Have the zlib compression code load ZLIB.DLL dynamically under
6604 *) New function BN_mod_sqrt for computing square roots modulo a prime
6605 (using the probabilistic Tonelli-Shanks algorithm unless
6606 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6607 be handled deterministically).
6608 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6610 *) Make BN_mod_inverse faster by explicitly handling small quotients
6611 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6612 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6615 *) New function BN_kronecker.
6618 *) Fix BN_gcd so that it works on negative inputs; the result is
6619 positive unless both parameters are zero.
6620 Previously something reasonably close to an infinite loop was
6621 possible because numbers could be growing instead of shrinking
6622 in the implementation of Euclid's algorithm.
6625 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6626 sign of the number in question.
6628 Fix BN_is_word(a,w) to work correctly for w == 0.
6630 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6631 because its test if the absolute value of 'a' equals 'w'.
6632 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6633 it exists mostly for use in the implementations of BN_is_zero(),
6634 BN_is_one(), and BN_is_word().
6637 *) New function BN_swap.
6640 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6641 the exponentiation functions are more likely to produce reasonable
6642 results on negative inputs.
6645 *) Change BN_mod_mul so that the result is always non-negative.
6646 Previously, it could be negative if one of the factors was negative;
6647 I don't think anyone really wanted that behaviour.
6650 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6651 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6652 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6653 and add new functions:
6662 BN_mod_lshift1_quick
6666 These functions always generate non-negative results.
6668 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6669 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6671 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6672 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6673 be reduced modulo m.
6674 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6677 The following entry accidentally appeared in the CHANGES file
6678 distributed with OpenSSL 0.9.7. The modifications described in
6679 it do *not* apply to OpenSSL 0.9.7.
6681 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6682 was actually never needed) and in BN_mul(). The removal in BN_mul()
6683 required a small change in bn_mul_part_recursive() and the addition
6684 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6685 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6686 bn_sub_words() and bn_add_words() except they take arrays with
6691 *) In 'openssl passwd', verify passwords read from the terminal
6692 unless the '-salt' option is used (which usually means that
6693 verification would just waste user's time since the resulting
6694 hash is going to be compared with some given password hash)
6695 or the new '-noverify' option is used.
6697 This is an incompatible change, but it does not affect
6698 non-interactive use of 'openssl passwd' (passwords on the command
6699 line, '-stdin' option, '-in ...' option) and thus should not
6703 *) Remove all references to RSAref, since there's no more need for it.
6706 *) Make DSO load along a path given through an environment variable
6707 (SHLIB_PATH) with shl_load().
6710 *) Constify the ENGINE code as a result of BIGNUM constification.
6711 Also constify the RSA code and most things related to it. In a
6712 few places, most notable in the depth of the ASN.1 code, ugly
6713 casts back to non-const were required (to be solved at a later
6717 *) Make it so the openssl application has all engines loaded by default.
6720 *) Constify the BIGNUM routines a little more.
6723 *) Add the following functions:
6725 ENGINE_load_cswift()
6727 ENGINE_load_atalla()
6729 ENGINE_load_builtin_engines()
6731 That way, an application can itself choose if external engines that
6732 are built-in in OpenSSL shall ever be used or not. The benefit is
6733 that applications won't have to be linked with libdl or other dso
6734 libraries unless it's really needed.
6736 Changed 'openssl engine' to load all engines on demand.
6737 Changed the engine header files to avoid the duplication of some
6738 declarations (they differed!).
6741 *) 'openssl engine' can now list capabilities.
6744 *) Better error reporting in 'openssl engine'.
6747 *) Never call load_dh_param(NULL) in s_server.
6750 *) Add engine application. It can currently list engines by name and
6751 identity, and test if they are actually available.
6754 *) Improve RPM specification file by forcing symbolic linking and making
6755 sure the installed documentation is also owned by root.root.
6756 [Damien Miller <djm@mindrot.org>]
6758 *) Give the OpenSSL applications more possibilities to make use of
6759 keys (public as well as private) handled by engines.
6762 *) Add OCSP code that comes from CertCo.
6765 *) Add VMS support for the Rijndael code.
6768 *) Added untested support for Nuron crypto accelerator.
6771 *) Add support for external cryptographic devices. This code was
6772 previously distributed separately as the "engine" branch.
6773 [Geoff Thorpe, Richard Levitte]
6775 *) Rework the filename-translation in the DSO code. It is now possible to
6776 have far greater control over how a "name" is turned into a filename
6777 depending on the operating environment and any oddities about the
6778 different shared library filenames on each system.
6781 *) Support threads on FreeBSD-elf in Configure.
6784 *) Fix for SHA1 assembly problem with MASM: it produces
6785 warnings about corrupt line number information when assembling
6786 with debugging information. This is caused by the overlapping
6788 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6791 NCONF_get_number() has no error checking at all. As a replacement,
6792 NCONF_get_number_e() is defined (_e for "error checking") and is
6793 promoted strongly. The old NCONF_get_number is kept around for
6794 binary backward compatibility.
6795 Make it possible for methods to load from something other than a BIO,
6796 by providing a function pointer that is given a name instead of a BIO.
6797 For example, this could be used to load configuration data from an
6801 *) Fix for non blocking accept BIOs. Added new I/O special reason
6802 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6803 with non blocking I/O was not possible because no retry code was
6804 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6808 *) Added the beginnings of Rijndael support.
6811 *) Fix for bug in DirectoryString mask setting. Add support for
6812 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6813 to allow certificate printing to more controllable, additional
6814 'certopt' option to 'x509' to allow new printing options to be
6818 *) Clean old EAY MD5 hack from e_os.h.
6821 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6823 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6824 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6825 [Joe Orton, Steve Henson]
6827 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6829 *) Fix additional bug revealed by the NISCC test suite:
6831 Stop bug triggering large recursion when presented with
6832 certain ASN.1 tags (CVE-2003-0851)
6835 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6837 *) Fix various bugs revealed by running the NISCC test suite:
6839 Stop out of bounds reads in the ASN1 code when presented with
6840 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6842 If verify callback ignores invalid public key errors don't try to check
6843 certificate signature with the NULL public key.
6847 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6848 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6852 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6853 extra data after the compression methods not only for TLS 1.0
6854 but also for SSL 3.0 (as required by the specification).
6855 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6857 *) Change X509_certificate_type() to mark the key as exported/exportable
6858 when it's 512 *bits* long, not 512 bytes.
6861 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6863 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6864 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6865 a protocol version number mismatch like a decryption error
6866 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6869 *) Turn on RSA blinding by default in the default implementation
6870 to avoid a timing attack. Applications that don't want it can call
6871 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6872 They would be ill-advised to do so in most cases.
6873 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6875 *) Change RSA blinding code so that it works when the PRNG is not
6876 seeded (in this case, the secret RSA exponent is abused as
6877 an unpredictable seed -- if it is not unpredictable, there
6878 is no point in blinding anyway). Make RSA blinding thread-safe
6879 by remembering the creator's thread ID in rsa->blinding and
6880 having all other threads use local one-time blinding factors
6881 (this requires more computation than sharing rsa->blinding, but
6882 avoids excessive locking; and if an RSA object is not shared
6883 between threads, blinding will still be very fast).
6886 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6888 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6889 via timing by performing a MAC computation even if incorrrect
6890 block cipher padding has been found. This is a countermeasure
6891 against active attacks where the attacker has to distinguish
6892 between bad padding and a MAC verification error. (CVE-2003-0078)
6894 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6895 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6896 Martin Vuagnoux (EPFL, Ilion)]
6898 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6900 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6901 memory from it's contents. This is done with a counter that will
6902 place alternating values in each byte. This can be used to solve
6903 two issues: 1) the removal of calls to memset() by highly optimizing
6904 compilers, and 2) cleansing with other values than 0, since those can
6905 be read through on certain media, for example a swap space on disk.
6908 *) Bugfix: client side session caching did not work with external caching,
6909 because the session->cipher setting was not restored when reloading
6910 from the external cache. This problem was masked, when
6911 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6912 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6915 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6916 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6917 [Zeev Lieber <zeev-l@yahoo.com>]
6919 *) Undo an undocumented change introduced in 0.9.6e which caused
6920 repeated calls to OpenSSL_add_all_ciphers() and
6921 OpenSSL_add_all_digests() to be ignored, even after calling
6925 *) Change the default configuration reader to deal with last line not
6926 being properly terminated.
6929 *) Change X509_NAME_cmp() so it applies the special rules on handling
6930 DN values that are of type PrintableString, as well as RDNs of type
6931 emailAddress where the value has the type ia5String.
6932 [stefank@valicert.com via Richard Levitte]
6934 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6935 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6936 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6937 the bitwise-OR of the two for use by the majority of applications
6938 wanting this behaviour, and update the docs. The documented
6939 behaviour and actual behaviour were inconsistent and had been
6940 changing anyway, so this is more a bug-fix than a behavioural
6942 [Geoff Thorpe, diagnosed by Nadav Har'El]
6944 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6945 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6948 *) Fix initialization code race conditions in
6949 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6950 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6951 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6952 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6953 ssl2_get_cipher_by_char(),
6954 ssl3_get_cipher_by_char().
6955 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6957 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6958 the cached sessions are flushed, as the remove_cb() might use ex_data
6959 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6960 (see [openssl.org #212]).
6961 [Geoff Thorpe, Lutz Jaenicke]
6963 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6964 length, instead of the encoding length to d2i_ASN1_OBJECT.
6967 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6969 *) [In 0.9.6g-engine release:]
6970 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6971 [Lynn Gazis <lgazis@rainbow.com>]
6973 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6975 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6976 and get fix the header length calculation.
6977 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6978 Alon Kantor <alonk@checkpoint.com> (and others),
6981 *) Use proper error handling instead of 'assertions' in buffer
6982 overflow checks added in 0.9.6e. This prevents DoS (the
6983 assertions could call abort()).
6984 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6986 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6988 *) Add various sanity checks to asn1_get_length() to reject
6989 the ASN1 length bytes if they exceed sizeof(long), will appear
6990 negative or the content length exceeds the length of the
6992 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6994 *) Fix cipher selection routines: ciphers without encryption had no flags
6995 for the cipher strength set and where therefore not handled correctly
6996 by the selection routines (PR #130).
6999 *) Fix EVP_dsa_sha macro.
7003 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
7004 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
7005 that was added in OpenSSL 0.9.6d.
7007 As the countermeasure turned out to be incompatible with some
7008 broken SSL implementations, the new option is part of SSL_OP_ALL.
7009 SSL_OP_ALL is usually employed when compatibility with weird SSL
7010 implementations is desired (e.g. '-bugs' option to 's_client' and
7011 's_server'), so the new option is automatically set in many
7015 *) Changes in security patch:
7017 Changes marked "(CHATS)" were sponsored by the Defense Advanced
7018 Research Projects Agency (DARPA) and Air Force Research Laboratory,
7019 Air Force Materiel Command, USAF, under agreement number
7022 *) Add various sanity checks to asn1_get_length() to reject
7023 the ASN1 length bytes if they exceed sizeof(long), will appear
7024 negative or the content length exceeds the length of the
7025 supplied buffer. (CVE-2002-0659)
7026 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7028 *) Assertions for various potential buffer overflows, not known to
7030 [Ben Laurie (CHATS)]
7032 *) Various temporary buffers to hold ASCII versions of integers were
7033 too small for 64 bit platforms. (CVE-2002-0655)
7034 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
7036 *) Remote buffer overflow in SSL3 protocol - an attacker could
7037 supply an oversized session ID to a client. (CVE-2002-0656)
7038 [Ben Laurie (CHATS)]
7040 *) Remote buffer overflow in SSL2 protocol - an attacker could
7041 supply an oversized client master key. (CVE-2002-0656)
7042 [Ben Laurie (CHATS)]
7044 Changes between 0.9.6c and 0.9.6d [9 May 2002]
7046 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
7047 encoded as NULL) with id-dsa-with-sha1.
7048 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
7050 *) Check various X509_...() return values in apps/req.c.
7051 [Nils Larsch <nla@trustcenter.de>]
7053 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
7054 an end-of-file condition would erronously be flagged, when the CRLF
7055 was just at the end of a processed block. The bug was discovered when
7056 processing data through a buffering memory BIO handing the data to a
7057 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
7058 <ptsekov@syntrex.com> and Nedelcho Stanev.
7061 *) Implement a countermeasure against a vulnerability recently found
7062 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
7063 before application data chunks to avoid the use of known IVs
7064 with data potentially chosen by the attacker.
7067 *) Fix length checks in ssl3_get_client_hello().
7070 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
7071 to prevent ssl3_read_internal() from incorrectly assuming that
7072 ssl3_read_bytes() found application data while handshake
7073 processing was enabled when in fact s->s3->in_read_app_data was
7074 merely automatically cleared during the initial handshake.
7075 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
7077 *) Fix object definitions for Private and Enterprise: they were not
7078 recognized in their shortname (=lowercase) representation. Extend
7079 obj_dat.pl to issue an error when using undefined keywords instead
7080 of silently ignoring the problem (Svenning Sorensen
7081 <sss@sss.dnsalias.net>).
7084 *) Fix DH_generate_parameters() so that it works for 'non-standard'
7085 generators, i.e. generators other than 2 and 5. (Previously, the
7086 code did not properly initialise the 'add' and 'rem' values to
7087 BN_generate_prime().)
7089 In the new general case, we do not insist that 'generator' is
7090 actually a primitive root: This requirement is rather pointless;
7091 a generator of the order-q subgroup is just as good, if not
7095 *) Map new X509 verification errors to alerts. Discovered and submitted by
7096 Tom Wu <tom@arcot.com>.
7099 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
7100 returning non-zero before the data has been completely received
7101 when using non-blocking I/O.
7102 [Bodo Moeller; problem pointed out by John Hughes]
7104 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
7105 [Ben Laurie, Lutz Jaenicke]
7107 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
7108 Yoram Zahavi <YoramZ@gilian.com>).
7111 *) Add information about CygWin 1.3 and on, and preserve proper
7112 configuration for the versions before that.
7113 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7115 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
7116 check whether we deal with a copy of a session and do not delete from
7117 the cache in this case. Problem reported by "Izhar Shoshani Levi"
7118 <izhar@checkpoint.com>.
7121 *) Do not store session data into the internal session cache, if it
7122 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
7123 flag is set). Proposed by Aslam <aslam@funk.com>.
7126 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
7130 *) [In 0.9.6d-engine release:]
7131 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7132 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7134 *) Add the configuration target linux-s390x.
7135 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7137 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7138 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7139 variable as an indication that a ClientHello message has been
7140 received. As the flag value will be lost between multiple
7141 invocations of ssl3_accept when using non-blocking I/O, the
7142 function may not be aware that a handshake has actually taken
7143 place, thus preventing a new session from being added to the
7146 To avoid this problem, we now set s->new_session to 2 instead of
7147 using a local variable.
7148 [Lutz Jaenicke, Bodo Moeller]
7150 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7151 if the SSL_R_LENGTH_MISMATCH error is detected.
7152 [Geoff Thorpe, Bodo Moeller]
7154 *) New 'shared_ldflag' column in Configure platform table.
7157 *) Fix EVP_CIPHER_mode macro.
7158 ["Dan S. Camper" <dan@bti.net>]
7160 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7161 type, we must throw them away by setting rr->length to 0.
7162 [D P Chang <dpc@qualys.com>]
7164 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7166 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7167 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7168 worked incorrectly for those cases where range = 10..._2 and
7169 3*range is two bits longer than range.)
7172 *) Only add signing time to PKCS7 structures if it is not already
7176 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7177 OBJ_ld_ce should be OBJ_id_ce.
7178 Also some ip-pda OIDs in crypto/objects/objects.txt were
7179 incorrect (cf. RFC 3039).
7180 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7182 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7183 returns early because it has nothing to do.
7184 [Andy Schneider <andy.schneider@bjss.co.uk>]
7186 *) [In 0.9.6c-engine release:]
7187 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7188 [Andy Schneider <andy.schneider@bjss.co.uk>]
7190 *) [In 0.9.6c-engine release:]
7191 Add support for Cryptographic Appliance's keyserver technology.
7192 (Use engine 'keyclient')
7193 [Cryptographic Appliances and Geoff Thorpe]
7195 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7196 is called via tools/c89.sh because arguments have to be
7197 rearranged (all '-L' options must appear before the first object
7199 [Richard Shapiro <rshapiro@abinitio.com>]
7201 *) [In 0.9.6c-engine release:]
7202 Add support for Broadcom crypto accelerator cards, backported
7204 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7206 *) [In 0.9.6c-engine release:]
7207 Add support for SureWare crypto accelerator cards from
7208 Baltimore Technologies. (Use engine 'sureware')
7209 [Baltimore Technologies and Mark Cox]
7211 *) [In 0.9.6c-engine release:]
7212 Add support for crypto accelerator cards from Accelerated
7213 Encryption Processing, www.aep.ie. (Use engine 'aep')
7214 [AEP Inc. and Mark Cox]
7216 *) Add a configuration entry for gcc on UnixWare.
