5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Added support for TLS extended master secret from
8 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
9 initial patch which was a great help during development.
12 *) All libssl internal structures have been removed from the public header
13 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
14 now redundant). Users should not attempt to access internal structures
15 directly. Instead they should use the provided API functions.
18 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
19 Access to deprecated functions can be re-enabled by running config with
20 "enable-deprecated". In addition applications wishing to use deprecated
21 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
22 will, by default, disable some transitive includes that previously existed
23 in the header files (e.g. ec.h will no longer, by default, include bn.h)
26 *) Added support for OCB mode. OpenSSL has been granted a patent license
27 compatible with the OpenSSL license for use of OCB. Details are available
28 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
29 for OCB can be removed by calling config with no-ocb.
32 *) SSLv2 support has been removed. It still supports receiving a SSLv2
33 compatible client hello.
36 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
37 done while fixing the error code for the key-too-small case.
38 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
40 *) Removed old DES API.
43 *) Remove various unsupported platforms:
49 Sinix/ReliantUNIX RM400
54 16-bit platforms such as WIN16
57 *) Clean up OPENSSL_NO_xxx #define's
58 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
59 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
60 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
61 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
62 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
63 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
64 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
65 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
66 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
67 Remove MS_STATIC; it's a relic from platforms <32 bits.
70 *) Cleaned up dead code
71 Remove all but one '#ifdef undef' which is to be looked at.
74 *) Clean up calling of xxx_free routines.
75 Just like free(), fix most of the xxx_free routines to accept
76 NULL. Remove the non-null checks from callers. Save much code.
79 *) Experimental support for a new, fast, unbiased prime candidate generator,
80 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
81 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
83 *) New output format NSS in the sess_id command line tool. This allows
84 exporting the session id and the master key in NSS keylog format.
85 [Martin Kaiser <martin@kaiser.cx>]
87 *) Harmonize version and its documentation. -f flag is used to display
89 [mancha <mancha1@zoho.com>]
91 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
93 [mancha <mancha1@zoho.com>]
95 *) Fix some double frees. These are not thought to be exploitable.
96 [mancha <mancha1@zoho.com>]
98 *) A missing bounds check in the handling of the TLS heartbeat extension
99 can be used to reveal up to 64k of memory to a connected client or
102 Thanks for Neel Mehta of Google Security for discovering this bug and to
103 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
104 preparing the fix (CVE-2014-0160)
105 [Adam Langley, Bodo Moeller]
107 *) Fix for the attack described in the paper "Recovering OpenSSL
108 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
109 by Yuval Yarom and Naomi Benger. Details can be obtained from:
110 http://eprint.iacr.org/2014/140
112 Thanks to Yuval Yarom and Naomi Benger for discovering this
113 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
114 [Yuval Yarom and Naomi Benger]
116 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
117 this fixes a limitation in previous versions of OpenSSL.
120 *) Experimental encrypt-then-mac support.
122 Experimental support for encrypt then mac from
123 draft-gutmann-tls-encrypt-then-mac-02.txt
125 To enable it set the appropriate extension number (0x42 for the test
126 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
128 For non-compliant peers (i.e. just about everything) this should have no
131 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
135 *) Add EVP support for key wrapping algorithms, to avoid problems with
136 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
137 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
138 algorithms and include tests cases.
141 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
145 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
146 MGF1 digest and OAEP label.
149 *) Make openssl verify return errors.
150 [Chris Palmer <palmer@google.com> and Ben Laurie]
152 *) New function ASN1_TIME_diff to calculate the difference between two
153 ASN1_TIME structures or one structure and the current time.
156 *) Update fips_test_suite to support multiple command line options. New
157 test to induce all self test errors in sequence and check expected
161 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
162 sign or verify all in one operation.
165 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
166 test programs and fips_test_suite. Includes functionality to parse
167 the minimal script output of fipsalgest.pl directly.
170 *) Add authorisation parameter to FIPS_module_mode_set().
173 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
176 *) Use separate DRBG fields for internal and external flags. New function
177 FIPS_drbg_health_check() to perform on demand health checking. Add
178 generation tests to fips_test_suite with reduced health check interval to
179 demonstrate periodic health checking. Add "nodh" option to
180 fips_test_suite to skip very slow DH test.
183 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
187 *) More extensive health check for DRBG checking many more failure modes.
188 New function FIPS_selftest_drbg_all() to handle every possible DRBG
189 combination: call this in fips_test_suite.
192 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
193 and POST to handle Dual EC cases.
196 *) Add support for canonical generation of DSA parameter 'g'. See
199 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
200 POST to handle HMAC cases.
203 *) Add functions FIPS_module_version() and FIPS_module_version_text()
204 to return numerical and string versions of the FIPS module number.
207 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
208 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
209 outside the validated module in the FIPS capable OpenSSL.
212 *) Minor change to DRBG entropy callback semantics. In some cases
213 there is no multiple of the block length between min_len and
214 max_len. Allow the callback to return more than max_len bytes
215 of entropy but discard any extra: it is the callback's responsibility
216 to ensure that the extra data discarded does not impact the
217 requested amount of entropy.
220 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
221 information in FIPS186-3, SP800-57 and SP800-131A.
224 *) CCM support via EVP. Interface is very similar to GCM case except we
225 must supply all data in one chunk (i.e. no update, final) and the
226 message length must be supplied if AAD is used. Add algorithm test
230 *) Initial version of POST overhaul. Add POST callback to allow the status
231 of POST to be monitored and/or failures induced. Modify fips_test_suite
232 to use callback. Always run all selftests even if one fails.
235 *) XTS support including algorithm test driver in the fips_gcmtest program.
236 Note: this does increase the maximum key length from 32 to 64 bytes but
237 there should be no binary compatibility issues as existing applications
238 will never use XTS mode.
241 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
242 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
243 performs algorithm blocking for unapproved PRNG types. Also do not
244 set PRNG type in FIPS_mode_set(): leave this to the application.
245 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
246 the standard OpenSSL PRNG: set additional data to a date time vector.
249 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
250 This shouldn't present any incompatibility problems because applications
251 shouldn't be using these directly and any that are will need to rethink
252 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
255 *) Extensive self tests and health checking required by SP800-90 DRBG.
256 Remove strength parameter from FIPS_drbg_instantiate and always
257 instantiate at maximum supported strength.
260 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
263 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
266 *) New function DH_compute_key_padded() to compute a DH key and pad with
267 leading zeroes if needed: this complies with SP800-56A et al.
270 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
271 anything, incomplete, subject to change and largely untested at present.
274 *) Modify fipscanisteronly build option to only build the necessary object
275 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
278 *) Add experimental option FIPSSYMS to give all symbols in
279 fipscanister.o and FIPS or fips prefix. This will avoid
280 conflicts with future versions of OpenSSL. Add perl script
281 util/fipsas.pl to preprocess assembly language source files
282 and rename any affected symbols.
285 *) Add selftest checks and algorithm block of non-fips algorithms in
286 FIPS mode. Remove DES2 from selftests.
289 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
290 return internal method without any ENGINE dependencies. Add new
291 tiny fips sign and verify functions.
294 *) New build option no-ec2m to disable characteristic 2 code.
297 *) New build option "fipscanisteronly". This only builds fipscanister.o
298 and (currently) associated fips utilities. Uses the file Makefile.fips
299 instead of Makefile.org as the prototype.
302 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
303 Update fips_gcmtest to use IV generator.
306 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
307 setting output buffer to NULL. The *Final function must be
308 called although it will not retrieve any additional data. The tag
309 can be set or retrieved with a ctrl. The IV length is by default 12
310 bytes (96 bits) but can be set to an alternative value. If the IV
311 length exceeds the maximum IV length (currently 16 bytes) it cannot be
315 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
316 underlying do_cipher function handles all cipher semantics itself
317 including padding and finalisation. This is useful if (for example)
318 an ENGINE cipher handles block padding itself. The behaviour of
319 do_cipher is subtly changed if this flag is set: the return value
320 is the number of characters written to the output buffer (zero is
321 no longer an error code) or a negative error code. Also if the
322 input buffer is NULL and length 0 finalisation should be performed.
325 *) If a candidate issuer certificate is already part of the constructed
326 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
329 *) Improve forward-security support: add functions
331 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
332 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
334 for use by SSL/TLS servers; the callback function will be called whenever a
335 new session is created, and gets to decide whether the session may be
336 cached to make it resumable (return 0) or not (return 1). (As by the
337 SSL/TLS protocol specifications, the session_id sent by the server will be
338 empty to indicate that the session is not resumable; also, the server will
339 not generate RFC 4507 (RFC 5077) session tickets.)
341 A simple reasonable callback implementation is to return is_forward_secure.
342 This parameter will be set to 1 or 0 depending on the ciphersuite selected
343 by the SSL/TLS server library, indicating whether it can provide forward
345 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
347 *) New -verify_name option in command line utilities to set verification
351 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
352 Add CMAC pkey methods.
355 *) Experimental renegotiation in s_server -www mode. If the client
356 browses /reneg connection is renegotiated. If /renegcert it is
357 renegotiated requesting a certificate.
360 *) Add an "external" session cache for debugging purposes to s_server. This
361 should help trace issues which normally are only apparent in deployed
362 multi-process servers.
365 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
366 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
367 BIO_set_cipher() and some obscure PEM functions were changed so they
368 can now return an error. The RAND changes required a change to the
369 RAND_METHOD structure.
372 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
373 a gcc attribute to warn if the result of a function is ignored. This
374 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
375 whose return value is often ignored.
378 Changes between 1.0.2 and 1.0.2a [xx XXX xxxx]
380 *) ClientHello sigalgs DoS fix
382 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
383 invalid signature algorithms extension a NULL pointer dereference will
384 occur. This can be exploited in a DoS attack against the server.
386 This issue was was reported to OpenSSL by David Ramos of Stanford
389 [Stephen Henson and Matt Caswell]
391 *) Multiblock corrupted pointer fix
393 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
394 feature only applies on 64 bit x86 architecture platforms that support AES
395 NI instructions. A defect in the implementation of "multiblock" can cause
396 OpenSSL's internal write buffer to become incorrectly set to NULL when
397 using non-blocking IO. Typically, when the user application is using a
398 socket BIO for writing, this will only result in a failed connection.
399 However if some other BIO is used then it is likely that a segmentation
400 fault will be triggered, thus enabling a potential DoS attack.
402 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
406 *) Segmentation fault in DTLSv1_listen fix
408 The DTLSv1_listen function is intended to be stateless and processes the
409 initial ClientHello from many peers. It is common for user code to loop
410 over the call to DTLSv1_listen until a valid ClientHello is received with
411 an associated cookie. A defect in the implementation of DTLSv1_listen means
412 that state is preserved in the SSL object from one invocation to the next
413 that can lead to a segmentation fault. Errors processing the initial
414 ClientHello can trigger this scenario. An example of such an error could be
415 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
418 This issue was reported to OpenSSL by Per Allansson.
422 *) Segmentation fault in ASN1_TYPE_cmp fix
424 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
425 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
426 certificate signature algorithm consistency this can be used to crash any
427 certificate verification operation and exploited in a DoS attack. Any
428 application which performs certificate verification is vulnerable including
429 OpenSSL clients and servers which enable client authentication.
433 *) Segmentation fault for invalid PSS parameters fix
435 The signature verification routines will crash with a NULL pointer
436 dereference if presented with an ASN.1 signature using the RSA PSS
437 algorithm and invalid parameters. Since these routines are used to verify
438 certificate signature algorithms this can be used to crash any
439 certificate verification operation and exploited in a DoS attack. Any
440 application which performs certificate verification is vulnerable including
441 OpenSSL clients and servers which enable client authentication.
443 This issue was was reported to OpenSSL by Brian Carpenter.
447 *) ASN.1 structure reuse memory corruption fix
449 Reusing a structure in ASN.1 parsing may allow an attacker to cause
450 memory corruption via an invalid write. Such reuse is and has been
451 strongly discouraged and is believed to be rare.
453 Applications that parse structures containing CHOICE or ANY DEFINED BY
454 components may be affected. Certificate parsing (d2i_X509 and related
455 functions) are however not affected. OpenSSL clients and servers are
460 *) PKCS7 NULL pointer dereferences fix
462 The PKCS#7 parsing code does not handle missing outer ContentInfo
463 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
464 missing content and trigger a NULL pointer dereference on parsing.
466 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
467 otherwise parse PKCS#7 structures from untrusted sources are
468 affected. OpenSSL clients and servers are not affected.
470 This issue was reported to OpenSSL by Michal Zalewski (Google).
474 *) DoS via reachable assert in SSLv2 servers fix
476 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
477 servers that both support SSLv2 and enable export cipher suites by sending
478 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
480 This issue was discovered by Sean Burford (Google) and Emilia Käsper
481 (OpenSSL development team).
485 *) Empty CKE with client auth and DHE fix
487 If client auth is used then a server can seg fault in the event of a DHE
488 ciphersuite being selected and a zero length ClientKeyExchange message
489 being sent by the client. This could be exploited in a DoS attack.
493 *) Handshake with unseeded PRNG fix
495 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
496 with an unseeded PRNG. The conditions are:
497 - The client is on a platform where the PRNG has not been seeded
498 automatically, and the user has not seeded manually
499 - A protocol specific client method version has been used (i.e. not
500 SSL_client_methodv23)
501 - A ciphersuite is used that does not require additional random data from
502 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
504 If the handshake succeeds then the client random that has been used will
505 have been generated from a PRNG with insufficient entropy and therefore the
506 output may be predictable.
508 For example using the following command with an unseeded openssl will
509 succeed on an unpatched platform:
511 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
515 *) Use After Free following d2i_ECPrivatekey error fix
517 A malformed EC private key file consumed via the d2i_ECPrivateKey function
518 could cause a use after free condition. This, in turn, could cause a double
519 free in several private key parsing functions (such as d2i_PrivateKey
520 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
521 for applications that receive EC private keys from untrusted
522 sources. This scenario is considered rare.
524 This issue was discovered by the BoringSSL project and fixed in their
529 *) X509_to_X509_REQ NULL pointer deref fix
531 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
532 the certificate key is invalid. This function is rarely used in practice.
534 This issue was discovered by Brian Carpenter.
538 *) Removed the export ciphers from the DEFAULT ciphers
541 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
543 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
544 ARMv5 through ARMv8, as opposite to "locking" it to single one.
545 So far those who have to target multiple plaforms would compromise
546 and argue that binary targeting say ARMv5 would still execute on
547 ARMv8. "Universal" build resolves this compromise by providing
548 near-optimal performance even on newer platforms.
551 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
552 (other platforms pending).
553 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
555 *) Add support for the SignedCertificateTimestampList certificate and
556 OCSP response extensions from RFC6962.
559 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
560 for corner cases. (Certain input points at infinity could lead to
561 bogus results, with non-infinity inputs mapped to infinity too.)
564 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
565 This covers AES, SHA256/512 and GHASH. "Initial" means that most
566 common cases are optimized and there still is room for further
567 improvements. Vector Permutation AES for Altivec is also added.
570 *) Add support for little-endian ppc64 Linux target.
571 [Marcelo Cerri (IBM)]
573 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
574 SHA1, SHA256 and GHASH. "Initial" means that most common cases
575 are optimized and there still is room for further improvements.
576 Both 32- and 64-bit modes are supported.
577 [Andy Polyakov, Ard Biesheuvel (Linaro)]
579 *) Improved ARMv7 NEON support.
582 *) Support for SPARC Architecture 2011 crypto extensions, first
583 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
584 SHA256/512, MD5, GHASH and modular exponentiation.
585 [Andy Polyakov, David Miller]
587 *) Accelerated modular exponentiation for Intel processors, a.k.a.
589 [Shay Gueron & Vlad Krasnov (Intel Corp)]
591 *) Support for new and upcoming Intel processors, including AVX2,
592 BMI and SHA ISA extensions. This includes additional "stitched"
593 implementations, AESNI-SHA256 and GCM, and multi-buffer support
596 This work was sponsored by Intel Corp.
599 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
600 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
601 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
604 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
605 this fixes a limiation in previous versions of OpenSSL.
608 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
609 MGF1 digest and OAEP label.
612 *) Add EVP support for key wrapping algorithms, to avoid problems with
613 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
614 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
615 algorithms and include tests cases.
618 *) Add functions to allocate and set the fields of an ECDSA_METHOD
620 [Douglas E. Engert, Steve Henson]
622 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
623 difference in days and seconds between two tm or ASN1_TIME structures.
626 *) Add -rev test option to s_server to just reverse order of characters
627 received by client and send back to server. Also prints an abbreviated
628 summary of the connection parameters.
631 *) New option -brief for s_client and s_server to print out a brief summary
632 of connection parameters.
635 *) Add callbacks for arbitrary TLS extensions.
636 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
638 *) New option -crl_download in several openssl utilities to download CRLs
639 from CRLDP extension in certificates.
642 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
645 *) New function X509_CRL_diff to generate a delta CRL from the difference
646 of two full CRLs. Add support to "crl" utility.
649 *) New functions to set lookup_crls function and to retrieve
650 X509_STORE from X509_STORE_CTX.
653 *) Print out deprecated issuer and subject unique ID fields in
657 *) Extend OCSP I/O functions so they can be used for simple general purpose
658 HTTP as well as OCSP. New wrapper function which can be used to download
659 CRLs using the OCSP API.
662 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
665 *) SSL_CONF* functions. These provide a common framework for application
666 configuration using configuration files or command lines.
669 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
670 message callback and prints the results. Needs compile time option
671 "enable-ssl-trace". New options to s_client and s_server to enable
675 *) New ctrl and macro to retrieve supported points extensions.
676 Print out extension in s_server and s_client.
679 *) New functions to retrieve certificate signature and signature
683 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
687 *) New Suite B modes for TLS code. These use and enforce the requirements
688 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
689 only use Suite B curves. The Suite B modes can be set by using the
690 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
693 *) New chain verification flags for Suite B levels of security. Check
694 algorithms are acceptable when flags are set in X509_verify_cert.
697 *) Make tls1_check_chain return a set of flags indicating checks passed
698 by a certificate chain. Add additional tests to handle client
699 certificates: checks for matching certificate type and issuer name
703 *) If an attempt is made to use a signature algorithm not in the peer
704 preference list abort the handshake. If client has no suitable
705 signature algorithms in response to a certificate request do not
709 *) If server EC tmp key is not in client preference list abort handshake.
712 *) Add support for certificate stores in CERT structure. This makes it
713 possible to have different stores per SSL structure or one store in
714 the parent SSL_CTX. Include distint stores for certificate chain
715 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
716 to build and store a certificate chain in CERT structure: returing
717 an error if the chain cannot be built: this will allow applications
718 to test if a chain is correctly configured.
720 Note: if the CERT based stores are not set then the parent SSL_CTX
721 store is used to retain compatibility with existing behaviour.
725 *) New function ssl_set_client_disabled to set a ciphersuite disabled
726 mask based on the current session, check mask when sending client
727 hello and checking the requested ciphersuite.
730 *) New ctrls to retrieve and set certificate types in a certificate
731 request message. Print out received values in s_client. If certificate
732 types is not set with custom values set sensible values based on
733 supported signature algorithms.
736 *) Support for distinct client and server supported signature algorithms.
739 *) Add certificate callback. If set this is called whenever a certificate
740 is required by client or server. An application can decide which
741 certificate chain to present based on arbitrary criteria: for example
742 supported signature algorithms. Add very simple example to s_server.
743 This fixes many of the problems and restrictions of the existing client
744 certificate callback: for example you can now clear an existing
745 certificate and specify the whole chain.
748 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
749 the certificate can be used for (if anything). Set valid_flags field
750 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
751 to have similar checks in it.
753 Add new "cert_flags" field to CERT structure and include a "strict mode".
754 This enforces some TLS certificate requirements (such as only permitting
755 certificate signature algorithms contained in the supported algorithms
756 extension) which some implementations ignore: this option should be used
757 with caution as it could cause interoperability issues.
760 *) Update and tidy signature algorithm extension processing. Work out
761 shared signature algorithms based on preferences and peer algorithms
762 and print them out in s_client and s_server. Abort handshake if no
763 shared signature algorithms.
766 *) Add new functions to allow customised supported signature algorithms
767 for SSL and SSL_CTX structures. Add options to s_client and s_server
771 *) New function SSL_certs_clear() to delete all references to certificates
772 from an SSL structure. Before this once a certificate had been added
773 it couldn't be removed.
776 *) Integrate hostname, email address and IP address checking with certificate
777 verification. New verify options supporting checking in opensl utility.
780 *) Fixes and wildcard matching support to hostname and email checking
781 functions. Add manual page.
782 [Florian Weimer (Red Hat Product Security Team)]
784 *) New functions to check a hostname email or IP address against a
785 certificate. Add options x509 utility to print results of checks against
789 *) Fix OCSP checking.
790 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
792 *) Initial experimental support for explicitly trusted non-root CAs.
793 OpenSSL still tries to build a complete chain to a root but if an
794 intermediate CA has a trust setting included that is used. The first
795 setting is used: whether to trust (e.g., -addtrust option to the x509
799 *) Add -trusted_first option which attempts to find certificates in the
800 trusted store even if an untrusted chain is also supplied.
803 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
804 platform support for Linux and Android.
807 *) Support for linux-x32, ILP32 environment in x86_64 framework.
810 *) Experimental multi-implementation support for FIPS capable OpenSSL.
811 When in FIPS mode the approved implementations are used as normal,
812 when not in FIPS mode the internal unapproved versions are used instead.
813 This means that the FIPS capable OpenSSL isn't forced to use the
814 (often lower performance) FIPS implementations outside FIPS mode.
817 *) Transparently support X9.42 DH parameters when calling
818 PEM_read_bio_DHparameters. This means existing applications can handle
819 the new parameter format automatically.
822 *) Initial experimental support for X9.42 DH parameter format: mainly
823 to support use of 'q' parameter for RFC5114 parameters.
826 *) Add DH parameters from RFC5114 including test data to dhtest.
829 *) Support for automatic EC temporary key parameter selection. If enabled
830 the most preferred EC parameters are automatically used instead of
831 hardcoded fixed parameters. Now a server just has to call:
832 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
833 support ECDH and use the most appropriate parameters.
836 *) Enhance and tidy EC curve and point format TLS extension code. Use
837 static structures instead of allocation if default values are used.
838 New ctrls to set curves we wish to support and to retrieve shared curves.
839 Print out shared curves in s_server. New options to s_server and s_client
840 to set list of supported curves.
843 *) New ctrls to retrieve supported signature algorithms and
844 supported curve values as an array of NIDs. Extend openssl utility
845 to print out received values.
848 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
849 between NIDs and the more common NIST names such as "P-256". Enhance
850 ecparam utility and ECC method to recognise the NIST names for curves.
853 *) Enhance SSL/TLS certificate chain handling to support different
854 chains for each certificate instead of one chain in the parent SSL_CTX.
857 *) Support for fixed DH ciphersuite client authentication: where both
858 server and client use DH certificates with common parameters.
861 *) Support for fixed DH ciphersuites: those requiring DH server
865 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
867 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
868 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
869 X509_CINF_get_signature were reverted post internal team review.
871 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
873 *) Build fixes for the Windows and OpenVMS platforms
874 [Matt Caswell and Richard Levitte]
876 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
878 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
879 message can cause a segmentation fault in OpenSSL due to a NULL pointer
880 dereference. This could lead to a Denial Of Service attack. Thanks to
881 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
885 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
886 dtls1_buffer_record function under certain conditions. In particular this
887 could occur if an attacker sent repeated DTLS records with the same
888 sequence number but for the next epoch. The memory leak could be exploited
889 by an attacker in a Denial of Service attack through memory exhaustion.
890 Thanks to Chris Mueller for reporting this issue.
894 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
895 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
896 method would be set to NULL which could later result in a NULL pointer
897 dereference. Thanks to Frank Schmirler for reporting this issue.
901 *) Abort handshake if server key exchange message is omitted for ephemeral
904 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
905 reporting this issue.
909 *) Remove non-export ephemeral RSA code on client and server. This code
910 violated the TLS standard by allowing the use of temporary RSA keys in
911 non-export ciphersuites and could be used by a server to effectively
912 downgrade the RSA key length used to a value smaller than the server
913 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
914 INRIA or reporting this issue.
918 *) Fixed issue where DH client certificates are accepted without verification.
919 An OpenSSL server will accept a DH certificate for client authentication
920 without the certificate verify message. This effectively allows a client to
921 authenticate without the use of a private key. This only affects servers
922 which trust a client certificate authority which issues certificates
923 containing DH keys: these are extremely rare and hardly ever encountered.
924 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
929 *) Ensure that the session ID context of an SSL is updated when its
930 SSL_CTX is updated via SSL_set_SSL_CTX.
932 The session ID context is typically set from the parent SSL_CTX,
933 and can vary with the CTX.
936 *) Fix various certificate fingerprint issues.
938 By using non-DER or invalid encodings outside the signed portion of a
939 certificate the fingerprint can be changed without breaking the signature.
940 Although no details of the signed portion of the certificate can be changed
941 this can cause problems with some applications: e.g. those using the
942 certificate fingerprint for blacklists.
944 1. Reject signatures with non zero unused bits.
946 If the BIT STRING containing the signature has non zero unused bits reject
947 the signature. All current signature algorithms require zero unused bits.
949 2. Check certificate algorithm consistency.
951 Check the AlgorithmIdentifier inside TBS matches the one in the
952 certificate signature. NB: this will result in signature failure
953 errors for some broken certificates.
955 Thanks to Konrad Kraszewski from Google for reporting this issue.
957 3. Check DSA/ECDSA signatures use DER.
959 Reencode DSA/ECDSA signatures and compare with the original received
960 signature. Return an error if there is a mismatch.
962 This will reject various cases including garbage after signature
963 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
964 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
965 (negative or with leading zeroes).
967 Further analysis was conducted and fixes were developed by Stephen Henson
968 of the OpenSSL core team.
973 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
974 results on some platforms, including x86_64. This bug occurs at random
975 with a very low probability, and is not known to be exploitable in any
976 way, though its exact impact is difficult to determine. Thanks to Pieter
977 Wuille (Blockstream) who reported this issue and also suggested an initial
978 fix. Further analysis was conducted by the OpenSSL development team and
979 Adam Langley of Google. The final fix was developed by Andy Polyakov of
980 the OpenSSL core team.
984 *) Do not resume sessions on the server if the negotiated protocol
985 version does not match the session's version. Resuming with a different
986 version, while not strictly forbidden by the RFC, is of questionable
987 sanity and breaks all known clients.
988 [David Benjamin, Emilia Käsper]
990 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
991 early CCS messages during renegotiation. (Note that because
992 renegotiation is encrypted, this early CCS was not exploitable.)
995 *) Tighten client-side session ticket handling during renegotiation:
996 ensure that the client only accepts a session ticket if the server sends
997 the extension anew in the ServerHello. Previously, a TLS client would
998 reuse the old extension state and thus accept a session ticket if one was
999 announced in the initial ServerHello.
1001 Similarly, ensure that the client requires a session ticket if one
1002 was advertised in the ServerHello. Previously, a TLS client would
1003 ignore a missing NewSessionTicket message.
1006 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
1008 *) SRTP Memory Leak.
1010 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
1011 sends a carefully crafted handshake message, to cause OpenSSL to fail
1012 to free up to 64k of memory causing a memory leak. This could be
1013 exploited in a Denial Of Service attack. This issue affects OpenSSL
1014 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
1015 whether SRTP is used or configured. Implementations of OpenSSL that
1016 have been compiled with OPENSSL_NO_SRTP defined are not affected.
1018 The fix was developed by the OpenSSL team.
1022 *) Session Ticket Memory Leak.
1024 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
1025 integrity of that ticket is first verified. In the event of a session
1026 ticket integrity check failing, OpenSSL will fail to free memory
1027 causing a memory leak. By sending a large number of invalid session
1028 tickets an attacker could exploit this issue in a Denial Of Service
1033 *) Build option no-ssl3 is incomplete.
1035 When OpenSSL is configured with "no-ssl3" as a build option, servers
1036 could accept and complete a SSL 3.0 handshake, and clients could be
1037 configured to send them.
1039 [Akamai and the OpenSSL team]
1041 *) Add support for TLS_FALLBACK_SCSV.
1042 Client applications doing fallback retries should call
1043 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
1045 [Adam Langley, Bodo Moeller]
1047 *) Add additional DigestInfo checks.
1049 Reencode DigestInto in DER and check against the original when
1050 verifying RSA signature: this will reject any improperly encoded
1051 DigestInfo structures.
1053 Note: this is a precautionary measure and no attacks are currently known.
1057 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
1059 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
1060 SRP code can be overrun an internal buffer. Add sanity check that
1061 g, A, B < N to SRP code.
1063 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
1064 Group for discovering this issue.
1068 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
1069 TLS 1.0 instead of higher protocol versions when the ClientHello message
1070 is badly fragmented. This allows a man-in-the-middle attacker to force a
1071 downgrade to TLS 1.0 even if both the server and the client support a
1072 higher protocol version, by modifying the client's TLS records.
1074 Thanks to David Benjamin and Adam Langley (Google) for discovering and
1075 researching this issue.
1079 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
1080 to a denial of service attack. A malicious server can crash the client
1081 with a null pointer dereference (read) by specifying an anonymous (EC)DH
1082 ciphersuite and sending carefully crafted handshake messages.
1084 Thanks to Felix Gröbert (Google) for discovering and researching this
1089 *) By sending carefully crafted DTLS packets an attacker could cause openssl
1090 to leak memory. This can be exploited through a Denial of Service attack.
1091 Thanks to Adam Langley for discovering and researching this issue.
1095 *) An attacker can force openssl to consume large amounts of memory whilst
1096 processing DTLS handshake messages. This can be exploited through a
1097 Denial of Service attack.
1098 Thanks to Adam Langley for discovering and researching this issue.
1102 *) An attacker can force an error condition which causes openssl to crash
1103 whilst processing DTLS packets due to memory being freed twice. This
1104 can be exploited through a Denial of Service attack.
1105 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
1110 *) If a multithreaded client connects to a malicious server using a resumed
1111 session and the server sends an ec point format extension it could write
1112 up to 255 bytes to freed memory.
1114 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
1119 *) A malicious server can crash an OpenSSL client with a null pointer
1120 dereference (read) by specifying an SRP ciphersuite even though it was not
1121 properly negotiated with the client. This can be exploited through a
1122 Denial of Service attack.
1124 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
1125 discovering and researching this issue.
1129 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
1130 X509_name_oneline, X509_name_print_ex et al. to leak some information
1131 from the stack. Applications may be affected if they echo pretty printing
1132 output to the attacker.
1134 Thanks to Ivan Fratric (Google) for discovering this issue.
1136 [Emilia Käsper, and Steve Henson]
1138 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1139 for corner cases. (Certain input points at infinity could lead to
1140 bogus results, with non-infinity inputs mapped to infinity too.)
1143 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
1145 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
1146 handshake can force the use of weak keying material in OpenSSL
1147 SSL/TLS clients and servers.
1149 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
1150 researching this issue. (CVE-2014-0224)
1151 [KIKUCHI Masashi, Steve Henson]
1153 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
1154 OpenSSL DTLS client the code can be made to recurse eventually crashing
1157 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
1159 [Imre Rad, Steve Henson]
1161 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
1162 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
1163 client or server. This is potentially exploitable to run arbitrary
1164 code on a vulnerable client or server.
1166 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
1167 [Jüri Aedla, Steve Henson]
1169 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
1170 are subject to a denial of service attack.
1172 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
1173 this issue. (CVE-2014-3470)
1174 [Felix Gröbert, Ivan Fratric, Steve Henson]
1176 *) Harmonize version and its documentation. -f flag is used to display
1178 [mancha <mancha1@zoho.com>]
1180 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1181 in i2d_ECPrivateKey.
1182 [mancha <mancha1@zoho.com>]
1184 *) Fix some double frees. These are not thought to be exploitable.
1185 [mancha <mancha1@zoho.com>]
1187 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
1189 *) A missing bounds check in the handling of the TLS heartbeat extension
1190 can be used to reveal up to 64k of memory to a connected client or
1193 Thanks for Neel Mehta of Google Security for discovering this bug and to
1194 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1195 preparing the fix (CVE-2014-0160)
1196 [Adam Langley, Bodo Moeller]
1198 *) Fix for the attack described in the paper "Recovering OpenSSL
1199 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1200 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1201 http://eprint.iacr.org/2014/140
1203 Thanks to Yuval Yarom and Naomi Benger for discovering this
1204 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1205 [Yuval Yarom and Naomi Benger]
1207 *) TLS pad extension: draft-agl-tls-padding-03
1209 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
1210 TLS client Hello record length value would otherwise be > 255 and
1211 less that 512 pad with a dummy extension containing zeroes so it
1212 is at least 512 bytes long.
1214 [Adam Langley, Steve Henson]
1216 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1218 *) Fix for TLS record tampering bug. A carefully crafted invalid
1219 handshake could crash OpenSSL with a NULL pointer exception.
1220 Thanks to Anton Johansson for reporting this issues.
1223 *) Keep original DTLS digest and encryption contexts in retransmission
1224 structures so we can use the previous session parameters if they need
1225 to be resent. (CVE-2013-6450)
1228 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1229 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1230 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1231 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1232 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1233 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1234 [Rob Stradling, Adam Langley]
1236 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1238 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1239 supporting platforms or when small records were transferred.
1240 [Andy Polyakov, Steve Henson]
1242 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1244 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1246 This addresses the flaw in CBC record processing discovered by
1247 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1248 at: http://www.isg.rhul.ac.uk/tls/
1250 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1251 Security Group at Royal Holloway, University of London
1252 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1253 Emilia Käsper for the initial patch.
1255 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1257 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1258 ciphersuites which can be exploited in a denial of service attack.
1259 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1260 and detecting this bug and to Wolfgang Ettlinger
1261 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1265 *) Return an error when checking OCSP signatures when key is NULL.
1266 This fixes a DoS attack. (CVE-2013-0166)
1269 *) Make openssl verify return errors.
1270 [Chris Palmer <palmer@google.com> and Ben Laurie]
1272 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1273 the right response is stapled. Also change SSL_get_certificate()
1274 so it returns the certificate actually sent.
1275 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1276 [Rob Stradling <rob.stradling@comodo.com>]
1278 *) Fix possible deadlock when decoding public keys.
1281 *) Don't use TLS 1.0 record version number in initial client hello
1285 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1287 *) Sanity check record length before skipping explicit IV in TLS
1288 1.2, 1.1 and DTLS to fix DoS attack.
1290 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1291 fuzzing as a service testing platform.
1295 *) Initialise tkeylen properly when encrypting CMS messages.
1296 Thanks to Solar Designer of Openwall for reporting this issue.
1299 *) In FIPS mode don't try to use composite ciphers as they are not
1303 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1305 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1306 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1307 mean any application compiled against OpenSSL 1.0.0 headers setting
1308 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1309 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1310 0x10000000L Any application which was previously compiled against
1311 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1312 will need to be recompiled as a result. Letting be results in
1313 inability to disable specifically TLS 1.1 and in client context,
1314 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1317 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1318 disable just protocol X, but all protocols above X *if* there are
1319 protocols *below* X still enabled. In more practical terms it means
1320 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1321 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1322 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1326 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1328 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1329 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1330 in CRYPTO_realloc_clean.
1332 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1333 issue and to Adam Langley <agl@chromium.org> for fixing it.
1335 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1337 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1340 *) Workarounds for some broken servers that "hang" if a client hello
1341 record length exceeds 255 bytes.
1343 1. Do not use record version number > TLS 1.0 in initial client
1344 hello: some (but not all) hanging servers will now work.
1345 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1346 the number of ciphers sent in the client hello. This should be
1347 set to an even number, such as 50, for example by passing:
1348 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1349 Most broken servers should now work.
1350 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1351 TLS 1.2 client support entirely.
1354 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1357 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1359 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1360 STRING form instead of a DigestInfo.
1363 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1364 and the RSA_sign/RSA_verify functions. This was made more apparent when
1365 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1366 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1367 the correct format in RSA_verify so both forms transparently work.
1370 *) Some servers which support TLS 1.0 can choke if we initially indicate
1371 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1372 encrypted premaster secret. As a workaround use the maximum permitted
1373 client version in client hello, this should keep such servers happy
1374 and still work with previous versions of OpenSSL.
1377 *) Add support for TLS/DTLS heartbeats.
1378 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1380 *) Add support for SCTP.
1381 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1383 *) Improved PRNG seeding for VOS.
1384 [Paul Green <Paul.Green@stratus.com>]
1386 *) Extensive assembler packs updates, most notably:
1388 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1389 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1390 - x86_64: bit-sliced AES implementation;
1391 - ARM: NEON support, contemporary platforms optimizations;
1392 - s390x: z196 support;
1393 - *: GHASH and GF(2^m) multiplication implementations;
1397 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1398 (removal of unnecessary code)
1399 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1401 *) Add TLS key material exporter from RFC 5705.
1404 *) Add DTLS-SRTP negotiation from RFC 5764.
1407 *) Add Next Protocol Negotiation,
1408 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1409 disabled with a no-npn flag to config or Configure. Code donated
1411 [Adam Langley <agl@google.com> and Ben Laurie]
1413 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1414 NIST-P256, NIST-P521, with constant-time single point multiplication on
1415 typical inputs. Compiler support for the nonstandard type __uint128_t is
1416 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1417 Code made available under Apache License version 2.0.
1419 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1420 line to include this in your build of OpenSSL, and run "make depend" (or
1421 "make update"). This enables the following EC_METHODs:
1423 EC_GFp_nistp224_method()
1424 EC_GFp_nistp256_method()
1425 EC_GFp_nistp521_method()
1427 EC_GROUP_new_by_curve_name() will automatically use these (while
1428 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1430 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1432 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1433 all platforms. Move ssize_t definition from e_os.h to the public
1434 header file e_os2.h as it now appears in public header file cms.h
1437 *) New -sigopt option to the ca, req and x509 utilities. Additional
1438 signature parameters can be passed using this option and in
1442 *) Add RSA PSS signing function. This will generate and set the
1443 appropriate AlgorithmIdentifiers for PSS based on those in the
1444 corresponding EVP_MD_CTX structure. No application support yet.
1447 *) Support for companion algorithm specific ASN1 signing routines.
1448 New function ASN1_item_sign_ctx() signs a pre-initialised
1449 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1450 the appropriate parameters.
1453 *) Add new algorithm specific ASN1 verification initialisation function
1454 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1455 handling will be the same no matter what EVP_PKEY_METHOD is used.
1456 Add a PSS handler to support verification of PSS signatures: checked
1457 against a number of sample certificates.
1460 *) Add signature printing for PSS. Add PSS OIDs.
1461 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1463 *) Add algorithm specific signature printing. An individual ASN1 method
1464 can now print out signatures instead of the standard hex dump.
