5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
6 *) Given the pervasive nature of TLS extensions it is inadvisable to run
7 OpenSSL without support for them. It also means that maintaining
8 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
9 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
12 *) Version negotiation has been rewritten. In particular SSLv23_method(),
13 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
14 and turned into macros which simply call the new preferred function names
15 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
16 should use the new names instead. Also as part of this change the ssl23.h
17 header file has been removed.
20 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
21 code and the associated standard is no longer considered fit-for-purpose.
24 *) Added HTTP GET support to the ocsp command.
27 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
30 *) Added support for TLS extended master secret from
31 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
32 initial patch which was a great help during development.
35 *) All libssl internal structures have been removed from the public header
36 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
37 now redundant). Users should not attempt to access internal structures
38 directly. Instead they should use the provided API functions.
41 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
42 Access to deprecated functions can be re-enabled by running config with
43 "enable-deprecated". In addition applications wishing to use deprecated
44 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
45 will, by default, disable some transitive includes that previously existed
46 in the header files (e.g. ec.h will no longer, by default, include bn.h)
49 *) Added support for OCB mode. OpenSSL has been granted a patent license
50 compatible with the OpenSSL license for use of OCB. Details are available
51 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
52 for OCB can be removed by calling config with no-ocb.
55 *) SSLv2 support has been removed. It still supports receiving a SSLv2
56 compatible client hello.
59 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
60 done while fixing the error code for the key-too-small case.
61 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
63 *) CA.sh has been removmed; use CA.pl instead.
66 *) Removed old DES API.
69 *) Remove various unsupported platforms:
75 Sinix/ReliantUNIX RM400
80 16-bit platforms such as WIN16
83 *) Clean up OPENSSL_NO_xxx #define's
84 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
85 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
86 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
87 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
88 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
89 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
90 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
91 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
92 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
93 Remove MS_STATIC; it's a relic from platforms <32 bits.
96 *) Cleaned up dead code
97 Remove all but one '#ifdef undef' which is to be looked at.
100 *) Clean up calling of xxx_free routines.
101 Just like free(), fix most of the xxx_free routines to accept
102 NULL. Remove the non-null checks from callers. Save much code.
105 *) Experimental support for a new, fast, unbiased prime candidate generator,
106 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
107 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
109 *) New output format NSS in the sess_id command line tool. This allows
110 exporting the session id and the master key in NSS keylog format.
111 [Martin Kaiser <martin@kaiser.cx>]
113 *) Harmonize version and its documentation. -f flag is used to display
115 [mancha <mancha1@zoho.com>]
117 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
119 [mancha <mancha1@zoho.com>]
121 *) Fix some double frees. These are not thought to be exploitable.
122 [mancha <mancha1@zoho.com>]
124 *) A missing bounds check in the handling of the TLS heartbeat extension
125 can be used to reveal up to 64k of memory to a connected client or
128 Thanks for Neel Mehta of Google Security for discovering this bug and to
129 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
130 preparing the fix (CVE-2014-0160)
131 [Adam Langley, Bodo Moeller]
133 *) Fix for the attack described in the paper "Recovering OpenSSL
134 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
135 by Yuval Yarom and Naomi Benger. Details can be obtained from:
136 http://eprint.iacr.org/2014/140
138 Thanks to Yuval Yarom and Naomi Benger for discovering this
139 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
140 [Yuval Yarom and Naomi Benger]
142 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
143 this fixes a limitation in previous versions of OpenSSL.
146 *) Experimental encrypt-then-mac support.
148 Experimental support for encrypt then mac from
149 draft-gutmann-tls-encrypt-then-mac-02.txt
151 To enable it set the appropriate extension number (0x42 for the test
152 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
154 For non-compliant peers (i.e. just about everything) this should have no
157 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
161 *) Add EVP support for key wrapping algorithms, to avoid problems with
162 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
163 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
164 algorithms and include tests cases.
167 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
171 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
172 MGF1 digest and OAEP label.
175 *) Make openssl verify return errors.
176 [Chris Palmer <palmer@google.com> and Ben Laurie]
178 *) New function ASN1_TIME_diff to calculate the difference between two
179 ASN1_TIME structures or one structure and the current time.
182 *) Update fips_test_suite to support multiple command line options. New
183 test to induce all self test errors in sequence and check expected
187 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
188 sign or verify all in one operation.
191 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
192 test programs and fips_test_suite. Includes functionality to parse
193 the minimal script output of fipsalgest.pl directly.
196 *) Add authorisation parameter to FIPS_module_mode_set().
199 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
202 *) Use separate DRBG fields for internal and external flags. New function
203 FIPS_drbg_health_check() to perform on demand health checking. Add
204 generation tests to fips_test_suite with reduced health check interval to
205 demonstrate periodic health checking. Add "nodh" option to
206 fips_test_suite to skip very slow DH test.
209 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
213 *) More extensive health check for DRBG checking many more failure modes.
214 New function FIPS_selftest_drbg_all() to handle every possible DRBG
215 combination: call this in fips_test_suite.
218 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
219 and POST to handle Dual EC cases.
222 *) Add support for canonical generation of DSA parameter 'g'. See
225 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
226 POST to handle HMAC cases.
229 *) Add functions FIPS_module_version() and FIPS_module_version_text()
230 to return numerical and string versions of the FIPS module number.
233 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
234 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
235 outside the validated module in the FIPS capable OpenSSL.
238 *) Minor change to DRBG entropy callback semantics. In some cases
239 there is no multiple of the block length between min_len and
240 max_len. Allow the callback to return more than max_len bytes
241 of entropy but discard any extra: it is the callback's responsibility
242 to ensure that the extra data discarded does not impact the
243 requested amount of entropy.
246 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
247 information in FIPS186-3, SP800-57 and SP800-131A.
250 *) CCM support via EVP. Interface is very similar to GCM case except we
251 must supply all data in one chunk (i.e. no update, final) and the
252 message length must be supplied if AAD is used. Add algorithm test
256 *) Initial version of POST overhaul. Add POST callback to allow the status
257 of POST to be monitored and/or failures induced. Modify fips_test_suite
258 to use callback. Always run all selftests even if one fails.
261 *) XTS support including algorithm test driver in the fips_gcmtest program.
262 Note: this does increase the maximum key length from 32 to 64 bytes but
263 there should be no binary compatibility issues as existing applications
264 will never use XTS mode.
267 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
268 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
269 performs algorithm blocking for unapproved PRNG types. Also do not
270 set PRNG type in FIPS_mode_set(): leave this to the application.
271 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
272 the standard OpenSSL PRNG: set additional data to a date time vector.
275 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
276 This shouldn't present any incompatibility problems because applications
277 shouldn't be using these directly and any that are will need to rethink
278 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
281 *) Extensive self tests and health checking required by SP800-90 DRBG.
282 Remove strength parameter from FIPS_drbg_instantiate and always
283 instantiate at maximum supported strength.
286 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
289 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
292 *) New function DH_compute_key_padded() to compute a DH key and pad with
293 leading zeroes if needed: this complies with SP800-56A et al.
296 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
297 anything, incomplete, subject to change and largely untested at present.
300 *) Modify fipscanisteronly build option to only build the necessary object
301 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
304 *) Add experimental option FIPSSYMS to give all symbols in
305 fipscanister.o and FIPS or fips prefix. This will avoid
306 conflicts with future versions of OpenSSL. Add perl script
307 util/fipsas.pl to preprocess assembly language source files
308 and rename any affected symbols.
311 *) Add selftest checks and algorithm block of non-fips algorithms in
312 FIPS mode. Remove DES2 from selftests.
315 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
316 return internal method without any ENGINE dependencies. Add new
317 tiny fips sign and verify functions.
320 *) New build option no-ec2m to disable characteristic 2 code.
323 *) New build option "fipscanisteronly". This only builds fipscanister.o
324 and (currently) associated fips utilities. Uses the file Makefile.fips
325 instead of Makefile.org as the prototype.
328 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
329 Update fips_gcmtest to use IV generator.
332 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
333 setting output buffer to NULL. The *Final function must be
334 called although it will not retrieve any additional data. The tag
335 can be set or retrieved with a ctrl. The IV length is by default 12
336 bytes (96 bits) but can be set to an alternative value. If the IV
337 length exceeds the maximum IV length (currently 16 bytes) it cannot be
341 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
342 underlying do_cipher function handles all cipher semantics itself
343 including padding and finalisation. This is useful if (for example)
344 an ENGINE cipher handles block padding itself. The behaviour of
345 do_cipher is subtly changed if this flag is set: the return value
346 is the number of characters written to the output buffer (zero is
347 no longer an error code) or a negative error code. Also if the
348 input buffer is NULL and length 0 finalisation should be performed.
351 *) If a candidate issuer certificate is already part of the constructed
352 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
355 *) Improve forward-security support: add functions
357 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
358 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
360 for use by SSL/TLS servers; the callback function will be called whenever a
361 new session is created, and gets to decide whether the session may be
362 cached to make it resumable (return 0) or not (return 1). (As by the
363 SSL/TLS protocol specifications, the session_id sent by the server will be
364 empty to indicate that the session is not resumable; also, the server will
365 not generate RFC 4507 (RFC 5077) session tickets.)
367 A simple reasonable callback implementation is to return is_forward_secure.
368 This parameter will be set to 1 or 0 depending on the ciphersuite selected
369 by the SSL/TLS server library, indicating whether it can provide forward
371 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
373 *) New -verify_name option in command line utilities to set verification
377 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
378 Add CMAC pkey methods.
381 *) Experimental renegotiation in s_server -www mode. If the client
382 browses /reneg connection is renegotiated. If /renegcert it is
383 renegotiated requesting a certificate.
386 *) Add an "external" session cache for debugging purposes to s_server. This
387 should help trace issues which normally are only apparent in deployed
388 multi-process servers.
391 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
392 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
393 BIO_set_cipher() and some obscure PEM functions were changed so they
394 can now return an error. The RAND changes required a change to the
395 RAND_METHOD structure.
398 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
399 a gcc attribute to warn if the result of a function is ignored. This
400 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
401 whose return value is often ignored.
404 Changes between 1.0.2a and 1.0.2b [xx XXX xxxx]
406 *) Only support 256-bit or stronger elliptic curves with the
407 'ecdh_auto' setting (server) or by default (client). Of supported
408 curves, prefer P-256 (both).
411 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
413 *) ClientHello sigalgs DoS fix
415 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
416 invalid signature algorithms extension a NULL pointer dereference will
417 occur. This can be exploited in a DoS attack against the server.
419 This issue was was reported to OpenSSL by David Ramos of Stanford
422 [Stephen Henson and Matt Caswell]
424 *) Multiblock corrupted pointer fix
426 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
427 feature only applies on 64 bit x86 architecture platforms that support AES
428 NI instructions. A defect in the implementation of "multiblock" can cause
429 OpenSSL's internal write buffer to become incorrectly set to NULL when
430 using non-blocking IO. Typically, when the user application is using a
431 socket BIO for writing, this will only result in a failed connection.
432 However if some other BIO is used then it is likely that a segmentation
433 fault will be triggered, thus enabling a potential DoS attack.
435 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
439 *) Segmentation fault in DTLSv1_listen fix
441 The DTLSv1_listen function is intended to be stateless and processes the
442 initial ClientHello from many peers. It is common for user code to loop
443 over the call to DTLSv1_listen until a valid ClientHello is received with
444 an associated cookie. A defect in the implementation of DTLSv1_listen means
445 that state is preserved in the SSL object from one invocation to the next
446 that can lead to a segmentation fault. Errors processing the initial
447 ClientHello can trigger this scenario. An example of such an error could be
448 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
451 This issue was reported to OpenSSL by Per Allansson.
455 *) Segmentation fault in ASN1_TYPE_cmp fix
457 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
458 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
459 certificate signature algorithm consistency this can be used to crash any
460 certificate verification operation and exploited in a DoS attack. Any
461 application which performs certificate verification is vulnerable including
462 OpenSSL clients and servers which enable client authentication.
466 *) Segmentation fault for invalid PSS parameters fix
468 The signature verification routines will crash with a NULL pointer
469 dereference if presented with an ASN.1 signature using the RSA PSS
470 algorithm and invalid parameters. Since these routines are used to verify
471 certificate signature algorithms this can be used to crash any
472 certificate verification operation and exploited in a DoS attack. Any
473 application which performs certificate verification is vulnerable including
474 OpenSSL clients and servers which enable client authentication.
476 This issue was was reported to OpenSSL by Brian Carpenter.
480 *) ASN.1 structure reuse memory corruption fix
482 Reusing a structure in ASN.1 parsing may allow an attacker to cause
483 memory corruption via an invalid write. Such reuse is and has been
484 strongly discouraged and is believed to be rare.
486 Applications that parse structures containing CHOICE or ANY DEFINED BY
487 components may be affected. Certificate parsing (d2i_X509 and related
488 functions) are however not affected. OpenSSL clients and servers are
493 *) PKCS7 NULL pointer dereferences fix
495 The PKCS#7 parsing code does not handle missing outer ContentInfo
496 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
497 missing content and trigger a NULL pointer dereference on parsing.
499 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
500 otherwise parse PKCS#7 structures from untrusted sources are
501 affected. OpenSSL clients and servers are not affected.
503 This issue was reported to OpenSSL by Michal Zalewski (Google).
507 *) DoS via reachable assert in SSLv2 servers fix
509 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
510 servers that both support SSLv2 and enable export cipher suites by sending
511 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
513 This issue was discovered by Sean Burford (Google) and Emilia Käsper
514 (OpenSSL development team).
518 *) Empty CKE with client auth and DHE fix
520 If client auth is used then a server can seg fault in the event of a DHE
521 ciphersuite being selected and a zero length ClientKeyExchange message
522 being sent by the client. This could be exploited in a DoS attack.
526 *) Handshake with unseeded PRNG fix
528 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
529 with an unseeded PRNG. The conditions are:
530 - The client is on a platform where the PRNG has not been seeded
531 automatically, and the user has not seeded manually
532 - A protocol specific client method version has been used (i.e. not
533 SSL_client_methodv23)
534 - A ciphersuite is used that does not require additional random data from
535 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
537 If the handshake succeeds then the client random that has been used will
538 have been generated from a PRNG with insufficient entropy and therefore the
539 output may be predictable.
541 For example using the following command with an unseeded openssl will
542 succeed on an unpatched platform:
544 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
548 *) Use After Free following d2i_ECPrivatekey error fix
550 A malformed EC private key file consumed via the d2i_ECPrivateKey function
551 could cause a use after free condition. This, in turn, could cause a double
552 free in several private key parsing functions (such as d2i_PrivateKey
553 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
554 for applications that receive EC private keys from untrusted
555 sources. This scenario is considered rare.
557 This issue was discovered by the BoringSSL project and fixed in their
562 *) X509_to_X509_REQ NULL pointer deref fix
564 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
565 the certificate key is invalid. This function is rarely used in practice.
567 This issue was discovered by Brian Carpenter.
571 *) Removed the export ciphers from the DEFAULT ciphers
574 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
576 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
577 ARMv5 through ARMv8, as opposite to "locking" it to single one.
578 So far those who have to target multiple plaforms would compromise
579 and argue that binary targeting say ARMv5 would still execute on
580 ARMv8. "Universal" build resolves this compromise by providing
581 near-optimal performance even on newer platforms.
584 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
585 (other platforms pending).
586 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
588 *) Add support for the SignedCertificateTimestampList certificate and
589 OCSP response extensions from RFC6962.
592 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
593 for corner cases. (Certain input points at infinity could lead to
594 bogus results, with non-infinity inputs mapped to infinity too.)
597 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
598 This covers AES, SHA256/512 and GHASH. "Initial" means that most
599 common cases are optimized and there still is room for further
600 improvements. Vector Permutation AES for Altivec is also added.
603 *) Add support for little-endian ppc64 Linux target.
604 [Marcelo Cerri (IBM)]
606 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
607 SHA1, SHA256 and GHASH. "Initial" means that most common cases
608 are optimized and there still is room for further improvements.
609 Both 32- and 64-bit modes are supported.
610 [Andy Polyakov, Ard Biesheuvel (Linaro)]
612 *) Improved ARMv7 NEON support.
615 *) Support for SPARC Architecture 2011 crypto extensions, first
616 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
617 SHA256/512, MD5, GHASH and modular exponentiation.
618 [Andy Polyakov, David Miller]
620 *) Accelerated modular exponentiation for Intel processors, a.k.a.
622 [Shay Gueron & Vlad Krasnov (Intel Corp)]
624 *) Support for new and upcoming Intel processors, including AVX2,
625 BMI and SHA ISA extensions. This includes additional "stitched"
626 implementations, AESNI-SHA256 and GCM, and multi-buffer support
629 This work was sponsored by Intel Corp.
632 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
633 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
634 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
637 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
638 this fixes a limiation in previous versions of OpenSSL.
641 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
642 MGF1 digest and OAEP label.
645 *) Add EVP support for key wrapping algorithms, to avoid problems with
646 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
647 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
648 algorithms and include tests cases.
651 *) Add functions to allocate and set the fields of an ECDSA_METHOD
653 [Douglas E. Engert, Steve Henson]
655 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
656 difference in days and seconds between two tm or ASN1_TIME structures.
659 *) Add -rev test option to s_server to just reverse order of characters
660 received by client and send back to server. Also prints an abbreviated
661 summary of the connection parameters.
664 *) New option -brief for s_client and s_server to print out a brief summary
665 of connection parameters.
668 *) Add callbacks for arbitrary TLS extensions.
669 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
671 *) New option -crl_download in several openssl utilities to download CRLs
672 from CRLDP extension in certificates.
675 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
678 *) New function X509_CRL_diff to generate a delta CRL from the difference
679 of two full CRLs. Add support to "crl" utility.
682 *) New functions to set lookup_crls function and to retrieve
683 X509_STORE from X509_STORE_CTX.
686 *) Print out deprecated issuer and subject unique ID fields in
690 *) Extend OCSP I/O functions so they can be used for simple general purpose
691 HTTP as well as OCSP. New wrapper function which can be used to download
692 CRLs using the OCSP API.
695 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
698 *) SSL_CONF* functions. These provide a common framework for application
699 configuration using configuration files or command lines.
702 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
703 message callback and prints the results. Needs compile time option
704 "enable-ssl-trace". New options to s_client and s_server to enable
708 *) New ctrl and macro to retrieve supported points extensions.
709 Print out extension in s_server and s_client.
712 *) New functions to retrieve certificate signature and signature
716 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
720 *) New Suite B modes for TLS code. These use and enforce the requirements
721 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
722 only use Suite B curves. The Suite B modes can be set by using the
723 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
726 *) New chain verification flags for Suite B levels of security. Check
727 algorithms are acceptable when flags are set in X509_verify_cert.
730 *) Make tls1_check_chain return a set of flags indicating checks passed
731 by a certificate chain. Add additional tests to handle client
732 certificates: checks for matching certificate type and issuer name
736 *) If an attempt is made to use a signature algorithm not in the peer
737 preference list abort the handshake. If client has no suitable
738 signature algorithms in response to a certificate request do not
742 *) If server EC tmp key is not in client preference list abort handshake.
745 *) Add support for certificate stores in CERT structure. This makes it
746 possible to have different stores per SSL structure or one store in
747 the parent SSL_CTX. Include distint stores for certificate chain
748 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
749 to build and store a certificate chain in CERT structure: returing
750 an error if the chain cannot be built: this will allow applications
751 to test if a chain is correctly configured.
753 Note: if the CERT based stores are not set then the parent SSL_CTX
754 store is used to retain compatibility with existing behaviour.
758 *) New function ssl_set_client_disabled to set a ciphersuite disabled
759 mask based on the current session, check mask when sending client
760 hello and checking the requested ciphersuite.
763 *) New ctrls to retrieve and set certificate types in a certificate
764 request message. Print out received values in s_client. If certificate
765 types is not set with custom values set sensible values based on
766 supported signature algorithms.
769 *) Support for distinct client and server supported signature algorithms.
772 *) Add certificate callback. If set this is called whenever a certificate
773 is required by client or server. An application can decide which
774 certificate chain to present based on arbitrary criteria: for example
775 supported signature algorithms. Add very simple example to s_server.
776 This fixes many of the problems and restrictions of the existing client
777 certificate callback: for example you can now clear an existing
778 certificate and specify the whole chain.
781 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
782 the certificate can be used for (if anything). Set valid_flags field
783 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
784 to have similar checks in it.
786 Add new "cert_flags" field to CERT structure and include a "strict mode".
787 This enforces some TLS certificate requirements (such as only permitting
788 certificate signature algorithms contained in the supported algorithms
789 extension) which some implementations ignore: this option should be used
790 with caution as it could cause interoperability issues.
793 *) Update and tidy signature algorithm extension processing. Work out
794 shared signature algorithms based on preferences and peer algorithms
795 and print them out in s_client and s_server. Abort handshake if no
796 shared signature algorithms.
799 *) Add new functions to allow customised supported signature algorithms
800 for SSL and SSL_CTX structures. Add options to s_client and s_server
804 *) New function SSL_certs_clear() to delete all references to certificates
805 from an SSL structure. Before this once a certificate had been added
806 it couldn't be removed.
809 *) Integrate hostname, email address and IP address checking with certificate
810 verification. New verify options supporting checking in opensl utility.
813 *) Fixes and wildcard matching support to hostname and email checking
814 functions. Add manual page.
815 [Florian Weimer (Red Hat Product Security Team)]
817 *) New functions to check a hostname email or IP address against a
818 certificate. Add options x509 utility to print results of checks against
822 *) Fix OCSP checking.
823 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
825 *) Initial experimental support for explicitly trusted non-root CAs.
826 OpenSSL still tries to build a complete chain to a root but if an
827 intermediate CA has a trust setting included that is used. The first
828 setting is used: whether to trust (e.g., -addtrust option to the x509
832 *) Add -trusted_first option which attempts to find certificates in the
833 trusted store even if an untrusted chain is also supplied.
836 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
837 platform support for Linux and Android.
840 *) Support for linux-x32, ILP32 environment in x86_64 framework.
843 *) Experimental multi-implementation support for FIPS capable OpenSSL.
844 When in FIPS mode the approved implementations are used as normal,
845 when not in FIPS mode the internal unapproved versions are used instead.
846 This means that the FIPS capable OpenSSL isn't forced to use the
847 (often lower performance) FIPS implementations outside FIPS mode.
850 *) Transparently support X9.42 DH parameters when calling
851 PEM_read_bio_DHparameters. This means existing applications can handle
852 the new parameter format automatically.
855 *) Initial experimental support for X9.42 DH parameter format: mainly
856 to support use of 'q' parameter for RFC5114 parameters.
859 *) Add DH parameters from RFC5114 including test data to dhtest.
862 *) Support for automatic EC temporary key parameter selection. If enabled
863 the most preferred EC parameters are automatically used instead of
864 hardcoded fixed parameters. Now a server just has to call:
865 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
866 support ECDH and use the most appropriate parameters.
869 *) Enhance and tidy EC curve and point format TLS extension code. Use
870 static structures instead of allocation if default values are used.
871 New ctrls to set curves we wish to support and to retrieve shared curves.
872 Print out shared curves in s_server. New options to s_server and s_client
873 to set list of supported curves.
876 *) New ctrls to retrieve supported signature algorithms and
877 supported curve values as an array of NIDs. Extend openssl utility
878 to print out received values.
881 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
882 between NIDs and the more common NIST names such as "P-256". Enhance
883 ecparam utility and ECC method to recognise the NIST names for curves.
886 *) Enhance SSL/TLS certificate chain handling to support different
887 chains for each certificate instead of one chain in the parent SSL_CTX.
890 *) Support for fixed DH ciphersuite client authentication: where both
891 server and client use DH certificates with common parameters.
894 *) Support for fixed DH ciphersuites: those requiring DH server
898 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
900 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
901 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
902 X509_CINF_get_signature were reverted post internal team review.
904 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
906 *) Build fixes for the Windows and OpenVMS platforms
907 [Matt Caswell and Richard Levitte]
909 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
911 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
912 message can cause a segmentation fault in OpenSSL due to a NULL pointer
913 dereference. This could lead to a Denial Of Service attack. Thanks to
914 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
918 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
919 dtls1_buffer_record function under certain conditions. In particular this
920 could occur if an attacker sent repeated DTLS records with the same
921 sequence number but for the next epoch. The memory leak could be exploited
922 by an attacker in a Denial of Service attack through memory exhaustion.
923 Thanks to Chris Mueller for reporting this issue.
927 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
928 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
929 method would be set to NULL which could later result in a NULL pointer
930 dereference. Thanks to Frank Schmirler for reporting this issue.
934 *) Abort handshake if server key exchange message is omitted for ephemeral
937 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
938 reporting this issue.
942 *) Remove non-export ephemeral RSA code on client and server. This code
943 violated the TLS standard by allowing the use of temporary RSA keys in
944 non-export ciphersuites and could be used by a server to effectively
945 downgrade the RSA key length used to a value smaller than the server
946 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
947 INRIA or reporting this issue.
951 *) Fixed issue where DH client certificates are accepted without verification.
952 An OpenSSL server will accept a DH certificate for client authentication
953 without the certificate verify message. This effectively allows a client to
954 authenticate without the use of a private key. This only affects servers
955 which trust a client certificate authority which issues certificates
956 containing DH keys: these are extremely rare and hardly ever encountered.
957 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
962 *) Ensure that the session ID context of an SSL is updated when its
963 SSL_CTX is updated via SSL_set_SSL_CTX.
965 The session ID context is typically set from the parent SSL_CTX,
966 and can vary with the CTX.
969 *) Fix various certificate fingerprint issues.
971 By using non-DER or invalid encodings outside the signed portion of a
972 certificate the fingerprint can be changed without breaking the signature.
973 Although no details of the signed portion of the certificate can be changed
974 this can cause problems with some applications: e.g. those using the
975 certificate fingerprint for blacklists.
977 1. Reject signatures with non zero unused bits.
979 If the BIT STRING containing the signature has non zero unused bits reject
980 the signature. All current signature algorithms require zero unused bits.
982 2. Check certificate algorithm consistency.
984 Check the AlgorithmIdentifier inside TBS matches the one in the
985 certificate signature. NB: this will result in signature failure
986 errors for some broken certificates.
988 Thanks to Konrad Kraszewski from Google for reporting this issue.
990 3. Check DSA/ECDSA signatures use DER.
992 Reencode DSA/ECDSA signatures and compare with the original received
993 signature. Return an error if there is a mismatch.
995 This will reject various cases including garbage after signature
996 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
997 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
998 (negative or with leading zeroes).
1000 Further analysis was conducted and fixes were developed by Stephen Henson
1001 of the OpenSSL core team.
1006 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
1007 results on some platforms, including x86_64. This bug occurs at random
1008 with a very low probability, and is not known to be exploitable in any
1009 way, though its exact impact is difficult to determine. Thanks to Pieter
1010 Wuille (Blockstream) who reported this issue and also suggested an initial
1011 fix. Further analysis was conducted by the OpenSSL development team and
1012 Adam Langley of Google. The final fix was developed by Andy Polyakov of
1013 the OpenSSL core team.
1017 *) Do not resume sessions on the server if the negotiated protocol
1018 version does not match the session's version. Resuming with a different
1019 version, while not strictly forbidden by the RFC, is of questionable
1020 sanity and breaks all known clients.
1021 [David Benjamin, Emilia Käsper]
1023 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
1024 early CCS messages during renegotiation. (Note that because
1025 renegotiation is encrypted, this early CCS was not exploitable.)
1028 *) Tighten client-side session ticket handling during renegotiation:
1029 ensure that the client only accepts a session ticket if the server sends
1030 the extension anew in the ServerHello. Previously, a TLS client would
1031 reuse the old extension state and thus accept a session ticket if one was
1032 announced in the initial ServerHello.
1034 Similarly, ensure that the client requires a session ticket if one
1035 was advertised in the ServerHello. Previously, a TLS client would
1036 ignore a missing NewSessionTicket message.
1039 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1041 *) SRTP Memory Leak.
1043 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1044 sends a carefully crafted handshake message, to cause OpenSSL to fail
1045 to free up to 64k of memory causing a memory leak. This could be
1046 exploited in a Denial Of Service attack. This issue affects OpenSSL
1047 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1048 whether SRTP is used or configured. Implementations of OpenSSL that
1049 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1051 The fix was developed by the OpenSSL team.
1055 *) Session Ticket Memory Leak.
1057 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1058 integrity of that ticket is first verified. In the event of a session
1059 ticket integrity check failing, OpenSSL will fail to free memory
1060 causing a memory leak. By sending a large number of invalid session
1061 tickets an attacker could exploit this issue in a Denial Of Service
1066 *) Build option no-ssl3 is incomplete.
1068 When OpenSSL is configured with "no-ssl3" as a build option, servers
1069 could accept and complete a SSL 3.0 handshake, and clients could be
1070 configured to send them.
1072 [Akamai and the OpenSSL team]
1074 *) Add support for TLS_FALLBACK_SCSV.
1075 Client applications doing fallback retries should call
1076 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1078 [Adam Langley, Bodo Moeller]
1080 *) Add additional DigestInfo checks.
1082 Reencode DigestInto in DER and check against the original when
1083 verifying RSA signature: this will reject any improperly encoded
1084 DigestInfo structures.
1086 Note: this is a precautionary measure and no attacks are currently known.
1090 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1092 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1093 SRP code can be overrun an internal buffer. Add sanity check that
1094 g, A, B < N to SRP code.
1096 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1097 Group for discovering this issue.
1101 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1102 TLS 1.0 instead of higher protocol versions when the ClientHello message
1103 is badly fragmented. This allows a man-in-the-middle attacker to force a
1104 downgrade to TLS 1.0 even if both the server and the client support a
1105 higher protocol version, by modifying the client's TLS records.
1107 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1108 researching this issue.
1112 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1113 to a denial of service attack. A malicious server can crash the client
1114 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1115 ciphersuite and sending carefully crafted handshake messages.
1117 Thanks to Felix Gröbert (Google) for discovering and researching this
1122 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1123 to leak memory. This can be exploited through a Denial of Service attack.
1124 Thanks to Adam Langley for discovering and researching this issue.
1128 *) An attacker can force openssl to consume large amounts of memory whilst
1129 processing DTLS handshake messages. This can be exploited through a
1130 Denial of Service attack.
1131 Thanks to Adam Langley for discovering and researching this issue.
1135 *) An attacker can force an error condition which causes openssl to crash
1136 whilst processing DTLS packets due to memory being freed twice. This
1137 can be exploited through a Denial of Service attack.
1138 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1143 *) If a multithreaded client connects to a malicious server using a resumed
1144 session and the server sends an ec point format extension it could write
1145 up to 255 bytes to freed memory.
1147 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1152 *) A malicious server can crash an OpenSSL client with a null pointer
1153 dereference (read) by specifying an SRP ciphersuite even though it was not
1154 properly negotiated with the client. This can be exploited through a
1155 Denial of Service attack.
1157 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1158 discovering and researching this issue.
1162 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1163 X509_name_oneline, X509_name_print_ex et al. to leak some information
1164 from the stack. Applications may be affected if they echo pretty printing
1165 output to the attacker.
1167 Thanks to Ivan Fratric (Google) for discovering this issue.
1169 [Emilia Käsper, and Steve Henson]
1171 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1172 for corner cases. (Certain input points at infinity could lead to
1173 bogus results, with non-infinity inputs mapped to infinity too.)
1176 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1178 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1179 handshake can force the use of weak keying material in OpenSSL
1180 SSL/TLS clients and servers.
1182 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1183 researching this issue. (CVE-2014-0224)
1184 [KIKUCHI Masashi, Steve Henson]
1186 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1187 OpenSSL DTLS client the code can be made to recurse eventually crashing
1190 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1192 [Imre Rad, Steve Henson]
1194 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1195 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1196 client or server. This is potentially exploitable to run arbitrary
1197 code on a vulnerable client or server.
1199 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1200 [Jüri Aedla, Steve Henson]
1202 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1203 are subject to a denial of service attack.
1205 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1206 this issue. (CVE-2014-3470)
1207 [Felix Gröbert, Ivan Fratric, Steve Henson]
1209 *) Harmonize version and its documentation. -f flag is used to display
1211 [mancha <mancha1@zoho.com>]
1213 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1214 in i2d_ECPrivateKey.
1215 [mancha <mancha1@zoho.com>]
1217 *) Fix some double frees. These are not thought to be exploitable.
1218 [mancha <mancha1@zoho.com>]
1220 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1222 *) A missing bounds check in the handling of the TLS heartbeat extension
1223 can be used to reveal up to 64k of memory to a connected client or
1226 Thanks for Neel Mehta of Google Security for discovering this bug and to
1227 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1228 preparing the fix (CVE-2014-0160)
1229 [Adam Langley, Bodo Moeller]
1231 *) Fix for the attack described in the paper "Recovering OpenSSL
1232 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1233 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1234 http://eprint.iacr.org/2014/140
1236 Thanks to Yuval Yarom and Naomi Benger for discovering this
1237 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1238 [Yuval Yarom and Naomi Benger]
1240 *) TLS pad extension: draft-agl-tls-padding-03
1242 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1243 TLS client Hello record length value would otherwise be > 255 and
1244 less that 512 pad with a dummy extension containing zeroes so it
1245 is at least 512 bytes long.
1247 [Adam Langley, Steve Henson]
1249 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1251 *) Fix for TLS record tampering bug. A carefully crafted invalid
1252 handshake could crash OpenSSL with a NULL pointer exception.
1253 Thanks to Anton Johansson for reporting this issues.
1256 *) Keep original DTLS digest and encryption contexts in retransmission
1257 structures so we can use the previous session parameters if they need
1258 to be resent. (CVE-2013-6450)
1261 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1262 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1263 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1264 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1265 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1266 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1267 [Rob Stradling, Adam Langley]
1269 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1271 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1272 supporting platforms or when small records were transferred.
1273 [Andy Polyakov, Steve Henson]
1275 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1277 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1279 This addresses the flaw in CBC record processing discovered by
1280 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1281 at: http://www.isg.rhul.ac.uk/tls/
1283 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1284 Security Group at Royal Holloway, University of London
1285 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1286 Emilia Käsper for the initial patch.
1288 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1290 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1291 ciphersuites which can be exploited in a denial of service attack.
1292 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1293 and detecting this bug and to Wolfgang Ettlinger
1294 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1298 *) Return an error when checking OCSP signatures when key is NULL.
1299 This fixes a DoS attack. (CVE-2013-0166)
1302 *) Make openssl verify return errors.
1303 [Chris Palmer <palmer@google.com> and Ben Laurie]
1305 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1306 the right response is stapled. Also change SSL_get_certificate()
1307 so it returns the certificate actually sent.
1308 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1309 [Rob Stradling <rob.stradling@comodo.com>]
1311 *) Fix possible deadlock when decoding public keys.
1314 *) Don't use TLS 1.0 record version number in initial client hello
1318 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1320 *) Sanity check record length before skipping explicit IV in TLS
1321 1.2, 1.1 and DTLS to fix DoS attack.
1323 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1324 fuzzing as a service testing platform.
1328 *) Initialise tkeylen properly when encrypting CMS messages.
1329 Thanks to Solar Designer of Openwall for reporting this issue.
1332 *) In FIPS mode don't try to use composite ciphers as they are not
1336 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1338 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1339 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1340 mean any application compiled against OpenSSL 1.0.0 headers setting
1341 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1342 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1343 0x10000000L Any application which was previously compiled against
1344 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1345 will need to be recompiled as a result. Letting be results in
1346 inability to disable specifically TLS 1.1 and in client context,
1347 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1350 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1351 disable just protocol X, but all protocols above X *if* there are
1352 protocols *below* X still enabled. In more practical terms it means
1353 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1354 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1355 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1359 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1361 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1362 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1363 in CRYPTO_realloc_clean.
1365 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1366 issue and to Adam Langley <agl@chromium.org> for fixing it.
1368 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1370 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1373 *) Workarounds for some broken servers that "hang" if a client hello
1374 record length exceeds 255 bytes.
1376 1. Do not use record version number > TLS 1.0 in initial client
1377 hello: some (but not all) hanging servers will now work.
1378 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1379 the number of ciphers sent in the client hello. This should be
1380 set to an even number, such as 50, for example by passing:
1381 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1382 Most broken servers should now work.
1383 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1384 TLS 1.2 client support entirely.
1387 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1390 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1392 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1393 STRING form instead of a DigestInfo.
1396 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1397 and the RSA_sign/RSA_verify functions. This was made more apparent when
1398 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1399 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1400 the correct format in RSA_verify so both forms transparently work.
1403 *) Some servers which support TLS 1.0 can choke if we initially indicate
1404 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1405 encrypted premaster secret. As a workaround use the maximum permitted
1406 client version in client hello, this should keep such servers happy
1407 and still work with previous versions of OpenSSL.
1410 *) Add support for TLS/DTLS heartbeats.
1411 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1413 *) Add support for SCTP.
1414 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1416 *) Improved PRNG seeding for VOS.
1417 [Paul Green <Paul.Green@stratus.com>]
1419 *) Extensive assembler packs updates, most notably:
1421 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1422 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1423 - x86_64: bit-sliced AES implementation;
1424 - ARM: NEON support, contemporary platforms optimizations;
1425 - s390x: z196 support;
1426 - *: GHASH and GF(2^m) multiplication implementations;
1430 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1431 (removal of unnecessary code)
1432 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1434 *) Add TLS key material exporter from RFC 5705.
1437 *) Add DTLS-SRTP negotiation from RFC 5764.
1440 *) Add Next Protocol Negotiation,
1441 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1442 disabled with a no-npn flag to config or Configure. Code donated
1444 [Adam Langley <agl@google.com> and Ben Laurie]
1446 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1447 NIST-P256, NIST-P521, with constant-time single point multiplication on
1448 typical inputs. Compiler support for the nonstandard type __uint128_t is
1449 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1450 Code made available under Apache License version 2.0.
1452 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1453 line to include this in your build of OpenSSL, and run "make depend" (or
1454 "make update"). This enables the following EC_METHODs:
1456 EC_GFp_nistp224_method()
1457 EC_GFp_nistp256_method()
1458 EC_GFp_nistp521_method()
1460 EC_GROUP_new_by_curve_name() will automatically use these (while
1461 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1463 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1465 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1466 all platforms. Move ssize_t definition from e_os.h to the public
1467 header file e_os2.h as it now appears in public header file cms.h
1470 *) New -sigopt option to the ca, req and x509 utilities. Additional
1471 signature parameters can be passed using this option and in
1475 *) Add RSA PSS signing function. This will generate and set the
1476 appropriate AlgorithmIdentifiers for PSS based on those in the
1477 corresponding EVP_MD_CTX structure. No application support yet.
1480 *) Support for companion algorithm specific ASN1 signing routines.
1481 New function ASN1_item_sign_ctx() signs a pre-initialised
1482 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1483 the appropriate parameters.