7217 [Gary Benson <gbenson@redhat.com>]
7219 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7220 messages are stored in a single piece (fixed-length part and
7221 variable-length part combined) and fix various bugs found on the way.
7224 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7225 instead. BIO_gethostbyname() does not know what timeouts are
7226 appropriate, so entries would stay in cache even when they have
7228 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7230 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7231 faced with a pathologically small ClientHello fragment that does
7232 not contain client_version: Instead of aborting with an error,
7233 simply choose the highest available protocol version (i.e.,
7234 TLS 1.0 unless it is disabled). In practice, ClientHello
7235 messages are never sent like this, but this change gives us
7236 strictly correct behaviour at least for TLS.
7239 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7240 never resets s->method to s->ctx->method when called from within
7241 one of the SSL handshake functions.
7242 [Bodo Moeller; problem pointed out by Niko Baric]
7244 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7245 (sent using the client's version number) if client_version is
7246 smaller than the protocol version in use. Also change
7247 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7248 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7249 the client will at least see that alert.
7252 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7256 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7257 client receives HelloRequest while in a handshake.
7258 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7260 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7261 should end in 'break', not 'goto end' which circumvents various
7262 cleanups done in state SSL_ST_OK. But session related stuff
7263 must be disabled for SSL_ST_OK in the case that we just sent a
7266 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7267 before just sending a HelloRequest.
7268 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7270 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7271 reveal whether illegal block cipher padding was found or a MAC
7272 verification error occured. (Neither SSLerr() codes nor alerts
7273 are directly visible to potential attackers, but the information
7274 may leak via logfiles.)
7276 Similar changes are not required for the SSL 2.0 implementation
7277 because the number of padding bytes is sent in clear for SSL 2.0,
7278 and the extra bytes are just ignored. However ssl/s2_pkt.c
7279 failed to verify that the purported number of padding bytes is in
7283 *) Add OpenUNIX-8 support including shared libraries
7284 (Boyd Lynn Gerber <gerberb@zenez.com>).
7287 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7288 'wristwatch attack' using huge encoding parameters (cf.
7289 James H. Manger's CRYPTO 2001 paper). Note that the
7290 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7291 encoding parameters and hence was not vulnerable.
7294 *) BN_sqr() bug fix.
7295 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7297 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7298 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7299 followed by modular reduction.
7300 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7302 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7303 equivalent based on BN_pseudo_rand() instead of BN_rand().
7306 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7307 This function was broken, as the check for a new client hello message
7308 to handle SGC did not allow these large messages.
7309 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7312 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7315 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7316 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7319 *) Rework the configuration and shared library support for Tru64 Unix.
7320 The configuration part makes use of modern compiler features and
7321 still retains old compiler behavior for those that run older versions
7322 of the OS. The shared library support part includes a variant that
7323 uses the RPATH feature, and is available through the special
7324 configuration target "alpha-cc-rpath", which will never be selected
7326 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7328 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7329 with the same message size as in ssl3_get_certificate_request().
7330 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7331 messages might inadvertently be reject as too long.
7332 [Petr Lampa <lampa@fee.vutbr.cz>]
7334 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7337 *) Modified SSL library such that the verify_callback that has been set
7338 specificly for an SSL object with SSL_set_verify() is actually being
7339 used. Before the change, a verify_callback set with this function was
7340 ignored and the verify_callback() set in the SSL_CTX at the time of
7341 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7342 to allow the necessary settings.
7345 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7346 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7347 done automatically (in contradiction to the requirements of the C
7348 standard). This made problems when used from OpenSSH.
7351 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7352 dh->length and always used
7354 BN_rand_range(priv_key, dh->p).
7356 BN_rand_range() is not necessary for Diffie-Hellman, and this
7357 specific range makes Diffie-Hellman unnecessarily inefficient if
7358 dh->length (recommended exponent length) is much smaller than the
7359 length of dh->p. We could use BN_rand_range() if the order of
7360 the subgroup was stored in the DH structure, but we only have
7365 BN_rand(priv_key, l, ...)
7367 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7373 RSA_eay_public_encrypt
7374 RSA_eay_private_decrypt
7375 RSA_eay_private_encrypt (signing)
7376 RSA_eay_public_decrypt (signature verification)
7378 (default implementations for RSA_public_encrypt,
7379 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7380 always reject numbers >= n.
7383 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7384 to synchronize access to 'locking_thread'. This is necessary on
7385 systems where access to 'locking_thread' (an 'unsigned long'
7386 variable) is not atomic.
7389 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7390 *before* setting the 'crypto_lock_rand' flag. The previous code had
7391 a race condition if 0 is a valid thread ID.
7392 [Travis Vitek <vitek@roguewave.com>]
7394 *) Add support for shared libraries under Irix.
7395 [Albert Chin-A-Young <china@thewrittenword.com>]
7397 *) Add configuration option to build on Linux on both big-endian and
7399 [Ralf Baechle <ralf@uni-koblenz.de>]
7401 *) Add the possibility to create shared libraries on HP-UX.
7404 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7406 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7407 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7408 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7409 PRNG state recovery was possible based on the output of
7410 one PRNG request appropriately sized to gain knowledge on
7411 'md' followed by enough consecutive 1-byte PRNG requests
7412 to traverse all of 'state'.
7414 1. When updating 'md_local' (the current thread's copy of 'md')
7415 during PRNG output generation, hash all of the previous
7416 'md_local' value, not just the half used for PRNG output.
7418 2. Make the number of bytes from 'state' included into the hash
7419 independent from the number of PRNG bytes requested.
7421 The first measure alone would be sufficient to avoid
7422 Markku-Juhani's attack. (Actually it had never occurred
7423 to me that the half of 'md_local' used for chaining was the
7424 half from which PRNG output bytes were taken -- I had always
7425 assumed that the secret half would be used.) The second
7426 measure makes sure that additional data from 'state' is never
7427 mixed into 'md_local' in small portions; this heuristically
7428 further strengthens the PRNG.
7431 *) Fix crypto/bn/asm/mips3.s.
7434 *) When only the key is given to "enc", the IV is undefined. Print out
7435 an error message in this case.
7438 *) Handle special case when X509_NAME is empty in X509 printing routines.
7441 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7442 positive and less than q.
7445 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7446 used: it isn't thread safe and the add_lock_callback should handle
7448 [Paul Rose <Paul.Rose@bridge.com>]
7450 *) Verify that incoming data obeys the block size in
7451 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7455 [Ulf Möller, Bodo Möller]
7457 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7458 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7459 when fixing the server behaviour for backwards-compatible 'client
7460 hello' messages. (Note that the attack is impractical against
7461 SSL 3.0 and TLS 1.0 anyway because length and version checking
7462 means that the probability of guessing a valid ciphertext is
7463 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7466 Before 0.9.5, the countermeasure (hide the error by generating a
7467 random 'decryption result') did not work properly because
7468 ERR_clear_error() was missing, meaning that SSL_get_error() would
7469 detect the supposedly ignored error.
7471 Both problems are now fixed.
7474 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7475 (previously it was 1024).
7478 *) Fix for compatibility mode trust settings: ignore trust settings
7479 unless some valid trust or reject settings are present.
7482 *) Fix for blowfish EVP: its a variable length cipher.
7485 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7486 parameters in DSA public key structures and return an error in the
7487 DSA routines if parameters are absent.
7490 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7491 in the current directory if neither $RANDFILE nor $HOME was set.
7492 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7493 caused some confusion to Windows users who haven't defined $HOME.
7494 Thus RAND_file_name() is changed again: e_os.h can define a
7495 DEFAULT_HOME, which will be used if $HOME is not set.
7496 For Windows, we use "C:"; on other platforms, we still require
7497 environment variables.
7499 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7500 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7501 having multiple threads call RAND_poll() concurrently.
7504 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7505 combination of a flag and a thread ID variable.
7506 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7507 flag), *other* threads can enter ssleay_add_bytes without obeying
7508 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7509 that they do not hold after the first thread unsets add_do_not_lock).
7512 *) Change bctest again: '-x' expressions are not available in all
7516 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7518 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7519 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7521 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7522 the default extension for executables, if any. Also, make the perl
7523 scripts that use symlink() to test if it really exists and use "cp"
7524 if it doesn't. All this made OpenSSL compilable and installable in
7528 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7529 If SEQUENCE is length is indefinite just set c->slen to the total
7530 amount of data available.
7531 [Steve Henson, reported by shige@FreeBSD.org]
7532 [This change does not apply to 0.9.7.]
7534 *) Change bctest to avoid here-documents inside command substitution
7535 (workaround for FreeBSD /bin/sh bug).
7536 For compatibility with Ultrix, avoid shell functions (introduced
7537 in the bctest version that searches along $PATH).
7540 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7541 with des_encrypt() defined on some operating systems, like Solaris
7545 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7546 On the Importance of Eliminating Errors in Cryptographic
7547 Computations, J. Cryptology 14 (2001) 2, 101-119,
7548 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7551 *) MIPS assembler BIGNUM division bug fix.
7554 *) Disabled incorrect Alpha assembler code.
7557 *) Fix PKCS#7 decode routines so they correctly update the length
7558 after reading an EOC for the EXPLICIT tag.
7560 [This change does not apply to 0.9.7.]
7562 *) Fix bug in PKCS#12 key generation routines. This was triggered
7563 if a 3DES key was generated with a 0 initial byte. Include
7564 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7565 (but broken) behaviour.
7568 *) Enhance bctest to search for a working bc along $PATH and print
7570 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7572 *) Fix memory leaks in err.c: free err_data string if necessary;
7573 don't write to the wrong index in ERR_set_error_data.
7576 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7580 *) Replace rdtsc with _emit statements for VC++ version 5.
7581 [Jeremy Cooper <jeremy@baymoo.org>]
7583 *) Make it possible to reuse SSLv2 sessions.
7586 *) In copy_email() check for >= 0 as a return value for
7587 X509_NAME_get_index_by_NID() since 0 is a valid index.
7588 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7590 *) Avoid coredump with unsupported or invalid public keys by checking if
7591 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7592 PKCS7_verify() fails with non detached data.
7595 *) Don't use getenv in library functions when run as setuid/setgid.
7596 New function OPENSSL_issetugid().
7599 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7600 due to incorrect handling of multi-threading:
7602 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7604 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7606 3. Count how many times MemCheck_off() has been called so that
7607 nested use can be treated correctly. This also avoids
7608 inband-signalling in the previous code (which relied on the
7609 assumption that thread ID 0 is impossible).
7612 *) Add "-rand" option also to s_client and s_server.
7615 *) Fix CPU detection on Irix 6.x.
7616 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7617 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7619 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7622 [This change does not apply to 0.9.7.]
7624 *) Use the cached encoding of an X509_NAME structure rather than
7625 copying it. This is apparently the reason for the libsafe "errors"
7626 but the code is actually correct.
7629 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7630 Bleichenbacher's DSA attack.
7631 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7632 to be set and top=0 forces the highest bit to be set; top=-1 is new
7633 and leaves the highest bit random.
7634 [Ulf Moeller, Bodo Moeller]
7636 *) In the NCONF_...-based implementations for CONF_... queries
7637 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7638 a temporary CONF structure with the data component set to NULL
7639 (which gives segmentation faults in lh_retrieve).
7640 Instead, use NULL for the CONF pointer in CONF_get_string and
7641 CONF_get_number (which may use environment variables) and directly
7642 return NULL from CONF_get_section.
7645 *) Fix potential buffer overrun for EBCDIC.
7648 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7649 keyUsage if basicConstraints absent for a CA.
7652 *) Make SMIME_write_PKCS7() write mail header values with a format that
7653 is more generally accepted (no spaces before the semicolon), since
7654 some programs can't parse those values properly otherwise. Also make
7655 sure BIO's that break lines after each write do not create invalid
7659 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7660 macros previously used would not encode an empty SEQUENCE OF
7661 and break the signature.
7663 [This change does not apply to 0.9.7.]
7665 *) Zero the premaster secret after deriving the master secret in
7669 *) Add some EVP_add_digest_alias registrations (as found in
7670 OpenSSL_add_all_digests()) to SSL_library_init()
7671 aka OpenSSL_add_ssl_algorithms(). This provides improved
7672 compatibility with peers using X.509 certificates
7673 with unconventional AlgorithmIdentifier OIDs.
7676 *) Fix for Irix with NO_ASM.
7677 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7679 *) ./config script fixes.
7680 [Ulf Moeller, Richard Levitte]
7682 *) Fix 'openssl passwd -1'.
7685 *) Change PKCS12_key_gen_asc() so it can cope with non null
7686 terminated strings whose length is passed in the passlen
7687 parameter, for example from PEM callbacks. This was done
7688 by adding an extra length parameter to asc2uni().
7689 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7691 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7692 call failed, free the DSA structure.
7695 *) Fix to uni2asc() to cope with zero length Unicode strings.
7696 These are present in some PKCS#12 files.
7699 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7700 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7701 when writing a 32767 byte record.
7702 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7704 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7705 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7707 (RSA objects have a reference count access to which is protected
7708 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7709 so they are meant to be shared between threads.)
7710 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7711 "Reddie, Steven" <Steven.Reddie@ca.com>]
7713 *) Fix a deadlock in CRYPTO_mem_leaks().
7716 *) Use better test patterns in bntest.
7719 *) rand_win.c fix for Borland C.
7722 *) BN_rshift bugfix for n == 0.
7725 *) Add a 'bctest' script that checks for some known 'bc' bugs
7726 so that 'make test' does not abort just because 'bc' is broken.
7729 *) Store verify_result within SSL_SESSION also for client side to
7730 avoid potential security hole. (Re-used sessions on the client side
7731 always resulted in verify_result==X509_V_OK, not using the original
7732 result of the server certificate verification.)
7735 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7736 SSL3_RT_APPLICATION_DATA, return 0.
7737 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7741 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7742 releases, have been re-implemented by renaming the previous
7743 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7744 and ssl3_read_internal, respectively, and adding 'peek' parameters
7745 to them. The new ssl[23]_{read,peek} functions are calls to
7746 ssl[23]_read_internal with the 'peek' flag set appropriately.
7747 A 'peek' parameter has also been added to ssl3_read_bytes, which
7748 does the actual work for ssl3_read_internal.
7751 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7752 the method-specific "init()" handler. Also clean up ex_data after
7753 calling the method-specific "finish()" handler. Previously, this was
7754 happening the other way round.
7757 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7758 The previous value, 12, was not always sufficient for BN_mod_exp().
7761 *) Make sure that shared libraries get the internal name engine with
7762 the full version number and not just 0. This should mark the
7763 shared libraries as not backward compatible. Of course, this should
7764 be changed again when we can guarantee backward binary compatibility.
7767 *) Fix typo in get_cert_by_subject() in by_dir.c
7768 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7770 *) Rework the system to generate shared libraries:
7772 - Make note of the expected extension for the shared libraries and
7773 if there is a need for symbolic links from for example libcrypto.so.0
7774 to libcrypto.so.0.9.7. There is extended info in Configure for
7777 - Make as few rebuilds of the shared libraries as possible.
7779 - Still avoid linking the OpenSSL programs with the shared libraries.
7781 - When installing, install the shared libraries separately from the
7785 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7787 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7788 and not in SSL_clear because the latter is also used by the
7789 accept/connect functions; previously, the settings made by
7790 SSL_set_read_ahead would be lost during the handshake.
7791 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7793 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7794 Previously, it would create entries for disabled algorithms no
7798 *) Added several new manual pages for SSL_* function.
7801 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7803 *) In ssl23_get_client_hello, generate an error message when faced
7804 with an initial SSL 3.0/TLS record that is too small to contain the
7805 first two bytes of the ClientHello message, i.e. client_version.
7806 (Note that this is a pathologic case that probably has never happened
7807 in real life.) The previous approach was to use the version number
7808 from the record header as a substitute; but our protocol choice
7809 should not depend on that one because it is not authenticated
7810 by the Finished messages.
7813 *) More robust randomness gathering functions for Windows.