1466 More complex signatures (e.g. PSS) can print out more meaningful
1467 information. Include DSA version that prints out the signature
1471 *) Password based recipient info support for CMS library: implementing
1475 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1476 neatly separates the code into cipher and PBE sections and is required
1477 for some algorithms that split PBES2 into separate pieces (such as
1478 password based CMS).
1481 *) Session-handling fixes:
1482 - Fix handling of connections that are resuming with a session ID,
1483 but also support Session Tickets.
1484 - Fix a bug that suppressed issuing of a new ticket if the client
1485 presented a ticket with an expired session.
1486 - Try to set the ticket lifetime hint to something reasonable.
1487 - Make tickets shorter by excluding irrelevant information.
1488 - On the client side, don't ignore renewed tickets.
1489 [Adam Langley, Bodo Moeller (Google)]
1491 *) Fix PSK session representation.
1494 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1496 This work was sponsored by Intel.
1499 *) Add GCM support to TLS library. Some custom code is needed to split
1500 the IV between the fixed (from PRF) and explicit (from TLS record)
1501 portions. This adds all GCM ciphersuites supported by RFC5288 and
1502 RFC5289. Generalise some AES* cipherstrings to include GCM and
1503 add a special AESGCM string for GCM only.
1506 *) Expand range of ctrls for AES GCM. Permit setting invocation
1507 field on decrypt and retrieval of invocation field only on encrypt.
1510 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1511 As required by RFC5289 these ciphersuites cannot be used if for
1512 versions of TLS earlier than 1.2.
1515 *) For FIPS capable OpenSSL interpret a NULL default public key method
1516 as unset and return the appropriate default but do *not* set the default.
1517 This means we can return the appropriate method in applications that
1518 switch between FIPS and non-FIPS modes.
1521 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1522 ENGINE is used then we cannot handle that in the FIPS module so we
1523 keep original code iff non-FIPS operations are allowed.
1526 *) Add -attime option to openssl utilities.
1527 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1529 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1532 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1533 FIPS EC methods unconditionally for now.
1536 *) New build option no-ec2m to disable characteristic 2 code.
1539 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1540 all cases can be covered as some introduce binary incompatibilities.
1543 *) Redirect RSA operations to FIPS module including keygen,
1544 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1547 *) Add similar low level API blocking to ciphers.
1550 *) Low level digest APIs are not approved in FIPS mode: any attempt
1551 to use these will cause a fatal error. Applications that *really* want
1552 to use them can use the private_* version instead.
1555 *) Redirect cipher operations to FIPS module for FIPS builds.
1558 *) Redirect digest operations to FIPS module for FIPS builds.
1561 *) Update build system to add "fips" flag which will link in fipscanister.o
1562 for static and shared library builds embedding a signature if needed.
1565 *) Output TLS supported curves in preference order instead of numerical
1566 order. This is currently hardcoded for the highest order curves first.
1567 This should be configurable so applications can judge speed vs strength.
1570 *) Add TLS v1.2 server support for client authentication.
1573 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1577 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1578 FIPS modules versions.
1581 *) Add TLS v1.2 client side support for client authentication. Keep cache
1582 of handshake records longer as we don't know the hash algorithm to use
1583 until after the certificate request message is received.
1586 *) Initial TLS v1.2 client support. Add a default signature algorithms
1587 extension including all the algorithms we support. Parse new signature
1588 format in client key exchange. Relax some ECC signing restrictions for
1589 TLS v1.2 as indicated in RFC5246.
1592 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1593 to new signature format when needed using client digest preference.
1594 All server ciphersuites should now work correctly in TLS v1.2. No client
1595 support yet and no support for client certificates.
1598 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1599 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1600 ciphersuites. At present only RSA key exchange ciphersuites work with
1601 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1602 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1603 and version checking.
1606 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1607 with this defined it will not be affected by any changes to ssl internal
1608 structures. Add several utility functions to allow openssl application
1609 to work with OPENSSL_NO_SSL_INTERN defined.
1613 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1615 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1618 *) Permit abbreviated handshakes when renegotiating using the function
1619 SSL_renegotiate_abbreviated().
1620 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1622 *) Add call to ENGINE_register_all_complete() to
1623 ENGINE_load_builtin_engines(), so some implementations get used
1624 automatically instead of needing explicit application support.
1627 *) Add support for TLS key exporter as described in RFC5705.
1628 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1630 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1631 a few changes are required:
1633 Add SSL_OP_NO_TLSv1_1 flag.
1634 Add TLSv1_1 methods.
1635 Update version checking logic to handle version 1.1.
1636 Add explicit IV handling (ported from DTLS code).
1637 Add command line options to s_client/s_server.
1640 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1642 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1643 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1644 content decryption and always return the same error. Note: this attack
1645 needs on average 2^20 messages so it only affects automated senders. The
1646 old behaviour can be reenabled in the CMS code by setting the
1647 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1648 an MMA defence is not necessary.
1649 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1650 this issue. (CVE-2012-0884)
1653 *) Fix CVE-2011-4619: make sure we really are receiving a
1654 client hello before rejecting multiple SGC restarts. Thanks to
1655 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1658 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1660 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1661 Thanks to Antonio Martin, Enterprise Secure Access Research and
1662 Development, Cisco Systems, Inc. for discovering this bug and
1663 preparing a fix. (CVE-2012-0050)
1666 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1668 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1669 of the Vaudenay padding oracle attack on CBC mode encryption
1670 which enables an efficient plaintext recovery attack against
1671 the OpenSSL implementation of DTLS. Their attack exploits timing
1672 differences arising during decryption processing. A research
1673 paper describing this attack can be found at:
1674 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1675 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1676 Security Group at Royal Holloway, University of London
1677 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1678 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1679 for preparing the fix. (CVE-2011-4108)
1680 [Robin Seggelmann, Michael Tuexen]
1682 *) Clear bytes used for block padding of SSL 3.0 records.
1684 [Adam Langley (Google)]
1686 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1687 Kadianakis <desnacked@gmail.com> for discovering this issue and
1688 Adam Langley for preparing the fix. (CVE-2011-4619)
1689 [Adam Langley (Google)]
1691 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1692 [Andrey Kulikov <amdeich@gmail.com>]
1694 *) Prevent malformed RFC3779 data triggering an assertion failure.
1695 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1696 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1697 [Rob Austein <sra@hactrn.net>]
1699 *) Improved PRNG seeding for VOS.
1700 [Paul Green <Paul.Green@stratus.com>]
1702 *) Fix ssl_ciph.c set-up race.
1703 [Adam Langley (Google)]
1705 *) Fix spurious failures in ecdsatest.c.
1706 [Emilia Käsper (Google)]
1708 *) Fix the BIO_f_buffer() implementation (which was mixing different
1709 interpretations of the '..._len' fields).
1710 [Adam Langley (Google)]
1712 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1713 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1714 threads won't reuse the same blinding coefficients.
1716 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1717 lock to call BN_BLINDING_invert_ex, and avoids one use of
1718 BN_BLINDING_update for each BN_BLINDING structure (previously,
1719 the last update always remained unused).
1720 [Emilia Käsper (Google)]
1722 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1723 [Bob Buckholz (Google)]
1725 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1727 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1728 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1729 [Kaspar Brand <ossl@velox.ch>]
1731 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1732 for multi-threaded use of ECDH. (CVE-2011-3210)
1733 [Adam Langley (Google)]
1735 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1738 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1739 signature public key algorithm by using OID xref utilities instead.
1740 Before this you could only use some ECC ciphersuites with SHA1 only.
1743 *) Add protection against ECDSA timing attacks as mentioned in the paper
1744 by Billy Bob Brumley and Nicola Tuveri, see:
1746 http://eprint.iacr.org/2011/232.pdf
1748 [Billy Bob Brumley and Nicola Tuveri]
1750 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1752 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1753 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1755 *) Fix bug in string printing code: if *any* escaping is enabled we must
1756 escape the escape character (backslash) or the resulting string is
1760 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1762 *) Disable code workaround for ancient and obsolete Netscape browsers
1763 and servers: an attacker can use it in a ciphersuite downgrade attack.
1764 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1767 *) Fixed J-PAKE implementation error, originally discovered by
1768 Sebastien Martini, further info and confirmation from Stefan
1769 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1772 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1774 *) Fix extension code to avoid race conditions which can result in a buffer
1775 overrun vulnerability: resumed sessions must not be modified as they can
1776 be shared by multiple threads. CVE-2010-3864
1779 *) Fix WIN32 build system to correctly link an ENGINE directory into
1783 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1785 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1787 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1789 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1791 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1792 context. The operation can be customised via the ctrl mechanism in
1793 case ENGINEs want to include additional functionality.
1796 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1799 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1800 output hashes compatible with older versions of OpenSSL.
1801 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1803 *) Fix compression algorithm handling: if resuming a session use the
1804 compression algorithm of the resumed session instead of determining
1805 it from client hello again. Don't allow server to change algorithm.
1808 *) Add load_crls() function to apps tidying load_certs() too. Add option
1809 to verify utility to allow additional CRLs to be included.
1812 *) Update OCSP request code to permit adding custom headers to the request:
1813 some responders need this.
1816 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1818 [Julia Lawall <julia@diku.dk>]
1820 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1821 needlessly dereferenced structures, used obsolete functions and
1822 didn't handle all updated verify codes correctly.
1825 *) Disable MD2 in the default configuration.
1828 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1829 indicate the initial BIO being pushed or popped. This makes it possible
1830 to determine whether the BIO is the one explicitly called or as a result
1831 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1832 it handles reference counts correctly and doesn't zero out the I/O bio
1833 when it is not being explicitly popped. WARNING: applications which
1834 included workarounds for the old buggy behaviour will need to be modified
1835 or they could free up already freed BIOs.
1838 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1839 renaming to all platforms (within the 0.9.8 branch, this was
1840 done conditionally on Netware platforms to avoid a name clash).
1841 [Guenter <lists@gknw.net>]
1843 *) Add ECDHE and PSK support to DTLS.
1844 [Michael Tuexen <tuexen@fh-muenster.de>]
1846 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1850 *) Add "missing" function EVP_MD_flags() (without this the only way to
1851 retrieve a digest flags is by accessing the structure directly. Update
1852 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1853 or cipher is registered as in the "from" argument. Print out all
1854 registered digests in the dgst usage message instead of manually
1855 attempting to work them out.
1858 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1859 this allows the use of compression and extensions. Change default cipher
1860 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1861 by default unless an application cipher string requests it.
1864 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1865 key ids to find matching certificates and keys but some PKCS#12 files
1866 don't follow the (somewhat unwritten) rules and this strategy fails.
1867 Now just gather all certificates together and the first private key
1868 then look for the first certificate that matches the key.
1871 *) Support use of registered digest and cipher names for dgst and cipher
1872 commands instead of having to add each one as a special case. So now
1879 openssl dgst -sha256 foo
1881 and this works for ENGINE based algorithms too.
1885 *) Update Gost ENGINE to support parameter files.
1886 [Victor B. Wagner <vitus@cryptocom.ru>]
1888 *) Support GeneralizedTime in ca utility.
1889 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1891 *) Enhance the hash format used for certificate directory links. The new
1892 form uses the canonical encoding (meaning equivalent names will work
1893 even if they aren't identical) and uses SHA1 instead of MD5. This form
1894 is incompatible with the older format and as a result c_rehash should
1895 be used to rebuild symbolic links.
1898 *) Make PKCS#8 the default write format for private keys, replacing the
1899 traditional format. This form is standardised, more secure and doesn't
1900 include an implicit MD5 dependency.
1903 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1904 committed to OpenSSL should pass this lot as a minimum.
1907 *) Add session ticket override functionality for use by EAP-FAST.
1908 [Jouni Malinen <j@w1.fi>]
1910 *) Modify HMAC functions to return a value. Since these can be implemented
1911 in an ENGINE errors can occur.
1914 *) Type-checked OBJ_bsearch_ex.
1917 *) Type-checked OBJ_bsearch. Also some constification necessitated
1918 by type-checking. Still to come: TXT_DB, bsearch(?),
1919 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1923 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1924 seconds to a tm structure directly, instead of going through OS
1925 specific date routines. This avoids any issues with OS routines such
1926 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1927 and X509_time_adj_ex() to cover the extended range. The existing
1928 X509_time_adj() is still usable and will no longer have any date issues.
1931 *) Delta CRL support. New use deltas option which will attempt to locate
1932 and search any appropriate delta CRLs available.
1934 This work was sponsored by Google.
1937 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1938 code and add additional score elements. Validate alternate CRL paths
1939 as part of the CRL checking and indicate a new error "CRL path validation
1940 error" in this case. Applications wanting additional details can use
1941 the verify callback and check the new "parent" field. If this is not
1942 NULL CRL path validation is taking place. Existing applications wont
1943 see this because it requires extended CRL support which is off by
1946 This work was sponsored by Google.
1949 *) Support for freshest CRL extension.
1951 This work was sponsored by Google.
1954 *) Initial indirect CRL support. Currently only supported in the CRLs
1955 passed directly and not via lookup. Process certificate issuer
1956 CRL entry extension and lookup CRL entries by bother issuer name
1957 and serial number. Check and process CRL issuer entry in IDP extension.
1959 This work was sponsored by Google.
1962 *) Add support for distinct certificate and CRL paths. The CRL issuer
1963 certificate is validated separately in this case. Only enabled if
1964 an extended CRL support flag is set: this flag will enable additional
1965 CRL functionality in future.
1967 This work was sponsored by Google.
1970 *) Add support for policy mappings extension.
1972 This work was sponsored by Google.
1975 *) Fixes to pathlength constraint, self issued certificate handling,
1976 policy processing to align with RFC3280 and PKITS tests.
1978 This work was sponsored by Google.
1981 *) Support for name constraints certificate extension. DN, email, DNS
1982 and URI types are currently supported.
1984 This work was sponsored by Google.
1987 *) To cater for systems that provide a pointer-based thread ID rather
1988 than numeric, deprecate the current numeric thread ID mechanism and
1989 replace it with a structure and associated callback type. This
1990 mechanism allows a numeric "hash" to be extracted from a thread ID in
1991 either case, and on platforms where pointers are larger than 'long',
1992 mixing is done to help ensure the numeric 'hash' is usable even if it
1993 can't be guaranteed unique. The default mechanism is to use "&errno"
1994 as a pointer-based thread ID to distinguish between threads.
1996 Applications that want to provide their own thread IDs should now use
1997 CRYPTO_THREADID_set_callback() to register a callback that will call
1998 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
2000 Note that ERR_remove_state() is now deprecated, because it is tied
2001 to the assumption that thread IDs are numeric. ERR_remove_state(0)
2002 to free the current thread's error state should be replaced by
2003 ERR_remove_thread_state(NULL).
2005 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
2006 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
2007 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
2008 application was previously providing a numeric thread callback that
2009 was inappropriate for distinguishing threads, then uniqueness might
2010 have been obtained with &errno that happened immediately in the
2011 intermediate development versions of OpenSSL; this is no longer the
2012 case, the numeric thread callback will now override the automatic use
2014 [Geoff Thorpe, with help from Bodo Moeller]
2016 *) Initial support for different CRL issuing certificates. This covers a
2017 simple case where the self issued certificates in the chain exist and
2018 the real CRL issuer is higher in the existing chain.
2020 This work was sponsored by Google.
2023 *) Removed effectively defunct crypto/store from the build.
2026 *) Revamp of STACK to provide stronger type-checking. Still to come:
2027 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
2028 ASN1_STRING, CONF_VALUE.
2031 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
2032 RAM on SSL connections. This option can save about 34k per idle SSL.
2035 *) Revamp of LHASH to provide stronger type-checking. Still to come:
2036 STACK, TXT_DB, bsearch, qsort.
2039 *) Initial support for Cryptographic Message Syntax (aka CMS) based
2040 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
2041 support for data, signedData, compressedData, digestedData and
2042 encryptedData, envelopedData types included. Scripts to check against
2043 RFC4134 examples draft and interop and consistency checks of many
2044 content types and variants.
2047 *) Add options to enc utility to support use of zlib compression BIO.
2050 *) Extend mk1mf to support importing of options and assembly language
2051 files from Configure script, currently only included in VC-WIN32.
2052 The assembly language rules can now optionally generate the source
2053 files from the associated perl scripts.
2056 *) Implement remaining functionality needed to support GOST ciphersuites.
2057 Interop testing has been performed using CryptoPro implementations.
2058 [Victor B. Wagner <vitus@cryptocom.ru>]
2060 *) s390x assembler pack.
2063 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
2067 *) Implement Opaque PRF Input TLS extension as specified in
2068 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
2069 official specification yet and no extension type assignment by
2070 IANA exists, this extension (for now) will have to be explicitly
2071 enabled when building OpenSSL by providing the extension number
2072 to use. For example, specify an option
2074 -DTLSEXT_TYPE_opaque_prf_input=0x9527
2076 to the "config" or "Configure" script to enable the extension,
2077 assuming extension number 0x9527 (which is a completely arbitrary
2078 and unofficial assignment based on the MD5 hash of the Internet
2079 Draft). Note that by doing so, you potentially lose
2080 interoperability with other TLS implementations since these might
2081 be using the same extension number for other purposes.
2083 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
2084 opaque PRF input value to use in the handshake. This will create
2085 an interal copy of the length-'len' string at 'src', and will
2086 return non-zero for success.
2088 To get more control and flexibility, provide a callback function
2091 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
2092 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
2096 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
2099 Callback function 'cb' will be called in handshakes, and is
2100 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
2101 Argument 'arg' is for application purposes (the value as given to
2102 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
2103 be provided to the callback function). The callback function
2104 has to return non-zero to report success: usually 1 to use opaque
2105 PRF input just if possible, or 2 to enforce use of the opaque PRF
2106 input. In the latter case, the library will abort the handshake
2107 if opaque PRF input is not successfully negotiated.
2109 Arguments 'peerinput' and 'len' given to the callback function
2110 will always be NULL and 0 in the case of a client. A server will
2111 see the client's opaque PRF input through these variables if
2112 available (NULL and 0 otherwise). Note that if the server
2113 provides an opaque PRF input, the length must be the same as the
2114 length of the client's opaque PRF input.
2116 Note that the callback function will only be called when creating
2117 a new session (session resumption can resume whatever was
2118 previously negotiated), and will not be called in SSL 2.0
2119 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
2120 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
2121 for applications that need to enforce opaque PRF input.
2125 *) Update ssl code to support digests other than SHA1+MD5 for handshake
2128 [Victor B. Wagner <vitus@cryptocom.ru>]
2130 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2131 RFC4507bis. The encrypted ticket format is an encrypted encoded
2132 SSL_SESSION structure, that way new session features are automatically
2135 If a client application caches session in an SSL_SESSION structure
2136 support is transparent because tickets are now stored in the encoded
2139 The SSL_CTX structure automatically generates keys for ticket
2140 protection in servers so again support should be possible
2141 with no application modification.
2143 If a client or server wishes to disable RFC4507 support then the option
2144 SSL_OP_NO_TICKET can be set.
2146 Add a TLS extension debugging callback to allow the contents of any client
2147 or server extensions to be examined.
2149 This work was sponsored by Google.
2152 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
2153 OpenSSL should now compile cleanly on gcc 4.2
2154 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
2156 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
2157 support including streaming MAC support: this is required for GOST
2158 ciphersuite support.
2159 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2161 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2162 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2163 to output in BER and PEM format.
2166 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2167 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2168 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2169 ENGINE support for HMAC keys which are unextractable. New -mac and
2170 -macopt options to dgst utility.
2173 *) New option -sigopt to dgst utility. Update dgst to use
2174 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2175 alternative signing paramaters such as X9.31 or PSS in the dgst
2179 *) Change ssl_cipher_apply_rule(), the internal function that does
2180 the work each time a ciphersuite string requests enabling
2181 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2182 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2183 the order of disabled ciphersuites such that those ciphersuites
2184 that most recently went from enabled to disabled not only stay
2185 in order with respect to each other, but also have higher priority
2186 than other disabled ciphersuites the next time ciphersuites are
2189 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2190 the same ciphersuites as with "HIGH" alone, but in a specific
2191 order where the PSK ciphersuites come first (since they are the
2192 most recently disabled ciphersuites when "HIGH" is parsed).
2194 Also, change ssl_create_cipher_list() (using this new
2195 funcionality) such that between otherwise identical
2196 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2200 *) Change ssl_create_cipher_list() so that it automatically
2201 arranges the ciphersuites in reasonable order before starting
2202 to process the rule string. Thus, the definition for "DEFAULT"
2203 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2204 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2205 This makes it much easier to arrive at a reasonable default order
2206 in applications for which anonymous ciphers are OK (meaning
2207 that you can't actually use DEFAULT).
2208 [Bodo Moeller; suggested by Victor Duchovni]
2210 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2211 processing) into multiple integers instead of setting
2212 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2213 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2214 (These masks as well as the individual bit definitions are hidden
2215 away into the non-exported interface ssl/ssl_locl.h, so this
2216 change to the definition of the SSL_CIPHER structure shouldn't
2217 affect applications.) This give us more bits for each of these
2218 categories, so there is no longer a need to coagulate AES128 and
2219 AES256 into a single algorithm bit, and to coagulate Camellia128
2220 and Camellia256 into a single algorithm bit, which has led to all
2223 Thus, among other things, the kludge introduced in 0.9.7m and
2224 0.9.8e for masking out AES256 independently of AES128 or masking
2225 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2227 With the change, we also introduce new ciphersuite aliases that
2228 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2232 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2233 Use the leftmost N bytes of the signature input if the input is
2234 larger than the prime q (with N being the size in bytes of q).
2237 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2238 it yet and it is largely untested.
2241 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2244 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2245 some compilers (gcc 4.2 and later) reject their use. Safestack is
2246 reimplemented. Update ASN1 to avoid use of legacy functions.
2249 *) Win32/64 targets are linked with Winsock2.
2252 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2253 to external functions. This can be used to increase CRL handling
2254 efficiency especially when CRLs are very large by (for example) storing
2255 the CRL revoked certificates in a database.
2258 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2259 new CRLs added to a directory can be used. New command line option
2260 -verify_return_error to s_client and s_server. This causes real errors
2261 to be returned by the verify callback instead of carrying on no matter
2262 what. This reflects the way a "real world" verify callback would behave.
2265 *) GOST engine, supporting several GOST algorithms and public key formats.
2266 Kindly donated by Cryptocom.
2269 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2270 partitioned by DP are handled but no indirect CRL or reason partitioning
2271 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2272 selected via a scoring technique which handles IDP and AKID in CRLs.
2275 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2276 will ultimately be used for all verify operations: this will remove the
2277 X509_STORE dependency on certificate verification and allow alternative
2278 lookup methods. X509_STORE based implementations of these two callbacks.
2281 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2282 Modify get_crl() to find a valid (unexpired) CRL if possible.
2285 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2286 this would be called X509_CRL_cmp() but that name is already used by
2287 a function that just compares CRL issuer names. Cache several CRL
2288 extensions in X509_CRL structure and cache CRLDP in X509.
2291 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2292 this maps equivalent X509_NAME structures into a consistent structure.
2293 Name comparison can then be performed rapidly using memcmp().
2296 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2300 *) Allow digests to supply their own micalg string for S/MIME type using
2301 the ctrl EVP_MD_CTRL_MICALG.
2304 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2305 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2306 ctrl. It can then customise the structure before and/or after signing
2310 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2311 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2312 to free up any added signature OIDs.
2315 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2316 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2317 digest and cipher tables. New options added to openssl utility:
2318 list-message-digest-algorithms and list-cipher-algorithms.
2321 *) Change the array representation of binary polynomials: the list
2322 of degrees of non-zero coefficients is now terminated with -1.
2323 Previously it was terminated with 0, which was also part of the
2324 value; thus, the array representation was not applicable to
2325 polynomials where t^0 has coefficient zero. This change makes
2326 the array representation useful in a more general context.
2329 *) Various modifications and fixes to SSL/TLS cipher string
2330 handling. For ECC, the code now distinguishes between fixed ECDH
2331 with RSA certificates on the one hand and with ECDSA certificates
2332 on the other hand, since these are separate ciphersuites. The
2333 unused code for Fortezza ciphersuites has been removed.
2335 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2336 (not "ECDHE"). For consistency with the code for DH
2337 certificates, use of ECDH certificates is now considered ECDH
2338 authentication, not RSA or ECDSA authentication (the latter is
2339 merely the CA's signing algorithm and not actively used in the
2342 The temporary ciphersuite alias "ECCdraft" is no longer
2343 available, and ECC ciphersuites are no longer excluded from "ALL"
2344 and "DEFAULT". The following aliases now exist for RFC 4492
2345 ciphersuites, most of these by analogy with the DH case:
2347 kECDHr - ECDH cert, signed with RSA
2348 kECDHe - ECDH cert, signed with ECDSA
2349 kECDH - ECDH cert (signed with either RSA or ECDSA)
2350 kEECDH - ephemeral ECDH
2351 ECDH - ECDH cert or ephemeral ECDH
2357 AECDH - anonymous ECDH
2358 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2362 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2363 Use correct micalg parameters depending on digest(s) in signed message.
2366 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2367 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2370 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2371 an engine to register a method. Add ENGINE lookups for methods and
2372 functional reference processing.
2375 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2376 EVP_{Sign,Verify}* which allow an application to customise the signature
2380 *) New -resign option to smime utility. This adds one or more signers
2381 to an existing PKCS#7 signedData structure. Also -md option to use an
2382 alternative message digest algorithm for signing.
2385 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2386 create PKCS7 structures containing multiple signers. Update smime
2387 application to support multiple signers.
2390 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2394 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2395 Reorganize PBE internals to lookup from a static table using NIDs,
2396 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2397 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2398 PRF which will be automatically used with PBES2.
2401 *) Replace the algorithm specific calls to generate keys in "req" with the
2405 *) Update PKCS#7 enveloped data routines to use new API. This is now
2406 supported by any public key method supporting the encrypt operation. A
2407 ctrl is added to allow the public key algorithm to examine or modify
2408 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2412 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2413 a default digest type to use. In most cases this will be SHA1 but some
2414 algorithms (such as GOST) need to specify an alternative digest. The
2415 return value indicates how strong the preference is 1 means optional and
2416 2 is mandatory (that is it is the only supported type). Modify
2417 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2418 use the default md. Update openssl utilities to use the default digest
2419 type for signing if it is not explicitly indicated.
2422 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2423 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2424 signing method from the key type. This effectively removes the link
2425 between digests and public key types.
2428 *) Add an OID cross reference table and utility functions. Its purpose is to
2429 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2430 rsaEncryption. This will allow some of the algorithm specific hackery
2431 needed to use the correct OID to be removed.
2434 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2435 structures for PKCS7_sign(). They are now set up by the relevant public
2439 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2442 *) Add support for key derivation (agreement) in the API, DH method and
2446 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2447 public and private key formats. As a side effect these add additional
2448 command line functionality not previously available: DSA signatures can be
2449 generated and verified using pkeyutl and DH key support and generation in
2454 [Oliver Tappe <zooey@hirschkaefer.de>]
2456 *) New make target "install_html_docs" installs HTML renditions of the
2458 [Oliver Tappe <zooey@hirschkaefer.de>]
2460 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2461 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2462 support key and parameter generation and add initial key generation
2463 functionality for RSA.
2466 *) Add functions for main EVP_PKEY_method operations. The undocumented
2467 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2468 EVP_PKEY_{encrypt,decrypt}_old.
2471 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2472 key API, doesn't do much yet.
2475 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2476 public key algorithms. New option to openssl utility:
2477 "list-public-key-algorithms" to print out info.
2480 *) Implement the Supported Elliptic Curves Extension for
2481 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2484 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2485 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2488 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2489 utilities such as rsa, dsa, dsaparam etc except they process any key
2493 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2494 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2495 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2499 *) Initial support for pluggable public key ASN1.
2500 De-spaghettify the public key ASN1 handling. Move public and private
2501 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2502 algorithm specific handling to a single module within the relevant
2503 algorithm directory. Add functions to allow (near) opaque processing
2504 of public and private key structures.
2507 *) Implement the Supported Point Formats Extension for
2508 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2511 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2512 for the psk identity [hint] and the psk callback functions to the
2513 SSL_SESSION, SSL and SSL_CTX structure.
2516 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2520 SSL_CTX_use_psk_identity_hint
2521 SSL_get_psk_identity_hint
2522 SSL_get_psk_identity
2523 SSL_use_psk_identity_hint
2525 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2527 *) Add RFC 3161 compliant time stamp request creation, response generation
2528 and response verification functionality.
2529 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2531 *) Add initial support for TLS extensions, specifically for the server_name
2532 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2533 have new members for a host name. The SSL data structure has an
2534 additional member SSL_CTX *initial_ctx so that new sessions can be
2535 stored in that context to allow for session resumption, even after the
2536 SSL has been switched to a new SSL_CTX in reaction to a client's
2537 server_name extension.
2539 New functions (subject to change):
2541 SSL_get_servername()
2542 SSL_get_servername_type()
2545 New CTRL codes and macros (subject to change):
2547 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2548 - SSL_CTX_set_tlsext_servername_callback()
2549 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2550 - SSL_CTX_set_tlsext_servername_arg()
2551 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2553 openssl s_client has a new '-servername ...' option.
2555 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2556 '-key2 ...', '-servername_fatal' (subject to change). This allows
2557 testing the HostName extension for a specific single host name ('-cert'
2558 and '-key' remain fallbacks for handshakes without HostName
2559 negotiation). If the unrecognized_name alert has to be sent, this by
2560 default is a warning; it becomes fatal with the '-servername_fatal'
2563 [Peter Sylvester, Remy Allais, Christophe Renou]
2565 *) Whirlpool hash implementation is added.
2568 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2569 bn(64,32). Because of instruction set limitations it doesn't have
2570 any negative impact on performance. This was done mostly in order
2571 to make it possible to share assembler modules, such as bn_mul_mont
2572 implementations, between 32- and 64-bit builds without hassle.
2575 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2576 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2580 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2581 dedicated Montgomery multiplication procedure, is introduced.
2582 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2583 "64-bit" performance on certain 32-bit targets.
2586 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2587 in SSL structures. New SSL ctrl to set maximum send fragment size.
2588 Save memory by seeting the I/O buffer sizes dynamically instead of
2589 using the maximum available value.
2592 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2593 in addition to the text details.
2596 *) Very, very preliminary EXPERIMENTAL support for printing of general
2597 ASN1 structures. This currently produces rather ugly output and doesn't
2598 handle several customised structures at all.
2601 *) Integrated support for PVK file format and some related formats such
2602 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2603 these in the 'rsa' and 'dsa' utilities.
2606 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2609 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2610 place for the (very old) "NETSCAPE" format certificates which are now
2611 handled using new ASN1 code equivalents.
2614 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2615 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2616 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2619 *) Modify CRL distribution points extension code to print out previously
2620 unsupported fields. Enhance extension setting code to allow setting of
2624 *) Add print and set support for Issuing Distribution Point CRL extension.
2627 *) Change 'Configure' script to enable Camellia by default.
2630 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2632 *) When rejecting SSL/TLS records due to an incorrect version number, never
2633 update s->server with a new major version number. As of
2634 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2635 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2636 the previous behavior could result in a read attempt at NULL when
2637 receiving specific incorrect SSL/TLS records once record payload
2638 protection is active. (CVE-2010-0740)
2639 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2641 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2642 could be crashed if the relevant tables were not present (e.g. chrooted).
2643 [Tomas Hoger <thoger@redhat.com>]
2645 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2647 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2648 [Martin Olsson, Neel Mehta]
2650 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2651 accommodate for stack sorting, always a write lock!).
2654 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2655 excessive delays in the RAND_poll(): over a minute. As a workaround
2656 include a time check in the inner Heap32Next loop too.
2659 *) The code that handled flushing of data in SSL/TLS originally used the
2660 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2661 the problem outlined in PR#1949. The fix suggested there however can
2662 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2663 of Apache). So instead simplify the code to flush unconditionally.
2664 This should be fine since flushing with no data to flush is a no op.
2667 *) Handle TLS versions 2.0 and later properly and correctly use the
2668 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2669 off ancient servers have a habit of sticking around for a while...
2672 *) Modify compression code so it frees up structures without using the
2673 ex_data callbacks. This works around a problem where some applications
2674 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2675 restarting) then use compression (e.g. SSL with compression) later.
2676 This results in significant per-connection memory leaks and
2677 has caused some security issues including CVE-2008-1678 and
2681 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2682 change when encrypting or decrypting.
2685 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2686 connect and renegotiate with servers which do not support RI.
2687 Until RI is more widely deployed this option is enabled by default.
2690 *) Add "missing" ssl ctrls to clear options and mode.
2693 *) If client attempts to renegotiate and doesn't support RI respond with
2694 a no_renegotiation alert as required by RFC5746. Some renegotiating
2695 TLS clients will continue a connection gracefully when they receive
2696 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2697 waiting for a server hello which it will never receive. Now we treat a
2698 received no_renegotiation alert as a fatal error. This is because
2699 applications requesting a renegotiation might well expect it to succeed
2700 and would have no code in place to handle the server denying it so the
2701 only safe thing to do is to terminate the connection.
2704 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2705 peer supports secure renegotiation and 0 otherwise. Print out peer
2706 renegotiation support in s_client/s_server.
2709 *) Replace the highly broken and deprecated SPKAC certification method with
2710 the updated NID creation version. This should correctly handle UTF8.
2713 *) Implement RFC5746. Re-enable renegotiation but require the extension
2714 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2715 turns out to be a bad idea. It has been replaced by
2716 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2717 SSL_CTX_set_options(). This is really not recommended unless you
2718 know what you are doing.
2719 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2721 *) Fixes to stateless session resumption handling. Use initial_ctx when
2722 issuing and attempting to decrypt tickets in case it has changed during
2723 servername handling. Use a non-zero length session ID when attempting
2724 stateless session resumption: this makes it possible to determine if
2725 a resumption has occurred immediately after receiving server hello
2726 (several places in OpenSSL subtly assume this) instead of later in
2730 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2731 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2732 fixes for a few places where the return code is not checked
2734 [Julia Lawall <julia@diku.dk>]
2736 *) Add --strict-warnings option to Configure script to include devteam
2737 warnings in other configurations.
2740 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2741 makes it possible to install openssl libraries in locations which
2742 have names other than "lib", for example "/usr/lib64" which some
2744 [Steve Henson, based on patch from Jeremy Utley]
2746 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2747 X690 8.9.12 and can produce some misleading textual output of OIDs.
2748 [Steve Henson, reported by Dan Kaminsky]
2750 *) Delete MD2 from algorithm tables. This follows the recommendation in
2751 several standards that it is not used in new applications due to
2752 several cryptographic weaknesses. For binary compatibility reasons
2753 the MD2 API is still compiled in by default.
2756 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2760 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2761 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2763 [Guenter <lists@gknw.net>]
2765 *) Fix the server certificate chain building code to use X509_verify_cert(),
2766 it used to have an ad-hoc builder which was unable to cope with anything
2767 other than a simple chain.
2768 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2770 *) Don't check self signed certificate signatures in X509_verify_cert()
2771 by default (a flag can override this): it just wastes time without
2772 adding any security. As a useful side effect self signed root CAs
2773 with non-FIPS digests are now usable in FIPS mode.
2776 *) In dtls1_process_out_of_seq_message() the check if the current message
2777 is already buffered was missing. For every new message was memory
2778 allocated, allowing an attacker to perform an denial of service attack
2779 with sending out of seq handshake messages until there is no memory
2780 left. Additionally every future messege was buffered, even if the
2781 sequence number made no sense and would be part of another handshake.
2782 So only messages with sequence numbers less than 10 in advance will be
2783 buffered. (CVE-2009-1378)
2784 [Robin Seggelmann, discovered by Daniel Mentz]
2786 *) Records are buffered if they arrive with a future epoch to be
2787 processed after finishing the corresponding handshake. There is
2788 currently no limitation to this buffer allowing an attacker to perform
2789 a DOS attack with sending records with future epochs until there is no
2790 memory left. This patch adds the pqueue_size() function to determine
2791 the size of a buffer and limits the record buffer to 100 entries.
2793 [Robin Seggelmann, discovered by Daniel Mentz]
2795 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2796 parent structure is freed. (CVE-2009-1379)
2799 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2800 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2803 [Ilya O. <vrghost@gmail.com>]
2805 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2807 *) Disable renegotiation completely - this fixes a severe security
2808 problem (CVE-2009-3555) at the cost of breaking all
2809 renegotiation. Renegotiation can be re-enabled by setting
2810 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2811 run-time. This is really not recommended unless you know what
2815 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2817 *) Don't set val to NULL when freeing up structures, it is freed up by
2818 underlying code. If sizeof(void *) > sizeof(long) this can result in
2819 zeroing past the valid field. (CVE-2009-0789)
2820 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2822 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2823 checked correctly. This would allow some invalid signed attributes to
2824 appear to verify correctly. (CVE-2009-0591)
2825 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2827 *) Reject UniversalString and BMPString types with invalid lengths. This
2828 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2829 a legal length. (CVE-2009-0590)
2832 *) Set S/MIME signing as the default purpose rather than setting it
2833 unconditionally. This allows applications to override it at the store
2837 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2838 to handle some structures.
2841 *) Improve efficiency of mem_gets: don't search whole buffer each time
2843 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2845 *) New -hex option for openssl rand.
2848 *) Print out UTF8String and NumericString when parsing ASN1.
2851 *) Support NumericString type for name components.
2854 *) Allow CC in the environment to override the automatically chosen
2855 compiler. Note that nothing is done to ensure flags work with the
2859 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2861 *) Properly check EVP_VerifyFinal() and similar return values
2863 [Ben Laurie, Bodo Moeller, Google Security Team]
2865 *) Enable TLS extensions by default.
2868 *) Allow the CHIL engine to be loaded, whether the application is
2869 multithreaded or not. (This does not release the developer from the
2870 obligation to set up the dynamic locking callbacks.)
2871 [Sander Temme <sander@temme.net>]
2873 *) Use correct exit code if there is an error in dgst command.
2874 [Steve Henson; problem pointed out by Roland Dirlewanger]
2876 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2877 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2880 *) Add experimental JPAKE support, including demo authentication in
2881 s_client and s_server.
2884 *) Set the comparison function in v3_addr_canonize().
2885 [Rob Austein <sra@hactrn.net>]
2887 *) Add support for XMPP STARTTLS in s_client.
2888 [Philip Paeps <philip@freebsd.org>]
2890 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2891 to ensure that even with this option, only ciphersuites in the
2892 server's preference list will be accepted. (Note that the option
2893 applies only when resuming a session, so the earlier behavior was
2894 just about the algorithm choice for symmetric cryptography.)
2897 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2899 *) Fix NULL pointer dereference if a DTLS server received
2900 ChangeCipherSpec as first record (CVE-2009-1386).
2903 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2904 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2907 *) The fix in 0.9.8c that supposedly got rid of unsafe
2908 double-checked locking was incomplete for RSA blinding,
2909 addressing just one layer of what turns out to have been
2910 doubly unsafe triple-checked locking.