1486 *) Add new algorithm specific ASN1 verification initialisation function
1487 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1488 handling will be the same no matter what EVP_PKEY_METHOD is used.
1489 Add a PSS handler to support verification of PSS signatures: checked
1490 against a number of sample certificates.
1493 *) Add signature printing for PSS. Add PSS OIDs.
1494 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1496 *) Add algorithm specific signature printing. An individual ASN1 method
1497 can now print out signatures instead of the standard hex dump.
1499 More complex signatures (e.g. PSS) can print out more meaningful
1500 information. Include DSA version that prints out the signature
1504 *) Password based recipient info support for CMS library: implementing
1508 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1509 neatly separates the code into cipher and PBE sections and is required
1510 for some algorithms that split PBES2 into separate pieces (such as
1511 password based CMS).
1514 *) Session-handling fixes:
1515 - Fix handling of connections that are resuming with a session ID,
1516 but also support Session Tickets.
1517 - Fix a bug that suppressed issuing of a new ticket if the client
1518 presented a ticket with an expired session.
1519 - Try to set the ticket lifetime hint to something reasonable.
1520 - Make tickets shorter by excluding irrelevant information.
1521 - On the client side, don't ignore renewed tickets.
1522 [Adam Langley, Bodo Moeller (Google)]
1524 *) Fix PSK session representation.
1527 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1529 This work was sponsored by Intel.
1532 *) Add GCM support to TLS library. Some custom code is needed to split
1533 the IV between the fixed (from PRF) and explicit (from TLS record)
1534 portions. This adds all GCM ciphersuites supported by RFC5288 and
1535 RFC5289. Generalise some AES* cipherstrings to include GCM and
1536 add a special AESGCM string for GCM only.
1539 *) Expand range of ctrls for AES GCM. Permit setting invocation
1540 field on decrypt and retrieval of invocation field only on encrypt.
1543 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1544 As required by RFC5289 these ciphersuites cannot be used if for
1545 versions of TLS earlier than 1.2.
1548 *) For FIPS capable OpenSSL interpret a NULL default public key method
1549 as unset and return the appropriate default but do *not* set the default.
1550 This means we can return the appropriate method in applications that
1551 switch between FIPS and non-FIPS modes.
1554 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1555 ENGINE is used then we cannot handle that in the FIPS module so we
1556 keep original code iff non-FIPS operations are allowed.
1559 *) Add -attime option to openssl utilities.
1560 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1562 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1565 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1566 FIPS EC methods unconditionally for now.
1569 *) New build option no-ec2m to disable characteristic 2 code.
1572 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1573 all cases can be covered as some introduce binary incompatibilities.
1576 *) Redirect RSA operations to FIPS module including keygen,
1577 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1580 *) Add similar low level API blocking to ciphers.
1583 *) Low level digest APIs are not approved in FIPS mode: any attempt
1584 to use these will cause a fatal error. Applications that *really* want
1585 to use them can use the private_* version instead.
1588 *) Redirect cipher operations to FIPS module for FIPS builds.
1591 *) Redirect digest operations to FIPS module for FIPS builds.
1594 *) Update build system to add "fips" flag which will link in fipscanister.o
1595 for static and shared library builds embedding a signature if needed.
1598 *) Output TLS supported curves in preference order instead of numerical
1599 order. This is currently hardcoded for the highest order curves first.
1600 This should be configurable so applications can judge speed vs strength.
1603 *) Add TLS v1.2 server support for client authentication.
1606 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1610 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1611 FIPS modules versions.
1614 *) Add TLS v1.2 client side support for client authentication. Keep cache
1615 of handshake records longer as we don't know the hash algorithm to use
1616 until after the certificate request message is received.
1619 *) Initial TLS v1.2 client support. Add a default signature algorithms
1620 extension including all the algorithms we support. Parse new signature
1621 format in client key exchange. Relax some ECC signing restrictions for
1622 TLS v1.2 as indicated in RFC5246.
1625 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1626 to new signature format when needed using client digest preference.
1627 All server ciphersuites should now work correctly in TLS v1.2. No client
1628 support yet and no support for client certificates.
1631 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1632 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1633 ciphersuites. At present only RSA key exchange ciphersuites work with
1634 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1635 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1636 and version checking.
1639 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1640 with this defined it will not be affected by any changes to ssl internal
1641 structures. Add several utility functions to allow openssl application
1642 to work with OPENSSL_NO_SSL_INTERN defined.
1646 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1648 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1651 *) Permit abbreviated handshakes when renegotiating using the function
1652 SSL_renegotiate_abbreviated().
1653 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1655 *) Add call to ENGINE_register_all_complete() to
1656 ENGINE_load_builtin_engines(), so some implementations get used
1657 automatically instead of needing explicit application support.
1660 *) Add support for TLS key exporter as described in RFC5705.
1661 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1663 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1664 a few changes are required:
1666 Add SSL_OP_NO_TLSv1_1 flag.
1667 Add TLSv1_1 methods.
1668 Update version checking logic to handle version 1.1.
1669 Add explicit IV handling (ported from DTLS code).
1670 Add command line options to s_client/s_server.
1673 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1675 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1676 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1677 content decryption and always return the same error. Note: this attack
1678 needs on average 2^20 messages so it only affects automated senders. The
1679 old behaviour can be reenabled in the CMS code by setting the
1680 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1681 an MMA defence is not necessary.
1682 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1683 this issue. (CVE-2012-0884)
1686 *) Fix CVE-2011-4619: make sure we really are receiving a
1687 client hello before rejecting multiple SGC restarts. Thanks to
1688 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1691 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1693 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1694 Thanks to Antonio Martin, Enterprise Secure Access Research and
1695 Development, Cisco Systems, Inc. for discovering this bug and
1696 preparing a fix. (CVE-2012-0050)
1699 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1701 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1702 of the Vaudenay padding oracle attack on CBC mode encryption
1703 which enables an efficient plaintext recovery attack against
1704 the OpenSSL implementation of DTLS. Their attack exploits timing
1705 differences arising during decryption processing. A research
1706 paper describing this attack can be found at:
1707 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1708 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1709 Security Group at Royal Holloway, University of London
1710 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1711 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1712 for preparing the fix. (CVE-2011-4108)
1713 [Robin Seggelmann, Michael Tuexen]
1715 *) Clear bytes used for block padding of SSL 3.0 records.
1717 [Adam Langley (Google)]
1719 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1720 Kadianakis <desnacked@gmail.com> for discovering this issue and
1721 Adam Langley for preparing the fix. (CVE-2011-4619)
1722 [Adam Langley (Google)]
1724 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1725 [Andrey Kulikov <amdeich@gmail.com>]
1727 *) Prevent malformed RFC3779 data triggering an assertion failure.
1728 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1729 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1730 [Rob Austein <sra@hactrn.net>]
1732 *) Improved PRNG seeding for VOS.
1733 [Paul Green <Paul.Green@stratus.com>]
1735 *) Fix ssl_ciph.c set-up race.
1736 [Adam Langley (Google)]
1738 *) Fix spurious failures in ecdsatest.c.
1739 [Emilia Käsper (Google)]
1741 *) Fix the BIO_f_buffer() implementation (which was mixing different
1742 interpretations of the '..._len' fields).
1743 [Adam Langley (Google)]
1745 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1746 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1747 threads won't reuse the same blinding coefficients.
1749 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1750 lock to call BN_BLINDING_invert_ex, and avoids one use of
1751 BN_BLINDING_update for each BN_BLINDING structure (previously,
1752 the last update always remained unused).
1753 [Emilia Käsper (Google)]
1755 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1756 [Bob Buckholz (Google)]
1758 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1760 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1761 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1762 [Kaspar Brand <ossl@velox.ch>]
1764 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1765 for multi-threaded use of ECDH. (CVE-2011-3210)
1766 [Adam Langley (Google)]
1768 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1771 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1772 signature public key algorithm by using OID xref utilities instead.
1773 Before this you could only use some ECC ciphersuites with SHA1 only.
1776 *) Add protection against ECDSA timing attacks as mentioned in the paper
1777 by Billy Bob Brumley and Nicola Tuveri, see:
1779 http://eprint.iacr.org/2011/232.pdf
1781 [Billy Bob Brumley and Nicola Tuveri]
1783 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1785 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1786 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1788 *) Fix bug in string printing code: if *any* escaping is enabled we must
1789 escape the escape character (backslash) or the resulting string is
1793 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1795 *) Disable code workaround for ancient and obsolete Netscape browsers
1796 and servers: an attacker can use it in a ciphersuite downgrade attack.
1797 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1800 *) Fixed J-PAKE implementation error, originally discovered by
1801 Sebastien Martini, further info and confirmation from Stefan
1802 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1805 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1807 *) Fix extension code to avoid race conditions which can result in a buffer
1808 overrun vulnerability: resumed sessions must not be modified as they can
1809 be shared by multiple threads. CVE-2010-3864
1812 *) Fix WIN32 build system to correctly link an ENGINE directory into
1816 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1818 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1820 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1822 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1824 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1825 context. The operation can be customised via the ctrl mechanism in
1826 case ENGINEs want to include additional functionality.
1829 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1832 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1833 output hashes compatible with older versions of OpenSSL.
1834 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1836 *) Fix compression algorithm handling: if resuming a session use the
1837 compression algorithm of the resumed session instead of determining
1838 it from client hello again. Don't allow server to change algorithm.
1841 *) Add load_crls() function to apps tidying load_certs() too. Add option
1842 to verify utility to allow additional CRLs to be included.
1845 *) Update OCSP request code to permit adding custom headers to the request:
1846 some responders need this.
1849 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1851 [Julia Lawall <julia@diku.dk>]
1853 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1854 needlessly dereferenced structures, used obsolete functions and
1855 didn't handle all updated verify codes correctly.
1858 *) Disable MD2 in the default configuration.
1861 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1862 indicate the initial BIO being pushed or popped. This makes it possible
1863 to determine whether the BIO is the one explicitly called or as a result
1864 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1865 it handles reference counts correctly and doesn't zero out the I/O bio
1866 when it is not being explicitly popped. WARNING: applications which
1867 included workarounds for the old buggy behaviour will need to be modified
1868 or they could free up already freed BIOs.
1871 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1872 renaming to all platforms (within the 0.9.8 branch, this was
1873 done conditionally on Netware platforms to avoid a name clash).
1874 [Guenter <lists@gknw.net>]
1876 *) Add ECDHE and PSK support to DTLS.
1877 [Michael Tuexen <tuexen@fh-muenster.de>]
1879 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1883 *) Add "missing" function EVP_MD_flags() (without this the only way to
1884 retrieve a digest flags is by accessing the structure directly. Update
1885 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1886 or cipher is registered as in the "from" argument. Print out all
1887 registered digests in the dgst usage message instead of manually
1888 attempting to work them out.
1891 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1892 this allows the use of compression and extensions. Change default cipher
1893 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1894 by default unless an application cipher string requests it.
1897 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1898 key ids to find matching certificates and keys but some PKCS#12 files
1899 don't follow the (somewhat unwritten) rules and this strategy fails.
1900 Now just gather all certificates together and the first private key
1901 then look for the first certificate that matches the key.
1904 *) Support use of registered digest and cipher names for dgst and cipher
1905 commands instead of having to add each one as a special case. So now
1912 openssl dgst -sha256 foo
1914 and this works for ENGINE based algorithms too.
1918 *) Update Gost ENGINE to support parameter files.
1919 [Victor B. Wagner <vitus@cryptocom.ru>]
1921 *) Support GeneralizedTime in ca utility.
1922 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1924 *) Enhance the hash format used for certificate directory links. The new
1925 form uses the canonical encoding (meaning equivalent names will work
1926 even if they aren't identical) and uses SHA1 instead of MD5. This form
1927 is incompatible with the older format and as a result c_rehash should
1928 be used to rebuild symbolic links.
1931 *) Make PKCS#8 the default write format for private keys, replacing the
1932 traditional format. This form is standardised, more secure and doesn't
1933 include an implicit MD5 dependency.
1936 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1937 committed to OpenSSL should pass this lot as a minimum.
1940 *) Add session ticket override functionality for use by EAP-FAST.
1941 [Jouni Malinen <j@w1.fi>]
1943 *) Modify HMAC functions to return a value. Since these can be implemented
1944 in an ENGINE errors can occur.
1947 *) Type-checked OBJ_bsearch_ex.
1950 *) Type-checked OBJ_bsearch. Also some constification necessitated
1951 by type-checking. Still to come: TXT_DB, bsearch(?),
1952 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1956 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1957 seconds to a tm structure directly, instead of going through OS
1958 specific date routines. This avoids any issues with OS routines such
1959 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1960 and X509_time_adj_ex() to cover the extended range. The existing
1961 X509_time_adj() is still usable and will no longer have any date issues.
1964 *) Delta CRL support. New use deltas option which will attempt to locate
1965 and search any appropriate delta CRLs available.
1967 This work was sponsored by Google.
1970 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1971 code and add additional score elements. Validate alternate CRL paths
1972 as part of the CRL checking and indicate a new error "CRL path validation
1973 error" in this case. Applications wanting additional details can use
1974 the verify callback and check the new "parent" field. If this is not
1975 NULL CRL path validation is taking place. Existing applications wont
1976 see this because it requires extended CRL support which is off by
1979 This work was sponsored by Google.
1982 *) Support for freshest CRL extension.
1984 This work was sponsored by Google.
1987 *) Initial indirect CRL support. Currently only supported in the CRLs
1988 passed directly and not via lookup. Process certificate issuer
1989 CRL entry extension and lookup CRL entries by bother issuer name
1990 and serial number. Check and process CRL issuer entry in IDP extension.
1992 This work was sponsored by Google.
1995 *) Add support for distinct certificate and CRL paths. The CRL issuer
1996 certificate is validated separately in this case. Only enabled if
1997 an extended CRL support flag is set: this flag will enable additional
1998 CRL functionality in future.
2000 This work was sponsored by Google.
2003 *) Add support for policy mappings extension.
2005 This work was sponsored by Google.
2008 *) Fixes to pathlength constraint, self issued certificate handling,
2009 policy processing to align with RFC3280 and PKITS tests.
2011 This work was sponsored by Google.
2014 *) Support for name constraints certificate extension. DN, email, DNS
2015 and URI types are currently supported.
2017 This work was sponsored by Google.
2020 *) To cater for systems that provide a pointer-based thread ID rather
2021 than numeric, deprecate the current numeric thread ID mechanism and
2022 replace it with a structure and associated callback type. This
2023 mechanism allows a numeric "hash" to be extracted from a thread ID in
2024 either case, and on platforms where pointers are larger than 'long',
2025 mixing is done to help ensure the numeric 'hash' is usable even if it
2026 can't be guaranteed unique. The default mechanism is to use "&errno"
2027 as a pointer-based thread ID to distinguish between threads.
2029 Applications that want to provide their own thread IDs should now use
2030 CRYPTO_THREADID_set_callback() to register a callback that will call
2031 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2033 Note that ERR_remove_state() is now deprecated, because it is tied
2034 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2035 to free the current thread's error state should be replaced by
2036 ERR_remove_thread_state(NULL).
2038 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2039 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2040 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2041 application was previously providing a numeric thread callback that
2042 was inappropriate for distinguishing threads, then uniqueness might
2043 have been obtained with &errno that happened immediately in the
2044 intermediate development versions of OpenSSL; this is no longer the
2045 case, the numeric thread callback will now override the automatic use
2047 [Geoff Thorpe, with help from Bodo Moeller]
2049 *) Initial support for different CRL issuing certificates. This covers a
2050 simple case where the self issued certificates in the chain exist and
2051 the real CRL issuer is higher in the existing chain.
2053 This work was sponsored by Google.
2056 *) Removed effectively defunct crypto/store from the build.
2059 *) Revamp of STACK to provide stronger type-checking. Still to come:
2060 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2061 ASN1_STRING, CONF_VALUE.
2064 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2065 RAM on SSL connections. This option can save about 34k per idle SSL.
2068 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2069 STACK, TXT_DB, bsearch, qsort.
2072 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2073 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2074 support for data, signedData, compressedData, digestedData and
2075 encryptedData, envelopedData types included. Scripts to check against
2076 RFC4134 examples draft and interop and consistency checks of many
2077 content types and variants.
2080 *) Add options to enc utility to support use of zlib compression BIO.
2083 *) Extend mk1mf to support importing of options and assembly language
2084 files from Configure script, currently only included in VC-WIN32.
2085 The assembly language rules can now optionally generate the source
2086 files from the associated perl scripts.
2089 *) Implement remaining functionality needed to support GOST ciphersuites.
2090 Interop testing has been performed using CryptoPro implementations.
2091 [Victor B. Wagner <vitus@cryptocom.ru>]
2093 *) s390x assembler pack.
2096 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2100 *) Implement Opaque PRF Input TLS extension as specified in
2101 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2102 official specification yet and no extension type assignment by
2103 IANA exists, this extension (for now) will have to be explicitly
2104 enabled when building OpenSSL by providing the extension number
2105 to use. For example, specify an option
2107 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2109 to the "config" or "Configure" script to enable the extension,
2110 assuming extension number 0x9527 (which is a completely arbitrary
2111 and unofficial assignment based on the MD5 hash of the Internet
2112 Draft). Note that by doing so, you potentially lose
2113 interoperability with other TLS implementations since these might
2114 be using the same extension number for other purposes.
2116 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2117 opaque PRF input value to use in the handshake. This will create
2118 an interal copy of the length-'len' string at 'src', and will
2119 return non-zero for success.
2121 To get more control and flexibility, provide a callback function
2124 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2125 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2129 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2132 Callback function 'cb' will be called in handshakes, and is
2133 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2134 Argument 'arg' is for application purposes (the value as given to
2135 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2136 be provided to the callback function). The callback function
2137 has to return non-zero to report success: usually 1 to use opaque
2138 PRF input just if possible, or 2 to enforce use of the opaque PRF
2139 input. In the latter case, the library will abort the handshake
2140 if opaque PRF input is not successfully negotiated.
2142 Arguments 'peerinput' and 'len' given to the callback function
2143 will always be NULL and 0 in the case of a client. A server will
2144 see the client's opaque PRF input through these variables if
2145 available (NULL and 0 otherwise). Note that if the server
2146 provides an opaque PRF input, the length must be the same as the
2147 length of the client's opaque PRF input.
2149 Note that the callback function will only be called when creating
2150 a new session (session resumption can resume whatever was
2151 previously negotiated), and will not be called in SSL 2.0
2152 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2153 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2154 for applications that need to enforce opaque PRF input.
2158 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2161 [Victor B. Wagner <vitus@cryptocom.ru>]
2163 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2164 RFC4507bis. The encrypted ticket format is an encrypted encoded
2165 SSL_SESSION structure, that way new session features are automatically
2168 If a client application caches session in an SSL_SESSION structure
2169 support is transparent because tickets are now stored in the encoded
2172 The SSL_CTX structure automatically generates keys for ticket
2173 protection in servers so again support should be possible
2174 with no application modification.
2176 If a client or server wishes to disable RFC4507 support then the option
2177 SSL_OP_NO_TICKET can be set.
2179 Add a TLS extension debugging callback to allow the contents of any client
2180 or server extensions to be examined.
2182 This work was sponsored by Google.
2185 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2186 OpenSSL should now compile cleanly on gcc 4.2
2187 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2189 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2190 support including streaming MAC support: this is required for GOST
2191 ciphersuite support.
2192 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2194 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2195 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2196 to output in BER and PEM format.
2199 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2200 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2201 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2202 ENGINE support for HMAC keys which are unextractable. New -mac and
2203 -macopt options to dgst utility.
2206 *) New option -sigopt to dgst utility. Update dgst to use
2207 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2208 alternative signing paramaters such as X9.31 or PSS in the dgst
2212 *) Change ssl_cipher_apply_rule(), the internal function that does
2213 the work each time a ciphersuite string requests enabling
2214 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2215 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2216 the order of disabled ciphersuites such that those ciphersuites
2217 that most recently went from enabled to disabled not only stay
2218 in order with respect to each other, but also have higher priority
2219 than other disabled ciphersuites the next time ciphersuites are
2222 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2223 the same ciphersuites as with "HIGH" alone, but in a specific
2224 order where the PSK ciphersuites come first (since they are the
2225 most recently disabled ciphersuites when "HIGH" is parsed).
2227 Also, change ssl_create_cipher_list() (using this new
2228 funcionality) such that between otherwise identical
2229 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2233 *) Change ssl_create_cipher_list() so that it automatically
2234 arranges the ciphersuites in reasonable order before starting
2235 to process the rule string. Thus, the definition for "DEFAULT"
2236 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2237 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2238 This makes it much easier to arrive at a reasonable default order
2239 in applications for which anonymous ciphers are OK (meaning
2240 that you can't actually use DEFAULT).
2241 [Bodo Moeller; suggested by Victor Duchovni]
2243 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2244 processing) into multiple integers instead of setting
2245 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2246 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2247 (These masks as well as the individual bit definitions are hidden
2248 away into the non-exported interface ssl/ssl_locl.h, so this
2249 change to the definition of the SSL_CIPHER structure shouldn't
2250 affect applications.) This give us more bits for each of these
2251 categories, so there is no longer a need to coagulate AES128 and
2252 AES256 into a single algorithm bit, and to coagulate Camellia128
2253 and Camellia256 into a single algorithm bit, which has led to all
2256 Thus, among other things, the kludge introduced in 0.9.7m and
2257 0.9.8e for masking out AES256 independently of AES128 or masking
2258 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2260 With the change, we also introduce new ciphersuite aliases that
2261 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2265 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2266 Use the leftmost N bytes of the signature input if the input is
2267 larger than the prime q (with N being the size in bytes of q).
2270 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2271 it yet and it is largely untested.
2274 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2277 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2278 some compilers (gcc 4.2 and later) reject their use. Safestack is
2279 reimplemented. Update ASN1 to avoid use of legacy functions.
2282 *) Win32/64 targets are linked with Winsock2.
2285 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2286 to external functions. This can be used to increase CRL handling
2287 efficiency especially when CRLs are very large by (for example) storing
2288 the CRL revoked certificates in a database.
2291 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2292 new CRLs added to a directory can be used. New command line option
2293 -verify_return_error to s_client and s_server. This causes real errors
2294 to be returned by the verify callback instead of carrying on no matter
2295 what. This reflects the way a "real world" verify callback would behave.
2298 *) GOST engine, supporting several GOST algorithms and public key formats.
2299 Kindly donated by Cryptocom.
2302 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2303 partitioned by DP are handled but no indirect CRL or reason partitioning
2304 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2305 selected via a scoring technique which handles IDP and AKID in CRLs.
2308 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2309 will ultimately be used for all verify operations: this will remove the
2310 X509_STORE dependency on certificate verification and allow alternative
2311 lookup methods. X509_STORE based implementations of these two callbacks.
2314 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2315 Modify get_crl() to find a valid (unexpired) CRL if possible.
2318 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2319 this would be called X509_CRL_cmp() but that name is already used by
2320 a function that just compares CRL issuer names. Cache several CRL
2321 extensions in X509_CRL structure and cache CRLDP in X509.
2324 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2325 this maps equivalent X509_NAME structures into a consistent structure.
2326 Name comparison can then be performed rapidly using memcmp().
2329 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2333 *) Allow digests to supply their own micalg string for S/MIME type using
2334 the ctrl EVP_MD_CTRL_MICALG.
2337 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2338 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2339 ctrl. It can then customise the structure before and/or after signing
2343 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2344 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2345 to free up any added signature OIDs.
2348 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2349 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2350 digest and cipher tables. New options added to openssl utility:
2351 list-message-digest-algorithms and list-cipher-algorithms.
2354 *) Change the array representation of binary polynomials: the list
2355 of degrees of non-zero coefficients is now terminated with -1.
2356 Previously it was terminated with 0, which was also part of the
2357 value; thus, the array representation was not applicable to
2358 polynomials where t^0 has coefficient zero. This change makes
2359 the array representation useful in a more general context.
2362 *) Various modifications and fixes to SSL/TLS cipher string
2363 handling. For ECC, the code now distinguishes between fixed ECDH
2364 with RSA certificates on the one hand and with ECDSA certificates
2365 on the other hand, since these are separate ciphersuites. The
2366 unused code for Fortezza ciphersuites has been removed.
2368 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2369 (not "ECDHE"). For consistency with the code for DH
2370 certificates, use of ECDH certificates is now considered ECDH
2371 authentication, not RSA or ECDSA authentication (the latter is
2372 merely the CA's signing algorithm and not actively used in the
2375 The temporary ciphersuite alias "ECCdraft" is no longer
2376 available, and ECC ciphersuites are no longer excluded from "ALL"
2377 and "DEFAULT". The following aliases now exist for RFC 4492
2378 ciphersuites, most of these by analogy with the DH case:
2380 kECDHr - ECDH cert, signed with RSA
2381 kECDHe - ECDH cert, signed with ECDSA
2382 kECDH - ECDH cert (signed with either RSA or ECDSA)
2383 kEECDH - ephemeral ECDH
2384 ECDH - ECDH cert or ephemeral ECDH
2390 AECDH - anonymous ECDH
2391 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2395 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2396 Use correct micalg parameters depending on digest(s) in signed message.
2399 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2400 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2403 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2404 an engine to register a method. Add ENGINE lookups for methods and
2405 functional reference processing.
2408 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2409 EVP_{Sign,Verify}* which allow an application to customise the signature
2413 *) New -resign option to smime utility. This adds one or more signers
2414 to an existing PKCS#7 signedData structure. Also -md option to use an
2415 alternative message digest algorithm for signing.
2418 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2419 create PKCS7 structures containing multiple signers. Update smime
2420 application to support multiple signers.
2423 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2427 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2428 Reorganize PBE internals to lookup from a static table using NIDs,
2429 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2430 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2431 PRF which will be automatically used with PBES2.
2434 *) Replace the algorithm specific calls to generate keys in "req" with the
2438 *) Update PKCS#7 enveloped data routines to use new API. This is now
2439 supported by any public key method supporting the encrypt operation. A
2440 ctrl is added to allow the public key algorithm to examine or modify
2441 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2445 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2446 a default digest type to use. In most cases this will be SHA1 but some
2447 algorithms (such as GOST) need to specify an alternative digest. The
2448 return value indicates how strong the preference is 1 means optional and
2449 2 is mandatory (that is it is the only supported type). Modify
2450 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2451 use the default md. Update openssl utilities to use the default digest
2452 type for signing if it is not explicitly indicated.
2455 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2456 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2457 signing method from the key type. This effectively removes the link
2458 between digests and public key types.
2461 *) Add an OID cross reference table and utility functions. Its purpose is to
2462 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2463 rsaEncryption. This will allow some of the algorithm specific hackery
2464 needed to use the correct OID to be removed.
2467 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2468 structures for PKCS7_sign(). They are now set up by the relevant public
2472 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2475 *) Add support for key derivation (agreement) in the API, DH method and
2479 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2480 public and private key formats. As a side effect these add additional
2481 command line functionality not previously available: DSA signatures can be
2482 generated and verified using pkeyutl and DH key support and generation in
2487 [Oliver Tappe <zooey@hirschkaefer.de>]
2489 *) New make target "install_html_docs" installs HTML renditions of the
2491 [Oliver Tappe <zooey@hirschkaefer.de>]
2493 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2494 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2495 support key and parameter generation and add initial key generation
2496 functionality for RSA.
2499 *) Add functions for main EVP_PKEY_method operations. The undocumented
2500 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2501 EVP_PKEY_{encrypt,decrypt}_old.
2504 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2505 key API, doesn't do much yet.
2508 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2509 public key algorithms. New option to openssl utility:
2510 "list-public-key-algorithms" to print out info.
2513 *) Implement the Supported Elliptic Curves Extension for
2514 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2517 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2518 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2521 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2522 utilities such as rsa, dsa, dsaparam etc except they process any key
2526 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2527 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2528 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2532 *) Initial support for pluggable public key ASN1.
2533 De-spaghettify the public key ASN1 handling. Move public and private
2534 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2535 algorithm specific handling to a single module within the relevant
2536 algorithm directory. Add functions to allow (near) opaque processing
2537 of public and private key structures.
2540 *) Implement the Supported Point Formats Extension for
2541 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2544 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2545 for the psk identity [hint] and the psk callback functions to the
2546 SSL_SESSION, SSL and SSL_CTX structure.
2549 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2553 SSL_CTX_use_psk_identity_hint
2554 SSL_get_psk_identity_hint
2555 SSL_get_psk_identity
2556 SSL_use_psk_identity_hint
2558 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2560 *) Add RFC 3161 compliant time stamp request creation, response generation
2561 and response verification functionality.
2562 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2564 *) Add initial support for TLS extensions, specifically for the server_name
2565 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2566 have new members for a host name. The SSL data structure has an
2567 additional member SSL_CTX *initial_ctx so that new sessions can be
2568 stored in that context to allow for session resumption, even after the
2569 SSL has been switched to a new SSL_CTX in reaction to a client's
2570 server_name extension.
2572 New functions (subject to change):
2574 SSL_get_servername()
2575 SSL_get_servername_type()
2578 New CTRL codes and macros (subject to change):
2580 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2581 - SSL_CTX_set_tlsext_servername_callback()
2582 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2583 - SSL_CTX_set_tlsext_servername_arg()
2584 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2586 openssl s_client has a new '-servername ...' option.
2588 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2589 '-key2 ...', '-servername_fatal' (subject to change). This allows
2590 testing the HostName extension for a specific single host name ('-cert'
2591 and '-key' remain fallbacks for handshakes without HostName
2592 negotiation). If the unrecognized_name alert has to be sent, this by
2593 default is a warning; it becomes fatal with the '-servername_fatal'
2596 [Peter Sylvester, Remy Allais, Christophe Renou]
2598 *) Whirlpool hash implementation is added.
2601 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2602 bn(64,32). Because of instruction set limitations it doesn't have
2603 any negative impact on performance. This was done mostly in order
2604 to make it possible to share assembler modules, such as bn_mul_mont
2605 implementations, between 32- and 64-bit builds without hassle.
2608 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2609 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2613 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2614 dedicated Montgomery multiplication procedure, is introduced.
2615 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2616 "64-bit" performance on certain 32-bit targets.
2619 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2620 in SSL structures. New SSL ctrl to set maximum send fragment size.
2621 Save memory by seeting the I/O buffer sizes dynamically instead of
2622 using the maximum available value.
2625 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2626 in addition to the text details.
2629 *) Very, very preliminary EXPERIMENTAL support for printing of general
2630 ASN1 structures. This currently produces rather ugly output and doesn't
2631 handle several customised structures at all.
2634 *) Integrated support for PVK file format and some related formats such
2635 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2636 these in the 'rsa' and 'dsa' utilities.
2639 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2642 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2643 place for the (very old) "NETSCAPE" format certificates which are now
2644 handled using new ASN1 code equivalents.
2647 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2648 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2649 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2652 *) Modify CRL distribution points extension code to print out previously
2653 unsupported fields. Enhance extension setting code to allow setting of
2657 *) Add print and set support for Issuing Distribution Point CRL extension.
2660 *) Change 'Configure' script to enable Camellia by default.
2663 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2665 *) When rejecting SSL/TLS records due to an incorrect version number, never
2666 update s->server with a new major version number. As of
2667 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2668 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2669 the previous behavior could result in a read attempt at NULL when
2670 receiving specific incorrect SSL/TLS records once record payload
2671 protection is active. (CVE-2010-0740)
2672 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2674 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2675 could be crashed if the relevant tables were not present (e.g. chrooted).
2676 [Tomas Hoger <thoger@redhat.com>]
2678 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2680 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2681 [Martin Olsson, Neel Mehta]
2683 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2684 accommodate for stack sorting, always a write lock!).
2687 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2688 excessive delays in the RAND_poll(): over a minute. As a workaround
2689 include a time check in the inner Heap32Next loop too.
2692 *) The code that handled flushing of data in SSL/TLS originally used the
2693 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2694 the problem outlined in PR#1949. The fix suggested there however can
2695 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2696 of Apache). So instead simplify the code to flush unconditionally.
2697 This should be fine since flushing with no data to flush is a no op.
2700 *) Handle TLS versions 2.0 and later properly and correctly use the
2701 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2702 off ancient servers have a habit of sticking around for a while...
2705 *) Modify compression code so it frees up structures without using the
2706 ex_data callbacks. This works around a problem where some applications
2707 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2708 restarting) then use compression (e.g. SSL with compression) later.
2709 This results in significant per-connection memory leaks and
2710 has caused some security issues including CVE-2008-1678 and
2714 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2715 change when encrypting or decrypting.
2718 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2719 connect and renegotiate with servers which do not support RI.
2720 Until RI is more widely deployed this option is enabled by default.
2723 *) Add "missing" ssl ctrls to clear options and mode.
2726 *) If client attempts to renegotiate and doesn't support RI respond with
2727 a no_renegotiation alert as required by RFC5746. Some renegotiating
2728 TLS clients will continue a connection gracefully when they receive
2729 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2730 waiting for a server hello which it will never receive. Now we treat a
2731 received no_renegotiation alert as a fatal error. This is because
2732 applications requesting a renegotiation might well expect it to succeed
2733 and would have no code in place to handle the server denying it so the
2734 only safe thing to do is to terminate the connection.
2737 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2738 peer supports secure renegotiation and 0 otherwise. Print out peer
2739 renegotiation support in s_client/s_server.
2742 *) Replace the highly broken and deprecated SPKAC certification method with
2743 the updated NID creation version. This should correctly handle UTF8.
2746 *) Implement RFC5746. Re-enable renegotiation but require the extension
2747 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2748 turns out to be a bad idea. It has been replaced by
2749 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2750 SSL_CTX_set_options(). This is really not recommended unless you
2751 know what you are doing.
2752 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2754 *) Fixes to stateless session resumption handling. Use initial_ctx when
2755 issuing and attempting to decrypt tickets in case it has changed during
2756 servername handling. Use a non-zero length session ID when attempting
2757 stateless session resumption: this makes it possible to determine if
2758 a resumption has occurred immediately after receiving server hello
2759 (several places in OpenSSL subtly assume this) instead of later in
2763 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2764 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2765 fixes for a few places where the return code is not checked
2767 [Julia Lawall <julia@diku.dk>]
2769 *) Add --strict-warnings option to Configure script to include devteam
2770 warnings in other configurations.
2773 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2774 makes it possible to install openssl libraries in locations which
2775 have names other than "lib", for example "/usr/lib64" which some
2777 [Steve Henson, based on patch from Jeremy Utley]
2779 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2780 X690 8.9.12 and can produce some misleading textual output of OIDs.
2781 [Steve Henson, reported by Dan Kaminsky]
2783 *) Delete MD2 from algorithm tables. This follows the recommendation in
2784 several standards that it is not used in new applications due to
2785 several cryptographic weaknesses. For binary compatibility reasons
2786 the MD2 API is still compiled in by default.
2789 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2793 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2794 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2796 [Guenter <lists@gknw.net>]
2798 *) Fix the server certificate chain building code to use X509_verify_cert(),
2799 it used to have an ad-hoc builder which was unable to cope with anything
2800 other than a simple chain.
2801 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2803 *) Don't check self signed certificate signatures in X509_verify_cert()
2804 by default (a flag can override this): it just wastes time without
2805 adding any security. As a useful side effect self signed root CAs
2806 with non-FIPS digests are now usable in FIPS mode.
2809 *) In dtls1_process_out_of_seq_message() the check if the current message
2810 is already buffered was missing. For every new message was memory
2811 allocated, allowing an attacker to perform an denial of service attack
2812 with sending out of seq handshake messages until there is no memory
2813 left. Additionally every future messege was buffered, even if the
2814 sequence number made no sense and would be part of another handshake.
2815 So only messages with sequence numbers less than 10 in advance will be
2816 buffered. (CVE-2009-1378)
2817 [Robin Seggelmann, discovered by Daniel Mentz]
2819 *) Records are buffered if they arrive with a future epoch to be
2820 processed after finishing the corresponding handshake. There is
2821 currently no limitation to this buffer allowing an attacker to perform
2822 a DOS attack with sending records with future epochs until there is no
2823 memory left. This patch adds the pqueue_size() function to determine
2824 the size of a buffer and limits the record buffer to 100 entries.
2826 [Robin Seggelmann, discovered by Daniel Mentz]
2828 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2829 parent structure is freed. (CVE-2009-1379)
2832 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2833 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2836 [Ilya O. <vrghost@gmail.com>]
2838 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2840 *) Disable renegotiation completely - this fixes a severe security
2841 problem (CVE-2009-3555) at the cost of breaking all
2842 renegotiation. Renegotiation can be re-enabled by setting
2843 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2844 run-time. This is really not recommended unless you know what
2848 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2850 *) Don't set val to NULL when freeing up structures, it is freed up by
2851 underlying code. If sizeof(void *) > sizeof(long) this can result in
2852 zeroing past the valid field. (CVE-2009-0789)
2853 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2855 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2856 checked correctly. This would allow some invalid signed attributes to
2857 appear to verify correctly. (CVE-2009-0591)
2858 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2860 *) Reject UniversalString and BMPString types with invalid lengths. This
2861 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2862 a legal length. (CVE-2009-0590)
2865 *) Set S/MIME signing as the default purpose rather than setting it
2866 unconditionally. This allows applications to override it at the store
2870 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2871 to handle some structures.
2874 *) Improve efficiency of mem_gets: don't search whole buffer each time
2876 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2878 *) New -hex option for openssl rand.
2881 *) Print out UTF8String and NumericString when parsing ASN1.
2884 *) Support NumericString type for name components.
2887 *) Allow CC in the environment to override the automatically chosen
2888 compiler. Note that nothing is done to ensure flags work with the
2892 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2894 *) Properly check EVP_VerifyFinal() and similar return values
2896 [Ben Laurie, Bodo Moeller, Google Security Team]
2898 *) Enable TLS extensions by default.
2901 *) Allow the CHIL engine to be loaded, whether the application is
2902 multithreaded or not. (This does not release the developer from the
2903 obligation to set up the dynamic locking callbacks.)
2904 [Sander Temme <sander@temme.net>]
2906 *) Use correct exit code if there is an error in dgst command.
2907 [Steve Henson; problem pointed out by Roland Dirlewanger]
2909 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2910 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2913 *) Add experimental JPAKE support, including demo authentication in
2914 s_client and s_server.
2917 *) Set the comparison function in v3_addr_canonize().
2918 [Rob Austein <sra@hactrn.net>]
2920 *) Add support for XMPP STARTTLS in s_client.
2921 [Philip Paeps <philip@freebsd.org>]
2923 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2924 to ensure that even with this option, only ciphersuites in the
2925 server's preference list will be accepted. (Note that the option
2926 applies only when resuming a session, so the earlier behavior was
2927 just about the algorithm choice for symmetric cryptography.)
2930 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2932 *) Fix NULL pointer dereference if a DTLS server received
2933 ChangeCipherSpec as first record (CVE-2009-1386).
2936 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2937 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2940 *) The fix in 0.9.8c that supposedly got rid of unsafe
2941 double-checked locking was incomplete for RSA blinding,
2942 addressing just one layer of what turns out to have been
2943 doubly unsafe triple-checked locking.