7814 [Jeffrey Altman <jaltman@columbia.edu>]
7816 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7817 not set then we don't setup the error code for issuer check errors
7818 to avoid possibly overwriting other errors which the callback does
7819 handle. If an application does set the flag then we assume it knows
7820 what it is doing and can handle the new informational codes
7824 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7825 a general "ANY" type, as such it should be able to decode anything
7826 including tagged types. However it didn't check the class so it would
7827 wrongly interpret tagged types in the same way as their universal
7828 counterpart and unknown types were just rejected. Changed so that the
7829 tagged and unknown types are handled in the same way as a SEQUENCE:
7830 that is the encoding is stored intact. There is also a new type
7831 "V_ASN1_OTHER" which is used when the class is not universal, in this
7832 case we have no idea what the actual type is so we just lump them all
7836 *) On VMS, stdout may very well lead to a file that is written to
7837 in a record-oriented fashion. That means that every write() will
7838 write a separate record, which will be read separately by the
7839 programs trying to read from it. This can be very confusing.
7841 The solution is to put a BIO filter in the way that will buffer
7842 text until a linefeed is reached, and then write everything a
7843 line at a time, so every record written will be an actual line,
7844 not chunks of lines and not (usually doesn't happen, but I've
7845 seen it once) several lines in one record. BIO_f_linebuffer() is
7848 Currently, it's a VMS-only method, because that's where it has
7849 been tested well enough.
7852 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7853 it can return incorrect results.
7854 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7855 but it was in 0.9.6-beta[12].)
7858 *) Disable the check for content being present when verifying detached
7859 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7860 include zero length content when signing messages.
7863 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7864 BIO_ctrl (for BIO pairs).
7867 *) Add DSO method for VMS.
7870 *) Bug fix: Montgomery multiplication could produce results with the
7874 *) Add RPM specification openssl.spec and modify it to build three
7875 packages. The default package contains applications, application
7876 documentation and run-time libraries. The devel package contains
7877 include files, static libraries and function documentation. The
7878 doc package contains the contents of the doc directory. The original
7879 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7882 *) Add a large number of documentation files for many SSL routines.
7883 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7885 *) Add a configuration entry for Sony News 4.
7886 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7888 *) Don't set the two most significant bits to one when generating a
7889 random number < q in the DSA library.
7892 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7893 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7894 the underlying transport is blocking) if a handshake took place.
7895 (The default behaviour is needed by applications such as s_client
7896 and s_server that use select() to determine when to use SSL_read;
7897 but for applications that know in advance when to expect data, it
7898 just makes things more complicated.)
7901 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7905 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7906 work better on such systems.
7907 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7909 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7910 Update PKCS12_parse() so it copies the friendlyName and the
7911 keyid to the certificates aux info.
7914 *) Fix bug in PKCS7_verify() which caused an infinite loop
7915 if there was more than one signature.
7916 [Sven Uszpelkat <su@celocom.de>]
7918 *) Major change in util/mkdef.pl to include extra information
7919 about each symbol, as well as presenting variables as well
7920 as functions. This change means that there's n more need
7921 to rebuild the .num files when some algorithms are excluded.
7924 *) Allow the verify time to be set by an application,
7925 rather than always using the current time.
7928 *) Phase 2 verify code reorganisation. The certificate
7929 verify code now looks up an issuer certificate by a
7930 number of criteria: subject name, authority key id
7931 and key usage. It also verifies self signed certificates
7932 by the same criteria. The main comparison function is
7933 X509_check_issued() which performs these checks.
7935 Lot of changes were necessary in order to support this
7936 without completely rewriting the lookup code.
7938 Authority and subject key identifier are now cached.
7940 The LHASH 'certs' is X509_STORE has now been replaced
7941 by a STACK_OF(X509_OBJECT). This is mainly because an
7942 LHASH can't store or retrieve multiple objects with
7943 the same hash value.
7945 As a result various functions (which were all internal
7946 use only) have changed to handle the new X509_STORE
7947 structure. This will break anything that messed round
7948 with X509_STORE internally.
7950 The functions X509_STORE_add_cert() now checks for an
7951 exact match, rather than just subject name.
7953 The X509_STORE API doesn't directly support the retrieval
7954 of multiple certificates matching a given criteria, however
7955 this can be worked round by performing a lookup first
7956 (which will fill the cache with candidate certificates)
7957 and then examining the cache for matches. This is probably
7958 the best we can do without throwing out X509_LOOKUP
7959 entirely (maybe later...).
7961 The X509_VERIFY_CTX structure has been enhanced considerably.
7963 All certificate lookup operations now go via a get_issuer()
7964 callback. Although this currently uses an X509_STORE it
7965 can be replaced by custom lookups. This is a simple way
7966 to bypass the X509_STORE hackery necessary to make this
7967 work and makes it possible to use more efficient techniques
7968 in future. A very simple version which uses a simple
7969 STACK for its trusted certificate store is also provided
7970 using X509_STORE_CTX_trusted_stack().
7972 The verify_cb() and verify() callbacks now have equivalents
7973 in the X509_STORE_CTX structure.
7975 X509_STORE_CTX also has a 'flags' field which can be used
7976 to customise the verify behaviour.
7979 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7980 excludes S/MIME capabilities.
7983 *) When a certificate request is read in keep a copy of the
7984 original encoding of the signed data and use it when outputing
7985 again. Signatures then use the original encoding rather than
7986 a decoded, encoded version which may cause problems if the
7987 request is improperly encoded.
7990 *) For consistency with other BIO_puts implementations, call
7991 buffer_write(b, ...) directly in buffer_puts instead of calling
7994 In BIO_puts, increment b->num_write as in BIO_write.
7995 [Peter.Sylvester@EdelWeb.fr]
7997 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7998 BN_zero, we may not return a BIGNUM with an array consisting of
8002 *) Avoid calling abort() from within the library when problems are
8003 detected, except if preprocessor symbols have been defined
8004 (such as REF_CHECK, BN_DEBUG etc.).
8007 *) New openssl application 'rsautl'. This utility can be
8008 used for low level RSA operations. DER public key
8009 BIO/fp routines also added.
8012 *) New Configure entry and patches for compiling on QNX 4.
8013 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
8015 *) A demo state-machine implementation was sponsored by
8016 Nuron (http://www.nuron.com/) and is now available in
8017 demos/state_machine.
8020 *) New options added to the 'dgst' utility for signature
8021 generation and verification.
8024 *) Unrecognized PKCS#7 content types are now handled via a
8025 catch all ASN1_TYPE structure. This allows unsupported
8026 types to be stored as a "blob" and an application can
8027 encode and decode it manually.
8030 *) Fix various signed/unsigned issues to make a_strex.c
8032 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
8034 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
8035 length if passed a buffer. ASN1_INTEGER_to_BN failed
8036 if passed a NULL BN and its argument was negative.
8037 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
8039 *) Modification to PKCS#7 encoding routines to output definite
8040 length encoding. Since currently the whole structures are in
8041 memory there's not real point in using indefinite length
8042 constructed encoding. However if OpenSSL is compiled with
8043 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
8046 *) Added BIO_vprintf() and BIO_vsnprintf().
8049 *) Added more prefixes to parse for in the the strings written
8050 through a logging bio, to cover all the levels that are available
8051 through syslog. The prefixes are now:
8053 PANIC, EMERG, EMR => LOG_EMERG
8054 ALERT, ALR => LOG_ALERT
8055 CRIT, CRI => LOG_CRIT
8056 ERROR, ERR => LOG_ERR
8057 WARNING, WARN, WAR => LOG_WARNING
8058 NOTICE, NOTE, NOT => LOG_NOTICE
8059 INFO, INF => LOG_INFO
8060 DEBUG, DBG => LOG_DEBUG
8062 and as before, if none of those prefixes are present at the
8063 beginning of the string, LOG_ERR is chosen.
8065 On Win32, the LOG_* levels are mapped according to this:
8067 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
8068 LOG_WARNING => EVENTLOG_WARNING_TYPE
8069 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
8073 *) Made it possible to reconfigure with just the configuration
8074 argument "reconf" or "reconfigure". The command line arguments
8075 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
8076 and are retrieved from there when reconfiguring.
8080 [Assar Westerlund <assar@sics.se>, Richard Levitte]
8082 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
8085 *) The obj_dat.pl script was messing up the sorting of object
8086 names. The reason was that it compared the quoted version
8087 of strings as a result "OCSP" > "OCSP Signing" because
8088 " > SPACE. Changed script to store unquoted versions of
8089 names and add quotes on output. It was also omitting some
8090 names from the lookup table if they were given a default
8091 value (that is if SN is missing it is given the same
8092 value as LN and vice versa), these are now added on the
8093 grounds that if an object has a name we should be able to
8094 look it up. Finally added warning output when duplicate
8095 short or long names are found.
8098 *) Changes needed for Tandem NSK.
8099 [Scott Uroff <scott@xypro.com>]
8101 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
8102 RSA_padding_check_SSLv23(), special padding was never detected
8103 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
8104 version rollback attacks was not effective.
8106 In s23_clnt.c, don't use special rollback-attack detection padding
8107 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
8108 client; similarly, in s23_srvr.c, don't do the rollback check if
8109 SSL 2.0 is the only protocol enabled in the server.
8112 *) Make it possible to get hexdumps of unprintable data with 'openssl
8113 asn1parse'. By implication, the functions ASN1_parse_dump() and
8114 BIO_dump_indent() are added.
8117 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
8118 these print out strings and name structures based on various
8119 flags including RFC2253 support and proper handling of
8120 multibyte characters. Added options to the 'x509' utility
8121 to allow the various flags to be set.
8124 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
8125 Also change the functions X509_cmp_current_time() and
8126 X509_gmtime_adj() work with an ASN1_TIME structure,
8127 this will enable certificates using GeneralizedTime in validity
8128 dates to be checked.
8131 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8132 negative public key encodings) on by default,
8133 NO_NEG_PUBKEY_BUG can be set to disable it.
8136 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8137 content octets. An i2c_ASN1_OBJECT is unnecessary because
8138 the encoding can be trivially obtained from the structure.
8141 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8142 not read locks (CRYPTO_r_[un]lock).
8145 *) A first attempt at creating official support for shared
8146 libraries through configuration. I've kept it so the
8147 default is static libraries only, and the OpenSSL programs
8148 are always statically linked for now, but there are
8149 preparations for dynamic linking in place.
8150 This has been tested on Linux and Tru64.
8153 *) Randomness polling function for Win9x, as described in:
8154 Peter Gutmann, Software Generation of Practically Strong
8158 *) Fix so PRNG is seeded in req if using an already existing
8162 *) New options to smime application. -inform and -outform
8163 allow alternative formats for the S/MIME message including
8164 PEM and DER. The -content option allows the content to be
8165 specified separately. This should allow things like Netscape
8166 form signing output easier to verify.
8169 *) Fix the ASN1 encoding of tags using the 'long form'.
8172 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8173 STRING types. These convert content octets to and from the
8174 underlying type. The actual tag and length octets are
8175 already assumed to have been read in and checked. These
8176 are needed because all other string types have virtually
8177 identical handling apart from the tag. By having versions
8178 of the ASN1 functions that just operate on content octets
8179 IMPLICIT tagging can be handled properly. It also allows
8180 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8181 and ASN1_INTEGER are identical apart from the tag.
8184 *) Change the handling of OID objects as follows:
8186 - New object identifiers are inserted in objects.txt, following
8187 the syntax given in objects.README.
8188 - objects.pl is used to process obj_mac.num and create a new
8190 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8193 This is currently kind of a hack, and the perl code in objects.pl
8194 isn't very elegant, but it works as I intended. The simplest way
8195 to check that it worked correctly is to look in obj_dat.h and
8196 check the array nid_objs and make sure the objects haven't moved
8197 around (this is important!). Additions are OK, as well as
8198 consistent name changes.
8201 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8204 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8205 The given file adds to whatever has already been seeded into the
8206 random pool through the RANDFILE configuration file option or
8207 environment variable, or the default random state file.
8210 *) mkstack.pl now sorts each macro group into lexical order.
8211 Previously the output order depended on the order the files
8212 appeared in the directory, resulting in needless rewriting
8216 *) Patches to make OpenSSL compile under Win32 again. Mostly
8217 work arounds for the VC++ problem that it treats func() as
8218 func(void). Also stripped out the parts of mkdef.pl that
8219 added extra typesafe functions: these no longer exist.
8222 *) Reorganisation of the stack code. The macros are now all
8223 collected in safestack.h . Each macro is defined in terms of
8224 a "stack macro" of the form SKM_<name>(type, a, b). The
8225 DEBUG_SAFESTACK is now handled in terms of function casts,
8226 this has the advantage of retaining type safety without the
8227 use of additional functions. If DEBUG_SAFESTACK is not defined
8228 then the non typesafe macros are used instead. Also modified the
8229 mkstack.pl script to handle the new form. Needs testing to see
8230 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8231 the default if no major problems. Similar behaviour for ASN1_SET_OF
8232 and PKCS12_STACK_OF.
8235 *) When some versions of IIS use the 'NET' form of private key the
8236 key derivation algorithm is different. Normally MD5(password) is
8237 used as a 128 bit RC4 key. In the modified case
8238 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8239 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8240 as the old Netscape_RSA functions except they have an additional
8241 'sgckey' parameter which uses the modified algorithm. Also added
8242 an -sgckey command line option to the rsa utility. Thanks to
8243 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8244 algorithm to openssl-dev.
8247 *) The evp_local.h macros were using 'c.##kname' which resulted in
8248 invalid expansion on some systems (SCO 5.0.5 for example).
8249 Corrected to 'c.kname'.
8250 [Phillip Porch <root@theporch.com>]
8252 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8253 a STACK of email addresses from a certificate or request, these look
8254 in the subject name and the subject alternative name extensions and
8255 omit any duplicate addresses.
8258 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8259 This makes DSA verification about 2 % faster.
8262 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8263 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8264 plus overhead for 1024 bit moduli).
8265 This makes exponentiations about 0.5 % faster for 1024 bit
8266 exponents (as measured by "openssl speed rsa2048").
8269 *) Rename memory handling macros to avoid conflicts with other
8271 Malloc => OPENSSL_malloc
8272 Malloc_locked => OPENSSL_malloc_locked
8273 Realloc => OPENSSL_realloc
8274 Free => OPENSSL_free
8277 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8278 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8281 *) CygWin32 support.
8282 [John Jarvie <jjarvie@newsguy.com>]
8284 *) The type-safe stack code has been rejigged. It is now only compiled
8285 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8286 by default all type-specific stack functions are "#define"d back to
8287 standard stack functions. This results in more streamlined output
8288 but retains the type-safety checking possibilities of the original
8292 *) The STACK code has been cleaned up, and certain type declarations
8293 that didn't make a lot of sense have been brought in line. This has
8294 also involved a cleanup of sorts in safestack.h to more correctly
8295 map type-safe stack functions onto their plain stack counterparts.
8296 This work has also resulted in a variety of "const"ifications of
8297 lots of the code, especially "_cmp" operations which should normally
8298 be prototyped with "const" parameters anyway.
8301 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8302 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8303 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8304 where all of 'md' is used each time the PRNG is used, but 'state'
8305 is used only indexed by a cyclic counter. As entropy may not be
8306 well distributed from the beginning, 'md' is important as a
8307 chaining variable. However, the output function chains only half
8308 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8309 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8310 in all of 'state' being rewritten, with the new values depending
8311 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8314 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8315 the handshake is continued after ssl_verify_cert_chain();
8316 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8317 can lead to 'unexplainable' connection aborts later.
8318 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8320 *) Major EVP API cipher revision.
8321 Add hooks for extra EVP features. This allows various cipher
8322 parameters to be set in the EVP interface. Support added for variable
8323 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8324 setting of RC2 and RC5 parameters.
8326 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8329 Remove lots of duplicated code from the EVP library. For example *every*
8330 cipher init() function handles the 'iv' in the same way according to the
8331 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8332 for CFB and OFB modes they zero ctx->num.
8334 New functionality allows removal of S/MIME code RC2 hack.
8336 Most of the routines have the same form and so can be declared in terms
8339 By shifting this to the top level EVP_CipherInit() it can be removed from
8340 all individual ciphers. If the cipher wants to handle IVs or keys
8341 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8344 Change lots of functions like EVP_EncryptUpdate() to now return a
8345 value: although software versions of the algorithms cannot fail
8346 any installed hardware versions can.
8349 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8350 this option is set, tolerate broken clients that send the negotiated
8351 protocol version number instead of the requested protocol version
8355 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8356 i.e. non-zero for export ciphersuites, zero otherwise.
8357 Previous versions had this flag inverted, inconsistent with
8358 rsa_tmp_cb (..._TMP_RSA_CB).