2912 So now fix this for real by retiring the MONT_HELPER macro
2913 in crypto/rsa/rsa_eay.c.
2915 [Bodo Moeller; problem pointed out by Marius Schilder]
2917 *) Various precautionary measures:
2919 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2921 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2922 (NB: This would require knowledge of the secret session ticket key
2923 to exploit, in which case you'd be SOL either way.)
2925 - Change bn_nist.c so that it will properly handle input BIGNUMs
2926 outside the expected range.
2928 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2931 [Neel Mehta, Bodo Moeller]
2933 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2934 the load fails. Useful for distros.
2935 [Ben Laurie and the FreeBSD team]
2937 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2940 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2943 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2945 This work was sponsored by Logica.
2948 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2949 keystores. Support for SSL/TLS client authentication too.
2950 Not compiled unless enable-capieng specified to Configure.
2952 This work was sponsored by Logica.
2955 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2956 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2957 attribute creation routines such as certifcate requests and PKCS#12
2961 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2963 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2964 handshake which could lead to a cilent crash as found using the
2965 Codenomicon TLS test suite (CVE-2008-1672)
2966 [Steve Henson, Mark Cox]
2968 *) Fix double free in TLS server name extensions which could lead to
2969 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2972 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2974 Clear the error queue to ensure that error entries left from
2975 older function calls do not interfere with the correct operation.
2976 [Lutz Jaenicke, Erik de Castro Lopo]
2978 *) Remove root CA certificates of commercial CAs:
2980 The OpenSSL project does not recommend any specific CA and does not
2981 have any policy with respect to including or excluding any CA.
2982 Therefore it does not make any sense to ship an arbitrary selection
2983 of root CA certificates with the OpenSSL software.
2986 *) RSA OAEP patches to fix two separate invalid memory reads.
2987 The first one involves inputs when 'lzero' is greater than
2988 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2989 before the beginning of from). The second one involves inputs where
2990 the 'db' section contains nothing but zeroes (there is a one-byte
2991 invalid read after the end of 'db').
2992 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2994 *) Partial backport from 0.9.9-dev:
2996 Introduce bn_mul_mont (dedicated Montgomery multiplication
2997 procedure) as a candidate for BIGNUM assembler implementation.
2998 While 0.9.9-dev uses assembler for various architectures, only
2999 x86_64 is available by default here in the 0.9.8 branch, and
3000 32-bit x86 is available through a compile-time setting.
3002 To try the 32-bit x86 assembler implementation, use Configure
3003 option "enable-montasm" (which exists only for this backport).
3005 As "enable-montasm" for 32-bit x86 disclaims code stability
3006 anyway, in this constellation we activate additional code
3007 backported from 0.9.9-dev for further performance improvements,
3008 namely BN_from_montgomery_word. (To enable this otherwise,
3009 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3011 [Andy Polyakov (backport partially by Bodo Moeller)]
3013 *) Add TLS session ticket callback. This allows an application to set
3014 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3015 values. This is useful for key rollover for example where several key
3016 sets may exist with different names.
3019 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3020 This was broken until now in 0.9.8 releases, such that the only way
3021 a registered ENGINE could be used (assuming it initialises
3022 successfully on the host) was to explicitly set it as the default
3023 for the relevant algorithms. This is in contradiction with 0.9.7
3024 behaviour and the documentation. With this fix, when an ENGINE is
3025 registered into a given algorithm's table of implementations, the
3026 'uptodate' flag is reset so that auto-discovery will be used next
3027 time a new context for that algorithm attempts to select an
3029 [Ian Lister (tweaked by Geoff Thorpe)]
3031 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3032 implemention in the following ways:
3034 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3037 Lack of BER streaming support means one pass streaming processing is
3038 only supported if data is detached: setting the streaming flag is
3039 ignored for embedded content.
3041 CMS support is disabled by default and must be explicitly enabled
3042 with the enable-cms configuration option.
3045 *) Update the GMP engine glue to do direct copies between BIGNUM and
3046 mpz_t when openssl and GMP use the same limb size. Otherwise the
3047 existing "conversion via a text string export" trick is still used.
3048 [Paul Sheer <paulsheer@gmail.com>]
3050 *) Zlib compression BIO. This is a filter BIO which compressed and
3051 uncompresses any data passed through it.
3054 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3055 RFC3394 compatible AES key wrapping.
3058 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3059 sets string data without copying. X509_ALGOR_set0() and
3060 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3061 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3062 from an X509_ATTRIBUTE structure optionally checking it occurs only
3063 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3067 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3068 to get the expected BN_FLG_CONSTTIME behavior.
3069 [Bodo Moeller (Google)]
3073 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3074 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3075 - added some more tests to do_tests.pl
3076 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3077 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3078 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3079 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3080 - various changes to netware.pl to enable gcc-cross builds on Win32
3082 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3083 - various changes to fix missing prototype warnings
3084 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3085 - added AES, WHIRLPOOL and CPUID assembler code to build files
3086 - added missing AES assembler make rules to mk1mf.pl
3087 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3088 [Guenter Knauf <eflash@gmx.net>]
3090 *) Implement certificate status request TLS extension defined in RFC3546.
3091 A client can set the appropriate parameters and receive the encoded
3092 OCSP response via a callback. A server can query the supplied parameters
3093 and set the encoded OCSP response in the callback. Add simplified examples
3094 to s_client and s_server.
3097 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3099 *) Fix various bugs:
3100 + Binary incompatibility of ssl_ctx_st structure
3101 + DTLS interoperation with non-compliant servers
3102 + Don't call get_session_cb() without proposed session
3103 + Fix ia64 assembler code
3104 [Andy Polyakov, Steve Henson]
3106 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3108 *) DTLS Handshake overhaul. There were longstanding issues with
3109 OpenSSL DTLS implementation, which were making it impossible for
3110 RFC 4347 compliant client to communicate with OpenSSL server.
3111 Unfortunately just fixing these incompatibilities would "cut off"
3112 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3113 server keeps tolerating non RFC compliant syntax. The opposite is
3114 not true, 0.9.8f client can not communicate with earlier server.
3115 This update even addresses CVE-2007-4995.
3118 *) Changes to avoid need for function casts in OpenSSL: some compilers
3119 (gcc 4.2 and later) reject their use.
3120 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3123 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3124 RFC4507bis. The encrypted ticket format is an encrypted encoded
3125 SSL_SESSION structure, that way new session features are automatically
3128 If a client application caches session in an SSL_SESSION structure
3129 support is transparent because tickets are now stored in the encoded
3132 The SSL_CTX structure automatically generates keys for ticket
3133 protection in servers so again support should be possible
3134 with no application modification.
3136 If a client or server wishes to disable RFC4507 support then the option
3137 SSL_OP_NO_TICKET can be set.
3139 Add a TLS extension debugging callback to allow the contents of any client
3140 or server extensions to be examined.
3142 This work was sponsored by Google.
3145 *) Add initial support for TLS extensions, specifically for the server_name
3146 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3147 have new members for a host name. The SSL data structure has an
3148 additional member SSL_CTX *initial_ctx so that new sessions can be
3149 stored in that context to allow for session resumption, even after the
3150 SSL has been switched to a new SSL_CTX in reaction to a client's
3151 server_name extension.
3153 New functions (subject to change):
3155 SSL_get_servername()
3156 SSL_get_servername_type()
3159 New CTRL codes and macros (subject to change):
3161 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3162 - SSL_CTX_set_tlsext_servername_callback()
3163 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3164 - SSL_CTX_set_tlsext_servername_arg()
3165 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3167 openssl s_client has a new '-servername ...' option.
3169 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3170 '-key2 ...', '-servername_fatal' (subject to change). This allows
3171 testing the HostName extension for a specific single host name ('-cert'
3172 and '-key' remain fallbacks for handshakes without HostName
3173 negotiation). If the unrecognized_name alert has to be sent, this by
3174 default is a warning; it becomes fatal with the '-servername_fatal'
3177 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3179 *) Add AES and SSE2 assembly language support to VC++ build.
3182 *) Mitigate attack on final subtraction in Montgomery reduction.
3185 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3186 (which previously caused an internal error).
3189 *) Squeeze another 10% out of IGE mode when in != out.
3192 *) AES IGE mode speedup.
3193 [Dean Gaudet (Google)]
3195 *) Add the Korean symmetric 128-bit cipher SEED (see
3196 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3197 add SEED ciphersuites from RFC 4162:
3199 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3200 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3201 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3202 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3204 To minimize changes between patchlevels in the OpenSSL 0.9.8
3205 series, SEED remains excluded from compilation unless OpenSSL
3206 is configured with 'enable-seed'.
3207 [KISA, Bodo Moeller]
3209 *) Mitigate branch prediction attacks, which can be practical if a
3210 single processor is shared, allowing a spy process to extract
3211 information. For detailed background information, see
3212 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3213 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3214 and Necessary Software Countermeasures"). The core of the change
3215 are new versions BN_div_no_branch() and
3216 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3217 respectively, which are slower, but avoid the security-relevant
3218 conditional branches. These are automatically called by BN_div()
3219 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3220 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3221 remove a conditional branch.
3223 BN_FLG_CONSTTIME is the new name for the previous
3224 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3225 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3226 in the exponent causes BN_mod_exp_mont() to use the alternative
3227 implementation in BN_mod_exp_mont_consttime().) The old name
3228 remains as a deprecated alias.
3230 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3231 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3232 constant-time implementations for more than just exponentiation.
3233 Here too the old name is kept as a deprecated alias.
3235 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3236 the BN_BLINDING structure gets an independent copy of the
3237 modulus. This means that the previous "BIGNUM *m" argument to
3238 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3239 essentially becomes "const BIGNUM *m", although we can't actually
3240 change this in the header file before 0.9.9. It allows
3241 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3242 enable BN_FLG_CONSTTIME.
3244 [Matthew D Wood (Intel Corp)]
3246 *) In the SSL/TLS server implementation, be strict about session ID
3247 context matching (which matters if an application uses a single
3248 external cache for different purposes). Previously,
3249 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3250 set. This did ensure strict client verification, but meant that,
3251 with applications using a single external cache for quite
3252 different requirements, clients could circumvent ciphersuite
3253 restrictions for a given session ID context by starting a session
3254 in a different context.
3257 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3258 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3259 authentication-only ciphersuites.
3262 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3263 not complete and could lead to a possible single byte overflow
3264 (CVE-2007-5135) [Ben Laurie]
3266 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3268 *) Since AES128 and AES256 (and similarly Camellia128 and
3269 Camellia256) share a single mask bit in the logic of
3270 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3271 kludge to work properly if AES128 is available and AES256 isn't
3272 (or if Camellia128 is available and Camellia256 isn't).
3275 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3276 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3277 When a point or a seed is encoded in a BIT STRING, we need to
3278 prevent the removal of trailing zero bits to get the proper DER
3279 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3280 of a NamedBitList, for which trailing 0 bits need to be removed.)
3283 *) Have SSL/TLS server implementation tolerate "mismatched" record
3284 protocol version while receiving ClientHello even if the
3285 ClientHello is fragmented. (The server can't insist on the
3286 particular protocol version it has chosen before the ServerHello
3287 message has informed the client about his choice.)
3290 *) Add RFC 3779 support.
3291 [Rob Austein for ARIN, Ben Laurie]
3293 *) Load error codes if they are not already present instead of using a
3294 static variable. This allows them to be cleanly unloaded and reloaded.
3295 Improve header file function name parsing.
3298 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3299 or CAPABILITY handshake as required by RFCs.
3302 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3304 *) Introduce limits to prevent malicious keys being able to
3305 cause a denial of service. (CVE-2006-2940)
3306 [Steve Henson, Bodo Moeller]
3308 *) Fix ASN.1 parsing of certain invalid structures that can result
3309 in a denial of service. (CVE-2006-2937) [Steve Henson]
3311 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3312 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3314 *) Fix SSL client code which could crash if connecting to a
3315 malicious SSLv2 server. (CVE-2006-4343)
3316 [Tavis Ormandy and Will Drewry, Google Security Team]
3318 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3319 match only those. Before that, "AES256-SHA" would be interpreted
3320 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3321 the same strength classification in 0.9.7h) as we currently only
3322 have a single AES bit in the ciphersuite description bitmap.
3323 That change, however, also applied to ciphersuite strings such as
3324 "RC4-MD5" that intentionally matched multiple ciphersuites --
3325 namely, SSL 2.0 ciphersuites in addition to the more common ones
3326 from SSL 3.0/TLS 1.0.
3328 So we change the selection algorithm again: Naming an explicit
3329 ciphersuite selects this one ciphersuite, and any other similar
3330 ciphersuite (same bitmap) from *other* protocol versions.
3331 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3332 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3334 Since SSL 2.0 does not have any ciphersuites for which the
3335 128/256 bit distinction would be relevant, this works for now.
3336 The proper fix will be to use different bits for AES128 and
3337 AES256, which would have avoided the problems from the beginning;
3338 however, bits are scarce, so we can only do this in a new release
3339 (not just a patchlevel) when we can change the SSL_CIPHER
3340 definition to split the single 'unsigned long mask' bitmap into
3341 multiple values to extend the available space.
3345 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3347 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3348 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3350 *) Add AES IGE and biIGE modes.
3353 *) Change the Unix randomness entropy gathering to use poll() when
3354 possible instead of select(), since the latter has some
3355 undesirable limitations.
3356 [Darryl Miles via Richard Levitte and Bodo Moeller]
3358 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3359 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3360 cannot be implicitly activated as part of, e.g., the "AES" alias.
3361 However, please upgrade to OpenSSL 0.9.9[-dev] for
3362 non-experimental use of the ECC ciphersuites to get TLS extension
3363 support, which is required for curve and point format negotiation
3364 to avoid potential handshake problems.
3367 *) Disable rogue ciphersuites:
3369 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3370 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3371 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3373 The latter two were purportedly from
3374 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3377 Also deactivate the remaining ciphersuites from
3378 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3379 unofficial, and the ID has long expired.
3382 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3383 dual-core machines) and other potential thread-safety issues.
3386 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3387 versions), which is now available for royalty-free use
3388 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3389 Also, add Camellia TLS ciphersuites from RFC 4132.
3391 To minimize changes between patchlevels in the OpenSSL 0.9.8
3392 series, Camellia remains excluded from compilation unless OpenSSL
3393 is configured with 'enable-camellia'.
3396 *) Disable the padding bug check when compression is in use. The padding
3397 bug check assumes the first packet is of even length, this is not
3398 necessarily true if compresssion is enabled and can result in false
3399 positives causing handshake failure. The actual bug test is ancient
3400 code so it is hoped that implementations will either have fixed it by
3401 now or any which still have the bug do not support compression.
3404 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3406 *) When applying a cipher rule check to see if string match is an explicit
3407 cipher suite and only match that one cipher suite if it is.
3410 *) Link in manifests for VC++ if needed.
3411 [Austin Ziegler <halostatue@gmail.com>]
3413 *) Update support for ECC-based TLS ciphersuites according to
3414 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3415 TLS extensions, which are supported starting with the 0.9.9
3416 branch, not in the OpenSSL 0.9.8 branch).
3419 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3420 opaque EVP_CIPHER_CTX handling.
3423 *) Fixes and enhancements to zlib compression code. We now only use
3424 "zlib1.dll" and use the default __cdecl calling convention on Win32
3425 to conform with the standards mentioned here:
3426 http://www.zlib.net/DLL_FAQ.txt
3427 Static zlib linking now works on Windows and the new --with-zlib-include
3428 --with-zlib-lib options to Configure can be used to supply the location
3429 of the headers and library. Gracefully handle case where zlib library
3433 *) Several fixes and enhancements to the OID generation code. The old code
3434 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3435 handle numbers larger than ULONG_MAX, truncated printing and had a
3436 non standard OBJ_obj2txt() behaviour.
3439 *) Add support for building of engines under engine/ as shared libraries
3440 under VC++ build system.
3443 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3444 Hopefully, we will not see any false combination of paths any more.
3447 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3449 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3450 (part of SSL_OP_ALL). This option used to disable the
3451 countermeasure against man-in-the-middle protocol-version
3452 rollback in the SSL 2.0 server implementation, which is a bad
3453 idea. (CVE-2005-2969)
3455 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3456 for Information Security, National Institute of Advanced Industrial
3457 Science and Technology [AIST], Japan)]
3459 *) Add two function to clear and return the verify parameter flags.
3462 *) Keep cipherlists sorted in the source instead of sorting them at
3463 runtime, thus removing the need for a lock.
3466 *) Avoid some small subgroup attacks in Diffie-Hellman.
3467 [Nick Mathewson and Ben Laurie]
3469 *) Add functions for well-known primes.
3472 *) Extended Windows CE support.
3473 [Satoshi Nakamura and Andy Polyakov]
3475 *) Initialize SSL_METHOD structures at compile time instead of during
3476 runtime, thus removing the need for a lock.
3479 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3480 attempting to decrypt each encrypted key in turn. Add support to
3484 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3486 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3489 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3492 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3493 key into the same file any more.
3496 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3499 *) Add -utf8 command line and config file option to 'ca'.
3500 [Stefan <stf@udoma.org]
3502 *) Removed the macro des_crypt(), as it seems to conflict with some
3503 libraries. Use DES_crypt().
3506 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3507 involves renaming the source and generated shared-libs for
3508 both. The engines will accept the corrected or legacy ids
3509 ('ncipher' and '4758_cca' respectively) when binding. NB,
3510 this only applies when building 'shared'.
3511 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3513 *) Add attribute functions to EVP_PKEY structure. Modify
3514 PKCS12_create() to recognize a CSP name attribute and
3515 use it. Make -CSP option work again in pkcs12 utility.
3518 *) Add new functionality to the bn blinding code:
3519 - automatic re-creation of the BN_BLINDING parameters after
3520 a fixed number of uses (currently 32)
3521 - add new function for parameter creation
3522 - introduce flags to control the update behaviour of the
3523 BN_BLINDING parameters
3524 - hide BN_BLINDING structure
3525 Add a second BN_BLINDING slot to the RSA structure to improve
3526 performance when a single RSA object is shared among several
3530 *) Add support for DTLS.
3531 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3533 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3534 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3537 *) Remove buggy and incomplete DH cert support from
3538 ssl/ssl_rsa.c and ssl/s3_both.c
3541 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3542 the apps/openssl applications.
3545 *) Compile clean with "-Wall -Wmissing-prototypes
3546 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3547 DEBUG_SAFESTACK must also be set.
3550 *) Change ./Configure so that certain algorithms can be disabled by default.
3551 The new counterpiece to "no-xxx" is "enable-xxx".
3553 The patented RC5 and MDC2 algorithms will now be disabled unless
3554 "enable-rc5" and "enable-mdc2", respectively, are specified.
3556 (IDEA remains enabled despite being patented. This is because IDEA
3557 is frequently required for interoperability, and there is no license
3558 fee for non-commercial use. As before, "no-idea" can be used to
3559 avoid this algorithm.)
3563 *) Add processing of proxy certificates (see RFC 3820). This work was
3564 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3565 EGEE (Enabling Grids for E-science in Europe).
3568 *) RC4 performance overhaul on modern architectures/implementations, such
3569 as Intel P4, IA-64 and AMD64.
3572 *) New utility extract-section.pl. This can be used specify an alternative
3573 section number in a pod file instead of having to treat each file as
3574 a separate case in Makefile. This can be done by adding two lines to the
3577 =for comment openssl_section:XXX
3579 The blank line is mandatory.
3583 *) New arguments -certform, -keyform and -pass for s_client and s_server
3584 to allow alternative format key and certificate files and passphrase
3588 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3589 update associated structures and add various utility functions.
3591 Add new policy related verify parameters, include policy checking in
3592 standard verify code. Enhance 'smime' application with extra parameters
3593 to support policy checking and print out.
3596 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3597 Nehemiah processors. These extensions support AES encryption in hardware
3598 as well as RNG (though RNG support is currently disabled).
3599 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3601 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3604 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3605 [Andy Polyakov and a number of other people]
3607 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3608 implementation contributed by IBM.
3609 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3611 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3612 exponent rather than 'unsigned long'. There is a corresponding change to
3613 the new 'rsa_keygen' element of the RSA_METHOD structure.
3614 [Jelte Jansen, Geoff Thorpe]
3616 *) Functionality for creating the initial serial number file is now
3617 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3619 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3620 number file to 1, which is bound to cause problems. To avoid
3621 the problems while respecting compatibility between different 0.9.7
3622 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3623 CA.pl for serial number initialization. With the new release 0.9.8,
3624 we can fix the problem directly in the 'ca' utility.)
3627 *) Reduced header interdepencies by declaring more opaque objects in
3628 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3629 give fewer recursive includes, which could break lazy source code - so
3630 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3631 developers should define this symbol when building and using openssl to
3632 ensure they track the recommended behaviour, interfaces, [etc], but
3633 backwards-compatible behaviour prevails when this isn't defined.
3636 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3639 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3640 This will generate a random key of the appropriate length based on the
3641 cipher context. The EVP_CIPHER can provide its own random key generation
3642 routine to support keys of a specific form. This is used in the des and
3643 3des routines to generate a key of the correct parity. Update S/MIME
3644 code to use new functions and hence generate correct parity DES keys.
3645 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3646 valid (weak or incorrect parity).
3649 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3650 as looking them up. This is useful when the verified structure may contain
3651 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3652 present unless the new PKCS7_NO_CRL flag is asserted.
3655 *) Extend ASN1 oid configuration module. It now additionally accepts the
3658 shortName = some long name, 1.2.3.4
3661 *) Reimplemented the BN_CTX implementation. There is now no more static
3662 limitation on the number of variables it can handle nor the depth of the
3663 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3664 information can now expand as required, and rather than having a single
3665 static array of bignums, BN_CTX now uses a linked-list of such arrays
3666 allowing it to expand on demand whilst maintaining the usefulness of
3667 BN_CTX's "bundling".
3670 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3671 to allow all RSA operations to function using a single BN_CTX.
3674 *) Preliminary support for certificate policy evaluation and checking. This
3675 is initially intended to pass the tests outlined in "Conformance Testing
3676 of Relying Party Client Certificate Path Processing Logic" v1.07.
3679 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3680 remained unused and not that useful. A variety of other little bignum
3681 tweaks and fixes have also been made continuing on from the audit (see
3685 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3686 associated ASN1, EVP and SSL functions and old ASN1 macros.
3689 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3690 and this should never fail. So the return value from the use of
3691 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3692 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3695 *) BN_CTX_get() should return zero-valued bignums, providing the same
3696 initialised value as BN_new().
3697 [Geoff Thorpe, suggested by Ulf Möller]
3699 *) Support for inhibitAnyPolicy certificate extension.
3702 *) An audit of the BIGNUM code is underway, for which debugging code is
3703 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3704 is considered valid when processing BIGNUMs, and causes execution to
3705 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3706 further steps are taken to deliberately pollute unused data in BIGNUM
3707 structures to try and expose faulty code further on. For now, openssl will
3708 (in its default mode of operation) continue to tolerate the inconsistent
3709 forms that it has tolerated in the past, but authors and packagers should
3710 consider trying openssl and their own applications when compiled with
3711 these debugging symbols defined. It will help highlight potential bugs in
3712 their own code, and will improve the test coverage for OpenSSL itself. At
3713 some point, these tighter rules will become openssl's default to improve
3714 maintainability, though the assert()s and other overheads will remain only
3715 in debugging configurations. See bn.h for more details.
3716 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3718 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3719 that can only be obtained through BN_CTX_new() (which implicitly
3720 initialises it). The presence of this function only made it possible
3721 to overwrite an existing structure (and cause memory leaks).
3724 *) Because of the callback-based approach for implementing LHASH as a
3725 template type, lh_insert() adds opaque objects to hash-tables and
3726 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3727 to clean up those corresponding objects before destroying the hash table
3728 (and losing the object pointers). So some over-zealous constifications in
3729 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3730 objects as "const" and the lh_doall[_arg] callback wrappers are not
3731 prototyped to have "const" restrictions on the object pointers they are
3732 given (and so aren't required to cast them away any more).
3735 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3736 (speed) prefers to use its own implementation. The two implementations
3737 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3738 its object type properly exposed (MS_TM) instead of casting to/from "char
3739 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3740 aren't necessarily the greatest nomenclatures - but this is what was used
3741 internally to the implementation so I've used that for now.
3744 *) Ensure that deprecated functions do not get compiled when
3745 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3746 the self-tests were still using deprecated key-generation functions so
3747 these have been updated also.
3750 *) Reorganise PKCS#7 code to separate the digest location functionality
3751 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3752 New function PKCS7_set_digest() to set the digest type for PKCS#7
3753 digestedData type. Add additional code to correctly generate the
3754 digestedData type and add support for this type in PKCS7 initialization
3758 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3759 structure of type "other".
3762 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3763 sure the loop does correctly stop and breaking ("division by zero")
3764 modulus operations are not performed. The (pre-generated) prime
3765 table crypto/bn/bn_prime.h was already correct, but it could not be
3766 re-generated on some platforms because of the "division by zero"
3767 situation in the script.
3768 [Ralf S. Engelschall]
3770 *) Update support for ECC-based TLS ciphersuites according to
3771 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3772 SHA-1 now is only used for "small" curves (where the
3773 representation of a field element takes up to 24 bytes); for
3774 larger curves, the field element resulting from ECDH is directly
3775 used as premaster secret.
3776 [Douglas Stebila (Sun Microsystems Laboratories)]
3778 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3779 curve secp160r1 to the tests.
3780 [Douglas Stebila (Sun Microsystems Laboratories)]
3782 *) Add the possibility to load symbols globally with DSO.
3783 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3785 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3786 control of the error stack.
3789 *) Add support for STORE in ENGINE.
3792 *) Add the STORE type. The intention is to provide a common interface
3793 to certificate and key stores, be they simple file-based stores, or
3794 HSM-type store, or LDAP stores, or...
3795 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3798 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3799 pass a list of arguments to any function as well as provide a way
3800 for a function to pass data back to the caller.
3803 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3804 works like BUF_strdup() but can be used to duplicate a portion of
3805 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3809 *) Add the function sk_find_ex() which works like sk_find(), but will
3810 return an index to an element even if an exact match couldn't be
3811 found. The index is guaranteed to point at the element where the
3812 searched-for key would be inserted to preserve sorting order.
3815 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3816 takes an extra flags argument for optional functionality. Currently,
3817 the following flags are defined:
3819 OBJ_BSEARCH_VALUE_ON_NOMATCH
3820 This one gets OBJ_bsearch_ex() to return a pointer to the first
3821 element where the comparing function returns a negative or zero
3824 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3825 This one gets OBJ_bsearch_ex() to return a pointer to the first
3826 element where the comparing function returns zero. This is useful
3827 if there are more than one element where the comparing function
3831 *) Make it possible to create self-signed certificates with 'openssl ca'
3832 in such a way that the self-signed certificate becomes part of the
3833 CA database and uses the same mechanisms for serial number generation
3834 as all other certificate signing. The new flag '-selfsign' enables
3835 this functionality. Adapt CA.sh and CA.pl.in.
3838 *) Add functionality to check the public key of a certificate request
3839 against a given private. This is useful to check that a certificate
3840 request can be signed by that key (self-signing).
3843 *) Make it possible to have multiple active certificates with the same
3844 subject in the CA index file. This is done only if the keyword
3845 'unique_subject' is set to 'no' in the main CA section (default
3846 if 'CA_default') of the configuration file. The value is saved
3847 with the database itself in a separate index attribute file,
3848 named like the index file with '.attr' appended to the name.
3851 *) Generate muti valued AVAs using '+' notation in config files for
3855 *) Support for nameConstraints certificate extension.
3858 *) Support for policyConstraints certificate extension.
3861 *) Support for policyMappings certificate extension.
3864 *) Make sure the default DSA_METHOD implementation only uses its
3865 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3866 and change its own handlers to be NULL so as to remove unnecessary
3867 indirection. This lets alternative implementations fallback to the
3868 default implementation more easily.
3871 *) Support for directoryName in GeneralName related extensions
3875 *) Make it possible to link applications using Makefile.shared.
3876 Make that possible even when linking against static libraries!
3879 *) Support for single pass processing for S/MIME signing. This now
3880 means that S/MIME signing can be done from a pipe, in addition
3881 cleartext signing (multipart/signed type) is effectively streaming
3882 and the signed data does not need to be all held in memory.
3884 This is done with a new flag PKCS7_STREAM. When this flag is set
3885 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3886 is done after the data is output (and digests calculated) in
3887 SMIME_write_PKCS7().
3890 *) Add full support for -rpath/-R, both in shared libraries and
3891 applications, at least on the platforms where it's known how
3895 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3896 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3897 will now compute a table of multiples of the generator that
3898 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3899 faster (notably in the case of a single point multiplication,
3900 scalar * generator).
3901 [Nils Larsch, Bodo Moeller]
3903 *) IPv6 support for certificate extensions. The various extensions
3904 which use the IP:a.b.c.d can now take IPv6 addresses using the
3905 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3909 *) Added an ENGINE that implements RSA by performing private key
3910 exponentiations with the GMP library. The conversions to and from
3911 GMP's mpz_t format aren't optimised nor are any montgomery forms
3912 cached, and on x86 it appears OpenSSL's own performance has caught up.
3913 However there are likely to be other architectures where GMP could
3914 provide a boost. This ENGINE is not built in by default, but it can be
3915 specified at Configure time and should be accompanied by the necessary
3916 linker additions, eg;
3917 ./config -DOPENSSL_USE_GMP -lgmp
3920 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3921 testing availability of engines with "-t" - the old behaviour is
3922 produced by increasing the feature's verbosity with "-tt".
3925 *) ECDSA routines: under certain error conditions uninitialized BN objects
3926 could be freed. Solution: make sure initialization is performed early
3927 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3931 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3932 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3933 software implementations. For DSA and DH, parameter generation can
3934 also be overriden by providing the appropriate method callbacks.
3937 *) Change the "progress" mechanism used in key-generation and
3938 primality testing to functions that take a new BN_GENCB pointer in
3939 place of callback/argument pairs. The new API functions have "_ex"
3940 postfixes and the older functions are reimplemented as wrappers for
3941 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3942 declarations of the old functions to help (graceful) attempts to
3943 migrate to the new functions. Also, the new key-generation API
3944 functions operate on a caller-supplied key-structure and return
3945 success/failure rather than returning a key or NULL - this is to
3946 help make "keygen" another member function of RSA_METHOD etc.
3948 Example for using the new callback interface:
3950 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3954 BN_GENCB_set(&my_cb, my_callback, my_arg);
3956 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3957 /* For the meaning of a, b in calls to my_callback(), see the
3958 * documentation of the function that calls the callback.
3959 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3960 * my_callback should return 1 if it wants BN_is_prime_ex()
3961 * to continue, or 0 to stop.
3966 *) Change the ZLIB compression method to be stateful, and make it
3967 available to TLS with the number defined in
3968 draft-ietf-tls-compression-04.txt.
3971 *) Add the ASN.1 structures and functions for CertificatePair, which
3972 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3974 CertificatePair ::= SEQUENCE {
3975 forward [0] Certificate OPTIONAL,
3976 reverse [1] Certificate OPTIONAL,
3977 -- at least one of the pair shall be present -- }
3979 Also implement the PEM functions to read and write certificate
3980 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3982 This needed to be defined, mostly for the sake of the LDAP
3983 attribute crossCertificatePair, but may prove useful elsewhere as
3987 *) Make it possible to inhibit symlinking of shared libraries in
3988 Makefile.shared, for Cygwin's sake.
3991 *) Extend the BIGNUM API by creating a function
3992 void BN_set_negative(BIGNUM *a, int neg);
3993 and a macro that behave like
3994 int BN_is_negative(const BIGNUM *a);
3996 to avoid the need to access 'a->neg' directly in applications.
3999 *) Implement fast modular reduction for pseudo-Mersenne primes
4000 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4001 EC_GROUP_new_curve_GFp() will now automatically use this
4003 [Nils Larsch <nla@trustcenter.de>]
4005 *) Add new lock type (CRYPTO_LOCK_BN).
4008 *) Change the ENGINE framework to automatically load engines
4009 dynamically from specific directories unless they could be
4010 found to already be built in or loaded. Move all the
4011 current engines except for the cryptodev one to a new
4013 The engines in engines/ are built as shared libraries if
4014 the "shared" options was given to ./Configure or ./config.
4015 Otherwise, they are inserted in libcrypto.a.
4016 /usr/local/ssl/engines is the default directory for dynamic
4017 engines, but that can be overriden at configure time through
4018 the usual use of --prefix and/or --openssldir, and at run
4019 time with the environment variable OPENSSL_ENGINES.
4020 [Geoff Thorpe and Richard Levitte]
4022 *) Add Makefile.shared, a helper makefile to build shared
4023 libraries. Adapt Makefile.org.
4026 *) Add version info to Win32 DLLs.
4027 [Peter 'Luna' Runestig" <peter@runestig.com>]
4029 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4030 can be added using this API to created arbitrary PKCS#12
4031 files while avoiding the low level API.
4033 New options to PKCS12_create(), key or cert can be NULL and
4034 will then be omitted from the output file. The encryption
4035 algorithm NIDs can be set to -1 for no encryption, the mac
4036 iteration count can be set to 0 to omit the mac.
4038 Enhance pkcs12 utility by making the -nokeys and -nocerts
4039 options work when creating a PKCS#12 file. New option -nomac
4040 to omit the mac, NONE can be set for an encryption algorithm.
4041 New code is modified to use the enhanced PKCS12_create()
4042 instead of the low level API.
4045 *) Extend ASN1 encoder to support indefinite length constructed
4046 encoding. This can output sequences tags and octet strings in
4047 this form. Modify pk7_asn1.c to support indefinite length
4048 encoding. This is experimental and needs additional code to
4049 be useful, such as an ASN1 bio and some enhanced streaming
4052 Extend template encode functionality so that tagging is passed
4053 down to the template encoder.
4056 *) Let 'openssl req' fail if an argument to '-newkey' is not
4057 recognized instead of using RSA as a default.
4060 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4061 As these are not official, they are not included in "ALL";
4062 the "ECCdraft" ciphersuite group alias can be used to select them.
4063 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4065 *) Add ECDH engine support.
4066 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4068 *) Add ECDH in new directory crypto/ecdh/.
4069 [Douglas Stebila (Sun Microsystems Laboratories)]
4071 *) Let BN_rand_range() abort with an error after 100 iterations
4072 without success (which indicates a broken PRNG).
4075 *) Change BN_mod_sqrt() so that it verifies that the input value
4076 is really the square of the return value. (Previously,
4077 BN_mod_sqrt would show GIGO behaviour.)
4080 *) Add named elliptic curves over binary fields from X9.62, SECG,
4081 and WAP/WTLS; add OIDs that were still missing.
4083 [Sheueling Chang Shantz and Douglas Stebila
4084 (Sun Microsystems Laboratories)]
4086 *) Extend the EC library for elliptic curves over binary fields
4087 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4090 EC_GF2m_simple_method
4094 EC_GROUP_new_curve_GF2m
4095 EC_GROUP_set_curve_GF2m
4096 EC_GROUP_get_curve_GF2m
4097 EC_POINT_set_affine_coordinates_GF2m
4098 EC_POINT_get_affine_coordinates_GF2m
4099 EC_POINT_set_compressed_coordinates_GF2m
4101 Point compression for binary fields is disabled by default for
4102 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4105 As binary polynomials are represented as BIGNUMs, various members
4106 of the EC_GROUP and EC_POINT data structures can be shared
4107 between the implementations for prime fields and binary fields;
4108 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4109 are essentially identical to their ..._GFp counterparts.
4110 (For simplicity, the '..._GFp' prefix has been dropped from
4111 various internal method names.)
4113 An internal 'field_div' method (similar to 'field_mul' and
4114 'field_sqr') has been added; this is used only for binary fields.
4116 [Sheueling Chang Shantz and Douglas Stebila
4117 (Sun Microsystems Laboratories)]
4119 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4120 through methods ('mul', 'precompute_mult').
4122 The generic implementations (now internally called 'ec_wNAF_mul'
4123 and 'ec_wNAF_precomputed_mult') remain the default if these
4124 methods are undefined.
4126 [Sheueling Chang Shantz and Douglas Stebila
4127 (Sun Microsystems Laboratories)]
4129 *) New function EC_GROUP_get_degree, which is defined through
4130 EC_METHOD. For curves over prime fields, this returns the bit
4131 length of the modulus.
4133 [Sheueling Chang Shantz and Douglas Stebila
4134 (Sun Microsystems Laboratories)]
4136 *) New functions EC_GROUP_dup, EC_POINT_dup.
4137 (These simply call ..._new and ..._copy).
4139 [Sheueling Chang Shantz and Douglas Stebila
4140 (Sun Microsystems Laboratories)]
4142 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4143 Polynomials are represented as BIGNUMs (where the sign bit is not
4144 used) in the following functions [macros]:
4147 BN_GF2m_sub [= BN_GF2m_add]
4148 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4149 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4150 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4152 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4153 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4154 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4155 BN_GF2m_cmp [= BN_ucmp]
4157 (Note that only the 'mod' functions are actually for fields GF(2^m).
4158 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4160 For some functions, an the irreducible polynomial defining a
4161 field can be given as an 'unsigned int[]' with strictly
4162 decreasing elements giving the indices of those bits that are set;
4163 i.e., p[] represents the polynomial
4164 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4166 p[0] > p[1] > ... > p[k] = 0.
4167 This applies to the following functions:
4172 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4173 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4175 BN_GF2m_mod_sqrt_arr
4176 BN_GF2m_mod_solve_quad_arr
4180 Conversion can be performed by the following functions:
4185 bntest.c has additional tests for binary polynomial arithmetic.
4187 Two implementations for BN_GF2m_mod_div() are available.
4188 The default algorithm simply uses BN_GF2m_mod_inv() and
4189 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4190 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4191 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4193 [Sheueling Chang Shantz and Douglas Stebila
4194 (Sun Microsystems Laboratories)]
4196 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4197 functionality is disabled at compile-time.
4198 [Douglas Stebila <douglas.stebila@sun.com>]
4200 *) Change default behaviour of 'openssl asn1parse' so that more
4201 information is visible when viewing, e.g., a certificate:
4203 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4204 mode the content of non-printable OCTET STRINGs is output in a
4205 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4206 avoid the appearance of a printable string.
4207 [Nils Larsch <nla@trustcenter.de>]
4209 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4211 EC_GROUP_set_asn1_flag()
4212 EC_GROUP_get_asn1_flag()
4213 EC_GROUP_set_point_conversion_form()
4214 EC_GROUP_get_point_conversion_form()
4215 These control ASN1 encoding details:
4216 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4217 has been set to OPENSSL_EC_NAMED_CURVE.
4218 - Points are encoded in uncompressed form by default; options for
4219 asn1_for are as for point2oct, namely
4220 POINT_CONVERSION_COMPRESSED
4221 POINT_CONVERSION_UNCOMPRESSED
4222 POINT_CONVERSION_HYBRID
4224 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4227 EC_GROUP_get0_seed()
4228 EC_GROUP_get_seed_len()
4229 This is used only for ASN1 purposes (so far).
4230 [Nils Larsch <nla@trustcenter.de>]
4232 *) Add 'field_type' member to EC_METHOD, which holds the NID
4233 of the appropriate field type OID. The new function
4234 EC_METHOD_get_field_type() returns this value.
4235 [Nils Larsch <nla@trustcenter.de>]
4240 EC_POINT_point2hex()
4241 EC_POINT_hex2point()
4242 providing useful interfaces to EC_POINT_point2oct() and
4243 EC_POINT_oct2point().
4244 [Nils Larsch <nla@trustcenter.de>]
4246 *) Change internals of the EC library so that the functions
4247 EC_GROUP_set_generator()
4248 EC_GROUP_get_generator()
4249 EC_GROUP_get_order()
4250 EC_GROUP_get_cofactor()
4251 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4252 to methods, which would lead to unnecessary code duplication when
4253 adding different types of curves.
4254 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4256 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4257 arithmetic, and such that modified wNAFs are generated
4258 (which avoid length expansion in many cases).
4261 *) Add a function EC_GROUP_check_discriminant() (defined via
4262 EC_METHOD) that verifies that the curve discriminant is non-zero.
4264 Add a function EC_GROUP_check() that makes some sanity tests
4265 on a EC_GROUP, its generator and order. This includes
4266 EC_GROUP_check_discriminant().
4267 [Nils Larsch <nla@trustcenter.de>]
4269 *) Add ECDSA in new directory crypto/ecdsa/.
4271 Add applications 'openssl ecparam' and 'openssl ecdsa'
4272 (these are based on 'openssl dsaparam' and 'openssl dsa').
4274 ECDSA support is also included in various other files across the
4275 library. Most notably,
4276 - 'openssl req' now has a '-newkey ecdsa:file' option;
4277 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4278 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4279 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4280 them suitable for ECDSA where domain parameters must be
4281 extracted before the specific public key;
4282 - ECDSA engine support has been added.
4283 [Nils Larsch <nla@trustcenter.de>]
4285 *) Include some named elliptic curves, and add OIDs from X9.62,
4286 SECG, and WAP/WTLS. Each curve can be obtained from the new
4288 EC_GROUP_new_by_curve_name(),
4289 and the list of available named curves can be obtained with
4290 EC_get_builtin_curves().
4291 Also add a 'curve_name' member to EC_GROUP objects, which can be
4293 EC_GROUP_set_curve_name()
4294 EC_GROUP_get_curve_name()
4295 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4297 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4298 was actually never needed) and in BN_mul(). The removal in BN_mul()
4299 required a small change in bn_mul_part_recursive() and the addition
4300 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4301 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4302 bn_sub_words() and bn_add_words() except they take arrays with
4306 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4308 *) Cleanse PEM buffers before freeing them since they may contain
4310 [Benjamin Bennett <ben@psc.edu>]
4312 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4313 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4314 authentication-only ciphersuites.
4317 *) Since AES128 and AES256 share a single mask bit in the logic of
4318 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4319 kludge to work properly if AES128 is available and AES256 isn't.
4322 *) Expand security boundary to match 1.1.1 module.
4325 *) Remove redundant features: hash file source, editing of test vectors
4326 modify fipsld to use external fips_premain.c signature.
4329 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4330 run algorithm test programs.
4333 *) Make algorithm test programs more tolerant of whitespace.
4336 *) Have SSL/TLS server implementation tolerate "mismatched" record
4337 protocol version while receiving ClientHello even if the
4338 ClientHello is fragmented. (The server can't insist on the
4339 particular protocol version it has chosen before the ServerHello
4340 message has informed the client about his choice.)
4343 *) Load error codes if they are not already present instead of using a
4344 static variable. This allows them to be cleanly unloaded and reloaded.
4347 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4349 *) Introduce limits to prevent malicious keys being able to
4350 cause a denial of service. (CVE-2006-2940)
4351 [Steve Henson, Bodo Moeller]
4353 *) Fix ASN.1 parsing of certain invalid structures that can result
4354 in a denial of service. (CVE-2006-2937) [Steve Henson]
4356 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4357 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4359 *) Fix SSL client code which could crash if connecting to a
4360 malicious SSLv2 server. (CVE-2006-4343)
4361 [Tavis Ormandy and Will Drewry, Google Security Team]
4363 *) Change ciphersuite string processing so that an explicit
4364 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4365 will no longer include "AES128-SHA"), and any other similar
4366 ciphersuite (same bitmap) from *other* protocol versions (so that
4367 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4368 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4369 changes from 0.9.8b and 0.9.8d.
4372 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4374 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4375 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4377 *) Change the Unix randomness entropy gathering to use poll() when
4378 possible instead of select(), since the latter has some
4379 undesirable limitations.
4380 [Darryl Miles via Richard Levitte and Bodo Moeller]
4382 *) Disable rogue ciphersuites:
4384 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4385 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4386 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4388 The latter two were purportedly from
4389 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4392 Also deactive the remaining ciphersuites from
4393 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4394 unofficial, and the ID has long expired.
4397 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4398 dual-core machines) and other potential thread-safety issues.
4401 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4403 *) Adapt fipsld and the build system to link against the validated FIPS
4404 module in FIPS mode.
4407 *) Fixes for VC++ 2005 build under Windows.
4410 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4411 from a Windows bash shell such as MSYS. It is autodetected from the
4412 "config" script when run from a VC++ environment. Modify standard VC++
4413 build to use fipscanister.o from the GNU make build.
4416 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4418 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4419 The value now differs depending on if you build for FIPS or not.
4420 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4421 safely run with a non-FIPSed libcrypto, as it may crash because of
4422 the difference induced by this change.
4425 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4427 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4428 (part of SSL_OP_ALL). This option used to disable the
4429 countermeasure against man-in-the-middle protocol-version
4430 rollback in the SSL 2.0 server implementation, which is a bad
4431 idea. (CVE-2005-2969)
4433 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4434 for Information Security, National Institute of Advanced Industrial
4435 Science and Technology [AIST], Japan)]
4437 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4438 mainly for FIPS compliance and not fully integrated at this stage.
4441 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4442 the exponentiation using a fixed-length exponent. (Otherwise,
4443 the information leaked through timing could expose the secret key
4444 after many signatures; cf. Bleichenbacher's attack on DSA with
4448 *) Make a new fixed-window mod_exp implementation the default for
4449 RSA, DSA, and DH private-key operations so that the sequence of
4450 squares and multiplies and the memory access pattern are
4451 independent of the particular secret key. This will mitigate
4452 cache-timing and potential related attacks.
4454 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4455 and this is automatically used by BN_mod_exp_mont() if the new flag
4456 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4457 will use this BN flag for private exponents unless the flag
4458 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4459 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4461 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4463 *) Change the client implementation for SSLv23_method() and
4464 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4465 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4466 (Previously, the SSL 2.0 backwards compatible Client Hello
4467 message format would be used even with SSL_OP_NO_SSLv2.)
4470 *) Add support for smime-type MIME parameter in S/MIME messages which some
4474 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4475 a threadsafe manner. Modify rsa code to use new function and add calls
4476 to dsa and dh code (which had race conditions before).
4479 *) Include the fixed error library code in the C error file definitions
4480 instead of fixing them up at runtime. This keeps the error code
4481 structures constant.
4484 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4486 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4489 *) Fixes for newer kerberos headers. NB: the casts are needed because
4490 the 'length' field is signed on one version and unsigned on another
4491 with no (?) obvious way to tell the difference, without these VC++
4492 complains. Also the "definition" of FAR (blank) is no longer included
4493 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4494 some needed definitions.
4497 *) Undo Cygwin change.
4500 *) Added support for proxy certificates according to RFC 3820.
4501 Because they may be a security thread to unaware applications,
4502 they must be explicitely allowed in run-time. See
4503 docs/HOWTO/proxy_certificates.txt for further information.
4506 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4508 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4509 server and client random values. Previously
4510 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4511 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4513 This change has negligible security impact because:
4515 1. Server and client random values still have 24 bytes of pseudo random
4518 2. Server and client random values are sent in the clear in the initial
4521 3. The master secret is derived using the premaster secret (48 bytes in
4522 size for static RSA ciphersuites) as well as client server and random
4525 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4528 [Stephen Henson, reported by UK NISCC]
4530 *) Use Windows randomness collection on Cygwin.
4533 *) Fix hang in EGD/PRNGD query when communication socket is closed
4534 prematurely by EGD/PRNGD.
4535 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4537 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4540 *) Back-port of selected performance improvements from development
4541 branch, as well as improved support for PowerPC platforms.
4544 *) Add lots of checks for memory allocation failure, error codes to indicate
4545 failure and freeing up memory if a failure occurs.
4546 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4548 *) Add new -passin argument to dgst.
4551 *) Perform some character comparisons of different types in X509_NAME_cmp:
4552 this is needed for some certificates that reencode DNs into UTF8Strings
4553 (in violation of RFC3280) and can't or wont issue name rollover
4557 *) Make an explicit check during certificate validation to see that
4558 the CA setting in each certificate on the chain is correct. As a
4559 side effect always do the following basic checks on extensions,
4560 not just when there's an associated purpose to the check:
4562 - if there is an unhandled critical extension (unless the user
4563 has chosen to ignore this fault)
4564 - if the path length has been exceeded (if one is set at all)
4565 - that certain extensions fit the associated purpose (if one has
4569 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4571 *) Avoid a race condition when CRLs are checked in a multi threaded
4572 environment. This would happen due to the reordering of the revoked
4573 entries during signature checking and serial number lookup. Now the
4574 encoding is cached and the serial number sort performed under a lock.
4575 Add new STACK function sk_is_sorted().
4578 *) Add Delta CRL to the extension code.
4581 *) Various fixes to s3_pkt.c so alerts are sent properly.
4582 [David Holmes <d.holmes@f5.com>]
4584 *) Reduce the chances of duplicate issuer name and serial numbers (in
4585 violation of RFC3280) using the OpenSSL certificate creation utilities.
4586 This is done by creating a random 64 bit value for the initial serial
4587 number when a serial number file is created or when a self signed
4588 certificate is created using 'openssl req -x509'. The initial serial
4589 number file is created using 'openssl x509 -next_serial' in CA.pl
4590 rather than being initialized to 1.
4593 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4595 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4596 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4597 [Joe Orton, Steve Henson]
4599 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4601 [Joe Orton, Steve Henson]
4603 *) Make it possible to have multiple active certificates with the same
4604 subject in the CA index file. This is done only if the keyword
4605 'unique_subject' is set to 'no' in the main CA section (default
4606 if 'CA_default') of the configuration file. The value is saved
4607 with the database itself in a separate index attribute file,
4608 named like the index file with '.attr' appended to the name.
4611 *) X509 verify fixes. Disable broken certificate workarounds when
4612 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4613 keyUsage extension present. Don't accept CRLs with unhandled critical
4614 extensions: since verify currently doesn't process CRL extensions this
4615 rejects a CRL with *any* critical extensions. Add new verify error codes
4619 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4620 A clarification of RFC2560 will require the use of OCTET STRINGs and
4621 some implementations cannot handle the current raw format. Since OpenSSL
4622 copies and compares OCSP nonces as opaque blobs without any attempt at
4623 parsing them this should not create any compatibility issues.
4626 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4627 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4628 this HMAC (and other) operations are several times slower than OpenSSL
4632 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4633 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4635 *) Use the correct content when signing type "other".
4638 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4640 *) Fix various bugs revealed by running the NISCC test suite:
4642 Stop out of bounds reads in the ASN1 code when presented with
4643 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4645 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4647 If verify callback ignores invalid public key errors don't try to check
4648 certificate signature with the NULL public key.
4652 *) New -ignore_err option in ocsp application to stop the server
4653 exiting on the first error in a request.
4656 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4657 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4661 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4662 extra data after the compression methods not only for TLS 1.0
4663 but also for SSL 3.0 (as required by the specification).
4664 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4666 *) Change X509_certificate_type() to mark the key as exported/exportable
4667 when it's 512 *bits* long, not 512 bytes.
4670 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4671 blocks during encryption.
4674 *) Various fixes to base64 BIO and non blocking I/O. On write
4675 flushes were not handled properly if the BIO retried. On read
4676 data was not being buffered properly and had various logic bugs.
4677 This also affects blocking I/O when the data being decoded is a
4681 *) Various S/MIME bugfixes and compatibility changes:
4682 output correct application/pkcs7 MIME type if
4683 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4684 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4685 of files as .eml work). Correctly handle very long lines in MIME
4689 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4691 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4692 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4693 a protocol version number mismatch like a decryption error
4694 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4697 *) Turn on RSA blinding by default in the default implementation
4698 to avoid a timing attack. Applications that don't want it can call
4699 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4700 They would be ill-advised to do so in most cases.
4701 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4703 *) Change RSA blinding code so that it works when the PRNG is not
4704 seeded (in this case, the secret RSA exponent is abused as
4705 an unpredictable seed -- if it is not unpredictable, there
4706 is no point in blinding anyway). Make RSA blinding thread-safe
4707 by remembering the creator's thread ID in rsa->blinding and
4708 having all other threads use local one-time blinding factors
4709 (this requires more computation than sharing rsa->blinding, but
4710 avoids excessive locking; and if an RSA object is not shared
4711 between threads, blinding will still be very fast).
4714 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4715 ENGINE as defaults for all supported algorithms irrespective of
4716 the 'flags' parameter. 'flags' is now honoured, so applications
4717 should make sure they are passing it correctly.
4720 *) Target "mingw" now allows native Windows code to be generated in
4721 the Cygwin environment as well as with the MinGW compiler.
4724 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4726 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4727 via timing by performing a MAC computation even if incorrect
4728 block cipher padding has been found. This is a countermeasure
4729 against active attacks where the attacker has to distinguish
4730 between bad padding and a MAC verification error. (CVE-2003-0078)
4732 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4733 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4734 Martin Vuagnoux (EPFL, Ilion)]
4736 *) Make the no-err option work as intended. The intention with no-err
4737 is not to have the whole error stack handling routines removed from
4738 libcrypto, it's only intended to remove all the function name and
4739 reason texts, thereby removing some of the footprint that may not
4740 be interesting if those errors aren't displayed anyway.
4742 NOTE: it's still possible for any application or module to have it's
4743 own set of error texts inserted. The routines are there, just not
4744 used by default when no-err is given.
4747 *) Add support for FreeBSD on IA64.
4748 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4750 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4751 Kerberos function mit_des_cbc_cksum(). Before this change,
4752 the value returned by DES_cbc_cksum() was like the one from
4753 mit_des_cbc_cksum(), except the bytes were swapped.
4754 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4756 *) Allow an application to disable the automatic SSL chain building.
4757 Before this a rather primitive chain build was always performed in
4758 ssl3_output_cert_chain(): an application had no way to send the
4759 correct chain if the automatic operation produced an incorrect result.
4761 Now the chain builder is disabled if either:
4763 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4765 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4767 The reasoning behind this is that an application would not want the
4768 auto chain building to take place if extra chain certificates are
4769 present and it might also want a means of sending no additional
4770 certificates (for example the chain has two certificates and the
4774 *) Add the possibility to build without the ENGINE framework.
4775 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4777 *) Under Win32 gmtime() can return NULL: check return value in
4778 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4781 *) DSA routines: under certain error conditions uninitialized BN objects
4782 could be freed. Solution: make sure initialization is performed early
4783 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4784 Nils Larsch <nla@trustcenter.de> via PR#459)
4787 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4788 checked on reconnect on the client side, therefore session resumption
4789 could still fail with a "ssl session id is different" error. This
4790 behaviour is masked when SSL_OP_ALL is used due to
4791 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4792 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4793 followup to PR #377.
4796 *) IA-32 assembler support enhancements: unified ELF targets, support
4797 for SCO/Caldera platforms, fix for Cygwin shared build.
4800 *) Add support for FreeBSD on sparc64. As a consequence, support for
4801 FreeBSD on non-x86 processors is separate from x86 processors on
4802 the config script, much like the NetBSD support.
4803 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4805 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4807 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4810 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4811 code (06) was taken as the first octet of the session ID and the last
4812 octet was ignored consequently. As a result SSLv2 client side session
4813 caching could not have worked due to the session ID mismatch between
4815 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4819 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4820 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4824 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4825 seems that in spite of existing for more than a year, many application
4826 author have done nothing to provide the necessary callbacks, which
4827 means that this particular engine will not work properly anywhere.
4828 This is a very unfortunate situation which forces us, in the name
4829 of usability, to give the hw_ncipher.c a static lock, which is part
4831 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4832 appear in 0.9.8 or later. We EXPECT application authors to have
4833 dealt properly with this when 0.9.8 is released (unless we actually
4834 make such changes in the libcrypto locking code that changes will
4835 have to be made anyway).
4838 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4839 octets have been read, EOF or an error occurs. Without this change
4840 some truncated ASN1 structures will not produce an error.
4843 *) Disable Heimdal support, since it hasn't been fully implemented.
4844 Still give the possibility to force the use of Heimdal, but with
4845 warnings and a request that patches get sent to openssl-dev.
4848 *) Add the VC-CE target, introduce the WINCE sysname, and add
4849 INSTALL.WCE and appropriate conditionals to make it build.
4850 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4852 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4853 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4854 edit numbers of the version.
4855 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4857 *) Introduce safe string copy and catenation functions
4858 (BUF_strlcpy() and BUF_strlcat()).
4859 [Ben Laurie (CHATS) and Richard Levitte]
4861 *) Avoid using fixed-size buffers for one-line DNs.
4862 [Ben Laurie (CHATS)]
4864 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4865 resizing buffers containing secrets, and use where appropriate.
4866 [Ben Laurie (CHATS)]
4868 *) Avoid using fixed size buffers for configuration file location.
4869 [Ben Laurie (CHATS)]
4871 *) Avoid filename truncation for various CA files.
4872 [Ben Laurie (CHATS)]
4874 *) Use sizeof in preference to magic numbers.
4875 [Ben Laurie (CHATS)]
4877 *) Avoid filename truncation in cert requests.
4878 [Ben Laurie (CHATS)]
4880 *) Add assertions to check for (supposedly impossible) buffer
4882 [Ben Laurie (CHATS)]
4884 *) Don't cache truncated DNS entries in the local cache (this could
4885 potentially lead to a spoofing attack).
4886 [Ben Laurie (CHATS)]
4888 *) Fix various buffers to be large enough for hex/decimal
4889 representations in a platform independent manner.
4890 [Ben Laurie (CHATS)]
4892 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4893 resizing buffers containing secrets, and use where appropriate.
4894 [Ben Laurie (CHATS)]
4896 *) Add BIO_indent() to avoid much slightly worrying code to do
4898 [Ben Laurie (CHATS)]
4900 *) Convert sprintf()/BIO_puts() to BIO_printf().
4901 [Ben Laurie (CHATS)]
4903 *) buffer_gets() could terminate with the buffer only half
4905 [Ben Laurie (CHATS)]
4907 *) Add assertions to prevent user-supplied crypto functions from
4908 overflowing internal buffers by having large block sizes, etc.
4909 [Ben Laurie (CHATS)]
4911 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4913 [Ben Laurie (CHATS)]
4915 *) Eliminate unused copy of key in RC4.
4916 [Ben Laurie (CHATS)]
4918 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4919 [Ben Laurie (CHATS)]
4921 *) Fix off-by-one error in EGD path.
4922 [Ben Laurie (CHATS)]
4924 *) If RANDFILE path is too long, ignore instead of truncating.
4925 [Ben Laurie (CHATS)]
4927 *) Eliminate unused and incorrectly sized X.509 structure
4929 [Ben Laurie (CHATS)]
4931 *) Eliminate unused and dangerous function knumber().
4932 [Ben Laurie (CHATS)]
4934 *) Eliminate unused and dangerous structure, KSSL_ERR.
4935 [Ben Laurie (CHATS)]
4937 *) Protect against overlong session ID context length in an encoded
4938 session object. Since these are local, this does not appear to be
4940 [Ben Laurie (CHATS)]
4942 *) Change from security patch (see 0.9.6e below) that did not affect
4943 the 0.9.6 release series:
4945 Remote buffer overflow in SSL3 protocol - an attacker could
4946 supply an oversized master key in Kerberos-enabled versions.
4948 [Ben Laurie (CHATS)]
4950 *) Change the SSL kerb5 codes to match RFC 2712.
4953 *) Make -nameopt work fully for req and add -reqopt switch.
4954 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4956 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4957 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4959 *) Make sure tests can be performed even if the corresponding algorithms
4960 have been removed entirely. This was also the last step to make
4961 OpenSSL compilable with DJGPP under all reasonable conditions.
4962 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4964 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4965 to allow version independent disabling of normally unselected ciphers,
4966 which may be activated as a side-effect of selecting a single cipher.
4968 (E.g., cipher list string "RSA" enables ciphersuites that are left
4969 out of "ALL" because they do not provide symmetric encryption.
4970 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4971 [Lutz Jaenicke, Bodo Moeller]
4973 *) Add appropriate support for separate platform-dependent build
4974 directories. The recommended way to make a platform-dependent
4975 build directory is the following (tested on Linux), maybe with
4978 # Place yourself outside of the OpenSSL source tree. In
4979 # this example, the environment variable OPENSSL_SOURCE
4980 # is assumed to contain the absolute OpenSSL source directory.
4981 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4982 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4983 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4984 mkdir -p `dirname $F`
4985 ln -s $OPENSSL_SOURCE/$F $F
4988 To be absolutely sure not to disturb the source tree, a "make clean"
4989 is a good thing. If it isn't successfull, don't worry about it,
4990 it probably means the source directory is very clean.
4993 *) Make sure any ENGINE control commands make local copies of string
4994 pointers passed to them whenever necessary. Otherwise it is possible
4995 the caller may have overwritten (or deallocated) the original string
4996 data when a later ENGINE operation tries to use the stored values.
4997 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4999 *) Improve diagnostics in file reading and command-line digests.
5000 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5002 *) Add AES modes CFB and OFB to the object database. Correct an
5003 error in AES-CFB decryption.
5006 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5007 allows existing EVP_CIPHER_CTX structures to be reused after
5008 calling EVP_*Final(). This behaviour is used by encryption
5009 BIOs and some applications. This has the side effect that
5010 applications must explicitly clean up cipher contexts with
5011 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5014 *) Check the values of dna and dnb in bn_mul_recursive before calling
5015 bn_mul_comba (a non zero value means the a or b arrays do not contain
5016 n2 elements) and fallback to bn_mul_normal if either is not zero.
5019 *) Fix escaping of non-ASCII characters when using the -subj option
5020 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5023 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5024 form for "surname", serialNumber has no short form.
5025 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5026 therefore remove "mail" short name for "internet 7".
5027 The OID for unique identifiers in X509 certificates is
5028 x500UniqueIdentifier, not uniqueIdentifier.
5029 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5032 *) Add an "init" command to the ENGINE config module and auto initialize
5033 ENGINEs. Without any "init" command the ENGINE will be initialized
5034 after all ctrl commands have been executed on it. If init=1 the
5035 ENGINE is initailized at that point (ctrls before that point are run
5036 on the uninitialized ENGINE and after on the initialized one). If
5037 init=0 then the ENGINE will not be iniatialized at all.
5040 *) Fix the 'app_verify_callback' interface so that the user-defined
5041 argument is actually passed to the callback: In the
5042 SSL_CTX_set_cert_verify_callback() prototype, the callback
5043 declaration has been changed from
5046 int (*cb)(X509_STORE_CTX *,void *);
5047 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5048 i=s->ctx->app_verify_callback(&ctx)
5049 has been changed into
5050 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5052 To update applications using SSL_CTX_set_cert_verify_callback(),
5053 a dummy argument can be added to their callback functions.
5054 [D. K. Smetters <smetters@parc.xerox.com>]
5056 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5057 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5059 *) Add and OPENSSL_LOAD_CONF define which will cause
5060 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5061 This allows older applications to transparently support certain
5062 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5063 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5064 load the config file and OPENSSL_add_all_algorithms_conf() which will
5065 always load it have also been added.
5068 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5069 Adjust NIDs and EVP layer.
5070 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5072 *) Config modules support in openssl utility.
5074 Most commands now load modules from the config file,
5075 though in a few (such as version) this isn't done
5076 because it couldn't be used for anything.
5078 In the case of ca and req the config file used is
5079 the same as the utility itself: that is the -config
5080 command line option can be used to specify an
5084 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5085 use "openssl_conf" if filename is NULL use default openssl config file.
5088 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5089 config section name. Add a new flag to tolerate a missing config file
5090 and move code to CONF_modules_load_file().
5093 *) Support for crypto accelerator cards from Accelerated Encryption
5094 Processing, www.aep.ie. (Use engine 'aep')
5095 The support was copied from 0.9.6c [engine] and adapted/corrected
5096 to work with the new engine framework.
5097 [AEP Inc. and Richard Levitte]
5099 *) Support for SureWare crypto accelerator cards from Baltimore
5100 Technologies. (Use engine 'sureware')
5101 The support was copied from 0.9.6c [engine] and adapted
5102 to work with the new engine framework.
5105 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5106 make the newer ENGINE framework commands for the CHIL engine work.
5107 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5109 *) Make it possible to produce shared libraries on ReliantUNIX.
5110 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5112 *) Add the configuration target debug-linux-ppro.
5113 Make 'openssl rsa' use the general key loading routines
5114 implemented in apps.c, and make those routines able to
5115 handle the key format FORMAT_NETSCAPE and the variant
5117 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5119 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5120 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5122 *) Add -keyform to rsautl, and document -engine.
5123 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5125 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5126 BIO_R_NO_SUCH_FILE error code rather than the generic
5127 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5130 *) Add new functions
5132 ERR_peek_last_error_line
5133 ERR_peek_last_error_line_data.
5134 These are similar to
5137 ERR_peek_error_line_data,
5138 but report on the latest error recorded rather than the first one
5139 still in the error queue.
5140 [Ben Laurie, Bodo Moeller]
5142 *) default_algorithms option in ENGINE config module. This allows things
5144 default_algorithms = ALL
5145 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5148 *) Preliminary ENGINE config module.
5151 *) New experimental application configuration code.
5154 *) Change the AES code to follow the same name structure as all other
5155 symmetric ciphers, and behave the same way. Move everything to
5156 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5157 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5159 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5160 [Ben Laurie and Theo de Raadt]
5162 *) Add option to output public keys in req command.
5163 [Massimiliano Pala madwolf@openca.org]
5165 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5166 (up to about 10% better than before for P-192 and P-224).
5169 *) New functions/macros
5171 SSL_CTX_set_msg_callback(ctx, cb)
5172 SSL_CTX_set_msg_callback_arg(ctx, arg)
5173 SSL_set_msg_callback(ssl, cb)
5174 SSL_set_msg_callback_arg(ssl, arg)
5176 to request calling a callback function
5178 void cb(int write_p, int version, int content_type,
5179 const void *buf, size_t len, SSL *ssl, void *arg)
5181 whenever a protocol message has been completely received
5182 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5183 protocol version according to which the SSL library interprets
5184 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5185 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5186 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5187 specification (change_cipher_spec(20), alert(21), handshake(22)).
5188 'buf' and 'len' point to the actual message, 'ssl' to the
5189 SSL object, and 'arg' is the application-defined value set by
5190 SSL[_CTX]_set_msg_callback_arg().
5192 'openssl s_client' and 'openssl s_server' have new '-msg' options
5193 to enable a callback that displays all protocol messages.
5196 *) Change the shared library support so shared libraries are built as
5197 soon as the corresponding static library is finished, and thereby get
5198 openssl and the test programs linked against the shared library.
5199 This still only happens when the keyword "shard" has been given to
5200 the configuration scripts.
5202 NOTE: shared library support is still an experimental thing, and
5203 backward binary compatibility is still not guaranteed.
5204 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5206 *) Add support for Subject Information Access extension.
5207 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5209 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5210 additional bytes when new memory had to be allocated, not just
5211 when reusing an existing buffer.
5214 *) New command line and configuration option 'utf8' for the req command.
5215 This allows field values to be specified as UTF8 strings.
5218 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5219 runs for the former and machine-readable output for the latter.
5222 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5223 of the e-mail address in the DN (i.e., it will go into a certificate
5224 extension only). The new configuration file option 'email_in_dn = no'
5225 has the same effect.
5226 [Massimiliano Pala madwolf@openca.org]
5228 *) Change all functions with names starting with des_ to be starting
5229 with DES_ instead. Add wrappers that are compatible with libdes,
5230 but are named _ossl_old_des_*. Finally, add macros that map the
5231 des_* symbols to the corresponding _ossl_old_des_* if libdes
5232 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5233 desired, the des_* symbols will be mapped to DES_*, with one
5236 Since we provide two compatibility mappings, the user needs to
5237 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5238 compatibility is desired. The default (i.e., when that macro
5239 isn't defined) is OpenSSL 0.9.6c compatibility.
5241 There are also macros that enable and disable the support of old
5242 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5243 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5244 are defined, the default will apply: to support the old des routines.
5246 In either case, one must include openssl/des.h to get the correct
5247 definitions. Do not try to just include openssl/des_old.h, that
5250 NOTE: This is a major break of an old API into a new one. Software
5251 authors are encouraged to switch to the DES_ style functions. Some
5252 time in the future, des_old.h and the libdes compatibility functions
5253 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5254 default), and then completely removed.
5257 *) Test for certificates which contain unsupported critical extensions.
5258 If such a certificate is found during a verify operation it is
5259 rejected by default: this behaviour can be overridden by either
5260 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5261 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5262 X509_supported_extension() has also been added which returns 1 if a
5263 particular extension is supported.
5266 *) Modify the behaviour of EVP cipher functions in similar way to digests
5267 to retain compatibility with existing code.
5270 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5271 compatibility with existing code. In particular the 'ctx' parameter does
5272 not have to be to be initialized before the call to EVP_DigestInit() and
5273 it is tidied up after a call to EVP_DigestFinal(). New function
5274 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5275 EVP_MD_CTX_copy() changed to not require the destination to be
5276 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5277 requires the destination to be valid.
5279 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5280 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5283 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5284 so that complete 'Handshake' protocol structures are kept in memory
5285 instead of overwriting 'msg_type' and 'length' with 'body' data.
5288 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5289 [Massimo Santin via Richard Levitte]
5291 *) Major restructuring to the underlying ENGINE code. This includes
5292 reduction of linker bloat, separation of pure "ENGINE" manipulation
5293 (initialisation, etc) from functionality dealing with implementations
5294 of specific crypto iterfaces. This change also introduces integrated
5295 support for symmetric ciphers and digest implementations - so ENGINEs
5296 can now accelerate these by providing EVP_CIPHER and EVP_MD
5297 implementations of their own. This is detailed in crypto/engine/README
5298 as it couldn't be adequately described here. However, there are a few
5299 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5300 were changed in the original introduction of ENGINE code have now
5301 reverted back - the hooking from this code to ENGINE is now a good
5302 deal more passive and at run-time, operations deal directly with
5303 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5304 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5305 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5306 they were not being used by the framework as there is no concept of a
5307 BIGNUM_METHOD and they could not be generalised to the new
5308 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5309 ENGINE_cpy() has been removed as it cannot be consistently defined in
5313 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5316 *) Change mkdef.pl to sort symbols that get the same entry number,
5317 and make sure the automatically generated functions ERR_load_*
5318 become part of libeay.num as well.
5321 *) New function SSL_renegotiate_pending(). This returns true once
5322 renegotiation has been requested (either SSL_renegotiate() call
5323 or HelloRequest/ClientHello received from the peer) and becomes
5324 false once a handshake has been completed.
5325 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5326 sends a HelloRequest, but does not ensure that a handshake takes
5327 place. SSL_renegotiate_pending() is useful for checking if the
5328 client has followed the request.)
5331 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5332 By default, clients may request session resumption even during
5333 renegotiation (if session ID contexts permit); with this option,
5334 session resumption is possible only in the first handshake.
5336 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5337 more bits available for options that should not be part of
5338 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5341 *) Add some demos for certificate and certificate request creation.
5344 *) Make maximum certificate chain size accepted from the peer application
5345 settable (SSL*_get/set_max_cert_list()), as proposed by
5346 "Douglas E. Engert" <deengert@anl.gov>.
5349 *) Add support for shared libraries for Unixware-7
5350 (Boyd Lynn Gerber <gerberb@zenez.com>).
5353 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5354 be done prior to destruction. Use this to unload error strings from
5355 ENGINEs that load their own error strings. NB: This adds two new API
5356 functions to "get" and "set" this destroy handler in an ENGINE.
5359 *) Alter all existing ENGINE implementations (except "openssl" and
5360 "openbsd") to dynamically instantiate their own error strings. This
5361 makes them more flexible to be built both as statically-linked ENGINEs
5362 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5363 Also, add stub code to each that makes building them as self-contained
5364 shared-libraries easier (see README.ENGINE).
5367 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5368 implementations into applications that are completely implemented in
5369 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5370 commands that can be used to configure what shared-library to load and
5371 to control aspects of the way it is handled. Also, made an update to
5372 the README.ENGINE file that brings its information up-to-date and
5373 provides some information and instructions on the "dynamic" ENGINE
5374 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5377 *) Make it possible to unload ranges of ERR strings with a new
5378 "ERR_unload_strings" function.
5381 *) Add a copy() function to EVP_MD.
5384 *) Make EVP_MD routines take a context pointer instead of just the
5385 md_data void pointer.
5388 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5389 that the digest can only process a single chunk of data
5390 (typically because it is provided by a piece of
5391 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5392 is only going to provide a single chunk of data, and hence the
5393 framework needn't accumulate the data for oneshot drivers.
5396 *) As with "ERR", make it possible to replace the underlying "ex_data"
5397 functions. This change also alters the storage and management of global
5398 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5399 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5400 index counters. The API functions that use this state have been changed
5401 to take a "class_index" rather than pointers to the class's local STACK
5402 and counter, and there is now an API function to dynamically create new
5403 classes. This centralisation allows us to (a) plug a lot of the
5404 thread-safety problems that existed, and (b) makes it possible to clean
5405 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5406 such data would previously have always leaked in application code and
5407 workarounds were in place to make the memory debugging turn a blind eye
5408 to it. Application code that doesn't use this new function will still
5409 leak as before, but their memory debugging output will announce it now
5410 rather than letting it slide.
5412 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5413 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5414 has a return value to indicate success or failure.
5417 *) Make it possible to replace the underlying "ERR" functions such that the
5418 global state (2 LHASH tables and 2 locks) is only used by the "default"
5419 implementation. This change also adds two functions to "get" and "set"
5420 the implementation prior to it being automatically set the first time
5421 any other ERR function takes place. Ie. an application can call "get",
5422 pass the return value to a module it has just loaded, and that module
5423 can call its own "set" function using that value. This means the
5424 module's "ERR" operations will use (and modify) the error state in the
5425 application and not in its own statically linked copy of OpenSSL code.
5428 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5429 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5430 the operation, and provides a more encapsulated way for external code
5431 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5432 to use these functions rather than manually incrementing the counts.
5434 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5437 *) Add EVP test program.
5440 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5443 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5444 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5445 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5446 These allow a CRL to be built without having to access X509_CRL fields
5447 directly. Modify 'ca' application to use new functions.
5450 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5451 bug workarounds. Rollback attack detection is a security feature.
5452 The problem will only arise on OpenSSL servers when TLSv1 is not
5453 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5454 Software authors not wanting to support TLSv1 will have special reasons
5455 for their choice and can explicitly enable this option.
5456 [Bodo Moeller, Lutz Jaenicke]
5458 *) Rationalise EVP so it can be extended: don't include a union of
5459 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5460 (similar to those existing for EVP_CIPHER_CTX).
5465 EVP_MD_CTX_init(&md); /* new function call */
5466 EVP_DigestInit(&md, EVP_sha1());
5467 EVP_DigestUpdate(&md, in, len);
5468 EVP_DigestFinal(&md, out, NULL);
5469 EVP_MD_CTX_cleanup(&md); /* new function call */
5473 *) Make DES key schedule conform to the usual scheme, as well as
5474 correcting its structure. This means that calls to DES functions
5475 now have to pass a pointer to a des_key_schedule instead of a
5476 plain des_key_schedule (which was actually always a pointer
5479 des_key_schedule ks;
5481 des_set_key_checked(..., &ks);
5482 des_ncbc_encrypt(..., &ks, ...);
5484 (Note that a later change renames 'des_...' into 'DES_...'.)
5487 *) Initial reduction of linker bloat: the use of some functions, such as
5488 PEM causes large amounts of unused functions to be linked in due to
5489 poor organisation. For example pem_all.c contains every PEM function
5490 which has a knock on effect of linking in large amounts of (unused)
5491 ASN1 code. Grouping together similar functions and splitting unrelated
5492 functions prevents this.
5495 *) Cleanup of EVP macros.
5498 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5499 correct _ecb suffix.
5502 *) Add initial OCSP responder support to ocsp application. The
5503 revocation information is handled using the text based index
5504 use by the ca application. The responder can either handle
5505 requests generated internally, supplied in files (for example
5506 via a CGI script) or using an internal minimal server.
5509 *) Add configuration choices to get zlib compression for TLS.
5512 *) Changes to Kerberos SSL for RFC 2712 compliance:
5513 1. Implemented real KerberosWrapper, instead of just using
5514 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5515 2. Implemented optional authenticator field of KerberosWrapper.
5517 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5518 and authenticator structs; see crypto/krb5/.
5520 Generalized Kerberos calls to support multiple Kerberos libraries.
5521 [Vern Staats <staatsvr@asc.hpc.mil>,
5522 Jeffrey Altman <jaltman@columbia.edu>
5523 via Richard Levitte]
5525 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5526 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5527 values for each of the key sizes rather than having just
5528 parameters (and 'speed' generating keys each time).
5531 *) Speed up EVP routines.
5534 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5535 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5536 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5537 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5539 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5540 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5541 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5544 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5546 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5549 *) Added the OS2-EMX target.
5550 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5552 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5553 to support NCONF routines in extension code. New function CONF_set_nconf()
5554 to allow functions which take an NCONF to also handle the old LHASH
5555 structure: this means that the old CONF compatible routines can be
5556 retained (in particular wrt extensions) without having to duplicate the
5557 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5560 *) Enhance the general user interface with mechanisms for inner control
5561 and with possibilities to have yes/no kind of prompts.
5564 *) Change all calls to low level digest routines in the library and
5565 applications to use EVP. Add missing calls to HMAC_cleanup() and
5566 don't assume HMAC_CTX can be copied using memcpy().
5567 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5569 *) Add the possibility to control engines through control names but with
5570 arbitrary arguments instead of just a string.
5571 Change the key loaders to take a UI_METHOD instead of a callback
5572 function pointer. NOTE: this breaks binary compatibility with earlier
5573 versions of OpenSSL [engine].
5574 Adapt the nCipher code for these new conditions and add a card insertion
5578 *) Enhance the general user interface with mechanisms to better support
5579 dialog box interfaces, application-defined prompts, the possibility
5580 to use defaults (for example default passwords from somewhere else)
5581 and interrupts/cancellations.
5584 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5585 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5588 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5589 tidy up some unnecessarily weird code in 'sk_new()').
5590 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5592 *) Change the key loading routines for ENGINEs to use the same kind
5593 callback (pem_password_cb) as all other routines that need this
5597 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5598 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5599 than this minimum value is recommended.
5602 *) New random seeder for OpenVMS, using the system process statistics
5603 that are easily reachable.
5606 *) Windows apparently can't transparently handle global
5607 variables defined in DLLs. Initialisations such as:
5609 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5611 wont compile. This is used by the any applications that need to
5612 declare their own ASN1 modules. This was fixed by adding the option
5613 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5614 needed for static libraries under Win32.
5617 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5618 setting of purpose and trust fields. New X509_STORE trust and
5619 purpose functions and tidy up setting in other SSL functions.
5622 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5623 structure. These are inherited by X509_STORE_CTX when it is
5624 initialised. This allows various defaults to be set in the
5625 X509_STORE structure (such as flags for CRL checking and custom
5626 purpose or trust settings) for functions which only use X509_STORE_CTX
5627 internally such as S/MIME.
5629 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5630 trust settings if they are not set in X509_STORE. This allows X509_STORE
5631 purposes and trust (in S/MIME for example) to override any set by default.
5633 Add command line options for CRL checking to smime, s_client and s_server
5637 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5638 are set then the CRL is looked up in the X509_STORE structure and
5639 its validity and signature checked, then if the certificate is found
5640 in the CRL the verify fails with a revoked error.
5642 Various new CRL related callbacks added to X509_STORE_CTX structure.
5644 Command line options added to 'verify' application to support this.
5646 This needs some additional work, such as being able to handle multiple
5647 CRLs with different times, extension based lookup (rather than just
5648 by subject name) and ultimately more complete V2 CRL extension
5652 *) Add a general user interface API (crypto/ui/). This is designed
5653 to replace things like des_read_password and friends (backward
5654 compatibility functions using this new API are provided).
5655 The purpose is to remove prompting functions from the DES code
5656 section as well as provide for prompting through dialog boxes in
5657 a window system and the like.
5660 *) Add "ex_data" support to ENGINE so implementations can add state at a
5661 per-structure level rather than having to store it globally.
5664 *) Make it possible for ENGINE structures to be copied when retrieved by
5665 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5666 This causes the "original" ENGINE structure to act like a template,
5667 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5668 operational state can be localised to each ENGINE structure, despite the
5669 fact they all share the same "methods". New ENGINE structures returned in
5670 this case have no functional references and the return value is the single
5671 structural reference. This matches the single structural reference returned
5672 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5676 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5677 needs to match any other type at all we need to manually clear the
5681 *) Changes to the "openssl engine" utility to include;
5682 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5683 about an ENGINE's available control commands.
5684 - executing control commands from command line arguments using the
5685 '-pre' and '-post' switches. '-post' is only used if '-t' is
5686 specified and the ENGINE is successfully initialised. The syntax for
5687 the individual commands are colon-separated, for example;
5688 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5691 *) New dynamic control command support for ENGINEs. ENGINEs can now
5692 declare their own commands (numbers), names (strings), descriptions,
5693 and input types for run-time discovery by calling applications. A
5694 subset of these commands are implicitly classed as "executable"
5695 depending on their input type, and only these can be invoked through
5696 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5697 can be based on user input, config files, etc). The distinction is
5698 that "executable" commands cannot return anything other than a boolean
5699 result and can only support numeric or string input, whereas some
5700 discoverable commands may only be for direct use through
5701 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5702 pointers, or other custom uses. The "executable" commands are to
5703 support parameterisations of ENGINE behaviour that can be
5704 unambiguously defined by ENGINEs and used consistently across any
5705 OpenSSL-based application. Commands have been added to all the
5706 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5707 control over shared-library paths without source code alterations.
5710 *) Changed all ENGINE implementations to dynamically allocate their
5711 ENGINEs rather than declaring them statically. Apart from this being
5712 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5713 this also allows the implementations to compile without using the
5714 internal engine_int.h header.
5717 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5718 'const' value. Any code that should be able to modify a RAND_METHOD
5719 should already have non-const pointers to it (ie. they should only
5720 modify their own ones).
5723 *) Made a variety of little tweaks to the ENGINE code.
5724 - "atalla" and "ubsec" string definitions were moved from header files
5725 to C code. "nuron" string definitions were placed in variables
5726 rather than hard-coded - allowing parameterisation of these values
5727 later on via ctrl() commands.
5728 - Removed unused "#if 0"'d code.
5729 - Fixed engine list iteration code so it uses ENGINE_free() to release
5730 structural references.
5731 - Constified the RAND_METHOD element of ENGINE structures.
5732 - Constified various get/set functions as appropriate and added
5733 missing functions (including a catch-all ENGINE_cpy that duplicates
5734 all ENGINE values onto a new ENGINE except reference counts/state).
5735 - Removed NULL parameter checks in get/set functions. Setting a method
5736 or function to NULL is a way of cancelling out a previously set
5737 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5738 and doesn't justify the extra error symbols and code.
5739 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5740 flags from engine_int.h to engine.h.
5741 - Changed prototypes for ENGINE handler functions (init(), finish(),
5742 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5745 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5746 to the algorithm using long division. The binary algorithm can be
5747 used only if the modulus is odd. On 32-bit systems, it is faster
5748 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5749 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5750 up to 450 bits. In 64-bit environments, the binary algorithm
5751 appears to be advantageous for much longer moduli; here we use it
5752 for moduli up to 2048 bits.
5755 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5756 could not support the combine flag in choice fields.
5759 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5760 extensions from a certificate request to the certificate.
5763 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5764 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5765 file: this allows the display of the certificate about to be
5766 signed to be customised, to allow certain fields to be included
5767 or excluded and extension details. The old system didn't display
5768 multicharacter strings properly, omitted fields not in the policy
5769 and couldn't display additional details such as extensions.
5772 *) Function EC_POINTs_mul for multiple scalar multiplication
5773 of an arbitrary number of elliptic curve points
5774 \sum scalars[i]*points[i],
5775 optionally including the generator defined for the EC_GROUP:
5776 scalar*generator + \sum scalars[i]*points[i].
5778 EC_POINT_mul is a simple wrapper function for the typical case
5779 that the point list has just one item (besides the optional
5783 *) First EC_METHODs for curves over GF(p):
5785 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5786 operations and provides various method functions that can also
5787 operate with faster implementations of modular arithmetic.
5789 EC_GFp_mont_method() reuses most functions that are part of
5790 EC_GFp_simple_method, but uses Montgomery arithmetic.
5792 [Bodo Moeller; point addition and point doubling
5793 implementation directly derived from source code provided by
5794 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5796 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5797 crypto/ec/ec_lib.c):
5799 Curves are EC_GROUP objects (with an optional group generator)
5800 based on EC_METHODs that are built into the library.
5802 Points are EC_POINT objects based on EC_GROUP objects.
5804 Most of the framework would be able to handle curves over arbitrary
5805 finite fields, but as there are no obvious types for fields other
5806 than GF(p), some functions are limited to that for now.
5809 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5810 that the file contains a complete HTTP response.
5813 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5814 change the def and num file printf format specifier from "%-40sXXX"
5815 to "%-39s XXX". The latter will always guarantee a space after the
5816 field while the former will cause them to run together if the field
5817 is 40 of more characters long.
5820 *) Constify the cipher and digest 'method' functions and structures
5821 and modify related functions to take constant EVP_MD and EVP_CIPHER
5825 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5826 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5829 *) Modify EVP_Digest*() routines so they now return values. Although the
5830 internal software routines can never fail additional hardware versions
5834 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5836 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5837 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5840 ERR_R_NESTED_ASN1_ERROR
5842 ERR_R_MISSING_ASN1_EOS
5843 were 4 .. 9, conflicting with
5844 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5846 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5847 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5849 Add new error code 'ERR_R_INTERNAL_ERROR'.
5852 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5856 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5857 sets the subject name for a new request or supersedes the
5858 subject name in a given request. Formats that can be parsed are
5859 'CN=Some Name, OU=myOU, C=IT'
5861 'CN=Some Name/OU=myOU/C=IT'.
5863 Add options '-batch' and '-verbose' to 'openssl req'.
5864 [Massimiliano Pala <madwolf@hackmasters.net>]
5866 *) Introduce the possibility to access global variables through
5867 functions on platform were that's the best way to handle exporting
5868 global variables in shared libraries. To enable this functionality,
5869 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5870 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5871 is normally done by Configure or something similar).
5873 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5874 in the source file (foo.c) like this:
5876 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5877 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5879 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5880 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5882 OPENSSL_DECLARE_GLOBAL(int,foo);
5883 #define foo OPENSSL_GLOBAL_REF(foo)
5884 OPENSSL_DECLARE_GLOBAL(double,bar);
5885 #define bar OPENSSL_GLOBAL_REF(bar)
5887 The #defines are very important, and therefore so is including the
5888 header file everywhere where the defined globals are used.
5890 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5891 of ASN.1 items, but that structure is a bit different.
5893 The largest change is in util/mkdef.pl which has been enhanced with
5894 better and easier to understand logic to choose which symbols should
5895 go into the Windows .def files as well as a number of fixes and code
5896 cleanup (among others, algorithm keywords are now sorted
5897 lexicographically to avoid constant rewrites).
5900 *) In BN_div() keep a copy of the sign of 'num' before writing the
5901 result to 'rm' because if rm==num the value will be overwritten
5902 and produce the wrong result if 'num' is negative: this caused
5903 problems with BN_mod() and BN_nnmod().
5906 *) Function OCSP_request_verify(). This checks the signature on an
5907 OCSP request and verifies the signer certificate. The signer
5908 certificate is just checked for a generic purpose and OCSP request
5912 *) Add OCSP_check_validity() function to check the validity of OCSP
5913 responses. OCSP responses are prepared in real time and may only
5914 be a few seconds old. Simply checking that the current time lies
5915 between thisUpdate and nextUpdate max reject otherwise valid responses
5916 caused by either OCSP responder or client clock inaccuracy. Instead
5917 we allow thisUpdate and nextUpdate to fall within a certain period of
5918 the current time. The age of the response can also optionally be
5919 checked. Two new options -validity_period and -status_age added to
5923 *) If signature or public key algorithm is unrecognized print out its
5924 OID rather that just UNKNOWN.
5927 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5928 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5929 ID to be generated from the issuer certificate alone which can then be
5930 passed to OCSP_id_issuer_cmp().
5933 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5934 ASN1 modules to export functions returning ASN1_ITEM pointers
5935 instead of the ASN1_ITEM structures themselves. This adds several
5936 new macros which allow the underlying ASN1 function/structure to
5937 be accessed transparently. As a result code should not use ASN1_ITEM
5938 references directly (such as &X509_it) but instead use the relevant
5939 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5940 use of the new ASN1 code on platforms where exporting structures
5941 is problematical (for example in shared libraries) but exporting
5942 functions returning pointers to structures is not.
5945 *) Add support for overriding the generation of SSL/TLS session IDs.
5946 These callbacks can be registered either in an SSL_CTX or per SSL.
5947 The purpose of this is to allow applications to control, if they wish,
5948 the arbitrary values chosen for use as session IDs, particularly as it
5949 can be useful for session caching in multiple-server environments. A
5950 command-line switch for testing this (and any client code that wishes
5951 to use such a feature) has been added to "s_server".
5952 [Geoff Thorpe, Lutz Jaenicke]
5954 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5955 of the form '#if defined(...) || defined(...) || ...' and
5956 '#if !defined(...) && !defined(...) && ...'. This also avoids
5957 the growing number of special cases it was previously handling.
5960 *) Make all configuration macros available for application by making
5961 sure they are available in opensslconf.h, by giving them names starting
5962 with "OPENSSL_" to avoid conflicts with other packages and by making
5963 sure e_os2.h will cover all platform-specific cases together with
5965 Additionally, it is now possible to define configuration/platform-
5966 specific names (called "system identities"). In the C code, these
5967 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5968 macro with the name beginning with "OPENSSL_SYS_", which is determined
5969 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5973 *) New option -set_serial to 'req' and 'x509' this allows the serial
5974 number to use to be specified on the command line. Previously self
5975 signed certificates were hard coded with serial number 0 and the
5976 CA options of 'x509' had to use a serial number in a file which was
5980 *) New options to 'ca' utility to support V2 CRL entry extensions.
5981 Currently CRL reason, invalidity date and hold instruction are
5982 supported. Add new CRL extensions to V3 code and some new objects.
5985 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5986 disable standard block padding (aka PKCS#5 padding) in the EVP
5987 API, which was previously mandatory. This means that the data is
5988 not padded in any way and so the total length much be a multiple
5989 of the block size, otherwise an error occurs.
5992 *) Initial (incomplete) OCSP SSL support.
5995 *) New function OCSP_parse_url(). This splits up a URL into its host,
5996 port and path components: primarily to parse OCSP URLs. New -url
5997 option to ocsp utility.
6000 *) New nonce behavior. The return value of OCSP_check_nonce() now
6001 reflects the various checks performed. Applications can decide
6002 whether to tolerate certain situations such as an absent nonce
6003 in a response when one was present in a request: the ocsp application
6004 just prints out a warning. New function OCSP_add1_basic_nonce()
6005 this is to allow responders to include a nonce in a response even if
6006 the request is nonce-less.
6009 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6010 skipped when using openssl x509 multiple times on a single input file,
6011 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6014 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6015 set string type: to handle setting ASN1_TIME structures. Fix ca
6016 utility to correctly initialize revocation date of CRLs.
6019 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6020 the clients preferred ciphersuites and rather use its own preferences.
6021 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6022 Internet Explorer by ensuring unchanged hash method during stepup.
6023 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6026 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6027 to aes and add a new 'exist' option to print out symbols that don't
6031 *) Additional options to ocsp utility to allow flags to be set and
6032 additional certificates supplied.
6035 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6036 OCSP client a number of certificate to only verify the response
6040 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6041 handle the new API. Currently only ECB, CBC modes supported. Add new
6044 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6045 Encryption Standard (AES) Ciphersuites for Transport Layer
6046 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6047 not enabled by default and were not part of the "ALL" ciphersuite
6048 alias because they were not yet official; they could be
6049 explicitly requested by specifying the "AESdraft" ciphersuite
6050 group alias. In the final release of OpenSSL 0.9.7, the group
6051 alias is called "AES" and is part of "ALL".)
6052 [Ben Laurie, Steve Henson, Bodo Moeller]
6054 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6055 request to response.
6058 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6059 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6060 extract information from a certificate request. OCSP_response_create()
6061 creates a response and optionally adds a basic response structure.
6062 OCSP_basic_add1_status() adds a complete single response to a basic
6063 response and returns the OCSP_SINGLERESP structure just added (to allow
6064 extensions to be included for example). OCSP_basic_add1_cert() adds a
6065 certificate to a basic response and OCSP_basic_sign() signs a basic
6066 response with various flags. New helper functions ASN1_TIME_check()
6067 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6068 (converts ASN1_TIME to GeneralizedTime).
6071 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6072 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6073 structure from a certificate. X509_pubkey_digest() digests the public_key
6074 contents: this is used in various key identifiers.
6077 *) Make sk_sort() tolerate a NULL argument.
6078 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6080 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6081 passed by the function are trusted implicitly. If any of them signed the
6082 response then it is assumed to be valid and is not verified.
6085 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6086 to data. This was previously part of the PKCS7 ASN1 code. This
6087 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6088 [Steve Henson, reported by Kenneth R. Robinette
6089 <support@securenetterm.com>]
6091 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6092 routines: without these tracing memory leaks is very painful.
6093 Fix leaks in PKCS12 and PKCS7 routines.
6096 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6097 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6098 effectively meant GeneralizedTime would never be used. Now it
6099 is initialised to -1 but X509_time_adj() now has to check the value
6100 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6101 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6102 [Steve Henson, reported by Kenneth R. Robinette
6103 <support@securenetterm.com>]
6105 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6106 result in a zero length in the ASN1_INTEGER structure which was
6107 not consistent with the structure when d2i_ASN1_INTEGER() was used
6108 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6109 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6110 where it did not print out a minus for negative ASN1_INTEGER.
6113 *) Add summary printout to ocsp utility. The various functions which
6114 convert status values to strings have been renamed to:
6115 OCSP_response_status_str(), OCSP_cert_status_str() and
6116 OCSP_crl_reason_str() and are no longer static. New options
6117 to verify nonce values and to disable verification. OCSP response
6118 printout format cleaned up.
6121 *) Add additional OCSP certificate checks. These are those specified
6122 in RFC2560. This consists of two separate checks: the CA of the
6123 certificate being checked must either be the OCSP signer certificate
6124 or the issuer of the OCSP signer certificate. In the latter case the
6125 OCSP signer certificate must contain the OCSP signing extended key
6126 usage. This check is performed by attempting to match the OCSP
6127 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6128 in the OCSP_CERTID structures of the response.
6131 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6132 and related routines. This uses the standard OpenSSL certificate
6133 verify routines to perform initial checks (just CA validity) and
6134 to obtain the certificate chain. Then additional checks will be
6135 performed on the chain. Currently the root CA is checked to see
6136 if it is explicitly trusted for OCSP signing. This is used to set
6137 a root CA as a global signing root: that is any certificate that
6138 chains to that CA is an acceptable OCSP signing certificate.
6141 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6142 extensions from a separate configuration file.
6143 As when reading extensions from the main configuration file,
6144 the '-extensions ...' option may be used for specifying the
6146 [Massimiliano Pala <madwolf@comune.modena.it>]
6148 *) New OCSP utility. Allows OCSP requests to be generated or
6149 read. The request can be sent to a responder and the output
6150 parsed, outputed or printed in text form. Not complete yet:
6151 still needs to check the OCSP response validity.
6154 *) New subcommands for 'openssl ca':
6155 'openssl ca -status <serial>' prints the status of the cert with
6156 the given serial number (according to the index file).
6157 'openssl ca -updatedb' updates the expiry status of certificates
6159 [Massimiliano Pala <madwolf@comune.modena.it>]
6161 *) New '-newreq-nodes' command option to CA.pl. This is like
6162 '-newreq', but calls 'openssl req' with the '-nodes' option
6163 so that the resulting key is not encrypted.
6164 [Damien Miller <djm@mindrot.org>]
6166 *) New configuration for the GNU Hurd.
6167 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6169 *) Initial code to implement OCSP basic response verify. This
6170 is currently incomplete. Currently just finds the signer's
6171 certificate and verifies the signature on the response.
6174 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6175 value of OPENSSLDIR. This is available via the new '-d' option
6176 to 'openssl version', and is also included in 'openssl version -a'.
6179 *) Allowing defining memory allocation callbacks that will be given
6180 file name and line number information in additional arguments
6181 (a const char* and an int). The basic functionality remains, as
6182 well as the original possibility to just replace malloc(),
6183 realloc() and free() by functions that do not know about these
6184 additional arguments. To register and find out the current
6185 settings for extended allocation functions, the following
6186 functions are provided:
6188 CRYPTO_set_mem_ex_functions
6189 CRYPTO_set_locked_mem_ex_functions
6190 CRYPTO_get_mem_ex_functions
6191 CRYPTO_get_locked_mem_ex_functions
6193 These work the same way as CRYPTO_set_mem_functions and friends.
6194 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6195 extended allocation function is enabled.
6196 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6197 a conventional allocation function is enabled.
6198 [Richard Levitte, Bodo Moeller]
6200 *) Finish off removing the remaining LHASH function pointer casts.
6201 There should no longer be any prototype-casting required when using
6202 the LHASH abstraction, and any casts that remain are "bugs". See
6203 the callback types and macros at the head of lhash.h for details
6204 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6207 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6208 If /dev/[u]random devices are not available or do not return enough
6209 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6211 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6212 /etc/entropy will be queried once each in this sequence, quering stops
6213 when enough entropy was collected without querying more sockets.
6216 *) Change the Unix RAND_poll() variant to be able to poll several
6217 random devices, as specified by DEVRANDOM, until a sufficient amount
6218 of data has been collected. We spend at most 10 ms on each file
6219 (select timeout) and read in non-blocking mode. DEVRANDOM now
6220 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6221 (previously it was just the string "/dev/urandom"), so on typical
6222 platforms the 10 ms delay will never occur.
6223 Also separate out the Unix variant to its own file, rand_unix.c.
6224 For VMS, there's a currently-empty rand_vms.c.
6227 *) Move OCSP client related routines to ocsp_cl.c. These
6228 provide utility functions which an application needing
6229 to issue a request to an OCSP responder and analyse the
6230 response will typically need: as opposed to those which an
6231 OCSP responder itself would need which will be added later.
6233 OCSP_request_sign() signs an OCSP request with an API similar
6234 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6235 response. OCSP_response_get1_basic() extracts basic response
6236 from response. OCSP_resp_find_status(): finds and extracts status
6237 information from an OCSP_CERTID structure (which will be created
6238 when the request structure is built). These are built from lower
6239 level functions which work on OCSP_SINGLERESP structures but
6240 wont normally be used unless the application wishes to examine
6241 extensions in the OCSP response for example.
6243 Replace nonce routines with a pair of functions.
6244 OCSP_request_add1_nonce() adds a nonce value and optionally
6245 generates a random value. OCSP_check_nonce() checks the
6246 validity of the nonce in an OCSP response.
6249 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6250 This doesn't copy the supplied OCSP_CERTID and avoids the
6251 need to free up the newly created id. Change return type
6252 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6253 This can then be used to add extensions to the request.
6254 Deleted OCSP_request_new(), since most of its functionality
6255 is now in OCSP_REQUEST_new() (and the case insensitive name
6256 clash) apart from the ability to set the request name which
6257 will be added elsewhere.
6260 *) Update OCSP API. Remove obsolete extensions argument from
6261 various functions. Extensions are now handled using the new
6262 OCSP extension code. New simple OCSP HTTP function which
6263 can be used to send requests and parse the response.
6266 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6267 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6268 uses the special reorder version of SET OF to sort the attributes
6269 and reorder them to match the encoded order. This resolves a long
6270 standing problem: a verify on a PKCS7 structure just after signing
6271 it used to fail because the attribute order did not match the
6272 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6273 it uses the received order. This is necessary to tolerate some broken
6274 software that does not order SET OF. This is handled by encoding
6275 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6276 to produce the required SET OF.
6279 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6280 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6281 files to get correct declarations of the ASN.1 item variables.
6284 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6285 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6286 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6287 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6288 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6289 ASN1_ITEM and no wrapper functions.
6292 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6293 replace the old function pointer based I/O routines. Change most of
6294 the *_d2i_bio() and *_d2i_fp() functions to use these.
6297 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6298 lines, recognice more "algorithms" that can be deselected, and make
6299 it complain about algorithm deselection that isn't recognised.
6302 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6303 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6304 to use new functions. Add NO_ASN1_OLD which can be set to remove
6305 some old style ASN1 functions: this can be used to determine if old
6306 code will still work when these eventually go away.
6309 *) New extension functions for OCSP structures, these follow the
6310 same conventions as certificates and CRLs.
6313 *) New function X509V3_add1_i2d(). This automatically encodes and
6314 adds an extension. Its behaviour can be customised with various
6315 flags to append, replace or delete. Various wrappers added for
6316 certifcates and CRLs.
6319 *) Fix to avoid calling the underlying ASN1 print routine when
6320 an extension cannot be parsed. Correct a typo in the
6321 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6324 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6325 entries for variables.
6328 *) Add functionality to apps/openssl.c for detecting locking
6329 problems: As the program is single-threaded, all we have
6330 to do is register a locking callback using an array for
6331 storing which locks are currently held by the program.
6334 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6335 SSL_get_ex_data_X509_STORE_idx(), which is used in
6336 ssl_verify_cert_chain() and thus can be called at any time
6337 during TLS/SSL handshakes so that thread-safety is essential.
6338 Unfortunately, the ex_data design is not at all suited
6339 for multi-threaded use, so it probably should be abolished.
6342 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6343 [Broadcom, tweaked and integrated by Geoff Thorpe]
6345 *) Move common extension printing code to new function
6346 X509V3_print_extensions(). Reorganise OCSP print routines and
6347 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6350 *) New function X509_signature_print() to remove duplication in some
6354 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6355 set (this was treated exactly the same as SET OF previously). This
6356 is used to reorder the STACK representing the structure to match the
6357 encoding. This will be used to get round a problem where a PKCS7
6358 structure which was signed could not be verified because the STACK
6359 order did not reflect the encoded order.
6362 *) Reimplement the OCSP ASN1 module using the new code.
6365 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6366 for its ASN1 operations. The old style function pointers still exist
6367 for now but they will eventually go away.
6370 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6371 completely replaces the old ASN1 functionality with a table driven
6372 encoder and decoder which interprets an ASN1_ITEM structure describing
6373 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6374 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6375 has also been converted to the new form.
6378 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6379 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6380 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6381 for negative moduli.
6384 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6385 of not touching the result's sign bit.
6388 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6392 *) Changed the LHASH code to use prototypes for callbacks, and created
6393 macros to declare and implement thin (optionally static) functions
6394 that provide type-safety and avoid function pointer casting for the
6395 type-specific callbacks.
6398 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6400 [Veers Staats <staatsvr@asc.hpc.mil>,
6401 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6403 *) Reformat the FAQ so the different questions and answers can be divided
6404 in sections depending on the subject.
6407 *) Have the zlib compression code load ZLIB.DLL dynamically under
6411 *) New function BN_mod_sqrt for computing square roots modulo a prime
6412 (using the probabilistic Tonelli-Shanks algorithm unless
6413 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6414 be handled deterministically).
6415 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6417 *) Make BN_mod_inverse faster by explicitly handling small quotients
6418 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6419 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6422 *) New function BN_kronecker.
6425 *) Fix BN_gcd so that it works on negative inputs; the result is
6426 positive unless both parameters are zero.
6427 Previously something reasonably close to an infinite loop was
6428 possible because numbers could be growing instead of shrinking
6429 in the implementation of Euclid's algorithm.
6432 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6433 sign of the number in question.
6435 Fix BN_is_word(a,w) to work correctly for w == 0.
6437 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6438 because its test if the absolute value of 'a' equals 'w'.
6439 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6440 it exists mostly for use in the implementations of BN_is_zero(),
6441 BN_is_one(), and BN_is_word().
6444 *) New function BN_swap.
6447 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6448 the exponentiation functions are more likely to produce reasonable
6449 results on negative inputs.
6452 *) Change BN_mod_mul so that the result is always non-negative.
6453 Previously, it could be negative if one of the factors was negative;
6454 I don't think anyone really wanted that behaviour.
6457 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6458 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6459 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6460 and add new functions:
6469 BN_mod_lshift1_quick
6473 These functions always generate non-negative results.
6475 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6476 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6478 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6479 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6480 be reduced modulo m.
6481 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6484 The following entry accidentally appeared in the CHANGES file
6485 distributed with OpenSSL 0.9.7. The modifications described in
6486 it do *not* apply to OpenSSL 0.9.7.
6488 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6489 was actually never needed) and in BN_mul(). The removal in BN_mul()
6490 required a small change in bn_mul_part_recursive() and the addition
6491 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6492 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6493 bn_sub_words() and bn_add_words() except they take arrays with
6498 *) In 'openssl passwd', verify passwords read from the terminal
6499 unless the '-salt' option is used (which usually means that
6500 verification would just waste user's time since the resulting
6501 hash is going to be compared with some given password hash)
6502 or the new '-noverify' option is used.
6504 This is an incompatible change, but it does not affect
6505 non-interactive use of 'openssl passwd' (passwords on the command
6506 line, '-stdin' option, '-in ...' option) and thus should not
6510 *) Remove all references to RSAref, since there's no more need for it.
6513 *) Make DSO load along a path given through an environment variable
6514 (SHLIB_PATH) with shl_load().
6517 *) Constify the ENGINE code as a result of BIGNUM constification.
6518 Also constify the RSA code and most things related to it. In a
6519 few places, most notable in the depth of the ASN.1 code, ugly
6520 casts back to non-const were required (to be solved at a later
6524 *) Make it so the openssl application has all engines loaded by default.
6527 *) Constify the BIGNUM routines a little more.
6530 *) Add the following functions:
6532 ENGINE_load_cswift()
6534 ENGINE_load_atalla()
6536 ENGINE_load_builtin_engines()
6538 That way, an application can itself choose if external engines that
6539 are built-in in OpenSSL shall ever be used or not. The benefit is
6540 that applications won't have to be linked with libdl or other dso
6541 libraries unless it's really needed.
6543 Changed 'openssl engine' to load all engines on demand.
6544 Changed the engine header files to avoid the duplication of some
6545 declarations (they differed!).
6548 *) 'openssl engine' can now list capabilities.
6551 *) Better error reporting in 'openssl engine'.
6554 *) Never call load_dh_param(NULL) in s_server.
6557 *) Add engine application. It can currently list engines by name and
6558 identity, and test if they are actually available.
6561 *) Improve RPM specification file by forcing symbolic linking and making
6562 sure the installed documentation is also owned by root.root.
6563 [Damien Miller <djm@mindrot.org>]
6565 *) Give the OpenSSL applications more possibilities to make use of
6566 keys (public as well as private) handled by engines.
6569 *) Add OCSP code that comes from CertCo.
6572 *) Add VMS support for the Rijndael code.
6575 *) Added untested support for Nuron crypto accelerator.
6578 *) Add support for external cryptographic devices. This code was
6579 previously distributed separately as the "engine" branch.
6580 [Geoff Thorpe, Richard Levitte]
6582 *) Rework the filename-translation in the DSO code. It is now possible to
6583 have far greater control over how a "name" is turned into a filename
6584 depending on the operating environment and any oddities about the
6585 different shared library filenames on each system.
6588 *) Support threads on FreeBSD-elf in Configure.
6591 *) Fix for SHA1 assembly problem with MASM: it produces
6592 warnings about corrupt line number information when assembling
6593 with debugging information. This is caused by the overlapping
6595 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6598 NCONF_get_number() has no error checking at all. As a replacement,
6599 NCONF_get_number_e() is defined (_e for "error checking") and is
6600 promoted strongly. The old NCONF_get_number is kept around for
6601 binary backward compatibility.
6602 Make it possible for methods to load from something other than a BIO,
6603 by providing a function pointer that is given a name instead of a BIO.
6604 For example, this could be used to load configuration data from an
6608 *) Fix for non blocking accept BIOs. Added new I/O special reason
6609 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6610 with non blocking I/O was not possible because no retry code was
6611 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6615 *) Added the beginnings of Rijndael support.
6618 *) Fix for bug in DirectoryString mask setting. Add support for
6619 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6620 to allow certificate printing to more controllable, additional
6621 'certopt' option to 'x509' to allow new printing options to be
6625 *) Clean old EAY MD5 hack from e_os.h.
6628 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6630 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6631 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6632 [Joe Orton, Steve Henson]
6634 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6636 *) Fix additional bug revealed by the NISCC test suite:
6638 Stop bug triggering large recursion when presented with
6639 certain ASN.1 tags (CVE-2003-0851)
6642 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6644 *) Fix various bugs revealed by running the NISCC test suite:
6646 Stop out of bounds reads in the ASN1 code when presented with
6647 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6649 If verify callback ignores invalid public key errors don't try to check
6650 certificate signature with the NULL public key.
6654 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6655 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6659 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6660 extra data after the compression methods not only for TLS 1.0
6661 but also for SSL 3.0 (as required by the specification).
6662 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6664 *) Change X509_certificate_type() to mark the key as exported/exportable
6665 when it's 512 *bits* long, not 512 bytes.
6668 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6670 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6671 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6672 a protocol version number mismatch like a decryption error
6673 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6676 *) Turn on RSA blinding by default in the default implementation
6677 to avoid a timing attack. Applications that don't want it can call
6678 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6679 They would be ill-advised to do so in most cases.
6680 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6682 *) Change RSA blinding code so that it works when the PRNG is not
6683 seeded (in this case, the secret RSA exponent is abused as
6684 an unpredictable seed -- if it is not unpredictable, there
6685 is no point in blinding anyway). Make RSA blinding thread-safe
6686 by remembering the creator's thread ID in rsa->blinding and
6687 having all other threads use local one-time blinding factors
6688 (this requires more computation than sharing rsa->blinding, but
6689 avoids excessive locking; and if an RSA object is not shared
6690 between threads, blinding will still be very fast).
6693 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6695 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6696 via timing by performing a MAC computation even if incorrrect
6697 block cipher padding has been found. This is a countermeasure
6698 against active attacks where the attacker has to distinguish
6699 between bad padding and a MAC verification error. (CVE-2003-0078)
6701 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6702 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6703 Martin Vuagnoux (EPFL, Ilion)]
6705 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6707 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6708 memory from it's contents. This is done with a counter that will
6709 place alternating values in each byte. This can be used to solve
6710 two issues: 1) the removal of calls to memset() by highly optimizing
6711 compilers, and 2) cleansing with other values than 0, since those can
6712 be read through on certain media, for example a swap space on disk.
6715 *) Bugfix: client side session caching did not work with external caching,
6716 because the session->cipher setting was not restored when reloading
6717 from the external cache. This problem was masked, when
6718 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6719 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6722 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6723 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6724 [Zeev Lieber <zeev-l@yahoo.com>]
6726 *) Undo an undocumented change introduced in 0.9.6e which caused
6727 repeated calls to OpenSSL_add_all_ciphers() and
6728 OpenSSL_add_all_digests() to be ignored, even after calling
6732 *) Change the default configuration reader to deal with last line not
6733 being properly terminated.
6736 *) Change X509_NAME_cmp() so it applies the special rules on handling
6737 DN values that are of type PrintableString, as well as RDNs of type
6738 emailAddress where the value has the type ia5String.
6739 [stefank@valicert.com via Richard Levitte]
6741 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6742 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6743 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6744 the bitwise-OR of the two for use by the majority of applications
6745 wanting this behaviour, and update the docs. The documented
6746 behaviour and actual behaviour were inconsistent and had been
6747 changing anyway, so this is more a bug-fix than a behavioural
6749 [Geoff Thorpe, diagnosed by Nadav Har'El]
6751 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6752 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6755 *) Fix initialization code race conditions in
6756 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6757 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6758 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6759 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6760 ssl2_get_cipher_by_char(),
6761 ssl3_get_cipher_by_char().
6762 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6764 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6765 the cached sessions are flushed, as the remove_cb() might use ex_data
6766 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6767 (see [openssl.org #212]).
6768 [Geoff Thorpe, Lutz Jaenicke]
6770 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6771 length, instead of the encoding length to d2i_ASN1_OBJECT.
6774 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6776 *) [In 0.9.6g-engine release:]
6777 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6778 [Lynn Gazis <lgazis@rainbow.com>]
6780 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6782 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6783 and get fix the header length calculation.
6784 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6785 Alon Kantor <alonk@checkpoint.com> (and others),
6788 *) Use proper error handling instead of 'assertions' in buffer
6789 overflow checks added in 0.9.6e. This prevents DoS (the
6790 assertions could call abort()).
6791 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6793 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6795 *) Add various sanity checks to asn1_get_length() to reject
6796 the ASN1 length bytes if they exceed sizeof(long), will appear
6797 negative or the content length exceeds the length of the
6799 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6801 *) Fix cipher selection routines: ciphers without encryption had no flags
6802 for the cipher strength set and where therefore not handled correctly
6803 by the selection routines (PR #130).
6806 *) Fix EVP_dsa_sha macro.
6810 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6811 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6812 that was added in OpenSSL 0.9.6d.
6814 As the countermeasure turned out to be incompatible with some
6815 broken SSL implementations, the new option is part of SSL_OP_ALL.
6816 SSL_OP_ALL is usually employed when compatibility with weird SSL
6817 implementations is desired (e.g. '-bugs' option to 's_client' and
6818 's_server'), so the new option is automatically set in many
6822 *) Changes in security patch:
6824 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6825 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6826 Air Force Materiel Command, USAF, under agreement number
6829 *) Add various sanity checks to asn1_get_length() to reject
6830 the ASN1 length bytes if they exceed sizeof(long), will appear
6831 negative or the content length exceeds the length of the
6832 supplied buffer. (CVE-2002-0659)
6833 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6835 *) Assertions for various potential buffer overflows, not known to
6837 [Ben Laurie (CHATS)]
6839 *) Various temporary buffers to hold ASCII versions of integers were
6840 too small for 64 bit platforms. (CVE-2002-0655)
6841 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6843 *) Remote buffer overflow in SSL3 protocol - an attacker could
6844 supply an oversized session ID to a client. (CVE-2002-0656)
6845 [Ben Laurie (CHATS)]
6847 *) Remote buffer overflow in SSL2 protocol - an attacker could
6848 supply an oversized client master key. (CVE-2002-0656)
6849 [Ben Laurie (CHATS)]
6851 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6853 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6854 encoded as NULL) with id-dsa-with-sha1.
6855 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6857 *) Check various X509_...() return values in apps/req.c.
6858 [Nils Larsch <nla@trustcenter.de>]
6860 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6861 an end-of-file condition would erronously be flagged, when the CRLF
6862 was just at the end of a processed block. The bug was discovered when
6863 processing data through a buffering memory BIO handing the data to a
6864 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6865 <ptsekov@syntrex.com> and Nedelcho Stanev.
6868 *) Implement a countermeasure against a vulnerability recently found
6869 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6870 before application data chunks to avoid the use of known IVs
6871 with data potentially chosen by the attacker.
6874 *) Fix length checks in ssl3_get_client_hello().
6877 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6878 to prevent ssl3_read_internal() from incorrectly assuming that
6879 ssl3_read_bytes() found application data while handshake
6880 processing was enabled when in fact s->s3->in_read_app_data was
6881 merely automatically cleared during the initial handshake.
6882 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6884 *) Fix object definitions for Private and Enterprise: they were not
6885 recognized in their shortname (=lowercase) representation. Extend
6886 obj_dat.pl to issue an error when using undefined keywords instead
6887 of silently ignoring the problem (Svenning Sorensen
6888 <sss@sss.dnsalias.net>).
6891 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6892 generators, i.e. generators other than 2 and 5. (Previously, the
6893 code did not properly initialise the 'add' and 'rem' values to
6894 BN_generate_prime().)
6896 In the new general case, we do not insist that 'generator' is
6897 actually a primitive root: This requirement is rather pointless;
6898 a generator of the order-q subgroup is just as good, if not
6902 *) Map new X509 verification errors to alerts. Discovered and submitted by
6903 Tom Wu <tom@arcot.com>.
6906 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6907 returning non-zero before the data has been completely received
6908 when using non-blocking I/O.
6909 [Bodo Moeller; problem pointed out by John Hughes]
6911 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6912 [Ben Laurie, Lutz Jaenicke]
6914 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6915 Yoram Zahavi <YoramZ@gilian.com>).
6918 *) Add information about CygWin 1.3 and on, and preserve proper
6919 configuration for the versions before that.
6920 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6922 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6923 check whether we deal with a copy of a session and do not delete from
6924 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6925 <izhar@checkpoint.com>.
6928 *) Do not store session data into the internal session cache, if it
6929 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6930 flag is set). Proposed by Aslam <aslam@funk.com>.
6933 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6937 *) [In 0.9.6d-engine release:]
6938 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6939 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6941 *) Add the configuration target linux-s390x.
6942 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6944 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6945 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6946 variable as an indication that a ClientHello message has been
6947 received. As the flag value will be lost between multiple
6948 invocations of ssl3_accept when using non-blocking I/O, the
6949 function may not be aware that a handshake has actually taken
6950 place, thus preventing a new session from being added to the
6953 To avoid this problem, we now set s->new_session to 2 instead of
6954 using a local variable.
6955 [Lutz Jaenicke, Bodo Moeller]
6957 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6958 if the SSL_R_LENGTH_MISMATCH error is detected.
6959 [Geoff Thorpe, Bodo Moeller]
6961 *) New 'shared_ldflag' column in Configure platform table.
6964 *) Fix EVP_CIPHER_mode macro.
6965 ["Dan S. Camper" <dan@bti.net>]
6967 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6968 type, we must throw them away by setting rr->length to 0.
6969 [D P Chang <dpc@qualys.com>]
6971 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6973 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6974 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6975 worked incorrectly for those cases where range = 10..._2 and
6976 3*range is two bits longer than range.)
6979 *) Only add signing time to PKCS7 structures if it is not already
6983 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6984 OBJ_ld_ce should be OBJ_id_ce.
6985 Also some ip-pda OIDs in crypto/objects/objects.txt were
6986 incorrect (cf. RFC 3039).
6987 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6989 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6990 returns early because it has nothing to do.
6991 [Andy Schneider <andy.schneider@bjss.co.uk>]
6993 *) [In 0.9.6c-engine release:]
6994 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6995 [Andy Schneider <andy.schneider@bjss.co.uk>]
6997 *) [In 0.9.6c-engine release:]
6998 Add support for Cryptographic Appliance's keyserver technology.
6999 (Use engine 'keyclient')
7000 [Cryptographic Appliances and Geoff Thorpe]
7002 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7003 is called via tools/c89.sh because arguments have to be
7004 rearranged (all '-L' options must appear before the first object
7006 [Richard Shapiro <rshapiro@abinitio.com>]
7008 *) [In 0.9.6c-engine release:]
7009 Add support for Broadcom crypto accelerator cards, backported
7011 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7013 *) [In 0.9.6c-engine release:]
7014 Add support for SureWare crypto accelerator cards from
7015 Baltimore Technologies. (Use engine 'sureware')
7016 [Baltimore Technologies and Mark Cox]
7018 *) [In 0.9.6c-engine release:]
7019 Add support for crypto accelerator cards from Accelerated
7020 Encryption Processing, www.aep.ie. (Use engine 'aep')
7021 [AEP Inc. and Mark Cox]
7023 *) Add a configuration entry for gcc on UnixWare.
7024 [Gary Benson <gbenson@redhat.com>]
7026 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7027 messages are stored in a single piece (fixed-length part and
7028 variable-length part combined) and fix various bugs found on the way.
7031 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7032 instead. BIO_gethostbyname() does not know what timeouts are
7033 appropriate, so entries would stay in cache even when they have
7035 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7037 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7038 faced with a pathologically small ClientHello fragment that does
7039 not contain client_version: Instead of aborting with an error,
7040 simply choose the highest available protocol version (i.e.,
7041 TLS 1.0 unless it is disabled). In practice, ClientHello
7042 messages are never sent like this, but this change gives us
7043 strictly correct behaviour at least for TLS.
7046 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7047 never resets s->method to s->ctx->method when called from within
7048 one of the SSL handshake functions.
7049 [Bodo Moeller; problem pointed out by Niko Baric]
7051 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7052 (sent using the client's version number) if client_version is
7053 smaller than the protocol version in use. Also change
7054 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7055 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7056 the client will at least see that alert.
7059 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7063 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7064 client receives HelloRequest while in a handshake.
7065 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7067 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7068 should end in 'break', not 'goto end' which circumvents various
7069 cleanups done in state SSL_ST_OK. But session related stuff
7070 must be disabled for SSL_ST_OK in the case that we just sent a
7073 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7074 before just sending a HelloRequest.
7075 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7077 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7078 reveal whether illegal block cipher padding was found or a MAC
7079 verification error occured. (Neither SSLerr() codes nor alerts
7080 are directly visible to potential attackers, but the information
7081 may leak via logfiles.)
7083 Similar changes are not required for the SSL 2.0 implementation
7084 because the number of padding bytes is sent in clear for SSL 2.0,
7085 and the extra bytes are just ignored. However ssl/s2_pkt.c
7086 failed to verify that the purported number of padding bytes is in
7090 *) Add OpenUNIX-8 support including shared libraries
7091 (Boyd Lynn Gerber <gerberb@zenez.com>).
7094 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7095 'wristwatch attack' using huge encoding parameters (cf.
7096 James H. Manger's CRYPTO 2001 paper). Note that the
7097 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7098 encoding parameters and hence was not vulnerable.
7101 *) BN_sqr() bug fix.
7102 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7104 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7105 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7106 followed by modular reduction.
7107 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7109 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7110 equivalent based on BN_pseudo_rand() instead of BN_rand().
7113 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7114 This function was broken, as the check for a new client hello message
7115 to handle SGC did not allow these large messages.
7116 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7119 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7122 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7123 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7126 *) Rework the configuration and shared library support for Tru64 Unix.
7127 The configuration part makes use of modern compiler features and
7128 still retains old compiler behavior for those that run older versions
7129 of the OS. The shared library support part includes a variant that
7130 uses the RPATH feature, and is available through the special
7131 configuration target "alpha-cc-rpath", which will never be selected
7133 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7135 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7136 with the same message size as in ssl3_get_certificate_request().
7137 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7138 messages might inadvertently be reject as too long.
7139 [Petr Lampa <lampa@fee.vutbr.cz>]
7141 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7144 *) Modified SSL library such that the verify_callback that has been set
7145 specificly for an SSL object with SSL_set_verify() is actually being
7146 used. Before the change, a verify_callback set with this function was
7147 ignored and the verify_callback() set in the SSL_CTX at the time of
7148 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7149 to allow the necessary settings.
7152 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7153 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7154 done automatically (in contradiction to the requirements of the C
7155 standard). This made problems when used from OpenSSH.
7158 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7159 dh->length and always used
7161 BN_rand_range(priv_key, dh->p).
7163 BN_rand_range() is not necessary for Diffie-Hellman, and this
7164 specific range makes Diffie-Hellman unnecessarily inefficient if
7165 dh->length (recommended exponent length) is much smaller than the
7166 length of dh->p. We could use BN_rand_range() if the order of
7167 the subgroup was stored in the DH structure, but we only have
7172 BN_rand(priv_key, l, ...)
7174 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7180 RSA_eay_public_encrypt
7181 RSA_eay_private_decrypt
7182 RSA_eay_private_encrypt (signing)
7183 RSA_eay_public_decrypt (signature verification)
7185 (default implementations for RSA_public_encrypt,
7186 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7187 always reject numbers >= n.
7190 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7191 to synchronize access to 'locking_thread'. This is necessary on
7192 systems where access to 'locking_thread' (an 'unsigned long'
7193 variable) is not atomic.
7196 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7197 *before* setting the 'crypto_lock_rand' flag. The previous code had
7198 a race condition if 0 is a valid thread ID.
7199 [Travis Vitek <vitek@roguewave.com>]
7201 *) Add support for shared libraries under Irix.
7202 [Albert Chin-A-Young <china@thewrittenword.com>]
7204 *) Add configuration option to build on Linux on both big-endian and
7206 [Ralf Baechle <ralf@uni-koblenz.de>]
7208 *) Add the possibility to create shared libraries on HP-UX.
7211 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7213 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7214 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7215 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7216 PRNG state recovery was possible based on the output of
7217 one PRNG request appropriately sized to gain knowledge on
7218 'md' followed by enough consecutive 1-byte PRNG requests
7219 to traverse all of 'state'.
7221 1. When updating 'md_local' (the current thread's copy of 'md')
7222 during PRNG output generation, hash all of the previous
7223 'md_local' value, not just the half used for PRNG output.
7225 2. Make the number of bytes from 'state' included into the hash
7226 independent from the number of PRNG bytes requested.
7228 The first measure alone would be sufficient to avoid
7229 Markku-Juhani's attack. (Actually it had never occurred
7230 to me that the half of 'md_local' used for chaining was the
7231 half from which PRNG output bytes were taken -- I had always
7232 assumed that the secret half would be used.) The second
7233 measure makes sure that additional data from 'state' is never
7234 mixed into 'md_local' in small portions; this heuristically
7235 further strengthens the PRNG.
7238 *) Fix crypto/bn/asm/mips3.s.
7241 *) When only the key is given to "enc", the IV is undefined. Print out
7242 an error message in this case.
7245 *) Handle special case when X509_NAME is empty in X509 printing routines.
7248 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7249 positive and less than q.
7252 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7253 used: it isn't thread safe and the add_lock_callback should handle
7255 [Paul Rose <Paul.Rose@bridge.com>]
7257 *) Verify that incoming data obeys the block size in
7258 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7262 [Ulf Möller, Bodo Möller]
7264 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7265 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7266 when fixing the server behaviour for backwards-compatible 'client
7267 hello' messages. (Note that the attack is impractical against
7268 SSL 3.0 and TLS 1.0 anyway because length and version checking
7269 means that the probability of guessing a valid ciphertext is
7270 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7273 Before 0.9.5, the countermeasure (hide the error by generating a
7274 random 'decryption result') did not work properly because
7275 ERR_clear_error() was missing, meaning that SSL_get_error() would
7276 detect the supposedly ignored error.
7278 Both problems are now fixed.
7281 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7282 (previously it was 1024).
7285 *) Fix for compatibility mode trust settings: ignore trust settings
7286 unless some valid trust or reject settings are present.
7289 *) Fix for blowfish EVP: its a variable length cipher.
7292 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7293 parameters in DSA public key structures and return an error in the
7294 DSA routines if parameters are absent.
7297 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7298 in the current directory if neither $RANDFILE nor $HOME was set.
7299 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7300 caused some confusion to Windows users who haven't defined $HOME.
7301 Thus RAND_file_name() is changed again: e_os.h can define a
7302 DEFAULT_HOME, which will be used if $HOME is not set.
7303 For Windows, we use "C:"; on other platforms, we still require
7304 environment variables.
7306 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7307 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7308 having multiple threads call RAND_poll() concurrently.
7311 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7312 combination of a flag and a thread ID variable.
7313 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7314 flag), *other* threads can enter ssleay_add_bytes without obeying
7315 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7316 that they do not hold after the first thread unsets add_do_not_lock).
7319 *) Change bctest again: '-x' expressions are not available in all
7323 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7325 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7326 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7328 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7329 the default extension for executables, if any. Also, make the perl
7330 scripts that use symlink() to test if it really exists and use "cp"
7331 if it doesn't. All this made OpenSSL compilable and installable in
7335 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7336 If SEQUENCE is length is indefinite just set c->slen to the total
7337 amount of data available.
7338 [Steve Henson, reported by shige@FreeBSD.org]
7339 [This change does not apply to 0.9.7.]
7341 *) Change bctest to avoid here-documents inside command substitution
7342 (workaround for FreeBSD /bin/sh bug).
7343 For compatibility with Ultrix, avoid shell functions (introduced
7344 in the bctest version that searches along $PATH).
7347 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7348 with des_encrypt() defined on some operating systems, like Solaris
7352 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7353 On the Importance of Eliminating Errors in Cryptographic
7354 Computations, J. Cryptology 14 (2001) 2, 101-119,
7355 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7358 *) MIPS assembler BIGNUM division bug fix.
7361 *) Disabled incorrect Alpha assembler code.
7364 *) Fix PKCS#7 decode routines so they correctly update the length
7365 after reading an EOC for the EXPLICIT tag.
7367 [This change does not apply to 0.9.7.]
7369 *) Fix bug in PKCS#12 key generation routines. This was triggered
7370 if a 3DES key was generated with a 0 initial byte. Include
7371 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7372 (but broken) behaviour.
7375 *) Enhance bctest to search for a working bc along $PATH and print
7377 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7379 *) Fix memory leaks in err.c: free err_data string if necessary;
7380 don't write to the wrong index in ERR_set_error_data.
7383 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7387 *) Replace rdtsc with _emit statements for VC++ version 5.
7388 [Jeremy Cooper <jeremy@baymoo.org>]
7390 *) Make it possible to reuse SSLv2 sessions.
7393 *) In copy_email() check for >= 0 as a return value for
7394 X509_NAME_get_index_by_NID() since 0 is a valid index.
7395 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7397 *) Avoid coredump with unsupported or invalid public keys by checking if
7398 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7399 PKCS7_verify() fails with non detached data.
7402 *) Don't use getenv in library functions when run as setuid/setgid.
7403 New function OPENSSL_issetugid().
7406 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7407 due to incorrect handling of multi-threading:
7409 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7411 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7413 3. Count how many times MemCheck_off() has been called so that
7414 nested use can be treated correctly. This also avoids
7415 inband-signalling in the previous code (which relied on the
7416 assumption that thread ID 0 is impossible).
7419 *) Add "-rand" option also to s_client and s_server.
7422 *) Fix CPU detection on Irix 6.x.
7423 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7424 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7426 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7429 [This change does not apply to 0.9.7.]
7431 *) Use the cached encoding of an X509_NAME structure rather than
7432 copying it. This is apparently the reason for the libsafe "errors"
7433 but the code is actually correct.
7436 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7437 Bleichenbacher's DSA attack.
7438 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7439 to be set and top=0 forces the highest bit to be set; top=-1 is new
7440 and leaves the highest bit random.
7441 [Ulf Moeller, Bodo Moeller]
7443 *) In the NCONF_...-based implementations for CONF_... queries
7444 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7445 a temporary CONF structure with the data component set to NULL
7446 (which gives segmentation faults in lh_retrieve).
7447 Instead, use NULL for the CONF pointer in CONF_get_string and
7448 CONF_get_number (which may use environment variables) and directly
7449 return NULL from CONF_get_section.
7452 *) Fix potential buffer overrun for EBCDIC.
7455 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7456 keyUsage if basicConstraints absent for a CA.
7459 *) Make SMIME_write_PKCS7() write mail header values with a format that
7460 is more generally accepted (no spaces before the semicolon), since
7461 some programs can't parse those values properly otherwise. Also make
7462 sure BIO's that break lines after each write do not create invalid
7466 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7467 macros previously used would not encode an empty SEQUENCE OF
7468 and break the signature.
7470 [This change does not apply to 0.9.7.]
7472 *) Zero the premaster secret after deriving the master secret in
7476 *) Add some EVP_add_digest_alias registrations (as found in
7477 OpenSSL_add_all_digests()) to SSL_library_init()
7478 aka OpenSSL_add_ssl_algorithms(). This provides improved
7479 compatibility with peers using X.509 certificates
7480 with unconventional AlgorithmIdentifier OIDs.
7483 *) Fix for Irix with NO_ASM.
7484 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7486 *) ./config script fixes.
7487 [Ulf Moeller, Richard Levitte]
7489 *) Fix 'openssl passwd -1'.
7492 *) Change PKCS12_key_gen_asc() so it can cope with non null
7493 terminated strings whose length is passed in the passlen
7494 parameter, for example from PEM callbacks. This was done
7495 by adding an extra length parameter to asc2uni().
7496 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7498 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7499 call failed, free the DSA structure.
7502 *) Fix to uni2asc() to cope with zero length Unicode strings.
7503 These are present in some PKCS#12 files.
7506 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7507 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7508 when writing a 32767 byte record.
7509 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7511 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7512 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7514 (RSA objects have a reference count access to which is protected
7515 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7516 so they are meant to be shared between threads.)
7517 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7518 "Reddie, Steven" <Steven.Reddie@ca.com>]
7520 *) Fix a deadlock in CRYPTO_mem_leaks().
7523 *) Use better test patterns in bntest.
7526 *) rand_win.c fix for Borland C.
7529 *) BN_rshift bugfix for n == 0.
7532 *) Add a 'bctest' script that checks for some known 'bc' bugs
7533 so that 'make test' does not abort just because 'bc' is broken.
7536 *) Store verify_result within SSL_SESSION also for client side to
7537 avoid potential security hole. (Re-used sessions on the client side
7538 always resulted in verify_result==X509_V_OK, not using the original
7539 result of the server certificate verification.)
7542 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7543 SSL3_RT_APPLICATION_DATA, return 0.
7544 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7548 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7549 releases, have been re-implemented by renaming the previous
7550 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7551 and ssl3_read_internal, respectively, and adding 'peek' parameters
7552 to them. The new ssl[23]_{read,peek} functions are calls to
7553 ssl[23]_read_internal with the 'peek' flag set appropriately.
7554 A 'peek' parameter has also been added to ssl3_read_bytes, which
7555 does the actual work for ssl3_read_internal.
7558 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7559 the method-specific "init()" handler. Also clean up ex_data after
7560 calling the method-specific "finish()" handler. Previously, this was
7561 happening the other way round.
7564 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7565 The previous value, 12, was not always sufficient for BN_mod_exp().
7568 *) Make sure that shared libraries get the internal name engine with
7569 the full version number and not just 0. This should mark the
7570 shared libraries as not backward compatible. Of course, this should
7571 be changed again when we can guarantee backward binary compatibility.
7574 *) Fix typo in get_cert_by_subject() in by_dir.c
7575 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7577 *) Rework the system to generate shared libraries:
7579 - Make note of the expected extension for the shared libraries and
7580 if there is a need for symbolic links from for example libcrypto.so.0
7581 to libcrypto.so.0.9.7. There is extended info in Configure for
7584 - Make as few rebuilds of the shared libraries as possible.
7586 - Still avoid linking the OpenSSL programs with the shared libraries.
7588 - When installing, install the shared libraries separately from the
7592 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7594 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7595 and not in SSL_clear because the latter is also used by the
7596 accept/connect functions; previously, the settings made by
7597 SSL_set_read_ahead would be lost during the handshake.
7598 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7600 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7601 Previously, it would create entries for disableed algorithms no
7605 *) Added several new manual pages for SSL_* function.
7608 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7610 *) In ssl23_get_client_hello, generate an error message when faced
7611 with an initial SSL 3.0/TLS record that is too small to contain the
7612 first two bytes of the ClientHello message, i.e. client_version.
7613 (Note that this is a pathologic case that probably has never happened
7614 in real life.) The previous approach was to use the version number
7615 from the record header as a substitute; but our protocol choice
7616 should not depend on that one because it is not authenticated
7617 by the Finished messages.
7620 *) More robust randomness gathering functions for Windows.
7621 [Jeffrey Altman <jaltman@columbia.edu>]
7623 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7624 not set then we don't setup the error code for issuer check errors
7625 to avoid possibly overwriting other errors which the callback does
7626 handle. If an application does set the flag then we assume it knows
7627 what it is doing and can handle the new informational codes
7631 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7632 a general "ANY" type, as such it should be able to decode anything
7633 including tagged types. However it didn't check the class so it would
7634 wrongly interpret tagged types in the same way as their universal
7635 counterpart and unknown types were just rejected. Changed so that the
7636 tagged and unknown types are handled in the same way as a SEQUENCE:
7637 that is the encoding is stored intact. There is also a new type
7638 "V_ASN1_OTHER" which is used when the class is not universal, in this
7639 case we have no idea what the actual type is so we just lump them all
7643 *) On VMS, stdout may very well lead to a file that is written to
7644 in a record-oriented fashion. That means that every write() will
7645 write a separate record, which will be read separately by the
7646 programs trying to read from it. This can be very confusing.
7648 The solution is to put a BIO filter in the way that will buffer
7649 text until a linefeed is reached, and then write everything a
7650 line at a time, so every record written will be an actual line,
7651 not chunks of lines and not (usually doesn't happen, but I've
7652 seen it once) several lines in one record. BIO_f_linebuffer() is
7655 Currently, it's a VMS-only method, because that's where it has
7656 been tested well enough.
7659 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7660 it can return incorrect results.
7661 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7662 but it was in 0.9.6-beta[12].)
7665 *) Disable the check for content being present when verifying detached
7666 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7667 include zero length content when signing messages.
7670 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7671 BIO_ctrl (for BIO pairs).
7674 *) Add DSO method for VMS.
7677 *) Bug fix: Montgomery multiplication could produce results with the
7681 *) Add RPM specification openssl.spec and modify it to build three
7682 packages. The default package contains applications, application
7683 documentation and run-time libraries. The devel package contains
7684 include files, static libraries and function documentation. The
7685 doc package contains the contents of the doc directory. The original
7686 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7689 *) Add a large number of documentation files for many SSL routines.
7690 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7692 *) Add a configuration entry for Sony News 4.
7693 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7695 *) Don't set the two most significant bits to one when generating a
7696 random number < q in the DSA library.
7699 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7700 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7701 the underlying transport is blocking) if a handshake took place.
7702 (The default behaviour is needed by applications such as s_client
7703 and s_server that use select() to determine when to use SSL_read;
7704 but for applications that know in advance when to expect data, it
7705 just makes things more complicated.)
7708 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7712 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7713 work better on such systems.
7714 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7716 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7717 Update PKCS12_parse() so it copies the friendlyName and the
7718 keyid to the certificates aux info.
7721 *) Fix bug in PKCS7_verify() which caused an infinite loop
7722 if there was more than one signature.
7723 [Sven Uszpelkat <su@celocom.de>]
7725 *) Major change in util/mkdef.pl to include extra information
7726 about each symbol, as well as presenting variables as well
7727 as functions. This change means that there's n more need
7728 to rebuild the .num files when some algorithms are excluded.
7731 *) Allow the verify time to be set by an application,
7732 rather than always using the current time.
7735 *) Phase 2 verify code reorganisation. The certificate
7736 verify code now looks up an issuer certificate by a
7737 number of criteria: subject name, authority key id
7738 and key usage. It also verifies self signed certificates
7739 by the same criteria. The main comparison function is
7740 X509_check_issued() which performs these checks.
7742 Lot of changes were necessary in order to support this
7743 without completely rewriting the lookup code.
7745 Authority and subject key identifier are now cached.
7747 The LHASH 'certs' is X509_STORE has now been replaced
7748 by a STACK_OF(X509_OBJECT). This is mainly because an
7749 LHASH can't store or retrieve multiple objects with
7750 the same hash value.
7752 As a result various functions (which were all internal
7753 use only) have changed to handle the new X509_STORE
7754 structure. This will break anything that messed round
7755 with X509_STORE internally.
7757 The functions X509_STORE_add_cert() now checks for an
7758 exact match, rather than just subject name.
7760 The X509_STORE API doesn't directly support the retrieval
7761 of multiple certificates matching a given criteria, however
7762 this can be worked round by performing a lookup first
7763 (which will fill the cache with candidate certificates)
7764 and then examining the cache for matches. This is probably
7765 the best we can do without throwing out X509_LOOKUP
7766 entirely (maybe later...).
7768 The X509_VERIFY_CTX structure has been enhanced considerably.
7770 All certificate lookup operations now go via a get_issuer()
7771 callback. Although this currently uses an X509_STORE it
7772 can be replaced by custom lookups. This is a simple way
7773 to bypass the X509_STORE hackery necessary to make this
7774 work and makes it possible to use more efficient techniques
7775 in future. A very simple version which uses a simple
7776 STACK for its trusted certificate store is also provided
7777 using X509_STORE_CTX_trusted_stack().
7779 The verify_cb() and verify() callbacks now have equivalents
7780 in the X509_STORE_CTX structure.
7782 X509_STORE_CTX also has a 'flags' field which can be used
7783 to customise the verify behaviour.
7786 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7787 excludes S/MIME capabilities.
7790 *) When a certificate request is read in keep a copy of the
7791 original encoding of the signed data and use it when outputing
7792 again. Signatures then use the original encoding rather than
7793 a decoded, encoded version which may cause problems if the
7794 request is improperly encoded.
7797 *) For consistency with other BIO_puts implementations, call
7798 buffer_write(b, ...) directly in buffer_puts instead of calling
7801 In BIO_puts, increment b->num_write as in BIO_write.
7802 [Peter.Sylvester@EdelWeb.fr]
7804 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7805 BN_zero, we may not return a BIGNUM with an array consisting of
7809 *) Avoid calling abort() from within the library when problems are
7810 detected, except if preprocessor symbols have been defined
7811 (such as REF_CHECK, BN_DEBUG etc.).
7814 *) New openssl application 'rsautl'. This utility can be
7815 used for low level RSA operations. DER public key
7816 BIO/fp routines also added.
7819 *) New Configure entry and patches for compiling on QNX 4.
7820 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7822 *) A demo state-machine implementation was sponsored by
7823 Nuron (http://www.nuron.com/) and is now available in
7824 demos/state_machine.
7827 *) New options added to the 'dgst' utility for signature
7828 generation and verification.
7831 *) Unrecognized PKCS#7 content types are now handled via a
7832 catch all ASN1_TYPE structure. This allows unsupported
7833 types to be stored as a "blob" and an application can
7834 encode and decode it manually.
7837 *) Fix various signed/unsigned issues to make a_strex.c
7839 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7841 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7842 length if passed a buffer. ASN1_INTEGER_to_BN failed
7843 if passed a NULL BN and its argument was negative.
7844 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7846 *) Modification to PKCS#7 encoding routines to output definite
7847 length encoding. Since currently the whole structures are in
7848 memory there's not real point in using indefinite length
7849 constructed encoding. However if OpenSSL is compiled with
7850 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7853 *) Added BIO_vprintf() and BIO_vsnprintf().
7856 *) Added more prefixes to parse for in the the strings written
7857 through a logging bio, to cover all the levels that are available
7858 through syslog. The prefixes are now:
7860 PANIC, EMERG, EMR => LOG_EMERG
7861 ALERT, ALR => LOG_ALERT
7862 CRIT, CRI => LOG_CRIT
7863 ERROR, ERR => LOG_ERR
7864 WARNING, WARN, WAR => LOG_WARNING
7865 NOTICE, NOTE, NOT => LOG_NOTICE
7866 INFO, INF => LOG_INFO
7867 DEBUG, DBG => LOG_DEBUG
7869 and as before, if none of those prefixes are present at the
7870 beginning of the string, LOG_ERR is chosen.
7872 On Win32, the LOG_* levels are mapped according to this:
7874 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7875 LOG_WARNING => EVENTLOG_WARNING_TYPE
7876 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7880 *) Made it possible to reconfigure with just the configuration
7881 argument "reconf" or "reconfigure". The command line arguments
7882 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7883 and are retrieved from there when reconfiguring.
7887 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7889 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7892 *) The obj_dat.pl script was messing up the sorting of object
7893 names. The reason was that it compared the quoted version
7894 of strings as a result "OCSP" > "OCSP Signing" because
7895 " > SPACE. Changed script to store unquoted versions of
7896 names and add quotes on output. It was also omitting some
7897 names from the lookup table if they were given a default
7898 value (that is if SN is missing it is given the same
7899 value as LN and vice versa), these are now added on the
7900 grounds that if an object has a name we should be able to
7901 look it up. Finally added warning output when duplicate
7902 short or long names are found.
7905 *) Changes needed for Tandem NSK.
7906 [Scott Uroff <scott@xypro.com>]
7908 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7909 RSA_padding_check_SSLv23(), special padding was never detected
7910 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7911 version rollback attacks was not effective.
7913 In s23_clnt.c, don't use special rollback-attack detection padding
7914 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7915 client; similarly, in s23_srvr.c, don't do the rollback check if
7916 SSL 2.0 is the only protocol enabled in the server.
7919 *) Make it possible to get hexdumps of unprintable data with 'openssl
7920 asn1parse'. By implication, the functions ASN1_parse_dump() and
7921 BIO_dump_indent() are added.
7924 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7925 these print out strings and name structures based on various
7926 flags including RFC2253 support and proper handling of
7927 multibyte characters. Added options to the 'x509' utility
7928 to allow the various flags to be set.
7931 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7932 Also change the functions X509_cmp_current_time() and
7933 X509_gmtime_adj() work with an ASN1_TIME structure,
7934 this will enable certificates using GeneralizedTime in validity
7935 dates to be checked.
7938 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7939 negative public key encodings) on by default,
7940 NO_NEG_PUBKEY_BUG can be set to disable it.
7943 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7944 content octets. An i2c_ASN1_OBJECT is unnecessary because
7945 the encoding can be trivially obtained from the structure.
7948 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7949 not read locks (CRYPTO_r_[un]lock).
7952 *) A first attempt at creating official support for shared
7953 libraries through configuration. I've kept it so the
7954 default is static libraries only, and the OpenSSL programs
7955 are always statically linked for now, but there are
7956 preparations for dynamic linking in place.
7957 This has been tested on Linux and Tru64.
7960 *) Randomness polling function for Win9x, as described in:
7961 Peter Gutmann, Software Generation of Practically Strong
7965 *) Fix so PRNG is seeded in req if using an already existing
7969 *) New options to smime application. -inform and -outform
7970 allow alternative formats for the S/MIME message including
7971 PEM and DER. The -content option allows the content to be
7972 specified separately. This should allow things like Netscape
7973 form signing output easier to verify.
7976 *) Fix the ASN1 encoding of tags using the 'long form'.
7979 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7980 STRING types. These convert content octets to and from the
7981 underlying type. The actual tag and length octets are
7982 already assumed to have been read in and checked. These
7983 are needed because all other string types have virtually
7984 identical handling apart from the tag. By having versions
7985 of the ASN1 functions that just operate on content octets
7986 IMPLICIT tagging can be handled properly. It also allows
7987 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7988 and ASN1_INTEGER are identical apart from the tag.
7991 *) Change the handling of OID objects as follows:
7993 - New object identifiers are inserted in objects.txt, following
7994 the syntax given in objects.README.
7995 - objects.pl is used to process obj_mac.num and create a new
7997 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8000 This is currently kind of a hack, and the perl code in objects.pl
8001 isn't very elegant, but it works as I intended. The simplest way
8002 to check that it worked correctly is to look in obj_dat.h and
8003 check the array nid_objs and make sure the objects haven't moved
8004 around (this is important!). Additions are OK, as well as
8005 consistent name changes.
8008 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8011 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8012 The given file adds to whatever has already been seeded into the
8013 random pool through the RANDFILE configuration file option or
8014 environment variable, or the default random state file.
8017 *) mkstack.pl now sorts each macro group into lexical order.
8018 Previously the output order depended on the order the files
8019 appeared in the directory, resulting in needless rewriting
8023 *) Patches to make OpenSSL compile under Win32 again. Mostly
8024 work arounds for the VC++ problem that it treats func() as
8025 func(void). Also stripped out the parts of mkdef.pl that
8026 added extra typesafe functions: these no longer exist.
8029 *) Reorganisation of the stack code. The macros are now all
8030 collected in safestack.h . Each macro is defined in terms of
8031 a "stack macro" of the form SKM_<name>(type, a, b). The
8032 DEBUG_SAFESTACK is now handled in terms of function casts,
8033 this has the advantage of retaining type safety without the
8034 use of additional functions. If DEBUG_SAFESTACK is not defined
8035 then the non typesafe macros are used instead. Also modified the
8036 mkstack.pl script to handle the new form. Needs testing to see
8037 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8038 the default if no major problems. Similar behaviour for ASN1_SET_OF
8039 and PKCS12_STACK_OF.
8042 *) When some versions of IIS use the 'NET' form of private key the
8043 key derivation algorithm is different. Normally MD5(password) is
8044 used as a 128 bit RC4 key. In the modified case
8045 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8046 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8047 as the old Netscape_RSA functions except they have an additional
8048 'sgckey' parameter which uses the modified algorithm. Also added
8049 an -sgckey command line option to the rsa utility. Thanks to
8050 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8051 algorithm to openssl-dev.
8054 *) The evp_local.h macros were using 'c.##kname' which resulted in
8055 invalid expansion on some systems (SCO 5.0.5 for example).
8056 Corrected to 'c.kname'.
8057 [Phillip Porch <root@theporch.com>]
8059 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8060 a STACK of email addresses from a certificate or request, these look
8061 in the subject name and the subject alternative name extensions and
8062 omit any duplicate addresses.
8065 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8066 This makes DSA verification about 2 % faster.
8069 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8070 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8071 plus overhead for 1024 bit moduli).
8072 This makes exponentiations about 0.5 % faster for 1024 bit
8073 exponents (as measured by "openssl speed rsa2048").
8076 *) Rename memory handling macros to avoid conflicts with other
8078 Malloc => OPENSSL_malloc
8079 Malloc_locked => OPENSSL_malloc_locked
8080 Realloc => OPENSSL_realloc
8081 Free => OPENSSL_free
8084 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8085 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8088 *) CygWin32 support.
8089 [John Jarvie <jjarvie@newsguy.com>]
8091 *) The type-safe stack code has been rejigged. It is now only compiled
8092 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8093 by default all type-specific stack functions are "#define"d back to
8094 standard stack functions. This results in more streamlined output
8095 but retains the type-safety checking possibilities of the original
8099 *) The STACK code has been cleaned up, and certain type declarations
8100 that didn't make a lot of sense have been brought in line. This has
8101 also involved a cleanup of sorts in safestack.h to more correctly
8102 map type-safe stack functions onto their plain stack counterparts.
8103 This work has also resulted in a variety of "const"ifications of
8104 lots of the code, especially "_cmp" operations which should normally
8105 be prototyped with "const" parameters anyway.
8108 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8109 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8110 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8111 where all of 'md' is used each time the PRNG is used, but 'state'
8112 is used only indexed by a cyclic counter. As entropy may not be
8113 well distributed from the beginning, 'md' is important as a
8114 chaining variable. However, the output function chains only half
8115 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8116 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8117 in all of 'state' being rewritten, with the new values depending
8118 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8121 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8122 the handshake is continued after ssl_verify_cert_chain();
8123 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8124 can lead to 'unexplainable' connection aborts later.
8125 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8127 *) Major EVP API cipher revision.
8128 Add hooks for extra EVP features. This allows various cipher
8129 parameters to be set in the EVP interface. Support added for variable
8130 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8131 setting of RC2 and RC5 parameters.
8133 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8136 Remove lots of duplicated code from the EVP library. For example *every*
8137 cipher init() function handles the 'iv' in the same way according to the
8138 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8139 for CFB and OFB modes they zero ctx->num.
8141 New functionality allows removal of S/MIME code RC2 hack.
8143 Most of the routines have the same form and so can be declared in terms
8146 By shifting this to the top level EVP_CipherInit() it can be removed from
8147 all individual ciphers. If the cipher wants to handle IVs or keys
8148 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8151 Change lots of functions like EVP_EncryptUpdate() to now return a
8152 value: although software versions of the algorithms cannot fail
8153 any installed hardware versions can.
8156 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8157 this option is set, tolerate broken clients that send the negotiated
8158 protocol version number instead of the requested protocol version
8162 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8163 i.e. non-zero for export ciphersuites, zero otherwise.
8164 Previous versions had this flag inverted, inconsistent with
8165 rsa_tmp_cb (..._TMP_RSA_CB).
8166 [Bodo Moeller; problem reported by Amit Chopra]
8168 *) Add missing DSA library text string. Work around for some IIS
8169 key files with invalid SEQUENCE encoding.
8172 *) Add a document (doc/standards.txt) that list all kinds of standards
8173 and so on that are implemented in OpenSSL.
8176 *) Enhance c_rehash script. Old version would mishandle certificates
8177 with the same subject name hash and wouldn't handle CRLs at all.
8178 Added -fingerprint option to crl utility, to support new c_rehash
8182 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8185 *) Fix for SSL server purpose checking. Server checking was
8186 rejecting certificates which had extended key usage present
8187 but no ssl client purpose.
8188 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8190 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8191 is a little unclear about how a blank password is handled.
8192 Since the password in encoded as a BMPString with terminating
8193 double NULL a zero length password would end up as just the
8194 double NULL. However no password at all is different and is
8195 handled differently in the PKCS#12 key generation code. NS
8196 treats a blank password as zero length. MSIE treats it as no
8197 password on export: but it will try both on import. We now do
8198 the same: PKCS12_parse() tries zero length and no password if
8199 the password is set to "" or NULL (NULL is now a valid password:
8200 it wasn't before) as does the pkcs12 application.
8203 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8204 perror when PEM_read_bio_X509_REQ fails, the error message must
8205 be obtained from the error queue.
8208 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8209 it in ERR_remove_state if appropriate, and change ERR_get_state
8210 accordingly to avoid race conditions (this is necessary because
8211 thread_hash is no longer constant once set).
8214 *) Bugfix for linux-elf makefile.one.
8217 *) RSA_get_default_method() will now cause a default
8218 RSA_METHOD to be chosen if one doesn't exist already.
8219 Previously this was only set during a call to RSA_new()
8220 or RSA_new_method(NULL) meaning it was possible for
8221 RSA_get_default_method() to return NULL.
8224 *) Added native name translation to the existing DSO code
8225 that will convert (if the flag to do so is set) filenames
8226 that are sufficiently small and have no path information
8227 into a canonical native form. Eg. "blah" converted to
8228 "libblah.so" or "blah.dll" etc.
8231 *) New function ERR_error_string_n(e, buf, len) which is like
8232 ERR_error_string(e, buf), but writes at most 'len' bytes
8233 including the 0 terminator. For ERR_error_string_n, 'buf'
8235 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8237 *) CONF library reworked to become more general. A new CONF
8238 configuration file reader "class" is implemented as well as a
8239 new functions (NCONF_*, for "New CONF") to handle it. The now
8240 old CONF_* functions are still there, but are reimplemented to
8241 work in terms of the new functions. Also, a set of functions
8242 to handle the internal storage of the configuration data is
8243 provided to make it easier to write new configuration file
8244 reader "classes" (I can definitely see something reading a
8245 configuration file in XML format, for example), called _CONF_*,
8246 or "the configuration storage API"...
8248 The new configuration file reading functions are:
8250 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8251 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8253 NCONF_default, NCONF_WIN32
8255 NCONF_dump_fp, NCONF_dump_bio
8257 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8258 NCONF_new creates a new CONF object. This works in the same way
8259 as other interfaces in OpenSSL, like the BIO interface.
8260 NCONF_dump_* dump the internal storage of the configuration file,
8261 which is useful for debugging. All other functions take the same
8262 arguments as the old CONF_* functions wth the exception of the
8263 first that must be a `CONF *' instead of a `LHASH *'.
8265 To make it easer to use the new classes with the old CONF_* functions,
8266 the function CONF_set_default_method is provided.
8269 *) Add '-tls1' option to 'openssl ciphers', which was already
8270 mentioned in the documentation but had not been implemented.
8271 (This option is not yet really useful because even the additional
8272 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8275 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8276 OpenSSL-based applications) load shared libraries and bind to
8277 them in a portable way.
8278 [Geoff Thorpe, with contributions from Richard Levitte]
8280 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8282 *) Make sure _lrotl and _lrotr are only used with MSVC.
8284 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8285 (the default implementation of RAND_status).
8287 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8288 to '-clrext' (= clear extensions), as intended and documented.
8289 [Bodo Moeller; inconsistency pointed out by Michael Attili
8292 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8293 was larger than the MD block size.
8294 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8296 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8297 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8298 using the passed key: if the passed key was a private key the result
8299 of X509_print(), for example, would be to print out all the private key
8303 *) des_quad_cksum() byte order bug fix.
8304 [Ulf Möller, using the problem description in krb4-0.9.7, where
8305 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8307 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8309 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8311 *) For easily testing in shell scripts whether some command
8312 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8313 returns with exit code 0 iff no command of the given name is available.
8314 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8315 the output goes to stdout and nothing is printed to stderr.
8316 Additional arguments are always ignored.
8318 Since for each cipher there is a command of the same name,
8319 the 'no-cipher' compilation switches can be tested this way.
8321 ('openssl no-XXX' is not able to detect pseudo-commands such
8322 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8325 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8328 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8329 is set; it will be thrown away anyway because each handshake creates
8331 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8332 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8333 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8334 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8337 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8338 'Q' and 'R' lose their special meanings (quit/renegotiate).
8339 This is part of what -quiet does; unlike -quiet, -ign_eof
8340 does not suppress any output.
8343 *) Add compatibility options to the purpose and trust code. The
8344 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8345 accepts a certificate or CA, this was the previous behaviour,
8346 with all the associated security issues.
8348 X509_TRUST_COMPAT is the old trust behaviour: only and
8349 automatically trust self signed roots in certificate store. A
8350 new trust setting X509_TRUST_DEFAULT is used to specify that
8351 a purpose has no associated trust setting and it should instead
8352 use the value in the default purpose.
8355 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8356 and fix a memory leak.
8359 *) In util/mkerr.pl (which implements 'make errors'), preserve
8360 reason strings from the previous version of the .c file, as
8361 the default to have only downcase letters (and digits) in
8362 automatically generated reasons codes is not always appropriate.
8365 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8366 using strerror. Previously, ERR_reason_error_string() returned
8367 library names as reason strings for SYSerr; but SYSerr is a special
8368 case where small numbers are errno values, not library numbers.
8371 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8372 converts DSA parameters into DH parameters. (When creating parameters,
8373 DSA_generate_parameters is used.)
8376 *) Include 'length' (recommended exponent length) in C code generated
8377 by 'openssl dhparam -C'.
8380 *) The second argument to set_label in perlasm was already being used
8381 so couldn't be used as a "file scope" flag. Moved to third argument
8385 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8386 instead of RAND_bytes for encryption IVs and salts.
8389 *) Include RAND_status() into RAND_METHOD instead of implementing
8390 it only for md_rand.c Otherwise replacing the PRNG by calling
8391 RAND_set_rand_method would be impossible.
8394 *) Don't let DSA_generate_key() enter an infinite loop if the random
8395 number generation fails.
8398 *) New 'rand' application for creating pseudo-random output.
8401 *) Added configuration support for Linux/IA64
8402 [Rolf Haberrecker <rolf@suse.de>]
8404 *) Assembler module support for Mingw32.
8407 *) Shared library support for HPUX (in shlib/).
8408 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8410 *) Shared library support for Solaris gcc.
8411 [Lutz Behnke <behnke@trustcenter.de>]
8413 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8415 *) PKCS7_encrypt() was adding text MIME headers twice because they
8416 were added manually and by SMIME_crlf_copy().
8419 *) In bntest.c don't call BN_rand with zero bits argument.
8420 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8422 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8423 case was implemented. This caused BN_div_recp() to fail occasionally.
8426 *) Add an optional second argument to the set_label() in the perl
8427 assembly language builder. If this argument exists and is set
8428 to 1 it signals that the assembler should use a symbol whose
8429 scope is the entire file, not just the current function. This
8430 is needed with MASM which uses the format label:: for this scope.
8431 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8433 *) Change the ASN1 types so they are typedefs by default. Before
8434 almost all types were #define'd to ASN1_STRING which was causing
8435 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8439 *) Change names of new functions to the new get1/get0 naming
8440 convention: After 'get1', the caller owns a reference count
8441 and has to call ..._free; 'get0' returns a pointer to some
8442 data structure without incrementing reference counters.
8443 (Some of the existing 'get' functions increment a reference
8444 counter, some don't.)
8445 Similarly, 'set1' and 'add1' functions increase reference
8446 counters or duplicate objects.
8449 *) Allow for the possibility of temp RSA key generation failure:
8450 the code used to assume it always worked and crashed on failure.
8453 *) Fix potential buffer overrun problem in BIO_printf().
8454 [Ulf Möller, using public domain code by Patrick Powell; problem
8455 pointed out by David Sacerdote <das33@cornell.edu>]
8457 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8458 RAND_egd() and RAND_status(). In the command line application,
8459 the EGD socket can be specified like a seed file using RANDFILE
8463 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8464 Some CAs (e.g. Verisign) distribute certificates in this form.
8467 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8468 list to exclude them. This means that no special compilation option
8469 is needed to use anonymous DH: it just needs to be included in the
8473 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8474 EVP_MD_type. The old functionality is available in a new macro called
8475 EVP_MD_md(). Change code that uses it and update docs.
8478 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8479 where the 'void *' argument is replaced by a function pointer argument.
8480 Previously 'void *' was abused to point to functions, which works on
8481 many platforms, but is not correct. As these functions are usually
8482 called by macros defined in OpenSSL header files, most source code
8483 should work without changes.
8486 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8487 sections with information on -D... compiler switches used for
8488 compiling the library so that applications can see them. To enable
8489 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8490 must be defined. E.g.,
8491 #define OPENSSL_ALGORITHM_DEFINES
8492 #include <openssl/opensslconf.h>
8493 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8494 [Richard Levitte, Ulf and Bodo Möller]
8496 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8500 *) Change the 'other' type in certificate aux info to a STACK_OF
8501 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8502 the required ASN1 format: arbitrary types determined by an OID.
8505 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8506 argument to 'req'. This is not because the function is newer or
8507 better than others it just uses the work 'NEW' in the certificate
8508 request header lines. Some software needs this.
8511 *) Reorganise password command line arguments: now passwords can be
8512 obtained from various sources. Delete the PEM_cb function and make
8513 it the default behaviour: i.e. if the callback is NULL and the
8514 usrdata argument is not NULL interpret it as a null terminated pass
8515 phrase. If usrdata and the callback are NULL then the pass phrase
8516 is prompted for as usual.
8519 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8520 the support is automatically enabled. The resulting binaries will
8521 autodetect the card and use it if present.
8522 [Ben Laurie and Compaq Inc.]
8524 *) Work around for Netscape hang bug. This sends certificate request
8525 and server done in one record. Since this is perfectly legal in the
8526 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8527 the bugs/SSLv3 entry for more info.
8530 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8533 *) Add -rand argument to smime and pkcs12 applications and read/write
8537 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8540 *) Add command line password options to the remaining applications.
8543 *) Bug fix for BN_div_recp() for numerators with an even number of
8547 *) More tests in bntest.c, and changed test_bn output.
8550 *) ./config recognizes MacOS X now.
8553 *) Bug fix for BN_div() when the first words of num and divsor are
8554 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8557 *) Add support for various broken PKCS#8 formats, and command line
8558 options to produce them.
8561 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8562 get temporary BIGNUMs from a BN_CTX.
8565 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8569 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8570 include a #define from the old name to the new. The original intent
8571 was that statically linked binaries could for example just call
8572 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8573 link with digests. This never worked because SSLeay_add_all_digests()
8574 and SSLeay_add_all_ciphers() were in the same source file so calling
8575 one would link with the other. They are now in separate source files.
8578 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8581 *) Use a less unusual form of the Miller-Rabin primality test (it used
8582 a binary algorithm for exponentiation integrated into the Miller-Rabin
8583 loop, our standard modexp algorithms are faster).
8586 *) Support for the EBCDIC character set completed.
8587 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8589 *) Source code cleanups: use const where appropriate, eliminate casts,
8590 use void * instead of char * in lhash.
8593 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8594 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8595 this the server could overwrite ephemeral keys that the client
8599 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8600 using 50 iterations of the Rabin-Miller test.
8602 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8603 iterations of the Rabin-Miller test as required by the appendix
8604 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8605 As BN_is_prime_fasttest includes trial division, DSA parameter
8606 generation becomes much faster.
8608 This implies a change for the callback functions in DSA_is_prime
8609 and DSA_generate_parameters: The callback function is called once
8610 for each positive witness in the Rabin-Miller test, not just
8611 occasionally in the inner loop; and the parameters to the
8612 callback function now provide an iteration count for the outer
8613 loop rather than for the current invocation of the inner loop.
8614 DSA_generate_parameters additionally can call the callback
8615 function with an 'iteration count' of -1, meaning that a
8616 candidate has passed the trial division test (when q is generated
8617 from an application-provided seed, trial division is skipped).
8620 *) New function BN_is_prime_fasttest that optionally does trial
8621 division before starting the Rabin-Miller test and has
8622 an additional BN_CTX * argument (whereas BN_is_prime always
8623 has to allocate at least one BN_CTX).
8624 'callback(1, -1, cb_arg)' is called when a number has passed the
8625 trial division stage.
8628 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8632 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8635 *) New function BN_pseudo_rand().
8638 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8639 bignum version of BN_from_montgomery() with the working code from
8640 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8644 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8645 made it impossible to use the same SSL_SESSION data structure in
8646 SSL2 clients in multiple threads.
8649 *) The return value of RAND_load_file() no longer counts bytes obtained
8650 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8651 to seed the PRNG (previously an explicit byte count was required).
8652 [Ulf Möller, Bodo Möller]
8654 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8655 used (char *) instead of (void *) and had casts all over the place.
8658 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8661 *) Retain source code compatibility for BN_prime_checks macro:
8662 BN_is_prime(..., BN_prime_checks, ...) now uses
8663 BN_prime_checks_for_size to determine the appropriate number of
8664 Rabin-Miller iterations.
8667 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8668 DH_CHECK_P_NOT_SAFE_PRIME.
8669 (Check if this is true? OpenPGP calls them "strong".)
8672 *) Merge the functionality of "dh" and "gendh" programs into a new program
8673 "dhparam". The old programs are retained for now but will handle DH keys
8674 (instead of parameters) in future.
8677 *) Make the ciphers, s_server and s_client programs check the return values
8678 when a new cipher list is set.
8681 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8682 ciphers. Before when the 56bit ciphers were enabled the sorting was
8685 The syntax for the cipher sorting has been extended to support sorting by
8686 cipher-strength (using the strength_bits hard coded in the tables).
8687 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8689 Fix a bug in the cipher-command parser: when supplying a cipher command
8690 string with an "undefined" symbol (neither command nor alphanumeric
8691 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8692 an error is flagged.
8694 Due to the strength-sorting extension, the code of the
8695 ssl_create_cipher_list() function was completely rearranged. I hope that
8696 the readability was also increased :-)
8697 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8699 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8700 for the first serial number and places 2 in the serial number file. This
8701 avoids problems when the root CA is created with serial number zero and
8702 the first user certificate has the same issuer name and serial number
8706 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8707 the new code. Add documentation for this stuff.
8710 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8711 X509_*() to X509at_*() on the grounds that they don't handle X509
8712 structures and behave in an analogous way to the X509v3 functions:
8713 they shouldn't be called directly but wrapper functions should be used
8716 So we also now have some wrapper functions that call the X509at functions
8717 when passed certificate requests. (TO DO: similar things can be done with
8718 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8719 things. Some of these need some d2i or i2d and print functionality
8720 because they handle more complex structures.)
8723 *) Add missing #ifndefs that caused missing symbols when building libssl
8724 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8725 NO_RSA in ssl/s2*.c.
8726 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8728 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8729 has a return value which indicates the quality of the random data
8730 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8731 error queue. New function RAND_pseudo_bytes() generates output that is
8732 guaranteed to be unique but not unpredictable. RAND_add is like
8733 RAND_seed, but takes an extra argument for an entropy estimate
8734 (RAND_seed always assumes full entropy).
8737 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8738 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8739 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8740 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8741 false-positive rate of at most 2^-80 for random input.
8744 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8747 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8748 in the 0.9.5 release), this returns the chain
8749 from an X509_CTX structure with a dup of the stack and all
8750 the X509 reference counts upped: so the stack will exist
8751 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8754 Also make SSL_SESSION_print() print out the verify return
8758 *) Add manpage for the pkcs12 command. Also change the default
8759 behaviour so MAC iteration counts are used unless the new
8760 -nomaciter option is used. This improves file security and
8761 only older versions of MSIE (4.0 for example) need it.
8764 *) Honor the no-xxx Configure options when creating .DEF files.
8767 *) Add PKCS#10 attributes to field table: challengePassword,
8768 unstructuredName and unstructuredAddress. These are taken from
8769 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8770 international characters are used.
8772 More changes to X509_ATTRIBUTE code: allow the setting of types
8773 based on strings. Remove the 'loc' parameter when adding
8774 attributes because these will be a SET OF encoding which is sorted
8778 *) Initial changes to the 'req' utility to allow request generation
8779 automation. This will allow an application to just generate a template
8780 file containing all the field values and have req construct the
8783 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8784 used all over the place including certificate requests and PKCS#7
8785 structures. They are currently handled manually where necessary with
8786 some primitive wrappers for PKCS#7. The new functions behave in a
8787 manner analogous to the X509 extension functions: they allow
8788 attributes to be looked up by NID and added.
8790 Later something similar to the X509V3 code would be desirable to
8791 automatically handle the encoding, decoding and printing of the
8792 more complex types. The string types like challengePassword can
8793 be handled by the string table functions.
8795 Also modified the multi byte string table handling. Now there is
8796 a 'global mask' which masks out certain types. The table itself
8797 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8798 is useful when for example there is only one permissible type
8799 (as in countryName) and using the mask might result in no valid
8803 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8804 SSL_get_peer_finished to allow applications to obtain the latest
8805 Finished messages sent to the peer or expected from the peer,
8806 respectively. (SSL_get_peer_finished is usually the Finished message
8807 actually received from the peer, otherwise the protocol will be aborted.)
8809 As the Finished message are message digests of the complete handshake
8810 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8811 be used for external authentication procedures when the authentication
8812 provided by SSL/TLS is not desired or is not enough.
8815 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8816 the host supports BWX extension and if Compaq C is present on the
8817 $PATH. Just exploiting of the BWX extension results in 20-30%
8818 performance kick for some algorithms, e.g. DES and RC4 to mention
8819 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8823 *) Add support for MS "fast SGC". This is arguably a violation of the
8824 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8825 weak crypto and after checking the certificate is SGC a second one
8826 with strong crypto. MS SGC stops the first handshake after receiving
8827 the server certificate message and sends a second client hello. Since
8828 a server will typically do all the time consuming operations before
8829 expecting any further messages from the client (server key exchange
8830 is the most expensive) there is little difference between the two.
8832 To get OpenSSL to support MS SGC we have to permit a second client
8833 hello message after we have sent server done. In addition we have to
8834 reset the MAC if we do get this second client hello.
8837 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8838 if a DER encoded private key is RSA or DSA traditional format. Changed
8839 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8840 format DER encoded private key. Newer code should use PKCS#8 format which
8841 has the key type encoded in the ASN1 structure. Added DER private key
8842 support to pkcs8 application.
8845 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8846 ciphersuites has been selected (as required by the SSL 3/TLS 1
8847 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8848 is set, we interpret this as a request to violate the specification
8849 (the worst that can happen is a handshake failure, and 'correct'
8850 behaviour would result in a handshake failure anyway).
8853 *) In SSL_CTX_add_session, take into account that there might be multiple
8854 SSL_SESSION structures with the same session ID (e.g. when two threads
8855 concurrently obtain them from an external cache).
8856 The internal cache can handle only one SSL_SESSION with a given ID,
8857 so if there's a conflict, we now throw out the old one to achieve
8861 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8862 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8863 some routines that use cipher OIDs: some ciphers do not have OIDs
8864 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8868 *) Simplify the trust setting structure and code. Now we just have
8869 two sequences of OIDs for trusted and rejected settings. These will
8870 typically have values the same as the extended key usage extension
8871 and any application specific purposes.
8873 The trust checking code now has a default behaviour: it will just
8874 check for an object with the same NID as the passed id. Functions can
8875 be provided to override either the default behaviour or the behaviour
8876 for a given id. SSL client, server and email already have functions
8877 in place for compatibility: they check the NID and also return "trusted"
8878 if the certificate is self signed.
8881 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8882 traditional format into an EVP_PKEY structure.
8885 *) Add a password callback function PEM_cb() which either prompts for
8886 a password if usr_data is NULL or otherwise assumes it is a null
8887 terminated password. Allow passwords to be passed on command line
8888 environment or config files in a few more utilities.
8891 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8892 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8893 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8894 Update documentation.
8897 *) Support for ASN1 "NULL" type. This could be handled before by using
8898 ASN1_TYPE but there wasn't any function that would try to read a NULL
8899 and produce an error if it couldn't. For compatibility we also have
8900 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8901 don't allocate anything because they don't need to.
8904 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8906 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8908 *) Rebuild of the memory allocation routines used by OpenSSL code and
8909 possibly others as well. The purpose is to make an interface that
8910 provide hooks so anyone can build a separate set of allocation and
8911 deallocation routines to be used by OpenSSL, for example memory
8912 pool implementations, or something else, which was previously hard
8913 since Malloc(), Realloc() and Free() were defined as macros having
8914 the values malloc, realloc and free, respectively (except for Win32
8915 compilations). The same is provided for memory debugging code.
8916 OpenSSL already comes with functionality to find memory leaks, but
8917 this gives people a chance to debug other memory problems.
8919 With these changes, a new set of functions and macros have appeared:
8921 CRYPTO_set_mem_debug_functions() [F]
8922 CRYPTO_get_mem_debug_functions() [F]
8923 CRYPTO_dbg_set_options() [F]
8924 CRYPTO_dbg_get_options() [F]
8925 CRYPTO_malloc_debug_init() [M]
8927 The memory debug functions are NULL by default, unless the library
8928 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8929 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8930 gives the standard debugging functions that come with OpenSSL) or
8931 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8932 provided by the library user) must be used. When the standard
8933 debugging functions are used, CRYPTO_dbg_set_options can be used to
8934 request additional information:
8935 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8936 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8938 Also, things like CRYPTO_set_mem_functions will always give the
8939 expected result (the new set of functions is used for allocation
8940 and deallocation) at all times, regardless of platform and compiler
8943 To finish it up, some functions that were never use in any other
8944 way than through macros have a new API and new semantic:
8947 CRYPTO_dbg_realloc()
8950 All macros of value have retained their old syntax.
8951 [Richard Levitte and Bodo Moeller]
8953 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8954 ordering of SMIMECapabilities wasn't in "strength order" and there
8955 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8959 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8960 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8961 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8963 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8964 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8965 functionality to handle multipart/signed properly) and a utility
8966 called 'smime' to call all this stuff. This is based on code I
8967 originally wrote for Celo who have kindly allowed it to be
8968 included in OpenSSL.
8971 *) Add variants des_set_key_checked and des_set_key_unchecked of
8972 des_set_key (aka des_key_sched). Global variable des_check_key
8973 decides which of these is called by des_set_key; this way
8974 des_check_key behaves as it always did, but applications and
8975 the library itself, which was buggy for des_check_key == 1,
8976 have a cleaner way to pick the version they need.
8979 *) New function PKCS12_newpass() which changes the password of a
8983 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8984 dynamic mix. In both cases the ids can be used as an index into the
8985 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8986 functions so they accept a list of the field values and the
8987 application doesn't need to directly manipulate the X509_TRUST
8991 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8995 *) Modify the way the V3 extension code looks up extensions. This now
8996 works in a similar way to the object code: we have some "standard"
8997 extensions in a static table which is searched with OBJ_bsearch()
8998 and the application can add dynamic ones if needed. The file
8999 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9000 updated whenever a new extension is added to the core code and kept
9001 in ext_nid order. There is a simple program 'tabtest.c' which checks
9002 this. New extensions are not added too often so this file can readily
9003 be maintained manually.
9005 There are two big advantages in doing things this way. The extensions
9006 can be looked up immediately and no longer need to be "added" using
9007 X509V3_add_standard_extensions(): this function now does nothing.
9008 [Side note: I get *lots* of email saying the extension code doesn't
9009 work because people forget to call this function]
9010 Also no dynamic allocation is done unless new extensions are added:
9011 so if we don't add custom extensions there is no need to call
9012 X509V3_EXT_cleanup().
9015 *) Modify enc utility's salting as follows: make salting the default. Add a
9016 magic header, so unsalted files fail gracefully instead of just decrypting
9017 to garbage. This is because not salting is a big security hole, so people
9018 should be discouraged from doing it.
9021 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9022 digest to be passed on the command line but it only used this
9023 parameter when signing a certificate. Modified so all relevant
9024 operations are affected by the digest parameter including the
9025 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9026 DSA key was used because it didn't fix the digest.
9029 *) Initial certificate chain verify code. Currently tests the untrusted
9030 certificates for consistency with the verify purpose (which is set
9031 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9033 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9034 this is because it will reject chains with invalid extensions whereas
9035 every previous version of OpenSSL and SSLeay made no checks at all.
9037 Trust code: checks the root CA for the relevant trust settings. Trust
9038 settings have an initial value consistent with the verify purpose: e.g.
9039 if the verify purpose is for SSL client use it expects the CA to be
9040 trusted for SSL client use. However the default value can be changed to
9041 permit custom trust settings: one example of this would be to only trust
9042 certificates from a specific "secure" set of CAs.
9044 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9045 which should be used for version portability: especially since the
9046 verify structure is likely to change more often now.
9048 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9049 to set them. If not set then assume SSL clients will verify SSL servers
9052 Two new options to the verify program: -untrusted allows a set of
9053 untrusted certificates to be passed in and -purpose which sets the
9054 intended purpose of the certificate. If a purpose is set then the
9055 new chain verify code is used to check extension consistency.
9058 *) Support for the authority information access extension.
9061 *) Modify RSA and DSA PEM read routines to transparently handle
9062 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9063 public keys in a format compatible with certificate
9064 SubjectPublicKeyInfo structures. Unfortunately there were already
9065 functions called *_PublicKey_* which used various odd formats so
9066 these are retained for compatibility: however the DSA variants were
9067 never in a public release so they have been deleted. Changed dsa/rsa
9068 utilities to handle the new format: note no releases ever handled public
9069 keys so we should be OK.
9071 The primary motivation for this change is to avoid the same fiasco
9072 that dogs private keys: there are several incompatible private key
9073 formats some of which are standard and some OpenSSL specific and
9074 require various evil hacks to allow partial transparent handling and
9075 even then it doesn't work with DER formats. Given the option anything
9076 other than PKCS#8 should be dumped: but the other formats have to
9077 stay in the name of compatibility.
9079 With public keys and the benefit of hindsight one standard format
9080 is used which works with EVP_PKEY, RSA or DSA structures: though
9081 it clearly returns an error if you try to read the wrong kind of key.
9083 Added a -pubkey option to the 'x509' utility to output the public key.
9084 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9085 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9086 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9087 that do the same as the EVP_PKEY_assign_*() except they up the
9088 reference count of the added key (they don't "swallow" the
9092 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9093 CRLs would fail if the file contained no certificates or no CRLs:
9094 added a new function to read in both types and return the number
9095 read: this means that if none are read it will be an error. The
9096 DER versions of the certificate and CRL reader would always fail
9097 because it isn't possible to mix certificates and CRLs in DER format
9098 without choking one or the other routine. Changed this to just read
9099 a certificate: this is the best we can do. Also modified the code
9100 in apps/verify.c to take notice of return codes: it was previously
9101 attempting to read in certificates from NULL pointers and ignoring
9102 any errors: this is one reason why the cert and CRL reader seemed
9103 to work. It doesn't check return codes from the default certificate
9104 routines: these may well fail if the certificates aren't installed.
9107 *) Code to support otherName option in GeneralName.
9110 *) First update to verify code. Change the verify utility
9111 so it warns if it is passed a self signed certificate:
9112 for consistency with the normal behaviour. X509_verify
9113 has been modified to it will now verify a self signed
9114 certificate if *exactly* the same certificate appears
9115 in the store: it was previously impossible to trust a
9116 single self signed certificate. This means that:
9117 openssl verify ss.pem
9118 now gives a warning about a self signed certificate but
9119 openssl verify -CAfile ss.pem ss.pem
9123 *) For servers, store verify_result in SSL_SESSION data structure
9124 (and add it to external session representation).
9125 This is needed when client certificate verifications fails,
9126 but an application-provided verification callback (set by
9127 SSL_CTX_set_cert_verify_callback) allows accepting the session
9128 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9129 but returns 1): When the session is reused, we have to set
9130 ssl->verify_result to the appropriate error code to avoid
9132 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9134 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9135 case in PKCS7_dataInit() where the signed PKCS7 structure
9136 didn't contain any existing data because it was being created.
9137 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9139 *) Add a salt to the key derivation routines in enc.c. This
9140 forms the first 8 bytes of the encrypted file. Also add a
9141 -S option to allow a salt to be input on the command line.
9144 *) New function X509_cmp(). Oddly enough there wasn't a function
9145 to compare two certificates. We do this by working out the SHA1
9146 hash and comparing that. X509_cmp() will be needed by the trust
9150 *) SSL_get1_session() is like SSL_get_session(), but increments
9151 the reference count in the SSL_SESSION returned.
9152 [Geoff Thorpe <geoff@eu.c2.net>]
9154 *) Fix for 'req': it was adding a null to request attributes.
9155 Also change the X509_LOOKUP and X509_INFO code to handle
9156 certificate auxiliary information.
9159 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9163 *) Add the possibility to add extra information to the memory leak
9164 detecting output, to form tracebacks, showing from where each
9165 allocation was originated: CRYPTO_push_info("constant string") adds
9166 the string plus current file name and line number to a per-thread
9167 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9168 is like calling CYRPTO_pop_info() until the stack is empty.
9169 Also updated memory leak detection code to be multi-thread-safe.
9172 *) Add options -text and -noout to pkcs7 utility and delete the
9173 encryption options which never did anything. Update docs.
9176 *) Add options to some of the utilities to allow the pass phrase
9177 to be included on either the command line (not recommended on
9178 OSes like Unix) or read from the environment. Update the
9179 manpages and fix a few bugs.
9182 *) Add a few manpages for some of the openssl commands.
9185 *) Fix the -revoke option in ca. It was freeing up memory twice,
9186 leaking and not finding already revoked certificates.
9189 *) Extensive changes to support certificate auxiliary information.
9190 This involves the use of X509_CERT_AUX structure and X509_AUX
9191 functions. An X509_AUX function such as PEM_read_X509_AUX()
9192 can still read in a certificate file in the usual way but it
9193 will also read in any additional "auxiliary information". By
9194 doing things this way a fair degree of compatibility can be
9195 retained: existing certificates can have this information added
9196 using the new 'x509' options.
9198 Current auxiliary information includes an "alias" and some trust
9199 settings. The trust settings will ultimately be used in enhanced
9200 certificate chain verification routines: currently a certificate
9201 can only be trusted if it is self signed and then it is trusted
9205 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9206 The problem was that one of the replacement routines had not been working
9207 since SSLeay releases. For now the offending routine has been replaced
9208 with non-optimised assembler. Even so, this now gives around 95%
9209 performance improvement for 1024 bit RSA signs.
9212 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9213 handling. Most clients have the effective key size in bits equal to
9214 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9215 A few however don't do this and instead use the size of the decrypted key
9216 to determine the RC2 key length and the AlgorithmIdentifier to determine
9217 the effective key length. In this case the effective key length can still
9218 be 40 bits but the key length can be 168 bits for example. This is fixed
9219 by manually forcing an RC2 key into the EVP_PKEY structure because the
9220 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9221 the key length and effective key length are equal.
9224 *) Add a bunch of functions that should simplify the creation of
9225 X509_NAME structures. Now you should be able to do:
9226 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9227 and have it automatically work out the correct field type and fill in
9228 the structures. The more adventurous can try:
9229 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9230 and it will (hopefully) work out the correct multibyte encoding.
9233 *) Change the 'req' utility to use the new field handling and multibyte
9234 copy routines. Before the DN field creation was handled in an ad hoc
9235 way in req, ca, and x509 which was rather broken and didn't support
9236 BMPStrings or UTF8Strings. Since some software doesn't implement
9237 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9238 using the dirstring_type option. See the new comment in the default
9239 openssl.cnf for more info.
9242 *) Make crypto/rand/md_rand.c more robust:
9243 - Assure unique random numbers after fork().
9244 - Make sure that concurrent threads access the global counter and
9245 md serializably so that we never lose entropy in them
9246 or use exactly the same state in multiple threads.
9247 Access to the large state is not always serializable because
9248 the additional locking could be a performance killer, and
9249 md should be large enough anyway.
9252 *) New file apps/app_rand.c with commonly needed functionality
9253 for handling the random seed file.
9255 Use the random seed file in some applications that previously did not:
9257 dsaparam -genkey (which also ignored its '-rand' option),
9260 x509 (when signing).
9261 Except on systems with /dev/urandom, it is crucial to have a random
9262 seed file at least for key creation, DSA signing, and for DH exchanges;
9263 for RSA signatures we could do without one.
9265 gendh and gendsa (unlike genrsa) used to read only the first byte
9266 of each file listed in the '-rand' option. The function as previously
9267 found in genrsa is now in app_rand.c and is used by all programs
9268 that support '-rand'.
9271 *) In RAND_write_file, use mode 0600 for creating files;
9272 don't just chmod when it may be too late.
9275 *) Report an error from X509_STORE_load_locations
9276 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9279 *) New function ASN1_mbstring_copy() this copies a string in either
9280 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9281 into an ASN1_STRING type. A mask of permissible types is passed
9282 and it chooses the "minimal" type to use or an error if not type
9286 *) Add function equivalents to the various macros in asn1.h. The old
9287 macros are retained with an M_ prefix. Code inside the library can
9288 use the M_ macros. External code (including the openssl utility)
9289 should *NOT* in order to be "shared library friendly".
9292 *) Add various functions that can check a certificate's extensions
9293 to see if it usable for various purposes such as SSL client,
9294 server or S/MIME and CAs of these types. This is currently
9295 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9296 verification. Also added a -purpose flag to x509 utility to
9297 print out all the purposes.
9300 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9304 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9305 for, obtain and decode and extension and obtain its critical flag.
9306 This allows all the necessary extension code to be handled in a
9307 single function call.
9310 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9311 platforms. See crypto/rc4/rc4_enc.c for further details.
9314 *) New -noout option to asn1parse. This causes no output to be produced
9315 its main use is when combined with -strparse and -out to extract data
9316 from a file (which may not be in ASN.1 format).
9319 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9320 when producing the local key id.
9321 [Richard Levitte <levitte@stacken.kth.se>]
9323 *) New option -dhparam in s_server. This allows a DH parameter file to be
9324 stated explicitly. If it is not stated then it tries the first server
9325 certificate file. The previous behaviour hard coded the filename
9329 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9330 a public key to be input or output. For example:
9331 openssl rsa -in key.pem -pubout -out pubkey.pem
9332 Also added necessary DSA public key functions to handle this.
9335 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9336 in the message. This was handled by allowing
9337 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9338 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9340 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9341 to the end of the strings whereas this didn't. This would cause problems
9342 if strings read with d2i_ASN1_bytes() were later modified.
9343 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9345 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9346 data and it contains EOF it will end up returning an error. This is
9347 caused by input 46 bytes long. The cause is due to the way base64
9348 BIOs find the start of base64 encoded data. They do this by trying a
9349 trial decode on each line until they find one that works. When they
9350 do a flag is set and it starts again knowing it can pass all the
9351 data directly through the decoder. Unfortunately it doesn't reset
9352 the context it uses. This means that if EOF is reached an attempt
9353 is made to pass two EOFs through the context and this causes the
9354 resulting error. This can also cause other problems as well. As is
9355 usual with these problems it takes *ages* to find and the fix is
9356 trivial: move one line.
9357 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9359 *) Ugly workaround to get s_client and s_server working under Windows. The
9360 old code wouldn't work because it needed to select() on sockets and the
9361 tty (for keypresses and to see if data could be written). Win32 only
9362 supports select() on sockets so we select() with a 1s timeout on the
9363 sockets and then see if any characters are waiting to be read, if none
9364 are present then we retry, we also assume we can always write data to
9365 the tty. This isn't nice because the code then blocks until we've
9366 received a complete line of data and it is effectively polling the
9367 keyboard at 1s intervals: however it's quite a bit better than not
9368 working at all :-) A dedicated Windows application might handle this
9369 with an event loop for example.
9372 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9373 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9374 will be called when RSA_sign() and RSA_verify() are used. This is useful
9375 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9376 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9377 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9378 This necessitated the support of an extra signature type NID_md5_sha1
9379 for SSL signatures and modifications to the SSL library to use it instead
9380 of calling RSA_public_decrypt() and RSA_private_encrypt().
9383 *) Add new -verify -CAfile and -CApath options to the crl program, these
9384 will lookup a CRL issuers certificate and verify the signature in a
9385 similar way to the verify program. Tidy up the crl program so it
9386 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9387 less strict. It will now permit CRL extensions even if it is not
9388 a V2 CRL: this will allow it to tolerate some broken CRLs.
9391 *) Initialize all non-automatic variables each time one of the openssl
9392 sub-programs is started (this is necessary as they may be started
9393 multiple times from the "OpenSSL>" prompt).
9394 [Lennart Bang, Bodo Moeller]
9396 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9397 removing all other RSA functionality (this is what NO_RSA does). This
9398 is so (for example) those in the US can disable those operations covered
9399 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9403 *) Non-copying interface to BIO pairs.
9404 (still largely untested)
9407 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9408 ASCII string. This was handled independently in various places before.
9411 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9412 UTF8 strings a character at a time.
9415 *) Use client_version from client hello to select the protocol
9416 (s23_srvr.c) and for RSA client key exchange verification
9417 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9420 *) Add various utility functions to handle SPKACs, these were previously
9421 handled by poking round in the structure internals. Added new function
9422 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9423 print, verify and generate SPKACs. Based on an original idea from
9424 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9427 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9430 *) Allow the config file extension section to be overwritten on the
9431 command line. Based on an original idea from Massimiliano Pala
9432 <madwolf@comune.modena.it>. The new option is called -extensions
9433 and can be applied to ca, req and x509. Also -reqexts to override
9434 the request extensions in req and -crlexts to override the crl extensions
9438 *) Add new feature to the SPKAC handling in ca. Now you can include
9439 the same field multiple times by preceding it by "XXXX." for example:
9442 this is the same syntax as used in the req config file.
9445 *) Allow certificate extensions to be added to certificate requests. These
9446 are specified in a 'req_extensions' option of the req section of the
9447 config file. They can be printed out with the -text option to req but
9448 are otherwise ignored at present.
9451 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9452 data read consists of only the final block it would not decrypted because
9453 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9454 A misplaced 'break' also meant the decrypted final block might not be
9455 copied until the next read.
9458 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9459 a few extra parameters to the DH structure: these will be useful if
9460 for example we want the value of 'q' or implement X9.42 DH.
9463 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9464 provides hooks that allow the default DSA functions or functions on a
9465 "per key" basis to be replaced. This allows hardware acceleration and
9466 hardware key storage to be handled without major modification to the
9467 library. Also added low level modexp hooks and CRYPTO_EX structure and
9468 associated functions.
9471 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9472 as "read only": it can't be written to and the buffer it points to will
9473 not be freed. Reading from a read only BIO is much more efficient than
9474 a normal memory BIO. This was added because there are several times when
9475 an area of memory needs to be read from a BIO. The previous method was
9476 to create a memory BIO and write the data to it, this results in two
9477 copies of the data and an O(n^2) reading algorithm. There is a new
9478 function BIO_new_mem_buf() which creates a read only memory BIO from
9479 an area of memory. Also modified the PKCS#7 routines to use read only
9483 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9484 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9485 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9486 but a retry condition occured while trying to read the rest.
9489 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9490 NID_pkcs7_encrypted by default: this was wrong since this should almost
9491 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9492 the encrypted data type: this is a more sensible place to put it and it
9493 allows the PKCS#12 code to be tidied up that duplicated this
9497 *) Changed obj_dat.pl script so it takes its input and output files on
9498 the command line. This should avoid shell escape redirection problems
9502 *) Initial support for certificate extension requests, these are included
9503 in things like Xenroll certificate requests. Included functions to allow
9504 extensions to be obtained and added.
9507 *) -crlf option to s_client and s_server for sending newlines as
9508 CRLF (as required by many protocols).
9511 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9513 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9514 [Ralf S. Engelschall]
9516 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9517 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9519 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9523 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9524 DH parameters/keys (q is lost during that conversion, but the resulting
9525 DH parameters contain its length).
9527 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9528 much faster than DH_generate_parameters (which creates parameters
9529 where p = 2*q + 1), and also the smaller q makes DH computations
9530 much more efficient (160-bit exponentiation instead of 1024-bit
9531 exponentiation); so this provides a convenient way to support DHE
9532 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9533 utter importance to use
9534 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9536 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9537 when such DH parameters are used, because otherwise small subgroup
9538 attacks may become possible!
9541 *) Avoid memory leak in i2d_DHparams.
9544 *) Allow the -k option to be used more than once in the enc program:
9545 this allows the same encrypted message to be read by multiple recipients.
9548 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9549 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9550 it will always use the numerical form of the OID, even if it has a short
9554 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9555 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9556 otherwise bn_mod_exp was called. In the case of hardware keys for example
9557 no private key components need be present and it might store extra data
9558 in the RSA structure, which cannot be accessed from bn_mod_exp.
9559 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9560 private key operations.
9563 *) Added support for SPARC Linux.
9566 *) pem_password_cb function type incompatibly changed from
9567 typedef int pem_password_cb(char *buf, int size, int rwflag);
9569 ....(char *buf, int size, int rwflag, void *userdata);
9570 so that applications can pass data to their callbacks:
9571 The PEM[_ASN1]_{read,write}... functions and macros now take an
9572 additional void * argument, which is just handed through whenever
9573 the password callback is called.
9574 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9576 New function SSL_CTX_set_default_passwd_cb_userdata.
9578 Compatibility note: As many C implementations push function arguments
9579 onto the stack in reverse order, the new library version is likely to
9580 interoperate with programs that have been compiled with the old
9581 pem_password_cb definition (PEM_whatever takes some data that
9582 happens to be on the stack as its last argument, and the callback
9583 just ignores this garbage); but there is no guarantee whatsoever that
9586 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9587 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9588 problems not only on Windows, but also on some Unix platforms.
9589 To avoid problematic command lines, these definitions are now in an
9590 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9591 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9594 *) MIPS III/IV assembler module is reimplemented.
9597 *) More DES library cleanups: remove references to srand/rand and
9598 delete an unused file.
9601 *) Add support for the the free Netwide assembler (NASM) under Win32,
9602 since not many people have MASM (ml) and it can be hard to obtain.
9603 This is currently experimental but it seems to work OK and pass all
9604 the tests. Check out INSTALL.W32 for info.
9607 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9608 without temporary keys kept an extra copy of the server key,
9609 and connections with temporary keys did not free everything in case
9613 *) New function RSA_check_key and new openssl rsa option -check
9614 for verifying the consistency of RSA keys.
9615 [Ulf Moeller, Bodo Moeller]
9617 *) Various changes to make Win32 compile work:
9618 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9619 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9620 comparison" warnings.
9621 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9624 *) Add a debugging option to PKCS#5 v2 key generation function: when
9625 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9626 derived keys are printed to stderr.
9629 *) Copy the flags in ASN1_STRING_dup().
9630 [Roman E. Pavlov <pre@mo.msk.ru>]
9632 *) The x509 application mishandled signing requests containing DSA
9633 keys when the signing key was also DSA and the parameters didn't match.
9635 It was supposed to omit the parameters when they matched the signing key:
9636 the verifying software was then supposed to automatically use the CA's
9637 parameters if they were absent from the end user certificate.
9639 Omitting parameters is no longer recommended. The test was also
9640 the wrong way round! This was probably due to unusual behaviour in
9641 EVP_cmp_parameters() which returns 1 if the parameters match.
9642 This meant that parameters were omitted when they *didn't* match and
9643 the certificate was useless. Certificates signed with 'ca' didn't have
9645 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9647 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9648 The interface is as follows:
9649 Applications can use
9650 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9651 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9652 "off" is now the default.
9653 The library internally uses
9654 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9655 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9656 to disable memory-checking temporarily.
9658 Some inconsistent states that previously were possible (and were
9659 even the default) are now avoided.
9661 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9662 with each memory chunk allocated; this is occasionally more helpful
9663 than just having a counter.
9665 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9667 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9671 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9672 which largely parallels "options", but is for changing API behaviour,
9673 whereas "options" are about protocol behaviour.
9674 Initial "mode" flags are:
9676 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9677 a single record has been written.
9678 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9679 retries use the same buffer location.
9680 (But all of the contents must be
9684 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9687 *) Fix problems with no-hmac etc.
9688 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9690 *) New functions RSA_get_default_method(), RSA_set_method() and
9691 RSA_get_method(). These allows replacement of RSA_METHODs without having
9692 to mess around with the internals of an RSA structure.
9695 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9696 Also really enable memory leak checks in openssl.c and in some
9698 [Chad C. Mulligan, Bodo Moeller]
9700 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9701 up the length of negative integers. This has now been simplified to just
9702 store the length when it is first determined and use it later, rather
9703 than trying to keep track of where data is copied and updating it to
9705 [Steve Henson, reported by Brien Wheeler
9706 <bwheeler@authentica-security.com>]
9708 *) Add a new function PKCS7_signatureVerify. This allows the verification
9709 of a PKCS#7 signature but with the signing certificate passed to the
9710 function itself. This contrasts with PKCS7_dataVerify which assumes the
9711 certificate is present in the PKCS#7 structure. This isn't always the
9712 case: certificates can be omitted from a PKCS#7 structure and be
9713 distributed by "out of band" means (such as a certificate database).
9716 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9717 function prototypes in pem.h, also change util/mkdef.pl to add the
9718 necessary function names.
9721 *) mk1mf.pl (used by Windows builds) did not properly read the
9722 options set by Configure in the top level Makefile, and Configure
9723 was not even able to write more than one option correctly.
9724 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9727 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9728 file to be loaded from a BIO or FILE pointer. The BIO version will
9729 for example allow memory BIOs to contain config info.
9732 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9733 Whoever hopes to achieve shared-library compatibility across versions
9734 must use this, not the compile-time macro.
9735 (Exercise 0.9.4: Which is the minimum library version required by
9737 Note: All this applies only to multi-threaded programs, others don't
9741 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9742 through a BIO pair triggered the default case, i.e.
9743 SSLerr(...,SSL_R_UNKNOWN_STATE).
9746 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9747 can use the SSL library even if none of the specific BIOs is
9751 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9752 for the encoded length.
9753 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9755 *) Add initial documentation of the X509V3 functions.
9758 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9759 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9760 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9761 secure PKCS#8 private key format with a high iteration count.
9764 *) Fix determination of Perl interpreter: A perl or perl5
9765 _directory_ in $PATH was also accepted as the interpreter.
9766 [Ralf S. Engelschall]
9768 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9769 wrong with it but it was very old and did things like calling
9770 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9774 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9775 to use the new extension code.
9778 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9779 with macros. This should make it easier to change their form, add extra
9780 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9784 *) Add to configuration table a new entry that can specify an alternative
9785 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9786 according to Mark Crispin <MRC@Panda.COM>.
9790 *) DES CBC did not update the IV. Weird.
9793 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9794 Changing the behaviour of the former might break existing programs --
9795 where IV updating is needed, des_ncbc_encrypt can be used.
9798 *) When bntest is run from "make test" it drives bc to check its
9799 calculations, as well as internally checking them. If an internal check
9800 fails, it needs to cause bc to give a non-zero result or make test carries
9801 on without noticing the failure. Fixed.
9804 *) DES library cleanups.
9807 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9808 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9809 ciphers. NOTE: although the key derivation function has been verified
9810 against some published test vectors it has not been extensively tested
9811 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9815 *) Instead of "mkdir -p", which is not fully portable, use new
9816 Perl script "util/mkdir-p.pl".
9819 *) Rewrite the way password based encryption (PBE) is handled. It used to
9820 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9821 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9822 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9823 the 'parameter' field of the AlgorithmIdentifier is passed to the
9824 underlying key generation function so it must do its own ASN1 parsing.
9825 This has also changed the EVP_PBE_CipherInit() function which now has a
9826 'parameter' argument instead of literal salt and iteration count values
9827 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9830 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9831 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9832 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9833 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9834 value was just used as a "magic string" and not used directly its
9835 value doesn't matter.
9838 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9842 *) "linux-sparc64" configuration (ultrapenguin).
9843 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9844 "linux-sparc" configuration.
9845 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9847 *) config now generates no-xxx options for missing ciphers.
9850 *) Support the EBCDIC character set (work in progress).
9851 File ebcdic.c not yet included because it has a different license.
9852 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9854 *) Support BS2000/OSD-POSIX.
9855 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9857 *) Make callbacks for key generation use void * instead of char *.
9860 *) Make S/MIME samples compile (not yet tested).
9863 *) Additional typesafe stacks.
9866 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9870 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9872 *) New configuration variant "sco5-gcc".
9874 *) Updated some demos.
9875 [Sean O Riordain, Wade Scholine]
9877 *) Add missing BIO_free at exit of pkcs12 application.
9880 *) Fix memory leak in conf.c.
9883 *) Updates for Win32 to assembler version of MD5.
9886 *) Set #! path to perl in apps/der_chop to where we found it
9887 instead of using a fixed path.
9890 *) SHA library changes for irix64-mips4-cc.
9893 *) Improvements for VMS support.
9897 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9899 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9900 This also avoids the problems with SC4.2 and unpatched SC5.
9901 [Andy Polyakov <appro@fy.chalmers.se>]
9903 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9904 These are required because of the typesafe stack would otherwise break
9905 existing code. If old code used a structure member which used to be STACK
9906 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9907 sk_num or sk_value it would produce an error because the num, data members
9908 are not present in STACK_OF. Now it just produces a warning. sk_set
9909 replaces the old method of assigning a value to sk_value
9910 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9911 that does this will no longer work (and should use sk_set instead) but
9912 this could be regarded as a "questionable" behaviour anyway.
9915 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9916 correctly handle encrypted S/MIME data.
9919 *) Change type of various DES function arguments from des_cblock
9920 (which means, in function argument declarations, pointer to char)
9921 to des_cblock * (meaning pointer to array with 8 char elements),
9922 which allows the compiler to do more typechecking; it was like
9923 that back in SSLeay, but with lots of ugly casts.
9925 Introduce new type const_des_cblock.
9928 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9929 problems: find RecipientInfo structure that matches recipient certificate
9930 and initialise the ASN1 structures properly based on passed cipher.
9933 *) Belatedly make the BN tests actually check the results.
9936 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9937 to and from BNs: it was completely broken. New compilation option
9938 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9939 key elements as negative integers.
9942 *) Reorganize and speed up MD5.
9943 [Andy Polyakov <appro@fy.chalmers.se>]
9946 [Richard Levitte <richard@levitte.org>]
9948 *) New option -out to asn1parse to allow the parsed structure to be
9949 output to a file. This is most useful when combined with the -strparse
9950 option to examine the output of things like OCTET STRINGS.
9953 *) Make SSL library a little more fool-proof by not requiring any longer
9954 that SSL_set_{accept,connect}_state be called before
9955 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9956 in many applications because usually everything *appeared* to work as
9957 intended anyway -- now it really works as intended).
9960 *) Move openssl.cnf out of lib/.
9963 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9964 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9965 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9966 [Ralf S. Engelschall]
9968 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9969 handle PKCS#7 enveloped data properly.
9970 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9972 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9973 copying pointers. The cert_st handling is changed by this in
9974 various ways (and thus what used to be known as ctx->default_cert
9975 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9976 any longer when s->cert does not give us what we need).
9977 ssl_cert_instantiate becomes obsolete by this change.
9978 As soon as we've got the new code right (possibly it already is?),
9979 we have solved a couple of bugs of the earlier code where s->cert
9980 was used as if it could not have been shared with other SSL structures.
9982 Note that using the SSL API in certain dirty ways now will result
9983 in different behaviour than observed with earlier library versions:
9984 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9985 does not influence s as it used to.
9987 In order to clean up things more thoroughly, inside SSL_SESSION
9988 we don't use CERT any longer, but a new structure SESS_CERT
9989 that holds per-session data (if available); currently, this is
9990 the peer's certificate chain and, for clients, the server's certificate
9991 and temporary key. CERT holds only those values that can have
9992 meaningful defaults in an SSL_CTX.
9995 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9996 from the internal representation. Various PKCS#7 fixes: remove some
9997 evil casts and set the enc_dig_alg field properly based on the signing
10001 *) Allow PKCS#12 password to be set from the command line or the
10002 environment. Let 'ca' get its config file name from the environment
10003 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10007 *) Allow certificate policies extension to use an IA5STRING for the
10008 organization field. This is contrary to the PKIX definition but
10009 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10013 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10014 without disallowing inline assembler and the like for non-pedantic builds.
10017 *) Support Borland C++ builder.
10018 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10020 *) Support Mingw32.
10023 *) SHA-1 cleanups and performance enhancements.
10024 [Andy Polyakov <appro@fy.chalmers.se>]
10026 *) Sparc v8plus assembler for the bignum library.
10027 [Andy Polyakov <appro@fy.chalmers.se>]
10029 *) Accept any -xxx and +xxx compiler options in Configure.
10032 *) Update HPUX configuration.
10035 *) Add missing sk_<type>_unshift() function to safestack.h
10036 [Ralf S. Engelschall]
10038 *) New function SSL_CTX_use_certificate_chain_file that sets the
10039 "extra_cert"s in addition to the certificate. (This makes sense
10040 only for "PEM" format files, as chains as a whole are not
10044 *) Support verify_depth from the SSL API.
10045 x509_vfy.c had what can be considered an off-by-one-error:
10046 Its depth (which was not part of the external interface)
10047 was actually counting the number of certificates in a chain;
10048 now it really counts the depth.
10051 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10052 instead of X509err, which often resulted in confusing error
10053 messages since the error codes are not globally unique
10054 (e.g. an alleged error in ssl3_accept when a certificate
10055 didn't match the private key).
10057 *) New function SSL_CTX_set_session_id_context that allows to set a default
10058 value (so that you don't need SSL_set_session_id_context for each
10059 connection using the SSL_CTX).
10062 *) OAEP decoding bug fix.
10065 *) Support INSTALL_PREFIX for package builders, as proposed by
10069 *) New Configure options "threads" and "no-threads". For systems
10070 where the proper compiler options are known (currently Solaris
10071 and Linux), "threads" is the default.
10074 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10077 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10078 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10079 such as /usr/local/bin.
10082 *) "make linux-shared" to build shared libraries.
10083 [Niels Poppe <niels@netbox.org>]
10085 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10088 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10089 extension adding in x509 utility.
10092 *) Remove NOPROTO sections and error code comments.
10095 *) Partial rewrite of the DEF file generator to now parse the ANSI
10099 *) New Configure options --prefix=DIR and --openssldir=DIR.
10102 *) Complete rewrite of the error code script(s). It is all now handled
10103 by one script at the top level which handles error code gathering,
10104 header rewriting and C source file generation. It should be much better
10105 than the old method: it now uses a modified version of Ulf's parser to
10106 read the ANSI prototypes in all header files (thus the old K&R definitions
10107 aren't needed for error creation any more) and do a better job of
10108 translating function codes into names. The old 'ASN1 error code imbedded
10109 in a comment' is no longer necessary and it doesn't use .err files which
10110 have now been deleted. Also the error code call doesn't have to appear all
10111 on one line (which resulted in some large lines...).
10114 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10117 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10118 0 (which usually indicates a closed connection), but continue reading.
10121 *) Fix some race conditions.
10124 *) Add support for CRL distribution points extension. Add Certificate
10125 Policies and CRL distribution points documentation.
10128 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10131 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10132 8 of keying material. Merlin has also confirmed interop with this fix
10133 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10134 [Merlin Hughes <merlin@baltimore.ie>]
10136 *) Fix lots of warnings.
10137 [Richard Levitte <levitte@stacken.kth.se>]
10139 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10140 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10141 [Richard Levitte <levitte@stacken.kth.se>]
10143 *) Fix problems with sizeof(long) == 8.
10144 [Andy Polyakov <appro@fy.chalmers.se>]
10146 *) Change functions to ANSI C.
10149 *) Fix typos in error codes.
10150 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10152 *) Remove defunct assembler files from Configure.
10155 *) SPARC v8 assembler BIGNUM implementation.
10156 [Andy Polyakov <appro@fy.chalmers.se>]
10158 *) Support for Certificate Policies extension: both print and set.
10159 Various additions to support the r2i method this uses.
10162 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10163 return a const string when you are expecting an allocated buffer.
10166 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10167 types DirectoryString and DisplayText.
10170 *) Add code to allow r2i extensions to access the configuration database,
10171 add an LHASH database driver and add several ctx helper functions.
10174 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10175 fail when they extended the size of a BIGNUM.
10178 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10179 support typesafe stack.
10182 *) Fix typo in SSL_[gs]et_options().
10183 [Nils Frostberg <nils@medcom.se>]
10185 *) Delete various functions and files that belonged to the (now obsolete)
10186 old X509V3 handling code.
10189 *) New Configure option "rsaref".
10192 *) Don't auto-generate pem.h.
10195 *) Introduce type-safe ASN.1 SETs.
10198 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10199 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10201 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10202 that links with OpenSSL (well at least cause lots of warnings), but fear
10203 not: the conversion is trivial, and it eliminates loads of evil casts. A
10204 few STACKed things have been converted already. Feel free to convert more.
10205 In the fullness of time, I'll do away with the STACK type altogether.
10208 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10209 specified in <certfile> by updating the entry in the index.txt file.
10210 This way one no longer has to edit the index.txt file manually for
10211 revoking a certificate. The -revoke option does the gory details now.
10212 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10214 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10215 `-text' option at all and this way the `-noout -text' combination was
10216 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10217 [Ralf S. Engelschall]
10219 *) Make sure a corresponding plain text error message exists for the
10220 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10221 verify callback function determined that a certificate was revoked.
10222 [Ralf S. Engelschall]
10224 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10225 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10226 all available cipers including rc5, which was forgotten until now.
10227 In order to let the testing shell script know which algorithms
10228 are available, a new (up to now undocumented) command
10229 "openssl list-cipher-commands" is used.
10232 *) Bugfix: s_client occasionally would sleep in select() when
10233 it should have checked SSL_pending() first.
10236 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10237 the raw DSA values prior to ASN.1 encoding.
10240 *) Tweaks to Configure
10241 [Niels Poppe <niels@netbox.org>]
10243 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10247 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10250 *) New config option to avoid instructions that are illegal on the 80386.
10251 The default code is faster, but requires at least a 486.
10254 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10255 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10256 same as SSL2_VERSION anyway.
10259 *) New "-showcerts" option for s_client.
10262 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10263 application. Various cleanups and fixes.
10266 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10267 modify error routines to work internally. Add error codes and PBE init
10268 to library startup routines.
10271 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10272 packing functions to asn1 and evp. Changed function names and error
10273 codes along the way.
10276 *) PKCS12 integration: and so it begins... First of several patches to
10277 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10278 objects to objects.h
10281 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10282 and display support for Thawte strong extranet extension.
10285 *) Add LinuxPPC support.
10286 [Jeff Dubrule <igor@pobox.org>]
10288 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10289 bn_div_words in alpha.s.
10290 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10292 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10293 OAEP isn't supported when OpenSSL is built with RSAref.
10294 [Ulf Moeller <ulf@fitug.de>]
10296 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10297 so they no longer are missing under -DNOPROTO.
10298 [Soren S. Jorvang <soren@t.dk>]
10301 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10303 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10304 doesn't work when the session is reused. Coming soon!
10307 *) Fix a security hole, that allows sessions to be reused in the wrong
10308 context thus bypassing client cert protection! All software that uses
10309 client certs and session caches in multiple contexts NEEDS PATCHING to
10310 allow session reuse! A fuller solution is in the works.
10311 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10313 *) Some more source tree cleanups (removed obsolete files
10314 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10315 permission on "config" script to be executable) and a fix for the INSTALL
10317 [Ulf Moeller <ulf@fitug.de>]
10319 *) Remove some legacy and erroneous uses of malloc, free instead of
10321 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10323 *) Make rsa_oaep_test return non-zero on error.
10324 [Ulf Moeller <ulf@fitug.de>]
10326 *) Add support for native Solaris shared libraries. Configure
10327 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10328 if someone would make that last step automatic.
10329 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10331 *) ctx_size was not built with the right compiler during "make links". Fixed.
10334 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10335 except NULL ciphers". This means the default cipher list will no longer
10336 enable NULL ciphers. They need to be specifically enabled e.g. with
10337 the string "DEFAULT:eNULL".
10340 *) Fix to RSA private encryption routines: if p < q then it would
10341 occasionally produce an invalid result. This will only happen with
10342 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10345 *) Be less restrictive and allow also `perl util/perlpath.pl
10346 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10347 because this way one can also use an interpreter named `perl5' (which is
10348 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10349 installed as `perl').
10350 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10352 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10353 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10355 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10356 advapi32.lib to Win32 build and change the pem test comparison
10357 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10358 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10359 and crypto/des/ede_cbcm_enc.c.
10362 *) DES quad checksum was broken on big-endian architectures. Fixed.
10365 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10366 Win32 test batch file so it (might) work again. The Win32 test batch file
10367 is horrible: I feel ill....
10370 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10371 in e_os.h. Audit of header files to check ANSI and non ANSI
10372 sections: 10 functions were absent from non ANSI section and not exported
10373 from Windows DLLs. Fixed up libeay.num for new functions.
10376 *) Make `openssl version' output lines consistent.
10377 [Ralf S. Engelschall]
10379 *) Fix Win32 symbol export lists for BIO functions: Added
10380 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10381 to ms/libeay{16,32}.def.
10382 [Ralf S. Engelschall]
10384 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10385 fine under Unix and passes some trivial tests I've now added. But the
10386 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10387 added to make sure no one expects that this stuff really works in the
10388 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10389 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10391 [Ralf S. Engelschall]
10393 *) Fix the generation of two part addresses in perl.
10394 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10396 *) Add config entry for Linux on MIPS.
10397 [John Tobey <jtobey@channel1.com>]
10399 *) Make links whenever Configure is run, unless we are on Windoze.
10402 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10403 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10407 *) Add a useful kludge to allow package maintainers to specify compiler and
10408 other platforms details on the command line without having to patch the
10409 Configure script everytime: One now can use ``perl Configure
10410 <id>:<details>'', i.e. platform ids are allowed to have details appended
10411 to them (separated by colons). This is treated as there would be a static
10412 pre-configured entry in Configure's %table under key <id> with value
10413 <details> and ``perl Configure <id>'' is called. So, when you want to
10414 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10415 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10416 now, which overrides the FreeBSD-elf entry on-the-fly.
10417 [Ralf S. Engelschall]
10419 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10422 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10423 on the `perl Configure ...' command line. This way one can compile
10424 OpenSSL libraries with Position Independent Code (PIC) which is needed
10425 for linking it into DSOs.
10426 [Ralf S. Engelschall]
10428 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10432 *) Cleaned up the LICENSE document: The official contact for any license
10433 questions now is the OpenSSL core team under openssl-core@openssl.org.
10434 And add a paragraph about the dual-license situation to make sure people
10435 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10436 to the OpenSSL toolkit.
10437 [Ralf S. Engelschall]
10439 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10440 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10441 Additionally cleaned up the `make links' target: Remove unnecessary
10442 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10443 to speed processing and no longer clutter the display with confusing
10444 stuff. Instead only the actually done links are displayed.
10445 [Ralf S. Engelschall]
10447 *) Permit null encryption ciphersuites, used for authentication only. It used
10448 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10449 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10453 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10454 signed attributes when verifying signatures (this would break them),
10455 the detached data encoding was wrong and public keys obtained using
10456 X509_get_pubkey() weren't freed.
10459 *) Add text documentation for the BUFFER functions. Also added a work around
10460 to a Win95 console bug. This was triggered by the password read stuff: the
10461 last character typed gets carried over to the next fread(). If you were
10462 generating a new cert request using 'req' for example then the last
10463 character of the passphrase would be CR which would then enter the first
10467 *) Added the new `Includes OpenSSL Cryptography Software' button as
10468 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10469 button and can be used by applications based on OpenSSL to show the
10470 relationship to the OpenSSL project.
10471 [Ralf S. Engelschall]
10473 *) Remove confusing variables in function signatures in files
10474 ssl/ssl_lib.c and ssl/ssl.h.
10475 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10477 *) Don't install bss_file.c under PREFIX/include/
10478 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10480 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10481 functions that return function pointers and has support for NT specific
10482 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10483 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10484 unsigned to signed types: this was killing the Win32 compile.
10487 *) Add new certificate file to stack functions,
10488 SSL_add_dir_cert_subjects_to_stack() and
10489 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10490 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10491 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10492 This means that Apache-SSL and similar packages don't have to mess around
10493 to add as many CAs as they want to the preferred list.
10496 *) Experiment with doxygen documentation. Currently only partially applied to
10498 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10499 openssl.doxy as the configuration file.
10502 *) Get rid of remaining C++-style comments which strict C compilers hate.
10503 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10505 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10506 compiled in by default: it has problems with large keys.
10509 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10510 DH private keys and/or callback functions which directly correspond to
10511 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10512 is needed for applications which have to configure certificates on a
10513 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10515 For the RSA certificate situation is makes no difference, but
10516 for the DSA certificate situation this fixes the "no shared cipher"
10517 problem where the OpenSSL cipher selection procedure failed because the
10518 temporary keys were not overtaken from the context and the API provided
10519 no way to reconfigure them.
10520 The new functions now let applications reconfigure the stuff and they
10521 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10522 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10523 non-public-API function ssl_cert_instantiate() is used as a helper
10524 function and also to reduce code redundancy inside ssl_rsa.c.
10525 [Ralf S. Engelschall]
10527 *) Move s_server -dcert and -dkey options out of the undocumented feature
10528 area because they are useful for the DSA situation and should be
10529 recognized by the users.
10530 [Ralf S. Engelschall]
10532 *) Fix the cipher decision scheme for export ciphers: the export bits are
10533 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10534 SSL_EXP_MASK. So, the original variable has to be used instead of the
10535 already masked variable.
10536 [Richard Levitte <levitte@stacken.kth.se>]
10538 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10539 [Richard Levitte <levitte@stacken.kth.se>]
10541 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10542 from `int' to `unsigned int' because it's a length and initialized by
10543 EVP_DigestFinal() which expects an `unsigned int *'.
10544 [Richard Levitte <levitte@stacken.kth.se>]
10546 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10547 script. Instead use the usual Shell->Perl transition trick.
10548 [Ralf S. Engelschall]
10550 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10551 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10552 -noout -modulus' as it's already the case for `openssl rsa -noout
10553 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10554 currently the public key is printed (a decision which was already done by
10555 `openssl dsa -modulus' in the past) which serves a similar purpose.
10556 Additionally the NO_RSA no longer completely removes the whole -modulus
10557 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10559 [Ralf S. Engelschall]
10561 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10562 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10563 [Arne Ansper <arne@ats.cyber.ee>]
10565 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10566 to be added. Now both 'req' and 'ca' can use new objects defined in the
10570 *) Add cool BIO that does syslog (or event log on NT).
10571 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10573 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10574 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10575 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10576 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10579 *) Add preliminary config info for new extension code.
10582 *) Make RSA_NO_PADDING really use no padding.
10583 [Ulf Moeller <ulf@fitug.de>]
10585 *) Generate errors when private/public key check is done.
10588 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10589 for some CRL extensions and new objects added.
10592 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10593 key usage extension and fuller support for authority key id.
10596 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10597 padding method for RSA, which is recommended for new applications in PKCS
10598 #1 v2.0 (RFC 2437, October 1998).
10599 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10600 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10601 against Bleichbacher's attack on RSA.
10602 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10605 *) Updates to the new SSL compression code
10606 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10608 *) Fix so that the version number in the master secret, when passed
10609 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10610 (because the server will not accept higher), that the version number
10611 is 0x03,0x01, not 0x03,0x00
10612 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10614 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10615 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10616 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10619 *) Support for RAW extensions where an arbitrary extension can be
10620 created by including its DER encoding. See apps/openssl.cnf for
10624 *) Make sure latest Perl versions don't interpret some generated C array
10625 code as Perl array code in the crypto/err/err_genc.pl script.
10626 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10628 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10629 not many people have the assembler. Various Win32 compilation fixes and
10630 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10631 build instructions.
10634 *) Modify configure script 'Configure' to automatically create crypto/date.h
10635 file under Win32 and also build pem.h from pem.org. New script
10636 util/mkfiles.pl to create the MINFO file on environments that can't do a
10637 'make files': perl util/mkfiles.pl >MINFO should work.
10640 *) Major rework of DES function declarations, in the pursuit of correctness
10641 and purity. As a result, many evil casts evaporated, and some weirdness,
10642 too. You may find this causes warnings in your code. Zapping your evil
10643 casts will probably fix them. Mostly.
10646 *) Fix for a typo in asn1.h. Bug fix to object creation script
10647 obj_dat.pl. It considered a zero in an object definition to mean
10648 "end of object": none of the objects in objects.h have any zeros
10649 so it wasn't spotted.
10650 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10652 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10653 Masking (CBCM). In the absence of test vectors, the best I have been able
10654 to do is check that the decrypt undoes the encrypt, so far. Send me test
10655 vectors if you have them.
10658 *) Correct calculation of key length for export ciphers (too much space was
10659 allocated for null ciphers). This has not been tested!
10662 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10663 message is now correct (it understands "crypto" and "ssl" on its
10664 command line). There is also now an "update" option. This will update
10665 the util/ssleay.num and util/libeay.num files with any new functions.
10667 perl util/mkdef.pl crypto ssl update
10668 it will update them.
10671 *) Overhauled the Perl interface (perl/*):
10672 - ported BN stuff to OpenSSL's different BN library
10673 - made the perl/ source tree CVS-aware
10674 - renamed the package from SSLeay to OpenSSL (the files still contain
10675 their history because I've copied them in the repository)
10676 - removed obsolete files (the test scripts will be replaced
10677 by better Test::Harness variants in the future)
10678 [Ralf S. Engelschall]
10680 *) First cut for a very conservative source tree cleanup:
10681 1. merge various obsolete readme texts into doc/ssleay.txt
10682 where we collect the old documents and readme texts.
10683 2. remove the first part of files where I'm already sure that we no
10684 longer need them because of three reasons: either they are just temporary
10685 files which were left by Eric or they are preserved original files where
10686 I've verified that the diff is also available in the CVS via "cvs diff
10687 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10688 the crypto/md/ stuff).
10689 [Ralf S. Engelschall]
10691 *) More extension code. Incomplete support for subject and issuer alt
10692 name, issuer and authority key id. Change the i2v function parameters
10693 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10694 what that's for :-) Fix to ASN1 macro which messed up
10695 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10698 *) Preliminary support for ENUMERATED type. This is largely copied from the
10702 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10703 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10705 *) Make sure `make rehash' target really finds the `openssl' program.
10706 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10708 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10709 like to hear about it if this slows down other processors.
10712 *) Add CygWin32 platform information to Configure script.
10713 [Alan Batie <batie@aahz.jf.intel.com>]
10715 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10716 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10718 *) New program nseq to manipulate netscape certificate sequences
10721 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10725 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10726 but the BN code had some problems that would cause failures when
10727 doing certificate verification and some other functions.
10728 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10730 *) Add ASN1 and PEM code to support netscape certificate sequences.
10733 *) Add ASN1 and PEM code to support netscape certificate sequences.
10736 *) Add several PKIX and private extended key usage OIDs.
10739 *) Modify the 'ca' program to handle the new extension code. Modify
10740 openssl.cnf for new extension format, add comments.
10743 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10744 and add a sample to openssl.cnf so req -x509 now adds appropriate
10748 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10749 error code, add initial support to X509_print() and x509 application.
10752 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10753 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10754 stuff is currently isolated and isn't even compiled yet.
10757 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10758 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10759 Removed the versions check from X509 routines when loading extensions:
10760 this allows certain broken certificates that don't set the version
10761 properly to be processed.
10764 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10765 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10766 can still be regenerated with "make depend".
10769 *) Spelling mistake in C version of CAST-128.
10770 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10772 *) Changes to the error generation code. The perl script err-code.pl
10773 now reads in the old error codes and retains the old numbers, only
10774 adding new ones if necessary. It also only changes the .err files if new
10775 codes are added. The makefiles have been modified to only insert errors
10776 when needed (to avoid needlessly modifying header files). This is done
10777 by only inserting errors if the .err file is newer than the auto generated
10778 C file. To rebuild all the error codes from scratch (the old behaviour)
10779 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10780 or delete all the .err files.
10783 *) CAST-128 was incorrectly implemented for short keys. The C version has
10784 been fixed, but is untested. The assembler versions are also fixed, but
10785 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10786 to regenerate it if needed.
10787 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10788 Hagino <itojun@kame.net>]
10790 *) File was opened incorrectly in randfile.c.
10791 [Ulf Möller <ulf@fitug.de>]
10793 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10794 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10795 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10796 al: it's just almost always a UTCTime. Note this patch adds new error
10797 codes so do a "make errors" if there are problems.
10800 *) Correct Linux 1 recognition in config.
10801 [Ulf Möller <ulf@fitug.de>]
10803 *) Remove pointless MD5 hash when using DSA keys in ca.
10804 [Anonymous <nobody@replay.com>]
10806 *) Generate an error if given an empty string as a cert directory. Also
10807 generate an error if handed NULL (previously returned 0 to indicate an
10808 error, but didn't set one).
10809 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10811 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10814 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10815 parameters. This was causing a warning which killed off the Win32 compile.
10818 *) Remove C++ style comments from crypto/bn/bn_local.h.
10819 [Neil Costigan <neil.costigan@celocom.com>]
10821 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10822 based on a text string, looking up short and long names and finally
10823 "dot" format. The "dot" format stuff didn't work. Added new function
10824 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10825 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10826 OID is not part of the table.
10829 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10830 X509_LOOKUP_by_alias().
10833 *) Sort openssl functions by name.
10836 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10837 encryption from sample DSA keys (in case anyone is interested the password
10841 *) Make _all_ *_free functions accept a NULL pointer.
10842 [Frans Heymans <fheymans@isaserver.be>]
10844 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10846 [Anonymous <nobody@replay.com>]
10848 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10849 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10851 *) Don't blow it for numeric -newkey arguments to apps/req.
10852 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10854 *) Temp key "for export" tests were wrong in s3_srvr.c.
10855 [Anonymous <nobody@replay.com>]
10857 *) Add prototype for temp key callback functions
10858 SSL_CTX_set_tmp_{rsa,dh}_callback().
10861 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10862 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10865 *) X509_name_add_entry() freed the wrong thing after an error.
10866 [Arne Ansper <arne@ats.cyber.ee>]
10868 *) rsa_eay.c would attempt to free a NULL context.
10869 [Arne Ansper <arne@ats.cyber.ee>]
10871 *) BIO_s_socket() had a broken should_retry() on Windoze.
10872 [Arne Ansper <arne@ats.cyber.ee>]
10874 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10875 [Arne Ansper <arne@ats.cyber.ee>]
10877 *) Make sure the already existing X509_STORE->depth variable is initialized
10878 in X509_STORE_new(), but document the fact that this variable is still
10879 unused in the certificate verification process.
10880 [Ralf S. Engelschall]
10882 *) Fix the various library and apps files to free up pkeys obtained from
10883 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10886 *) Fix reference counting in X509_PUBKEY_get(). This makes
10887 demos/maurice/example2.c work, amongst others, probably.
10888 [Steve Henson and Ben Laurie]
10890 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10891 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10892 are no longer created. This way we have a single and consistent command
10893 line interface `openssl <command>', similar to `cvs <command>'.
10894 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10896 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10897 BIT STRING wrapper always have zero unused bits.
10900 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10903 *) Make the top-level INSTALL documentation easier to understand.
10906 *) Makefiles updated to exit if an error occurs in a sub-directory
10907 make (including if user presses ^C) [Paul Sutton]
10909 *) Make Montgomery context stuff explicit in RSA data structure.
10912 *) Fix build order of pem and err to allow for generated pem.h.
10915 *) Fix renumbering bug in X509_NAME_delete_entry().
10918 *) Enhanced the err-ins.pl script so it makes the error library number
10919 global and can add a library name. This is needed for external ASN1 and
10920 other error libraries.
10923 *) Fixed sk_insert which never worked properly.
10926 *) Fix ASN1 macros so they can handle indefinite length constructed
10927 EXPLICIT tags. Some non standard certificates use these: they can now
10931 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10932 into a single doc/ssleay.txt bundle. This way the information is still
10933 preserved but no longer messes up this directory. Now it's new room for
10934 the new set of documentation files.
10935 [Ralf S. Engelschall]
10937 *) SETs were incorrectly DER encoded. This was a major pain, because they
10938 shared code with SEQUENCEs, which aren't coded the same. This means that
10939 almost everything to do with SETs or SEQUENCEs has either changed name or
10940 number of arguments.
10941 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10943 *) Fix test data to work with the above.
10946 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10947 was already fixed by Eric for 0.9.1 it seems.
10948 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10950 *) Autodetect FreeBSD3.
10953 *) Fix various bugs in Configure. This affects the following platforms:
10957 unixware-2.0-pentium
10961 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10962 before they are needed.
10965 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10969 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10971 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10972 changed SSLeay to OpenSSL in version strings.
10973 [Ralf S. Engelschall]
10975 *) Some fixups to the top-level documents.
10978 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10979 because the symlink to include/ was missing.
10980 [Ralf S. Engelschall]
10982 *) Incorporated the popular no-RSA/DSA-only patches
10983 which allow to compile a RSA-free SSLeay.
10984 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10986 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10987 when "ssleay" is still not found.
10988 [Ralf S. Engelschall]
10990 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10991 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10993 *) Updated the README file.
10994 [Ralf S. Engelschall]
10996 *) Added various .cvsignore files in the CVS repository subdirs
10997 to make a "cvs update" really silent.
10998 [Ralf S. Engelschall]
11000 *) Recompiled the error-definition header files and added
11001 missing symbols to the Win32 linker tables.
11002 [Ralf S. Engelschall]
11004 *) Cleaned up the top-level documents;
11005 o new files: CHANGES and LICENSE
11006 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11007 o merged COPYRIGHT into LICENSE
11008 o removed obsolete TODO file
11009 o renamed MICROSOFT to INSTALL.W32
11010 [Ralf S. Engelschall]
11012 *) Removed dummy files from the 0.9.1b source tree:
11013 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11014 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11015 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11016 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11017 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11018 [Ralf S. Engelschall]
11020 *) Added various platform portability fixes.
11023 *) The Genesis of the OpenSSL rpject:
11024 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11025 Young and Tim J. Hudson created while they were working for C2Net until
11027 [The OpenSSL Project]
11030 Changes between 0.9.0b and 0.9.1b [not released]
11032 *) Updated a few CA certificates under certs/
11035 *) Changed some BIGNUM api stuff.
11038 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11039 DGUX x86, Linux Alpha, etc.
11042 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11043 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11047 *) Add -strparse option to asn1pars program which parses nested
11049 [Dr Stephen Henson <shenson@bigfoot.com>]
11051 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11054 *) DSA fix for "ca" program.
11057 *) Added "-genkey" option to "dsaparam" program.
11060 *) Added RIPE MD160 (rmd160) message digest.
11063 *) Added -a (all) option to "ssleay version" command.
11066 *) Added PLATFORM define which is the id given to Configure.
11069 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11072 *) Extended the ASN.1 parser routines.
11075 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11078 *) Added a BN_CTX to the BN library.
11081 *) Fixed the weak key values in DES library
11084 *) Changed API in EVP library for cipher aliases.
11087 *) Added support for RC2/64bit cipher.
11090 *) Converted the lhash library to the crypto/mem.c functions.
11093 *) Added more recognized ASN.1 object ids.
11096 *) Added more RSA padding checks for SSL/TLS.
11099 *) Added BIO proxy/filter functionality.
11102 *) Added extra_certs to SSL_CTX which can be used
11103 send extra CA certificates to the client in the CA cert chain sending
11104 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11107 *) Now Fortezza is denied in the authentication phase because
11108 this is key exchange mechanism is not supported by SSLeay at all.
11111 *) Additional PKCS1 checks.
11114 *) Support the string "TLSv1" for all TLS v1 ciphers.
11117 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11118 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11121 *) Fixed a few memory leaks.
11124 *) Fixed various code and comment typos.
11127 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11128 bytes sent in the client random.
11129 [Edward Bishop <ebishop@spyglass.com>]