2945 So now fix this for real by retiring the MONT_HELPER macro
2946 in crypto/rsa/rsa_eay.c.
2948 [Bodo Moeller; problem pointed out by Marius Schilder]
2950 *) Various precautionary measures:
2952 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2954 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2955 (NB: This would require knowledge of the secret session ticket key
2956 to exploit, in which case you'd be SOL either way.)
2958 - Change bn_nist.c so that it will properly handle input BIGNUMs
2959 outside the expected range.
2961 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2964 [Neel Mehta, Bodo Moeller]
2966 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2967 the load fails. Useful for distros.
2968 [Ben Laurie and the FreeBSD team]
2970 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2973 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2976 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2978 This work was sponsored by Logica.
2981 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2982 keystores. Support for SSL/TLS client authentication too.
2983 Not compiled unless enable-capieng specified to Configure.
2985 This work was sponsored by Logica.
2988 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2989 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2990 attribute creation routines such as certifcate requests and PKCS#12
2994 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2996 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2997 handshake which could lead to a cilent crash as found using the
2998 Codenomicon TLS test suite (CVE-2008-1672)
2999 [Steve Henson, Mark Cox]
3001 *) Fix double free in TLS server name extensions which could lead to
3002 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3005 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3007 Clear the error queue to ensure that error entries left from
3008 older function calls do not interfere with the correct operation.
3009 [Lutz Jaenicke, Erik de Castro Lopo]
3011 *) Remove root CA certificates of commercial CAs:
3013 The OpenSSL project does not recommend any specific CA and does not
3014 have any policy with respect to including or excluding any CA.
3015 Therefore it does not make any sense to ship an arbitrary selection
3016 of root CA certificates with the OpenSSL software.
3019 *) RSA OAEP patches to fix two separate invalid memory reads.
3020 The first one involves inputs when 'lzero' is greater than
3021 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3022 before the beginning of from). The second one involves inputs where
3023 the 'db' section contains nothing but zeroes (there is a one-byte
3024 invalid read after the end of 'db').
3025 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3027 *) Partial backport from 0.9.9-dev:
3029 Introduce bn_mul_mont (dedicated Montgomery multiplication
3030 procedure) as a candidate for BIGNUM assembler implementation.
3031 While 0.9.9-dev uses assembler for various architectures, only
3032 x86_64 is available by default here in the 0.9.8 branch, and
3033 32-bit x86 is available through a compile-time setting.
3035 To try the 32-bit x86 assembler implementation, use Configure
3036 option "enable-montasm" (which exists only for this backport).
3038 As "enable-montasm" for 32-bit x86 disclaims code stability
3039 anyway, in this constellation we activate additional code
3040 backported from 0.9.9-dev for further performance improvements,
3041 namely BN_from_montgomery_word. (To enable this otherwise,
3042 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3044 [Andy Polyakov (backport partially by Bodo Moeller)]
3046 *) Add TLS session ticket callback. This allows an application to set
3047 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3048 values. This is useful for key rollover for example where several key
3049 sets may exist with different names.
3052 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3053 This was broken until now in 0.9.8 releases, such that the only way
3054 a registered ENGINE could be used (assuming it initialises
3055 successfully on the host) was to explicitly set it as the default
3056 for the relevant algorithms. This is in contradiction with 0.9.7
3057 behaviour and the documentation. With this fix, when an ENGINE is
3058 registered into a given algorithm's table of implementations, the
3059 'uptodate' flag is reset so that auto-discovery will be used next
3060 time a new context for that algorithm attempts to select an
3062 [Ian Lister (tweaked by Geoff Thorpe)]
3064 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3065 implemention in the following ways:
3067 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3070 Lack of BER streaming support means one pass streaming processing is
3071 only supported if data is detached: setting the streaming flag is
3072 ignored for embedded content.
3074 CMS support is disabled by default and must be explicitly enabled
3075 with the enable-cms configuration option.
3078 *) Update the GMP engine glue to do direct copies between BIGNUM and
3079 mpz_t when openssl and GMP use the same limb size. Otherwise the
3080 existing "conversion via a text string export" trick is still used.
3081 [Paul Sheer <paulsheer@gmail.com>]
3083 *) Zlib compression BIO. This is a filter BIO which compressed and
3084 uncompresses any data passed through it.
3087 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3088 RFC3394 compatible AES key wrapping.
3091 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3092 sets string data without copying. X509_ALGOR_set0() and
3093 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3094 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3095 from an X509_ATTRIBUTE structure optionally checking it occurs only
3096 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3100 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3101 to get the expected BN_FLG_CONSTTIME behavior.
3102 [Bodo Moeller (Google)]
3106 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3107 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3108 - added some more tests to do_tests.pl
3109 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3110 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3111 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3112 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3113 - various changes to netware.pl to enable gcc-cross builds on Win32
3115 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3116 - various changes to fix missing prototype warnings
3117 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3118 - added AES, WHIRLPOOL and CPUID assembler code to build files
3119 - added missing AES assembler make rules to mk1mf.pl
3120 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3121 [Guenter Knauf <eflash@gmx.net>]
3123 *) Implement certificate status request TLS extension defined in RFC3546.
3124 A client can set the appropriate parameters and receive the encoded
3125 OCSP response via a callback. A server can query the supplied parameters
3126 and set the encoded OCSP response in the callback. Add simplified examples
3127 to s_client and s_server.
3130 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3132 *) Fix various bugs:
3133 + Binary incompatibility of ssl_ctx_st structure
3134 + DTLS interoperation with non-compliant servers
3135 + Don't call get_session_cb() without proposed session
3136 + Fix ia64 assembler code
3137 [Andy Polyakov, Steve Henson]
3139 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3141 *) DTLS Handshake overhaul. There were longstanding issues with
3142 OpenSSL DTLS implementation, which were making it impossible for
3143 RFC 4347 compliant client to communicate with OpenSSL server.
3144 Unfortunately just fixing these incompatibilities would "cut off"
3145 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3146 server keeps tolerating non RFC compliant syntax. The opposite is
3147 not true, 0.9.8f client can not communicate with earlier server.
3148 This update even addresses CVE-2007-4995.
3151 *) Changes to avoid need for function casts in OpenSSL: some compilers
3152 (gcc 4.2 and later) reject their use.
3153 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3156 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3157 RFC4507bis. The encrypted ticket format is an encrypted encoded
3158 SSL_SESSION structure, that way new session features are automatically
3161 If a client application caches session in an SSL_SESSION structure
3162 support is transparent because tickets are now stored in the encoded
3165 The SSL_CTX structure automatically generates keys for ticket
3166 protection in servers so again support should be possible
3167 with no application modification.
3169 If a client or server wishes to disable RFC4507 support then the option
3170 SSL_OP_NO_TICKET can be set.
3172 Add a TLS extension debugging callback to allow the contents of any client
3173 or server extensions to be examined.
3175 This work was sponsored by Google.
3178 *) Add initial support for TLS extensions, specifically for the server_name
3179 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3180 have new members for a host name. The SSL data structure has an
3181 additional member SSL_CTX *initial_ctx so that new sessions can be
3182 stored in that context to allow for session resumption, even after the
3183 SSL has been switched to a new SSL_CTX in reaction to a client's
3184 server_name extension.
3186 New functions (subject to change):
3188 SSL_get_servername()
3189 SSL_get_servername_type()
3192 New CTRL codes and macros (subject to change):
3194 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3195 - SSL_CTX_set_tlsext_servername_callback()
3196 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3197 - SSL_CTX_set_tlsext_servername_arg()
3198 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3200 openssl s_client has a new '-servername ...' option.
3202 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3203 '-key2 ...', '-servername_fatal' (subject to change). This allows
3204 testing the HostName extension for a specific single host name ('-cert'
3205 and '-key' remain fallbacks for handshakes without HostName
3206 negotiation). If the unrecognized_name alert has to be sent, this by
3207 default is a warning; it becomes fatal with the '-servername_fatal'
3210 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3212 *) Add AES and SSE2 assembly language support to VC++ build.
3215 *) Mitigate attack on final subtraction in Montgomery reduction.
3218 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3219 (which previously caused an internal error).
3222 *) Squeeze another 10% out of IGE mode when in != out.
3225 *) AES IGE mode speedup.
3226 [Dean Gaudet (Google)]
3228 *) Add the Korean symmetric 128-bit cipher SEED (see
3229 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3230 add SEED ciphersuites from RFC 4162:
3232 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3233 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3234 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3235 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3237 To minimize changes between patchlevels in the OpenSSL 0.9.8
3238 series, SEED remains excluded from compilation unless OpenSSL
3239 is configured with 'enable-seed'.
3240 [KISA, Bodo Moeller]
3242 *) Mitigate branch prediction attacks, which can be practical if a
3243 single processor is shared, allowing a spy process to extract
3244 information. For detailed background information, see
3245 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3246 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3247 and Necessary Software Countermeasures"). The core of the change
3248 are new versions BN_div_no_branch() and
3249 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3250 respectively, which are slower, but avoid the security-relevant
3251 conditional branches. These are automatically called by BN_div()
3252 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3253 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3254 remove a conditional branch.
3256 BN_FLG_CONSTTIME is the new name for the previous
3257 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3258 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3259 in the exponent causes BN_mod_exp_mont() to use the alternative
3260 implementation in BN_mod_exp_mont_consttime().) The old name
3261 remains as a deprecated alias.
3263 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3264 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3265 constant-time implementations for more than just exponentiation.
3266 Here too the old name is kept as a deprecated alias.
3268 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3269 the BN_BLINDING structure gets an independent copy of the
3270 modulus. This means that the previous "BIGNUM *m" argument to
3271 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3272 essentially becomes "const BIGNUM *m", although we can't actually
3273 change this in the header file before 0.9.9. It allows
3274 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3275 enable BN_FLG_CONSTTIME.
3277 [Matthew D Wood (Intel Corp)]
3279 *) In the SSL/TLS server implementation, be strict about session ID
3280 context matching (which matters if an application uses a single
3281 external cache for different purposes). Previously,
3282 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3283 set. This did ensure strict client verification, but meant that,
3284 with applications using a single external cache for quite
3285 different requirements, clients could circumvent ciphersuite
3286 restrictions for a given session ID context by starting a session
3287 in a different context.
3290 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3291 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3292 authentication-only ciphersuites.
3295 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3296 not complete and could lead to a possible single byte overflow
3297 (CVE-2007-5135) [Ben Laurie]
3299 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3301 *) Since AES128 and AES256 (and similarly Camellia128 and
3302 Camellia256) share a single mask bit in the logic of
3303 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3304 kludge to work properly if AES128 is available and AES256 isn't
3305 (or if Camellia128 is available and Camellia256 isn't).
3308 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3309 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3310 When a point or a seed is encoded in a BIT STRING, we need to
3311 prevent the removal of trailing zero bits to get the proper DER
3312 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3313 of a NamedBitList, for which trailing 0 bits need to be removed.)
3316 *) Have SSL/TLS server implementation tolerate "mismatched" record
3317 protocol version while receiving ClientHello even if the
3318 ClientHello is fragmented. (The server can't insist on the
3319 particular protocol version it has chosen before the ServerHello
3320 message has informed the client about his choice.)
3323 *) Add RFC 3779 support.
3324 [Rob Austein for ARIN, Ben Laurie]
3326 *) Load error codes if they are not already present instead of using a
3327 static variable. This allows them to be cleanly unloaded and reloaded.
3328 Improve header file function name parsing.
3331 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3332 or CAPABILITY handshake as required by RFCs.
3335 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3337 *) Introduce limits to prevent malicious keys being able to
3338 cause a denial of service. (CVE-2006-2940)
3339 [Steve Henson, Bodo Moeller]
3341 *) Fix ASN.1 parsing of certain invalid structures that can result
3342 in a denial of service. (CVE-2006-2937) [Steve Henson]
3344 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3345 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3347 *) Fix SSL client code which could crash if connecting to a
3348 malicious SSLv2 server. (CVE-2006-4343)
3349 [Tavis Ormandy and Will Drewry, Google Security Team]
3351 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3352 match only those. Before that, "AES256-SHA" would be interpreted
3353 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3354 the same strength classification in 0.9.7h) as we currently only
3355 have a single AES bit in the ciphersuite description bitmap.
3356 That change, however, also applied to ciphersuite strings such as
3357 "RC4-MD5" that intentionally matched multiple ciphersuites --
3358 namely, SSL 2.0 ciphersuites in addition to the more common ones
3359 from SSL 3.0/TLS 1.0.
3361 So we change the selection algorithm again: Naming an explicit
3362 ciphersuite selects this one ciphersuite, and any other similar
3363 ciphersuite (same bitmap) from *other* protocol versions.
3364 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3365 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3367 Since SSL 2.0 does not have any ciphersuites for which the
3368 128/256 bit distinction would be relevant, this works for now.
3369 The proper fix will be to use different bits for AES128 and
3370 AES256, which would have avoided the problems from the beginning;
3371 however, bits are scarce, so we can only do this in a new release
3372 (not just a patchlevel) when we can change the SSL_CIPHER
3373 definition to split the single 'unsigned long mask' bitmap into
3374 multiple values to extend the available space.
3378 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3380 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3381 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3383 *) Add AES IGE and biIGE modes.
3386 *) Change the Unix randomness entropy gathering to use poll() when
3387 possible instead of select(), since the latter has some
3388 undesirable limitations.
3389 [Darryl Miles via Richard Levitte and Bodo Moeller]
3391 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3392 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3393 cannot be implicitly activated as part of, e.g., the "AES" alias.
3394 However, please upgrade to OpenSSL 0.9.9[-dev] for
3395 non-experimental use of the ECC ciphersuites to get TLS extension
3396 support, which is required for curve and point format negotiation
3397 to avoid potential handshake problems.
3400 *) Disable rogue ciphersuites:
3402 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3403 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3404 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3406 The latter two were purportedly from
3407 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3410 Also deactivate the remaining ciphersuites from
3411 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3412 unofficial, and the ID has long expired.
3415 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3416 dual-core machines) and other potential thread-safety issues.
3419 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3420 versions), which is now available for royalty-free use
3421 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3422 Also, add Camellia TLS ciphersuites from RFC 4132.
3424 To minimize changes between patchlevels in the OpenSSL 0.9.8
3425 series, Camellia remains excluded from compilation unless OpenSSL
3426 is configured with 'enable-camellia'.
3429 *) Disable the padding bug check when compression is in use. The padding
3430 bug check assumes the first packet is of even length, this is not
3431 necessarily true if compresssion is enabled and can result in false
3432 positives causing handshake failure. The actual bug test is ancient
3433 code so it is hoped that implementations will either have fixed it by
3434 now or any which still have the bug do not support compression.
3437 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3439 *) When applying a cipher rule check to see if string match is an explicit
3440 cipher suite and only match that one cipher suite if it is.
3443 *) Link in manifests for VC++ if needed.
3444 [Austin Ziegler <halostatue@gmail.com>]
3446 *) Update support for ECC-based TLS ciphersuites according to
3447 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3448 TLS extensions, which are supported starting with the 0.9.9
3449 branch, not in the OpenSSL 0.9.8 branch).
3452 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3453 opaque EVP_CIPHER_CTX handling.
3456 *) Fixes and enhancements to zlib compression code. We now only use
3457 "zlib1.dll" and use the default __cdecl calling convention on Win32
3458 to conform with the standards mentioned here:
3459 http://www.zlib.net/DLL_FAQ.txt
3460 Static zlib linking now works on Windows and the new --with-zlib-include
3461 --with-zlib-lib options to Configure can be used to supply the location
3462 of the headers and library. Gracefully handle case where zlib library
3466 *) Several fixes and enhancements to the OID generation code. The old code
3467 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3468 handle numbers larger than ULONG_MAX, truncated printing and had a
3469 non standard OBJ_obj2txt() behaviour.
3472 *) Add support for building of engines under engine/ as shared libraries
3473 under VC++ build system.
3476 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3477 Hopefully, we will not see any false combination of paths any more.
3480 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3482 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3483 (part of SSL_OP_ALL). This option used to disable the
3484 countermeasure against man-in-the-middle protocol-version
3485 rollback in the SSL 2.0 server implementation, which is a bad
3486 idea. (CVE-2005-2969)
3488 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3489 for Information Security, National Institute of Advanced Industrial
3490 Science and Technology [AIST], Japan)]
3492 *) Add two function to clear and return the verify parameter flags.
3495 *) Keep cipherlists sorted in the source instead of sorting them at
3496 runtime, thus removing the need for a lock.
3499 *) Avoid some small subgroup attacks in Diffie-Hellman.
3500 [Nick Mathewson and Ben Laurie]
3502 *) Add functions for well-known primes.
3505 *) Extended Windows CE support.
3506 [Satoshi Nakamura and Andy Polyakov]
3508 *) Initialize SSL_METHOD structures at compile time instead of during
3509 runtime, thus removing the need for a lock.
3512 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3513 attempting to decrypt each encrypted key in turn. Add support to
3517 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3519 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3522 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3525 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3526 key into the same file any more.
3529 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3532 *) Add -utf8 command line and config file option to 'ca'.
3533 [Stefan <stf@udoma.org]
3535 *) Removed the macro des_crypt(), as it seems to conflict with some
3536 libraries. Use DES_crypt().
3539 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3540 involves renaming the source and generated shared-libs for
3541 both. The engines will accept the corrected or legacy ids
3542 ('ncipher' and '4758_cca' respectively) when binding. NB,
3543 this only applies when building 'shared'.
3544 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3546 *) Add attribute functions to EVP_PKEY structure. Modify
3547 PKCS12_create() to recognize a CSP name attribute and
3548 use it. Make -CSP option work again in pkcs12 utility.
3551 *) Add new functionality to the bn blinding code:
3552 - automatic re-creation of the BN_BLINDING parameters after
3553 a fixed number of uses (currently 32)
3554 - add new function for parameter creation
3555 - introduce flags to control the update behaviour of the
3556 BN_BLINDING parameters
3557 - hide BN_BLINDING structure
3558 Add a second BN_BLINDING slot to the RSA structure to improve
3559 performance when a single RSA object is shared among several
3563 *) Add support for DTLS.
3564 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3566 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3567 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3570 *) Remove buggy and incomplete DH cert support from
3571 ssl/ssl_rsa.c and ssl/s3_both.c
3574 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3575 the apps/openssl applications.
3578 *) Compile clean with "-Wall -Wmissing-prototypes
3579 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3580 DEBUG_SAFESTACK must also be set.
3583 *) Change ./Configure so that certain algorithms can be disabled by default.
3584 The new counterpiece to "no-xxx" is "enable-xxx".
3586 The patented RC5 and MDC2 algorithms will now be disabled unless
3587 "enable-rc5" and "enable-mdc2", respectively, are specified.
3589 (IDEA remains enabled despite being patented. This is because IDEA
3590 is frequently required for interoperability, and there is no license
3591 fee for non-commercial use. As before, "no-idea" can be used to
3592 avoid this algorithm.)
3596 *) Add processing of proxy certificates (see RFC 3820). This work was
3597 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3598 EGEE (Enabling Grids for E-science in Europe).
3601 *) RC4 performance overhaul on modern architectures/implementations, such
3602 as Intel P4, IA-64 and AMD64.
3605 *) New utility extract-section.pl. This can be used specify an alternative
3606 section number in a pod file instead of having to treat each file as
3607 a separate case in Makefile. This can be done by adding two lines to the
3610 =for comment openssl_section:XXX
3612 The blank line is mandatory.
3616 *) New arguments -certform, -keyform and -pass for s_client and s_server
3617 to allow alternative format key and certificate files and passphrase
3621 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3622 update associated structures and add various utility functions.
3624 Add new policy related verify parameters, include policy checking in
3625 standard verify code. Enhance 'smime' application with extra parameters
3626 to support policy checking and print out.
3629 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3630 Nehemiah processors. These extensions support AES encryption in hardware
3631 as well as RNG (though RNG support is currently disabled).
3632 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3634 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3637 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3638 [Andy Polyakov and a number of other people]
3640 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3641 implementation contributed by IBM.
3642 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3644 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3645 exponent rather than 'unsigned long'. There is a corresponding change to
3646 the new 'rsa_keygen' element of the RSA_METHOD structure.
3647 [Jelte Jansen, Geoff Thorpe]
3649 *) Functionality for creating the initial serial number file is now
3650 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3652 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3653 number file to 1, which is bound to cause problems. To avoid
3654 the problems while respecting compatibility between different 0.9.7
3655 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3656 CA.pl for serial number initialization. With the new release 0.9.8,
3657 we can fix the problem directly in the 'ca' utility.)
3660 *) Reduced header interdepencies by declaring more opaque objects in
3661 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3662 give fewer recursive includes, which could break lazy source code - so
3663 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3664 developers should define this symbol when building and using openssl to
3665 ensure they track the recommended behaviour, interfaces, [etc], but
3666 backwards-compatible behaviour prevails when this isn't defined.
3669 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3672 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3673 This will generate a random key of the appropriate length based on the
3674 cipher context. The EVP_CIPHER can provide its own random key generation
3675 routine to support keys of a specific form. This is used in the des and
3676 3des routines to generate a key of the correct parity. Update S/MIME
3677 code to use new functions and hence generate correct parity DES keys.
3678 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3679 valid (weak or incorrect parity).
3682 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3683 as looking them up. This is useful when the verified structure may contain
3684 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3685 present unless the new PKCS7_NO_CRL flag is asserted.
3688 *) Extend ASN1 oid configuration module. It now additionally accepts the
3691 shortName = some long name, 1.2.3.4
3694 *) Reimplemented the BN_CTX implementation. There is now no more static
3695 limitation on the number of variables it can handle nor the depth of the
3696 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3697 information can now expand as required, and rather than having a single
3698 static array of bignums, BN_CTX now uses a linked-list of such arrays
3699 allowing it to expand on demand whilst maintaining the usefulness of
3700 BN_CTX's "bundling".
3703 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3704 to allow all RSA operations to function using a single BN_CTX.
3707 *) Preliminary support for certificate policy evaluation and checking. This
3708 is initially intended to pass the tests outlined in "Conformance Testing
3709 of Relying Party Client Certificate Path Processing Logic" v1.07.
3712 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3713 remained unused and not that useful. A variety of other little bignum
3714 tweaks and fixes have also been made continuing on from the audit (see
3718 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3719 associated ASN1, EVP and SSL functions and old ASN1 macros.
3722 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3723 and this should never fail. So the return value from the use of
3724 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3725 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3728 *) BN_CTX_get() should return zero-valued bignums, providing the same
3729 initialised value as BN_new().
3730 [Geoff Thorpe, suggested by Ulf Möller]
3732 *) Support for inhibitAnyPolicy certificate extension.
3735 *) An audit of the BIGNUM code is underway, for which debugging code is
3736 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3737 is considered valid when processing BIGNUMs, and causes execution to
3738 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3739 further steps are taken to deliberately pollute unused data in BIGNUM
3740 structures to try and expose faulty code further on. For now, openssl will
3741 (in its default mode of operation) continue to tolerate the inconsistent
3742 forms that it has tolerated in the past, but authors and packagers should
3743 consider trying openssl and their own applications when compiled with
3744 these debugging symbols defined. It will help highlight potential bugs in
3745 their own code, and will improve the test coverage for OpenSSL itself. At
3746 some point, these tighter rules will become openssl's default to improve
3747 maintainability, though the assert()s and other overheads will remain only
3748 in debugging configurations. See bn.h for more details.
3749 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3751 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3752 that can only be obtained through BN_CTX_new() (which implicitly
3753 initialises it). The presence of this function only made it possible
3754 to overwrite an existing structure (and cause memory leaks).
3757 *) Because of the callback-based approach for implementing LHASH as a
3758 template type, lh_insert() adds opaque objects to hash-tables and
3759 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3760 to clean up those corresponding objects before destroying the hash table
3761 (and losing the object pointers). So some over-zealous constifications in
3762 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3763 objects as "const" and the lh_doall[_arg] callback wrappers are not
3764 prototyped to have "const" restrictions on the object pointers they are
3765 given (and so aren't required to cast them away any more).
3768 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3769 (speed) prefers to use its own implementation. The two implementations
3770 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3771 its object type properly exposed (MS_TM) instead of casting to/from "char
3772 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3773 aren't necessarily the greatest nomenclatures - but this is what was used
3774 internally to the implementation so I've used that for now.
3777 *) Ensure that deprecated functions do not get compiled when
3778 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3779 the self-tests were still using deprecated key-generation functions so
3780 these have been updated also.
3783 *) Reorganise PKCS#7 code to separate the digest location functionality
3784 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3785 New function PKCS7_set_digest() to set the digest type for PKCS#7
3786 digestedData type. Add additional code to correctly generate the
3787 digestedData type and add support for this type in PKCS7 initialization
3791 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3792 structure of type "other".
3795 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3796 sure the loop does correctly stop and breaking ("division by zero")
3797 modulus operations are not performed. The (pre-generated) prime
3798 table crypto/bn/bn_prime.h was already correct, but it could not be
3799 re-generated on some platforms because of the "division by zero"
3800 situation in the script.
3801 [Ralf S. Engelschall]
3803 *) Update support for ECC-based TLS ciphersuites according to
3804 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3805 SHA-1 now is only used for "small" curves (where the
3806 representation of a field element takes up to 24 bytes); for
3807 larger curves, the field element resulting from ECDH is directly
3808 used as premaster secret.
3809 [Douglas Stebila (Sun Microsystems Laboratories)]
3811 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3812 curve secp160r1 to the tests.
3813 [Douglas Stebila (Sun Microsystems Laboratories)]
3815 *) Add the possibility to load symbols globally with DSO.
3816 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3818 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3819 control of the error stack.
3822 *) Add support for STORE in ENGINE.
3825 *) Add the STORE type. The intention is to provide a common interface
3826 to certificate and key stores, be they simple file-based stores, or
3827 HSM-type store, or LDAP stores, or...
3828 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3831 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3832 pass a list of arguments to any function as well as provide a way
3833 for a function to pass data back to the caller.
3836 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3837 works like BUF_strdup() but can be used to duplicate a portion of
3838 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3842 *) Add the function sk_find_ex() which works like sk_find(), but will
3843 return an index to an element even if an exact match couldn't be
3844 found. The index is guaranteed to point at the element where the
3845 searched-for key would be inserted to preserve sorting order.
3848 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3849 takes an extra flags argument for optional functionality. Currently,
3850 the following flags are defined:
3852 OBJ_BSEARCH_VALUE_ON_NOMATCH
3853 This one gets OBJ_bsearch_ex() to return a pointer to the first
3854 element where the comparing function returns a negative or zero
3857 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3858 This one gets OBJ_bsearch_ex() to return a pointer to the first
3859 element where the comparing function returns zero. This is useful
3860 if there are more than one element where the comparing function
3864 *) Make it possible to create self-signed certificates with 'openssl ca'
3865 in such a way that the self-signed certificate becomes part of the
3866 CA database and uses the same mechanisms for serial number generation
3867 as all other certificate signing. The new flag '-selfsign' enables
3868 this functionality. Adapt CA.sh and CA.pl.in.
3871 *) Add functionality to check the public key of a certificate request
3872 against a given private. This is useful to check that a certificate
3873 request can be signed by that key (self-signing).
3876 *) Make it possible to have multiple active certificates with the same
3877 subject in the CA index file. This is done only if the keyword
3878 'unique_subject' is set to 'no' in the main CA section (default
3879 if 'CA_default') of the configuration file. The value is saved
3880 with the database itself in a separate index attribute file,
3881 named like the index file with '.attr' appended to the name.
3884 *) Generate muti valued AVAs using '+' notation in config files for
3888 *) Support for nameConstraints certificate extension.
3891 *) Support for policyConstraints certificate extension.
3894 *) Support for policyMappings certificate extension.
3897 *) Make sure the default DSA_METHOD implementation only uses its
3898 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3899 and change its own handlers to be NULL so as to remove unnecessary
3900 indirection. This lets alternative implementations fallback to the
3901 default implementation more easily.
3904 *) Support for directoryName in GeneralName related extensions
3908 *) Make it possible to link applications using Makefile.shared.
3909 Make that possible even when linking against static libraries!
3912 *) Support for single pass processing for S/MIME signing. This now
3913 means that S/MIME signing can be done from a pipe, in addition
3914 cleartext signing (multipart/signed type) is effectively streaming
3915 and the signed data does not need to be all held in memory.
3917 This is done with a new flag PKCS7_STREAM. When this flag is set
3918 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3919 is done after the data is output (and digests calculated) in
3920 SMIME_write_PKCS7().
3923 *) Add full support for -rpath/-R, both in shared libraries and
3924 applications, at least on the platforms where it's known how
3928 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3929 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3930 will now compute a table of multiples of the generator that
3931 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3932 faster (notably in the case of a single point multiplication,
3933 scalar * generator).
3934 [Nils Larsch, Bodo Moeller]
3936 *) IPv6 support for certificate extensions. The various extensions
3937 which use the IP:a.b.c.d can now take IPv6 addresses using the
3938 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3942 *) Added an ENGINE that implements RSA by performing private key
3943 exponentiations with the GMP library. The conversions to and from
3944 GMP's mpz_t format aren't optimised nor are any montgomery forms
3945 cached, and on x86 it appears OpenSSL's own performance has caught up.
3946 However there are likely to be other architectures where GMP could
3947 provide a boost. This ENGINE is not built in by default, but it can be
3948 specified at Configure time and should be accompanied by the necessary
3949 linker additions, eg;
3950 ./config -DOPENSSL_USE_GMP -lgmp
3953 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3954 testing availability of engines with "-t" - the old behaviour is
3955 produced by increasing the feature's verbosity with "-tt".
3958 *) ECDSA routines: under certain error conditions uninitialized BN objects
3959 could be freed. Solution: make sure initialization is performed early
3960 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3964 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3965 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3966 software implementations. For DSA and DH, parameter generation can
3967 also be overriden by providing the appropriate method callbacks.
3970 *) Change the "progress" mechanism used in key-generation and
3971 primality testing to functions that take a new BN_GENCB pointer in
3972 place of callback/argument pairs. The new API functions have "_ex"
3973 postfixes and the older functions are reimplemented as wrappers for
3974 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3975 declarations of the old functions to help (graceful) attempts to
3976 migrate to the new functions. Also, the new key-generation API
3977 functions operate on a caller-supplied key-structure and return
3978 success/failure rather than returning a key or NULL - this is to
3979 help make "keygen" another member function of RSA_METHOD etc.
3981 Example for using the new callback interface:
3983 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3987 BN_GENCB_set(&my_cb, my_callback, my_arg);
3989 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3990 /* For the meaning of a, b in calls to my_callback(), see the
3991 * documentation of the function that calls the callback.
3992 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3993 * my_callback should return 1 if it wants BN_is_prime_ex()
3994 * to continue, or 0 to stop.
3999 *) Change the ZLIB compression method to be stateful, and make it
4000 available to TLS with the number defined in
4001 draft-ietf-tls-compression-04.txt.
4004 *) Add the ASN.1 structures and functions for CertificatePair, which
4005 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4007 CertificatePair ::= SEQUENCE {
4008 forward [0] Certificate OPTIONAL,
4009 reverse [1] Certificate OPTIONAL,
4010 -- at least one of the pair shall be present -- }
4012 Also implement the PEM functions to read and write certificate
4013 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4015 This needed to be defined, mostly for the sake of the LDAP
4016 attribute crossCertificatePair, but may prove useful elsewhere as
4020 *) Make it possible to inhibit symlinking of shared libraries in
4021 Makefile.shared, for Cygwin's sake.
4024 *) Extend the BIGNUM API by creating a function
4025 void BN_set_negative(BIGNUM *a, int neg);
4026 and a macro that behave like
4027 int BN_is_negative(const BIGNUM *a);
4029 to avoid the need to access 'a->neg' directly in applications.
4032 *) Implement fast modular reduction for pseudo-Mersenne primes
4033 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4034 EC_GROUP_new_curve_GFp() will now automatically use this
4036 [Nils Larsch <nla@trustcenter.de>]
4038 *) Add new lock type (CRYPTO_LOCK_BN).
4041 *) Change the ENGINE framework to automatically load engines
4042 dynamically from specific directories unless they could be
4043 found to already be built in or loaded. Move all the
4044 current engines except for the cryptodev one to a new
4046 The engines in engines/ are built as shared libraries if
4047 the "shared" options was given to ./Configure or ./config.
4048 Otherwise, they are inserted in libcrypto.a.
4049 /usr/local/ssl/engines is the default directory for dynamic
4050 engines, but that can be overriden at configure time through
4051 the usual use of --prefix and/or --openssldir, and at run
4052 time with the environment variable OPENSSL_ENGINES.
4053 [Geoff Thorpe and Richard Levitte]
4055 *) Add Makefile.shared, a helper makefile to build shared
4056 libraries. Adapt Makefile.org.
4059 *) Add version info to Win32 DLLs.
4060 [Peter 'Luna' Runestig" <peter@runestig.com>]
4062 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4063 can be added using this API to created arbitrary PKCS#12
4064 files while avoiding the low level API.
4066 New options to PKCS12_create(), key or cert can be NULL and
4067 will then be omitted from the output file. The encryption
4068 algorithm NIDs can be set to -1 for no encryption, the mac
4069 iteration count can be set to 0 to omit the mac.
4071 Enhance pkcs12 utility by making the -nokeys and -nocerts
4072 options work when creating a PKCS#12 file. New option -nomac
4073 to omit the mac, NONE can be set for an encryption algorithm.
4074 New code is modified to use the enhanced PKCS12_create()
4075 instead of the low level API.
4078 *) Extend ASN1 encoder to support indefinite length constructed
4079 encoding. This can output sequences tags and octet strings in
4080 this form. Modify pk7_asn1.c to support indefinite length
4081 encoding. This is experimental and needs additional code to
4082 be useful, such as an ASN1 bio and some enhanced streaming
4085 Extend template encode functionality so that tagging is passed
4086 down to the template encoder.
4089 *) Let 'openssl req' fail if an argument to '-newkey' is not
4090 recognized instead of using RSA as a default.
4093 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4094 As these are not official, they are not included in "ALL";
4095 the "ECCdraft" ciphersuite group alias can be used to select them.
4096 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4098 *) Add ECDH engine support.
4099 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4101 *) Add ECDH in new directory crypto/ecdh/.
4102 [Douglas Stebila (Sun Microsystems Laboratories)]
4104 *) Let BN_rand_range() abort with an error after 100 iterations
4105 without success (which indicates a broken PRNG).
4108 *) Change BN_mod_sqrt() so that it verifies that the input value
4109 is really the square of the return value. (Previously,
4110 BN_mod_sqrt would show GIGO behaviour.)
4113 *) Add named elliptic curves over binary fields from X9.62, SECG,
4114 and WAP/WTLS; add OIDs that were still missing.
4116 [Sheueling Chang Shantz and Douglas Stebila
4117 (Sun Microsystems Laboratories)]
4119 *) Extend the EC library for elliptic curves over binary fields
4120 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4123 EC_GF2m_simple_method
4127 EC_GROUP_new_curve_GF2m
4128 EC_GROUP_set_curve_GF2m
4129 EC_GROUP_get_curve_GF2m
4130 EC_POINT_set_affine_coordinates_GF2m
4131 EC_POINT_get_affine_coordinates_GF2m
4132 EC_POINT_set_compressed_coordinates_GF2m
4134 Point compression for binary fields is disabled by default for
4135 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4138 As binary polynomials are represented as BIGNUMs, various members
4139 of the EC_GROUP and EC_POINT data structures can be shared
4140 between the implementations for prime fields and binary fields;
4141 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4142 are essentially identical to their ..._GFp counterparts.
4143 (For simplicity, the '..._GFp' prefix has been dropped from
4144 various internal method names.)
4146 An internal 'field_div' method (similar to 'field_mul' and
4147 'field_sqr') has been added; this is used only for binary fields.
4149 [Sheueling Chang Shantz and Douglas Stebila
4150 (Sun Microsystems Laboratories)]
4152 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4153 through methods ('mul', 'precompute_mult').
4155 The generic implementations (now internally called 'ec_wNAF_mul'
4156 and 'ec_wNAF_precomputed_mult') remain the default if these
4157 methods are undefined.
4159 [Sheueling Chang Shantz and Douglas Stebila
4160 (Sun Microsystems Laboratories)]
4162 *) New function EC_GROUP_get_degree, which is defined through
4163 EC_METHOD. For curves over prime fields, this returns the bit
4164 length of the modulus.
4166 [Sheueling Chang Shantz and Douglas Stebila
4167 (Sun Microsystems Laboratories)]
4169 *) New functions EC_GROUP_dup, EC_POINT_dup.
4170 (These simply call ..._new and ..._copy).
4172 [Sheueling Chang Shantz and Douglas Stebila
4173 (Sun Microsystems Laboratories)]
4175 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4176 Polynomials are represented as BIGNUMs (where the sign bit is not
4177 used) in the following functions [macros]:
4180 BN_GF2m_sub [= BN_GF2m_add]
4181 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4182 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4183 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4185 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4186 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4187 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4188 BN_GF2m_cmp [= BN_ucmp]
4190 (Note that only the 'mod' functions are actually for fields GF(2^m).
4191 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4193 For some functions, an the irreducible polynomial defining a
4194 field can be given as an 'unsigned int[]' with strictly
4195 decreasing elements giving the indices of those bits that are set;
4196 i.e., p[] represents the polynomial
4197 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4199 p[0] > p[1] > ... > p[k] = 0.
4200 This applies to the following functions:
4205 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4206 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4208 BN_GF2m_mod_sqrt_arr
4209 BN_GF2m_mod_solve_quad_arr
4213 Conversion can be performed by the following functions:
4218 bntest.c has additional tests for binary polynomial arithmetic.
4220 Two implementations for BN_GF2m_mod_div() are available.
4221 The default algorithm simply uses BN_GF2m_mod_inv() and
4222 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4223 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4224 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4226 [Sheueling Chang Shantz and Douglas Stebila
4227 (Sun Microsystems Laboratories)]
4229 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4230 functionality is disabled at compile-time.
4231 [Douglas Stebila <douglas.stebila@sun.com>]
4233 *) Change default behaviour of 'openssl asn1parse' so that more
4234 information is visible when viewing, e.g., a certificate:
4236 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4237 mode the content of non-printable OCTET STRINGs is output in a
4238 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4239 avoid the appearance of a printable string.
4240 [Nils Larsch <nla@trustcenter.de>]
4242 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4244 EC_GROUP_set_asn1_flag()
4245 EC_GROUP_get_asn1_flag()
4246 EC_GROUP_set_point_conversion_form()
4247 EC_GROUP_get_point_conversion_form()
4248 These control ASN1 encoding details:
4249 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4250 has been set to OPENSSL_EC_NAMED_CURVE.