8359 [Bodo Moeller; problem reported by Amit Chopra]
8361 *) Add missing DSA library text string. Work around for some IIS
8362 key files with invalid SEQUENCE encoding.
8365 *) Add a document (doc/standards.txt) that list all kinds of standards
8366 and so on that are implemented in OpenSSL.
8369 *) Enhance c_rehash script. Old version would mishandle certificates
8370 with the same subject name hash and wouldn't handle CRLs at all.
8371 Added -fingerprint option to crl utility, to support new c_rehash
8375 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8378 *) Fix for SSL server purpose checking. Server checking was
8379 rejecting certificates which had extended key usage present
8380 but no ssl client purpose.
8381 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8383 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8384 is a little unclear about how a blank password is handled.
8385 Since the password in encoded as a BMPString with terminating
8386 double NULL a zero length password would end up as just the
8387 double NULL. However no password at all is different and is
8388 handled differently in the PKCS#12 key generation code. NS
8389 treats a blank password as zero length. MSIE treats it as no
8390 password on export: but it will try both on import. We now do
8391 the same: PKCS12_parse() tries zero length and no password if
8392 the password is set to "" or NULL (NULL is now a valid password:
8393 it wasn't before) as does the pkcs12 application.
8396 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8397 perror when PEM_read_bio_X509_REQ fails, the error message must
8398 be obtained from the error queue.
8401 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8402 it in ERR_remove_state if appropriate, and change ERR_get_state
8403 accordingly to avoid race conditions (this is necessary because
8404 thread_hash is no longer constant once set).
8407 *) Bugfix for linux-elf makefile.one.
8410 *) RSA_get_default_method() will now cause a default
8411 RSA_METHOD to be chosen if one doesn't exist already.
8412 Previously this was only set during a call to RSA_new()
8413 or RSA_new_method(NULL) meaning it was possible for
8414 RSA_get_default_method() to return NULL.
8417 *) Added native name translation to the existing DSO code
8418 that will convert (if the flag to do so is set) filenames
8419 that are sufficiently small and have no path information
8420 into a canonical native form. Eg. "blah" converted to
8421 "libblah.so" or "blah.dll" etc.
8424 *) New function ERR_error_string_n(e, buf, len) which is like
8425 ERR_error_string(e, buf), but writes at most 'len' bytes
8426 including the 0 terminator. For ERR_error_string_n, 'buf'
8428 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8430 *) CONF library reworked to become more general. A new CONF
8431 configuration file reader "class" is implemented as well as a
8432 new functions (NCONF_*, for "New CONF") to handle it. The now
8433 old CONF_* functions are still there, but are reimplemented to
8434 work in terms of the new functions. Also, a set of functions
8435 to handle the internal storage of the configuration data is
8436 provided to make it easier to write new configuration file
8437 reader "classes" (I can definitely see something reading a
8438 configuration file in XML format, for example), called _CONF_*,
8439 or "the configuration storage API"...
8441 The new configuration file reading functions are:
8443 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8444 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8446 NCONF_default, NCONF_WIN32
8448 NCONF_dump_fp, NCONF_dump_bio
8450 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8451 NCONF_new creates a new CONF object. This works in the same way
8452 as other interfaces in OpenSSL, like the BIO interface.
8453 NCONF_dump_* dump the internal storage of the configuration file,
8454 which is useful for debugging. All other functions take the same
8455 arguments as the old CONF_* functions wth the exception of the
8456 first that must be a `CONF *' instead of a `LHASH *'.
8458 To make it easer to use the new classes with the old CONF_* functions,
8459 the function CONF_set_default_method is provided.
8462 *) Add '-tls1' option to 'openssl ciphers', which was already
8463 mentioned in the documentation but had not been implemented.
8464 (This option is not yet really useful because even the additional
8465 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8468 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8469 OpenSSL-based applications) load shared libraries and bind to
8470 them in a portable way.
8471 [Geoff Thorpe, with contributions from Richard Levitte]
8473 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8475 *) Make sure _lrotl and _lrotr are only used with MSVC.
8477 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8478 (the default implementation of RAND_status).
8480 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8481 to '-clrext' (= clear extensions), as intended and documented.
8482 [Bodo Moeller; inconsistency pointed out by Michael Attili
8485 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8486 was larger than the MD block size.
8487 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8489 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8490 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8491 using the passed key: if the passed key was a private key the result
8492 of X509_print(), for example, would be to print out all the private key
8496 *) des_quad_cksum() byte order bug fix.
8497 [Ulf Möller, using the problem description in krb4-0.9.7, where
8498 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8500 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8502 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8504 *) For easily testing in shell scripts whether some command
8505 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8506 returns with exit code 0 iff no command of the given name is available.
8507 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8508 the output goes to stdout and nothing is printed to stderr.
8509 Additional arguments are always ignored.
8511 Since for each cipher there is a command of the same name,
8512 the 'no-cipher' compilation switches can be tested this way.
8514 ('openssl no-XXX' is not able to detect pseudo-commands such
8515 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8518 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8521 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8522 is set; it will be thrown away anyway because each handshake creates
8524 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8525 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8526 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8527 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8530 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8531 'Q' and 'R' lose their special meanings (quit/renegotiate).
8532 This is part of what -quiet does; unlike -quiet, -ign_eof
8533 does not suppress any output.
8536 *) Add compatibility options to the purpose and trust code. The
8537 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8538 accepts a certificate or CA, this was the previous behaviour,
8539 with all the associated security issues.
8541 X509_TRUST_COMPAT is the old trust behaviour: only and
8542 automatically trust self signed roots in certificate store. A
8543 new trust setting X509_TRUST_DEFAULT is used to specify that
8544 a purpose has no associated trust setting and it should instead
8545 use the value in the default purpose.
8548 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8549 and fix a memory leak.
8552 *) In util/mkerr.pl (which implements 'make errors'), preserve
8553 reason strings from the previous version of the .c file, as
8554 the default to have only downcase letters (and digits) in
8555 automatically generated reasons codes is not always appropriate.
8558 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8559 using strerror. Previously, ERR_reason_error_string() returned
8560 library names as reason strings for SYSerr; but SYSerr is a special
8561 case where small numbers are errno values, not library numbers.
8564 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8565 converts DSA parameters into DH parameters. (When creating parameters,
8566 DSA_generate_parameters is used.)
8569 *) Include 'length' (recommended exponent length) in C code generated
8570 by 'openssl dhparam -C'.
8573 *) The second argument to set_label in perlasm was already being used
8574 so couldn't be used as a "file scope" flag. Moved to third argument
8578 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8579 instead of RAND_bytes for encryption IVs and salts.
8582 *) Include RAND_status() into RAND_METHOD instead of implementing
8583 it only for md_rand.c Otherwise replacing the PRNG by calling
8584 RAND_set_rand_method would be impossible.
8587 *) Don't let DSA_generate_key() enter an infinite loop if the random
8588 number generation fails.
8591 *) New 'rand' application for creating pseudo-random output.
8594 *) Added configuration support for Linux/IA64
8595 [Rolf Haberrecker <rolf@suse.de>]
8597 *) Assembler module support for Mingw32.
8600 *) Shared library support for HPUX (in shlib/).
8601 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8603 *) Shared library support for Solaris gcc.
8604 [Lutz Behnke <behnke@trustcenter.de>]
8606 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8608 *) PKCS7_encrypt() was adding text MIME headers twice because they
8609 were added manually and by SMIME_crlf_copy().
8612 *) In bntest.c don't call BN_rand with zero bits argument.
8613 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8615 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8616 case was implemented. This caused BN_div_recp() to fail occasionally.
8619 *) Add an optional second argument to the set_label() in the perl
8620 assembly language builder. If this argument exists and is set
8621 to 1 it signals that the assembler should use a symbol whose
8622 scope is the entire file, not just the current function. This
8623 is needed with MASM which uses the format label:: for this scope.
8624 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8626 *) Change the ASN1 types so they are typedefs by default. Before
8627 almost all types were #define'd to ASN1_STRING which was causing
8628 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8632 *) Change names of new functions to the new get1/get0 naming
8633 convention: After 'get1', the caller owns a reference count
8634 and has to call ..._free; 'get0' returns a pointer to some
8635 data structure without incrementing reference counters.
8636 (Some of the existing 'get' functions increment a reference
8637 counter, some don't.)
8638 Similarly, 'set1' and 'add1' functions increase reference
8639 counters or duplicate objects.
8642 *) Allow for the possibility of temp RSA key generation failure:
8643 the code used to assume it always worked and crashed on failure.
8646 *) Fix potential buffer overrun problem in BIO_printf().
8647 [Ulf Möller, using public domain code by Patrick Powell; problem
8648 pointed out by David Sacerdote <das33@cornell.edu>]
8650 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8651 RAND_egd() and RAND_status(). In the command line application,
8652 the EGD socket can be specified like a seed file using RANDFILE
8656 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8657 Some CAs (e.g. Verisign) distribute certificates in this form.
8660 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8661 list to exclude them. This means that no special compilation option
8662 is needed to use anonymous DH: it just needs to be included in the
8666 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8667 EVP_MD_type. The old functionality is available in a new macro called
8668 EVP_MD_md(). Change code that uses it and update docs.
8671 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8672 where the 'void *' argument is replaced by a function pointer argument.
8673 Previously 'void *' was abused to point to functions, which works on
8674 many platforms, but is not correct. As these functions are usually
8675 called by macros defined in OpenSSL header files, most source code
8676 should work without changes.
8679 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8680 sections with information on -D... compiler switches used for
8681 compiling the library so that applications can see them. To enable
8682 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8683 must be defined. E.g.,
8684 #define OPENSSL_ALGORITHM_DEFINES
8685 #include <openssl/opensslconf.h>
8686 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8687 [Richard Levitte, Ulf and Bodo Möller]
8689 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8693 *) Change the 'other' type in certificate aux info to a STACK_OF
8694 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8695 the required ASN1 format: arbitrary types determined by an OID.
8698 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8699 argument to 'req'. This is not because the function is newer or
8700 better than others it just uses the work 'NEW' in the certificate
8701 request header lines. Some software needs this.
8704 *) Reorganise password command line arguments: now passwords can be
8705 obtained from various sources. Delete the PEM_cb function and make
8706 it the default behaviour: i.e. if the callback is NULL and the
8707 usrdata argument is not NULL interpret it as a null terminated pass
8708 phrase. If usrdata and the callback are NULL then the pass phrase
8709 is prompted for as usual.
8712 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8713 the support is automatically enabled. The resulting binaries will
8714 autodetect the card and use it if present.
8715 [Ben Laurie and Compaq Inc.]
8717 *) Work around for Netscape hang bug. This sends certificate request
8718 and server done in one record. Since this is perfectly legal in the
8719 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8720 the bugs/SSLv3 entry for more info.
8723 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8726 *) Add -rand argument to smime and pkcs12 applications and read/write
8730 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8733 *) Add command line password options to the remaining applications.
8736 *) Bug fix for BN_div_recp() for numerators with an even number of
8740 *) More tests in bntest.c, and changed test_bn output.
8743 *) ./config recognizes MacOS X now.
8746 *) Bug fix for BN_div() when the first words of num and divsor are
8747 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8750 *) Add support for various broken PKCS#8 formats, and command line
8751 options to produce them.
8754 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8755 get temporary BIGNUMs from a BN_CTX.
8758 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8762 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8763 include a #define from the old name to the new. The original intent
8764 was that statically linked binaries could for example just call
8765 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8766 link with digests. This never worked because SSLeay_add_all_digests()
8767 and SSLeay_add_all_ciphers() were in the same source file so calling
8768 one would link with the other. They are now in separate source files.
8771 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8774 *) Use a less unusual form of the Miller-Rabin primality test (it used
8775 a binary algorithm for exponentiation integrated into the Miller-Rabin
8776 loop, our standard modexp algorithms are faster).
8779 *) Support for the EBCDIC character set completed.
8780 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8782 *) Source code cleanups: use const where appropriate, eliminate casts,
8783 use void * instead of char * in lhash.
8786 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8787 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8788 this the server could overwrite ephemeral keys that the client
8792 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8793 using 50 iterations of the Rabin-Miller test.
8795 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8796 iterations of the Rabin-Miller test as required by the appendix
8797 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8798 As BN_is_prime_fasttest includes trial division, DSA parameter
8799 generation becomes much faster.
8801 This implies a change for the callback functions in DSA_is_prime
8802 and DSA_generate_parameters: The callback function is called once
8803 for each positive witness in the Rabin-Miller test, not just
8804 occasionally in the inner loop; and the parameters to the
8805 callback function now provide an iteration count for the outer
8806 loop rather than for the current invocation of the inner loop.
8807 DSA_generate_parameters additionally can call the callback
8808 function with an 'iteration count' of -1, meaning that a
8809 candidate has passed the trial division test (when q is generated
8810 from an application-provided seed, trial division is skipped).
8813 *) New function BN_is_prime_fasttest that optionally does trial
8814 division before starting the Rabin-Miller test and has
8815 an additional BN_CTX * argument (whereas BN_is_prime always
8816 has to allocate at least one BN_CTX).
8817 'callback(1, -1, cb_arg)' is called when a number has passed the
8818 trial division stage.
8821 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8825 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8828 *) New function BN_pseudo_rand().
8831 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8832 bignum version of BN_from_montgomery() with the working code from
8833 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8837 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8838 made it impossible to use the same SSL_SESSION data structure in
8839 SSL2 clients in multiple threads.
8842 *) The return value of RAND_load_file() no longer counts bytes obtained
8843 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8844 to seed the PRNG (previously an explicit byte count was required).
8845 [Ulf Möller, Bodo Möller]
8847 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8848 used (char *) instead of (void *) and had casts all over the place.
8851 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8854 *) Retain source code compatibility for BN_prime_checks macro:
8855 BN_is_prime(..., BN_prime_checks, ...) now uses
8856 BN_prime_checks_for_size to determine the appropriate number of
8857 Rabin-Miller iterations.
8860 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8861 DH_CHECK_P_NOT_SAFE_PRIME.
8862 (Check if this is true? OpenPGP calls them "strong".)
8865 *) Merge the functionality of "dh" and "gendh" programs into a new program
8866 "dhparam". The old programs are retained for now but will handle DH keys
8867 (instead of parameters) in future.
8870 *) Make the ciphers, s_server and s_client programs check the return values
8871 when a new cipher list is set.
8874 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8875 ciphers. Before when the 56bit ciphers were enabled the sorting was
8878 The syntax for the cipher sorting has been extended to support sorting by
8879 cipher-strength (using the strength_bits hard coded in the tables).
8880 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8882 Fix a bug in the cipher-command parser: when supplying a cipher command
8883 string with an "undefined" symbol (neither command nor alphanumeric
8884 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8885 an error is flagged.
8887 Due to the strength-sorting extension, the code of the
8888 ssl_create_cipher_list() function was completely rearranged. I hope that
8889 the readability was also increased :-)
8890 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8892 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8893 for the first serial number and places 2 in the serial number file. This
8894 avoids problems when the root CA is created with serial number zero and
8895 the first user certificate has the same issuer name and serial number
8899 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8900 the new code. Add documentation for this stuff.
8903 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8904 X509_*() to X509at_*() on the grounds that they don't handle X509
8905 structures and behave in an analogous way to the X509v3 functions:
8906 they shouldn't be called directly but wrapper functions should be used
8909 So we also now have some wrapper functions that call the X509at functions
8910 when passed certificate requests. (TO DO: similar things can be done with
8911 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8912 things. Some of these need some d2i or i2d and print functionality
8913 because they handle more complex structures.)
8916 *) Add missing #ifndefs that caused missing symbols when building libssl
8917 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8918 NO_RSA in ssl/s2*.c.
8919 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8921 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8922 has a return value which indicates the quality of the random data
8923 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8924 error queue. New function RAND_pseudo_bytes() generates output that is
8925 guaranteed to be unique but not unpredictable. RAND_add is like
8926 RAND_seed, but takes an extra argument for an entropy estimate
8927 (RAND_seed always assumes full entropy).
8930 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8931 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8932 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8933 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8934 false-positive rate of at most 2^-80 for random input.
8937 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8940 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8941 in the 0.9.5 release), this returns the chain
8942 from an X509_CTX structure with a dup of the stack and all
8943 the X509 reference counts upped: so the stack will exist
8944 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8947 Also make SSL_SESSION_print() print out the verify return
8951 *) Add manpage for the pkcs12 command. Also change the default
8952 behaviour so MAC iteration counts are used unless the new
8953 -nomaciter option is used. This improves file security and
8954 only older versions of MSIE (4.0 for example) need it.