4251 - Points are encoded in uncompressed form by default; options for
4252 asn1_for are as for point2oct, namely
4253 POINT_CONVERSION_COMPRESSED
4254 POINT_CONVERSION_UNCOMPRESSED
4255 POINT_CONVERSION_HYBRID
4257 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4260 EC_GROUP_get0_seed()
4261 EC_GROUP_get_seed_len()
4262 This is used only for ASN1 purposes (so far).
4263 [Nils Larsch <nla@trustcenter.de>]
4265 *) Add 'field_type' member to EC_METHOD, which holds the NID
4266 of the appropriate field type OID. The new function
4267 EC_METHOD_get_field_type() returns this value.
4268 [Nils Larsch <nla@trustcenter.de>]
4273 EC_POINT_point2hex()
4274 EC_POINT_hex2point()
4275 providing useful interfaces to EC_POINT_point2oct() and
4276 EC_POINT_oct2point().
4277 [Nils Larsch <nla@trustcenter.de>]
4279 *) Change internals of the EC library so that the functions
4280 EC_GROUP_set_generator()
4281 EC_GROUP_get_generator()
4282 EC_GROUP_get_order()
4283 EC_GROUP_get_cofactor()
4284 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4285 to methods, which would lead to unnecessary code duplication when
4286 adding different types of curves.
4287 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4289 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4290 arithmetic, and such that modified wNAFs are generated
4291 (which avoid length expansion in many cases).
4294 *) Add a function EC_GROUP_check_discriminant() (defined via
4295 EC_METHOD) that verifies that the curve discriminant is non-zero.
4297 Add a function EC_GROUP_check() that makes some sanity tests
4298 on a EC_GROUP, its generator and order. This includes
4299 EC_GROUP_check_discriminant().
4300 [Nils Larsch <nla@trustcenter.de>]
4302 *) Add ECDSA in new directory crypto/ecdsa/.
4304 Add applications 'openssl ecparam' and 'openssl ecdsa'
4305 (these are based on 'openssl dsaparam' and 'openssl dsa').
4307 ECDSA support is also included in various other files across the
4308 library. Most notably,
4309 - 'openssl req' now has a '-newkey ecdsa:file' option;
4310 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4311 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4312 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4313 them suitable for ECDSA where domain parameters must be
4314 extracted before the specific public key;
4315 - ECDSA engine support has been added.
4316 [Nils Larsch <nla@trustcenter.de>]
4318 *) Include some named elliptic curves, and add OIDs from X9.62,
4319 SECG, and WAP/WTLS. Each curve can be obtained from the new
4321 EC_GROUP_new_by_curve_name(),
4322 and the list of available named curves can be obtained with
4323 EC_get_builtin_curves().
4324 Also add a 'curve_name' member to EC_GROUP objects, which can be
4326 EC_GROUP_set_curve_name()
4327 EC_GROUP_get_curve_name()
4328 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4330 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4331 was actually never needed) and in BN_mul(). The removal in BN_mul()
4332 required a small change in bn_mul_part_recursive() and the addition
4333 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4334 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4335 bn_sub_words() and bn_add_words() except they take arrays with
4339 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4341 *) Cleanse PEM buffers before freeing them since they may contain
4343 [Benjamin Bennett <ben@psc.edu>]
4345 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4346 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4347 authentication-only ciphersuites.
4350 *) Since AES128 and AES256 share a single mask bit in the logic of
4351 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4352 kludge to work properly if AES128 is available and AES256 isn't.
4355 *) Expand security boundary to match 1.1.1 module.
4358 *) Remove redundant features: hash file source, editing of test vectors
4359 modify fipsld to use external fips_premain.c signature.
4362 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4363 run algorithm test programs.
4366 *) Make algorithm test programs more tolerant of whitespace.
4369 *) Have SSL/TLS server implementation tolerate "mismatched" record
4370 protocol version while receiving ClientHello even if the
4371 ClientHello is fragmented. (The server can't insist on the
4372 particular protocol version it has chosen before the ServerHello
4373 message has informed the client about his choice.)
4376 *) Load error codes if they are not already present instead of using a
4377 static variable. This allows them to be cleanly unloaded and reloaded.
4380 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4382 *) Introduce limits to prevent malicious keys being able to
4383 cause a denial of service. (CVE-2006-2940)
4384 [Steve Henson, Bodo Moeller]
4386 *) Fix ASN.1 parsing of certain invalid structures that can result
4387 in a denial of service. (CVE-2006-2937) [Steve Henson]
4389 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4390 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4392 *) Fix SSL client code which could crash if connecting to a
4393 malicious SSLv2 server. (CVE-2006-4343)
4394 [Tavis Ormandy and Will Drewry, Google Security Team]
4396 *) Change ciphersuite string processing so that an explicit
4397 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4398 will no longer include "AES128-SHA"), and any other similar
4399 ciphersuite (same bitmap) from *other* protocol versions (so that
4400 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4401 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4402 changes from 0.9.8b and 0.9.8d.
4405 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4407 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4408 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4410 *) Change the Unix randomness entropy gathering to use poll() when
4411 possible instead of select(), since the latter has some
4412 undesirable limitations.
4413 [Darryl Miles via Richard Levitte and Bodo Moeller]
4415 *) Disable rogue ciphersuites:
4417 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4418 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4419 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4421 The latter two were purportedly from
4422 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4425 Also deactive the remaining ciphersuites from
4426 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4427 unofficial, and the ID has long expired.
4430 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4431 dual-core machines) and other potential thread-safety issues.
4434 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4436 *) Adapt fipsld and the build system to link against the validated FIPS
4437 module in FIPS mode.
4440 *) Fixes for VC++ 2005 build under Windows.
4443 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4444 from a Windows bash shell such as MSYS. It is autodetected from the
4445 "config" script when run from a VC++ environment. Modify standard VC++
4446 build to use fipscanister.o from the GNU make build.
4449 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4451 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4452 The value now differs depending on if you build for FIPS or not.
4453 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4454 safely run with a non-FIPSed libcrypto, as it may crash because of
4455 the difference induced by this change.
4458 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4460 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4461 (part of SSL_OP_ALL). This option used to disable the
4462 countermeasure against man-in-the-middle protocol-version
4463 rollback in the SSL 2.0 server implementation, which is a bad
4464 idea. (CVE-2005-2969)
4466 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4467 for Information Security, National Institute of Advanced Industrial
4468 Science and Technology [AIST], Japan)]
4470 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4471 mainly for FIPS compliance and not fully integrated at this stage.
4474 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4475 the exponentiation using a fixed-length exponent. (Otherwise,
4476 the information leaked through timing could expose the secret key
4477 after many signatures; cf. Bleichenbacher's attack on DSA with
4481 *) Make a new fixed-window mod_exp implementation the default for
4482 RSA, DSA, and DH private-key operations so that the sequence of
4483 squares and multiplies and the memory access pattern are
4484 independent of the particular secret key. This will mitigate
4485 cache-timing and potential related attacks.
4487 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4488 and this is automatically used by BN_mod_exp_mont() if the new flag
4489 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4490 will use this BN flag for private exponents unless the flag
4491 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4492 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4494 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4496 *) Change the client implementation for SSLv23_method() and
4497 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4498 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4499 (Previously, the SSL 2.0 backwards compatible Client Hello
4500 message format would be used even with SSL_OP_NO_SSLv2.)
4503 *) Add support for smime-type MIME parameter in S/MIME messages which some
4507 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4508 a threadsafe manner. Modify rsa code to use new function and add calls
4509 to dsa and dh code (which had race conditions before).
4512 *) Include the fixed error library code in the C error file definitions
4513 instead of fixing them up at runtime. This keeps the error code
4514 structures constant.
4517 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4519 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4522 *) Fixes for newer kerberos headers. NB: the casts are needed because
4523 the 'length' field is signed on one version and unsigned on another
4524 with no (?) obvious way to tell the difference, without these VC++
4525 complains. Also the "definition" of FAR (blank) is no longer included
4526 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4527 some needed definitions.
4530 *) Undo Cygwin change.
4533 *) Added support for proxy certificates according to RFC 3820.
4534 Because they may be a security thread to unaware applications,
4535 they must be explicitely allowed in run-time. See
4536 docs/HOWTO/proxy_certificates.txt for further information.
4539 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4541 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4542 server and client random values. Previously
4543 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4544 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4546 This change has negligible security impact because:
4548 1. Server and client random values still have 24 bytes of pseudo random
4551 2. Server and client random values are sent in the clear in the initial
4554 3. The master secret is derived using the premaster secret (48 bytes in
4555 size for static RSA ciphersuites) as well as client server and random
4558 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4561 [Stephen Henson, reported by UK NISCC]
4563 *) Use Windows randomness collection on Cygwin.
4566 *) Fix hang in EGD/PRNGD query when communication socket is closed
4567 prematurely by EGD/PRNGD.
4568 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4570 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4573 *) Back-port of selected performance improvements from development
4574 branch, as well as improved support for PowerPC platforms.
4577 *) Add lots of checks for memory allocation failure, error codes to indicate
4578 failure and freeing up memory if a failure occurs.
4579 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4581 *) Add new -passin argument to dgst.
4584 *) Perform some character comparisons of different types in X509_NAME_cmp:
4585 this is needed for some certificates that reencode DNs into UTF8Strings
4586 (in violation of RFC3280) and can't or wont issue name rollover
4590 *) Make an explicit check during certificate validation to see that
4591 the CA setting in each certificate on the chain is correct. As a
4592 side effect always do the following basic checks on extensions,
4593 not just when there's an associated purpose to the check:
4595 - if there is an unhandled critical extension (unless the user
4596 has chosen to ignore this fault)
4597 - if the path length has been exceeded (if one is set at all)
4598 - that certain extensions fit the associated purpose (if one has
4602 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4604 *) Avoid a race condition when CRLs are checked in a multi threaded
4605 environment. This would happen due to the reordering of the revoked
4606 entries during signature checking and serial number lookup. Now the
4607 encoding is cached and the serial number sort performed under a lock.
4608 Add new STACK function sk_is_sorted().
4611 *) Add Delta CRL to the extension code.
4614 *) Various fixes to s3_pkt.c so alerts are sent properly.
4615 [David Holmes <d.holmes@f5.com>]
4617 *) Reduce the chances of duplicate issuer name and serial numbers (in
4618 violation of RFC3280) using the OpenSSL certificate creation utilities.
4619 This is done by creating a random 64 bit value for the initial serial
4620 number when a serial number file is created or when a self signed
4621 certificate is created using 'openssl req -x509'. The initial serial
4622 number file is created using 'openssl x509 -next_serial' in CA.pl
4623 rather than being initialized to 1.
4626 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4628 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4629 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4630 [Joe Orton, Steve Henson]
4632 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4634 [Joe Orton, Steve Henson]
4636 *) Make it possible to have multiple active certificates with the same
4637 subject in the CA index file. This is done only if the keyword
4638 'unique_subject' is set to 'no' in the main CA section (default
4639 if 'CA_default') of the configuration file. The value is saved
4640 with the database itself in a separate index attribute file,
4641 named like the index file with '.attr' appended to the name.
4644 *) X509 verify fixes. Disable broken certificate workarounds when
4645 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4646 keyUsage extension present. Don't accept CRLs with unhandled critical
4647 extensions: since verify currently doesn't process CRL extensions this
4648 rejects a CRL with *any* critical extensions. Add new verify error codes
4652 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4653 A clarification of RFC2560 will require the use of OCTET STRINGs and
4654 some implementations cannot handle the current raw format. Since OpenSSL
4655 copies and compares OCSP nonces as opaque blobs without any attempt at
4656 parsing them this should not create any compatibility issues.
4659 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4660 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4661 this HMAC (and other) operations are several times slower than OpenSSL
4665 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4666 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4668 *) Use the correct content when signing type "other".
4671 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4673 *) Fix various bugs revealed by running the NISCC test suite:
4675 Stop out of bounds reads in the ASN1 code when presented with
4676 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4678 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4680 If verify callback ignores invalid public key errors don't try to check
4681 certificate signature with the NULL public key.
4685 *) New -ignore_err option in ocsp application to stop the server
4686 exiting on the first error in a request.
4689 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4690 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4694 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4695 extra data after the compression methods not only for TLS 1.0
4696 but also for SSL 3.0 (as required by the specification).
4697 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4699 *) Change X509_certificate_type() to mark the key as exported/exportable
4700 when it's 512 *bits* long, not 512 bytes.
4703 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4704 blocks during encryption.
4707 *) Various fixes to base64 BIO and non blocking I/O. On write
4708 flushes were not handled properly if the BIO retried. On read
4709 data was not being buffered properly and had various logic bugs.
4710 This also affects blocking I/O when the data being decoded is a
4714 *) Various S/MIME bugfixes and compatibility changes:
4715 output correct application/pkcs7 MIME type if
4716 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4717 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4718 of files as .eml work). Correctly handle very long lines in MIME
4722 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4724 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4725 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4726 a protocol version number mismatch like a decryption error
4727 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4730 *) Turn on RSA blinding by default in the default implementation
4731 to avoid a timing attack. Applications that don't want it can call
4732 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4733 They would be ill-advised to do so in most cases.
4734 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4736 *) Change RSA blinding code so that it works when the PRNG is not
4737 seeded (in this case, the secret RSA exponent is abused as
4738 an unpredictable seed -- if it is not unpredictable, there
4739 is no point in blinding anyway). Make RSA blinding thread-safe
4740 by remembering the creator's thread ID in rsa->blinding and
4741 having all other threads use local one-time blinding factors
4742 (this requires more computation than sharing rsa->blinding, but
4743 avoids excessive locking; and if an RSA object is not shared
4744 between threads, blinding will still be very fast).
4747 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4748 ENGINE as defaults for all supported algorithms irrespective of
4749 the 'flags' parameter. 'flags' is now honoured, so applications
4750 should make sure they are passing it correctly.
4753 *) Target "mingw" now allows native Windows code to be generated in
4754 the Cygwin environment as well as with the MinGW compiler.
4757 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4759 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4760 via timing by performing a MAC computation even if incorrect
4761 block cipher padding has been found. This is a countermeasure
4762 against active attacks where the attacker has to distinguish
4763 between bad padding and a MAC verification error. (CVE-2003-0078)
4765 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4766 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4767 Martin Vuagnoux (EPFL, Ilion)]
4769 *) Make the no-err option work as intended. The intention with no-err
4770 is not to have the whole error stack handling routines removed from
4771 libcrypto, it's only intended to remove all the function name and
4772 reason texts, thereby removing some of the footprint that may not
4773 be interesting if those errors aren't displayed anyway.
4775 NOTE: it's still possible for any application or module to have it's
4776 own set of error texts inserted. The routines are there, just not
4777 used by default when no-err is given.
4780 *) Add support for FreeBSD on IA64.
4781 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4783 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4784 Kerberos function mit_des_cbc_cksum(). Before this change,
4785 the value returned by DES_cbc_cksum() was like the one from
4786 mit_des_cbc_cksum(), except the bytes were swapped.
4787 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4789 *) Allow an application to disable the automatic SSL chain building.
4790 Before this a rather primitive chain build was always performed in
4791 ssl3_output_cert_chain(): an application had no way to send the
4792 correct chain if the automatic operation produced an incorrect result.
4794 Now the chain builder is disabled if either:
4796 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4798 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4800 The reasoning behind this is that an application would not want the
4801 auto chain building to take place if extra chain certificates are
4802 present and it might also want a means of sending no additional
4803 certificates (for example the chain has two certificates and the
4807 *) Add the possibility to build without the ENGINE framework.
4808 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4810 *) Under Win32 gmtime() can return NULL: check return value in
4811 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4814 *) DSA routines: under certain error conditions uninitialized BN objects
4815 could be freed. Solution: make sure initialization is performed early
4816 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4817 Nils Larsch <nla@trustcenter.de> via PR#459)
4820 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4821 checked on reconnect on the client side, therefore session resumption
4822 could still fail with a "ssl session id is different" error. This
4823 behaviour is masked when SSL_OP_ALL is used due to
4824 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4825 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4826 followup to PR #377.
4829 *) IA-32 assembler support enhancements: unified ELF targets, support
4830 for SCO/Caldera platforms, fix for Cygwin shared build.
4833 *) Add support for FreeBSD on sparc64. As a consequence, support for
4834 FreeBSD on non-x86 processors is separate from x86 processors on
4835 the config script, much like the NetBSD support.
4836 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4838 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4840 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4843 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4844 code (06) was taken as the first octet of the session ID and the last
4845 octet was ignored consequently. As a result SSLv2 client side session
4846 caching could not have worked due to the session ID mismatch between
4848 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4852 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4853 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4857 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4858 seems that in spite of existing for more than a year, many application
4859 author have done nothing to provide the necessary callbacks, which
4860 means that this particular engine will not work properly anywhere.
4861 This is a very unfortunate situation which forces us, in the name
4862 of usability, to give the hw_ncipher.c a static lock, which is part
4864 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4865 appear in 0.9.8 or later. We EXPECT application authors to have
4866 dealt properly with this when 0.9.8 is released (unless we actually
4867 make such changes in the libcrypto locking code that changes will
4868 have to be made anyway).
4871 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4872 octets have been read, EOF or an error occurs. Without this change
4873 some truncated ASN1 structures will not produce an error.
4876 *) Disable Heimdal support, since it hasn't been fully implemented.
4877 Still give the possibility to force the use of Heimdal, but with
4878 warnings and a request that patches get sent to openssl-dev.
4881 *) Add the VC-CE target, introduce the WINCE sysname, and add
4882 INSTALL.WCE and appropriate conditionals to make it build.
4883 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4885 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4886 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4887 edit numbers of the version.
4888 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4890 *) Introduce safe string copy and catenation functions
4891 (BUF_strlcpy() and BUF_strlcat()).
4892 [Ben Laurie (CHATS) and Richard Levitte]
4894 *) Avoid using fixed-size buffers for one-line DNs.
4895 [Ben Laurie (CHATS)]
4897 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4898 resizing buffers containing secrets, and use where appropriate.
4899 [Ben Laurie (CHATS)]
4901 *) Avoid using fixed size buffers for configuration file location.
4902 [Ben Laurie (CHATS)]
4904 *) Avoid filename truncation for various CA files.
4905 [Ben Laurie (CHATS)]
4907 *) Use sizeof in preference to magic numbers.
4908 [Ben Laurie (CHATS)]
4910 *) Avoid filename truncation in cert requests.
4911 [Ben Laurie (CHATS)]
4913 *) Add assertions to check for (supposedly impossible) buffer
4915 [Ben Laurie (CHATS)]
4917 *) Don't cache truncated DNS entries in the local cache (this could
4918 potentially lead to a spoofing attack).
4919 [Ben Laurie (CHATS)]
4921 *) Fix various buffers to be large enough for hex/decimal
4922 representations in a platform independent manner.
4923 [Ben Laurie (CHATS)]
4925 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4926 resizing buffers containing secrets, and use where appropriate.
4927 [Ben Laurie (CHATS)]
4929 *) Add BIO_indent() to avoid much slightly worrying code to do
4931 [Ben Laurie (CHATS)]
4933 *) Convert sprintf()/BIO_puts() to BIO_printf().
4934 [Ben Laurie (CHATS)]
4936 *) buffer_gets() could terminate with the buffer only half
4938 [Ben Laurie (CHATS)]
4940 *) Add assertions to prevent user-supplied crypto functions from
4941 overflowing internal buffers by having large block sizes, etc.
4942 [Ben Laurie (CHATS)]
4944 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4946 [Ben Laurie (CHATS)]
4948 *) Eliminate unused copy of key in RC4.
4949 [Ben Laurie (CHATS)]
4951 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4952 [Ben Laurie (CHATS)]
4954 *) Fix off-by-one error in EGD path.
4955 [Ben Laurie (CHATS)]
4957 *) If RANDFILE path is too long, ignore instead of truncating.
4958 [Ben Laurie (CHATS)]
4960 *) Eliminate unused and incorrectly sized X.509 structure
4962 [Ben Laurie (CHATS)]
4964 *) Eliminate unused and dangerous function knumber().
4965 [Ben Laurie (CHATS)]
4967 *) Eliminate unused and dangerous structure, KSSL_ERR.
4968 [Ben Laurie (CHATS)]
4970 *) Protect against overlong session ID context length in an encoded
4971 session object. Since these are local, this does not appear to be
4973 [Ben Laurie (CHATS)]
4975 *) Change from security patch (see 0.9.6e below) that did not affect
4976 the 0.9.6 release series:
4978 Remote buffer overflow in SSL3 protocol - an attacker could
4979 supply an oversized master key in Kerberos-enabled versions.
4981 [Ben Laurie (CHATS)]
4983 *) Change the SSL kerb5 codes to match RFC 2712.
4986 *) Make -nameopt work fully for req and add -reqopt switch.
4987 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4989 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4990 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4992 *) Make sure tests can be performed even if the corresponding algorithms
4993 have been removed entirely. This was also the last step to make
4994 OpenSSL compilable with DJGPP under all reasonable conditions.
4995 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4997 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4998 to allow version independent disabling of normally unselected ciphers,
4999 which may be activated as a side-effect of selecting a single cipher.
5001 (E.g., cipher list string "RSA" enables ciphersuites that are left
5002 out of "ALL" because they do not provide symmetric encryption.
5003 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5004 [Lutz Jaenicke, Bodo Moeller]
5006 *) Add appropriate support for separate platform-dependent build
5007 directories. The recommended way to make a platform-dependent
5008 build directory is the following (tested on Linux), maybe with
5011 # Place yourself outside of the OpenSSL source tree. In
5012 # this example, the environment variable OPENSSL_SOURCE
5013 # is assumed to contain the absolute OpenSSL source directory.
5014 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5015 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5016 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5017 mkdir -p `dirname $F`
5018 ln -s $OPENSSL_SOURCE/$F $F
5021 To be absolutely sure not to disturb the source tree, a "make clean"
5022 is a good thing. If it isn't successfull, don't worry about it,
5023 it probably means the source directory is very clean.
5026 *) Make sure any ENGINE control commands make local copies of string
5027 pointers passed to them whenever necessary. Otherwise it is possible
5028 the caller may have overwritten (or deallocated) the original string
5029 data when a later ENGINE operation tries to use the stored values.
5030 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5032 *) Improve diagnostics in file reading and command-line digests.
5033 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5035 *) Add AES modes CFB and OFB to the object database. Correct an
5036 error in AES-CFB decryption.
5039 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5040 allows existing EVP_CIPHER_CTX structures to be reused after
5041 calling EVP_*Final(). This behaviour is used by encryption
5042 BIOs and some applications. This has the side effect that
5043 applications must explicitly clean up cipher contexts with
5044 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5047 *) Check the values of dna and dnb in bn_mul_recursive before calling
5048 bn_mul_comba (a non zero value means the a or b arrays do not contain
5049 n2 elements) and fallback to bn_mul_normal if either is not zero.
5052 *) Fix escaping of non-ASCII characters when using the -subj option
5053 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5056 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5057 form for "surname", serialNumber has no short form.
5058 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5059 therefore remove "mail" short name for "internet 7".
5060 The OID for unique identifiers in X509 certificates is
5061 x500UniqueIdentifier, not uniqueIdentifier.
5062 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5065 *) Add an "init" command to the ENGINE config module and auto initialize
5066 ENGINEs. Without any "init" command the ENGINE will be initialized
5067 after all ctrl commands have been executed on it. If init=1 the
5068 ENGINE is initailized at that point (ctrls before that point are run
5069 on the uninitialized ENGINE and after on the initialized one). If
5070 init=0 then the ENGINE will not be iniatialized at all.
5073 *) Fix the 'app_verify_callback' interface so that the user-defined
5074 argument is actually passed to the callback: In the
5075 SSL_CTX_set_cert_verify_callback() prototype, the callback
5076 declaration has been changed from
5079 int (*cb)(X509_STORE_CTX *,void *);
5080 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5081 i=s->ctx->app_verify_callback(&ctx)
5082 has been changed into
5083 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5085 To update applications using SSL_CTX_set_cert_verify_callback(),
5086 a dummy argument can be added to their callback functions.
5087 [D. K. Smetters <smetters@parc.xerox.com>]
5089 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5090 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5092 *) Add and OPENSSL_LOAD_CONF define which will cause
5093 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5094 This allows older applications to transparently support certain
5095 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5096 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5097 load the config file and OPENSSL_add_all_algorithms_conf() which will
5098 always load it have also been added.
5101 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5102 Adjust NIDs and EVP layer.
5103 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5105 *) Config modules support in openssl utility.
5107 Most commands now load modules from the config file,
5108 though in a few (such as version) this isn't done
5109 because it couldn't be used for anything.
5111 In the case of ca and req the config file used is
5112 the same as the utility itself: that is the -config
5113 command line option can be used to specify an
5117 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5118 use "openssl_conf" if filename is NULL use default openssl config file.
5121 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5122 config section name. Add a new flag to tolerate a missing config file
5123 and move code to CONF_modules_load_file().
5126 *) Support for crypto accelerator cards from Accelerated Encryption
5127 Processing, www.aep.ie. (Use engine 'aep')
5128 The support was copied from 0.9.6c [engine] and adapted/corrected
5129 to work with the new engine framework.
5130 [AEP Inc. and Richard Levitte]
5132 *) Support for SureWare crypto accelerator cards from Baltimore
5133 Technologies. (Use engine 'sureware')
5134 The support was copied from 0.9.6c [engine] and adapted
5135 to work with the new engine framework.
5138 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5139 make the newer ENGINE framework commands for the CHIL engine work.
5140 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5142 *) Make it possible to produce shared libraries on ReliantUNIX.
5143 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5145 *) Add the configuration target debug-linux-ppro.
5146 Make 'openssl rsa' use the general key loading routines
5147 implemented in apps.c, and make those routines able to
5148 handle the key format FORMAT_NETSCAPE and the variant
5150 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5152 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5153 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5155 *) Add -keyform to rsautl, and document -engine.
5156 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5158 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5159 BIO_R_NO_SUCH_FILE error code rather than the generic
5160 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5163 *) Add new functions
5165 ERR_peek_last_error_line
5166 ERR_peek_last_error_line_data.
5167 These are similar to
5170 ERR_peek_error_line_data,
5171 but report on the latest error recorded rather than the first one
5172 still in the error queue.
5173 [Ben Laurie, Bodo Moeller]
5175 *) default_algorithms option in ENGINE config module. This allows things
5177 default_algorithms = ALL
5178 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5181 *) Preliminary ENGINE config module.
5184 *) New experimental application configuration code.
5187 *) Change the AES code to follow the same name structure as all other
5188 symmetric ciphers, and behave the same way. Move everything to
5189 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5190 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5192 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5193 [Ben Laurie and Theo de Raadt]
5195 *) Add option to output public keys in req command.
5196 [Massimiliano Pala madwolf@openca.org]
5198 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5199 (up to about 10% better than before for P-192 and P-224).
5202 *) New functions/macros
5204 SSL_CTX_set_msg_callback(ctx, cb)
5205 SSL_CTX_set_msg_callback_arg(ctx, arg)
5206 SSL_set_msg_callback(ssl, cb)
5207 SSL_set_msg_callback_arg(ssl, arg)
5209 to request calling a callback function
5211 void cb(int write_p, int version, int content_type,
5212 const void *buf, size_t len, SSL *ssl, void *arg)
5214 whenever a protocol message has been completely received
5215 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5216 protocol version according to which the SSL library interprets
5217 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5218 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5219 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5220 specification (change_cipher_spec(20), alert(21), handshake(22)).
5221 'buf' and 'len' point to the actual message, 'ssl' to the
5222 SSL object, and 'arg' is the application-defined value set by
5223 SSL[_CTX]_set_msg_callback_arg().
5225 'openssl s_client' and 'openssl s_server' have new '-msg' options
5226 to enable a callback that displays all protocol messages.
5229 *) Change the shared library support so shared libraries are built as
5230 soon as the corresponding static library is finished, and thereby get
5231 openssl and the test programs linked against the shared library.
5232 This still only happens when the keyword "shard" has been given to
5233 the configuration scripts.
5235 NOTE: shared library support is still an experimental thing, and
5236 backward binary compatibility is still not guaranteed.
5237 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5239 *) Add support for Subject Information Access extension.
5240 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5242 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5243 additional bytes when new memory had to be allocated, not just
5244 when reusing an existing buffer.
5247 *) New command line and configuration option 'utf8' for the req command.
5248 This allows field values to be specified as UTF8 strings.
5251 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5252 runs for the former and machine-readable output for the latter.
5255 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5256 of the e-mail address in the DN (i.e., it will go into a certificate
5257 extension only). The new configuration file option 'email_in_dn = no'
5258 has the same effect.
5259 [Massimiliano Pala madwolf@openca.org]
5261 *) Change all functions with names starting with des_ to be starting
5262 with DES_ instead. Add wrappers that are compatible with libdes,
5263 but are named _ossl_old_des_*. Finally, add macros that map the
5264 des_* symbols to the corresponding _ossl_old_des_* if libdes
5265 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5266 desired, the des_* symbols will be mapped to DES_*, with one
5269 Since we provide two compatibility mappings, the user needs to
5270 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5271 compatibility is desired. The default (i.e., when that macro
5272 isn't defined) is OpenSSL 0.9.6c compatibility.
5274 There are also macros that enable and disable the support of old
5275 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5276 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5277 are defined, the default will apply: to support the old des routines.
5279 In either case, one must include openssl/des.h to get the correct
5280 definitions. Do not try to just include openssl/des_old.h, that
5283 NOTE: This is a major break of an old API into a new one. Software
5284 authors are encouraged to switch to the DES_ style functions. Some
5285 time in the future, des_old.h and the libdes compatibility functions
5286 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5287 default), and then completely removed.
5290 *) Test for certificates which contain unsupported critical extensions.
5291 If such a certificate is found during a verify operation it is
5292 rejected by default: this behaviour can be overridden by either
5293 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5294 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5295 X509_supported_extension() has also been added which returns 1 if a
5296 particular extension is supported.
5299 *) Modify the behaviour of EVP cipher functions in similar way to digests
5300 to retain compatibility with existing code.
5303 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5304 compatibility with existing code. In particular the 'ctx' parameter does
5305 not have to be to be initialized before the call to EVP_DigestInit() and
5306 it is tidied up after a call to EVP_DigestFinal(). New function
5307 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5308 EVP_MD_CTX_copy() changed to not require the destination to be
5309 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5310 requires the destination to be valid.
5312 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5313 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5316 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5317 so that complete 'Handshake' protocol structures are kept in memory
5318 instead of overwriting 'msg_type' and 'length' with 'body' data.
5321 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5322 [Massimo Santin via Richard Levitte]
5324 *) Major restructuring to the underlying ENGINE code. This includes
5325 reduction of linker bloat, separation of pure "ENGINE" manipulation
5326 (initialisation, etc) from functionality dealing with implementations
5327 of specific crypto iterfaces. This change also introduces integrated
5328 support for symmetric ciphers and digest implementations - so ENGINEs
5329 can now accelerate these by providing EVP_CIPHER and EVP_MD
5330 implementations of their own. This is detailed in crypto/engine/README
5331 as it couldn't be adequately described here. However, there are a few
5332 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5333 were changed in the original introduction of ENGINE code have now
5334 reverted back - the hooking from this code to ENGINE is now a good
5335 deal more passive and at run-time, operations deal directly with
5336 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5337 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5338 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5339 they were not being used by the framework as there is no concept of a
5340 BIGNUM_METHOD and they could not be generalised to the new
5341 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5342 ENGINE_cpy() has been removed as it cannot be consistently defined in
5346 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5349 *) Change mkdef.pl to sort symbols that get the same entry number,
5350 and make sure the automatically generated functions ERR_load_*
5351 become part of libeay.num as well.
5354 *) New function SSL_renegotiate_pending(). This returns true once
5355 renegotiation has been requested (either SSL_renegotiate() call
5356 or HelloRequest/ClientHello received from the peer) and becomes
5357 false once a handshake has been completed.
5358 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5359 sends a HelloRequest, but does not ensure that a handshake takes
5360 place. SSL_renegotiate_pending() is useful for checking if the
5361 client has followed the request.)
5364 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5365 By default, clients may request session resumption even during
5366 renegotiation (if session ID contexts permit); with this option,
5367 session resumption is possible only in the first handshake.
5369 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5370 more bits available for options that should not be part of
5371 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5374 *) Add some demos for certificate and certificate request creation.
5377 *) Make maximum certificate chain size accepted from the peer application
5378 settable (SSL*_get/set_max_cert_list()), as proposed by
5379 "Douglas E. Engert" <deengert@anl.gov>.
5382 *) Add support for shared libraries for Unixware-7
5383 (Boyd Lynn Gerber <gerberb@zenez.com>).
5386 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5387 be done prior to destruction. Use this to unload error strings from
5388 ENGINEs that load their own error strings. NB: This adds two new API
5389 functions to "get" and "set" this destroy handler in an ENGINE.
5392 *) Alter all existing ENGINE implementations (except "openssl" and
5393 "openbsd") to dynamically instantiate their own error strings. This
5394 makes them more flexible to be built both as statically-linked ENGINEs
5395 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5396 Also, add stub code to each that makes building them as self-contained
5397 shared-libraries easier (see README.ENGINE).
5400 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5401 implementations into applications that are completely implemented in
5402 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5403 commands that can be used to configure what shared-library to load and
5404 to control aspects of the way it is handled. Also, made an update to
5405 the README.ENGINE file that brings its information up-to-date and
5406 provides some information and instructions on the "dynamic" ENGINE
5407 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5410 *) Make it possible to unload ranges of ERR strings with a new
5411 "ERR_unload_strings" function.
5414 *) Add a copy() function to EVP_MD.
5417 *) Make EVP_MD routines take a context pointer instead of just the
5418 md_data void pointer.
5421 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5422 that the digest can only process a single chunk of data
5423 (typically because it is provided by a piece of
5424 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5425 is only going to provide a single chunk of data, and hence the
5426 framework needn't accumulate the data for oneshot drivers.
5429 *) As with "ERR", make it possible to replace the underlying "ex_data"
5430 functions. This change also alters the storage and management of global
5431 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5432 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5433 index counters. The API functions that use this state have been changed
5434 to take a "class_index" rather than pointers to the class's local STACK
5435 and counter, and there is now an API function to dynamically create new
5436 classes. This centralisation allows us to (a) plug a lot of the
5437 thread-safety problems that existed, and (b) makes it possible to clean
5438 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5439 such data would previously have always leaked in application code and
5440 workarounds were in place to make the memory debugging turn a blind eye
5441 to it. Application code that doesn't use this new function will still
5442 leak as before, but their memory debugging output will announce it now
5443 rather than letting it slide.
5445 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5446 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5447 has a return value to indicate success or failure.
5450 *) Make it possible to replace the underlying "ERR" functions such that the
5451 global state (2 LHASH tables and 2 locks) is only used by the "default"
5452 implementation. This change also adds two functions to "get" and "set"
5453 the implementation prior to it being automatically set the first time
5454 any other ERR function takes place. Ie. an application can call "get",
5455 pass the return value to a module it has just loaded, and that module
5456 can call its own "set" function using that value. This means the
5457 module's "ERR" operations will use (and modify) the error state in the
5458 application and not in its own statically linked copy of OpenSSL code.
5461 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5462 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5463 the operation, and provides a more encapsulated way for external code
5464 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5465 to use these functions rather than manually incrementing the counts.
5467 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5470 *) Add EVP test program.
5473 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5476 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5477 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5478 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5479 These allow a CRL to be built without having to access X509_CRL fields
5480 directly. Modify 'ca' application to use new functions.
5483 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5484 bug workarounds. Rollback attack detection is a security feature.
5485 The problem will only arise on OpenSSL servers when TLSv1 is not
5486 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5487 Software authors not wanting to support TLSv1 will have special reasons
5488 for their choice and can explicitly enable this option.
5489 [Bodo Moeller, Lutz Jaenicke]
5491 *) Rationalise EVP so it can be extended: don't include a union of
5492 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5493 (similar to those existing for EVP_CIPHER_CTX).
5498 EVP_MD_CTX_init(&md); /* new function call */
5499 EVP_DigestInit(&md, EVP_sha1());
5500 EVP_DigestUpdate(&md, in, len);
5501 EVP_DigestFinal(&md, out, NULL);
5502 EVP_MD_CTX_cleanup(&md); /* new function call */
5506 *) Make DES key schedule conform to the usual scheme, as well as
5507 correcting its structure. This means that calls to DES functions
5508 now have to pass a pointer to a des_key_schedule instead of a
5509 plain des_key_schedule (which was actually always a pointer
5512 des_key_schedule ks;
5514 des_set_key_checked(..., &ks);
5515 des_ncbc_encrypt(..., &ks, ...);
5517 (Note that a later change renames 'des_...' into 'DES_...'.)
5520 *) Initial reduction of linker bloat: the use of some functions, such as
5521 PEM causes large amounts of unused functions to be linked in due to
5522 poor organisation. For example pem_all.c contains every PEM function
5523 which has a knock on effect of linking in large amounts of (unused)
5524 ASN1 code. Grouping together similar functions and splitting unrelated
5525 functions prevents this.
5528 *) Cleanup of EVP macros.
5531 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5532 correct _ecb suffix.
5535 *) Add initial OCSP responder support to ocsp application. The
5536 revocation information is handled using the text based index
5537 use by the ca application. The responder can either handle
5538 requests generated internally, supplied in files (for example
5539 via a CGI script) or using an internal minimal server.
5542 *) Add configuration choices to get zlib compression for TLS.
5545 *) Changes to Kerberos SSL for RFC 2712 compliance:
5546 1. Implemented real KerberosWrapper, instead of just using
5547 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5548 2. Implemented optional authenticator field of KerberosWrapper.
5550 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5551 and authenticator structs; see crypto/krb5/.
5553 Generalized Kerberos calls to support multiple Kerberos libraries.
5554 [Vern Staats <staatsvr@asc.hpc.mil>,
5555 Jeffrey Altman <jaltman@columbia.edu>
5556 via Richard Levitte]
5558 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5559 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5560 values for each of the key sizes rather than having just
5561 parameters (and 'speed' generating keys each time).
5564 *) Speed up EVP routines.
5567 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5568 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5569 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5570 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5572 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5573 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5574 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5577 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5579 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5582 *) Added the OS2-EMX target.
5583 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5585 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5586 to support NCONF routines in extension code. New function CONF_set_nconf()
5587 to allow functions which take an NCONF to also handle the old LHASH
5588 structure: this means that the old CONF compatible routines can be
5589 retained (in particular wrt extensions) without having to duplicate the
5590 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5593 *) Enhance the general user interface with mechanisms for inner control
5594 and with possibilities to have yes/no kind of prompts.
5597 *) Change all calls to low level digest routines in the library and
5598 applications to use EVP. Add missing calls to HMAC_cleanup() and
5599 don't assume HMAC_CTX can be copied using memcpy().
5600 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5602 *) Add the possibility to control engines through control names but with
5603 arbitrary arguments instead of just a string.
5604 Change the key loaders to take a UI_METHOD instead of a callback
5605 function pointer. NOTE: this breaks binary compatibility with earlier
5606 versions of OpenSSL [engine].
5607 Adapt the nCipher code for these new conditions and add a card insertion
5611 *) Enhance the general user interface with mechanisms to better support
5612 dialog box interfaces, application-defined prompts, the possibility
5613 to use defaults (for example default passwords from somewhere else)
5614 and interrupts/cancellations.
5617 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5618 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5621 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5622 tidy up some unnecessarily weird code in 'sk_new()').
5623 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5625 *) Change the key loading routines for ENGINEs to use the same kind
5626 callback (pem_password_cb) as all other routines that need this
5630 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5631 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5632 than this minimum value is recommended.
5635 *) New random seeder for OpenVMS, using the system process statistics
5636 that are easily reachable.
5639 *) Windows apparently can't transparently handle global
5640 variables defined in DLLs. Initialisations such as:
5642 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5644 wont compile. This is used by the any applications that need to
5645 declare their own ASN1 modules. This was fixed by adding the option
5646 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5647 needed for static libraries under Win32.
5650 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5651 setting of purpose and trust fields. New X509_STORE trust and
5652 purpose functions and tidy up setting in other SSL functions.
5655 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5656 structure. These are inherited by X509_STORE_CTX when it is
5657 initialised. This allows various defaults to be set in the
5658 X509_STORE structure (such as flags for CRL checking and custom
5659 purpose or trust settings) for functions which only use X509_STORE_CTX
5660 internally such as S/MIME.
5662 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5663 trust settings if they are not set in X509_STORE. This allows X509_STORE
5664 purposes and trust (in S/MIME for example) to override any set by default.
5666 Add command line options for CRL checking to smime, s_client and s_server
5670 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5671 are set then the CRL is looked up in the X509_STORE structure and
5672 its validity and signature checked, then if the certificate is found
5673 in the CRL the verify fails with a revoked error.
5675 Various new CRL related callbacks added to X509_STORE_CTX structure.
5677 Command line options added to 'verify' application to support this.
5679 This needs some additional work, such as being able to handle multiple
5680 CRLs with different times, extension based lookup (rather than just
5681 by subject name) and ultimately more complete V2 CRL extension
5685 *) Add a general user interface API (crypto/ui/). This is designed
5686 to replace things like des_read_password and friends (backward
5687 compatibility functions using this new API are provided).
5688 The purpose is to remove prompting functions from the DES code
5689 section as well as provide for prompting through dialog boxes in
5690 a window system and the like.
5693 *) Add "ex_data" support to ENGINE so implementations can add state at a
5694 per-structure level rather than having to store it globally.
5697 *) Make it possible for ENGINE structures to be copied when retrieved by
5698 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5699 This causes the "original" ENGINE structure to act like a template,
5700 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5701 operational state can be localised to each ENGINE structure, despite the
5702 fact they all share the same "methods". New ENGINE structures returned in
5703 this case have no functional references and the return value is the single
5704 structural reference. This matches the single structural reference returned
5705 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5709 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5710 needs to match any other type at all we need to manually clear the
5714 *) Changes to the "openssl engine" utility to include;
5715 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5716 about an ENGINE's available control commands.
5717 - executing control commands from command line arguments using the
5718 '-pre' and '-post' switches. '-post' is only used if '-t' is
5719 specified and the ENGINE is successfully initialised. The syntax for
5720 the individual commands are colon-separated, for example;
5721 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5724 *) New dynamic control command support for ENGINEs. ENGINEs can now
5725 declare their own commands (numbers), names (strings), descriptions,
5726 and input types for run-time discovery by calling applications. A
5727 subset of these commands are implicitly classed as "executable"
5728 depending on their input type, and only these can be invoked through
5729 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5730 can be based on user input, config files, etc). The distinction is
5731 that "executable" commands cannot return anything other than a boolean
5732 result and can only support numeric or string input, whereas some
5733 discoverable commands may only be for direct use through
5734 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5735 pointers, or other custom uses. The "executable" commands are to
5736 support parameterisations of ENGINE behaviour that can be
5737 unambiguously defined by ENGINEs and used consistently across any
5738 OpenSSL-based application. Commands have been added to all the
5739 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5740 control over shared-library paths without source code alterations.
5743 *) Changed all ENGINE implementations to dynamically allocate their
5744 ENGINEs rather than declaring them statically. Apart from this being
5745 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5746 this also allows the implementations to compile without using the
5747 internal engine_int.h header.
5750 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5751 'const' value. Any code that should be able to modify a RAND_METHOD
5752 should already have non-const pointers to it (ie. they should only
5753 modify their own ones).
5756 *) Made a variety of little tweaks to the ENGINE code.
5757 - "atalla" and "ubsec" string definitions were moved from header files
5758 to C code. "nuron" string definitions were placed in variables
5759 rather than hard-coded - allowing parameterisation of these values
5760 later on via ctrl() commands.
5761 - Removed unused "#if 0"'d code.
5762 - Fixed engine list iteration code so it uses ENGINE_free() to release
5763 structural references.
5764 - Constified the RAND_METHOD element of ENGINE structures.
5765 - Constified various get/set functions as appropriate and added
5766 missing functions (including a catch-all ENGINE_cpy that duplicates
5767 all ENGINE values onto a new ENGINE except reference counts/state).
5768 - Removed NULL parameter checks in get/set functions. Setting a method
5769 or function to NULL is a way of cancelling out a previously set
5770 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5771 and doesn't justify the extra error symbols and code.
5772 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5773 flags from engine_int.h to engine.h.
5774 - Changed prototypes for ENGINE handler functions (init(), finish(),
5775 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5778 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5779 to the algorithm using long division. The binary algorithm can be
5780 used only if the modulus is odd. On 32-bit systems, it is faster
5781 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5782 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5783 up to 450 bits. In 64-bit environments, the binary algorithm
5784 appears to be advantageous for much longer moduli; here we use it
5785 for moduli up to 2048 bits.
5788 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5789 could not support the combine flag in choice fields.
5792 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5793 extensions from a certificate request to the certificate.
5796 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5797 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5798 file: this allows the display of the certificate about to be
5799 signed to be customised, to allow certain fields to be included
5800 or excluded and extension details. The old system didn't display
5801 multicharacter strings properly, omitted fields not in the policy
5802 and couldn't display additional details such as extensions.
5805 *) Function EC_POINTs_mul for multiple scalar multiplication
5806 of an arbitrary number of elliptic curve points
5807 \sum scalars[i]*points[i],
5808 optionally including the generator defined for the EC_GROUP:
5809 scalar*generator + \sum scalars[i]*points[i].
5811 EC_POINT_mul is a simple wrapper function for the typical case
5812 that the point list has just one item (besides the optional
5816 *) First EC_METHODs for curves over GF(p):
5818 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5819 operations and provides various method functions that can also
5820 operate with faster implementations of modular arithmetic.
5822 EC_GFp_mont_method() reuses most functions that are part of
5823 EC_GFp_simple_method, but uses Montgomery arithmetic.
5825 [Bodo Moeller; point addition and point doubling
5826 implementation directly derived from source code provided by
5827 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5829 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5830 crypto/ec/ec_lib.c):
5832 Curves are EC_GROUP objects (with an optional group generator)
5833 based on EC_METHODs that are built into the library.
5835 Points are EC_POINT objects based on EC_GROUP objects.
5837 Most of the framework would be able to handle curves over arbitrary
5838 finite fields, but as there are no obvious types for fields other
5839 than GF(p), some functions are limited to that for now.
5842 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5843 that the file contains a complete HTTP response.
5846 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5847 change the def and num file printf format specifier from "%-40sXXX"
5848 to "%-39s XXX". The latter will always guarantee a space after the
5849 field while the former will cause them to run together if the field
5850 is 40 of more characters long.
5853 *) Constify the cipher and digest 'method' functions and structures
5854 and modify related functions to take constant EVP_MD and EVP_CIPHER
5858 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5859 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5862 *) Modify EVP_Digest*() routines so they now return values. Although the
5863 internal software routines can never fail additional hardware versions
5867 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5869 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5870 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5873 ERR_R_NESTED_ASN1_ERROR
5875 ERR_R_MISSING_ASN1_EOS
5876 were 4 .. 9, conflicting with
5877 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5879 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5880 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5882 Add new error code 'ERR_R_INTERNAL_ERROR'.
5885 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5889 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5890 sets the subject name for a new request or supersedes the
5891 subject name in a given request. Formats that can be parsed are
5892 'CN=Some Name, OU=myOU, C=IT'
5894 'CN=Some Name/OU=myOU/C=IT'.
5896 Add options '-batch' and '-verbose' to 'openssl req'.
5897 [Massimiliano Pala <madwolf@hackmasters.net>]
5899 *) Introduce the possibility to access global variables through
5900 functions on platform were that's the best way to handle exporting
5901 global variables in shared libraries. To enable this functionality,
5902 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5903 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5904 is normally done by Configure or something similar).
5906 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5907 in the source file (foo.c) like this:
5909 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5910 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5912 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5913 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5915 OPENSSL_DECLARE_GLOBAL(int,foo);
5916 #define foo OPENSSL_GLOBAL_REF(foo)
5917 OPENSSL_DECLARE_GLOBAL(double,bar);
5918 #define bar OPENSSL_GLOBAL_REF(bar)
5920 The #defines are very important, and therefore so is including the
5921 header file everywhere where the defined globals are used.
5923 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5924 of ASN.1 items, but that structure is a bit different.
5926 The largest change is in util/mkdef.pl which has been enhanced with
5927 better and easier to understand logic to choose which symbols should
5928 go into the Windows .def files as well as a number of fixes and code
5929 cleanup (among others, algorithm keywords are now sorted
5930 lexicographically to avoid constant rewrites).
5933 *) In BN_div() keep a copy of the sign of 'num' before writing the
5934 result to 'rm' because if rm==num the value will be overwritten
5935 and produce the wrong result if 'num' is negative: this caused
5936 problems with BN_mod() and BN_nnmod().
5939 *) Function OCSP_request_verify(). This checks the signature on an
5940 OCSP request and verifies the signer certificate. The signer
5941 certificate is just checked for a generic purpose and OCSP request
5945 *) Add OCSP_check_validity() function to check the validity of OCSP
5946 responses. OCSP responses are prepared in real time and may only
5947 be a few seconds old. Simply checking that the current time lies
5948 between thisUpdate and nextUpdate max reject otherwise valid responses
5949 caused by either OCSP responder or client clock inaccuracy. Instead
5950 we allow thisUpdate and nextUpdate to fall within a certain period of
5951 the current time. The age of the response can also optionally be
5952 checked. Two new options -validity_period and -status_age added to
5956 *) If signature or public key algorithm is unrecognized print out its
5957 OID rather that just UNKNOWN.
5960 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5961 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5962 ID to be generated from the issuer certificate alone which can then be
5963 passed to OCSP_id_issuer_cmp().
5966 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5967 ASN1 modules to export functions returning ASN1_ITEM pointers
5968 instead of the ASN1_ITEM structures themselves. This adds several
5969 new macros which allow the underlying ASN1 function/structure to
5970 be accessed transparently. As a result code should not use ASN1_ITEM
5971 references directly (such as &X509_it) but instead use the relevant
5972 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5973 use of the new ASN1 code on platforms where exporting structures
5974 is problematical (for example in shared libraries) but exporting
5975 functions returning pointers to structures is not.
5978 *) Add support for overriding the generation of SSL/TLS session IDs.
5979 These callbacks can be registered either in an SSL_CTX or per SSL.
5980 The purpose of this is to allow applications to control, if they wish,
5981 the arbitrary values chosen for use as session IDs, particularly as it
5982 can be useful for session caching in multiple-server environments. A
5983 command-line switch for testing this (and any client code that wishes
5984 to use such a feature) has been added to "s_server".
5985 [Geoff Thorpe, Lutz Jaenicke]
5987 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5988 of the form '#if defined(...) || defined(...) || ...' and
5989 '#if !defined(...) && !defined(...) && ...'. This also avoids
5990 the growing number of special cases it was previously handling.
5993 *) Make all configuration macros available for application by making
5994 sure they are available in opensslconf.h, by giving them names starting
5995 with "OPENSSL_" to avoid conflicts with other packages and by making
5996 sure e_os2.h will cover all platform-specific cases together with
5998 Additionally, it is now possible to define configuration/platform-
5999 specific names (called "system identities"). In the C code, these
6000 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6001 macro with the name beginning with "OPENSSL_SYS_", which is determined
6002 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6006 *) New option -set_serial to 'req' and 'x509' this allows the serial
6007 number to use to be specified on the command line. Previously self
6008 signed certificates were hard coded with serial number 0 and the
6009 CA options of 'x509' had to use a serial number in a file which was
6013 *) New options to 'ca' utility to support V2 CRL entry extensions.
6014 Currently CRL reason, invalidity date and hold instruction are
6015 supported. Add new CRL extensions to V3 code and some new objects.
6018 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6019 disable standard block padding (aka PKCS#5 padding) in the EVP
6020 API, which was previously mandatory. This means that the data is
6021 not padded in any way and so the total length much be a multiple
6022 of the block size, otherwise an error occurs.
6025 *) Initial (incomplete) OCSP SSL support.
6028 *) New function OCSP_parse_url(). This splits up a URL into its host,
6029 port and path components: primarily to parse OCSP URLs. New -url
6030 option to ocsp utility.
6033 *) New nonce behavior. The return value of OCSP_check_nonce() now
6034 reflects the various checks performed. Applications can decide
6035 whether to tolerate certain situations such as an absent nonce
6036 in a response when one was present in a request: the ocsp application
6037 just prints out a warning. New function OCSP_add1_basic_nonce()
6038 this is to allow responders to include a nonce in a response even if
6039 the request is nonce-less.
6042 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6043 skipped when using openssl x509 multiple times on a single input file,
6044 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6047 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6048 set string type: to handle setting ASN1_TIME structures. Fix ca
6049 utility to correctly initialize revocation date of CRLs.
6052 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6053 the clients preferred ciphersuites and rather use its own preferences.
6054 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6055 Internet Explorer by ensuring unchanged hash method during stepup.
6056 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6059 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6060 to aes and add a new 'exist' option to print out symbols that don't
6064 *) Additional options to ocsp utility to allow flags to be set and
6065 additional certificates supplied.
6068 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6069 OCSP client a number of certificate to only verify the response
6073 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6074 handle the new API. Currently only ECB, CBC modes supported. Add new
6077 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6078 Encryption Standard (AES) Ciphersuites for Transport Layer
6079 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6080 not enabled by default and were not part of the "ALL" ciphersuite
6081 alias because they were not yet official; they could be
6082 explicitly requested by specifying the "AESdraft" ciphersuite
6083 group alias. In the final release of OpenSSL 0.9.7, the group
6084 alias is called "AES" and is part of "ALL".)
6085 [Ben Laurie, Steve Henson, Bodo Moeller]
6087 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6088 request to response.
6091 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6092 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6093 extract information from a certificate request. OCSP_response_create()
6094 creates a response and optionally adds a basic response structure.
6095 OCSP_basic_add1_status() adds a complete single response to a basic
6096 response and returns the OCSP_SINGLERESP structure just added (to allow
6097 extensions to be included for example). OCSP_basic_add1_cert() adds a
6098 certificate to a basic response and OCSP_basic_sign() signs a basic
6099 response with various flags. New helper functions ASN1_TIME_check()
6100 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6101 (converts ASN1_TIME to GeneralizedTime).
6104 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6105 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6106 structure from a certificate. X509_pubkey_digest() digests the public_key
6107 contents: this is used in various key identifiers.
6110 *) Make sk_sort() tolerate a NULL argument.
6111 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6113 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6114 passed by the function are trusted implicitly. If any of them signed the
6115 response then it is assumed to be valid and is not verified.
6118 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6119 to data. This was previously part of the PKCS7 ASN1 code. This
6120 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6121 [Steve Henson, reported by Kenneth R. Robinette
6122 <support@securenetterm.com>]
6124 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6125 routines: without these tracing memory leaks is very painful.
6126 Fix leaks in PKCS12 and PKCS7 routines.
6129 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6130 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6131 effectively meant GeneralizedTime would never be used. Now it
6132 is initialised to -1 but X509_time_adj() now has to check the value
6133 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6134 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6135 [Steve Henson, reported by Kenneth R. Robinette
6136 <support@securenetterm.com>]
6138 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6139 result in a zero length in the ASN1_INTEGER structure which was
6140 not consistent with the structure when d2i_ASN1_INTEGER() was used
6141 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6142 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6143 where it did not print out a minus for negative ASN1_INTEGER.
6146 *) Add summary printout to ocsp utility. The various functions which
6147 convert status values to strings have been renamed to:
6148 OCSP_response_status_str(), OCSP_cert_status_str() and
6149 OCSP_crl_reason_str() and are no longer static. New options
6150 to verify nonce values and to disable verification. OCSP response
6151 printout format cleaned up.
6154 *) Add additional OCSP certificate checks. These are those specified
6155 in RFC2560. This consists of two separate checks: the CA of the
6156 certificate being checked must either be the OCSP signer certificate
6157 or the issuer of the OCSP signer certificate. In the latter case the
6158 OCSP signer certificate must contain the OCSP signing extended key
6159 usage. This check is performed by attempting to match the OCSP
6160 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6161 in the OCSP_CERTID structures of the response.
6164 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6165 and related routines. This uses the standard OpenSSL certificate
6166 verify routines to perform initial checks (just CA validity) and
6167 to obtain the certificate chain. Then additional checks will be
6168 performed on the chain. Currently the root CA is checked to see
6169 if it is explicitly trusted for OCSP signing. This is used to set
6170 a root CA as a global signing root: that is any certificate that
6171 chains to that CA is an acceptable OCSP signing certificate.
6174 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6175 extensions from a separate configuration file.
6176 As when reading extensions from the main configuration file,
6177 the '-extensions ...' option may be used for specifying the
6179 [Massimiliano Pala <madwolf@comune.modena.it>]
6181 *) New OCSP utility. Allows OCSP requests to be generated or
6182 read. The request can be sent to a responder and the output
6183 parsed, outputed or printed in text form. Not complete yet:
6184 still needs to check the OCSP response validity.
6187 *) New subcommands for 'openssl ca':
6188 'openssl ca -status <serial>' prints the status of the cert with
6189 the given serial number (according to the index file).
6190 'openssl ca -updatedb' updates the expiry status of certificates
6192 [Massimiliano Pala <madwolf@comune.modena.it>]
6194 *) New '-newreq-nodes' command option to CA.pl. This is like
6195 '-newreq', but calls 'openssl req' with the '-nodes' option
6196 so that the resulting key is not encrypted.
6197 [Damien Miller <djm@mindrot.org>]
6199 *) New configuration for the GNU Hurd.
6200 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6202 *) Initial code to implement OCSP basic response verify. This
6203 is currently incomplete. Currently just finds the signer's
6204 certificate and verifies the signature on the response.
6207 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6208 value of OPENSSLDIR. This is available via the new '-d' option
6209 to 'openssl version', and is also included in 'openssl version -a'.
6212 *) Allowing defining memory allocation callbacks that will be given
6213 file name and line number information in additional arguments
6214 (a const char* and an int). The basic functionality remains, as
6215 well as the original possibility to just replace malloc(),
6216 realloc() and free() by functions that do not know about these
6217 additional arguments. To register and find out the current
6218 settings for extended allocation functions, the following
6219 functions are provided:
6221 CRYPTO_set_mem_ex_functions
6222 CRYPTO_set_locked_mem_ex_functions
6223 CRYPTO_get_mem_ex_functions
6224 CRYPTO_get_locked_mem_ex_functions
6226 These work the same way as CRYPTO_set_mem_functions and friends.
6227 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6228 extended allocation function is enabled.
6229 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6230 a conventional allocation function is enabled.
6231 [Richard Levitte, Bodo Moeller]
6233 *) Finish off removing the remaining LHASH function pointer casts.
6234 There should no longer be any prototype-casting required when using
6235 the LHASH abstraction, and any casts that remain are "bugs". See
6236 the callback types and macros at the head of lhash.h for details
6237 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6240 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6241 If /dev/[u]random devices are not available or do not return enough
6242 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6244 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6245 /etc/entropy will be queried once each in this sequence, quering stops
6246 when enough entropy was collected without querying more sockets.
6249 *) Change the Unix RAND_poll() variant to be able to poll several
6250 random devices, as specified by DEVRANDOM, until a sufficient amount
6251 of data has been collected. We spend at most 10 ms on each file
6252 (select timeout) and read in non-blocking mode. DEVRANDOM now
6253 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6254 (previously it was just the string "/dev/urandom"), so on typical
6255 platforms the 10 ms delay will never occur.
6256 Also separate out the Unix variant to its own file, rand_unix.c.
6257 For VMS, there's a currently-empty rand_vms.c.
6260 *) Move OCSP client related routines to ocsp_cl.c. These
6261 provide utility functions which an application needing
6262 to issue a request to an OCSP responder and analyse the
6263 response will typically need: as opposed to those which an
6264 OCSP responder itself would need which will be added later.
6266 OCSP_request_sign() signs an OCSP request with an API similar
6267 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6268 response. OCSP_response_get1_basic() extracts basic response
6269 from response. OCSP_resp_find_status(): finds and extracts status
6270 information from an OCSP_CERTID structure (which will be created
6271 when the request structure is built). These are built from lower
6272 level functions which work on OCSP_SINGLERESP structures but
6273 wont normally be used unless the application wishes to examine
6274 extensions in the OCSP response for example.
6276 Replace nonce routines with a pair of functions.
6277 OCSP_request_add1_nonce() adds a nonce value and optionally
6278 generates a random value. OCSP_check_nonce() checks the
6279 validity of the nonce in an OCSP response.
6282 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6283 This doesn't copy the supplied OCSP_CERTID and avoids the
6284 need to free up the newly created id. Change return type
6285 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6286 This can then be used to add extensions to the request.
6287 Deleted OCSP_request_new(), since most of its functionality
6288 is now in OCSP_REQUEST_new() (and the case insensitive name
6289 clash) apart from the ability to set the request name which
6290 will be added elsewhere.
6293 *) Update OCSP API. Remove obsolete extensions argument from
6294 various functions. Extensions are now handled using the new
6295 OCSP extension code. New simple OCSP HTTP function which
6296 can be used to send requests and parse the response.
6299 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6300 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6301 uses the special reorder version of SET OF to sort the attributes
6302 and reorder them to match the encoded order. This resolves a long
6303 standing problem: a verify on a PKCS7 structure just after signing
6304 it used to fail because the attribute order did not match the
6305 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6306 it uses the received order. This is necessary to tolerate some broken
6307 software that does not order SET OF. This is handled by encoding
6308 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6309 to produce the required SET OF.
6312 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6313 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6314 files to get correct declarations of the ASN.1 item variables.
6317 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6318 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6319 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6320 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6321 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6322 ASN1_ITEM and no wrapper functions.
6325 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6326 replace the old function pointer based I/O routines. Change most of
6327 the *_d2i_bio() and *_d2i_fp() functions to use these.
6330 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6331 lines, recognice more "algorithms" that can be deselected, and make
6332 it complain about algorithm deselection that isn't recognised.
6335 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6336 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6337 to use new functions. Add NO_ASN1_OLD which can be set to remove
6338 some old style ASN1 functions: this can be used to determine if old
6339 code will still work when these eventually go away.
6342 *) New extension functions for OCSP structures, these follow the
6343 same conventions as certificates and CRLs.
6346 *) New function X509V3_add1_i2d(). This automatically encodes and
6347 adds an extension. Its behaviour can be customised with various
6348 flags to append, replace or delete. Various wrappers added for
6349 certifcates and CRLs.
6352 *) Fix to avoid calling the underlying ASN1 print routine when
6353 an extension cannot be parsed. Correct a typo in the
6354 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6357 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6358 entries for variables.
6361 *) Add functionality to apps/openssl.c for detecting locking
6362 problems: As the program is single-threaded, all we have
6363 to do is register a locking callback using an array for
6364 storing which locks are currently held by the program.
6367 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6368 SSL_get_ex_data_X509_STORE_idx(), which is used in
6369 ssl_verify_cert_chain() and thus can be called at any time
6370 during TLS/SSL handshakes so that thread-safety is essential.
6371 Unfortunately, the ex_data design is not at all suited
6372 for multi-threaded use, so it probably should be abolished.
6375 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6376 [Broadcom, tweaked and integrated by Geoff Thorpe]
6378 *) Move common extension printing code to new function
6379 X509V3_print_extensions(). Reorganise OCSP print routines and
6380 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6383 *) New function X509_signature_print() to remove duplication in some
6387 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6388 set (this was treated exactly the same as SET OF previously). This
6389 is used to reorder the STACK representing the structure to match the
6390 encoding. This will be used to get round a problem where a PKCS7
6391 structure which was signed could not be verified because the STACK
6392 order did not reflect the encoded order.
6395 *) Reimplement the OCSP ASN1 module using the new code.
6398 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6399 for its ASN1 operations. The old style function pointers still exist
6400 for now but they will eventually go away.
6403 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6404 completely replaces the old ASN1 functionality with a table driven
6405 encoder and decoder which interprets an ASN1_ITEM structure describing
6406 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6407 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6408 has also been converted to the new form.
6411 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6412 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6413 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6414 for negative moduli.
6417 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6418 of not touching the result's sign bit.
6421 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6425 *) Changed the LHASH code to use prototypes for callbacks, and created
6426 macros to declare and implement thin (optionally static) functions
6427 that provide type-safety and avoid function pointer casting for the
6428 type-specific callbacks.
6431 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6433 [Veers Staats <staatsvr@asc.hpc.mil>,
6434 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6436 *) Reformat the FAQ so the different questions and answers can be divided
6437 in sections depending on the subject.
6440 *) Have the zlib compression code load ZLIB.DLL dynamically under
6444 *) New function BN_mod_sqrt for computing square roots modulo a prime
6445 (using the probabilistic Tonelli-Shanks algorithm unless
6446 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6447 be handled deterministically).
6448 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6450 *) Make BN_mod_inverse faster by explicitly handling small quotients
6451 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6452 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6455 *) New function BN_kronecker.
6458 *) Fix BN_gcd so that it works on negative inputs; the result is
6459 positive unless both parameters are zero.
6460 Previously something reasonably close to an infinite loop was
6461 possible because numbers could be growing instead of shrinking
6462 in the implementation of Euclid's algorithm.
6465 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6466 sign of the number in question.
6468 Fix BN_is_word(a,w) to work correctly for w == 0.
6470 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6471 because its test if the absolute value of 'a' equals 'w'.
6472 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6473 it exists mostly for use in the implementations of BN_is_zero(),
6474 BN_is_one(), and BN_is_word().
6477 *) New function BN_swap.
6480 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6481 the exponentiation functions are more likely to produce reasonable
6482 results on negative inputs.
6485 *) Change BN_mod_mul so that the result is always non-negative.
6486 Previously, it could be negative if one of the factors was negative;
6487 I don't think anyone really wanted that behaviour.
6490 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6491 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6492 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6493 and add new functions:
6502 BN_mod_lshift1_quick
6506 These functions always generate non-negative results.
6508 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6509 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6511 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6512 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6513 be reduced modulo m.
6514 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6517 The following entry accidentally appeared in the CHANGES file
6518 distributed with OpenSSL 0.9.7. The modifications described in
6519 it do *not* apply to OpenSSL 0.9.7.
6521 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6522 was actually never needed) and in BN_mul(). The removal in BN_mul()
6523 required a small change in bn_mul_part_recursive() and the addition
6524 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6525 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6526 bn_sub_words() and bn_add_words() except they take arrays with
6531 *) In 'openssl passwd', verify passwords read from the terminal
6532 unless the '-salt' option is used (which usually means that
6533 verification would just waste user's time since the resulting
6534 hash is going to be compared with some given password hash)
6535 or the new '-noverify' option is used.
6537 This is an incompatible change, but it does not affect
6538 non-interactive use of 'openssl passwd' (passwords on the command
6539 line, '-stdin' option, '-in ...' option) and thus should not
6543 *) Remove all references to RSAref, since there's no more need for it.
6546 *) Make DSO load along a path given through an environment variable
6547 (SHLIB_PATH) with shl_load().
6550 *) Constify the ENGINE code as a result of BIGNUM constification.
6551 Also constify the RSA code and most things related to it. In a
6552 few places, most notable in the depth of the ASN.1 code, ugly
6553 casts back to non-const were required (to be solved at a later
6557 *) Make it so the openssl application has all engines loaded by default.
6560 *) Constify the BIGNUM routines a little more.
6563 *) Add the following functions:
6565 ENGINE_load_cswift()
6567 ENGINE_load_atalla()
6569 ENGINE_load_builtin_engines()
6571 That way, an application can itself choose if external engines that
6572 are built-in in OpenSSL shall ever be used or not. The benefit is
6573 that applications won't have to be linked with libdl or other dso
6574 libraries unless it's really needed.
6576 Changed 'openssl engine' to load all engines on demand.
6577 Changed the engine header files to avoid the duplication of some
6578 declarations (they differed!).
6581 *) 'openssl engine' can now list capabilities.
6584 *) Better error reporting in 'openssl engine'.
6587 *) Never call load_dh_param(NULL) in s_server.
6590 *) Add engine application. It can currently list engines by name and
6591 identity, and test if they are actually available.
6594 *) Improve RPM specification file by forcing symbolic linking and making
6595 sure the installed documentation is also owned by root.root.
6596 [Damien Miller <djm@mindrot.org>]
6598 *) Give the OpenSSL applications more possibilities to make use of
6599 keys (public as well as private) handled by engines.
6602 *) Add OCSP code that comes from CertCo.
6605 *) Add VMS support for the Rijndael code.
6608 *) Added untested support for Nuron crypto accelerator.
6611 *) Add support for external cryptographic devices. This code was
6612 previously distributed separately as the "engine" branch.
6613 [Geoff Thorpe, Richard Levitte]
6615 *) Rework the filename-translation in the DSO code. It is now possible to
6616 have far greater control over how a "name" is turned into a filename
6617 depending on the operating environment and any oddities about the
6618 different shared library filenames on each system.
6621 *) Support threads on FreeBSD-elf in Configure.
6624 *) Fix for SHA1 assembly problem with MASM: it produces
6625 warnings about corrupt line number information when assembling
6626 with debugging information. This is caused by the overlapping
6628 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6631 NCONF_get_number() has no error checking at all. As a replacement,
6632 NCONF_get_number_e() is defined (_e for "error checking") and is
6633 promoted strongly. The old NCONF_get_number is kept around for
6634 binary backward compatibility.
6635 Make it possible for methods to load from something other than a BIO,
6636 by providing a function pointer that is given a name instead of a BIO.
6637 For example, this could be used to load configuration data from an
6641 *) Fix for non blocking accept BIOs. Added new I/O special reason
6642 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6643 with non blocking I/O was not possible because no retry code was
6644 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6648 *) Added the beginnings of Rijndael support.
6651 *) Fix for bug in DirectoryString mask setting. Add support for
6652 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6653 to allow certificate printing to more controllable, additional
6654 'certopt' option to 'x509' to allow new printing options to be
6658 *) Clean old EAY MD5 hack from e_os.h.
6661 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6663 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6664 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6665 [Joe Orton, Steve Henson]
6667 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6669 *) Fix additional bug revealed by the NISCC test suite:
6671 Stop bug triggering large recursion when presented with
6672 certain ASN.1 tags (CVE-2003-0851)
6675 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6677 *) Fix various bugs revealed by running the NISCC test suite:
6679 Stop out of bounds reads in the ASN1 code when presented with
6680 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6682 If verify callback ignores invalid public key errors don't try to check
6683 certificate signature with the NULL public key.
6687 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6688 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6692 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6693 extra data after the compression methods not only for TLS 1.0
6694 but also for SSL 3.0 (as required by the specification).
6695 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6697 *) Change X509_certificate_type() to mark the key as exported/exportable
6698 when it's 512 *bits* long, not 512 bytes.
6701 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6703 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6704 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6705 a protocol version number mismatch like a decryption error
6706 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6709 *) Turn on RSA blinding by default in the default implementation
6710 to avoid a timing attack. Applications that don't want it can call
6711 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6712 They would be ill-advised to do so in most cases.
6713 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6715 *) Change RSA blinding code so that it works when the PRNG is not
6716 seeded (in this case, the secret RSA exponent is abused as
6717 an unpredictable seed -- if it is not unpredictable, there
6718 is no point in blinding anyway). Make RSA blinding thread-safe
6719 by remembering the creator's thread ID in rsa->blinding and
6720 having all other threads use local one-time blinding factors
6721 (this requires more computation than sharing rsa->blinding, but
6722 avoids excessive locking; and if an RSA object is not shared
6723 between threads, blinding will still be very fast).
6726 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6728 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6729 via timing by performing a MAC computation even if incorrrect
6730 block cipher padding has been found. This is a countermeasure
6731 against active attacks where the attacker has to distinguish
6732 between bad padding and a MAC verification error. (CVE-2003-0078)
6734 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6735 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6736 Martin Vuagnoux (EPFL, Ilion)]
6738 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6740 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6741 memory from it's contents. This is done with a counter that will
6742 place alternating values in each byte. This can be used to solve
6743 two issues: 1) the removal of calls to memset() by highly optimizing
6744 compilers, and 2) cleansing with other values than 0, since those can
6745 be read through on certain media, for example a swap space on disk.
6748 *) Bugfix: client side session caching did not work with external caching,
6749 because the session->cipher setting was not restored when reloading
6750 from the external cache. This problem was masked, when
6751 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6752 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6755 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6756 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6757 [Zeev Lieber <zeev-l@yahoo.com>]
6759 *) Undo an undocumented change introduced in 0.9.6e which caused
6760 repeated calls to OpenSSL_add_all_ciphers() and
6761 OpenSSL_add_all_digests() to be ignored, even after calling
6765 *) Change the default configuration reader to deal with last line not
6766 being properly terminated.
6769 *) Change X509_NAME_cmp() so it applies the special rules on handling
6770 DN values that are of type PrintableString, as well as RDNs of type
6771 emailAddress where the value has the type ia5String.
6772 [stefank@valicert.com via Richard Levitte]
6774 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6775 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6776 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6777 the bitwise-OR of the two for use by the majority of applications
6778 wanting this behaviour, and update the docs. The documented
6779 behaviour and actual behaviour were inconsistent and had been
6780 changing anyway, so this is more a bug-fix than a behavioural
6782 [Geoff Thorpe, diagnosed by Nadav Har'El]
6784 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6785 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6788 *) Fix initialization code race conditions in
6789 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6790 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6791 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6792 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6793 ssl2_get_cipher_by_char(),
6794 ssl3_get_cipher_by_char().
6795 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6797 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6798 the cached sessions are flushed, as the remove_cb() might use ex_data
6799 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6800 (see [openssl.org #212]).
6801 [Geoff Thorpe, Lutz Jaenicke]
6803 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6804 length, instead of the encoding length to d2i_ASN1_OBJECT.
6807 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6809 *) [In 0.9.6g-engine release:]
6810 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6811 [Lynn Gazis <lgazis@rainbow.com>]
6813 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6815 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6816 and get fix the header length calculation.
6817 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6818 Alon Kantor <alonk@checkpoint.com> (and others),
6821 *) Use proper error handling instead of 'assertions' in buffer
6822 overflow checks added in 0.9.6e. This prevents DoS (the
6823 assertions could call abort()).
6824 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6826 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6828 *) Add various sanity checks to asn1_get_length() to reject
6829 the ASN1 length bytes if they exceed sizeof(long), will appear
6830 negative or the content length exceeds the length of the
6832 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6834 *) Fix cipher selection routines: ciphers without encryption had no flags
6835 for the cipher strength set and where therefore not handled correctly
6836 by the selection routines (PR #130).
6839 *) Fix EVP_dsa_sha macro.
6843 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6844 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6845 that was added in OpenSSL 0.9.6d.
6847 As the countermeasure turned out to be incompatible with some
6848 broken SSL implementations, the new option is part of SSL_OP_ALL.
6849 SSL_OP_ALL is usually employed when compatibility with weird SSL
6850 implementations is desired (e.g. '-bugs' option to 's_client' and
6851 's_server'), so the new option is automatically set in many
6855 *) Changes in security patch:
6857 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6858 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6859 Air Force Materiel Command, USAF, under agreement number
6862 *) Add various sanity checks to asn1_get_length() to reject
6863 the ASN1 length bytes if they exceed sizeof(long), will appear
6864 negative or the content length exceeds the length of the
6865 supplied buffer. (CVE-2002-0659)
6866 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6868 *) Assertions for various potential buffer overflows, not known to
6870 [Ben Laurie (CHATS)]
6872 *) Various temporary buffers to hold ASCII versions of integers were
6873 too small for 64 bit platforms. (CVE-2002-0655)
6874 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6876 *) Remote buffer overflow in SSL3 protocol - an attacker could
6877 supply an oversized session ID to a client. (CVE-2002-0656)
6878 [Ben Laurie (CHATS)]
6880 *) Remote buffer overflow in SSL2 protocol - an attacker could
6881 supply an oversized client master key. (CVE-2002-0656)
6882 [Ben Laurie (CHATS)]
6884 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6886 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6887 encoded as NULL) with id-dsa-with-sha1.
6888 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6890 *) Check various X509_...() return values in apps/req.c.
6891 [Nils Larsch <nla@trustcenter.de>]
6893 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6894 an end-of-file condition would erronously be flagged, when the CRLF
6895 was just at the end of a processed block. The bug was discovered when
6896 processing data through a buffering memory BIO handing the data to a
6897 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6898 <ptsekov@syntrex.com> and Nedelcho Stanev.
6901 *) Implement a countermeasure against a vulnerability recently found
6902 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6903 before application data chunks to avoid the use of known IVs
6904 with data potentially chosen by the attacker.
6907 *) Fix length checks in ssl3_get_client_hello().
6910 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6911 to prevent ssl3_read_internal() from incorrectly assuming that
6912 ssl3_read_bytes() found application data while handshake
6913 processing was enabled when in fact s->s3->in_read_app_data was
6914 merely automatically cleared during the initial handshake.
6915 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6917 *) Fix object definitions for Private and Enterprise: they were not
6918 recognized in their shortname (=lowercase) representation. Extend
6919 obj_dat.pl to issue an error when using undefined keywords instead
6920 of silently ignoring the problem (Svenning Sorensen
6921 <sss@sss.dnsalias.net>).
6924 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6925 generators, i.e. generators other than 2 and 5. (Previously, the
6926 code did not properly initialise the 'add' and 'rem' values to
6927 BN_generate_prime().)
6929 In the new general case, we do not insist that 'generator' is
6930 actually a primitive root: This requirement is rather pointless;
6931 a generator of the order-q subgroup is just as good, if not
6935 *) Map new X509 verification errors to alerts. Discovered and submitted by
6936 Tom Wu <tom@arcot.com>.
6939 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6940 returning non-zero before the data has been completely received
6941 when using non-blocking I/O.