8957 *) Honor the no-xxx Configure options when creating .DEF files.
8960 *) Add PKCS#10 attributes to field table: challengePassword,
8961 unstructuredName and unstructuredAddress. These are taken from
8962 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8963 international characters are used.
8965 More changes to X509_ATTRIBUTE code: allow the setting of types
8966 based on strings. Remove the 'loc' parameter when adding
8967 attributes because these will be a SET OF encoding which is sorted
8971 *) Initial changes to the 'req' utility to allow request generation
8972 automation. This will allow an application to just generate a template
8973 file containing all the field values and have req construct the
8976 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8977 used all over the place including certificate requests and PKCS#7
8978 structures. They are currently handled manually where necessary with
8979 some primitive wrappers for PKCS#7. The new functions behave in a
8980 manner analogous to the X509 extension functions: they allow
8981 attributes to be looked up by NID and added.
8983 Later something similar to the X509V3 code would be desirable to
8984 automatically handle the encoding, decoding and printing of the
8985 more complex types. The string types like challengePassword can
8986 be handled by the string table functions.
8988 Also modified the multi byte string table handling. Now there is
8989 a 'global mask' which masks out certain types. The table itself
8990 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8991 is useful when for example there is only one permissible type
8992 (as in countryName) and using the mask might result in no valid
8996 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8997 SSL_get_peer_finished to allow applications to obtain the latest
8998 Finished messages sent to the peer or expected from the peer,
8999 respectively. (SSL_get_peer_finished is usually the Finished message
9000 actually received from the peer, otherwise the protocol will be aborted.)
9002 As the Finished message are message digests of the complete handshake
9003 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
9004 be used for external authentication procedures when the authentication
9005 provided by SSL/TLS is not desired or is not enough.
9008 *) Enhanced support for Alpha Linux is added. Now ./config checks if
9009 the host supports BWX extension and if Compaq C is present on the
9010 $PATH. Just exploiting of the BWX extension results in 20-30%
9011 performance kick for some algorithms, e.g. DES and RC4 to mention
9012 a couple. Compaq C in turn generates ~20% faster code for MD5 and
9016 *) Add support for MS "fast SGC". This is arguably a violation of the
9017 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
9018 weak crypto and after checking the certificate is SGC a second one
9019 with strong crypto. MS SGC stops the first handshake after receiving
9020 the server certificate message and sends a second client hello. Since
9021 a server will typically do all the time consuming operations before
9022 expecting any further messages from the client (server key exchange
9023 is the most expensive) there is little difference between the two.
9025 To get OpenSSL to support MS SGC we have to permit a second client
9026 hello message after we have sent server done. In addition we have to
9027 reset the MAC if we do get this second client hello.
9030 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
9031 if a DER encoded private key is RSA or DSA traditional format. Changed
9032 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
9033 format DER encoded private key. Newer code should use PKCS#8 format which
9034 has the key type encoded in the ASN1 structure. Added DER private key
9035 support to pkcs8 application.
9038 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
9039 ciphersuites has been selected (as required by the SSL 3/TLS 1
9040 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
9041 is set, we interpret this as a request to violate the specification
9042 (the worst that can happen is a handshake failure, and 'correct'
9043 behaviour would result in a handshake failure anyway).
9046 *) In SSL_CTX_add_session, take into account that there might be multiple
9047 SSL_SESSION structures with the same session ID (e.g. when two threads
9048 concurrently obtain them from an external cache).
9049 The internal cache can handle only one SSL_SESSION with a given ID,
9050 so if there's a conflict, we now throw out the old one to achieve
9054 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
9055 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
9056 some routines that use cipher OIDs: some ciphers do not have OIDs
9057 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
9061 *) Simplify the trust setting structure and code. Now we just have
9062 two sequences of OIDs for trusted and rejected settings. These will
9063 typically have values the same as the extended key usage extension
9064 and any application specific purposes.
9066 The trust checking code now has a default behaviour: it will just
9067 check for an object with the same NID as the passed id. Functions can
9068 be provided to override either the default behaviour or the behaviour
9069 for a given id. SSL client, server and email already have functions
9070 in place for compatibility: they check the NID and also return "trusted"
9071 if the certificate is self signed.
9074 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
9075 traditional format into an EVP_PKEY structure.
9078 *) Add a password callback function PEM_cb() which either prompts for
9079 a password if usr_data is NULL or otherwise assumes it is a null
9080 terminated password. Allow passwords to be passed on command line
9081 environment or config files in a few more utilities.
9084 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
9085 keys. Add some short names for PKCS#8 PBE algorithms and allow them
9086 to be specified on the command line for the pkcs8 and pkcs12 utilities.
9087 Update documentation.
9090 *) Support for ASN1 "NULL" type. This could be handled before by using
9091 ASN1_TYPE but there wasn't any function that would try to read a NULL
9092 and produce an error if it couldn't. For compatibility we also have
9093 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
9094 don't allocate anything because they don't need to.
9097 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
9099 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
9101 *) Rebuild of the memory allocation routines used by OpenSSL code and
9102 possibly others as well. The purpose is to make an interface that
9103 provide hooks so anyone can build a separate set of allocation and
9104 deallocation routines to be used by OpenSSL, for example memory
9105 pool implementations, or something else, which was previously hard
9106 since Malloc(), Realloc() and Free() were defined as macros having
9107 the values malloc, realloc and free, respectively (except for Win32
9108 compilations). The same is provided for memory debugging code.
9109 OpenSSL already comes with functionality to find memory leaks, but
9110 this gives people a chance to debug other memory problems.
9112 With these changes, a new set of functions and macros have appeared:
9114 CRYPTO_set_mem_debug_functions() [F]
9115 CRYPTO_get_mem_debug_functions() [F]
9116 CRYPTO_dbg_set_options() [F]
9117 CRYPTO_dbg_get_options() [F]
9118 CRYPTO_malloc_debug_init() [M]
9120 The memory debug functions are NULL by default, unless the library
9121 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
9122 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
9123 gives the standard debugging functions that come with OpenSSL) or
9124 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
9125 provided by the library user) must be used. When the standard
9126 debugging functions are used, CRYPTO_dbg_set_options can be used to
9127 request additional information:
9128 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
9129 the CRYPTO_MDEBUG_xxx macro when compiling the library.
9131 Also, things like CRYPTO_set_mem_functions will always give the
9132 expected result (the new set of functions is used for allocation
9133 and deallocation) at all times, regardless of platform and compiler
9136 To finish it up, some functions that were never use in any other
9137 way than through macros have a new API and new semantic:
9140 CRYPTO_dbg_realloc()
9143 All macros of value have retained their old syntax.
9144 [Richard Levitte and Bodo Moeller]
9146 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9147 ordering of SMIMECapabilities wasn't in "strength order" and there
9148 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9152 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9153 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9154 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9156 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9157 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9158 functionality to handle multipart/signed properly) and a utility
9159 called 'smime' to call all this stuff. This is based on code I
9160 originally wrote for Celo who have kindly allowed it to be
9161 included in OpenSSL.
9164 *) Add variants des_set_key_checked and des_set_key_unchecked of
9165 des_set_key (aka des_key_sched). Global variable des_check_key
9166 decides which of these is called by des_set_key; this way
9167 des_check_key behaves as it always did, but applications and
9168 the library itself, which was buggy for des_check_key == 1,
9169 have a cleaner way to pick the version they need.
9172 *) New function PKCS12_newpass() which changes the password of a
9176 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9177 dynamic mix. In both cases the ids can be used as an index into the
9178 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9179 functions so they accept a list of the field values and the
9180 application doesn't need to directly manipulate the X509_TRUST
9184 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9188 *) Modify the way the V3 extension code looks up extensions. This now
9189 works in a similar way to the object code: we have some "standard"
9190 extensions in a static table which is searched with OBJ_bsearch()
9191 and the application can add dynamic ones if needed. The file
9192 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9193 updated whenever a new extension is added to the core code and kept
9194 in ext_nid order. There is a simple program 'tabtest.c' which checks
9195 this. New extensions are not added too often so this file can readily
9196 be maintained manually.
9198 There are two big advantages in doing things this way. The extensions
9199 can be looked up immediately and no longer need to be "added" using
9200 X509V3_add_standard_extensions(): this function now does nothing.
9201 [Side note: I get *lots* of email saying the extension code doesn't
9202 work because people forget to call this function]
9203 Also no dynamic allocation is done unless new extensions are added:
9204 so if we don't add custom extensions there is no need to call
9205 X509V3_EXT_cleanup().
9208 *) Modify enc utility's salting as follows: make salting the default. Add a
9209 magic header, so unsalted files fail gracefully instead of just decrypting
9210 to garbage. This is because not salting is a big security hole, so people
9211 should be discouraged from doing it.
9214 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9215 digest to be passed on the command line but it only used this
9216 parameter when signing a certificate. Modified so all relevant
9217 operations are affected by the digest parameter including the
9218 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9219 DSA key was used because it didn't fix the digest.
9222 *) Initial certificate chain verify code. Currently tests the untrusted
9223 certificates for consistency with the verify purpose (which is set
9224 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9226 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9227 this is because it will reject chains with invalid extensions whereas
9228 every previous version of OpenSSL and SSLeay made no checks at all.
9230 Trust code: checks the root CA for the relevant trust settings. Trust
9231 settings have an initial value consistent with the verify purpose: e.g.
9232 if the verify purpose is for SSL client use it expects the CA to be
9233 trusted for SSL client use. However the default value can be changed to
9234 permit custom trust settings: one example of this would be to only trust
9235 certificates from a specific "secure" set of CAs.
9237 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9238 which should be used for version portability: especially since the
9239 verify structure is likely to change more often now.
9241 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9242 to set them. If not set then assume SSL clients will verify SSL servers
9245 Two new options to the verify program: -untrusted allows a set of
9246 untrusted certificates to be passed in and -purpose which sets the
9247 intended purpose of the certificate. If a purpose is set then the
9248 new chain verify code is used to check extension consistency.
9251 *) Support for the authority information access extension.
9254 *) Modify RSA and DSA PEM read routines to transparently handle
9255 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9256 public keys in a format compatible with certificate
9257 SubjectPublicKeyInfo structures. Unfortunately there were already
9258 functions called *_PublicKey_* which used various odd formats so
9259 these are retained for compatibility: however the DSA variants were
9260 never in a public release so they have been deleted. Changed dsa/rsa
9261 utilities to handle the new format: note no releases ever handled public
9262 keys so we should be OK.
9264 The primary motivation for this change is to avoid the same fiasco
9265 that dogs private keys: there are several incompatible private key
9266 formats some of which are standard and some OpenSSL specific and
9267 require various evil hacks to allow partial transparent handling and
9268 even then it doesn't work with DER formats. Given the option anything
9269 other than PKCS#8 should be dumped: but the other formats have to
9270 stay in the name of compatibility.
9272 With public keys and the benefit of hindsight one standard format
9273 is used which works with EVP_PKEY, RSA or DSA structures: though
9274 it clearly returns an error if you try to read the wrong kind of key.
9276 Added a -pubkey option to the 'x509' utility to output the public key.
9277 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9278 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9279 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9280 that do the same as the EVP_PKEY_assign_*() except they up the
9281 reference count of the added key (they don't "swallow" the
9285 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9286 CRLs would fail if the file contained no certificates or no CRLs:
9287 added a new function to read in both types and return the number
9288 read: this means that if none are read it will be an error. The
9289 DER versions of the certificate and CRL reader would always fail
9290 because it isn't possible to mix certificates and CRLs in DER format
9291 without choking one or the other routine. Changed this to just read
9292 a certificate: this is the best we can do. Also modified the code
9293 in apps/verify.c to take notice of return codes: it was previously
9294 attempting to read in certificates from NULL pointers and ignoring
9295 any errors: this is one reason why the cert and CRL reader seemed
9296 to work. It doesn't check return codes from the default certificate
9297 routines: these may well fail if the certificates aren't installed.
9300 *) Code to support otherName option in GeneralName.
9303 *) First update to verify code. Change the verify utility
9304 so it warns if it is passed a self signed certificate:
9305 for consistency with the normal behaviour. X509_verify
9306 has been modified to it will now verify a self signed
9307 certificate if *exactly* the same certificate appears
9308 in the store: it was previously impossible to trust a
9309 single self signed certificate. This means that:
9310 openssl verify ss.pem
9311 now gives a warning about a self signed certificate but
9312 openssl verify -CAfile ss.pem ss.pem
9316 *) For servers, store verify_result in SSL_SESSION data structure
9317 (and add it to external session representation).
9318 This is needed when client certificate verifications fails,
9319 but an application-provided verification callback (set by
9320 SSL_CTX_set_cert_verify_callback) allows accepting the session
9321 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9322 but returns 1): When the session is reused, we have to set
9323 ssl->verify_result to the appropriate error code to avoid
9325 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9327 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9328 case in PKCS7_dataInit() where the signed PKCS7 structure
9329 didn't contain any existing data because it was being created.
9330 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9332 *) Add a salt to the key derivation routines in enc.c. This
9333 forms the first 8 bytes of the encrypted file. Also add a
9334 -S option to allow a salt to be input on the command line.
9337 *) New function X509_cmp(). Oddly enough there wasn't a function
9338 to compare two certificates. We do this by working out the SHA1
9339 hash and comparing that. X509_cmp() will be needed by the trust
9343 *) SSL_get1_session() is like SSL_get_session(), but increments
9344 the reference count in the SSL_SESSION returned.
9345 [Geoff Thorpe <geoff@eu.c2.net>]
9347 *) Fix for 'req': it was adding a null to request attributes.
9348 Also change the X509_LOOKUP and X509_INFO code to handle
9349 certificate auxiliary information.
9352 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9356 *) Add the possibility to add extra information to the memory leak
9357 detecting output, to form tracebacks, showing from where each
9358 allocation was originated: CRYPTO_push_info("constant string") adds
9359 the string plus current file name and line number to a per-thread
9360 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9361 is like calling CYRPTO_pop_info() until the stack is empty.
9362 Also updated memory leak detection code to be multi-thread-safe.
9365 *) Add options -text and -noout to pkcs7 utility and delete the
9366 encryption options which never did anything. Update docs.
9369 *) Add options to some of the utilities to allow the pass phrase
9370 to be included on either the command line (not recommended on
9371 OSes like Unix) or read from the environment. Update the
9372 manpages and fix a few bugs.
9375 *) Add a few manpages for some of the openssl commands.
9378 *) Fix the -revoke option in ca. It was freeing up memory twice,
9379 leaking and not finding already revoked certificates.
9382 *) Extensive changes to support certificate auxiliary information.
9383 This involves the use of X509_CERT_AUX structure and X509_AUX
9384 functions. An X509_AUX function such as PEM_read_X509_AUX()
9385 can still read in a certificate file in the usual way but it
9386 will also read in any additional "auxiliary information". By
9387 doing things this way a fair degree of compatibility can be
9388 retained: existing certificates can have this information added
9389 using the new 'x509' options.
9391 Current auxiliary information includes an "alias" and some trust
9392 settings. The trust settings will ultimately be used in enhanced
9393 certificate chain verification routines: currently a certificate
9394 can only be trusted if it is self signed and then it is trusted
9398 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9399 The problem was that one of the replacement routines had not been working
9400 since SSLeay releases. For now the offending routine has been replaced
9401 with non-optimised assembler. Even so, this now gives around 95%
9402 performance improvement for 1024 bit RSA signs.
9405 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9406 handling. Most clients have the effective key size in bits equal to
9407 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9408 A few however don't do this and instead use the size of the decrypted key
9409 to determine the RC2 key length and the AlgorithmIdentifier to determine
9410 the effective key length. In this case the effective key length can still
9411 be 40 bits but the key length can be 168 bits for example. This is fixed
9412 by manually forcing an RC2 key into the EVP_PKEY structure because the
9413 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9414 the key length and effective key length are equal.
9417 *) Add a bunch of functions that should simplify the creation of
9418 X509_NAME structures. Now you should be able to do:
9419 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9420 and have it automatically work out the correct field type and fill in
9421 the structures. The more adventurous can try:
9422 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9423 and it will (hopefully) work out the correct multibyte encoding.
9426 *) Change the 'req' utility to use the new field handling and multibyte
9427 copy routines. Before the DN field creation was handled in an ad hoc
9428 way in req, ca, and x509 which was rather broken and didn't support
9429 BMPStrings or UTF8Strings. Since some software doesn't implement
9430 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9431 using the dirstring_type option. See the new comment in the default
9432 openssl.cnf for more info.
9435 *) Make crypto/rand/md_rand.c more robust:
9436 - Assure unique random numbers after fork().
9437 - Make sure that concurrent threads access the global counter and
9438 md serializably so that we never lose entropy in them
9439 or use exactly the same state in multiple threads.
9440 Access to the large state is not always serializable because
9441 the additional locking could be a performance killer, and
9442 md should be large enough anyway.
9445 *) New file apps/app_rand.c with commonly needed functionality
9446 for handling the random seed file.
9448 Use the random seed file in some applications that previously did not:
9450 dsaparam -genkey (which also ignored its '-rand' option),
9453 x509 (when signing).
9454 Except on systems with /dev/urandom, it is crucial to have a random
9455 seed file at least for key creation, DSA signing, and for DH exchanges;
9456 for RSA signatures we could do without one.
9458 gendh and gendsa (unlike genrsa) used to read only the first byte
9459 of each file listed in the '-rand' option. The function as previously
9460 found in genrsa is now in app_rand.c and is used by all programs
9461 that support '-rand'.
9464 *) In RAND_write_file, use mode 0600 for creating files;
9465 don't just chmod when it may be too late.
9468 *) Report an error from X509_STORE_load_locations
9469 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9472 *) New function ASN1_mbstring_copy() this copies a string in either
9473 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9474 into an ASN1_STRING type. A mask of permissible types is passed
9475 and it chooses the "minimal" type to use or an error if not type
9479 *) Add function equivalents to the various macros in asn1.h. The old
9480 macros are retained with an M_ prefix. Code inside the library can
9481 use the M_ macros. External code (including the openssl utility)
9482 should *NOT* in order to be "shared library friendly".
9485 *) Add various functions that can check a certificate's extensions
9486 to see if it usable for various purposes such as SSL client,
9487 server or S/MIME and CAs of these types. This is currently
9488 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9489 verification. Also added a -purpose flag to x509 utility to
9490 print out all the purposes.
9493 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9497 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9498 for, obtain and decode and extension and obtain its critical flag.
9499 This allows all the necessary extension code to be handled in a
9500 single function call.
9503 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9504 platforms. See crypto/rc4/rc4_enc.c for further details.
9507 *) New -noout option to asn1parse. This causes no output to be produced
9508 its main use is when combined with -strparse and -out to extract data
9509 from a file (which may not be in ASN.1 format).
9512 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9513 when producing the local key id.
9514 [Richard Levitte <levitte@stacken.kth.se>]
9516 *) New option -dhparam in s_server. This allows a DH parameter file to be
9517 stated explicitly. If it is not stated then it tries the first server
9518 certificate file. The previous behaviour hard coded the filename
9522 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9523 a public key to be input or output. For example:
9524 openssl rsa -in key.pem -pubout -out pubkey.pem
9525 Also added necessary DSA public key functions to handle this.
9528 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9529 in the message. This was handled by allowing
9530 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9531 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9533 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9534 to the end of the strings whereas this didn't. This would cause problems
9535 if strings read with d2i_ASN1_bytes() were later modified.
9536 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9538 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9539 data and it contains EOF it will end up returning an error. This is
9540 caused by input 46 bytes long. The cause is due to the way base64
9541 BIOs find the start of base64 encoded data. They do this by trying a
9542 trial decode on each line until they find one that works. When they
9543 do a flag is set and it starts again knowing it can pass all the
9544 data directly through the decoder. Unfortunately it doesn't reset
9545 the context it uses. This means that if EOF is reached an attempt
9546 is made to pass two EOFs through the context and this causes the
9547 resulting error. This can also cause other problems as well. As is
9548 usual with these problems it takes *ages* to find and the fix is
9549 trivial: move one line.
9550 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9552 *) Ugly workaround to get s_client and s_server working under Windows. The
9553 old code wouldn't work because it needed to select() on sockets and the
9554 tty (for keypresses and to see if data could be written). Win32 only
9555 supports select() on sockets so we select() with a 1s timeout on the
9556 sockets and then see if any characters are waiting to be read, if none
9557 are present then we retry, we also assume we can always write data to
9558 the tty. This isn't nice because the code then blocks until we've
9559 received a complete line of data and it is effectively polling the
9560 keyboard at 1s intervals: however it's quite a bit better than not
9561 working at all :-) A dedicated Windows application might handle this
9562 with an event loop for example.
9565 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9566 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9567 will be called when RSA_sign() and RSA_verify() are used. This is useful
9568 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9569 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9570 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9571 This necessitated the support of an extra signature type NID_md5_sha1
9572 for SSL signatures and modifications to the SSL library to use it instead
9573 of calling RSA_public_decrypt() and RSA_private_encrypt().
9576 *) Add new -verify -CAfile and -CApath options to the crl program, these
9577 will lookup a CRL issuers certificate and verify the signature in a
9578 similar way to the verify program. Tidy up the crl program so it
9579 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9580 less strict. It will now permit CRL extensions even if it is not
9581 a V2 CRL: this will allow it to tolerate some broken CRLs.
9584 *) Initialize all non-automatic variables each time one of the openssl
9585 sub-programs is started (this is necessary as they may be started
9586 multiple times from the "OpenSSL>" prompt).
9587 [Lennart Bang, Bodo Moeller]
9589 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9590 removing all other RSA functionality (this is what NO_RSA does). This
9591 is so (for example) those in the US can disable those operations covered
9592 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9596 *) Non-copying interface to BIO pairs.
9597 (still largely untested)
9600 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9601 ASCII string. This was handled independently in various places before.
9604 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9605 UTF8 strings a character at a time.
9608 *) Use client_version from client hello to select the protocol
9609 (s23_srvr.c) and for RSA client key exchange verification
9610 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9613 *) Add various utility functions to handle SPKACs, these were previously
9614 handled by poking round in the structure internals. Added new function
9615 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9616 print, verify and generate SPKACs. Based on an original idea from
9617 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9620 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9623 *) Allow the config file extension section to be overwritten on the
9624 command line. Based on an original idea from Massimiliano Pala
9625 <madwolf@comune.modena.it>. The new option is called -extensions
9626 and can be applied to ca, req and x509. Also -reqexts to override
9627 the request extensions in req and -crlexts to override the crl extensions
9631 *) Add new feature to the SPKAC handling in ca. Now you can include
9632 the same field multiple times by preceding it by "XXXX." for example:
9635 this is the same syntax as used in the req config file.
9638 *) Allow certificate extensions to be added to certificate requests. These
9639 are specified in a 'req_extensions' option of the req section of the
9640 config file. They can be printed out with the -text option to req but
9641 are otherwise ignored at present.
9644 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9645 data read consists of only the final block it would not decrypted because
9646 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9647 A misplaced 'break' also meant the decrypted final block might not be
9648 copied until the next read.
9651 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9652 a few extra parameters to the DH structure: these will be useful if
9653 for example we want the value of 'q' or implement X9.42 DH.
9656 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9657 provides hooks that allow the default DSA functions or functions on a
9658 "per key" basis to be replaced. This allows hardware acceleration and
9659 hardware key storage to be handled without major modification to the
9660 library. Also added low level modexp hooks and CRYPTO_EX structure and
9661 associated functions.
9664 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9665 as "read only": it can't be written to and the buffer it points to will
9666 not be freed. Reading from a read only BIO is much more efficient than
9667 a normal memory BIO. This was added because there are several times when
9668 an area of memory needs to be read from a BIO. The previous method was
9669 to create a memory BIO and write the data to it, this results in two
9670 copies of the data and an O(n^2) reading algorithm. There is a new
9671 function BIO_new_mem_buf() which creates a read only memory BIO from
9672 an area of memory. Also modified the PKCS#7 routines to use read only
9676 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9677 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9678 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9679 but a retry condition occured while trying to read the rest.
9682 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9683 NID_pkcs7_encrypted by default: this was wrong since this should almost
9684 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9685 the encrypted data type: this is a more sensible place to put it and it
9686 allows the PKCS#12 code to be tidied up that duplicated this
9690 *) Changed obj_dat.pl script so it takes its input and output files on
9691 the command line. This should avoid shell escape redirection problems
9695 *) Initial support for certificate extension requests, these are included
9696 in things like Xenroll certificate requests. Included functions to allow
9697 extensions to be obtained and added.
9700 *) -crlf option to s_client and s_server for sending newlines as
9701 CRLF (as required by many protocols).
9704 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9706 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9707 [Ralf S. Engelschall]
9709 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9710 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9712 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9716 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9717 DH parameters/keys (q is lost during that conversion, but the resulting
9718 DH parameters contain its length).
9720 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9721 much faster than DH_generate_parameters (which creates parameters
9722 where p = 2*q + 1), and also the smaller q makes DH computations
9723 much more efficient (160-bit exponentiation instead of 1024-bit
9724 exponentiation); so this provides a convenient way to support DHE
9725 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9726 utter importance to use
9727 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9729 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9730 when such DH parameters are used, because otherwise small subgroup
9731 attacks may become possible!
9734 *) Avoid memory leak in i2d_DHparams.
9737 *) Allow the -k option to be used more than once in the enc program:
9738 this allows the same encrypted message to be read by multiple recipients.
9741 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9742 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9743 it will always use the numerical form of the OID, even if it has a short
9747 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9748 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9749 otherwise bn_mod_exp was called. In the case of hardware keys for example
9750 no private key components need be present and it might store extra data
9751 in the RSA structure, which cannot be accessed from bn_mod_exp.
9752 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9753 private key operations.
9756 *) Added support for SPARC Linux.
9759 *) pem_password_cb function type incompatibly changed from
9760 typedef int pem_password_cb(char *buf, int size, int rwflag);
9762 ....(char *buf, int size, int rwflag, void *userdata);
9763 so that applications can pass data to their callbacks:
9764 The PEM[_ASN1]_{read,write}... functions and macros now take an
9765 additional void * argument, which is just handed through whenever
9766 the password callback is called.
9767 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9769 New function SSL_CTX_set_default_passwd_cb_userdata.
9771 Compatibility note: As many C implementations push function arguments
9772 onto the stack in reverse order, the new library version is likely to
9773 interoperate with programs that have been compiled with the old
9774 pem_password_cb definition (PEM_whatever takes some data that
9775 happens to be on the stack as its last argument, and the callback
9776 just ignores this garbage); but there is no guarantee whatsoever that
9779 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9780 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9781 problems not only on Windows, but also on some Unix platforms.
9782 To avoid problematic command lines, these definitions are now in an
9783 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9784 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9787 *) MIPS III/IV assembler module is reimplemented.
9790 *) More DES library cleanups: remove references to srand/rand and
9791 delete an unused file.
9794 *) Add support for the the free Netwide assembler (NASM) under Win32,
9795 since not many people have MASM (ml) and it can be hard to obtain.
9796 This is currently experimental but it seems to work OK and pass all
9797 the tests. Check out INSTALL.W32 for info.
9800 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9801 without temporary keys kept an extra copy of the server key,
9802 and connections with temporary keys did not free everything in case
9806 *) New function RSA_check_key and new openssl rsa option -check
9807 for verifying the consistency of RSA keys.
9808 [Ulf Moeller, Bodo Moeller]
9810 *) Various changes to make Win32 compile work:
9811 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9812 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9813 comparison" warnings.
9814 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9817 *) Add a debugging option to PKCS#5 v2 key generation function: when
9818 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9819 derived keys are printed to stderr.
9822 *) Copy the flags in ASN1_STRING_dup().
9823 [Roman E. Pavlov <pre@mo.msk.ru>]
9825 *) The x509 application mishandled signing requests containing DSA
9826 keys when the signing key was also DSA and the parameters didn't match.
9828 It was supposed to omit the parameters when they matched the signing key:
9829 the verifying software was then supposed to automatically use the CA's
9830 parameters if they were absent from the end user certificate.
9832 Omitting parameters is no longer recommended. The test was also
9833 the wrong way round! This was probably due to unusual behaviour in
9834 EVP_cmp_parameters() which returns 1 if the parameters match.
9835 This meant that parameters were omitted when they *didn't* match and
9836 the certificate was useless. Certificates signed with 'ca' didn't have
9838 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9840 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9841 The interface is as follows:
9842 Applications can use
9843 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9844 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9845 "off" is now the default.
9846 The library internally uses
9847 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9848 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9849 to disable memory-checking temporarily.
9851 Some inconsistent states that previously were possible (and were
9852 even the default) are now avoided.
9854 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9855 with each memory chunk allocated; this is occasionally more helpful
9856 than just having a counter.
9858 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9860 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9864 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9865 which largely parallels "options", but is for changing API behaviour,
9866 whereas "options" are about protocol behaviour.
9867 Initial "mode" flags are:
9869 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9870 a single record has been written.
9871 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9872 retries use the same buffer location.
9873 (But all of the contents must be
9877 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9880 *) Fix problems with no-hmac etc.
9881 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9883 *) New functions RSA_get_default_method(), RSA_set_method() and
9884 RSA_get_method(). These allows replacement of RSA_METHODs without having
9885 to mess around with the internals of an RSA structure.
9888 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9889 Also really enable memory leak checks in openssl.c and in some
9891 [Chad C. Mulligan, Bodo Moeller]
9893 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9894 up the length of negative integers. This has now been simplified to just
9895 store the length when it is first determined and use it later, rather
9896 than trying to keep track of where data is copied and updating it to
9898 [Steve Henson, reported by Brien Wheeler
9899 <bwheeler@authentica-security.com>]
9901 *) Add a new function PKCS7_signatureVerify. This allows the verification
9902 of a PKCS#7 signature but with the signing certificate passed to the
9903 function itself. This contrasts with PKCS7_dataVerify which assumes the
9904 certificate is present in the PKCS#7 structure. This isn't always the
9905 case: certificates can be omitted from a PKCS#7 structure and be
9906 distributed by "out of band" means (such as a certificate database).
9909 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9910 function prototypes in pem.h, also change util/mkdef.pl to add the
9911 necessary function names.
9914 *) mk1mf.pl (used by Windows builds) did not properly read the
9915 options set by Configure in the top level Makefile, and Configure
9916 was not even able to write more than one option correctly.
9917 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9920 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9921 file to be loaded from a BIO or FILE pointer. The BIO version will
9922 for example allow memory BIOs to contain config info.
9925 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9926 Whoever hopes to achieve shared-library compatibility across versions
9927 must use this, not the compile-time macro.
9928 (Exercise 0.9.4: Which is the minimum library version required by
9930 Note: All this applies only to multi-threaded programs, others don't
9934 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9935 through a BIO pair triggered the default case, i.e.
9936 SSLerr(...,SSL_R_UNKNOWN_STATE).
9939 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9940 can use the SSL library even if none of the specific BIOs is
9944 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9945 for the encoded length.
9946 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9948 *) Add initial documentation of the X509V3 functions.
9951 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9952 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9953 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9954 secure PKCS#8 private key format with a high iteration count.
9957 *) Fix determination of Perl interpreter: A perl or perl5
9958 _directory_ in $PATH was also accepted as the interpreter.
9959 [Ralf S. Engelschall]
9961 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9962 wrong with it but it was very old and did things like calling
9963 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9967 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9968 to use the new extension code.
9971 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9972 with macros. This should make it easier to change their form, add extra
9973 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9977 *) Add to configuration table a new entry that can specify an alternative
9978 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9979 according to Mark Crispin <MRC@Panda.COM>.
9983 *) DES CBC did not update the IV. Weird.
9986 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9987 Changing the behaviour of the former might break existing programs --
9988 where IV updating is needed, des_ncbc_encrypt can be used.
9991 *) When bntest is run from "make test" it drives bc to check its
9992 calculations, as well as internally checking them. If an internal check
9993 fails, it needs to cause bc to give a non-zero result or make test carries
9994 on without noticing the failure. Fixed.
9997 *) DES library cleanups.
10000 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
10001 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
10002 ciphers. NOTE: although the key derivation function has been verified
10003 against some published test vectors it has not been extensively tested
10004 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
10008 *) Instead of "mkdir -p", which is not fully portable, use new
10009 Perl script "util/mkdir-p.pl".
10012 *) Rewrite the way password based encryption (PBE) is handled. It used to
10013 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
10014 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
10015 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
10016 the 'parameter' field of the AlgorithmIdentifier is passed to the
10017 underlying key generation function so it must do its own ASN1 parsing.
10018 This has also changed the EVP_PBE_CipherInit() function which now has a
10019 'parameter' argument instead of literal salt and iteration count values
10020 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
10023 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
10024 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
10025 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
10026 KEY" because this clashed with PKCS#8 unencrypted string. Since this
10027 value was just used as a "magic string" and not used directly its
10028 value doesn't matter.
10031 *) Introduce some semblance of const correctness to BN. Shame C doesn't
10035 *) "linux-sparc64" configuration (ultrapenguin).
10036 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
10037 "linux-sparc" configuration.
10038 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
10040 *) config now generates no-xxx options for missing ciphers.
10043 *) Support the EBCDIC character set (work in progress).
10044 File ebcdic.c not yet included because it has a different license.
10045 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10047 *) Support BS2000/OSD-POSIX.
10048 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10050 *) Make callbacks for key generation use void * instead of char *.
10053 *) Make S/MIME samples compile (not yet tested).
10056 *) Additional typesafe stacks.
10059 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
10063 Changes between 0.9.3 and 0.9.3a [29 May 1999]
10065 *) New configuration variant "sco5-gcc".
10067 *) Updated some demos.
10068 [Sean O Riordain, Wade Scholine]
10070 *) Add missing BIO_free at exit of pkcs12 application.
10073 *) Fix memory leak in conf.c.
10076 *) Updates for Win32 to assembler version of MD5.
10079 *) Set #! path to perl in apps/der_chop to where we found it
10080 instead of using a fixed path.
10083 *) SHA library changes for irix64-mips4-cc.
10086 *) Improvements for VMS support.
10090 Changes between 0.9.2b and 0.9.3 [24 May 1999]
10092 *) Bignum library bug fix. IRIX 6 passes "make test" now!
10093 This also avoids the problems with SC4.2 and unpatched SC5.
10094 [Andy Polyakov <appro@fy.chalmers.se>]
10096 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
10097 These are required because of the typesafe stack would otherwise break
10098 existing code. If old code used a structure member which used to be STACK
10099 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
10100 sk_num or sk_value it would produce an error because the num, data members
10101 are not present in STACK_OF. Now it just produces a warning. sk_set
10102 replaces the old method of assigning a value to sk_value
10103 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
10104 that does this will no longer work (and should use sk_set instead) but
10105 this could be regarded as a "questionable" behaviour anyway.
10108 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
10109 correctly handle encrypted S/MIME data.
10112 *) Change type of various DES function arguments from des_cblock
10113 (which means, in function argument declarations, pointer to char)
10114 to des_cblock * (meaning pointer to array with 8 char elements),
10115 which allows the compiler to do more typechecking; it was like
10116 that back in SSLeay, but with lots of ugly casts.
10118 Introduce new type const_des_cblock.
10121 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
10122 problems: find RecipientInfo structure that matches recipient certificate
10123 and initialise the ASN1 structures properly based on passed cipher.
10126 *) Belatedly make the BN tests actually check the results.
10129 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
10130 to and from BNs: it was completely broken. New compilation option
10131 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10132 key elements as negative integers.
10135 *) Reorganize and speed up MD5.
10136 [Andy Polyakov <appro@fy.chalmers.se>]
10139 [Richard Levitte <richard@levitte.org>]
10141 *) New option -out to asn1parse to allow the parsed structure to be
10142 output to a file. This is most useful when combined with the -strparse
10143 option to examine the output of things like OCTET STRINGS.
10146 *) Make SSL library a little more fool-proof by not requiring any longer
10147 that SSL_set_{accept,connect}_state be called before
10148 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10149 in many applications because usually everything *appeared* to work as
10150 intended anyway -- now it really works as intended).
10153 *) Move openssl.cnf out of lib/.
10156 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10157 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10158 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10159 [Ralf S. Engelschall]
10161 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10162 handle PKCS#7 enveloped data properly.
10163 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10165 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10166 copying pointers. The cert_st handling is changed by this in
10167 various ways (and thus what used to be known as ctx->default_cert
10168 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10169 any longer when s->cert does not give us what we need).
10170 ssl_cert_instantiate becomes obsolete by this change.
10171 As soon as we've got the new code right (possibly it already is?),
10172 we have solved a couple of bugs of the earlier code where s->cert
10173 was used as if it could not have been shared with other SSL structures.
10175 Note that using the SSL API in certain dirty ways now will result
10176 in different behaviour than observed with earlier library versions:
10177 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10178 does not influence s as it used to.
10180 In order to clean up things more thoroughly, inside SSL_SESSION
10181 we don't use CERT any longer, but a new structure SESS_CERT
10182 that holds per-session data (if available); currently, this is
10183 the peer's certificate chain and, for clients, the server's certificate
10184 and temporary key. CERT holds only those values that can have
10185 meaningful defaults in an SSL_CTX.
10188 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10189 from the internal representation. Various PKCS#7 fixes: remove some
10190 evil casts and set the enc_dig_alg field properly based on the signing
10194 *) Allow PKCS#12 password to be set from the command line or the
10195 environment. Let 'ca' get its config file name from the environment
10196 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10200 *) Allow certificate policies extension to use an IA5STRING for the
10201 organization field. This is contrary to the PKIX definition but
10202 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10206 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10207 without disallowing inline assembler and the like for non-pedantic builds.
10210 *) Support Borland C++ builder.
10211 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10213 *) Support Mingw32.
10216 *) SHA-1 cleanups and performance enhancements.
10217 [Andy Polyakov <appro@fy.chalmers.se>]
10219 *) Sparc v8plus assembler for the bignum library.
10220 [Andy Polyakov <appro@fy.chalmers.se>]
10222 *) Accept any -xxx and +xxx compiler options in Configure.
10225 *) Update HPUX configuration.
10228 *) Add missing sk_<type>_unshift() function to safestack.h
10229 [Ralf S. Engelschall]
10231 *) New function SSL_CTX_use_certificate_chain_file that sets the
10232 "extra_cert"s in addition to the certificate. (This makes sense
10233 only for "PEM" format files, as chains as a whole are not
10237 *) Support verify_depth from the SSL API.
10238 x509_vfy.c had what can be considered an off-by-one-error:
10239 Its depth (which was not part of the external interface)
10240 was actually counting the number of certificates in a chain;
10241 now it really counts the depth.
10244 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10245 instead of X509err, which often resulted in confusing error
10246 messages since the error codes are not globally unique
10247 (e.g. an alleged error in ssl3_accept when a certificate
10248 didn't match the private key).
10250 *) New function SSL_CTX_set_session_id_context that allows to set a default
10251 value (so that you don't need SSL_set_session_id_context for each
10252 connection using the SSL_CTX).
10255 *) OAEP decoding bug fix.
10258 *) Support INSTALL_PREFIX for package builders, as proposed by
10262 *) New Configure options "threads" and "no-threads". For systems
10263 where the proper compiler options are known (currently Solaris
10264 and Linux), "threads" is the default.
10267 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10270 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10271 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10272 such as /usr/local/bin.
10275 *) "make linux-shared" to build shared libraries.
10276 [Niels Poppe <niels@netbox.org>]
10278 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10281 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10282 extension adding in x509 utility.
10285 *) Remove NOPROTO sections and error code comments.
10288 *) Partial rewrite of the DEF file generator to now parse the ANSI
10292 *) New Configure options --prefix=DIR and --openssldir=DIR.
10295 *) Complete rewrite of the error code script(s). It is all now handled
10296 by one script at the top level which handles error code gathering,
10297 header rewriting and C source file generation. It should be much better
10298 than the old method: it now uses a modified version of Ulf's parser to
10299 read the ANSI prototypes in all header files (thus the old K&R definitions
10300 aren't needed for error creation any more) and do a better job of
10301 translating function codes into names. The old 'ASN1 error code imbedded
10302 in a comment' is no longer necessary and it doesn't use .err files which
10303 have now been deleted. Also the error code call doesn't have to appear all
10304 on one line (which resulted in some large lines...).
10307 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10310 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10311 0 (which usually indicates a closed connection), but continue reading.
10314 *) Fix some race conditions.
10317 *) Add support for CRL distribution points extension. Add Certificate
10318 Policies and CRL distribution points documentation.
10321 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10324 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10325 8 of keying material. Merlin has also confirmed interop with this fix
10326 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10327 [Merlin Hughes <merlin@baltimore.ie>]
10329 *) Fix lots of warnings.
10330 [Richard Levitte <levitte@stacken.kth.se>]
10332 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10333 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10334 [Richard Levitte <levitte@stacken.kth.se>]
10336 *) Fix problems with sizeof(long) == 8.
10337 [Andy Polyakov <appro@fy.chalmers.se>]
10339 *) Change functions to ANSI C.
10342 *) Fix typos in error codes.
10343 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10345 *) Remove defunct assembler files from Configure.
10348 *) SPARC v8 assembler BIGNUM implementation.
10349 [Andy Polyakov <appro@fy.chalmers.se>]
10351 *) Support for Certificate Policies extension: both print and set.
10352 Various additions to support the r2i method this uses.
10355 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10356 return a const string when you are expecting an allocated buffer.
10359 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10360 types DirectoryString and DisplayText.
10363 *) Add code to allow r2i extensions to access the configuration database,
10364 add an LHASH database driver and add several ctx helper functions.
10367 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10368 fail when they extended the size of a BIGNUM.
10371 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10372 support typesafe stack.
10375 *) Fix typo in SSL_[gs]et_options().
10376 [Nils Frostberg <nils@medcom.se>]
10378 *) Delete various functions and files that belonged to the (now obsolete)
10379 old X509V3 handling code.
10382 *) New Configure option "rsaref".
10385 *) Don't auto-generate pem.h.
10388 *) Introduce type-safe ASN.1 SETs.
10391 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10392 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10394 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10395 that links with OpenSSL (well at least cause lots of warnings), but fear
10396 not: the conversion is trivial, and it eliminates loads of evil casts. A
10397 few STACKed things have been converted already. Feel free to convert more.
10398 In the fullness of time, I'll do away with the STACK type altogether.
10401 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10402 specified in <certfile> by updating the entry in the index.txt file.
10403 This way one no longer has to edit the index.txt file manually for
10404 revoking a certificate. The -revoke option does the gory details now.
10405 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10407 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10408 `-text' option at all and this way the `-noout -text' combination was
10409 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10410 [Ralf S. Engelschall]
10412 *) Make sure a corresponding plain text error message exists for the
10413 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10414 verify callback function determined that a certificate was revoked.
10415 [Ralf S. Engelschall]
10417 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10418 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10419 all available cipers including rc5, which was forgotten until now.
10420 In order to let the testing shell script know which algorithms
10421 are available, a new (up to now undocumented) command
10422 "openssl list-cipher-commands" is used.
10425 *) Bugfix: s_client occasionally would sleep in select() when
10426 it should have checked SSL_pending() first.
10429 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10430 the raw DSA values prior to ASN.1 encoding.
10433 *) Tweaks to Configure
10434 [Niels Poppe <niels@netbox.org>]
10436 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10440 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10443 *) New config option to avoid instructions that are illegal on the 80386.
10444 The default code is faster, but requires at least a 486.
10447 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10448 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10449 same as SSL2_VERSION anyway.
10452 *) New "-showcerts" option for s_client.
10455 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10456 application. Various cleanups and fixes.
10459 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10460 modify error routines to work internally. Add error codes and PBE init
10461 to library startup routines.
10464 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10465 packing functions to asn1 and evp. Changed function names and error
10466 codes along the way.
10469 *) PKCS12 integration: and so it begins... First of several patches to
10470 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10471 objects to objects.h
10474 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10475 and display support for Thawte strong extranet extension.
10478 *) Add LinuxPPC support.
10479 [Jeff Dubrule <igor@pobox.org>]
10481 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10482 bn_div_words in alpha.s.
10483 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10485 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10486 OAEP isn't supported when OpenSSL is built with RSAref.
10487 [Ulf Moeller <ulf@fitug.de>]
10489 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10490 so they no longer are missing under -DNOPROTO.
10491 [Soren S. Jorvang <soren@t.dk>]
10494 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10496 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10497 doesn't work when the session is reused. Coming soon!
10500 *) Fix a security hole, that allows sessions to be reused in the wrong
10501 context thus bypassing client cert protection! All software that uses
10502 client certs and session caches in multiple contexts NEEDS PATCHING to
10503 allow session reuse! A fuller solution is in the works.
10504 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10506 *) Some more source tree cleanups (removed obsolete files
10507 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10508 permission on "config" script to be executable) and a fix for the INSTALL
10510 [Ulf Moeller <ulf@fitug.de>]
10512 *) Remove some legacy and erroneous uses of malloc, free instead of
10514 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10516 *) Make rsa_oaep_test return non-zero on error.
10517 [Ulf Moeller <ulf@fitug.de>]
10519 *) Add support for native Solaris shared libraries. Configure
10520 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10521 if someone would make that last step automatic.
10522 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10524 *) ctx_size was not built with the right compiler during "make links". Fixed.
10527 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10528 except NULL ciphers". This means the default cipher list will no longer
10529 enable NULL ciphers. They need to be specifically enabled e.g. with
10530 the string "DEFAULT:eNULL".
10533 *) Fix to RSA private encryption routines: if p < q then it would
10534 occasionally produce an invalid result. This will only happen with
10535 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10538 *) Be less restrictive and allow also `perl util/perlpath.pl
10539 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10540 because this way one can also use an interpreter named `perl5' (which is
10541 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10542 installed as `perl').
10543 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10545 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10546 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10548 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10549 advapi32.lib to Win32 build and change the pem test comparison
10550 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10551 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10552 and crypto/des/ede_cbcm_enc.c.
10555 *) DES quad checksum was broken on big-endian architectures. Fixed.
10558 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10559 Win32 test batch file so it (might) work again. The Win32 test batch file
10560 is horrible: I feel ill....
10563 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10564 in e_os.h. Audit of header files to check ANSI and non ANSI
10565 sections: 10 functions were absent from non ANSI section and not exported
10566 from Windows DLLs. Fixed up libeay.num for new functions.
10569 *) Make `openssl version' output lines consistent.
10570 [Ralf S. Engelschall]
10572 *) Fix Win32 symbol export lists for BIO functions: Added
10573 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10574 to ms/libeay{16,32}.def.
10575 [Ralf S. Engelschall]
10577 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10578 fine under Unix and passes some trivial tests I've now added. But the
10579 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10580 added to make sure no one expects that this stuff really works in the
10581 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10582 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10584 [Ralf S. Engelschall]
10586 *) Fix the generation of two part addresses in perl.
10587 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10589 *) Add config entry for Linux on MIPS.
10590 [John Tobey <jtobey@channel1.com>]
10592 *) Make links whenever Configure is run, unless we are on Windoze.
10595 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10596 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10600 *) Add a useful kludge to allow package maintainers to specify compiler and
10601 other platforms details on the command line without having to patch the
10602 Configure script everytime: One now can use ``perl Configure
10603 <id>:<details>'', i.e. platform ids are allowed to have details appended
10604 to them (separated by colons). This is treated as there would be a static
10605 pre-configured entry in Configure's %table under key <id> with value
10606 <details> and ``perl Configure <id>'' is called. So, when you want to
10607 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10608 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10609 now, which overrides the FreeBSD-elf entry on-the-fly.
10610 [Ralf S. Engelschall]
10612 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10615 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10616 on the `perl Configure ...' command line. This way one can compile
10617 OpenSSL libraries with Position Independent Code (PIC) which is needed
10618 for linking it into DSOs.
10619 [Ralf S. Engelschall]
10621 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10625 *) Cleaned up the LICENSE document: The official contact for any license
10626 questions now is the OpenSSL core team under openssl-core@openssl.org.
10627 And add a paragraph about the dual-license situation to make sure people
10628 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10629 to the OpenSSL toolkit.
10630 [Ralf S. Engelschall]
10632 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10633 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10634 Additionally cleaned up the `make links' target: Remove unnecessary
10635 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10636 to speed processing and no longer clutter the display with confusing
10637 stuff. Instead only the actually done links are displayed.
10638 [Ralf S. Engelschall]
10640 *) Permit null encryption ciphersuites, used for authentication only. It used
10641 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10642 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10646 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10647 signed attributes when verifying signatures (this would break them),
10648 the detached data encoding was wrong and public keys obtained using
10649 X509_get_pubkey() weren't freed.
10652 *) Add text documentation for the BUFFER functions. Also added a work around
10653 to a Win95 console bug. This was triggered by the password read stuff: the
10654 last character typed gets carried over to the next fread(). If you were
10655 generating a new cert request using 'req' for example then the last
10656 character of the passphrase would be CR which would then enter the first
10660 *) Added the new `Includes OpenSSL Cryptography Software' button as
10661 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10662 button and can be used by applications based on OpenSSL to show the
10663 relationship to the OpenSSL project.
10664 [Ralf S. Engelschall]
10666 *) Remove confusing variables in function signatures in files
10667 ssl/ssl_lib.c and ssl/ssl.h.