6942 [Bodo Moeller; problem pointed out by John Hughes]
6944 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6945 [Ben Laurie, Lutz Jaenicke]
6947 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6948 Yoram Zahavi <YoramZ@gilian.com>).
6951 *) Add information about CygWin 1.3 and on, and preserve proper
6952 configuration for the versions before that.
6953 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6955 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6956 check whether we deal with a copy of a session and do not delete from
6957 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6958 <izhar@checkpoint.com>.
6961 *) Do not store session data into the internal session cache, if it
6962 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6963 flag is set). Proposed by Aslam <aslam@funk.com>.
6966 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6970 *) [In 0.9.6d-engine release:]
6971 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6972 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6974 *) Add the configuration target linux-s390x.
6975 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6977 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6978 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6979 variable as an indication that a ClientHello message has been
6980 received. As the flag value will be lost between multiple
6981 invocations of ssl3_accept when using non-blocking I/O, the
6982 function may not be aware that a handshake has actually taken
6983 place, thus preventing a new session from being added to the
6986 To avoid this problem, we now set s->new_session to 2 instead of
6987 using a local variable.
6988 [Lutz Jaenicke, Bodo Moeller]
6990 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6991 if the SSL_R_LENGTH_MISMATCH error is detected.
6992 [Geoff Thorpe, Bodo Moeller]
6994 *) New 'shared_ldflag' column in Configure platform table.
6997 *) Fix EVP_CIPHER_mode macro.
6998 ["Dan S. Camper" <dan@bti.net>]
7000 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7001 type, we must throw them away by setting rr->length to 0.
7002 [D P Chang <dpc@qualys.com>]
7004 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7006 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7007 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7008 worked incorrectly for those cases where range = 10..._2 and
7009 3*range is two bits longer than range.)
7012 *) Only add signing time to PKCS7 structures if it is not already
7016 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7017 OBJ_ld_ce should be OBJ_id_ce.
7018 Also some ip-pda OIDs in crypto/objects/objects.txt were
7019 incorrect (cf. RFC 3039).
7020 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7022 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7023 returns early because it has nothing to do.
7024 [Andy Schneider <andy.schneider@bjss.co.uk>]
7026 *) [In 0.9.6c-engine release:]
7027 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7028 [Andy Schneider <andy.schneider@bjss.co.uk>]
7030 *) [In 0.9.6c-engine release:]
7031 Add support for Cryptographic Appliance's keyserver technology.
7032 (Use engine 'keyclient')
7033 [Cryptographic Appliances and Geoff Thorpe]
7035 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7036 is called via tools/c89.sh because arguments have to be
7037 rearranged (all '-L' options must appear before the first object
7039 [Richard Shapiro <rshapiro@abinitio.com>]
7041 *) [In 0.9.6c-engine release:]
7042 Add support for Broadcom crypto accelerator cards, backported
7044 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7046 *) [In 0.9.6c-engine release:]
7047 Add support for SureWare crypto accelerator cards from
7048 Baltimore Technologies. (Use engine 'sureware')
7049 [Baltimore Technologies and Mark Cox]
7051 *) [In 0.9.6c-engine release:]
7052 Add support for crypto accelerator cards from Accelerated
7053 Encryption Processing, www.aep.ie. (Use engine 'aep')
7054 [AEP Inc. and Mark Cox]
7056 *) Add a configuration entry for gcc on UnixWare.
7057 [Gary Benson <gbenson@redhat.com>]
7059 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7060 messages are stored in a single piece (fixed-length part and
7061 variable-length part combined) and fix various bugs found on the way.
7064 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7065 instead. BIO_gethostbyname() does not know what timeouts are
7066 appropriate, so entries would stay in cache even when they have
7068 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7070 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7071 faced with a pathologically small ClientHello fragment that does
7072 not contain client_version: Instead of aborting with an error,
7073 simply choose the highest available protocol version (i.e.,
7074 TLS 1.0 unless it is disabled). In practice, ClientHello
7075 messages are never sent like this, but this change gives us
7076 strictly correct behaviour at least for TLS.
7079 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7080 never resets s->method to s->ctx->method when called from within
7081 one of the SSL handshake functions.
7082 [Bodo Moeller; problem pointed out by Niko Baric]
7084 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7085 (sent using the client's version number) if client_version is
7086 smaller than the protocol version in use. Also change
7087 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7088 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7089 the client will at least see that alert.
7092 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7096 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7097 client receives HelloRequest while in a handshake.
7098 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7100 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7101 should end in 'break', not 'goto end' which circumvents various
7102 cleanups done in state SSL_ST_OK. But session related stuff
7103 must be disabled for SSL_ST_OK in the case that we just sent a
7106 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7107 before just sending a HelloRequest.
7108 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7110 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7111 reveal whether illegal block cipher padding was found or a MAC
7112 verification error occured. (Neither SSLerr() codes nor alerts
7113 are directly visible to potential attackers, but the information
7114 may leak via logfiles.)
7116 Similar changes are not required for the SSL 2.0 implementation
7117 because the number of padding bytes is sent in clear for SSL 2.0,
7118 and the extra bytes are just ignored. However ssl/s2_pkt.c
7119 failed to verify that the purported number of padding bytes is in
7123 *) Add OpenUNIX-8 support including shared libraries
7124 (Boyd Lynn Gerber <gerberb@zenez.com>).
7127 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7128 'wristwatch attack' using huge encoding parameters (cf.
7129 James H. Manger's CRYPTO 2001 paper). Note that the
7130 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7131 encoding parameters and hence was not vulnerable.
7134 *) BN_sqr() bug fix.
7135 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7137 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7138 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7139 followed by modular reduction.
7140 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7142 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7143 equivalent based on BN_pseudo_rand() instead of BN_rand().
7146 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7147 This function was broken, as the check for a new client hello message
7148 to handle SGC did not allow these large messages.
7149 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7152 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7155 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7156 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7159 *) Rework the configuration and shared library support for Tru64 Unix.
7160 The configuration part makes use of modern compiler features and
7161 still retains old compiler behavior for those that run older versions
7162 of the OS. The shared library support part includes a variant that
7163 uses the RPATH feature, and is available through the special
7164 configuration target "alpha-cc-rpath", which will never be selected
7166 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7168 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7169 with the same message size as in ssl3_get_certificate_request().
7170 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7171 messages might inadvertently be reject as too long.
7172 [Petr Lampa <lampa@fee.vutbr.cz>]
7174 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7177 *) Modified SSL library such that the verify_callback that has been set
7178 specificly for an SSL object with SSL_set_verify() is actually being
7179 used. Before the change, a verify_callback set with this function was
7180 ignored and the verify_callback() set in the SSL_CTX at the time of
7181 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7182 to allow the necessary settings.
7185 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7186 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7187 done automatically (in contradiction to the requirements of the C
7188 standard). This made problems when used from OpenSSH.
7191 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7192 dh->length and always used
7194 BN_rand_range(priv_key, dh->p).
7196 BN_rand_range() is not necessary for Diffie-Hellman, and this
7197 specific range makes Diffie-Hellman unnecessarily inefficient if
7198 dh->length (recommended exponent length) is much smaller than the
7199 length of dh->p. We could use BN_rand_range() if the order of
7200 the subgroup was stored in the DH structure, but we only have
7205 BN_rand(priv_key, l, ...)
7207 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7213 RSA_eay_public_encrypt
7214 RSA_eay_private_decrypt
7215 RSA_eay_private_encrypt (signing)
7216 RSA_eay_public_decrypt (signature verification)
7218 (default implementations for RSA_public_encrypt,
7219 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7220 always reject numbers >= n.
7223 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7224 to synchronize access to 'locking_thread'. This is necessary on
7225 systems where access to 'locking_thread' (an 'unsigned long'
7226 variable) is not atomic.
7229 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7230 *before* setting the 'crypto_lock_rand' flag. The previous code had
7231 a race condition if 0 is a valid thread ID.
7232 [Travis Vitek <vitek@roguewave.com>]
7234 *) Add support for shared libraries under Irix.
7235 [Albert Chin-A-Young <china@thewrittenword.com>]
7237 *) Add configuration option to build on Linux on both big-endian and
7239 [Ralf Baechle <ralf@uni-koblenz.de>]
7241 *) Add the possibility to create shared libraries on HP-UX.
7244 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7246 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7247 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7248 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7249 PRNG state recovery was possible based on the output of
7250 one PRNG request appropriately sized to gain knowledge on
7251 'md' followed by enough consecutive 1-byte PRNG requests
7252 to traverse all of 'state'.
7254 1. When updating 'md_local' (the current thread's copy of 'md')
7255 during PRNG output generation, hash all of the previous
7256 'md_local' value, not just the half used for PRNG output.
7258 2. Make the number of bytes from 'state' included into the hash
7259 independent from the number of PRNG bytes requested.
7261 The first measure alone would be sufficient to avoid
7262 Markku-Juhani's attack. (Actually it had never occurred
7263 to me that the half of 'md_local' used for chaining was the
7264 half from which PRNG output bytes were taken -- I had always
7265 assumed that the secret half would be used.) The second
7266 measure makes sure that additional data from 'state' is never
7267 mixed into 'md_local' in small portions; this heuristically
7268 further strengthens the PRNG.
7271 *) Fix crypto/bn/asm/mips3.s.
7274 *) When only the key is given to "enc", the IV is undefined. Print out
7275 an error message in this case.
7278 *) Handle special case when X509_NAME is empty in X509 printing routines.
7281 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7282 positive and less than q.
7285 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7286 used: it isn't thread safe and the add_lock_callback should handle
7288 [Paul Rose <Paul.Rose@bridge.com>]
7290 *) Verify that incoming data obeys the block size in
7291 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7295 [Ulf Möller, Bodo Möller]
7297 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7298 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7299 when fixing the server behaviour for backwards-compatible 'client
7300 hello' messages. (Note that the attack is impractical against
7301 SSL 3.0 and TLS 1.0 anyway because length and version checking
7302 means that the probability of guessing a valid ciphertext is
7303 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7306 Before 0.9.5, the countermeasure (hide the error by generating a
7307 random 'decryption result') did not work properly because
7308 ERR_clear_error() was missing, meaning that SSL_get_error() would
7309 detect the supposedly ignored error.
7311 Both problems are now fixed.
7314 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7315 (previously it was 1024).
7318 *) Fix for compatibility mode trust settings: ignore trust settings
7319 unless some valid trust or reject settings are present.
7322 *) Fix for blowfish EVP: its a variable length cipher.
7325 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7326 parameters in DSA public key structures and return an error in the
7327 DSA routines if parameters are absent.
7330 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7331 in the current directory if neither $RANDFILE nor $HOME was set.
7332 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7333 caused some confusion to Windows users who haven't defined $HOME.
7334 Thus RAND_file_name() is changed again: e_os.h can define a
7335 DEFAULT_HOME, which will be used if $HOME is not set.
7336 For Windows, we use "C:"; on other platforms, we still require
7337 environment variables.
7339 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7340 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7341 having multiple threads call RAND_poll() concurrently.
7344 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7345 combination of a flag and a thread ID variable.
7346 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7347 flag), *other* threads can enter ssleay_add_bytes without obeying
7348 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7349 that they do not hold after the first thread unsets add_do_not_lock).
7352 *) Change bctest again: '-x' expressions are not available in all
7356 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7358 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7359 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7361 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7362 the default extension for executables, if any. Also, make the perl
7363 scripts that use symlink() to test if it really exists and use "cp"
7364 if it doesn't. All this made OpenSSL compilable and installable in
7368 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7369 If SEQUENCE is length is indefinite just set c->slen to the total
7370 amount of data available.
7371 [Steve Henson, reported by shige@FreeBSD.org]
7372 [This change does not apply to 0.9.7.]
7374 *) Change bctest to avoid here-documents inside command substitution
7375 (workaround for FreeBSD /bin/sh bug).
7376 For compatibility with Ultrix, avoid shell functions (introduced
7377 in the bctest version that searches along $PATH).
7380 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7381 with des_encrypt() defined on some operating systems, like Solaris
7385 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7386 On the Importance of Eliminating Errors in Cryptographic
7387 Computations, J. Cryptology 14 (2001) 2, 101-119,
7388 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7391 *) MIPS assembler BIGNUM division bug fix.
7394 *) Disabled incorrect Alpha assembler code.
7397 *) Fix PKCS#7 decode routines so they correctly update the length
7398 after reading an EOC for the EXPLICIT tag.
7400 [This change does not apply to 0.9.7.]
7402 *) Fix bug in PKCS#12 key generation routines. This was triggered
7403 if a 3DES key was generated with a 0 initial byte. Include
7404 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7405 (but broken) behaviour.
7408 *) Enhance bctest to search for a working bc along $PATH and print
7410 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7412 *) Fix memory leaks in err.c: free err_data string if necessary;
7413 don't write to the wrong index in ERR_set_error_data.
7416 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7420 *) Replace rdtsc with _emit statements for VC++ version 5.
7421 [Jeremy Cooper <jeremy@baymoo.org>]
7423 *) Make it possible to reuse SSLv2 sessions.
7426 *) In copy_email() check for >= 0 as a return value for
7427 X509_NAME_get_index_by_NID() since 0 is a valid index.
7428 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7430 *) Avoid coredump with unsupported or invalid public keys by checking if
7431 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7432 PKCS7_verify() fails with non detached data.
7435 *) Don't use getenv in library functions when run as setuid/setgid.
7436 New function OPENSSL_issetugid().
7439 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7440 due to incorrect handling of multi-threading:
7442 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7444 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7446 3. Count how many times MemCheck_off() has been called so that
7447 nested use can be treated correctly. This also avoids
7448 inband-signalling in the previous code (which relied on the
7449 assumption that thread ID 0 is impossible).
7452 *) Add "-rand" option also to s_client and s_server.
7455 *) Fix CPU detection on Irix 6.x.
7456 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7457 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7459 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7462 [This change does not apply to 0.9.7.]
7464 *) Use the cached encoding of an X509_NAME structure rather than
7465 copying it. This is apparently the reason for the libsafe "errors"
7466 but the code is actually correct.
7469 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7470 Bleichenbacher's DSA attack.
7471 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7472 to be set and top=0 forces the highest bit to be set; top=-1 is new
7473 and leaves the highest bit random.
7474 [Ulf Moeller, Bodo Moeller]
7476 *) In the NCONF_...-based implementations for CONF_... queries
7477 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7478 a temporary CONF structure with the data component set to NULL
7479 (which gives segmentation faults in lh_retrieve).
7480 Instead, use NULL for the CONF pointer in CONF_get_string and
7481 CONF_get_number (which may use environment variables) and directly
7482 return NULL from CONF_get_section.
7485 *) Fix potential buffer overrun for EBCDIC.
7488 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7489 keyUsage if basicConstraints absent for a CA.
7492 *) Make SMIME_write_PKCS7() write mail header values with a format that
7493 is more generally accepted (no spaces before the semicolon), since
7494 some programs can't parse those values properly otherwise. Also make
7495 sure BIO's that break lines after each write do not create invalid
7499 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7500 macros previously used would not encode an empty SEQUENCE OF
7501 and break the signature.
7503 [This change does not apply to 0.9.7.]
7505 *) Zero the premaster secret after deriving the master secret in
7509 *) Add some EVP_add_digest_alias registrations (as found in
7510 OpenSSL_add_all_digests()) to SSL_library_init()
7511 aka OpenSSL_add_ssl_algorithms(). This provides improved
7512 compatibility with peers using X.509 certificates
7513 with unconventional AlgorithmIdentifier OIDs.
7516 *) Fix for Irix with NO_ASM.
7517 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7519 *) ./config script fixes.
7520 [Ulf Moeller, Richard Levitte]
7522 *) Fix 'openssl passwd -1'.
7525 *) Change PKCS12_key_gen_asc() so it can cope with non null
7526 terminated strings whose length is passed in the passlen
7527 parameter, for example from PEM callbacks. This was done
7528 by adding an extra length parameter to asc2uni().
7529 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7531 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7532 call failed, free the DSA structure.
7535 *) Fix to uni2asc() to cope with zero length Unicode strings.
7536 These are present in some PKCS#12 files.
7539 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7540 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7541 when writing a 32767 byte record.
7542 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7544 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7545 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7547 (RSA objects have a reference count access to which is protected
7548 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7549 so they are meant to be shared between threads.)
7550 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7551 "Reddie, Steven" <Steven.Reddie@ca.com>]
7553 *) Fix a deadlock in CRYPTO_mem_leaks().
7556 *) Use better test patterns in bntest.
7559 *) rand_win.c fix for Borland C.
7562 *) BN_rshift bugfix for n == 0.
7565 *) Add a 'bctest' script that checks for some known 'bc' bugs
7566 so that 'make test' does not abort just because 'bc' is broken.
7569 *) Store verify_result within SSL_SESSION also for client side to
7570 avoid potential security hole. (Re-used sessions on the client side
7571 always resulted in verify_result==X509_V_OK, not using the original
7572 result of the server certificate verification.)
7575 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7576 SSL3_RT_APPLICATION_DATA, return 0.
7577 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7581 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7582 releases, have been re-implemented by renaming the previous
7583 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7584 and ssl3_read_internal, respectively, and adding 'peek' parameters
7585 to them. The new ssl[23]_{read,peek} functions are calls to
7586 ssl[23]_read_internal with the 'peek' flag set appropriately.
7587 A 'peek' parameter has also been added to ssl3_read_bytes, which
7588 does the actual work for ssl3_read_internal.
7591 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7592 the method-specific "init()" handler. Also clean up ex_data after
7593 calling the method-specific "finish()" handler. Previously, this was
7594 happening the other way round.
7597 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7598 The previous value, 12, was not always sufficient for BN_mod_exp().
7601 *) Make sure that shared libraries get the internal name engine with
7602 the full version number and not just 0. This should mark the
7603 shared libraries as not backward compatible. Of course, this should
7604 be changed again when we can guarantee backward binary compatibility.
7607 *) Fix typo in get_cert_by_subject() in by_dir.c
7608 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7610 *) Rework the system to generate shared libraries:
7612 - Make note of the expected extension for the shared libraries and
7613 if there is a need for symbolic links from for example libcrypto.so.0
7614 to libcrypto.so.0.9.7. There is extended info in Configure for
7617 - Make as few rebuilds of the shared libraries as possible.
7619 - Still avoid linking the OpenSSL programs with the shared libraries.
7621 - When installing, install the shared libraries separately from the
7625 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7627 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7628 and not in SSL_clear because the latter is also used by the
7629 accept/connect functions; previously, the settings made by
7630 SSL_set_read_ahead would be lost during the handshake.
7631 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7633 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7634 Previously, it would create entries for disabled algorithms no
7638 *) Added several new manual pages for SSL_* function.
7641 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7643 *) In ssl23_get_client_hello, generate an error message when faced
7644 with an initial SSL 3.0/TLS record that is too small to contain the
7645 first two bytes of the ClientHello message, i.e. client_version.
7646 (Note that this is a pathologic case that probably has never happened
7647 in real life.) The previous approach was to use the version number
7648 from the record header as a substitute; but our protocol choice
7649 should not depend on that one because it is not authenticated
7650 by the Finished messages.
7653 *) More robust randomness gathering functions for Windows.
7654 [Jeffrey Altman <jaltman@columbia.edu>]
7656 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7657 not set then we don't setup the error code for issuer check errors
7658 to avoid possibly overwriting other errors which the callback does
7659 handle. If an application does set the flag then we assume it knows
7660 what it is doing and can handle the new informational codes
7664 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7665 a general "ANY" type, as such it should be able to decode anything
7666 including tagged types. However it didn't check the class so it would
7667 wrongly interpret tagged types in the same way as their universal
7668 counterpart and unknown types were just rejected. Changed so that the
7669 tagged and unknown types are handled in the same way as a SEQUENCE:
7670 that is the encoding is stored intact. There is also a new type
7671 "V_ASN1_OTHER" which is used when the class is not universal, in this
7672 case we have no idea what the actual type is so we just lump them all
7676 *) On VMS, stdout may very well lead to a file that is written to
7677 in a record-oriented fashion. That means that every write() will
7678 write a separate record, which will be read separately by the
7679 programs trying to read from it. This can be very confusing.
7681 The solution is to put a BIO filter in the way that will buffer
7682 text until a linefeed is reached, and then write everything a
7683 line at a time, so every record written will be an actual line,
7684 not chunks of lines and not (usually doesn't happen, but I've
7685 seen it once) several lines in one record. BIO_f_linebuffer() is
7688 Currently, it's a VMS-only method, because that's where it has
7689 been tested well enough.
7692 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7693 it can return incorrect results.
7694 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7695 but it was in 0.9.6-beta[12].)
7698 *) Disable the check for content being present when verifying detached
7699 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7700 include zero length content when signing messages.
7703 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7704 BIO_ctrl (for BIO pairs).
7707 *) Add DSO method for VMS.
7710 *) Bug fix: Montgomery multiplication could produce results with the
7714 *) Add RPM specification openssl.spec and modify it to build three
7715 packages. The default package contains applications, application
7716 documentation and run-time libraries. The devel package contains
7717 include files, static libraries and function documentation. The
7718 doc package contains the contents of the doc directory. The original
7719 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7722 *) Add a large number of documentation files for many SSL routines.
7723 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7725 *) Add a configuration entry for Sony News 4.
7726 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7728 *) Don't set the two most significant bits to one when generating a
7729 random number < q in the DSA library.
7732 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7733 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7734 the underlying transport is blocking) if a handshake took place.
7735 (The default behaviour is needed by applications such as s_client
7736 and s_server that use select() to determine when to use SSL_read;
7737 but for applications that know in advance when to expect data, it
7738 just makes things more complicated.)
7741 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7745 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7746 work better on such systems.
7747 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7749 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7750 Update PKCS12_parse() so it copies the friendlyName and the
7751 keyid to the certificates aux info.
7754 *) Fix bug in PKCS7_verify() which caused an infinite loop
7755 if there was more than one signature.
7756 [Sven Uszpelkat <su@celocom.de>]
7758 *) Major change in util/mkdef.pl to include extra information
7759 about each symbol, as well as presenting variables as well
7760 as functions. This change means that there's n more need
7761 to rebuild the .num files when some algorithms are excluded.
7764 *) Allow the verify time to be set by an application,
7765 rather than always using the current time.
7768 *) Phase 2 verify code reorganisation. The certificate
7769 verify code now looks up an issuer certificate by a
7770 number of criteria: subject name, authority key id
7771 and key usage. It also verifies self signed certificates
7772 by the same criteria. The main comparison function is
7773 X509_check_issued() which performs these checks.
7775 Lot of changes were necessary in order to support this
7776 without completely rewriting the lookup code.
7778 Authority and subject key identifier are now cached.
7780 The LHASH 'certs' is X509_STORE has now been replaced
7781 by a STACK_OF(X509_OBJECT). This is mainly because an
7782 LHASH can't store or retrieve multiple objects with
7783 the same hash value.
7785 As a result various functions (which were all internal
7786 use only) have changed to handle the new X509_STORE
7787 structure. This will break anything that messed round
7788 with X509_STORE internally.
7790 The functions X509_STORE_add_cert() now checks for an
7791 exact match, rather than just subject name.
7793 The X509_STORE API doesn't directly support the retrieval
7794 of multiple certificates matching a given criteria, however
7795 this can be worked round by performing a lookup first
7796 (which will fill the cache with candidate certificates)
7797 and then examining the cache for matches. This is probably
7798 the best we can do without throwing out X509_LOOKUP
7799 entirely (maybe later...).
7801 The X509_VERIFY_CTX structure has been enhanced considerably.
7803 All certificate lookup operations now go via a get_issuer()
7804 callback. Although this currently uses an X509_STORE it
7805 can be replaced by custom lookups. This is a simple way
7806 to bypass the X509_STORE hackery necessary to make this
7807 work and makes it possible to use more efficient techniques
7808 in future. A very simple version which uses a simple
7809 STACK for its trusted certificate store is also provided
7810 using X509_STORE_CTX_trusted_stack().
7812 The verify_cb() and verify() callbacks now have equivalents
7813 in the X509_STORE_CTX structure.
7815 X509_STORE_CTX also has a 'flags' field which can be used
7816 to customise the verify behaviour.
7819 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7820 excludes S/MIME capabilities.
7823 *) When a certificate request is read in keep a copy of the
7824 original encoding of the signed data and use it when outputing
7825 again. Signatures then use the original encoding rather than
7826 a decoded, encoded version which may cause problems if the
7827 request is improperly encoded.
7830 *) For consistency with other BIO_puts implementations, call
7831 buffer_write(b, ...) directly in buffer_puts instead of calling
7834 In BIO_puts, increment b->num_write as in BIO_write.
7835 [Peter.Sylvester@EdelWeb.fr]
7837 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7838 BN_zero, we may not return a BIGNUM with an array consisting of
7842 *) Avoid calling abort() from within the library when problems are
7843 detected, except if preprocessor symbols have been defined
7844 (such as REF_CHECK, BN_DEBUG etc.).
7847 *) New openssl application 'rsautl'. This utility can be
7848 used for low level RSA operations. DER public key
7849 BIO/fp routines also added.
7852 *) New Configure entry and patches for compiling on QNX 4.
7853 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7855 *) A demo state-machine implementation was sponsored by
7856 Nuron (http://www.nuron.com/) and is now available in
7857 demos/state_machine.
7860 *) New options added to the 'dgst' utility for signature
7861 generation and verification.
7864 *) Unrecognized PKCS#7 content types are now handled via a
7865 catch all ASN1_TYPE structure. This allows unsupported
7866 types to be stored as a "blob" and an application can
7867 encode and decode it manually.
7870 *) Fix various signed/unsigned issues to make a_strex.c
7872 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7874 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7875 length if passed a buffer. ASN1_INTEGER_to_BN failed
7876 if passed a NULL BN and its argument was negative.
7877 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7879 *) Modification to PKCS#7 encoding routines to output definite
7880 length encoding. Since currently the whole structures are in
7881 memory there's not real point in using indefinite length
7882 constructed encoding. However if OpenSSL is compiled with
7883 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7886 *) Added BIO_vprintf() and BIO_vsnprintf().
7889 *) Added more prefixes to parse for in the the strings written
7890 through a logging bio, to cover all the levels that are available
7891 through syslog. The prefixes are now:
7893 PANIC, EMERG, EMR => LOG_EMERG
7894 ALERT, ALR => LOG_ALERT
7895 CRIT, CRI => LOG_CRIT
7896 ERROR, ERR => LOG_ERR
7897 WARNING, WARN, WAR => LOG_WARNING
7898 NOTICE, NOTE, NOT => LOG_NOTICE
7899 INFO, INF => LOG_INFO
7900 DEBUG, DBG => LOG_DEBUG
7902 and as before, if none of those prefixes are present at the
7903 beginning of the string, LOG_ERR is chosen.
7905 On Win32, the LOG_* levels are mapped according to this:
7907 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7908 LOG_WARNING => EVENTLOG_WARNING_TYPE
7909 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7913 *) Made it possible to reconfigure with just the configuration
7914 argument "reconf" or "reconfigure". The command line arguments
7915 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7916 and are retrieved from there when reconfiguring.
7920 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7922 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7925 *) The obj_dat.pl script was messing up the sorting of object
7926 names. The reason was that it compared the quoted version
7927 of strings as a result "OCSP" > "OCSP Signing" because
7928 " > SPACE. Changed script to store unquoted versions of
7929 names and add quotes on output. It was also omitting some
7930 names from the lookup table if they were given a default
7931 value (that is if SN is missing it is given the same
7932 value as LN and vice versa), these are now added on the
7933 grounds that if an object has a name we should be able to
7934 look it up. Finally added warning output when duplicate
7935 short or long names are found.
7938 *) Changes needed for Tandem NSK.
7939 [Scott Uroff <scott@xypro.com>]
7941 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7942 RSA_padding_check_SSLv23(), special padding was never detected
7943 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7944 version rollback attacks was not effective.
7946 In s23_clnt.c, don't use special rollback-attack detection padding
7947 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7948 client; similarly, in s23_srvr.c, don't do the rollback check if
7949 SSL 2.0 is the only protocol enabled in the server.
7952 *) Make it possible to get hexdumps of unprintable data with 'openssl
7953 asn1parse'. By implication, the functions ASN1_parse_dump() and
7954 BIO_dump_indent() are added.
7957 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7958 these print out strings and name structures based on various
7959 flags including RFC2253 support and proper handling of
7960 multibyte characters. Added options to the 'x509' utility
7961 to allow the various flags to be set.
7964 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7965 Also change the functions X509_cmp_current_time() and
7966 X509_gmtime_adj() work with an ASN1_TIME structure,
7967 this will enable certificates using GeneralizedTime in validity
7968 dates to be checked.
7971 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7972 negative public key encodings) on by default,
7973 NO_NEG_PUBKEY_BUG can be set to disable it.
7976 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7977 content octets. An i2c_ASN1_OBJECT is unnecessary because
7978 the encoding can be trivially obtained from the structure.
7981 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7982 not read locks (CRYPTO_r_[un]lock).
7985 *) A first attempt at creating official support for shared
7986 libraries through configuration. I've kept it so the
7987 default is static libraries only, and the OpenSSL programs
7988 are always statically linked for now, but there are
7989 preparations for dynamic linking in place.
7990 This has been tested on Linux and Tru64.
7993 *) Randomness polling function for Win9x, as described in:
7994 Peter Gutmann, Software Generation of Practically Strong
7998 *) Fix so PRNG is seeded in req if using an already existing
8002 *) New options to smime application. -inform and -outform
8003 allow alternative formats for the S/MIME message including
8004 PEM and DER. The -content option allows the content to be
8005 specified separately. This should allow things like Netscape
8006 form signing output easier to verify.
8009 *) Fix the ASN1 encoding of tags using the 'long form'.
8012 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8013 STRING types. These convert content octets to and from the
8014 underlying type. The actual tag and length octets are
8015 already assumed to have been read in and checked. These
8016 are needed because all other string types have virtually
8017 identical handling apart from the tag. By having versions
8018 of the ASN1 functions that just operate on content octets
8019 IMPLICIT tagging can be handled properly. It also allows
8020 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8021 and ASN1_INTEGER are identical apart from the tag.
8024 *) Change the handling of OID objects as follows:
8026 - New object identifiers are inserted in objects.txt, following
8027 the syntax given in objects.README.
8028 - objects.pl is used to process obj_mac.num and create a new
8030 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8033 This is currently kind of a hack, and the perl code in objects.pl
8034 isn't very elegant, but it works as I intended. The simplest way
8035 to check that it worked correctly is to look in obj_dat.h and
8036 check the array nid_objs and make sure the objects haven't moved
8037 around (this is important!). Additions are OK, as well as
8038 consistent name changes.
8041 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8044 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8045 The given file adds to whatever has already been seeded into the
8046 random pool through the RANDFILE configuration file option or
8047 environment variable, or the default random state file.
8050 *) mkstack.pl now sorts each macro group into lexical order.
8051 Previously the output order depended on the order the files
8052 appeared in the directory, resulting in needless rewriting
8056 *) Patches to make OpenSSL compile under Win32 again. Mostly
8057 work arounds for the VC++ problem that it treats func() as
8058 func(void). Also stripped out the parts of mkdef.pl that
8059 added extra typesafe functions: these no longer exist.
8062 *) Reorganisation of the stack code. The macros are now all
8063 collected in safestack.h . Each macro is defined in terms of
8064 a "stack macro" of the form SKM_<name>(type, a, b). The
8065 DEBUG_SAFESTACK is now handled in terms of function casts,
8066 this has the advantage of retaining type safety without the
8067 use of additional functions. If DEBUG_SAFESTACK is not defined
8068 then the non typesafe macros are used instead. Also modified the
8069 mkstack.pl script to handle the new form. Needs testing to see
8070 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8071 the default if no major problems. Similar behaviour for ASN1_SET_OF
8072 and PKCS12_STACK_OF.
8075 *) When some versions of IIS use the 'NET' form of private key the
8076 key derivation algorithm is different. Normally MD5(password) is
8077 used as a 128 bit RC4 key. In the modified case
8078 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8079 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8080 as the old Netscape_RSA functions except they have an additional
8081 'sgckey' parameter which uses the modified algorithm. Also added
8082 an -sgckey command line option to the rsa utility. Thanks to
8083 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8084 algorithm to openssl-dev.
8087 *) The evp_local.h macros were using 'c.##kname' which resulted in
8088 invalid expansion on some systems (SCO 5.0.5 for example).
8089 Corrected to 'c.kname'.
8090 [Phillip Porch <root@theporch.com>]
8092 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8093 a STACK of email addresses from a certificate or request, these look
8094 in the subject name and the subject alternative name extensions and
8095 omit any duplicate addresses.
8098 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8099 This makes DSA verification about 2 % faster.
8102 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8103 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8104 plus overhead for 1024 bit moduli).
8105 This makes exponentiations about 0.5 % faster for 1024 bit
8106 exponents (as measured by "openssl speed rsa2048").
8109 *) Rename memory handling macros to avoid conflicts with other
8111 Malloc => OPENSSL_malloc
8112 Malloc_locked => OPENSSL_malloc_locked
8113 Realloc => OPENSSL_realloc
8114 Free => OPENSSL_free
8117 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8118 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8121 *) CygWin32 support.
8122 [John Jarvie <jjarvie@newsguy.com>]
8124 *) The type-safe stack code has been rejigged. It is now only compiled
8125 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8126 by default all type-specific stack functions are "#define"d back to
8127 standard stack functions. This results in more streamlined output
8128 but retains the type-safety checking possibilities of the original
8132 *) The STACK code has been cleaned up, and certain type declarations
8133 that didn't make a lot of sense have been brought in line. This has
8134 also involved a cleanup of sorts in safestack.h to more correctly
8135 map type-safe stack functions onto their plain stack counterparts.
8136 This work has also resulted in a variety of "const"ifications of
8137 lots of the code, especially "_cmp" operations which should normally
8138 be prototyped with "const" parameters anyway.
8141 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8142 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8143 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8144 where all of 'md' is used each time the PRNG is used, but 'state'
8145 is used only indexed by a cyclic counter. As entropy may not be
8146 well distributed from the beginning, 'md' is important as a
8147 chaining variable. However, the output function chains only half
8148 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8149 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8150 in all of 'state' being rewritten, with the new values depending
8151 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8154 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8155 the handshake is continued after ssl_verify_cert_chain();
8156 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8157 can lead to 'unexplainable' connection aborts later.
8158 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8160 *) Major EVP API cipher revision.
8161 Add hooks for extra EVP features. This allows various cipher
8162 parameters to be set in the EVP interface. Support added for variable
8163 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8164 setting of RC2 and RC5 parameters.
8166 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8169 Remove lots of duplicated code from the EVP library. For example *every*
8170 cipher init() function handles the 'iv' in the same way according to the
8171 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8172 for CFB and OFB modes they zero ctx->num.
8174 New functionality allows removal of S/MIME code RC2 hack.
8176 Most of the routines have the same form and so can be declared in terms
8179 By shifting this to the top level EVP_CipherInit() it can be removed from
8180 all individual ciphers. If the cipher wants to handle IVs or keys
8181 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8184 Change lots of functions like EVP_EncryptUpdate() to now return a
8185 value: although software versions of the algorithms cannot fail
8186 any installed hardware versions can.
8189 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8190 this option is set, tolerate broken clients that send the negotiated
8191 protocol version number instead of the requested protocol version
8195 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8196 i.e. non-zero for export ciphersuites, zero otherwise.
8197 Previous versions had this flag inverted, inconsistent with
8198 rsa_tmp_cb (..._TMP_RSA_CB).
8199 [Bodo Moeller; problem reported by Amit Chopra]
8201 *) Add missing DSA library text string. Work around for some IIS
8202 key files with invalid SEQUENCE encoding.
8205 *) Add a document (doc/standards.txt) that list all kinds of standards
8206 and so on that are implemented in OpenSSL.
8209 *) Enhance c_rehash script. Old version would mishandle certificates
8210 with the same subject name hash and wouldn't handle CRLs at all.
8211 Added -fingerprint option to crl utility, to support new c_rehash
8215 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8218 *) Fix for SSL server purpose checking. Server checking was
8219 rejecting certificates which had extended key usage present
8220 but no ssl client purpose.
8221 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8223 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8224 is a little unclear about how a blank password is handled.
8225 Since the password in encoded as a BMPString with terminating
8226 double NULL a zero length password would end up as just the
8227 double NULL. However no password at all is different and is
8228 handled differently in the PKCS#12 key generation code. NS
8229 treats a blank password as zero length. MSIE treats it as no
8230 password on export: but it will try both on import. We now do
8231 the same: PKCS12_parse() tries zero length and no password if
8232 the password is set to "" or NULL (NULL is now a valid password:
8233 it wasn't before) as does the pkcs12 application.
8236 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8237 perror when PEM_read_bio_X509_REQ fails, the error message must
8238 be obtained from the error queue.
8241 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8242 it in ERR_remove_state if appropriate, and change ERR_get_state
8243 accordingly to avoid race conditions (this is necessary because
8244 thread_hash is no longer constant once set).
8247 *) Bugfix for linux-elf makefile.one.
8250 *) RSA_get_default_method() will now cause a default
8251 RSA_METHOD to be chosen if one doesn't exist already.
8252 Previously this was only set during a call to RSA_new()
8253 or RSA_new_method(NULL) meaning it was possible for
8254 RSA_get_default_method() to return NULL.
8257 *) Added native name translation to the existing DSO code
8258 that will convert (if the flag to do so is set) filenames
8259 that are sufficiently small and have no path information
8260 into a canonical native form. Eg. "blah" converted to
8261 "libblah.so" or "blah.dll" etc.
8264 *) New function ERR_error_string_n(e, buf, len) which is like
8265 ERR_error_string(e, buf), but writes at most 'len' bytes
8266 including the 0 terminator. For ERR_error_string_n, 'buf'
8268 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8270 *) CONF library reworked to become more general. A new CONF
8271 configuration file reader "class" is implemented as well as a
8272 new functions (NCONF_*, for "New CONF") to handle it. The now
8273 old CONF_* functions are still there, but are reimplemented to
8274 work in terms of the new functions. Also, a set of functions
8275 to handle the internal storage of the configuration data is
8276 provided to make it easier to write new configuration file
8277 reader "classes" (I can definitely see something reading a
8278 configuration file in XML format, for example), called _CONF_*,
8279 or "the configuration storage API"...
8281 The new configuration file reading functions are:
8283 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8284 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8286 NCONF_default, NCONF_WIN32
8288 NCONF_dump_fp, NCONF_dump_bio
8290 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8291 NCONF_new creates a new CONF object. This works in the same way
8292 as other interfaces in OpenSSL, like the BIO interface.
8293 NCONF_dump_* dump the internal storage of the configuration file,
8294 which is useful for debugging. All other functions take the same
8295 arguments as the old CONF_* functions wth the exception of the
8296 first that must be a `CONF *' instead of a `LHASH *'.
8298 To make it easer to use the new classes with the old CONF_* functions,
8299 the function CONF_set_default_method is provided.
8302 *) Add '-tls1' option to 'openssl ciphers', which was already
8303 mentioned in the documentation but had not been implemented.
8304 (This option is not yet really useful because even the additional
8305 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8308 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8309 OpenSSL-based applications) load shared libraries and bind to
8310 them in a portable way.
8311 [Geoff Thorpe, with contributions from Richard Levitte]
8313 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8315 *) Make sure _lrotl and _lrotr are only used with MSVC.
8317 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8318 (the default implementation of RAND_status).
8320 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8321 to '-clrext' (= clear extensions), as intended and documented.
8322 [Bodo Moeller; inconsistency pointed out by Michael Attili
8325 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8326 was larger than the MD block size.
8327 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8329 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8330 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8331 using the passed key: if the passed key was a private key the result
8332 of X509_print(), for example, would be to print out all the private key
8336 *) des_quad_cksum() byte order bug fix.
8337 [Ulf Möller, using the problem description in krb4-0.9.7, where
8338 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8340 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8342 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8344 *) For easily testing in shell scripts whether some command
8345 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8346 returns with exit code 0 iff no command of the given name is available.
8347 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8348 the output goes to stdout and nothing is printed to stderr.
8349 Additional arguments are always ignored.
8351 Since for each cipher there is a command of the same name,
8352 the 'no-cipher' compilation switches can be tested this way.
8354 ('openssl no-XXX' is not able to detect pseudo-commands such
8355 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8358 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8361 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8362 is set; it will be thrown away anyway because each handshake creates
8364 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8365 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8366 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8367 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8370 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8371 'Q' and 'R' lose their special meanings (quit/renegotiate).
8372 This is part of what -quiet does; unlike -quiet, -ign_eof
8373 does not suppress any output.
8376 *) Add compatibility options to the purpose and trust code. The
8377 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8378 accepts a certificate or CA, this was the previous behaviour,
8379 with all the associated security issues.
8381 X509_TRUST_COMPAT is the old trust behaviour: only and
8382 automatically trust self signed roots in certificate store. A
8383 new trust setting X509_TRUST_DEFAULT is used to specify that
8384 a purpose has no associated trust setting and it should instead
8385 use the value in the default purpose.
8388 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8389 and fix a memory leak.
8392 *) In util/mkerr.pl (which implements 'make errors'), preserve
8393 reason strings from the previous version of the .c file, as
8394 the default to have only downcase letters (and digits) in
8395 automatically generated reasons codes is not always appropriate.
8398 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8399 using strerror. Previously, ERR_reason_error_string() returned
8400 library names as reason strings for SYSerr; but SYSerr is a special
8401 case where small numbers are errno values, not library numbers.
8404 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8405 converts DSA parameters into DH parameters. (When creating parameters,
8406 DSA_generate_parameters is used.)
8409 *) Include 'length' (recommended exponent length) in C code generated
8410 by 'openssl dhparam -C'.
8413 *) The second argument to set_label in perlasm was already being used
8414 so couldn't be used as a "file scope" flag. Moved to third argument
8418 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8419 instead of RAND_bytes for encryption IVs and salts.
8422 *) Include RAND_status() into RAND_METHOD instead of implementing
8423 it only for md_rand.c Otherwise replacing the PRNG by calling
8424 RAND_set_rand_method would be impossible.
8427 *) Don't let DSA_generate_key() enter an infinite loop if the random
8428 number generation fails.
8431 *) New 'rand' application for creating pseudo-random output.
8434 *) Added configuration support for Linux/IA64
8435 [Rolf Haberrecker <rolf@suse.de>]
8437 *) Assembler module support for Mingw32.
8440 *) Shared library support for HPUX (in shlib/).
8441 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8443 *) Shared library support for Solaris gcc.
8444 [Lutz Behnke <behnke@trustcenter.de>]
8446 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8448 *) PKCS7_encrypt() was adding text MIME headers twice because they
8449 were added manually and by SMIME_crlf_copy().
8452 *) In bntest.c don't call BN_rand with zero bits argument.
8453 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8455 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8456 case was implemented. This caused BN_div_recp() to fail occasionally.
8459 *) Add an optional second argument to the set_label() in the perl
8460 assembly language builder. If this argument exists and is set
8461 to 1 it signals that the assembler should use a symbol whose
8462 scope is the entire file, not just the current function. This
8463 is needed with MASM which uses the format label:: for this scope.
8464 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8466 *) Change the ASN1 types so they are typedefs by default. Before
8467 almost all types were #define'd to ASN1_STRING which was causing
8468 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8472 *) Change names of new functions to the new get1/get0 naming
8473 convention: After 'get1', the caller owns a reference count
8474 and has to call ..._free; 'get0' returns a pointer to some
8475 data structure without incrementing reference counters.
8476 (Some of the existing 'get' functions increment a reference
8477 counter, some don't.)
8478 Similarly, 'set1' and 'add1' functions increase reference
8479 counters or duplicate objects.
8482 *) Allow for the possibility of temp RSA key generation failure:
8483 the code used to assume it always worked and crashed on failure.
8486 *) Fix potential buffer overrun problem in BIO_printf().
8487 [Ulf Möller, using public domain code by Patrick Powell; problem
8488 pointed out by David Sacerdote <das33@cornell.edu>]
8490 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8491 RAND_egd() and RAND_status(). In the command line application,
8492 the EGD socket can be specified like a seed file using RANDFILE
8496 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8497 Some CAs (e.g. Verisign) distribute certificates in this form.
8500 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8501 list to exclude them. This means that no special compilation option
8502 is needed to use anonymous DH: it just needs to be included in the
8506 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8507 EVP_MD_type. The old functionality is available in a new macro called
8508 EVP_MD_md(). Change code that uses it and update docs.
8511 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8512 where the 'void *' argument is replaced by a function pointer argument.
8513 Previously 'void *' was abused to point to functions, which works on
8514 many platforms, but is not correct. As these functions are usually
8515 called by macros defined in OpenSSL header files, most source code
8516 should work without changes.
8519 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8520 sections with information on -D... compiler switches used for
8521 compiling the library so that applications can see them. To enable
8522 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8523 must be defined. E.g.,
8524 #define OPENSSL_ALGORITHM_DEFINES
8525 #include <openssl/opensslconf.h>
8526 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8527 [Richard Levitte, Ulf and Bodo Möller]
8529 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8533 *) Change the 'other' type in certificate aux info to a STACK_OF
8534 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8535 the required ASN1 format: arbitrary types determined by an OID.
8538 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8539 argument to 'req'. This is not because the function is newer or
8540 better than others it just uses the work 'NEW' in the certificate
8541 request header lines. Some software needs this.
8544 *) Reorganise password command line arguments: now passwords can be
8545 obtained from various sources. Delete the PEM_cb function and make
8546 it the default behaviour: i.e. if the callback is NULL and the
8547 usrdata argument is not NULL interpret it as a null terminated pass
8548 phrase. If usrdata and the callback are NULL then the pass phrase
8549 is prompted for as usual.
8552 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8553 the support is automatically enabled. The resulting binaries will
8554 autodetect the card and use it if present.
8555 [Ben Laurie and Compaq Inc.]
8557 *) Work around for Netscape hang bug. This sends certificate request
8558 and server done in one record. Since this is perfectly legal in the
8559 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8560 the bugs/SSLv3 entry for more info.
8563 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8566 *) Add -rand argument to smime and pkcs12 applications and read/write
8570 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8573 *) Add command line password options to the remaining applications.
8576 *) Bug fix for BN_div_recp() for numerators with an even number of
8580 *) More tests in bntest.c, and changed test_bn output.
8583 *) ./config recognizes MacOS X now.
8586 *) Bug fix for BN_div() when the first words of num and divsor are
8587 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8590 *) Add support for various broken PKCS#8 formats, and command line
8591 options to produce them.
8594 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8595 get temporary BIGNUMs from a BN_CTX.
8598 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8602 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8603 include a #define from the old name to the new. The original intent
8604 was that statically linked binaries could for example just call
8605 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8606 link with digests. This never worked because SSLeay_add_all_digests()
8607 and SSLeay_add_all_ciphers() were in the same source file so calling
8608 one would link with the other. They are now in separate source files.
8611 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8614 *) Use a less unusual form of the Miller-Rabin primality test (it used
8615 a binary algorithm for exponentiation integrated into the Miller-Rabin
8616 loop, our standard modexp algorithms are faster).
8619 *) Support for the EBCDIC character set completed.
8620 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8622 *) Source code cleanups: use const where appropriate, eliminate casts,
8623 use void * instead of char * in lhash.
8626 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8627 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8628 this the server could overwrite ephemeral keys that the client
8632 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8633 using 50 iterations of the Rabin-Miller test.
8635 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8636 iterations of the Rabin-Miller test as required by the appendix
8637 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8638 As BN_is_prime_fasttest includes trial division, DSA parameter
8639 generation becomes much faster.
8641 This implies a change for the callback functions in DSA_is_prime
8642 and DSA_generate_parameters: The callback function is called once
8643 for each positive witness in the Rabin-Miller test, not just
8644 occasionally in the inner loop; and the parameters to the
8645 callback function now provide an iteration count for the outer
8646 loop rather than for the current invocation of the inner loop.
8647 DSA_generate_parameters additionally can call the callback
8648 function with an 'iteration count' of -1, meaning that a
8649 candidate has passed the trial division test (when q is generated
8650 from an application-provided seed, trial division is skipped).
8653 *) New function BN_is_prime_fasttest that optionally does trial
8654 division before starting the Rabin-Miller test and has
8655 an additional BN_CTX * argument (whereas BN_is_prime always
8656 has to allocate at least one BN_CTX).
8657 'callback(1, -1, cb_arg)' is called when a number has passed the
8658 trial division stage.
8661 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8665 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8668 *) New function BN_pseudo_rand().
8671 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8672 bignum version of BN_from_montgomery() with the working code from
8673 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8677 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8678 made it impossible to use the same SSL_SESSION data structure in
8679 SSL2 clients in multiple threads.
8682 *) The return value of RAND_load_file() no longer counts bytes obtained
8683 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8684 to seed the PRNG (previously an explicit byte count was required).
8685 [Ulf Möller, Bodo Möller]
8687 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8688 used (char *) instead of (void *) and had casts all over the place.
8691 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8694 *) Retain source code compatibility for BN_prime_checks macro:
8695 BN_is_prime(..., BN_prime_checks, ...) now uses
8696 BN_prime_checks_for_size to determine the appropriate number of
8697 Rabin-Miller iterations.
8700 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8701 DH_CHECK_P_NOT_SAFE_PRIME.
8702 (Check if this is true? OpenPGP calls them "strong".)
8705 *) Merge the functionality of "dh" and "gendh" programs into a new program
8706 "dhparam". The old programs are retained for now but will handle DH keys
8707 (instead of parameters) in future.
8710 *) Make the ciphers, s_server and s_client programs check the return values
8711 when a new cipher list is set.
8714 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8715 ciphers. Before when the 56bit ciphers were enabled the sorting was
8718 The syntax for the cipher sorting has been extended to support sorting by
8719 cipher-strength (using the strength_bits hard coded in the tables).
8720 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8722 Fix a bug in the cipher-command parser: when supplying a cipher command
8723 string with an "undefined" symbol (neither command nor alphanumeric
8724 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8725 an error is flagged.
8727 Due to the strength-sorting extension, the code of the
8728 ssl_create_cipher_list() function was completely rearranged. I hope that
8729 the readability was also increased :-)
8730 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8732 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8733 for the first serial number and places 2 in the serial number file. This
8734 avoids problems when the root CA is created with serial number zero and
8735 the first user certificate has the same issuer name and serial number
8739 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8740 the new code. Add documentation for this stuff.
8743 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8744 X509_*() to X509at_*() on the grounds that they don't handle X509
8745 structures and behave in an analogous way to the X509v3 functions:
8746 they shouldn't be called directly but wrapper functions should be used
8749 So we also now have some wrapper functions that call the X509at functions
8750 when passed certificate requests. (TO DO: similar things can be done with
8751 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8752 things. Some of these need some d2i or i2d and print functionality
8753 because they handle more complex structures.)
8756 *) Add missing #ifndefs that caused missing symbols when building libssl
8757 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8758 NO_RSA in ssl/s2*.c.
8759 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8761 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8762 has a return value which indicates the quality of the random data
8763 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8764 error queue. New function RAND_pseudo_bytes() generates output that is
8765 guaranteed to be unique but not unpredictable. RAND_add is like
8766 RAND_seed, but takes an extra argument for an entropy estimate
8767 (RAND_seed always assumes full entropy).
8770 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8771 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8772 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8773 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8774 false-positive rate of at most 2^-80 for random input.
8777 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8780 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8781 in the 0.9.5 release), this returns the chain
8782 from an X509_CTX structure with a dup of the stack and all
8783 the X509 reference counts upped: so the stack will exist
8784 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8787 Also make SSL_SESSION_print() print out the verify return
8791 *) Add manpage for the pkcs12 command. Also change the default
8792 behaviour so MAC iteration counts are used unless the new
8793 -nomaciter option is used. This improves file security and
8794 only older versions of MSIE (4.0 for example) need it.
8797 *) Honor the no-xxx Configure options when creating .DEF files.
8800 *) Add PKCS#10 attributes to field table: challengePassword,
8801 unstructuredName and unstructuredAddress. These are taken from
8802 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8803 international characters are used.
8805 More changes to X509_ATTRIBUTE code: allow the setting of types
8806 based on strings. Remove the 'loc' parameter when adding
8807 attributes because these will be a SET OF encoding which is sorted
8811 *) Initial changes to the 'req' utility to allow request generation
8812 automation. This will allow an application to just generate a template
8813 file containing all the field values and have req construct the
8816 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8817 used all over the place including certificate requests and PKCS#7
8818 structures. They are currently handled manually where necessary with
8819 some primitive wrappers for PKCS#7. The new functions behave in a
8820 manner analogous to the X509 extension functions: they allow
8821 attributes to be looked up by NID and added.
8823 Later something similar to the X509V3 code would be desirable to
8824 automatically handle the encoding, decoding and printing of the
8825 more complex types. The string types like challengePassword can
8826 be handled by the string table functions.
8828 Also modified the multi byte string table handling. Now there is
8829 a 'global mask' which masks out certain types. The table itself
8830 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8831 is useful when for example there is only one permissible type
8832 (as in countryName) and using the mask might result in no valid
8836 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8837 SSL_get_peer_finished to allow applications to obtain the latest
8838 Finished messages sent to the peer or expected from the peer,
8839 respectively. (SSL_get_peer_finished is usually the Finished message
8840 actually received from the peer, otherwise the protocol will be aborted.)
8842 As the Finished message are message digests of the complete handshake
8843 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8844 be used for external authentication procedures when the authentication
8845 provided by SSL/TLS is not desired or is not enough.
8848 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8849 the host supports BWX extension and if Compaq C is present on the
8850 $PATH. Just exploiting of the BWX extension results in 20-30%
8851 performance kick for some algorithms, e.g. DES and RC4 to mention
8852 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8856 *) Add support for MS "fast SGC". This is arguably a violation of the
8857 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8858 weak crypto and after checking the certificate is SGC a second one
8859 with strong crypto. MS SGC stops the first handshake after receiving
8860 the server certificate message and sends a second client hello. Since
8861 a server will typically do all the time consuming operations before
8862 expecting any further messages from the client (server key exchange
8863 is the most expensive) there is little difference between the two.
8865 To get OpenSSL to support MS SGC we have to permit a second client
8866 hello message after we have sent server done. In addition we have to
8867 reset the MAC if we do get this second client hello.
8870 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8871 if a DER encoded private key is RSA or DSA traditional format. Changed
8872 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8873 format DER encoded private key. Newer code should use PKCS#8 format which
8874 has the key type encoded in the ASN1 structure. Added DER private key
8875 support to pkcs8 application.
8878 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8879 ciphersuites has been selected (as required by the SSL 3/TLS 1
8880 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8881 is set, we interpret this as a request to violate the specification
8882 (the worst that can happen is a handshake failure, and 'correct'
8883 behaviour would result in a handshake failure anyway).
8886 *) In SSL_CTX_add_session, take into account that there might be multiple
8887 SSL_SESSION structures with the same session ID (e.g. when two threads
8888 concurrently obtain them from an external cache).
8889 The internal cache can handle only one SSL_SESSION with a given ID,
8890 so if there's a conflict, we now throw out the old one to achieve
8894 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8895 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8896 some routines that use cipher OIDs: some ciphers do not have OIDs
8897 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8901 *) Simplify the trust setting structure and code. Now we just have
8902 two sequences of OIDs for trusted and rejected settings. These will
8903 typically have values the same as the extended key usage extension
8904 and any application specific purposes.
8906 The trust checking code now has a default behaviour: it will just
8907 check for an object with the same NID as the passed id. Functions can
8908 be provided to override either the default behaviour or the behaviour
8909 for a given id. SSL client, server and email already have functions
8910 in place for compatibility: they check the NID and also return "trusted"
8911 if the certificate is self signed.
8914 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8915 traditional format into an EVP_PKEY structure.
8918 *) Add a password callback function PEM_cb() which either prompts for
8919 a password if usr_data is NULL or otherwise assumes it is a null
8920 terminated password. Allow passwords to be passed on command line
8921 environment or config files in a few more utilities.
8924 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8925 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8926 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8927 Update documentation.
8930 *) Support for ASN1 "NULL" type. This could be handled before by using
8931 ASN1_TYPE but there wasn't any function that would try to read a NULL
8932 and produce an error if it couldn't. For compatibility we also have
8933 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8934 don't allocate anything because they don't need to.
8937 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8939 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8941 *) Rebuild of the memory allocation routines used by OpenSSL code and
8942 possibly others as well. The purpose is to make an interface that
8943 provide hooks so anyone can build a separate set of allocation and
8944 deallocation routines to be used by OpenSSL, for example memory
8945 pool implementations, or something else, which was previously hard
8946 since Malloc(), Realloc() and Free() were defined as macros having
8947 the values malloc, realloc and free, respectively (except for Win32
8948 compilations). The same is provided for memory debugging code.
8949 OpenSSL already comes with functionality to find memory leaks, but
8950 this gives people a chance to debug other memory problems.
8952 With these changes, a new set of functions and macros have appeared:
8954 CRYPTO_set_mem_debug_functions() [F]
8955 CRYPTO_get_mem_debug_functions() [F]
8956 CRYPTO_dbg_set_options() [F]
8957 CRYPTO_dbg_get_options() [F]
8958 CRYPTO_malloc_debug_init() [M]
8960 The memory debug functions are NULL by default, unless the library
8961 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8962 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8963 gives the standard debugging functions that come with OpenSSL) or
8964 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8965 provided by the library user) must be used. When the standard
8966 debugging functions are used, CRYPTO_dbg_set_options can be used to
8967 request additional information:
8968 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8969 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8971 Also, things like CRYPTO_set_mem_functions will always give the
8972 expected result (the new set of functions is used for allocation
8973 and deallocation) at all times, regardless of platform and compiler
8976 To finish it up, some functions that were never use in any other
8977 way than through macros have a new API and new semantic:
8980 CRYPTO_dbg_realloc()
8983 All macros of value have retained their old syntax.
8984 [Richard Levitte and Bodo Moeller]
8986 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8987 ordering of SMIMECapabilities wasn't in "strength order" and there
8988 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8992 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8993 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8994 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8996 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8997 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8998 functionality to handle multipart/signed properly) and a utility
8999 called 'smime' to call all this stuff. This is based on code I
9000 originally wrote for Celo who have kindly allowed it to be
9001 included in OpenSSL.
9004 *) Add variants des_set_key_checked and des_set_key_unchecked of
9005 des_set_key (aka des_key_sched). Global variable des_check_key
9006 decides which of these is called by des_set_key; this way
9007 des_check_key behaves as it always did, but applications and
9008 the library itself, which was buggy for des_check_key == 1,
9009 have a cleaner way to pick the version they need.
9012 *) New function PKCS12_newpass() which changes the password of a
9016 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9017 dynamic mix. In both cases the ids can be used as an index into the
9018 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9019 functions so they accept a list of the field values and the
9020 application doesn't need to directly manipulate the X509_TRUST
9024 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9028 *) Modify the way the V3 extension code looks up extensions. This now
9029 works in a similar way to the object code: we have some "standard"
9030 extensions in a static table which is searched with OBJ_bsearch()
9031 and the application can add dynamic ones if needed. The file
9032 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9033 updated whenever a new extension is added to the core code and kept
9034 in ext_nid order. There is a simple program 'tabtest.c' which checks
9035 this. New extensions are not added too often so this file can readily
9036 be maintained manually.
9038 There are two big advantages in doing things this way. The extensions
9039 can be looked up immediately and no longer need to be "added" using
9040 X509V3_add_standard_extensions(): this function now does nothing.
9041 [Side note: I get *lots* of email saying the extension code doesn't
9042 work because people forget to call this function]
9043 Also no dynamic allocation is done unless new extensions are added:
9044 so if we don't add custom extensions there is no need to call
9045 X509V3_EXT_cleanup().
9048 *) Modify enc utility's salting as follows: make salting the default. Add a
9049 magic header, so unsalted files fail gracefully instead of just decrypting
9050 to garbage. This is because not salting is a big security hole, so people
9051 should be discouraged from doing it.
9054 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9055 digest to be passed on the command line but it only used this
9056 parameter when signing a certificate. Modified so all relevant
9057 operations are affected by the digest parameter including the
9058 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9059 DSA key was used because it didn't fix the digest.
9062 *) Initial certificate chain verify code. Currently tests the untrusted
9063 certificates for consistency with the verify purpose (which is set
9064 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9066 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9067 this is because it will reject chains with invalid extensions whereas
9068 every previous version of OpenSSL and SSLeay made no checks at all.
9070 Trust code: checks the root CA for the relevant trust settings. Trust
9071 settings have an initial value consistent with the verify purpose: e.g.
9072 if the verify purpose is for SSL client use it expects the CA to be
9073 trusted for SSL client use. However the default value can be changed to
9074 permit custom trust settings: one example of this would be to only trust
9075 certificates from a specific "secure" set of CAs.
9077 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9078 which should be used for version portability: especially since the
9079 verify structure is likely to change more often now.
9081 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9082 to set them. If not set then assume SSL clients will verify SSL servers
9085 Two new options to the verify program: -untrusted allows a set of
9086 untrusted certificates to be passed in and -purpose which sets the
9087 intended purpose of the certificate. If a purpose is set then the
9088 new chain verify code is used to check extension consistency.
9091 *) Support for the authority information access extension.
9094 *) Modify RSA and DSA PEM read routines to transparently handle
9095 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9096 public keys in a format compatible with certificate
9097 SubjectPublicKeyInfo structures. Unfortunately there were already
9098 functions called *_PublicKey_* which used various odd formats so
9099 these are retained for compatibility: however the DSA variants were
9100 never in a public release so they have been deleted. Changed dsa/rsa
9101 utilities to handle the new format: note no releases ever handled public
9102 keys so we should be OK.
9104 The primary motivation for this change is to avoid the same fiasco
9105 that dogs private keys: there are several incompatible private key
9106 formats some of which are standard and some OpenSSL specific and
9107 require various evil hacks to allow partial transparent handling and
9108 even then it doesn't work with DER formats. Given the option anything
9109 other than PKCS#8 should be dumped: but the other formats have to
9110 stay in the name of compatibility.
9112 With public keys and the benefit of hindsight one standard format
9113 is used which works with EVP_PKEY, RSA or DSA structures: though
9114 it clearly returns an error if you try to read the wrong kind of key.
9116 Added a -pubkey option to the 'x509' utility to output the public key.
9117 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9118 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9119 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9120 that do the same as the EVP_PKEY_assign_*() except they up the
9121 reference count of the added key (they don't "swallow" the
9125 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9126 CRLs would fail if the file contained no certificates or no CRLs:
9127 added a new function to read in both types and return the number
9128 read: this means that if none are read it will be an error. The
9129 DER versions of the certificate and CRL reader would always fail
9130 because it isn't possible to mix certificates and CRLs in DER format
9131 without choking one or the other routine. Changed this to just read
9132 a certificate: this is the best we can do. Also modified the code
9133 in apps/verify.c to take notice of return codes: it was previously
9134 attempting to read in certificates from NULL pointers and ignoring
9135 any errors: this is one reason why the cert and CRL reader seemed
9136 to work. It doesn't check return codes from the default certificate
9137 routines: these may well fail if the certificates aren't installed.
9140 *) Code to support otherName option in GeneralName.
9143 *) First update to verify code. Change the verify utility
9144 so it warns if it is passed a self signed certificate:
9145 for consistency with the normal behaviour. X509_verify
9146 has been modified to it will now verify a self signed
9147 certificate if *exactly* the same certificate appears
9148 in the store: it was previously impossible to trust a
9149 single self signed certificate. This means that:
9150 openssl verify ss.pem
9151 now gives a warning about a self signed certificate but
9152 openssl verify -CAfile ss.pem ss.pem
9156 *) For servers, store verify_result in SSL_SESSION data structure
9157 (and add it to external session representation).
9158 This is needed when client certificate verifications fails,
9159 but an application-provided verification callback (set by
9160 SSL_CTX_set_cert_verify_callback) allows accepting the session
9161 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9162 but returns 1): When the session is reused, we have to set
9163 ssl->verify_result to the appropriate error code to avoid
9165 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9167 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9168 case in PKCS7_dataInit() where the signed PKCS7 structure
9169 didn't contain any existing data because it was being created.
9170 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9172 *) Add a salt to the key derivation routines in enc.c. This
9173 forms the first 8 bytes of the encrypted file. Also add a
9174 -S option to allow a salt to be input on the command line.
9177 *) New function X509_cmp(). Oddly enough there wasn't a function
9178 to compare two certificates. We do this by working out the SHA1
9179 hash and comparing that. X509_cmp() will be needed by the trust
9183 *) SSL_get1_session() is like SSL_get_session(), but increments
9184 the reference count in the SSL_SESSION returned.
9185 [Geoff Thorpe <geoff@eu.c2.net>]
9187 *) Fix for 'req': it was adding a null to request attributes.
9188 Also change the X509_LOOKUP and X509_INFO code to handle
9189 certificate auxiliary information.
9192 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9196 *) Add the possibility to add extra information to the memory leak
9197 detecting output, to form tracebacks, showing from where each
9198 allocation was originated: CRYPTO_push_info("constant string") adds
9199 the string plus current file name and line number to a per-thread
9200 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9201 is like calling CYRPTO_pop_info() until the stack is empty.
9202 Also updated memory leak detection code to be multi-thread-safe.
9205 *) Add options -text and -noout to pkcs7 utility and delete the
9206 encryption options which never did anything. Update docs.
9209 *) Add options to some of the utilities to allow the pass phrase
9210 to be included on either the command line (not recommended on
9211 OSes like Unix) or read from the environment. Update the
9212 manpages and fix a few bugs.
9215 *) Add a few manpages for some of the openssl commands.
9218 *) Fix the -revoke option in ca. It was freeing up memory twice,
9219 leaking and not finding already revoked certificates.
9222 *) Extensive changes to support certificate auxiliary information.
9223 This involves the use of X509_CERT_AUX structure and X509_AUX
9224 functions. An X509_AUX function such as PEM_read_X509_AUX()
9225 can still read in a certificate file in the usual way but it
9226 will also read in any additional "auxiliary information". By
9227 doing things this way a fair degree of compatibility can be
9228 retained: existing certificates can have this information added
9229 using the new 'x509' options.
9231 Current auxiliary information includes an "alias" and some trust
9232 settings. The trust settings will ultimately be used in enhanced
9233 certificate chain verification routines: currently a certificate
9234 can only be trusted if it is self signed and then it is trusted
9238 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9239 The problem was that one of the replacement routines had not been working
9240 since SSLeay releases. For now the offending routine has been replaced
9241 with non-optimised assembler. Even so, this now gives around 95%
9242 performance improvement for 1024 bit RSA signs.
9245 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9246 handling. Most clients have the effective key size in bits equal to
9247 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9248 A few however don't do this and instead use the size of the decrypted key
9249 to determine the RC2 key length and the AlgorithmIdentifier to determine
9250 the effective key length. In this case the effective key length can still
9251 be 40 bits but the key length can be 168 bits for example. This is fixed
9252 by manually forcing an RC2 key into the EVP_PKEY structure because the
9253 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9254 the key length and effective key length are equal.
9257 *) Add a bunch of functions that should simplify the creation of
9258 X509_NAME structures. Now you should be able to do:
9259 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9260 and have it automatically work out the correct field type and fill in
9261 the structures. The more adventurous can try:
9262 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9263 and it will (hopefully) work out the correct multibyte encoding.
9266 *) Change the 'req' utility to use the new field handling and multibyte
9267 copy routines. Before the DN field creation was handled in an ad hoc
9268 way in req, ca, and x509 which was rather broken and didn't support
9269 BMPStrings or UTF8Strings. Since some software doesn't implement
9270 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9271 using the dirstring_type option. See the new comment in the default
9272 openssl.cnf for more info.
9275 *) Make crypto/rand/md_rand.c more robust:
9276 - Assure unique random numbers after fork().
9277 - Make sure that concurrent threads access the global counter and
9278 md serializably so that we never lose entropy in them
9279 or use exactly the same state in multiple threads.
9280 Access to the large state is not always serializable because
9281 the additional locking could be a performance killer, and
9282 md should be large enough anyway.
9285 *) New file apps/app_rand.c with commonly needed functionality
9286 for handling the random seed file.
9288 Use the random seed file in some applications that previously did not:
9290 dsaparam -genkey (which also ignored its '-rand' option),
9293 x509 (when signing).
9294 Except on systems with /dev/urandom, it is crucial to have a random
9295 seed file at least for key creation, DSA signing, and for DH exchanges;
9296 for RSA signatures we could do without one.
9298 gendh and gendsa (unlike genrsa) used to read only the first byte
9299 of each file listed in the '-rand' option. The function as previously
9300 found in genrsa is now in app_rand.c and is used by all programs
9301 that support '-rand'.
9304 *) In RAND_write_file, use mode 0600 for creating files;
9305 don't just chmod when it may be too late.
9308 *) Report an error from X509_STORE_load_locations
9309 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9312 *) New function ASN1_mbstring_copy() this copies a string in either
9313 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9314 into an ASN1_STRING type. A mask of permissible types is passed
9315 and it chooses the "minimal" type to use or an error if not type
9319 *) Add function equivalents to the various macros in asn1.h. The old
9320 macros are retained with an M_ prefix. Code inside the library can
9321 use the M_ macros. External code (including the openssl utility)
9322 should *NOT* in order to be "shared library friendly".
9325 *) Add various functions that can check a certificate's extensions
9326 to see if it usable for various purposes such as SSL client,
9327 server or S/MIME and CAs of these types. This is currently
9328 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9329 verification. Also added a -purpose flag to x509 utility to
9330 print out all the purposes.
9333 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9337 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9338 for, obtain and decode and extension and obtain its critical flag.
9339 This allows all the necessary extension code to be handled in a
9340 single function call.
9343 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9344 platforms. See crypto/rc4/rc4_enc.c for further details.
9347 *) New -noout option to asn1parse. This causes no output to be produced
9348 its main use is when combined with -strparse and -out to extract data
9349 from a file (which may not be in ASN.1 format).
9352 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9353 when producing the local key id.
9354 [Richard Levitte <levitte@stacken.kth.se>]
9356 *) New option -dhparam in s_server. This allows a DH parameter file to be
9357 stated explicitly. If it is not stated then it tries the first server
9358 certificate file. The previous behaviour hard coded the filename
9362 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9363 a public key to be input or output. For example:
9364 openssl rsa -in key.pem -pubout -out pubkey.pem
9365 Also added necessary DSA public key functions to handle this.
9368 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9369 in the message. This was handled by allowing
9370 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9371 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9373 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9374 to the end of the strings whereas this didn't. This would cause problems
9375 if strings read with d2i_ASN1_bytes() were later modified.
9376 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9378 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9379 data and it contains EOF it will end up returning an error. This is
9380 caused by input 46 bytes long. The cause is due to the way base64
9381 BIOs find the start of base64 encoded data. They do this by trying a
9382 trial decode on each line until they find one that works. When they
9383 do a flag is set and it starts again knowing it can pass all the
9384 data directly through the decoder. Unfortunately it doesn't reset
9385 the context it uses. This means that if EOF is reached an attempt
9386 is made to pass two EOFs through the context and this causes the
9387 resulting error. This can also cause other problems as well. As is
9388 usual with these problems it takes *ages* to find and the fix is
9389 trivial: move one line.
9390 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9392 *) Ugly workaround to get s_client and s_server working under Windows. The
9393 old code wouldn't work because it needed to select() on sockets and the
9394 tty (for keypresses and to see if data could be written). Win32 only
9395 supports select() on sockets so we select() with a 1s timeout on the
9396 sockets and then see if any characters are waiting to be read, if none
9397 are present then we retry, we also assume we can always write data to
9398 the tty. This isn't nice because the code then blocks until we've
9399 received a complete line of data and it is effectively polling the
9400 keyboard at 1s intervals: however it's quite a bit better than not
9401 working at all :-) A dedicated Windows application might handle this
9402 with an event loop for example.
9405 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9406 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9407 will be called when RSA_sign() and RSA_verify() are used. This is useful
9408 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9409 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9410 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9411 This necessitated the support of an extra signature type NID_md5_sha1
9412 for SSL signatures and modifications to the SSL library to use it instead
9413 of calling RSA_public_decrypt() and RSA_private_encrypt().
9416 *) Add new -verify -CAfile and -CApath options to the crl program, these
9417 will lookup a CRL issuers certificate and verify the signature in a
9418 similar way to the verify program. Tidy up the crl program so it
9419 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9420 less strict. It will now permit CRL extensions even if it is not
9421 a V2 CRL: this will allow it to tolerate some broken CRLs.
9424 *) Initialize all non-automatic variables each time one of the openssl
9425 sub-programs is started (this is necessary as they may be started
9426 multiple times from the "OpenSSL>" prompt).
9427 [Lennart Bang, Bodo Moeller]
9429 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9430 removing all other RSA functionality (this is what NO_RSA does). This
9431 is so (for example) those in the US can disable those operations covered
9432 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9436 *) Non-copying interface to BIO pairs.
9437 (still largely untested)
9440 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9441 ASCII string. This was handled independently in various places before.
9444 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9445 UTF8 strings a character at a time.
9448 *) Use client_version from client hello to select the protocol
9449 (s23_srvr.c) and for RSA client key exchange verification
9450 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9453 *) Add various utility functions to handle SPKACs, these were previously
9454 handled by poking round in the structure internals. Added new function
9455 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9456 print, verify and generate SPKACs. Based on an original idea from
9457 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9460 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9463 *) Allow the config file extension section to be overwritten on the
9464 command line. Based on an original idea from Massimiliano Pala
9465 <madwolf@comune.modena.it>. The new option is called -extensions
9466 and can be applied to ca, req and x509. Also -reqexts to override
9467 the request extensions in req and -crlexts to override the crl extensions
9471 *) Add new feature to the SPKAC handling in ca. Now you can include
9472 the same field multiple times by preceding it by "XXXX." for example:
9475 this is the same syntax as used in the req config file.
9478 *) Allow certificate extensions to be added to certificate requests. These
9479 are specified in a 'req_extensions' option of the req section of the
9480 config file. They can be printed out with the -text option to req but
9481 are otherwise ignored at present.
9484 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9485 data read consists of only the final block it would not decrypted because
9486 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9487 A misplaced 'break' also meant the decrypted final block might not be
9488 copied until the next read.
9491 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9492 a few extra parameters to the DH structure: these will be useful if
9493 for example we want the value of 'q' or implement X9.42 DH.
9496 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9497 provides hooks that allow the default DSA functions or functions on a
9498 "per key" basis to be replaced. This allows hardware acceleration and
9499 hardware key storage to be handled without major modification to the
9500 library. Also added low level modexp hooks and CRYPTO_EX structure and
9501 associated functions.
9504 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9505 as "read only": it can't be written to and the buffer it points to will
9506 not be freed. Reading from a read only BIO is much more efficient than
9507 a normal memory BIO. This was added because there are several times when
9508 an area of memory needs to be read from a BIO. The previous method was
9509 to create a memory BIO and write the data to it, this results in two
9510 copies of the data and an O(n^2) reading algorithm. There is a new
9511 function BIO_new_mem_buf() which creates a read only memory BIO from
9512 an area of memory. Also modified the PKCS#7 routines to use read only
9516 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9517 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9518 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9519 but a retry condition occured while trying to read the rest.
9522 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9523 NID_pkcs7_encrypted by default: this was wrong since this should almost
9524 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9525 the encrypted data type: this is a more sensible place to put it and it
9526 allows the PKCS#12 code to be tidied up that duplicated this
9530 *) Changed obj_dat.pl script so it takes its input and output files on
9531 the command line. This should avoid shell escape redirection problems
9535 *) Initial support for certificate extension requests, these are included
9536 in things like Xenroll certificate requests. Included functions to allow
9537 extensions to be obtained and added.
9540 *) -crlf option to s_client and s_server for sending newlines as
9541 CRLF (as required by many protocols).
9544 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9546 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9547 [Ralf S. Engelschall]
9549 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9550 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9552 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9556 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9557 DH parameters/keys (q is lost during that conversion, but the resulting
9558 DH parameters contain its length).
9560 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9561 much faster than DH_generate_parameters (which creates parameters
9562 where p = 2*q + 1), and also the smaller q makes DH computations
9563 much more efficient (160-bit exponentiation instead of 1024-bit
9564 exponentiation); so this provides a convenient way to support DHE
9565 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9566 utter importance to use
9567 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9569 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9570 when such DH parameters are used, because otherwise small subgroup
9571 attacks may become possible!
9574 *) Avoid memory leak in i2d_DHparams.
9577 *) Allow the -k option to be used more than once in the enc program:
9578 this allows the same encrypted message to be read by multiple recipients.
9581 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9582 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9583 it will always use the numerical form of the OID, even if it has a short
9587 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9588 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9589 otherwise bn_mod_exp was called. In the case of hardware keys for example
9590 no private key components need be present and it might store extra data
9591 in the RSA structure, which cannot be accessed from bn_mod_exp.
9592 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9593 private key operations.
9596 *) Added support for SPARC Linux.
9599 *) pem_password_cb function type incompatibly changed from
9600 typedef int pem_password_cb(char *buf, int size, int rwflag);
9602 ....(char *buf, int size, int rwflag, void *userdata);
9603 so that applications can pass data to their callbacks:
9604 The PEM[_ASN1]_{read,write}... functions and macros now take an
9605 additional void * argument, which is just handed through whenever
9606 the password callback is called.
9607 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9609 New function SSL_CTX_set_default_passwd_cb_userdata.