10668 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10670 *) Don't install bss_file.c under PREFIX/include/
10671 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10673 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10674 functions that return function pointers and has support for NT specific
10675 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10676 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10677 unsigned to signed types: this was killing the Win32 compile.
10680 *) Add new certificate file to stack functions,
10681 SSL_add_dir_cert_subjects_to_stack() and
10682 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10683 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10684 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10685 This means that Apache-SSL and similar packages don't have to mess around
10686 to add as many CAs as they want to the preferred list.
10689 *) Experiment with doxygen documentation. Currently only partially applied to
10691 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10692 openssl.doxy as the configuration file.
10695 *) Get rid of remaining C++-style comments which strict C compilers hate.
10696 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10698 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10699 compiled in by default: it has problems with large keys.
10702 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10703 DH private keys and/or callback functions which directly correspond to
10704 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10705 is needed for applications which have to configure certificates on a
10706 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10708 For the RSA certificate situation is makes no difference, but
10709 for the DSA certificate situation this fixes the "no shared cipher"
10710 problem where the OpenSSL cipher selection procedure failed because the
10711 temporary keys were not overtaken from the context and the API provided
10712 no way to reconfigure them.
10713 The new functions now let applications reconfigure the stuff and they
10714 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10715 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10716 non-public-API function ssl_cert_instantiate() is used as a helper
10717 function and also to reduce code redundancy inside ssl_rsa.c.
10718 [Ralf S. Engelschall]
10720 *) Move s_server -dcert and -dkey options out of the undocumented feature
10721 area because they are useful for the DSA situation and should be
10722 recognized by the users.
10723 [Ralf S. Engelschall]
10725 *) Fix the cipher decision scheme for export ciphers: the export bits are
10726 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10727 SSL_EXP_MASK. So, the original variable has to be used instead of the
10728 already masked variable.
10729 [Richard Levitte <levitte@stacken.kth.se>]
10731 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10732 [Richard Levitte <levitte@stacken.kth.se>]
10734 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10735 from `int' to `unsigned int' because it's a length and initialized by
10736 EVP_DigestFinal() which expects an `unsigned int *'.
10737 [Richard Levitte <levitte@stacken.kth.se>]
10739 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10740 script. Instead use the usual Shell->Perl transition trick.
10741 [Ralf S. Engelschall]
10743 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10744 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10745 -noout -modulus' as it's already the case for `openssl rsa -noout
10746 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10747 currently the public key is printed (a decision which was already done by
10748 `openssl dsa -modulus' in the past) which serves a similar purpose.
10749 Additionally the NO_RSA no longer completely removes the whole -modulus
10750 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10752 [Ralf S. Engelschall]
10754 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10755 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10756 [Arne Ansper <arne@ats.cyber.ee>]
10758 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10759 to be added. Now both 'req' and 'ca' can use new objects defined in the
10763 *) Add cool BIO that does syslog (or event log on NT).
10764 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10766 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10767 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10768 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10769 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10772 *) Add preliminary config info for new extension code.
10775 *) Make RSA_NO_PADDING really use no padding.
10776 [Ulf Moeller <ulf@fitug.de>]
10778 *) Generate errors when private/public key check is done.
10781 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10782 for some CRL extensions and new objects added.
10785 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10786 key usage extension and fuller support for authority key id.
10789 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10790 padding method for RSA, which is recommended for new applications in PKCS
10791 #1 v2.0 (RFC 2437, October 1998).
10792 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10793 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10794 against Bleichbacher's attack on RSA.
10795 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10798 *) Updates to the new SSL compression code
10799 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10801 *) Fix so that the version number in the master secret, when passed
10802 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10803 (because the server will not accept higher), that the version number
10804 is 0x03,0x01, not 0x03,0x00
10805 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10807 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10808 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10809 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10812 *) Support for RAW extensions where an arbitrary extension can be
10813 created by including its DER encoding. See apps/openssl.cnf for
10817 *) Make sure latest Perl versions don't interpret some generated C array
10818 code as Perl array code in the crypto/err/err_genc.pl script.
10819 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10821 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10822 not many people have the assembler. Various Win32 compilation fixes and
10823 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10824 build instructions.
10827 *) Modify configure script 'Configure' to automatically create crypto/date.h
10828 file under Win32 and also build pem.h from pem.org. New script
10829 util/mkfiles.pl to create the MINFO file on environments that can't do a
10830 'make files': perl util/mkfiles.pl >MINFO should work.
10833 *) Major rework of DES function declarations, in the pursuit of correctness
10834 and purity. As a result, many evil casts evaporated, and some weirdness,
10835 too. You may find this causes warnings in your code. Zapping your evil
10836 casts will probably fix them. Mostly.
10839 *) Fix for a typo in asn1.h. Bug fix to object creation script
10840 obj_dat.pl. It considered a zero in an object definition to mean
10841 "end of object": none of the objects in objects.h have any zeros
10842 so it wasn't spotted.
10843 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10845 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10846 Masking (CBCM). In the absence of test vectors, the best I have been able
10847 to do is check that the decrypt undoes the encrypt, so far. Send me test
10848 vectors if you have them.
10851 *) Correct calculation of key length for export ciphers (too much space was
10852 allocated for null ciphers). This has not been tested!
10855 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10856 message is now correct (it understands "crypto" and "ssl" on its
10857 command line). There is also now an "update" option. This will update
10858 the util/ssleay.num and util/libeay.num files with any new functions.
10860 perl util/mkdef.pl crypto ssl update
10861 it will update them.
10864 *) Overhauled the Perl interface (perl/*):
10865 - ported BN stuff to OpenSSL's different BN library
10866 - made the perl/ source tree CVS-aware
10867 - renamed the package from SSLeay to OpenSSL (the files still contain
10868 their history because I've copied them in the repository)
10869 - removed obsolete files (the test scripts will be replaced
10870 by better Test::Harness variants in the future)
10871 [Ralf S. Engelschall]
10873 *) First cut for a very conservative source tree cleanup:
10874 1. merge various obsolete readme texts into doc/ssleay.txt
10875 where we collect the old documents and readme texts.
10876 2. remove the first part of files where I'm already sure that we no
10877 longer need them because of three reasons: either they are just temporary
10878 files which were left by Eric or they are preserved original files where
10879 I've verified that the diff is also available in the CVS via "cvs diff
10880 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10881 the crypto/md/ stuff).
10882 [Ralf S. Engelschall]
10884 *) More extension code. Incomplete support for subject and issuer alt
10885 name, issuer and authority key id. Change the i2v function parameters
10886 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10887 what that's for :-) Fix to ASN1 macro which messed up
10888 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10891 *) Preliminary support for ENUMERATED type. This is largely copied from the
10895 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10896 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10898 *) Make sure `make rehash' target really finds the `openssl' program.
10899 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10901 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10902 like to hear about it if this slows down other processors.
10905 *) Add CygWin32 platform information to Configure script.
10906 [Alan Batie <batie@aahz.jf.intel.com>]
10908 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10909 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10911 *) New program nseq to manipulate netscape certificate sequences
10914 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10918 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10919 but the BN code had some problems that would cause failures when
10920 doing certificate verification and some other functions.
10921 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10923 *) Add ASN1 and PEM code to support netscape certificate sequences.
10926 *) Add ASN1 and PEM code to support netscape certificate sequences.
10929 *) Add several PKIX and private extended key usage OIDs.
10932 *) Modify the 'ca' program to handle the new extension code. Modify
10933 openssl.cnf for new extension format, add comments.
10936 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10937 and add a sample to openssl.cnf so req -x509 now adds appropriate
10941 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10942 error code, add initial support to X509_print() and x509 application.
10945 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10946 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10947 stuff is currently isolated and isn't even compiled yet.
10950 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10951 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10952 Removed the versions check from X509 routines when loading extensions:
10953 this allows certain broken certificates that don't set the version
10954 properly to be processed.
10957 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10958 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10959 can still be regenerated with "make depend".
10962 *) Spelling mistake in C version of CAST-128.
10963 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10965 *) Changes to the error generation code. The perl script err-code.pl
10966 now reads in the old error codes and retains the old numbers, only
10967 adding new ones if necessary. It also only changes the .err files if new
10968 codes are added. The makefiles have been modified to only insert errors
10969 when needed (to avoid needlessly modifying header files). This is done
10970 by only inserting errors if the .err file is newer than the auto generated
10971 C file. To rebuild all the error codes from scratch (the old behaviour)
10972 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10973 or delete all the .err files.
10976 *) CAST-128 was incorrectly implemented for short keys. The C version has
10977 been fixed, but is untested. The assembler versions are also fixed, but
10978 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10979 to regenerate it if needed.
10980 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10981 Hagino <itojun@kame.net>]
10983 *) File was opened incorrectly in randfile.c.
10984 [Ulf Möller <ulf@fitug.de>]
10986 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10987 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10988 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10989 al: it's just almost always a UTCTime. Note this patch adds new error
10990 codes so do a "make errors" if there are problems.
10993 *) Correct Linux 1 recognition in config.
10994 [Ulf Möller <ulf@fitug.de>]
10996 *) Remove pointless MD5 hash when using DSA keys in ca.
10997 [Anonymous <nobody@replay.com>]
10999 *) Generate an error if given an empty string as a cert directory. Also
11000 generate an error if handed NULL (previously returned 0 to indicate an
11001 error, but didn't set one).
11002 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
11004 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
11007 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
11008 parameters. This was causing a warning which killed off the Win32 compile.
11011 *) Remove C++ style comments from crypto/bn/bn_local.h.
11012 [Neil Costigan <neil.costigan@celocom.com>]
11014 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
11015 based on a text string, looking up short and long names and finally
11016 "dot" format. The "dot" format stuff didn't work. Added new function
11017 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
11018 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
11019 OID is not part of the table.
11022 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
11023 X509_LOOKUP_by_alias().
11026 *) Sort openssl functions by name.
11029 *) Get the gendsa program working (hopefully) and add it to app list. Remove
11030 encryption from sample DSA keys (in case anyone is interested the password
11034 *) Make _all_ *_free functions accept a NULL pointer.
11035 [Frans Heymans <fheymans@isaserver.be>]
11037 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
11039 [Anonymous <nobody@replay.com>]
11041 *) s_server should send the CAfile as acceptable CAs, not its own cert.
11042 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11044 *) Don't blow it for numeric -newkey arguments to apps/req.
11045 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
11047 *) Temp key "for export" tests were wrong in s3_srvr.c.
11048 [Anonymous <nobody@replay.com>]
11050 *) Add prototype for temp key callback functions
11051 SSL_CTX_set_tmp_{rsa,dh}_callback().
11054 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
11055 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
11058 *) X509_name_add_entry() freed the wrong thing after an error.
11059 [Arne Ansper <arne@ats.cyber.ee>]
11061 *) rsa_eay.c would attempt to free a NULL context.
11062 [Arne Ansper <arne@ats.cyber.ee>]
11064 *) BIO_s_socket() had a broken should_retry() on Windoze.
11065 [Arne Ansper <arne@ats.cyber.ee>]
11067 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
11068 [Arne Ansper <arne@ats.cyber.ee>]
11070 *) Make sure the already existing X509_STORE->depth variable is initialized
11071 in X509_STORE_new(), but document the fact that this variable is still
11072 unused in the certificate verification process.
11073 [Ralf S. Engelschall]
11075 *) Fix the various library and apps files to free up pkeys obtained from
11076 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
11079 *) Fix reference counting in X509_PUBKEY_get(). This makes
11080 demos/maurice/example2.c work, amongst others, probably.
11081 [Steve Henson and Ben Laurie]
11083 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
11084 `openssl' and second, the shortcut symlinks for the `openssl <command>'
11085 are no longer created. This way we have a single and consistent command
11086 line interface `openssl <command>', similar to `cvs <command>'.
11087 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
11089 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
11090 BIT STRING wrapper always have zero unused bits.
11093 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
11096 *) Make the top-level INSTALL documentation easier to understand.
11099 *) Makefiles updated to exit if an error occurs in a sub-directory
11100 make (including if user presses ^C) [Paul Sutton]
11102 *) Make Montgomery context stuff explicit in RSA data structure.
11105 *) Fix build order of pem and err to allow for generated pem.h.
11108 *) Fix renumbering bug in X509_NAME_delete_entry().
11111 *) Enhanced the err-ins.pl script so it makes the error library number
11112 global and can add a library name. This is needed for external ASN1 and
11113 other error libraries.
11116 *) Fixed sk_insert which never worked properly.
11119 *) Fix ASN1 macros so they can handle indefinite length constructed
11120 EXPLICIT tags. Some non standard certificates use these: they can now
11124 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
11125 into a single doc/ssleay.txt bundle. This way the information is still
11126 preserved but no longer messes up this directory. Now it's new room for
11127 the new set of documentation files.
11128 [Ralf S. Engelschall]
11130 *) SETs were incorrectly DER encoded. This was a major pain, because they
11131 shared code with SEQUENCEs, which aren't coded the same. This means that
11132 almost everything to do with SETs or SEQUENCEs has either changed name or
11133 number of arguments.
11134 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11136 *) Fix test data to work with the above.
11139 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11140 was already fixed by Eric for 0.9.1 it seems.
11141 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11143 *) Autodetect FreeBSD3.
11146 *) Fix various bugs in Configure. This affects the following platforms:
11150 unixware-2.0-pentium
11154 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11155 before they are needed.
11158 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11162 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11164 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11165 changed SSLeay to OpenSSL in version strings.
11166 [Ralf S. Engelschall]
11168 *) Some fixups to the top-level documents.
11171 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11172 because the symlink to include/ was missing.
11173 [Ralf S. Engelschall]
11175 *) Incorporated the popular no-RSA/DSA-only patches
11176 which allow to compile a RSA-free SSLeay.
11177 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11179 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11180 when "ssleay" is still not found.
11181 [Ralf S. Engelschall]
11183 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11184 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11186 *) Updated the README file.
11187 [Ralf S. Engelschall]
11189 *) Added various .cvsignore files in the CVS repository subdirs
11190 to make a "cvs update" really silent.
11191 [Ralf S. Engelschall]
11193 *) Recompiled the error-definition header files and added
11194 missing symbols to the Win32 linker tables.
11195 [Ralf S. Engelschall]
11197 *) Cleaned up the top-level documents;
11198 o new files: CHANGES and LICENSE
11199 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11200 o merged COPYRIGHT into LICENSE
11201 o removed obsolete TODO file
11202 o renamed MICROSOFT to INSTALL.W32
11203 [Ralf S. Engelschall]
11205 *) Removed dummy files from the 0.9.1b source tree:
11206 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11207 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11208 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11209 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11210 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11211 [Ralf S. Engelschall]
11213 *) Added various platform portability fixes.
11216 *) The Genesis of the OpenSSL rpject:
11217 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11218 Young and Tim J. Hudson created while they were working for C2Net until
11220 [The OpenSSL Project]
11223 Changes between 0.9.0b and 0.9.1b [not released]
11225 *) Updated a few CA certificates under certs/
11228 *) Changed some BIGNUM api stuff.
11231 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11232 DGUX x86, Linux Alpha, etc.
11235 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11236 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11240 *) Add -strparse option to asn1pars program which parses nested
11242 [Dr Stephen Henson <shenson@bigfoot.com>]
11244 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11247 *) DSA fix for "ca" program.
11250 *) Added "-genkey" option to "dsaparam" program.
11253 *) Added RIPE MD160 (rmd160) message digest.
11256 *) Added -a (all) option to "ssleay version" command.
11259 *) Added PLATFORM define which is the id given to Configure.
11262 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11265 *) Extended the ASN.1 parser routines.
11268 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11271 *) Added a BN_CTX to the BN library.
11274 *) Fixed the weak key values in DES library
11277 *) Changed API in EVP library for cipher aliases.
11280 *) Added support for RC2/64bit cipher.
11283 *) Converted the lhash library to the crypto/mem.c functions.
11286 *) Added more recognized ASN.1 object ids.
11289 *) Added more RSA padding checks for SSL/TLS.
11292 *) Added BIO proxy/filter functionality.
11295 *) Added extra_certs to SSL_CTX which can be used
11296 send extra CA certificates to the client in the CA cert chain sending
11297 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11300 *) Now Fortezza is denied in the authentication phase because
11301 this is key exchange mechanism is not supported by SSLeay at all.
11304 *) Additional PKCS1 checks.
11307 *) Support the string "TLSv1" for all TLS v1 ciphers.
11310 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11311 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11314 *) Fixed a few memory leaks.
11317 *) Fixed various code and comment typos.
11320 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11321 bytes sent in the client random.
11322 [Edward Bishop <ebishop@spyglass.com>]