9611 Compatibility note: As many C implementations push function arguments
9612 onto the stack in reverse order, the new library version is likely to
9613 interoperate with programs that have been compiled with the old
9614 pem_password_cb definition (PEM_whatever takes some data that
9615 happens to be on the stack as its last argument, and the callback
9616 just ignores this garbage); but there is no guarantee whatsoever that
9619 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9620 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9621 problems not only on Windows, but also on some Unix platforms.
9622 To avoid problematic command lines, these definitions are now in an
9623 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9624 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9627 *) MIPS III/IV assembler module is reimplemented.
9630 *) More DES library cleanups: remove references to srand/rand and
9631 delete an unused file.
9634 *) Add support for the the free Netwide assembler (NASM) under Win32,
9635 since not many people have MASM (ml) and it can be hard to obtain.
9636 This is currently experimental but it seems to work OK and pass all
9637 the tests. Check out INSTALL.W32 for info.
9640 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9641 without temporary keys kept an extra copy of the server key,
9642 and connections with temporary keys did not free everything in case
9646 *) New function RSA_check_key and new openssl rsa option -check
9647 for verifying the consistency of RSA keys.
9648 [Ulf Moeller, Bodo Moeller]
9650 *) Various changes to make Win32 compile work:
9651 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9652 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9653 comparison" warnings.
9654 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9657 *) Add a debugging option to PKCS#5 v2 key generation function: when
9658 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9659 derived keys are printed to stderr.
9662 *) Copy the flags in ASN1_STRING_dup().
9663 [Roman E. Pavlov <pre@mo.msk.ru>]
9665 *) The x509 application mishandled signing requests containing DSA
9666 keys when the signing key was also DSA and the parameters didn't match.
9668 It was supposed to omit the parameters when they matched the signing key:
9669 the verifying software was then supposed to automatically use the CA's
9670 parameters if they were absent from the end user certificate.
9672 Omitting parameters is no longer recommended. The test was also
9673 the wrong way round! This was probably due to unusual behaviour in
9674 EVP_cmp_parameters() which returns 1 if the parameters match.
9675 This meant that parameters were omitted when they *didn't* match and
9676 the certificate was useless. Certificates signed with 'ca' didn't have
9678 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9680 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9681 The interface is as follows:
9682 Applications can use
9683 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9684 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9685 "off" is now the default.
9686 The library internally uses
9687 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9688 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9689 to disable memory-checking temporarily.
9691 Some inconsistent states that previously were possible (and were
9692 even the default) are now avoided.
9694 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9695 with each memory chunk allocated; this is occasionally more helpful
9696 than just having a counter.
9698 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9700 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9704 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9705 which largely parallels "options", but is for changing API behaviour,
9706 whereas "options" are about protocol behaviour.
9707 Initial "mode" flags are:
9709 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9710 a single record has been written.
9711 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9712 retries use the same buffer location.
9713 (But all of the contents must be
9717 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9720 *) Fix problems with no-hmac etc.
9721 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9723 *) New functions RSA_get_default_method(), RSA_set_method() and
9724 RSA_get_method(). These allows replacement of RSA_METHODs without having
9725 to mess around with the internals of an RSA structure.
9728 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9729 Also really enable memory leak checks in openssl.c and in some
9731 [Chad C. Mulligan, Bodo Moeller]
9733 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9734 up the length of negative integers. This has now been simplified to just
9735 store the length when it is first determined and use it later, rather
9736 than trying to keep track of where data is copied and updating it to
9738 [Steve Henson, reported by Brien Wheeler
9739 <bwheeler@authentica-security.com>]
9741 *) Add a new function PKCS7_signatureVerify. This allows the verification
9742 of a PKCS#7 signature but with the signing certificate passed to the
9743 function itself. This contrasts with PKCS7_dataVerify which assumes the
9744 certificate is present in the PKCS#7 structure. This isn't always the
9745 case: certificates can be omitted from a PKCS#7 structure and be
9746 distributed by "out of band" means (such as a certificate database).
9749 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9750 function prototypes in pem.h, also change util/mkdef.pl to add the
9751 necessary function names.
9754 *) mk1mf.pl (used by Windows builds) did not properly read the
9755 options set by Configure in the top level Makefile, and Configure
9756 was not even able to write more than one option correctly.
9757 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9760 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9761 file to be loaded from a BIO or FILE pointer. The BIO version will
9762 for example allow memory BIOs to contain config info.
9765 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9766 Whoever hopes to achieve shared-library compatibility across versions
9767 must use this, not the compile-time macro.
9768 (Exercise 0.9.4: Which is the minimum library version required by
9770 Note: All this applies only to multi-threaded programs, others don't
9774 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9775 through a BIO pair triggered the default case, i.e.
9776 SSLerr(...,SSL_R_UNKNOWN_STATE).
9779 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9780 can use the SSL library even if none of the specific BIOs is
9784 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9785 for the encoded length.
9786 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9788 *) Add initial documentation of the X509V3 functions.
9791 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9792 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9793 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9794 secure PKCS#8 private key format with a high iteration count.
9797 *) Fix determination of Perl interpreter: A perl or perl5
9798 _directory_ in $PATH was also accepted as the interpreter.
9799 [Ralf S. Engelschall]
9801 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9802 wrong with it but it was very old and did things like calling
9803 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9807 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9808 to use the new extension code.
9811 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9812 with macros. This should make it easier to change their form, add extra
9813 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9817 *) Add to configuration table a new entry that can specify an alternative
9818 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9819 according to Mark Crispin <MRC@Panda.COM>.
9823 *) DES CBC did not update the IV. Weird.
9826 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9827 Changing the behaviour of the former might break existing programs --
9828 where IV updating is needed, des_ncbc_encrypt can be used.
9831 *) When bntest is run from "make test" it drives bc to check its
9832 calculations, as well as internally checking them. If an internal check
9833 fails, it needs to cause bc to give a non-zero result or make test carries
9834 on without noticing the failure. Fixed.
9837 *) DES library cleanups.
9840 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9841 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9842 ciphers. NOTE: although the key derivation function has been verified
9843 against some published test vectors it has not been extensively tested
9844 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9848 *) Instead of "mkdir -p", which is not fully portable, use new
9849 Perl script "util/mkdir-p.pl".
9852 *) Rewrite the way password based encryption (PBE) is handled. It used to
9853 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9854 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9855 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9856 the 'parameter' field of the AlgorithmIdentifier is passed to the
9857 underlying key generation function so it must do its own ASN1 parsing.
9858 This has also changed the EVP_PBE_CipherInit() function which now has a
9859 'parameter' argument instead of literal salt and iteration count values
9860 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9863 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9864 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9865 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9866 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9867 value was just used as a "magic string" and not used directly its
9868 value doesn't matter.
9871 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9875 *) "linux-sparc64" configuration (ultrapenguin).
9876 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9877 "linux-sparc" configuration.
9878 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9880 *) config now generates no-xxx options for missing ciphers.
9883 *) Support the EBCDIC character set (work in progress).
9884 File ebcdic.c not yet included because it has a different license.
9885 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9887 *) Support BS2000/OSD-POSIX.
9888 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9890 *) Make callbacks for key generation use void * instead of char *.
9893 *) Make S/MIME samples compile (not yet tested).
9896 *) Additional typesafe stacks.
9899 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9903 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9905 *) New configuration variant "sco5-gcc".
9907 *) Updated some demos.
9908 [Sean O Riordain, Wade Scholine]
9910 *) Add missing BIO_free at exit of pkcs12 application.
9913 *) Fix memory leak in conf.c.
9916 *) Updates for Win32 to assembler version of MD5.
9919 *) Set #! path to perl in apps/der_chop to where we found it
9920 instead of using a fixed path.
9923 *) SHA library changes for irix64-mips4-cc.
9926 *) Improvements for VMS support.
9930 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9932 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9933 This also avoids the problems with SC4.2 and unpatched SC5.
9934 [Andy Polyakov <appro@fy.chalmers.se>]
9936 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9937 These are required because of the typesafe stack would otherwise break
9938 existing code. If old code used a structure member which used to be STACK
9939 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9940 sk_num or sk_value it would produce an error because the num, data members
9941 are not present in STACK_OF. Now it just produces a warning. sk_set
9942 replaces the old method of assigning a value to sk_value
9943 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9944 that does this will no longer work (and should use sk_set instead) but
9945 this could be regarded as a "questionable" behaviour anyway.
9948 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9949 correctly handle encrypted S/MIME data.
9952 *) Change type of various DES function arguments from des_cblock
9953 (which means, in function argument declarations, pointer to char)
9954 to des_cblock * (meaning pointer to array with 8 char elements),
9955 which allows the compiler to do more typechecking; it was like
9956 that back in SSLeay, but with lots of ugly casts.
9958 Introduce new type const_des_cblock.
9961 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9962 problems: find RecipientInfo structure that matches recipient certificate
9963 and initialise the ASN1 structures properly based on passed cipher.
9966 *) Belatedly make the BN tests actually check the results.
9969 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9970 to and from BNs: it was completely broken. New compilation option
9971 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9972 key elements as negative integers.
9975 *) Reorganize and speed up MD5.
9976 [Andy Polyakov <appro@fy.chalmers.se>]
9979 [Richard Levitte <richard@levitte.org>]
9981 *) New option -out to asn1parse to allow the parsed structure to be
9982 output to a file. This is most useful when combined with the -strparse
9983 option to examine the output of things like OCTET STRINGS.
9986 *) Make SSL library a little more fool-proof by not requiring any longer
9987 that SSL_set_{accept,connect}_state be called before
9988 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9989 in many applications because usually everything *appeared* to work as
9990 intended anyway -- now it really works as intended).
9993 *) Move openssl.cnf out of lib/.
9996 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9997 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9998 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9999 [Ralf S. Engelschall]
10001 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10002 handle PKCS#7 enveloped data properly.
10003 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10005 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10006 copying pointers. The cert_st handling is changed by this in
10007 various ways (and thus what used to be known as ctx->default_cert
10008 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10009 any longer when s->cert does not give us what we need).
10010 ssl_cert_instantiate becomes obsolete by this change.
10011 As soon as we've got the new code right (possibly it already is?),
10012 we have solved a couple of bugs of the earlier code where s->cert
10013 was used as if it could not have been shared with other SSL structures.
10015 Note that using the SSL API in certain dirty ways now will result
10016 in different behaviour than observed with earlier library versions:
10017 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10018 does not influence s as it used to.
10020 In order to clean up things more thoroughly, inside SSL_SESSION
10021 we don't use CERT any longer, but a new structure SESS_CERT
10022 that holds per-session data (if available); currently, this is
10023 the peer's certificate chain and, for clients, the server's certificate
10024 and temporary key. CERT holds only those values that can have
10025 meaningful defaults in an SSL_CTX.
10028 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10029 from the internal representation. Various PKCS#7 fixes: remove some
10030 evil casts and set the enc_dig_alg field properly based on the signing
10034 *) Allow PKCS#12 password to be set from the command line or the
10035 environment. Let 'ca' get its config file name from the environment
10036 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10040 *) Allow certificate policies extension to use an IA5STRING for the
10041 organization field. This is contrary to the PKIX definition but
10042 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10046 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10047 without disallowing inline assembler and the like for non-pedantic builds.
10050 *) Support Borland C++ builder.
10051 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10053 *) Support Mingw32.
10056 *) SHA-1 cleanups and performance enhancements.
10057 [Andy Polyakov <appro@fy.chalmers.se>]
10059 *) Sparc v8plus assembler for the bignum library.
10060 [Andy Polyakov <appro@fy.chalmers.se>]
10062 *) Accept any -xxx and +xxx compiler options in Configure.
10065 *) Update HPUX configuration.
10068 *) Add missing sk_<type>_unshift() function to safestack.h
10069 [Ralf S. Engelschall]
10071 *) New function SSL_CTX_use_certificate_chain_file that sets the
10072 "extra_cert"s in addition to the certificate. (This makes sense
10073 only for "PEM" format files, as chains as a whole are not
10077 *) Support verify_depth from the SSL API.
10078 x509_vfy.c had what can be considered an off-by-one-error:
10079 Its depth (which was not part of the external interface)
10080 was actually counting the number of certificates in a chain;
10081 now it really counts the depth.
10084 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10085 instead of X509err, which often resulted in confusing error
10086 messages since the error codes are not globally unique
10087 (e.g. an alleged error in ssl3_accept when a certificate
10088 didn't match the private key).
10090 *) New function SSL_CTX_set_session_id_context that allows to set a default
10091 value (so that you don't need SSL_set_session_id_context for each
10092 connection using the SSL_CTX).
10095 *) OAEP decoding bug fix.
10098 *) Support INSTALL_PREFIX for package builders, as proposed by
10102 *) New Configure options "threads" and "no-threads". For systems
10103 where the proper compiler options are known (currently Solaris
10104 and Linux), "threads" is the default.
10107 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10110 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10111 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10112 such as /usr/local/bin.
10115 *) "make linux-shared" to build shared libraries.
10116 [Niels Poppe <niels@netbox.org>]
10118 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10121 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10122 extension adding in x509 utility.
10125 *) Remove NOPROTO sections and error code comments.
10128 *) Partial rewrite of the DEF file generator to now parse the ANSI
10132 *) New Configure options --prefix=DIR and --openssldir=DIR.
10135 *) Complete rewrite of the error code script(s). It is all now handled
10136 by one script at the top level which handles error code gathering,
10137 header rewriting and C source file generation. It should be much better
10138 than the old method: it now uses a modified version of Ulf's parser to
10139 read the ANSI prototypes in all header files (thus the old K&R definitions
10140 aren't needed for error creation any more) and do a better job of
10141 translating function codes into names. The old 'ASN1 error code imbedded
10142 in a comment' is no longer necessary and it doesn't use .err files which
10143 have now been deleted. Also the error code call doesn't have to appear all
10144 on one line (which resulted in some large lines...).
10147 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10150 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10151 0 (which usually indicates a closed connection), but continue reading.
10154 *) Fix some race conditions.
10157 *) Add support for CRL distribution points extension. Add Certificate
10158 Policies and CRL distribution points documentation.
10161 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10164 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10165 8 of keying material. Merlin has also confirmed interop with this fix
10166 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10167 [Merlin Hughes <merlin@baltimore.ie>]
10169 *) Fix lots of warnings.
10170 [Richard Levitte <levitte@stacken.kth.se>]
10172 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10173 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10174 [Richard Levitte <levitte@stacken.kth.se>]
10176 *) Fix problems with sizeof(long) == 8.
10177 [Andy Polyakov <appro@fy.chalmers.se>]
10179 *) Change functions to ANSI C.
10182 *) Fix typos in error codes.
10183 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10185 *) Remove defunct assembler files from Configure.
10188 *) SPARC v8 assembler BIGNUM implementation.
10189 [Andy Polyakov <appro@fy.chalmers.se>]
10191 *) Support for Certificate Policies extension: both print and set.
10192 Various additions to support the r2i method this uses.
10195 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10196 return a const string when you are expecting an allocated buffer.
10199 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10200 types DirectoryString and DisplayText.
10203 *) Add code to allow r2i extensions to access the configuration database,
10204 add an LHASH database driver and add several ctx helper functions.
10207 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10208 fail when they extended the size of a BIGNUM.
10211 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10212 support typesafe stack.
10215 *) Fix typo in SSL_[gs]et_options().
10216 [Nils Frostberg <nils@medcom.se>]
10218 *) Delete various functions and files that belonged to the (now obsolete)
10219 old X509V3 handling code.
10222 *) New Configure option "rsaref".
10225 *) Don't auto-generate pem.h.
10228 *) Introduce type-safe ASN.1 SETs.
10231 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10232 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10234 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10235 that links with OpenSSL (well at least cause lots of warnings), but fear
10236 not: the conversion is trivial, and it eliminates loads of evil casts. A
10237 few STACKed things have been converted already. Feel free to convert more.
10238 In the fullness of time, I'll do away with the STACK type altogether.
10241 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10242 specified in <certfile> by updating the entry in the index.txt file.
10243 This way one no longer has to edit the index.txt file manually for
10244 revoking a certificate. The -revoke option does the gory details now.
10245 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10247 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10248 `-text' option at all and this way the `-noout -text' combination was
10249 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10250 [Ralf S. Engelschall]
10252 *) Make sure a corresponding plain text error message exists for the
10253 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10254 verify callback function determined that a certificate was revoked.
10255 [Ralf S. Engelschall]
10257 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10258 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10259 all available cipers including rc5, which was forgotten until now.
10260 In order to let the testing shell script know which algorithms
10261 are available, a new (up to now undocumented) command
10262 "openssl list-cipher-commands" is used.
10265 *) Bugfix: s_client occasionally would sleep in select() when
10266 it should have checked SSL_pending() first.
10269 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10270 the raw DSA values prior to ASN.1 encoding.
10273 *) Tweaks to Configure
10274 [Niels Poppe <niels@netbox.org>]
10276 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10280 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10283 *) New config option to avoid instructions that are illegal on the 80386.
10284 The default code is faster, but requires at least a 486.
10287 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10288 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10289 same as SSL2_VERSION anyway.
10292 *) New "-showcerts" option for s_client.
10295 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10296 application. Various cleanups and fixes.
10299 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10300 modify error routines to work internally. Add error codes and PBE init
10301 to library startup routines.
10304 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10305 packing functions to asn1 and evp. Changed function names and error
10306 codes along the way.
10309 *) PKCS12 integration: and so it begins... First of several patches to
10310 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10311 objects to objects.h
10314 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10315 and display support for Thawte strong extranet extension.
10318 *) Add LinuxPPC support.
10319 [Jeff Dubrule <igor@pobox.org>]
10321 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10322 bn_div_words in alpha.s.
10323 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10325 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10326 OAEP isn't supported when OpenSSL is built with RSAref.
10327 [Ulf Moeller <ulf@fitug.de>]
10329 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10330 so they no longer are missing under -DNOPROTO.
10331 [Soren S. Jorvang <soren@t.dk>]
10334 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10336 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10337 doesn't work when the session is reused. Coming soon!
10340 *) Fix a security hole, that allows sessions to be reused in the wrong
10341 context thus bypassing client cert protection! All software that uses
10342 client certs and session caches in multiple contexts NEEDS PATCHING to
10343 allow session reuse! A fuller solution is in the works.
10344 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10346 *) Some more source tree cleanups (removed obsolete files
10347 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10348 permission on "config" script to be executable) and a fix for the INSTALL
10350 [Ulf Moeller <ulf@fitug.de>]
10352 *) Remove some legacy and erroneous uses of malloc, free instead of
10354 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10356 *) Make rsa_oaep_test return non-zero on error.
10357 [Ulf Moeller <ulf@fitug.de>]
10359 *) Add support for native Solaris shared libraries. Configure
10360 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10361 if someone would make that last step automatic.
10362 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10364 *) ctx_size was not built with the right compiler during "make links". Fixed.
10367 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10368 except NULL ciphers". This means the default cipher list will no longer
10369 enable NULL ciphers. They need to be specifically enabled e.g. with
10370 the string "DEFAULT:eNULL".
10373 *) Fix to RSA private encryption routines: if p < q then it would
10374 occasionally produce an invalid result. This will only happen with
10375 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10378 *) Be less restrictive and allow also `perl util/perlpath.pl
10379 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10380 because this way one can also use an interpreter named `perl5' (which is
10381 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10382 installed as `perl').
10383 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10385 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10386 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10388 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10389 advapi32.lib to Win32 build and change the pem test comparison
10390 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10391 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10392 and crypto/des/ede_cbcm_enc.c.
10395 *) DES quad checksum was broken on big-endian architectures. Fixed.
10398 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10399 Win32 test batch file so it (might) work again. The Win32 test batch file
10400 is horrible: I feel ill....
10403 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10404 in e_os.h. Audit of header files to check ANSI and non ANSI
10405 sections: 10 functions were absent from non ANSI section and not exported
10406 from Windows DLLs. Fixed up libeay.num for new functions.
10409 *) Make `openssl version' output lines consistent.
10410 [Ralf S. Engelschall]
10412 *) Fix Win32 symbol export lists for BIO functions: Added
10413 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10414 to ms/libeay{16,32}.def.
10415 [Ralf S. Engelschall]
10417 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10418 fine under Unix and passes some trivial tests I've now added. But the
10419 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10420 added to make sure no one expects that this stuff really works in the
10421 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10422 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10424 [Ralf S. Engelschall]
10426 *) Fix the generation of two part addresses in perl.
10427 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10429 *) Add config entry for Linux on MIPS.
10430 [John Tobey <jtobey@channel1.com>]
10432 *) Make links whenever Configure is run, unless we are on Windoze.
10435 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10436 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10440 *) Add a useful kludge to allow package maintainers to specify compiler and
10441 other platforms details on the command line without having to patch the
10442 Configure script everytime: One now can use ``perl Configure
10443 <id>:<details>'', i.e. platform ids are allowed to have details appended
10444 to them (separated by colons). This is treated as there would be a static
10445 pre-configured entry in Configure's %table under key <id> with value
10446 <details> and ``perl Configure <id>'' is called. So, when you want to
10447 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10448 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10449 now, which overrides the FreeBSD-elf entry on-the-fly.
10450 [Ralf S. Engelschall]
10452 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10455 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10456 on the `perl Configure ...' command line. This way one can compile
10457 OpenSSL libraries with Position Independent Code (PIC) which is needed
10458 for linking it into DSOs.
10459 [Ralf S. Engelschall]
10461 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10465 *) Cleaned up the LICENSE document: The official contact for any license
10466 questions now is the OpenSSL core team under openssl-core@openssl.org.
10467 And add a paragraph about the dual-license situation to make sure people
10468 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10469 to the OpenSSL toolkit.
10470 [Ralf S. Engelschall]
10472 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10473 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10474 Additionally cleaned up the `make links' target: Remove unnecessary
10475 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10476 to speed processing and no longer clutter the display with confusing
10477 stuff. Instead only the actually done links are displayed.
10478 [Ralf S. Engelschall]
10480 *) Permit null encryption ciphersuites, used for authentication only. It used
10481 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10482 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10486 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10487 signed attributes when verifying signatures (this would break them),
10488 the detached data encoding was wrong and public keys obtained using
10489 X509_get_pubkey() weren't freed.
10492 *) Add text documentation for the BUFFER functions. Also added a work around
10493 to a Win95 console bug. This was triggered by the password read stuff: the
10494 last character typed gets carried over to the next fread(). If you were
10495 generating a new cert request using 'req' for example then the last
10496 character of the passphrase would be CR which would then enter the first
10500 *) Added the new `Includes OpenSSL Cryptography Software' button as
10501 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10502 button and can be used by applications based on OpenSSL to show the
10503 relationship to the OpenSSL project.
10504 [Ralf S. Engelschall]
10506 *) Remove confusing variables in function signatures in files
10507 ssl/ssl_lib.c and ssl/ssl.h.
10508 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10510 *) Don't install bss_file.c under PREFIX/include/
10511 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10513 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10514 functions that return function pointers and has support for NT specific
10515 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10516 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10517 unsigned to signed types: this was killing the Win32 compile.
10520 *) Add new certificate file to stack functions,
10521 SSL_add_dir_cert_subjects_to_stack() and
10522 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10523 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10524 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10525 This means that Apache-SSL and similar packages don't have to mess around
10526 to add as many CAs as they want to the preferred list.
10529 *) Experiment with doxygen documentation. Currently only partially applied to
10531 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10532 openssl.doxy as the configuration file.
10535 *) Get rid of remaining C++-style comments which strict C compilers hate.
10536 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10538 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10539 compiled in by default: it has problems with large keys.
10542 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10543 DH private keys and/or callback functions which directly correspond to
10544 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10545 is needed for applications which have to configure certificates on a
10546 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10548 For the RSA certificate situation is makes no difference, but
10549 for the DSA certificate situation this fixes the "no shared cipher"
10550 problem where the OpenSSL cipher selection procedure failed because the
10551 temporary keys were not overtaken from the context and the API provided
10552 no way to reconfigure them.
10553 The new functions now let applications reconfigure the stuff and they
10554 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10555 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10556 non-public-API function ssl_cert_instantiate() is used as a helper
10557 function and also to reduce code redundancy inside ssl_rsa.c.
10558 [Ralf S. Engelschall]
10560 *) Move s_server -dcert and -dkey options out of the undocumented feature
10561 area because they are useful for the DSA situation and should be
10562 recognized by the users.
10563 [Ralf S. Engelschall]
10565 *) Fix the cipher decision scheme for export ciphers: the export bits are
10566 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10567 SSL_EXP_MASK. So, the original variable has to be used instead of the
10568 already masked variable.
10569 [Richard Levitte <levitte@stacken.kth.se>]
10571 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10572 [Richard Levitte <levitte@stacken.kth.se>]
10574 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10575 from `int' to `unsigned int' because it's a length and initialized by
10576 EVP_DigestFinal() which expects an `unsigned int *'.
10577 [Richard Levitte <levitte@stacken.kth.se>]
10579 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10580 script. Instead use the usual Shell->Perl transition trick.
10581 [Ralf S. Engelschall]
10583 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10584 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10585 -noout -modulus' as it's already the case for `openssl rsa -noout
10586 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10587 currently the public key is printed (a decision which was already done by
10588 `openssl dsa -modulus' in the past) which serves a similar purpose.
10589 Additionally the NO_RSA no longer completely removes the whole -modulus
10590 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10592 [Ralf S. Engelschall]
10594 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10595 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10596 [Arne Ansper <arne@ats.cyber.ee>]
10598 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10599 to be added. Now both 'req' and 'ca' can use new objects defined in the
10603 *) Add cool BIO that does syslog (or event log on NT).
10604 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10606 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10607 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10608 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10609 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10612 *) Add preliminary config info for new extension code.
10615 *) Make RSA_NO_PADDING really use no padding.
10616 [Ulf Moeller <ulf@fitug.de>]
10618 *) Generate errors when private/public key check is done.
10621 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10622 for some CRL extensions and new objects added.
10625 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10626 key usage extension and fuller support for authority key id.
10629 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10630 padding method for RSA, which is recommended for new applications in PKCS
10631 #1 v2.0 (RFC 2437, October 1998).
10632 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10633 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10634 against Bleichbacher's attack on RSA.
10635 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10638 *) Updates to the new SSL compression code
10639 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10641 *) Fix so that the version number in the master secret, when passed
10642 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10643 (because the server will not accept higher), that the version number
10644 is 0x03,0x01, not 0x03,0x00
10645 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10647 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10648 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10649 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10652 *) Support for RAW extensions where an arbitrary extension can be
10653 created by including its DER encoding. See apps/openssl.cnf for
10657 *) Make sure latest Perl versions don't interpret some generated C array
10658 code as Perl array code in the crypto/err/err_genc.pl script.
10659 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10661 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10662 not many people have the assembler. Various Win32 compilation fixes and
10663 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10664 build instructions.
10667 *) Modify configure script 'Configure' to automatically create crypto/date.h
10668 file under Win32 and also build pem.h from pem.org. New script
10669 util/mkfiles.pl to create the MINFO file on environments that can't do a
10670 'make files': perl util/mkfiles.pl >MINFO should work.
10673 *) Major rework of DES function declarations, in the pursuit of correctness
10674 and purity. As a result, many evil casts evaporated, and some weirdness,
10675 too. You may find this causes warnings in your code. Zapping your evil
10676 casts will probably fix them. Mostly.
10679 *) Fix for a typo in asn1.h. Bug fix to object creation script
10680 obj_dat.pl. It considered a zero in an object definition to mean
10681 "end of object": none of the objects in objects.h have any zeros
10682 so it wasn't spotted.
10683 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10685 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10686 Masking (CBCM). In the absence of test vectors, the best I have been able
10687 to do is check that the decrypt undoes the encrypt, so far. Send me test
10688 vectors if you have them.
10691 *) Correct calculation of key length for export ciphers (too much space was
10692 allocated for null ciphers). This has not been tested!
10695 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10696 message is now correct (it understands "crypto" and "ssl" on its
10697 command line). There is also now an "update" option. This will update
10698 the util/ssleay.num and util/libeay.num files with any new functions.
10700 perl util/mkdef.pl crypto ssl update
10701 it will update them.
10704 *) Overhauled the Perl interface (perl/*):
10705 - ported BN stuff to OpenSSL's different BN library
10706 - made the perl/ source tree CVS-aware
10707 - renamed the package from SSLeay to OpenSSL (the files still contain
10708 their history because I've copied them in the repository)
10709 - removed obsolete files (the test scripts will be replaced
10710 by better Test::Harness variants in the future)
10711 [Ralf S. Engelschall]
10713 *) First cut for a very conservative source tree cleanup:
10714 1. merge various obsolete readme texts into doc/ssleay.txt
10715 where we collect the old documents and readme texts.
10716 2. remove the first part of files where I'm already sure that we no
10717 longer need them because of three reasons: either they are just temporary
10718 files which were left by Eric or they are preserved original files where
10719 I've verified that the diff is also available in the CVS via "cvs diff
10720 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10721 the crypto/md/ stuff).
10722 [Ralf S. Engelschall]
10724 *) More extension code. Incomplete support for subject and issuer alt
10725 name, issuer and authority key id. Change the i2v function parameters
10726 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10727 what that's for :-) Fix to ASN1 macro which messed up
10728 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10731 *) Preliminary support for ENUMERATED type. This is largely copied from the
10735 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10736 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10738 *) Make sure `make rehash' target really finds the `openssl' program.
10739 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10741 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10742 like to hear about it if this slows down other processors.
10745 *) Add CygWin32 platform information to Configure script.
10746 [Alan Batie <batie@aahz.jf.intel.com>]
10748 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10749 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10751 *) New program nseq to manipulate netscape certificate sequences
10754 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10758 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10759 but the BN code had some problems that would cause failures when
10760 doing certificate verification and some other functions.
10761 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10763 *) Add ASN1 and PEM code to support netscape certificate sequences.
10766 *) Add ASN1 and PEM code to support netscape certificate sequences.
10769 *) Add several PKIX and private extended key usage OIDs.
10772 *) Modify the 'ca' program to handle the new extension code. Modify
10773 openssl.cnf for new extension format, add comments.
10776 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10777 and add a sample to openssl.cnf so req -x509 now adds appropriate
10781 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10782 error code, add initial support to X509_print() and x509 application.
10785 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10786 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10787 stuff is currently isolated and isn't even compiled yet.
10790 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10791 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10792 Removed the versions check from X509 routines when loading extensions:
10793 this allows certain broken certificates that don't set the version
10794 properly to be processed.
10797 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10798 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10799 can still be regenerated with "make depend".
10802 *) Spelling mistake in C version of CAST-128.
10803 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10805 *) Changes to the error generation code. The perl script err-code.pl
10806 now reads in the old error codes and retains the old numbers, only
10807 adding new ones if necessary. It also only changes the .err files if new
10808 codes are added. The makefiles have been modified to only insert errors
10809 when needed (to avoid needlessly modifying header files). This is done
10810 by only inserting errors if the .err file is newer than the auto generated
10811 C file. To rebuild all the error codes from scratch (the old behaviour)
10812 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10813 or delete all the .err files.
10816 *) CAST-128 was incorrectly implemented for short keys. The C version has
10817 been fixed, but is untested. The assembler versions are also fixed, but
10818 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10819 to regenerate it if needed.
10820 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10821 Hagino <itojun@kame.net>]
10823 *) File was opened incorrectly in randfile.c.
10824 [Ulf Möller <ulf@fitug.de>]
10826 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10827 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10828 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10829 al: it's just almost always a UTCTime. Note this patch adds new error
10830 codes so do a "make errors" if there are problems.
10833 *) Correct Linux 1 recognition in config.
10834 [Ulf Möller <ulf@fitug.de>]
10836 *) Remove pointless MD5 hash when using DSA keys in ca.
10837 [Anonymous <nobody@replay.com>]
10839 *) Generate an error if given an empty string as a cert directory. Also
10840 generate an error if handed NULL (previously returned 0 to indicate an
10841 error, but didn't set one).
10842 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10844 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10847 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10848 parameters. This was causing a warning which killed off the Win32 compile.
10851 *) Remove C++ style comments from crypto/bn/bn_local.h.
10852 [Neil Costigan <neil.costigan@celocom.com>]
10854 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10855 based on a text string, looking up short and long names and finally
10856 "dot" format. The "dot" format stuff didn't work. Added new function
10857 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10858 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10859 OID is not part of the table.
10862 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10863 X509_LOOKUP_by_alias().
10866 *) Sort openssl functions by name.
10869 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10870 encryption from sample DSA keys (in case anyone is interested the password
10874 *) Make _all_ *_free functions accept a NULL pointer.
10875 [Frans Heymans <fheymans@isaserver.be>]
10877 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10879 [Anonymous <nobody@replay.com>]
10881 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10882 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10884 *) Don't blow it for numeric -newkey arguments to apps/req.
10885 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10887 *) Temp key "for export" tests were wrong in s3_srvr.c.
10888 [Anonymous <nobody@replay.com>]
10890 *) Add prototype for temp key callback functions
10891 SSL_CTX_set_tmp_{rsa,dh}_callback().
10894 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10895 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10898 *) X509_name_add_entry() freed the wrong thing after an error.
10899 [Arne Ansper <arne@ats.cyber.ee>]
10901 *) rsa_eay.c would attempt to free a NULL context.
10902 [Arne Ansper <arne@ats.cyber.ee>]
10904 *) BIO_s_socket() had a broken should_retry() on Windoze.
10905 [Arne Ansper <arne@ats.cyber.ee>]
10907 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10908 [Arne Ansper <arne@ats.cyber.ee>]
10910 *) Make sure the already existing X509_STORE->depth variable is initialized
10911 in X509_STORE_new(), but document the fact that this variable is still
10912 unused in the certificate verification process.
10913 [Ralf S. Engelschall]
10915 *) Fix the various library and apps files to free up pkeys obtained from
10916 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10919 *) Fix reference counting in X509_PUBKEY_get(). This makes
10920 demos/maurice/example2.c work, amongst others, probably.
10921 [Steve Henson and Ben Laurie]
10923 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10924 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10925 are no longer created. This way we have a single and consistent command
10926 line interface `openssl <command>', similar to `cvs <command>'.
10927 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10929 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10930 BIT STRING wrapper always have zero unused bits.
10933 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10936 *) Make the top-level INSTALL documentation easier to understand.
10939 *) Makefiles updated to exit if an error occurs in a sub-directory
10940 make (including if user presses ^C) [Paul Sutton]
10942 *) Make Montgomery context stuff explicit in RSA data structure.
10945 *) Fix build order of pem and err to allow for generated pem.h.
10948 *) Fix renumbering bug in X509_NAME_delete_entry().
10951 *) Enhanced the err-ins.pl script so it makes the error library number
10952 global and can add a library name. This is needed for external ASN1 and
10953 other error libraries.
10956 *) Fixed sk_insert which never worked properly.
10959 *) Fix ASN1 macros so they can handle indefinite length constructed
10960 EXPLICIT tags. Some non standard certificates use these: they can now
10964 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10965 into a single doc/ssleay.txt bundle. This way the information is still
10966 preserved but no longer messes up this directory. Now it's new room for
10967 the new set of documentation files.
10968 [Ralf S. Engelschall]
10970 *) SETs were incorrectly DER encoded. This was a major pain, because they
10971 shared code with SEQUENCEs, which aren't coded the same. This means that
10972 almost everything to do with SETs or SEQUENCEs has either changed name or
10973 number of arguments.
10974 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10976 *) Fix test data to work with the above.
10979 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10980 was already fixed by Eric for 0.9.1 it seems.
10981 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10983 *) Autodetect FreeBSD3.
10986 *) Fix various bugs in Configure. This affects the following platforms:
10990 unixware-2.0-pentium
10994 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10995 before they are needed.
10998 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11002 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11004 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11005 changed SSLeay to OpenSSL in version strings.
11006 [Ralf S. Engelschall]
11008 *) Some fixups to the top-level documents.
11011 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11012 because the symlink to include/ was missing.
11013 [Ralf S. Engelschall]
11015 *) Incorporated the popular no-RSA/DSA-only patches
11016 which allow to compile a RSA-free SSLeay.
11017 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11019 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11020 when "ssleay" is still not found.
11021 [Ralf S. Engelschall]
11023 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11024 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11026 *) Updated the README file.
11027 [Ralf S. Engelschall]
11029 *) Added various .cvsignore files in the CVS repository subdirs
11030 to make a "cvs update" really silent.
11031 [Ralf S. Engelschall]
11033 *) Recompiled the error-definition header files and added
11034 missing symbols to the Win32 linker tables.
11035 [Ralf S. Engelschall]
11037 *) Cleaned up the top-level documents;
11038 o new files: CHANGES and LICENSE
11039 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11040 o merged COPYRIGHT into LICENSE
11041 o removed obsolete TODO file
11042 o renamed MICROSOFT to INSTALL.W32
11043 [Ralf S. Engelschall]
11045 *) Removed dummy files from the 0.9.1b source tree:
11046 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11047 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11048 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11049 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11050 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11051 [Ralf S. Engelschall]
11053 *) Added various platform portability fixes.
11056 *) The Genesis of the OpenSSL rpject:
11057 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11058 Young and Tim J. Hudson created while they were working for C2Net until
11060 [The OpenSSL Project]
11063 Changes between 0.9.0b and 0.9.1b [not released]
11065 *) Updated a few CA certificates under certs/
11068 *) Changed some BIGNUM api stuff.
11071 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11072 DGUX x86, Linux Alpha, etc.
11075 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11076 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11080 *) Add -strparse option to asn1pars program which parses nested
11082 [Dr Stephen Henson <shenson@bigfoot.com>]
11084 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11087 *) DSA fix for "ca" program.
11090 *) Added "-genkey" option to "dsaparam" program.
11093 *) Added RIPE MD160 (rmd160) message digest.
11096 *) Added -a (all) option to "ssleay version" command.
11099 *) Added PLATFORM define which is the id given to Configure.
11102 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11105 *) Extended the ASN.1 parser routines.
11108 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11111 *) Added a BN_CTX to the BN library.
11114 *) Fixed the weak key values in DES library
11117 *) Changed API in EVP library for cipher aliases.
11120 *) Added support for RC2/64bit cipher.
11123 *) Converted the lhash library to the crypto/mem.c functions.
11126 *) Added more recognized ASN.1 object ids.
11129 *) Added more RSA padding checks for SSL/TLS.
11132 *) Added BIO proxy/filter functionality.
11135 *) Added extra_certs to SSL_CTX which can be used
11136 send extra CA certificates to the client in the CA cert chain sending
11137 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11140 *) Now Fortezza is denied in the authentication phase because
11141 this is key exchange mechanism is not supported by SSLeay at all.
11144 *) Additional PKCS1 checks.
11147 *) Support the string "TLSv1" for all TLS v1 ciphers.
11150 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11151 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11154 *) Fixed a few memory leaks.
11157 *) Fixed various code and comment typos.
11160 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11161 bytes sent in the client random.
11162 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob