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 *) Start cleaning up OPENSSL_NO_xxx #define's
58 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
59 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
60 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
61 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
62 Remove MS_STATIC; it's a relic from platforms <32 bits.
65 *) Start cleaning up dead code
66 Remove all but one '#ifdef undef' which is to be looked at.
69 *) Experimental support for a new, fast, unbiased prime candidate generator,
70 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
71 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
73 *) New output format NSS in the sess_id command line tool. This allows
74 exporting the session id and the master key in NSS keylog format.
75 [Martin Kaiser <martin@kaiser.cx>]
77 *) Harmonize version and its documentation. -f flag is used to display
79 [mancha <mancha1@zoho.com>]
81 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
83 [mancha <mancha1@zoho.com>]
85 *) Fix some double frees. These are not thought to be exploitable.
86 [mancha <mancha1@zoho.com>]
88 *) A missing bounds check in the handling of the TLS heartbeat extension
89 can be used to reveal up to 64k of memory to a connected client or
92 Thanks for Neel Mehta of Google Security for discovering this bug and to
93 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
94 preparing the fix (CVE-2014-0160)
95 [Adam Langley, Bodo Moeller]
97 *) Fix for the attack described in the paper "Recovering OpenSSL
98 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
99 by Yuval Yarom and Naomi Benger. Details can be obtained from:
100 http://eprint.iacr.org/2014/140
102 Thanks to Yuval Yarom and Naomi Benger for discovering this
103 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
104 [Yuval Yarom and Naomi Benger]
106 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
107 this fixes a limitation in previous versions of OpenSSL.
110 *) Experimental encrypt-then-mac support.
112 Experimental support for encrypt then mac from
113 draft-gutmann-tls-encrypt-then-mac-02.txt
115 To enable it set the appropriate extension number (0x42 for the test
116 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
118 For non-compliant peers (i.e. just about everything) this should have no
121 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
125 *) Add EVP support for key wrapping algorithms, to avoid problems with
126 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
127 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
128 algorithms and include tests cases.
131 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
135 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
136 MGF1 digest and OAEP label.
139 *) Make openssl verify return errors.
140 [Chris Palmer <palmer@google.com> and Ben Laurie]
142 *) New function ASN1_TIME_diff to calculate the difference between two
143 ASN1_TIME structures or one structure and the current time.
146 *) Update fips_test_suite to support multiple command line options. New
147 test to induce all self test errors in sequence and check expected
151 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
152 sign or verify all in one operation.
155 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
156 test programs and fips_test_suite. Includes functionality to parse
157 the minimal script output of fipsalgest.pl directly.
160 *) Add authorisation parameter to FIPS_module_mode_set().
163 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
166 *) Use separate DRBG fields for internal and external flags. New function
167 FIPS_drbg_health_check() to perform on demand health checking. Add
168 generation tests to fips_test_suite with reduced health check interval to
169 demonstrate periodic health checking. Add "nodh" option to
170 fips_test_suite to skip very slow DH test.
173 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
177 *) More extensive health check for DRBG checking many more failure modes.
178 New function FIPS_selftest_drbg_all() to handle every possible DRBG
179 combination: call this in fips_test_suite.
182 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
183 and POST to handle Dual EC cases.
186 *) Add support for canonical generation of DSA parameter 'g'. See
189 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
190 POST to handle HMAC cases.
193 *) Add functions FIPS_module_version() and FIPS_module_version_text()
194 to return numerical and string versions of the FIPS module number.
197 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
198 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
199 outside the validated module in the FIPS capable OpenSSL.
202 *) Minor change to DRBG entropy callback semantics. In some cases
203 there is no multiple of the block length between min_len and
204 max_len. Allow the callback to return more than max_len bytes
205 of entropy but discard any extra: it is the callback's responsibility
206 to ensure that the extra data discarded does not impact the
207 requested amount of entropy.
210 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
211 information in FIPS186-3, SP800-57 and SP800-131A.
214 *) CCM support via EVP. Interface is very similar to GCM case except we
215 must supply all data in one chunk (i.e. no update, final) and the
216 message length must be supplied if AAD is used. Add algorithm test
220 *) Initial version of POST overhaul. Add POST callback to allow the status
221 of POST to be monitored and/or failures induced. Modify fips_test_suite
222 to use callback. Always run all selftests even if one fails.
225 *) XTS support including algorithm test driver in the fips_gcmtest program.
226 Note: this does increase the maximum key length from 32 to 64 bytes but
227 there should be no binary compatibility issues as existing applications
228 will never use XTS mode.
231 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
232 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
233 performs algorithm blocking for unapproved PRNG types. Also do not
234 set PRNG type in FIPS_mode_set(): leave this to the application.
235 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
236 the standard OpenSSL PRNG: set additional data to a date time vector.
239 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
240 This shouldn't present any incompatibility problems because applications
241 shouldn't be using these directly and any that are will need to rethink
242 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
245 *) Extensive self tests and health checking required by SP800-90 DRBG.
246 Remove strength parameter from FIPS_drbg_instantiate and always
247 instantiate at maximum supported strength.
250 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
253 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
256 *) New function DH_compute_key_padded() to compute a DH key and pad with
257 leading zeroes if needed: this complies with SP800-56A et al.
260 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
261 anything, incomplete, subject to change and largely untested at present.
264 *) Modify fipscanisteronly build option to only build the necessary object
265 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
268 *) Add experimental option FIPSSYMS to give all symbols in
269 fipscanister.o and FIPS or fips prefix. This will avoid
270 conflicts with future versions of OpenSSL. Add perl script
271 util/fipsas.pl to preprocess assembly language source files
272 and rename any affected symbols.
275 *) Add selftest checks and algorithm block of non-fips algorithms in
276 FIPS mode. Remove DES2 from selftests.
279 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
280 return internal method without any ENGINE dependencies. Add new
281 tiny fips sign and verify functions.
284 *) New build option no-ec2m to disable characteristic 2 code.
287 *) New build option "fipscanisteronly". This only builds fipscanister.o
288 and (currently) associated fips utilities. Uses the file Makefile.fips
289 instead of Makefile.org as the prototype.
292 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
293 Update fips_gcmtest to use IV generator.
296 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
297 setting output buffer to NULL. The *Final function must be
298 called although it will not retrieve any additional data. The tag
299 can be set or retrieved with a ctrl. The IV length is by default 12
300 bytes (96 bits) but can be set to an alternative value. If the IV
301 length exceeds the maximum IV length (currently 16 bytes) it cannot be
305 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
306 underlying do_cipher function handles all cipher semantics itself
307 including padding and finalisation. This is useful if (for example)
308 an ENGINE cipher handles block padding itself. The behaviour of
309 do_cipher is subtly changed if this flag is set: the return value
310 is the number of characters written to the output buffer (zero is
311 no longer an error code) or a negative error code. Also if the
312 input buffer is NULL and length 0 finalisation should be performed.
315 *) If a candidate issuer certificate is already part of the constructed
316 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
319 *) Improve forward-security support: add functions
321 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
322 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
324 for use by SSL/TLS servers; the callback function will be called whenever a
325 new session is created, and gets to decide whether the session may be
326 cached to make it resumable (return 0) or not (return 1). (As by the
327 SSL/TLS protocol specifications, the session_id sent by the server will be
328 empty to indicate that the session is not resumable; also, the server will
329 not generate RFC 4507 (RFC 5077) session tickets.)
331 A simple reasonable callback implementation is to return is_forward_secure.
332 This parameter will be set to 1 or 0 depending on the ciphersuite selected
333 by the SSL/TLS server library, indicating whether it can provide forward
335 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
337 *) New -verify_name option in command line utilities to set verification
341 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
342 Add CMAC pkey methods.
345 *) Experimental renegotiation in s_server -www mode. If the client
346 browses /reneg connection is renegotiated. If /renegcert it is
347 renegotiated requesting a certificate.
350 *) Add an "external" session cache for debugging purposes to s_server. This
351 should help trace issues which normally are only apparent in deployed
352 multi-process servers.
355 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
356 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
357 BIO_set_cipher() and some obscure PEM functions were changed so they
358 can now return an error. The RAND changes required a change to the
359 RAND_METHOD structure.
362 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
363 a gcc attribute to warn if the result of a function is ignored. This
364 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
365 whose return value is often ignored.
368 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
370 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
371 ARMv5 through ARMv8, as opposite to "locking" it to single one.
372 So far those who have to target multiple plaforms would compromise
373 and argue that binary targeting say ARMv5 would still execute on
374 ARMv8. "Universal" build resolves this compromise by providing
375 near-optimal performance even on newer platforms.
378 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
379 (other platforms pending).
380 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
382 *) Add support for the SignedCertificateTimestampList certificate and
383 OCSP response extensions from RFC6962.
386 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
387 for corner cases. (Certain input points at infinity could lead to
388 bogus results, with non-infinity inputs mapped to infinity too.)
391 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
392 This covers AES, SHA256/512 and GHASH. "Initial" means that most
393 common cases are optimized and there still is room for further
394 improvements. Vector Permutation AES for Altivec is also added.
397 *) Add support for little-endian ppc64 Linux target.
398 [Marcelo Cerri (IBM)]
400 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
401 SHA1, SHA256 and GHASH. "Initial" means that most common cases
402 are optimized and there still is room for further improvements.
403 Both 32- and 64-bit modes are supported.
404 [Andy Polyakov, Ard Biesheuvel (Linaro)]
406 *) Improved ARMv7 NEON support.
409 *) Support for SPARC Architecture 2011 crypto extensions, first
410 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
411 SHA256/512, MD5, GHASH and modular exponentiation.
412 [Andy Polyakov, David Miller]
414 *) Accelerated modular exponentiation for Intel processors, a.k.a.
416 [Shay Gueron & Vlad Krasnov (Intel Corp)]
418 *) Support for new and upcoming Intel processors, including AVX2,
419 BMI and SHA ISA extensions. This includes additional "stitched"
420 implementations, AESNI-SHA256 and GCM, and multi-buffer support
423 This work was sponsored by Intel Corp.
426 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
427 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
428 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
431 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
432 this fixes a limiation in previous versions of OpenSSL.
435 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
436 MGF1 digest and OAEP label.
439 *) Add EVP support for key wrapping algorithms, to avoid problems with
440 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
441 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
442 algorithms and include tests cases.
445 *) Add functions to allocate and set the fields of an ECDSA_METHOD
447 [Douglas E. Engert, Steve Henson]
449 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
450 difference in days and seconds between two tm or ASN1_TIME structures.
453 *) Add -rev test option to s_server to just reverse order of characters
454 received by client and send back to server. Also prints an abbreviated
455 summary of the connection parameters.
458 *) New option -brief for s_client and s_server to print out a brief summary
459 of connection parameters.
462 *) Add callbacks for arbitrary TLS extensions.
463 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
465 *) New option -crl_download in several openssl utilities to download CRLs
466 from CRLDP extension in certificates.
469 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
472 *) New function X509_CRL_diff to generate a delta CRL from the difference
473 of two full CRLs. Add support to "crl" utility.
476 *) New functions to set lookup_crls function and to retrieve
477 X509_STORE from X509_STORE_CTX.
480 *) Print out deprecated issuer and subject unique ID fields in
484 *) Extend OCSP I/O functions so they can be used for simple general purpose
485 HTTP as well as OCSP. New wrapper function which can be used to download
486 CRLs using the OCSP API.
489 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
492 *) SSL_CONF* functions. These provide a common framework for application
493 configuration using configuration files or command lines.
496 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
497 message callback and prints the results. Needs compile time option
498 "enable-ssl-trace". New options to s_client and s_server to enable
502 *) New ctrl and macro to retrieve supported points extensions.
503 Print out extension in s_server and s_client.
506 *) New functions to retrieve certificate signature and signature
510 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
514 *) New Suite B modes for TLS code. These use and enforce the requirements
515 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
516 only use Suite B curves. The Suite B modes can be set by using the
517 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
520 *) New chain verification flags for Suite B levels of security. Check
521 algorithms are acceptable when flags are set in X509_verify_cert.
524 *) Make tls1_check_chain return a set of flags indicating checks passed
525 by a certificate chain. Add additional tests to handle client
526 certificates: checks for matching certificate type and issuer name
530 *) If an attempt is made to use a signature algorithm not in the peer
531 preference list abort the handshake. If client has no suitable
532 signature algorithms in response to a certificate request do not
536 *) If server EC tmp key is not in client preference list abort handshake.
539 *) Add support for certificate stores in CERT structure. This makes it
540 possible to have different stores per SSL structure or one store in
541 the parent SSL_CTX. Include distint stores for certificate chain
542 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
543 to build and store a certificate chain in CERT structure: returing
544 an error if the chain cannot be built: this will allow applications
545 to test if a chain is correctly configured.
547 Note: if the CERT based stores are not set then the parent SSL_CTX
548 store is used to retain compatibility with existing behaviour.
552 *) New function ssl_set_client_disabled to set a ciphersuite disabled
553 mask based on the current session, check mask when sending client
554 hello and checking the requested ciphersuite.
557 *) New ctrls to retrieve and set certificate types in a certificate
558 request message. Print out received values in s_client. If certificate
559 types is not set with custom values set sensible values based on
560 supported signature algorithms.
563 *) Support for distinct client and server supported signature algorithms.
566 *) Add certificate callback. If set this is called whenever a certificate
567 is required by client or server. An application can decide which
568 certificate chain to present based on arbitrary criteria: for example
569 supported signature algorithms. Add very simple example to s_server.
570 This fixes many of the problems and restrictions of the existing client
571 certificate callback: for example you can now clear an existing
572 certificate and specify the whole chain.
575 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
576 the certificate can be used for (if anything). Set valid_flags field
577 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
578 to have similar checks in it.
580 Add new "cert_flags" field to CERT structure and include a "strict mode".
581 This enforces some TLS certificate requirements (such as only permitting
582 certificate signature algorithms contained in the supported algorithms
583 extension) which some implementations ignore: this option should be used
584 with caution as it could cause interoperability issues.
587 *) Update and tidy signature algorithm extension processing. Work out
588 shared signature algorithms based on preferences and peer algorithms
589 and print them out in s_client and s_server. Abort handshake if no
590 shared signature algorithms.
593 *) Add new functions to allow customised supported signature algorithms
594 for SSL and SSL_CTX structures. Add options to s_client and s_server
598 *) New function SSL_certs_clear() to delete all references to certificates
599 from an SSL structure. Before this once a certificate had been added
600 it couldn't be removed.
603 *) Integrate hostname, email address and IP address checking with certificate
604 verification. New verify options supporting checking in opensl utility.
607 *) Fixes and wildcard matching support to hostname and email checking
608 functions. Add manual page.
609 [Florian Weimer (Red Hat Product Security Team)]
611 *) New functions to check a hostname email or IP address against a
612 certificate. Add options x509 utility to print results of checks against
616 *) Fix OCSP checking.
617 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
619 *) Initial experimental support for explicitly trusted non-root CAs.
620 OpenSSL still tries to build a complete chain to a root but if an
621 intermediate CA has a trust setting included that is used. The first
622 setting is used: whether to trust (e.g., -addtrust option to the x509
626 *) Add -trusted_first option which attempts to find certificates in the
627 trusted store even if an untrusted chain is also supplied.
630 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
631 platform support for Linux and Android.
634 *) Support for linux-x32, ILP32 environment in x86_64 framework.
637 *) Experimental multi-implementation support for FIPS capable OpenSSL.
638 When in FIPS mode the approved implementations are used as normal,
639 when not in FIPS mode the internal unapproved versions are used instead.
640 This means that the FIPS capable OpenSSL isn't forced to use the
641 (often lower performance) FIPS implementations outside FIPS mode.
644 *) Transparently support X9.42 DH parameters when calling
645 PEM_read_bio_DHparameters. This means existing applications can handle
646 the new parameter format automatically.
649 *) Initial experimental support for X9.42 DH parameter format: mainly
650 to support use of 'q' parameter for RFC5114 parameters.
653 *) Add DH parameters from RFC5114 including test data to dhtest.
656 *) Support for automatic EC temporary key parameter selection. If enabled
657 the most preferred EC parameters are automatically used instead of
658 hardcoded fixed parameters. Now a server just has to call:
659 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
660 support ECDH and use the most appropriate parameters.
663 *) Enhance and tidy EC curve and point format TLS extension code. Use
664 static structures instead of allocation if default values are used.
665 New ctrls to set curves we wish to support and to retrieve shared curves.
666 Print out shared curves in s_server. New options to s_server and s_client
667 to set list of supported curves.
670 *) New ctrls to retrieve supported signature algorithms and
671 supported curve values as an array of NIDs. Extend openssl utility
672 to print out received values.
675 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
676 between NIDs and the more common NIST names such as "P-256". Enhance
677 ecparam utility and ECC method to recognise the NIST names for curves.
680 *) Enhance SSL/TLS certificate chain handling to support different
681 chains for each certificate instead of one chain in the parent SSL_CTX.
684 *) Support for fixed DH ciphersuite client authentication: where both
685 server and client use DH certificates with common parameters.
688 *) Support for fixed DH ciphersuites: those requiring DH server
692 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
694 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
695 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
696 X509_CINF_get_signature were reverted post internal team review.
698 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
700 *) Abort handshake if server key exchange message is omitted for ephemeral
703 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
704 reporting this issue.
708 *) Remove non-export ephemeral RSA code on client and server. This code
709 violated the TLS standard by allowing the use of temporary RSA keys in
710 non-export ciphersuites and could be used by a server to effectively
711 downgrade the RSA key length used to a value smaller than the server
712 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
713 INRIA or reporting this issue.
717 *) Ensure that the session ID context of an SSL is updated when its
718 SSL_CTX is updated via SSL_set_SSL_CTX.
720 The session ID context is typically set from the parent SSL_CTX,
721 and can vary with the CTX.
724 *) Fix various certificate fingerprint issues.
726 By using non-DER or invalid encodings outside the signed portion of a
727 certificate the fingerprint can be changed without breaking the signature.
728 Although no details of the signed portion of the certificate can be changed
729 this can cause problems with some applications: e.g. those using the
730 certificate fingerprint for blacklists.
732 1. Reject signatures with non zero unused bits.
734 If the BIT STRING containing the signature has non zero unused bits reject
735 the signature. All current signature algorithms require zero unused bits.
737 2. Check certificate algorithm consistency.
739 Check the AlgorithmIdentifier inside TBS matches the one in the
740 certificate signature. NB: this will result in signature failure
741 errors for some broken certificates.
743 Thanks to Konrad Kraszewski from Google for reporting this issue.
745 3. Check DSA/ECDSA signatures use DER.
747 Reencode DSA/ECDSA signatures and compare with the original received
748 signature. Return an error if there is a mismatch.
750 This will reject various cases including garbage after signature
751 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
752 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
753 (negative or with leading zeroes).
755 Further analysis was conducted and fixes were developed by Stephen Henson
756 of the OpenSSL core team.
761 *) Do not resume sessions on the server if the negotiated protocol
762 version does not match the session's version. Resuming with a different
763 version, while not strictly forbidden by the RFC, is of questionable
764 sanity and breaks all known clients.
765 [David Benjamin, Emilia Käsper]
767 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
768 early CCS messages during renegotiation. (Note that because
769 renegotiation is encrypted, this early CCS was not exploitable.)
772 *) Tighten client-side session ticket handling during renegotiation:
773 ensure that the client only accepts a session ticket if the server sends
774 the extension anew in the ServerHello. Previously, a TLS client would
775 reuse the old extension state and thus accept a session ticket if one was
776 announced in the initial ServerHello.
778 Similarly, ensure that the client requires a session ticket if one
779 was advertised in the ServerHello. Previously, a TLS client would
780 ignore a missing NewSessionTicket message.
783 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
787 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
788 sends a carefully crafted handshake message, to cause OpenSSL to fail
789 to free up to 64k of memory causing a memory leak. This could be
790 exploited in a Denial Of Service attack. This issue affects OpenSSL
791 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
792 whether SRTP is used or configured. Implementations of OpenSSL that
793 have been compiled with OPENSSL_NO_SRTP defined are not affected.
795 The fix was developed by the OpenSSL team.
799 *) Session Ticket Memory Leak.
801 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
802 integrity of that ticket is first verified. In the event of a session
803 ticket integrity check failing, OpenSSL will fail to free memory
804 causing a memory leak. By sending a large number of invalid session
805 tickets an attacker could exploit this issue in a Denial Of Service
810 *) Build option no-ssl3 is incomplete.
812 When OpenSSL is configured with "no-ssl3" as a build option, servers
813 could accept and complete a SSL 3.0 handshake, and clients could be
814 configured to send them.
816 [Akamai and the OpenSSL team]
818 *) Add support for TLS_FALLBACK_SCSV.
819 Client applications doing fallback retries should call
820 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
822 [Adam Langley, Bodo Moeller]
824 *) Add additional DigestInfo checks.
826 Reencode DigestInto in DER and check against the original when
827 verifying RSA signature: this will reject any improperly encoded
828 DigestInfo structures.
830 Note: this is a precautionary measure and no attacks are currently known.
834 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
836 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
837 SRP code can be overrun an internal buffer. Add sanity check that
838 g, A, B < N to SRP code.
840 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
841 Group for discovering this issue.
845 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
846 TLS 1.0 instead of higher protocol versions when the ClientHello message
847 is badly fragmented. This allows a man-in-the-middle attacker to force a
848 downgrade to TLS 1.0 even if both the server and the client support a
849 higher protocol version, by modifying the client's TLS records.
851 Thanks to David Benjamin and Adam Langley (Google) for discovering and
852 researching this issue.
856 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
857 to a denial of service attack. A malicious server can crash the client
858 with a null pointer dereference (read) by specifying an anonymous (EC)DH
859 ciphersuite and sending carefully crafted handshake messages.
861 Thanks to Felix Gröbert (Google) for discovering and researching this
866 *) By sending carefully crafted DTLS packets an attacker could cause openssl
867 to leak memory. This can be exploited through a Denial of Service attack.
868 Thanks to Adam Langley for discovering and researching this issue.
872 *) An attacker can force openssl to consume large amounts of memory whilst
873 processing DTLS handshake messages. This can be exploited through a
874 Denial of Service attack.
875 Thanks to Adam Langley for discovering and researching this issue.
879 *) An attacker can force an error condition which causes openssl to crash
880 whilst processing DTLS packets due to memory being freed twice. This
881 can be exploited through a Denial of Service attack.
882 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
887 *) If a multithreaded client connects to a malicious server using a resumed
888 session and the server sends an ec point format extension it could write
889 up to 255 bytes to freed memory.
891 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
896 *) A malicious server can crash an OpenSSL client with a null pointer
897 dereference (read) by specifying an SRP ciphersuite even though it was not
898 properly negotiated with the client. This can be exploited through a
899 Denial of Service attack.
901 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
902 discovering and researching this issue.
906 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
907 X509_name_oneline, X509_name_print_ex et al. to leak some information
908 from the stack. Applications may be affected if they echo pretty printing
909 output to the attacker.
911 Thanks to Ivan Fratric (Google) for discovering this issue.
913 [Emilia Käsper, and Steve Henson]
915 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
916 for corner cases. (Certain input points at infinity could lead to
917 bogus results, with non-infinity inputs mapped to infinity too.)
920 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
922 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
923 handshake can force the use of weak keying material in OpenSSL
924 SSL/TLS clients and servers.
926 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
927 researching this issue. (CVE-2014-0224)
928 [KIKUCHI Masashi, Steve Henson]
930 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
931 OpenSSL DTLS client the code can be made to recurse eventually crashing
934 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
936 [Imre Rad, Steve Henson]
938 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
939 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
940 client or server. This is potentially exploitable to run arbitrary
941 code on a vulnerable client or server.
943 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
944 [Jüri Aedla, Steve Henson]
946 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
947 are subject to a denial of service attack.
949 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
950 this issue. (CVE-2014-3470)
951 [Felix Gröbert, Ivan Fratric, Steve Henson]
953 *) Harmonize version and its documentation. -f flag is used to display
955 [mancha <mancha1@zoho.com>]
957 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
959 [mancha <mancha1@zoho.com>]
961 *) Fix some double frees. These are not thought to be exploitable.
962 [mancha <mancha1@zoho.com>]
964 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
966 *) A missing bounds check in the handling of the TLS heartbeat extension
967 can be used to reveal up to 64k of memory to a connected client or
970 Thanks for Neel Mehta of Google Security for discovering this bug and to
971 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
972 preparing the fix (CVE-2014-0160)
973 [Adam Langley, Bodo Moeller]
975 *) Fix for the attack described in the paper "Recovering OpenSSL
976 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
977 by Yuval Yarom and Naomi Benger. Details can be obtained from:
978 http://eprint.iacr.org/2014/140
980 Thanks to Yuval Yarom and Naomi Benger for discovering this
981 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
982 [Yuval Yarom and Naomi Benger]
984 *) TLS pad extension: draft-agl-tls-padding-03
986 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
987 TLS client Hello record length value would otherwise be > 255 and
988 less that 512 pad with a dummy extension containing zeroes so it
989 is at least 512 bytes long.
991 [Adam Langley, Steve Henson]
993 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
995 *) Fix for TLS record tampering bug. A carefully crafted invalid
996 handshake could crash OpenSSL with a NULL pointer exception.
997 Thanks to Anton Johansson for reporting this issues.
1000 *) Keep original DTLS digest and encryption contexts in retransmission
1001 structures so we can use the previous session parameters if they need
1002 to be resent. (CVE-2013-6450)
1005 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1006 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1007 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1008 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1009 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1010 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1011 [Rob Stradling, Adam Langley]
1013 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1015 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1016 supporting platforms or when small records were transferred.
1017 [Andy Polyakov, Steve Henson]
1019 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1021 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1023 This addresses the flaw in CBC record processing discovered by
1024 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1025 at: http://www.isg.rhul.ac.uk/tls/
1027 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1028 Security Group at Royal Holloway, University of London
1029 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1030 Emilia Käsper for the initial patch.
1032 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1034 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1035 ciphersuites which can be exploited in a denial of service attack.
1036 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1037 and detecting this bug and to Wolfgang Ettlinger
1038 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1042 *) Return an error when checking OCSP signatures when key is NULL.
1043 This fixes a DoS attack. (CVE-2013-0166)
1046 *) Make openssl verify return errors.
1047 [Chris Palmer <palmer@google.com> and Ben Laurie]
1049 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1050 the right response is stapled. Also change SSL_get_certificate()
1051 so it returns the certificate actually sent.
1052 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1053 [Rob Stradling <rob.stradling@comodo.com>]
1055 *) Fix possible deadlock when decoding public keys.
1058 *) Don't use TLS 1.0 record version number in initial client hello
1062 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1064 *) Sanity check record length before skipping explicit IV in TLS
1065 1.2, 1.1 and DTLS to fix DoS attack.
1067 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1068 fuzzing as a service testing platform.
1072 *) Initialise tkeylen properly when encrypting CMS messages.
1073 Thanks to Solar Designer of Openwall for reporting this issue.
1076 *) In FIPS mode don't try to use composite ciphers as they are not
1080 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1082 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1083 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1084 mean any application compiled against OpenSSL 1.0.0 headers setting
1085 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1086 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1087 0x10000000L Any application which was previously compiled against
1088 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1089 will need to be recompiled as a result. Letting be results in
1090 inability to disable specifically TLS 1.1 and in client context,
1091 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1094 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1095 disable just protocol X, but all protocols above X *if* there are
1096 protocols *below* X still enabled. In more practical terms it means
1097 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1098 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1099 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1103 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1105 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1106 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1107 in CRYPTO_realloc_clean.
1109 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1110 issue and to Adam Langley <agl@chromium.org> for fixing it.
1112 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1114 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1117 *) Workarounds for some broken servers that "hang" if a client hello
1118 record length exceeds 255 bytes.
1120 1. Do not use record version number > TLS 1.0 in initial client
1121 hello: some (but not all) hanging servers will now work.
1122 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1123 the number of ciphers sent in the client hello. This should be
1124 set to an even number, such as 50, for example by passing:
1125 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1126 Most broken servers should now work.
1127 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1128 TLS 1.2 client support entirely.
1131 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1134 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1136 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1137 STRING form instead of a DigestInfo.
1140 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1141 and the RSA_sign/RSA_verify functions. This was made more apparent when
1142 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1143 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1144 the correct format in RSA_verify so both forms transparently work.
1147 *) Some servers which support TLS 1.0 can choke if we initially indicate
1148 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1149 encrypted premaster secret. As a workaround use the maximum permitted
1150 client version in client hello, this should keep such servers happy
1151 and still work with previous versions of OpenSSL.
1154 *) Add support for TLS/DTLS heartbeats.
1155 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1157 *) Add support for SCTP.
1158 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1160 *) Improved PRNG seeding for VOS.
1161 [Paul Green <Paul.Green@stratus.com>]
1163 *) Extensive assembler packs updates, most notably:
1165 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1166 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1167 - x86_64: bit-sliced AES implementation;
1168 - ARM: NEON support, contemporary platforms optimizations;
1169 - s390x: z196 support;
1170 - *: GHASH and GF(2^m) multiplication implementations;
1174 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1175 (removal of unnecessary code)
1176 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1178 *) Add TLS key material exporter from RFC 5705.
1181 *) Add DTLS-SRTP negotiation from RFC 5764.
1184 *) Add Next Protocol Negotiation,
1185 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1186 disabled with a no-npn flag to config or Configure. Code donated
1188 [Adam Langley <agl@google.com> and Ben Laurie]
1190 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1191 NIST-P256, NIST-P521, with constant-time single point multiplication on
1192 typical inputs. Compiler support for the nonstandard type __uint128_t is
1193 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1194 Code made available under Apache License version 2.0.
1196 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1197 line to include this in your build of OpenSSL, and run "make depend" (or
1198 "make update"). This enables the following EC_METHODs:
1200 EC_GFp_nistp224_method()
1201 EC_GFp_nistp256_method()
1202 EC_GFp_nistp521_method()
1204 EC_GROUP_new_by_curve_name() will automatically use these (while
1205 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1207 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1209 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1210 all platforms. Move ssize_t definition from e_os.h to the public
1211 header file e_os2.h as it now appears in public header file cms.h
1214 *) New -sigopt option to the ca, req and x509 utilities. Additional
1215 signature parameters can be passed using this option and in
1219 *) Add RSA PSS signing function. This will generate and set the
1220 appropriate AlgorithmIdentifiers for PSS based on those in the
1221 corresponding EVP_MD_CTX structure. No application support yet.
1224 *) Support for companion algorithm specific ASN1 signing routines.
1225 New function ASN1_item_sign_ctx() signs a pre-initialised
1226 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1227 the appropriate parameters.
1230 *) Add new algorithm specific ASN1 verification initialisation function
1231 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1232 handling will be the same no matter what EVP_PKEY_METHOD is used.
1233 Add a PSS handler to support verification of PSS signatures: checked
1234 against a number of sample certificates.
1237 *) Add signature printing for PSS. Add PSS OIDs.
1238 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1240 *) Add algorithm specific signature printing. An individual ASN1 method
1241 can now print out signatures instead of the standard hex dump.
1243 More complex signatures (e.g. PSS) can print out more meaningful
1244 information. Include DSA version that prints out the signature
1248 *) Password based recipient info support for CMS library: implementing
1252 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1253 neatly separates the code into cipher and PBE sections and is required
1254 for some algorithms that split PBES2 into separate pieces (such as
1255 password based CMS).
1258 *) Session-handling fixes:
1259 - Fix handling of connections that are resuming with a session ID,
1260 but also support Session Tickets.
1261 - Fix a bug that suppressed issuing of a new ticket if the client
1262 presented a ticket with an expired session.
1263 - Try to set the ticket lifetime hint to something reasonable.
1264 - Make tickets shorter by excluding irrelevant information.
1265 - On the client side, don't ignore renewed tickets.
1266 [Adam Langley, Bodo Moeller (Google)]
1268 *) Fix PSK session representation.
1271 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1273 This work was sponsored by Intel.
1276 *) Add GCM support to TLS library. Some custom code is needed to split
1277 the IV between the fixed (from PRF) and explicit (from TLS record)
1278 portions. This adds all GCM ciphersuites supported by RFC5288 and
1279 RFC5289. Generalise some AES* cipherstrings to include GCM and
1280 add a special AESGCM string for GCM only.
1283 *) Expand range of ctrls for AES GCM. Permit setting invocation
1284 field on decrypt and retrieval of invocation field only on encrypt.
1287 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1288 As required by RFC5289 these ciphersuites cannot be used if for
1289 versions of TLS earlier than 1.2.
1292 *) For FIPS capable OpenSSL interpret a NULL default public key method
1293 as unset and return the appropriate default but do *not* set the default.
1294 This means we can return the appropriate method in applications that
1295 switch between FIPS and non-FIPS modes.
1298 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1299 ENGINE is used then we cannot handle that in the FIPS module so we
1300 keep original code iff non-FIPS operations are allowed.
1303 *) Add -attime option to openssl utilities.
1304 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1306 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1309 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1310 FIPS EC methods unconditionally for now.
1313 *) New build option no-ec2m to disable characteristic 2 code.
1316 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1317 all cases can be covered as some introduce binary incompatibilities.
1320 *) Redirect RSA operations to FIPS module including keygen,
1321 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1324 *) Add similar low level API blocking to ciphers.
1327 *) Low level digest APIs are not approved in FIPS mode: any attempt
1328 to use these will cause a fatal error. Applications that *really* want
1329 to use them can use the private_* version instead.
1332 *) Redirect cipher operations to FIPS module for FIPS builds.
1335 *) Redirect digest operations to FIPS module for FIPS builds.
1338 *) Update build system to add "fips" flag which will link in fipscanister.o
1339 for static and shared library builds embedding a signature if needed.
1342 *) Output TLS supported curves in preference order instead of numerical
1343 order. This is currently hardcoded for the highest order curves first.
1344 This should be configurable so applications can judge speed vs strength.
1347 *) Add TLS v1.2 server support for client authentication.
1350 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1354 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1355 FIPS modules versions.
1358 *) Add TLS v1.2 client side support for client authentication. Keep cache
1359 of handshake records longer as we don't know the hash algorithm to use
1360 until after the certificate request message is received.
1363 *) Initial TLS v1.2 client support. Add a default signature algorithms
1364 extension including all the algorithms we support. Parse new signature
1365 format in client key exchange. Relax some ECC signing restrictions for
1366 TLS v1.2 as indicated in RFC5246.
1369 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1370 to new signature format when needed using client digest preference.
1371 All server ciphersuites should now work correctly in TLS v1.2. No client
1372 support yet and no support for client certificates.
1375 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1376 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1377 ciphersuites. At present only RSA key exchange ciphersuites work with
1378 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1379 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1380 and version checking.
1383 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1384 with this defined it will not be affected by any changes to ssl internal
1385 structures. Add several utility functions to allow openssl application
1386 to work with OPENSSL_NO_SSL_INTERN defined.
1390 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1392 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1395 *) Permit abbreviated handshakes when renegotiating using the function
1396 SSL_renegotiate_abbreviated().
1397 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1399 *) Add call to ENGINE_register_all_complete() to
1400 ENGINE_load_builtin_engines(), so some implementations get used
1401 automatically instead of needing explicit application support.
1404 *) Add support for TLS key exporter as described in RFC5705.
1405 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1407 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1408 a few changes are required:
1410 Add SSL_OP_NO_TLSv1_1 flag.
1411 Add TLSv1_1 methods.
1412 Update version checking logic to handle version 1.1.
1413 Add explicit IV handling (ported from DTLS code).
1414 Add command line options to s_client/s_server.
1417 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1419 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1421 This addresses the flaw in CBC record processing discovered by
1422 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1423 at: http://www.isg.rhul.ac.uk/tls/
1425 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1426 Security Group at Royal Holloway, University of London
1427 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1428 Emilia Käsper for the initial patch.
1430 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1432 *) Return an error when checking OCSP signatures when key is NULL.
1433 This fixes a DoS attack. (CVE-2013-0166)
1436 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1437 the right response is stapled. Also change SSL_get_certificate()
1438 so it returns the certificate actually sent.
1439 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1440 (This is a backport)
1441 [Rob Stradling <rob.stradling@comodo.com>]
1443 *) Fix possible deadlock when decoding public keys.
1446 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1448 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1451 *) Sanity check record length before skipping explicit IV in DTLS
1454 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1455 fuzzing as a service testing platform.
1459 *) Initialise tkeylen properly when encrypting CMS messages.
1460 Thanks to Solar Designer of Openwall for reporting this issue.
1463 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1465 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1466 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1467 in CRYPTO_realloc_clean.
1469 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1470 issue and to Adam Langley <agl@chromium.org> for fixing it.
1472 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1474 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1476 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1477 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1478 content decryption and always return the same error. Note: this attack
1479 needs on average 2^20 messages so it only affects automated senders. The
1480 old behaviour can be reenabled in the CMS code by setting the
1481 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1482 an MMA defence is not necessary.
1483 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1484 this issue. (CVE-2012-0884)
1487 *) Fix CVE-2011-4619: make sure we really are receiving a
1488 client hello before rejecting multiple SGC restarts. Thanks to
1489 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1492 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1494 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1495 Thanks to Antonio Martin, Enterprise Secure Access Research and
1496 Development, Cisco Systems, Inc. for discovering this bug and
1497 preparing a fix. (CVE-2012-0050)
1500 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1502 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1503 of the Vaudenay padding oracle attack on CBC mode encryption
1504 which enables an efficient plaintext recovery attack against
1505 the OpenSSL implementation of DTLS. Their attack exploits timing
1506 differences arising during decryption processing. A research
1507 paper describing this attack can be found at:
1508 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1509 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1510 Security Group at Royal Holloway, University of London
1511 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1512 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1513 for preparing the fix. (CVE-2011-4108)
1514 [Robin Seggelmann, Michael Tuexen]
1516 *) Clear bytes used for block padding of SSL 3.0 records.
1518 [Adam Langley (Google)]
1520 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1521 Kadianakis <desnacked@gmail.com> for discovering this issue and
1522 Adam Langley for preparing the fix. (CVE-2011-4619)
1523 [Adam Langley (Google)]
1525 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1526 [Andrey Kulikov <amdeich@gmail.com>]
1528 *) Prevent malformed RFC3779 data triggering an assertion failure.
1529 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1530 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1531 [Rob Austein <sra@hactrn.net>]
1533 *) Improved PRNG seeding for VOS.
1534 [Paul Green <Paul.Green@stratus.com>]
1536 *) Fix ssl_ciph.c set-up race.
1537 [Adam Langley (Google)]
1539 *) Fix spurious failures in ecdsatest.c.
1540 [Emilia Käsper (Google)]
1542 *) Fix the BIO_f_buffer() implementation (which was mixing different
1543 interpretations of the '..._len' fields).
1544 [Adam Langley (Google)]
1546 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1547 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1548 threads won't reuse the same blinding coefficients.
1550 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1551 lock to call BN_BLINDING_invert_ex, and avoids one use of
1552 BN_BLINDING_update for each BN_BLINDING structure (previously,
1553 the last update always remained unused).
1554 [Emilia Käsper (Google)]
1556 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1557 [Bob Buckholz (Google)]
1559 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1561 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1562 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1563 [Kaspar Brand <ossl@velox.ch>]
1565 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1566 for multi-threaded use of ECDH. (CVE-2011-3210)
1567 [Adam Langley (Google)]
1569 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1572 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1573 signature public key algorithm by using OID xref utilities instead.
1574 Before this you could only use some ECC ciphersuites with SHA1 only.
1577 *) Add protection against ECDSA timing attacks as mentioned in the paper
1578 by Billy Bob Brumley and Nicola Tuveri, see:
1580 http://eprint.iacr.org/2011/232.pdf
1582 [Billy Bob Brumley and Nicola Tuveri]
1584 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1586 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1587 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1589 *) Fix bug in string printing code: if *any* escaping is enabled we must
1590 escape the escape character (backslash) or the resulting string is
1594 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1596 *) Disable code workaround for ancient and obsolete Netscape browsers
1597 and servers: an attacker can use it in a ciphersuite downgrade attack.
1598 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1601 *) Fixed J-PAKE implementation error, originally discovered by
1602 Sebastien Martini, further info and confirmation from Stefan
1603 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1606 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1608 *) Fix extension code to avoid race conditions which can result in a buffer
1609 overrun vulnerability: resumed sessions must not be modified as they can
1610 be shared by multiple threads. CVE-2010-3864
1613 *) Fix WIN32 build system to correctly link an ENGINE directory into
1617 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1619 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1621 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1623 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1625 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1626 context. The operation can be customised via the ctrl mechanism in
1627 case ENGINEs want to include additional functionality.
1630 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1633 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1634 output hashes compatible with older versions of OpenSSL.
1635 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1637 *) Fix compression algorithm handling: if resuming a session use the
1638 compression algorithm of the resumed session instead of determining
1639 it from client hello again. Don't allow server to change algorithm.
1642 *) Add load_crls() function to apps tidying load_certs() too. Add option
1643 to verify utility to allow additional CRLs to be included.
1646 *) Update OCSP request code to permit adding custom headers to the request:
1647 some responders need this.
1650 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1652 [Julia Lawall <julia@diku.dk>]
1654 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1655 needlessly dereferenced structures, used obsolete functions and
1656 didn't handle all updated verify codes correctly.
1659 *) Disable MD2 in the default configuration.
1662 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1663 indicate the initial BIO being pushed or popped. This makes it possible
1664 to determine whether the BIO is the one explicitly called or as a result
1665 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1666 it handles reference counts correctly and doesn't zero out the I/O bio
1667 when it is not being explicitly popped. WARNING: applications which
1668 included workarounds for the old buggy behaviour will need to be modified
1669 or they could free up already freed BIOs.
1672 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1673 renaming to all platforms (within the 0.9.8 branch, this was
1674 done conditionally on Netware platforms to avoid a name clash).
1675 [Guenter <lists@gknw.net>]
1677 *) Add ECDHE and PSK support to DTLS.
1678 [Michael Tuexen <tuexen@fh-muenster.de>]
1680 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1684 *) Add "missing" function EVP_MD_flags() (without this the only way to
1685 retrieve a digest flags is by accessing the structure directly. Update
1686 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1687 or cipher is registered as in the "from" argument. Print out all
1688 registered digests in the dgst usage message instead of manually
1689 attempting to work them out.
1692 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1693 this allows the use of compression and extensions. Change default cipher
1694 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1695 by default unless an application cipher string requests it.
1698 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1699 key ids to find matching certificates and keys but some PKCS#12 files
1700 don't follow the (somewhat unwritten) rules and this strategy fails.
1701 Now just gather all certificates together and the first private key
1702 then look for the first certificate that matches the key.
1705 *) Support use of registered digest and cipher names for dgst and cipher
1706 commands instead of having to add each one as a special case. So now
1713 openssl dgst -sha256 foo
1715 and this works for ENGINE based algorithms too.
1719 *) Update Gost ENGINE to support parameter files.
1720 [Victor B. Wagner <vitus@cryptocom.ru>]
1722 *) Support GeneralizedTime in ca utility.
1723 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1725 *) Enhance the hash format used for certificate directory links. The new
1726 form uses the canonical encoding (meaning equivalent names will work
1727 even if they aren't identical) and uses SHA1 instead of MD5. This form
1728 is incompatible with the older format and as a result c_rehash should
1729 be used to rebuild symbolic links.
1732 *) Make PKCS#8 the default write format for private keys, replacing the
1733 traditional format. This form is standardised, more secure and doesn't
1734 include an implicit MD5 dependency.
1737 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1738 committed to OpenSSL should pass this lot as a minimum.
1741 *) Add session ticket override functionality for use by EAP-FAST.
1742 [Jouni Malinen <j@w1.fi>]
1744 *) Modify HMAC functions to return a value. Since these can be implemented
1745 in an ENGINE errors can occur.
1748 *) Type-checked OBJ_bsearch_ex.
1751 *) Type-checked OBJ_bsearch. Also some constification necessitated
1752 by type-checking. Still to come: TXT_DB, bsearch(?),
1753 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1757 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1758 seconds to a tm structure directly, instead of going through OS
1759 specific date routines. This avoids any issues with OS routines such
1760 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1761 and X509_time_adj_ex() to cover the extended range. The existing
1762 X509_time_adj() is still usable and will no longer have any date issues.
1765 *) Delta CRL support. New use deltas option which will attempt to locate
1766 and search any appropriate delta CRLs available.
1768 This work was sponsored by Google.
1771 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1772 code and add additional score elements. Validate alternate CRL paths
1773 as part of the CRL checking and indicate a new error "CRL path validation
1774 error" in this case. Applications wanting additional details can use
1775 the verify callback and check the new "parent" field. If this is not
1776 NULL CRL path validation is taking place. Existing applications wont
1777 see this because it requires extended CRL support which is off by
1780 This work was sponsored by Google.
1783 *) Support for freshest CRL extension.
1785 This work was sponsored by Google.
1788 *) Initial indirect CRL support. Currently only supported in the CRLs
1789 passed directly and not via lookup. Process certificate issuer
1790 CRL entry extension and lookup CRL entries by bother issuer name
1791 and serial number. Check and process CRL issuer entry in IDP extension.
1793 This work was sponsored by Google.
1796 *) Add support for distinct certificate and CRL paths. The CRL issuer
1797 certificate is validated separately in this case. Only enabled if
1798 an extended CRL support flag is set: this flag will enable additional
1799 CRL functionality in future.
1801 This work was sponsored by Google.
1804 *) Add support for policy mappings extension.
1806 This work was sponsored by Google.
1809 *) Fixes to pathlength constraint, self issued certificate handling,
1810 policy processing to align with RFC3280 and PKITS tests.
1812 This work was sponsored by Google.
1815 *) Support for name constraints certificate extension. DN, email, DNS
1816 and URI types are currently supported.
1818 This work was sponsored by Google.
1821 *) To cater for systems that provide a pointer-based thread ID rather
1822 than numeric, deprecate the current numeric thread ID mechanism and
1823 replace it with a structure and associated callback type. This
1824 mechanism allows a numeric "hash" to be extracted from a thread ID in
1825 either case, and on platforms where pointers are larger than 'long',
1826 mixing is done to help ensure the numeric 'hash' is usable even if it
1827 can't be guaranteed unique. The default mechanism is to use "&errno"
1828 as a pointer-based thread ID to distinguish between threads.
1830 Applications that want to provide their own thread IDs should now use
1831 CRYPTO_THREADID_set_callback() to register a callback that will call
1832 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1834 Note that ERR_remove_state() is now deprecated, because it is tied
1835 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1836 to free the current thread's error state should be replaced by
1837 ERR_remove_thread_state(NULL).
1839 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1840 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1841 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1842 application was previously providing a numeric thread callback that
1843 was inappropriate for distinguishing threads, then uniqueness might
1844 have been obtained with &errno that happened immediately in the
1845 intermediate development versions of OpenSSL; this is no longer the
1846 case, the numeric thread callback will now override the automatic use
1848 [Geoff Thorpe, with help from Bodo Moeller]
1850 *) Initial support for different CRL issuing certificates. This covers a
1851 simple case where the self issued certificates in the chain exist and
1852 the real CRL issuer is higher in the existing chain.
1854 This work was sponsored by Google.
1857 *) Removed effectively defunct crypto/store from the build.
1860 *) Revamp of STACK to provide stronger type-checking. Still to come:
1861 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1862 ASN1_STRING, CONF_VALUE.
1865 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1866 RAM on SSL connections. This option can save about 34k per idle SSL.
1869 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1870 STACK, TXT_DB, bsearch, qsort.
1873 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1874 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1875 support for data, signedData, compressedData, digestedData and
1876 encryptedData, envelopedData types included. Scripts to check against
1877 RFC4134 examples draft and interop and consistency checks of many
1878 content types and variants.
1881 *) Add options to enc utility to support use of zlib compression BIO.
1884 *) Extend mk1mf to support importing of options and assembly language
1885 files from Configure script, currently only included in VC-WIN32.
1886 The assembly language rules can now optionally generate the source
1887 files from the associated perl scripts.
1890 *) Implement remaining functionality needed to support GOST ciphersuites.
1891 Interop testing has been performed using CryptoPro implementations.
1892 [Victor B. Wagner <vitus@cryptocom.ru>]
1894 *) s390x assembler pack.
1897 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1901 *) Implement Opaque PRF Input TLS extension as specified in
1902 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1903 official specification yet and no extension type assignment by
1904 IANA exists, this extension (for now) will have to be explicitly
1905 enabled when building OpenSSL by providing the extension number
1906 to use. For example, specify an option
1908 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1910 to the "config" or "Configure" script to enable the extension,
1911 assuming extension number 0x9527 (which is a completely arbitrary
1912 and unofficial assignment based on the MD5 hash of the Internet
1913 Draft). Note that by doing so, you potentially lose
1914 interoperability with other TLS implementations since these might
1915 be using the same extension number for other purposes.
1917 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1918 opaque PRF input value to use in the handshake. This will create
1919 an interal copy of the length-'len' string at 'src', and will
1920 return non-zero for success.
1922 To get more control and flexibility, provide a callback function
1925 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1926 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1930 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1933 Callback function 'cb' will be called in handshakes, and is
1934 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1935 Argument 'arg' is for application purposes (the value as given to
1936 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1937 be provided to the callback function). The callback function
1938 has to return non-zero to report success: usually 1 to use opaque
1939 PRF input just if possible, or 2 to enforce use of the opaque PRF
1940 input. In the latter case, the library will abort the handshake
1941 if opaque PRF input is not successfully negotiated.
1943 Arguments 'peerinput' and 'len' given to the callback function
1944 will always be NULL and 0 in the case of a client. A server will
1945 see the client's opaque PRF input through these variables if
1946 available (NULL and 0 otherwise). Note that if the server
1947 provides an opaque PRF input, the length must be the same as the
1948 length of the client's opaque PRF input.
1950 Note that the callback function will only be called when creating
1951 a new session (session resumption can resume whatever was
1952 previously negotiated), and will not be called in SSL 2.0
1953 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1954 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1955 for applications that need to enforce opaque PRF input.
1959 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1962 [Victor B. Wagner <vitus@cryptocom.ru>]
1964 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1965 RFC4507bis. The encrypted ticket format is an encrypted encoded
1966 SSL_SESSION structure, that way new session features are automatically
1969 If a client application caches session in an SSL_SESSION structure
1970 support is transparent because tickets are now stored in the encoded
1973 The SSL_CTX structure automatically generates keys for ticket
1974 protection in servers so again support should be possible
1975 with no application modification.
1977 If a client or server wishes to disable RFC4507 support then the option
1978 SSL_OP_NO_TICKET can be set.
1980 Add a TLS extension debugging callback to allow the contents of any client
1981 or server extensions to be examined.
1983 This work was sponsored by Google.
1986 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1987 OpenSSL should now compile cleanly on gcc 4.2
1988 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1990 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1991 support including streaming MAC support: this is required for GOST
1992 ciphersuite support.
1993 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1995 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1996 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1997 to output in BER and PEM format.
2000 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2001 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2002 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2003 ENGINE support for HMAC keys which are unextractable. New -mac and
2004 -macopt options to dgst utility.
2007 *) New option -sigopt to dgst utility. Update dgst to use
2008 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2009 alternative signing paramaters such as X9.31 or PSS in the dgst
2013 *) Change ssl_cipher_apply_rule(), the internal function that does
2014 the work each time a ciphersuite string requests enabling
2015 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2016 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2017 the order of disabled ciphersuites such that those ciphersuites
2018 that most recently went from enabled to disabled not only stay
2019 in order with respect to each other, but also have higher priority
2020 than other disabled ciphersuites the next time ciphersuites are
2023 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2024 the same ciphersuites as with "HIGH" alone, but in a specific
2025 order where the PSK ciphersuites come first (since they are the
2026 most recently disabled ciphersuites when "HIGH" is parsed).
2028 Also, change ssl_create_cipher_list() (using this new
2029 funcionality) such that between otherwise identical
2030 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2034 *) Change ssl_create_cipher_list() so that it automatically
2035 arranges the ciphersuites in reasonable order before starting
2036 to process the rule string. Thus, the definition for "DEFAULT"
2037 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2038 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2039 This makes it much easier to arrive at a reasonable default order
2040 in applications for which anonymous ciphers are OK (meaning
2041 that you can't actually use DEFAULT).
2042 [Bodo Moeller; suggested by Victor Duchovni]
2044 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2045 processing) into multiple integers instead of setting
2046 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2047 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2048 (These masks as well as the individual bit definitions are hidden
2049 away into the non-exported interface ssl/ssl_locl.h, so this
2050 change to the definition of the SSL_CIPHER structure shouldn't
2051 affect applications.) This give us more bits for each of these
2052 categories, so there is no longer a need to coagulate AES128 and
2053 AES256 into a single algorithm bit, and to coagulate Camellia128
2054 and Camellia256 into a single algorithm bit, which has led to all
2057 Thus, among other things, the kludge introduced in 0.9.7m and
2058 0.9.8e for masking out AES256 independently of AES128 or masking
2059 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2061 With the change, we also introduce new ciphersuite aliases that
2062 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2066 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2067 Use the leftmost N bytes of the signature input if the input is
2068 larger than the prime q (with N being the size in bytes of q).
2071 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2072 it yet and it is largely untested.
2075 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2078 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2079 some compilers (gcc 4.2 and later) reject their use. Safestack is
2080 reimplemented. Update ASN1 to avoid use of legacy functions.
2083 *) Win32/64 targets are linked with Winsock2.
2086 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2087 to external functions. This can be used to increase CRL handling
2088 efficiency especially when CRLs are very large by (for example) storing
2089 the CRL revoked certificates in a database.
2092 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2093 new CRLs added to a directory can be used. New command line option
2094 -verify_return_error to s_client and s_server. This causes real errors
2095 to be returned by the verify callback instead of carrying on no matter
2096 what. This reflects the way a "real world" verify callback would behave.
2099 *) GOST engine, supporting several GOST algorithms and public key formats.
2100 Kindly donated by Cryptocom.
2103 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2104 partitioned by DP are handled but no indirect CRL or reason partitioning
2105 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2106 selected via a scoring technique which handles IDP and AKID in CRLs.
2109 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2110 will ultimately be used for all verify operations: this will remove the
2111 X509_STORE dependency on certificate verification and allow alternative
2112 lookup methods. X509_STORE based implementations of these two callbacks.
2115 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2116 Modify get_crl() to find a valid (unexpired) CRL if possible.
2119 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2120 this would be called X509_CRL_cmp() but that name is already used by
2121 a function that just compares CRL issuer names. Cache several CRL
2122 extensions in X509_CRL structure and cache CRLDP in X509.
2125 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2126 this maps equivalent X509_NAME structures into a consistent structure.
2127 Name comparison can then be performed rapidly using memcmp().
2130 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2134 *) Allow digests to supply their own micalg string for S/MIME type using
2135 the ctrl EVP_MD_CTRL_MICALG.
2138 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2139 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2140 ctrl. It can then customise the structure before and/or after signing
2144 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2145 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2146 to free up any added signature OIDs.
2149 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2150 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2151 digest and cipher tables. New options added to openssl utility:
2152 list-message-digest-algorithms and list-cipher-algorithms.
2155 *) Change the array representation of binary polynomials: the list
2156 of degrees of non-zero coefficients is now terminated with -1.
2157 Previously it was terminated with 0, which was also part of the
2158 value; thus, the array representation was not applicable to
2159 polynomials where t^0 has coefficient zero. This change makes
2160 the array representation useful in a more general context.
2163 *) Various modifications and fixes to SSL/TLS cipher string
2164 handling. For ECC, the code now distinguishes between fixed ECDH
2165 with RSA certificates on the one hand and with ECDSA certificates
2166 on the other hand, since these are separate ciphersuites. The
2167 unused code for Fortezza ciphersuites has been removed.
2169 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2170 (not "ECDHE"). For consistency with the code for DH
2171 certificates, use of ECDH certificates is now considered ECDH
2172 authentication, not RSA or ECDSA authentication (the latter is
2173 merely the CA's signing algorithm and not actively used in the
2176 The temporary ciphersuite alias "ECCdraft" is no longer
2177 available, and ECC ciphersuites are no longer excluded from "ALL"
2178 and "DEFAULT". The following aliases now exist for RFC 4492
2179 ciphersuites, most of these by analogy with the DH case:
2181 kECDHr - ECDH cert, signed with RSA
2182 kECDHe - ECDH cert, signed with ECDSA
2183 kECDH - ECDH cert (signed with either RSA or ECDSA)
2184 kEECDH - ephemeral ECDH
2185 ECDH - ECDH cert or ephemeral ECDH
2191 AECDH - anonymous ECDH
2192 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2196 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2197 Use correct micalg parameters depending on digest(s) in signed message.
2200 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2201 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2204 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2205 an engine to register a method. Add ENGINE lookups for methods and
2206 functional reference processing.
2209 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2210 EVP_{Sign,Verify}* which allow an application to customise the signature
2214 *) New -resign option to smime utility. This adds one or more signers
2215 to an existing PKCS#7 signedData structure. Also -md option to use an
2216 alternative message digest algorithm for signing.
2219 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2220 create PKCS7 structures containing multiple signers. Update smime
2221 application to support multiple signers.
2224 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2228 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2229 Reorganize PBE internals to lookup from a static table using NIDs,
2230 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2231 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2232 PRF which will be automatically used with PBES2.
2235 *) Replace the algorithm specific calls to generate keys in "req" with the
2239 *) Update PKCS#7 enveloped data routines to use new API. This is now
2240 supported by any public key method supporting the encrypt operation. A
2241 ctrl is added to allow the public key algorithm to examine or modify
2242 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2246 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2247 a default digest type to use. In most cases this will be SHA1 but some
2248 algorithms (such as GOST) need to specify an alternative digest. The
2249 return value indicates how strong the preference is 1 means optional and
2250 2 is mandatory (that is it is the only supported type). Modify
2251 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2252 use the default md. Update openssl utilities to use the default digest
2253 type for signing if it is not explicitly indicated.
2256 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2257 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2258 signing method from the key type. This effectively removes the link
2259 between digests and public key types.
2262 *) Add an OID cross reference table and utility functions. Its purpose is to
2263 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2264 rsaEncryption. This will allow some of the algorithm specific hackery
2265 needed to use the correct OID to be removed.
2268 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2269 structures for PKCS7_sign(). They are now set up by the relevant public
2273 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2276 *) Add support for key derivation (agreement) in the API, DH method and
2280 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2281 public and private key formats. As a side effect these add additional
2282 command line functionality not previously available: DSA signatures can be
2283 generated and verified using pkeyutl and DH key support and generation in
2288 [Oliver Tappe <zooey@hirschkaefer.de>]
2290 *) New make target "install_html_docs" installs HTML renditions of the
2292 [Oliver Tappe <zooey@hirschkaefer.de>]
2294 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2295 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2296 support key and parameter generation and add initial key generation
2297 functionality for RSA.
2300 *) Add functions for main EVP_PKEY_method operations. The undocumented
2301 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2302 EVP_PKEY_{encrypt,decrypt}_old.
2305 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2306 key API, doesn't do much yet.
2309 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2310 public key algorithms. New option to openssl utility:
2311 "list-public-key-algorithms" to print out info.
2314 *) Implement the Supported Elliptic Curves Extension for
2315 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2318 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2319 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2322 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2323 utilities such as rsa, dsa, dsaparam etc except they process any key
2327 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2328 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2329 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2333 *) Initial support for pluggable public key ASN1.
2334 De-spaghettify the public key ASN1 handling. Move public and private
2335 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2336 algorithm specific handling to a single module within the relevant
2337 algorithm directory. Add functions to allow (near) opaque processing
2338 of public and private key structures.
2341 *) Implement the Supported Point Formats Extension for
2342 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2345 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2346 for the psk identity [hint] and the psk callback functions to the
2347 SSL_SESSION, SSL and SSL_CTX structure.
2350 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2354 SSL_CTX_use_psk_identity_hint
2355 SSL_get_psk_identity_hint
2356 SSL_get_psk_identity
2357 SSL_use_psk_identity_hint
2359 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2361 *) Add RFC 3161 compliant time stamp request creation, response generation
2362 and response verification functionality.
2363 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2365 *) Add initial support for TLS extensions, specifically for the server_name
2366 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2367 have new members for a host name. The SSL data structure has an
2368 additional member SSL_CTX *initial_ctx so that new sessions can be
2369 stored in that context to allow for session resumption, even after the
2370 SSL has been switched to a new SSL_CTX in reaction to a client's
2371 server_name extension.
2373 New functions (subject to change):
2375 SSL_get_servername()
2376 SSL_get_servername_type()
2379 New CTRL codes and macros (subject to change):
2381 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2382 - SSL_CTX_set_tlsext_servername_callback()
2383 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2384 - SSL_CTX_set_tlsext_servername_arg()
2385 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2387 openssl s_client has a new '-servername ...' option.
2389 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2390 '-key2 ...', '-servername_fatal' (subject to change). This allows
2391 testing the HostName extension for a specific single host name ('-cert'
2392 and '-key' remain fallbacks for handshakes without HostName
2393 negotiation). If the unrecognized_name alert has to be sent, this by
2394 default is a warning; it becomes fatal with the '-servername_fatal'
2397 [Peter Sylvester, Remy Allais, Christophe Renou]
2399 *) Whirlpool hash implementation is added.
2402 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2403 bn(64,32). Because of instruction set limitations it doesn't have
2404 any negative impact on performance. This was done mostly in order
2405 to make it possible to share assembler modules, such as bn_mul_mont
2406 implementations, between 32- and 64-bit builds without hassle.
2409 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2410 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2414 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2415 dedicated Montgomery multiplication procedure, is introduced.
2416 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2417 "64-bit" performance on certain 32-bit targets.
2420 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2421 in SSL structures. New SSL ctrl to set maximum send fragment size.
2422 Save memory by seeting the I/O buffer sizes dynamically instead of
2423 using the maximum available value.
2426 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2427 in addition to the text details.
2430 *) Very, very preliminary EXPERIMENTAL support for printing of general
2431 ASN1 structures. This currently produces rather ugly output and doesn't
2432 handle several customised structures at all.
2435 *) Integrated support for PVK file format and some related formats such
2436 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2437 these in the 'rsa' and 'dsa' utilities.
2440 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2443 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2444 place for the (very old) "NETSCAPE" format certificates which are now
2445 handled using new ASN1 code equivalents.
2448 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2449 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2450 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2453 *) Modify CRL distribution points extension code to print out previously
2454 unsupported fields. Enhance extension setting code to allow setting of
2458 *) Add print and set support for Issuing Distribution Point CRL extension.
2461 *) Change 'Configure' script to enable Camellia by default.
2464 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2466 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2468 This addresses the flaw in CBC record processing discovered by
2469 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2470 at: http://www.isg.rhul.ac.uk/tls/
2472 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2473 Security Group at Royal Holloway, University of London
2474 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2475 Emilia Käsper for the initial patch.
2477 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2479 *) Return an error when checking OCSP signatures when key is NULL.
2480 This fixes a DoS attack. (CVE-2013-0166)
2483 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2484 the right response is stapled. Also change SSL_get_certificate()
2485 so it returns the certificate actually sent.
2486 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2487 (This is a backport)
2488 [Rob Stradling <rob.stradling@comodo.com>]
2490 *) Fix possible deadlock when decoding public keys.
2493 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2495 *) Sanity check record length before skipping explicit IV in DTLS
2498 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2499 fuzzing as a service testing platform.
2503 *) Initialise tkeylen properly when encrypting CMS messages.
2504 Thanks to Solar Designer of Openwall for reporting this issue.
2507 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2509 *) The fix for CVE-2012-2110 did not take into account that the
2510 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2511 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2512 rejecting negative len parameter. (CVE-2012-2131)
2513 [Tomas Hoger <thoger@redhat.com>]
2515 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2517 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2518 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2519 in CRYPTO_realloc_clean.
2521 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2522 issue and to Adam Langley <agl@chromium.org> for fixing it.
2524 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2526 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2528 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2529 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2530 content decryption and always return the same error. Note: this attack
2531 needs on average 2^20 messages so it only affects automated senders. The
2532 old behaviour can be reenabled in the CMS code by setting the
2533 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2534 an MMA defence is not necessary.
2535 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2536 this issue. (CVE-2012-0884)
2539 *) Fix CVE-2011-4619: make sure we really are receiving a
2540 client hello before rejecting multiple SGC restarts. Thanks to
2541 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2544 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2546 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2547 Thanks to Antonio Martin, Enterprise Secure Access Research and
2548 Development, Cisco Systems, Inc. for discovering this bug and
2549 preparing a fix. (CVE-2012-0050)
2552 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2554 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2555 of the Vaudenay padding oracle attack on CBC mode encryption
2556 which enables an efficient plaintext recovery attack against
2557 the OpenSSL implementation of DTLS. Their attack exploits timing
2558 differences arising during decryption processing. A research
2559 paper describing this attack can be found at:
2560 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2561 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2562 Security Group at Royal Holloway, University of London
2563 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2564 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2565 for preparing the fix. (CVE-2011-4108)
2566 [Robin Seggelmann, Michael Tuexen]
2568 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2569 [Ben Laurie, Kasper <ekasper@google.com>]
2571 *) Clear bytes used for block padding of SSL 3.0 records.
2573 [Adam Langley (Google)]
2575 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2576 Kadianakis <desnacked@gmail.com> for discovering this issue and
2577 Adam Langley for preparing the fix. (CVE-2011-4619)
2578 [Adam Langley (Google)]
2580 *) Prevent malformed RFC3779 data triggering an assertion failure.
2581 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2582 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2583 [Rob Austein <sra@hactrn.net>]
2585 *) Fix ssl_ciph.c set-up race.
2586 [Adam Langley (Google)]
2588 *) Fix spurious failures in ecdsatest.c.
2589 [Emilia Käsper (Google)]
2591 *) Fix the BIO_f_buffer() implementation (which was mixing different
2592 interpretations of the '..._len' fields).
2593 [Adam Langley (Google)]
2595 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2596 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2597 threads won't reuse the same blinding coefficients.
2599 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2600 lock to call BN_BLINDING_invert_ex, and avoids one use of
2601 BN_BLINDING_update for each BN_BLINDING structure (previously,
2602 the last update always remained unused).
2603 [Emilia Käsper (Google)]
2605 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2606 for multi-threaded use of ECDH.
2607 [Adam Langley (Google)]
2609 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2612 *) Add protection against ECDSA timing attacks as mentioned in the paper
2613 by Billy Bob Brumley and Nicola Tuveri, see:
2615 http://eprint.iacr.org/2011/232.pdf
2617 [Billy Bob Brumley and Nicola Tuveri]
2619 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2621 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2622 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2624 *) Fix bug in string printing code: if *any* escaping is enabled we must
2625 escape the escape character (backslash) or the resulting string is
2629 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2631 *) Disable code workaround for ancient and obsolete Netscape browsers
2632 and servers: an attacker can use it in a ciphersuite downgrade attack.
2633 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2636 *) Fixed J-PAKE implementation error, originally discovered by
2637 Sebastien Martini, further info and confirmation from Stefan
2638 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2641 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2643 *) Fix extension code to avoid race conditions which can result in a buffer
2644 overrun vulnerability: resumed sessions must not be modified as they can
2645 be shared by multiple threads. CVE-2010-3864
2648 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2651 *) Don't reencode certificate when calculating signature: cache and use
2652 the original encoding instead. This makes signature verification of
2653 some broken encodings work correctly.
2656 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2657 is also one of the inputs.
2658 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2660 *) Don't repeatedly append PBE algorithms to table if they already exist.
2661 Sort table on each new add. This effectively makes the table read only
2662 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2666 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2668 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2671 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2672 access or freeing data twice (CVE-2010-0742)
2673 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2675 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2676 common in certificates and some applications which only call
2677 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2681 Reduce copying into .apps and .test in makevms.com
2682 Don't try to use blank CA certificate in CA.com
2683 Allow use of C files from original directories in maketests.com
2684 [Steven M. Schweda" <sms@antinode.info>]
2686 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2688 *) When rejecting SSL/TLS records due to an incorrect version number, never
2689 update s->server with a new major version number. As of
2690 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2691 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2692 the previous behavior could result in a read attempt at NULL when
2693 receiving specific incorrect SSL/TLS records once record payload
2694 protection is active. (CVE-2010-0740)
2695 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2697 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2698 could be crashed if the relevant tables were not present (e.g. chrooted).
2699 [Tomas Hoger <thoger@redhat.com>]
2701 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2703 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2704 [Martin Olsson, Neel Mehta]
2706 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2707 accommodate for stack sorting, always a write lock!).
2710 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2711 excessive delays in the RAND_poll(): over a minute. As a workaround
2712 include a time check in the inner Heap32Next loop too.
2715 *) The code that handled flushing of data in SSL/TLS originally used the
2716 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2717 the problem outlined in PR#1949. The fix suggested there however can
2718 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2719 of Apache). So instead simplify the code to flush unconditionally.
2720 This should be fine since flushing with no data to flush is a no op.
2723 *) Handle TLS versions 2.0 and later properly and correctly use the
2724 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2725 off ancient servers have a habit of sticking around for a while...
2728 *) Modify compression code so it frees up structures without using the
2729 ex_data callbacks. This works around a problem where some applications
2730 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2731 restarting) then use compression (e.g. SSL with compression) later.
2732 This results in significant per-connection memory leaks and
2733 has caused some security issues including CVE-2008-1678 and
2737 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2738 change when encrypting or decrypting.
2741 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2742 connect and renegotiate with servers which do not support RI.
2743 Until RI is more widely deployed this option is enabled by default.
2746 *) Add "missing" ssl ctrls to clear options and mode.
2749 *) If client attempts to renegotiate and doesn't support RI respond with
2750 a no_renegotiation alert as required by RFC5746. Some renegotiating
2751 TLS clients will continue a connection gracefully when they receive
2752 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2753 waiting for a server hello which it will never receive. Now we treat a
2754 received no_renegotiation alert as a fatal error. This is because
2755 applications requesting a renegotiation might well expect it to succeed
2756 and would have no code in place to handle the server denying it so the
2757 only safe thing to do is to terminate the connection.
2760 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2761 peer supports secure renegotiation and 0 otherwise. Print out peer
2762 renegotiation support in s_client/s_server.
2765 *) Replace the highly broken and deprecated SPKAC certification method with
2766 the updated NID creation version. This should correctly handle UTF8.
2769 *) Implement RFC5746. Re-enable renegotiation but require the extension
2770 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2771 turns out to be a bad idea. It has been replaced by
2772 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2773 SSL_CTX_set_options(). This is really not recommended unless you
2774 know what you are doing.
2775 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2777 *) Fixes to stateless session resumption handling. Use initial_ctx when
2778 issuing and attempting to decrypt tickets in case it has changed during
2779 servername handling. Use a non-zero length session ID when attempting
2780 stateless session resumption: this makes it possible to determine if
2781 a resumption has occurred immediately after receiving server hello
2782 (several places in OpenSSL subtly assume this) instead of later in
2786 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2787 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2788 fixes for a few places where the return code is not checked
2790 [Julia Lawall <julia@diku.dk>]
2792 *) Add --strict-warnings option to Configure script to include devteam
2793 warnings in other configurations.
2796 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2797 makes it possible to install openssl libraries in locations which
2798 have names other than "lib", for example "/usr/lib64" which some
2800 [Steve Henson, based on patch from Jeremy Utley]
2802 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2803 X690 8.9.12 and can produce some misleading textual output of OIDs.
2804 [Steve Henson, reported by Dan Kaminsky]
2806 *) Delete MD2 from algorithm tables. This follows the recommendation in
2807 several standards that it is not used in new applications due to
2808 several cryptographic weaknesses. For binary compatibility reasons
2809 the MD2 API is still compiled in by default.
2812 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2816 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2817 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2819 [Guenter <lists@gknw.net>]
2821 *) Fix the server certificate chain building code to use X509_verify_cert(),
2822 it used to have an ad-hoc builder which was unable to cope with anything
2823 other than a simple chain.
2824 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2826 *) Don't check self signed certificate signatures in X509_verify_cert()
2827 by default (a flag can override this): it just wastes time without
2828 adding any security. As a useful side effect self signed root CAs
2829 with non-FIPS digests are now usable in FIPS mode.
2832 *) In dtls1_process_out_of_seq_message() the check if the current message
2833 is already buffered was missing. For every new message was memory
2834 allocated, allowing an attacker to perform an denial of service attack
2835 with sending out of seq handshake messages until there is no memory
2836 left. Additionally every future messege was buffered, even if the
2837 sequence number made no sense and would be part of another handshake.
2838 So only messages with sequence numbers less than 10 in advance will be
2839 buffered. (CVE-2009-1378)
2840 [Robin Seggelmann, discovered by Daniel Mentz]
2842 *) Records are buffered if they arrive with a future epoch to be
2843 processed after finishing the corresponding handshake. There is
2844 currently no limitation to this buffer allowing an attacker to perform
2845 a DOS attack with sending records with future epochs until there is no
2846 memory left. This patch adds the pqueue_size() function to determine
2847 the size of a buffer and limits the record buffer to 100 entries.
2849 [Robin Seggelmann, discovered by Daniel Mentz]
2851 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2852 parent structure is freed. (CVE-2009-1379)
2855 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2856 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2859 [Ilya O. <vrghost@gmail.com>]
2861 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2863 *) Disable renegotiation completely - this fixes a severe security
2864 problem (CVE-2009-3555) at the cost of breaking all
2865 renegotiation. Renegotiation can be re-enabled by setting
2866 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2867 run-time. This is really not recommended unless you know what
2871 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2873 *) Don't set val to NULL when freeing up structures, it is freed up by
2874 underlying code. If sizeof(void *) > sizeof(long) this can result in
2875 zeroing past the valid field. (CVE-2009-0789)
2876 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2878 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2879 checked correctly. This would allow some invalid signed attributes to
2880 appear to verify correctly. (CVE-2009-0591)
2881 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2883 *) Reject UniversalString and BMPString types with invalid lengths. This
2884 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2885 a legal length. (CVE-2009-0590)
2888 *) Set S/MIME signing as the default purpose rather than setting it
2889 unconditionally. This allows applications to override it at the store
2893 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2894 to handle some structures.
2897 *) Improve efficiency of mem_gets: don't search whole buffer each time
2899 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2901 *) New -hex option for openssl rand.
2904 *) Print out UTF8String and NumericString when parsing ASN1.
2907 *) Support NumericString type for name components.
2910 *) Allow CC in the environment to override the automatically chosen
2911 compiler. Note that nothing is done to ensure flags work with the
2915 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2917 *) Properly check EVP_VerifyFinal() and similar return values
2919 [Ben Laurie, Bodo Moeller, Google Security Team]
2921 *) Enable TLS extensions by default.
2924 *) Allow the CHIL engine to be loaded, whether the application is
2925 multithreaded or not. (This does not release the developer from the
2926 obligation to set up the dynamic locking callbacks.)
2927 [Sander Temme <sander@temme.net>]
2929 *) Use correct exit code if there is an error in dgst command.
2930 [Steve Henson; problem pointed out by Roland Dirlewanger]
2932 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2933 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2936 *) Add experimental JPAKE support, including demo authentication in
2937 s_client and s_server.
2940 *) Set the comparison function in v3_addr_canonize().
2941 [Rob Austein <sra@hactrn.net>]
2943 *) Add support for XMPP STARTTLS in s_client.
2944 [Philip Paeps <philip@freebsd.org>]
2946 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2947 to ensure that even with this option, only ciphersuites in the
2948 server's preference list will be accepted. (Note that the option
2949 applies only when resuming a session, so the earlier behavior was
2950 just about the algorithm choice for symmetric cryptography.)
2953 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2955 *) Fix NULL pointer dereference if a DTLS server received
2956 ChangeCipherSpec as first record (CVE-2009-1386).
2959 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2960 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2963 *) The fix in 0.9.8c that supposedly got rid of unsafe
2964 double-checked locking was incomplete for RSA blinding,
2965 addressing just one layer of what turns out to have been
2966 doubly unsafe triple-checked locking.
2968 So now fix this for real by retiring the MONT_HELPER macro
2969 in crypto/rsa/rsa_eay.c.
2971 [Bodo Moeller; problem pointed out by Marius Schilder]
2973 *) Various precautionary measures:
2975 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2977 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2978 (NB: This would require knowledge of the secret session ticket key
2979 to exploit, in which case you'd be SOL either way.)
2981 - Change bn_nist.c so that it will properly handle input BIGNUMs
2982 outside the expected range.
2984 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2987 [Neel Mehta, Bodo Moeller]
2989 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2990 the load fails. Useful for distros.
2991 [Ben Laurie and the FreeBSD team]
2993 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2996 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2999 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3001 This work was sponsored by Logica.
3004 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3005 keystores. Support for SSL/TLS client authentication too.
3006 Not compiled unless enable-capieng specified to Configure.
3008 This work was sponsored by Logica.
3011 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3012 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3013 attribute creation routines such as certifcate requests and PKCS#12
3017 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3019 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3020 handshake which could lead to a cilent crash as found using the
3021 Codenomicon TLS test suite (CVE-2008-1672)
3022 [Steve Henson, Mark Cox]
3024 *) Fix double free in TLS server name extensions which could lead to
3025 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3028 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3030 Clear the error queue to ensure that error entries left from
3031 older function calls do not interfere with the correct operation.
3032 [Lutz Jaenicke, Erik de Castro Lopo]
3034 *) Remove root CA certificates of commercial CAs:
3036 The OpenSSL project does not recommend any specific CA and does not
3037 have any policy with respect to including or excluding any CA.
3038 Therefore it does not make any sense to ship an arbitrary selection
3039 of root CA certificates with the OpenSSL software.
3042 *) RSA OAEP patches to fix two separate invalid memory reads.
3043 The first one involves inputs when 'lzero' is greater than
3044 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3045 before the beginning of from). The second one involves inputs where
3046 the 'db' section contains nothing but zeroes (there is a one-byte
3047 invalid read after the end of 'db').
3048 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3050 *) Partial backport from 0.9.9-dev:
3052 Introduce bn_mul_mont (dedicated Montgomery multiplication
3053 procedure) as a candidate for BIGNUM assembler implementation.
3054 While 0.9.9-dev uses assembler for various architectures, only
3055 x86_64 is available by default here in the 0.9.8 branch, and
3056 32-bit x86 is available through a compile-time setting.
3058 To try the 32-bit x86 assembler implementation, use Configure
3059 option "enable-montasm" (which exists only for this backport).
3061 As "enable-montasm" for 32-bit x86 disclaims code stability
3062 anyway, in this constellation we activate additional code
3063 backported from 0.9.9-dev for further performance improvements,
3064 namely BN_from_montgomery_word. (To enable this otherwise,
3065 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3067 [Andy Polyakov (backport partially by Bodo Moeller)]
3069 *) Add TLS session ticket callback. This allows an application to set
3070 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3071 values. This is useful for key rollover for example where several key
3072 sets may exist with different names.
3075 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3076 This was broken until now in 0.9.8 releases, such that the only way
3077 a registered ENGINE could be used (assuming it initialises
3078 successfully on the host) was to explicitly set it as the default
3079 for the relevant algorithms. This is in contradiction with 0.9.7
3080 behaviour and the documentation. With this fix, when an ENGINE is
3081 registered into a given algorithm's table of implementations, the
3082 'uptodate' flag is reset so that auto-discovery will be used next
3083 time a new context for that algorithm attempts to select an
3085 [Ian Lister (tweaked by Geoff Thorpe)]
3087 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3088 implemention in the following ways:
3090 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3093 Lack of BER streaming support means one pass streaming processing is
3094 only supported if data is detached: setting the streaming flag is
3095 ignored for embedded content.
3097 CMS support is disabled by default and must be explicitly enabled
3098 with the enable-cms configuration option.
3101 *) Update the GMP engine glue to do direct copies between BIGNUM and
3102 mpz_t when openssl and GMP use the same limb size. Otherwise the
3103 existing "conversion via a text string export" trick is still used.
3104 [Paul Sheer <paulsheer@gmail.com>]
3106 *) Zlib compression BIO. This is a filter BIO which compressed and
3107 uncompresses any data passed through it.
3110 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3111 RFC3394 compatible AES key wrapping.
3114 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3115 sets string data without copying. X509_ALGOR_set0() and
3116 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3117 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3118 from an X509_ATTRIBUTE structure optionally checking it occurs only
3119 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3123 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3124 to get the expected BN_FLG_CONSTTIME behavior.
3125 [Bodo Moeller (Google)]
3129 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3130 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3131 - added some more tests to do_tests.pl
3132 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3133 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3134 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3135 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3136 - various changes to netware.pl to enable gcc-cross builds on Win32
3138 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3139 - various changes to fix missing prototype warnings
3140 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3141 - added AES, WHIRLPOOL and CPUID assembler code to build files
3142 - added missing AES assembler make rules to mk1mf.pl
3143 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3144 [Guenter Knauf <eflash@gmx.net>]
3146 *) Implement certificate status request TLS extension defined in RFC3546.
3147 A client can set the appropriate parameters and receive the encoded
3148 OCSP response via a callback. A server can query the supplied parameters
3149 and set the encoded OCSP response in the callback. Add simplified examples
3150 to s_client and s_server.
3153 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3155 *) Fix various bugs:
3156 + Binary incompatibility of ssl_ctx_st structure
3157 + DTLS interoperation with non-compliant servers
3158 + Don't call get_session_cb() without proposed session
3159 + Fix ia64 assembler code
3160 [Andy Polyakov, Steve Henson]
3162 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3164 *) DTLS Handshake overhaul. There were longstanding issues with
3165 OpenSSL DTLS implementation, which were making it impossible for
3166 RFC 4347 compliant client to communicate with OpenSSL server.
3167 Unfortunately just fixing these incompatibilities would "cut off"
3168 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3169 server keeps tolerating non RFC compliant syntax. The opposite is
3170 not true, 0.9.8f client can not communicate with earlier server.
3171 This update even addresses CVE-2007-4995.
3174 *) Changes to avoid need for function casts in OpenSSL: some compilers
3175 (gcc 4.2 and later) reject their use.
3176 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3179 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3180 RFC4507bis. The encrypted ticket format is an encrypted encoded
3181 SSL_SESSION structure, that way new session features are automatically
3184 If a client application caches session in an SSL_SESSION structure
3185 support is transparent because tickets are now stored in the encoded
3188 The SSL_CTX structure automatically generates keys for ticket
3189 protection in servers so again support should be possible
3190 with no application modification.
3192 If a client or server wishes to disable RFC4507 support then the option
3193 SSL_OP_NO_TICKET can be set.
3195 Add a TLS extension debugging callback to allow the contents of any client
3196 or server extensions to be examined.
3198 This work was sponsored by Google.
3201 *) Add initial support for TLS extensions, specifically for the server_name
3202 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3203 have new members for a host name. The SSL data structure has an
3204 additional member SSL_CTX *initial_ctx so that new sessions can be
3205 stored in that context to allow for session resumption, even after the
3206 SSL has been switched to a new SSL_CTX in reaction to a client's
3207 server_name extension.
3209 New functions (subject to change):
3211 SSL_get_servername()
3212 SSL_get_servername_type()
3215 New CTRL codes and macros (subject to change):
3217 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3218 - SSL_CTX_set_tlsext_servername_callback()
3219 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3220 - SSL_CTX_set_tlsext_servername_arg()
3221 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3223 openssl s_client has a new '-servername ...' option.
3225 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3226 '-key2 ...', '-servername_fatal' (subject to change). This allows
3227 testing the HostName extension for a specific single host name ('-cert'
3228 and '-key' remain fallbacks for handshakes without HostName
3229 negotiation). If the unrecognized_name alert has to be sent, this by
3230 default is a warning; it becomes fatal with the '-servername_fatal'
3233 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3235 *) Add AES and SSE2 assembly language support to VC++ build.
3238 *) Mitigate attack on final subtraction in Montgomery reduction.
3241 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3242 (which previously caused an internal error).
3245 *) Squeeze another 10% out of IGE mode when in != out.
3248 *) AES IGE mode speedup.
3249 [Dean Gaudet (Google)]
3251 *) Add the Korean symmetric 128-bit cipher SEED (see
3252 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3253 add SEED ciphersuites from RFC 4162:
3255 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3256 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3257 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3258 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3260 To minimize changes between patchlevels in the OpenSSL 0.9.8
3261 series, SEED remains excluded from compilation unless OpenSSL
3262 is configured with 'enable-seed'.
3263 [KISA, Bodo Moeller]
3265 *) Mitigate branch prediction attacks, which can be practical if a
3266 single processor is shared, allowing a spy process to extract
3267 information. For detailed background information, see
3268 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3269 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3270 and Necessary Software Countermeasures"). The core of the change
3271 are new versions BN_div_no_branch() and
3272 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3273 respectively, which are slower, but avoid the security-relevant
3274 conditional branches. These are automatically called by BN_div()
3275 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3276 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3277 remove a conditional branch.
3279 BN_FLG_CONSTTIME is the new name for the previous
3280 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3281 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3282 in the exponent causes BN_mod_exp_mont() to use the alternative
3283 implementation in BN_mod_exp_mont_consttime().) The old name
3284 remains as a deprecated alias.
3286 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3287 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3288 constant-time implementations for more than just exponentiation.
3289 Here too the old name is kept as a deprecated alias.
3291 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3292 the BN_BLINDING structure gets an independent copy of the
3293 modulus. This means that the previous "BIGNUM *m" argument to
3294 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3295 essentially becomes "const BIGNUM *m", although we can't actually
3296 change this in the header file before 0.9.9. It allows
3297 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3298 enable BN_FLG_CONSTTIME.
3300 [Matthew D Wood (Intel Corp)]
3302 *) In the SSL/TLS server implementation, be strict about session ID
3303 context matching (which matters if an application uses a single
3304 external cache for different purposes). Previously,
3305 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3306 set. This did ensure strict client verification, but meant that,
3307 with applications using a single external cache for quite
3308 different requirements, clients could circumvent ciphersuite
3309 restrictions for a given session ID context by starting a session
3310 in a different context.
3313 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3314 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3315 authentication-only ciphersuites.
3318 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3319 not complete and could lead to a possible single byte overflow
3320 (CVE-2007-5135) [Ben Laurie]
3322 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3324 *) Since AES128 and AES256 (and similarly Camellia128 and
3325 Camellia256) share a single mask bit in the logic of
3326 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3327 kludge to work properly if AES128 is available and AES256 isn't
3328 (or if Camellia128 is available and Camellia256 isn't).
3331 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3332 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3333 When a point or a seed is encoded in a BIT STRING, we need to
3334 prevent the removal of trailing zero bits to get the proper DER
3335 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3336 of a NamedBitList, for which trailing 0 bits need to be removed.)
3339 *) Have SSL/TLS server implementation tolerate "mismatched" record
3340 protocol version while receiving ClientHello even if the
3341 ClientHello is fragmented. (The server can't insist on the
3342 particular protocol version it has chosen before the ServerHello
3343 message has informed the client about his choice.)
3346 *) Add RFC 3779 support.
3347 [Rob Austein for ARIN, Ben Laurie]
3349 *) Load error codes if they are not already present instead of using a
3350 static variable. This allows them to be cleanly unloaded and reloaded.
3351 Improve header file function name parsing.
3354 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3355 or CAPABILITY handshake as required by RFCs.
3358 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3360 *) Introduce limits to prevent malicious keys being able to
3361 cause a denial of service. (CVE-2006-2940)
3362 [Steve Henson, Bodo Moeller]
3364 *) Fix ASN.1 parsing of certain invalid structures that can result
3365 in a denial of service. (CVE-2006-2937) [Steve Henson]
3367 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3368 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3370 *) Fix SSL client code which could crash if connecting to a
3371 malicious SSLv2 server. (CVE-2006-4343)
3372 [Tavis Ormandy and Will Drewry, Google Security Team]
3374 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3375 match only those. Before that, "AES256-SHA" would be interpreted
3376 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3377 the same strength classification in 0.9.7h) as we currently only
3378 have a single AES bit in the ciphersuite description bitmap.
3379 That change, however, also applied to ciphersuite strings such as
3380 "RC4-MD5" that intentionally matched multiple ciphersuites --
3381 namely, SSL 2.0 ciphersuites in addition to the more common ones
3382 from SSL 3.0/TLS 1.0.
3384 So we change the selection algorithm again: Naming an explicit
3385 ciphersuite selects this one ciphersuite, and any other similar
3386 ciphersuite (same bitmap) from *other* protocol versions.
3387 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3388 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3390 Since SSL 2.0 does not have any ciphersuites for which the
3391 128/256 bit distinction would be relevant, this works for now.
3392 The proper fix will be to use different bits for AES128 and
3393 AES256, which would have avoided the problems from the beginning;
3394 however, bits are scarce, so we can only do this in a new release
3395 (not just a patchlevel) when we can change the SSL_CIPHER
3396 definition to split the single 'unsigned long mask' bitmap into
3397 multiple values to extend the available space.
3401 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3403 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3404 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3406 *) Add AES IGE and biIGE modes.
3409 *) Change the Unix randomness entropy gathering to use poll() when
3410 possible instead of select(), since the latter has some
3411 undesirable limitations.
3412 [Darryl Miles via Richard Levitte and Bodo Moeller]
3414 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3415 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3416 cannot be implicitly activated as part of, e.g., the "AES" alias.
3417 However, please upgrade to OpenSSL 0.9.9[-dev] for
3418 non-experimental use of the ECC ciphersuites to get TLS extension
3419 support, which is required for curve and point format negotiation
3420 to avoid potential handshake problems.
3423 *) Disable rogue ciphersuites:
3425 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3426 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3427 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3429 The latter two were purportedly from
3430 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3433 Also deactivate the remaining ciphersuites from
3434 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3435 unofficial, and the ID has long expired.
3438 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3439 dual-core machines) and other potential thread-safety issues.
3442 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3443 versions), which is now available for royalty-free use
3444 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3445 Also, add Camellia TLS ciphersuites from RFC 4132.
3447 To minimize changes between patchlevels in the OpenSSL 0.9.8
3448 series, Camellia remains excluded from compilation unless OpenSSL
3449 is configured with 'enable-camellia'.
3452 *) Disable the padding bug check when compression is in use. The padding
3453 bug check assumes the first packet is of even length, this is not
3454 necessarily true if compresssion is enabled and can result in false
3455 positives causing handshake failure. The actual bug test is ancient
3456 code so it is hoped that implementations will either have fixed it by
3457 now or any which still have the bug do not support compression.
3460 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3462 *) When applying a cipher rule check to see if string match is an explicit
3463 cipher suite and only match that one cipher suite if it is.
3466 *) Link in manifests for VC++ if needed.
3467 [Austin Ziegler <halostatue@gmail.com>]
3469 *) Update support for ECC-based TLS ciphersuites according to
3470 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3471 TLS extensions, which are supported starting with the 0.9.9
3472 branch, not in the OpenSSL 0.9.8 branch).
3475 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3476 opaque EVP_CIPHER_CTX handling.
3479 *) Fixes and enhancements to zlib compression code. We now only use
3480 "zlib1.dll" and use the default __cdecl calling convention on Win32
3481 to conform with the standards mentioned here:
3482 http://www.zlib.net/DLL_FAQ.txt
3483 Static zlib linking now works on Windows and the new --with-zlib-include
3484 --with-zlib-lib options to Configure can be used to supply the location
3485 of the headers and library. Gracefully handle case where zlib library
3489 *) Several fixes and enhancements to the OID generation code. The old code
3490 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3491 handle numbers larger than ULONG_MAX, truncated printing and had a
3492 non standard OBJ_obj2txt() behaviour.
3495 *) Add support for building of engines under engine/ as shared libraries
3496 under VC++ build system.
3499 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3500 Hopefully, we will not see any false combination of paths any more.
3503 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3505 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3506 (part of SSL_OP_ALL). This option used to disable the
3507 countermeasure against man-in-the-middle protocol-version
3508 rollback in the SSL 2.0 server implementation, which is a bad
3509 idea. (CVE-2005-2969)
3511 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3512 for Information Security, National Institute of Advanced Industrial
3513 Science and Technology [AIST], Japan)]
3515 *) Add two function to clear and return the verify parameter flags.
3518 *) Keep cipherlists sorted in the source instead of sorting them at
3519 runtime, thus removing the need for a lock.
3522 *) Avoid some small subgroup attacks in Diffie-Hellman.
3523 [Nick Mathewson and Ben Laurie]
3525 *) Add functions for well-known primes.
3528 *) Extended Windows CE support.
3529 [Satoshi Nakamura and Andy Polyakov]
3531 *) Initialize SSL_METHOD structures at compile time instead of during
3532 runtime, thus removing the need for a lock.
3535 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3536 attempting to decrypt each encrypted key in turn. Add support to
3540 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3542 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3545 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3548 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3549 key into the same file any more.
3552 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3555 *) Add -utf8 command line and config file option to 'ca'.
3556 [Stefan <stf@udoma.org]
3558 *) Removed the macro des_crypt(), as it seems to conflict with some
3559 libraries. Use DES_crypt().
3562 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3563 involves renaming the source and generated shared-libs for
3564 both. The engines will accept the corrected or legacy ids
3565 ('ncipher' and '4758_cca' respectively) when binding. NB,
3566 this only applies when building 'shared'.
3567 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3569 *) Add attribute functions to EVP_PKEY structure. Modify
3570 PKCS12_create() to recognize a CSP name attribute and
3571 use it. Make -CSP option work again in pkcs12 utility.
3574 *) Add new functionality to the bn blinding code:
3575 - automatic re-creation of the BN_BLINDING parameters after
3576 a fixed number of uses (currently 32)
3577 - add new function for parameter creation
3578 - introduce flags to control the update behaviour of the
3579 BN_BLINDING parameters
3580 - hide BN_BLINDING structure
3581 Add a second BN_BLINDING slot to the RSA structure to improve
3582 performance when a single RSA object is shared among several
3586 *) Add support for DTLS.
3587 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3589 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3590 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3593 *) Remove buggy and incomplete DH cert support from
3594 ssl/ssl_rsa.c and ssl/s3_both.c
3597 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3598 the apps/openssl applications.
3601 *) Compile clean with "-Wall -Wmissing-prototypes
3602 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3603 DEBUG_SAFESTACK must also be set.
3606 *) Change ./Configure so that certain algorithms can be disabled by default.
3607 The new counterpiece to "no-xxx" is "enable-xxx".
3609 The patented RC5 and MDC2 algorithms will now be disabled unless
3610 "enable-rc5" and "enable-mdc2", respectively, are specified.
3612 (IDEA remains enabled despite being patented. This is because IDEA
3613 is frequently required for interoperability, and there is no license
3614 fee for non-commercial use. As before, "no-idea" can be used to
3615 avoid this algorithm.)
3619 *) Add processing of proxy certificates (see RFC 3820). This work was
3620 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3621 EGEE (Enabling Grids for E-science in Europe).
3624 *) RC4 performance overhaul on modern architectures/implementations, such
3625 as Intel P4, IA-64 and AMD64.
3628 *) New utility extract-section.pl. This can be used specify an alternative
3629 section number in a pod file instead of having to treat each file as
3630 a separate case in Makefile. This can be done by adding two lines to the
3633 =for comment openssl_section:XXX
3635 The blank line is mandatory.
3639 *) New arguments -certform, -keyform and -pass for s_client and s_server
3640 to allow alternative format key and certificate files and passphrase
3644 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3645 update associated structures and add various utility functions.
3647 Add new policy related verify parameters, include policy checking in
3648 standard verify code. Enhance 'smime' application with extra parameters
3649 to support policy checking and print out.
3652 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3653 Nehemiah processors. These extensions support AES encryption in hardware
3654 as well as RNG (though RNG support is currently disabled).
3655 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3657 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3660 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3661 [Andy Polyakov and a number of other people]
3663 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3664 implementation contributed by IBM.
3665 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3667 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3668 exponent rather than 'unsigned long'. There is a corresponding change to
3669 the new 'rsa_keygen' element of the RSA_METHOD structure.
3670 [Jelte Jansen, Geoff Thorpe]
3672 *) Functionality for creating the initial serial number file is now
3673 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3675 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3676 number file to 1, which is bound to cause problems. To avoid
3677 the problems while respecting compatibility between different 0.9.7
3678 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3679 CA.pl for serial number initialization. With the new release 0.9.8,
3680 we can fix the problem directly in the 'ca' utility.)
3683 *) Reduced header interdepencies by declaring more opaque objects in
3684 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3685 give fewer recursive includes, which could break lazy source code - so
3686 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3687 developers should define this symbol when building and using openssl to
3688 ensure they track the recommended behaviour, interfaces, [etc], but
3689 backwards-compatible behaviour prevails when this isn't defined.
3692 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3695 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3696 This will generate a random key of the appropriate length based on the
3697 cipher context. The EVP_CIPHER can provide its own random key generation
3698 routine to support keys of a specific form. This is used in the des and
3699 3des routines to generate a key of the correct parity. Update S/MIME
3700 code to use new functions and hence generate correct parity DES keys.
3701 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3702 valid (weak or incorrect parity).
3705 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3706 as looking them up. This is useful when the verified structure may contain
3707 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3708 present unless the new PKCS7_NO_CRL flag is asserted.
3711 *) Extend ASN1 oid configuration module. It now additionally accepts the
3714 shortName = some long name, 1.2.3.4
3717 *) Reimplemented the BN_CTX implementation. There is now no more static
3718 limitation on the number of variables it can handle nor the depth of the
3719 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3720 information can now expand as required, and rather than having a single
3721 static array of bignums, BN_CTX now uses a linked-list of such arrays
3722 allowing it to expand on demand whilst maintaining the usefulness of
3723 BN_CTX's "bundling".
3726 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3727 to allow all RSA operations to function using a single BN_CTX.
3730 *) Preliminary support for certificate policy evaluation and checking. This
3731 is initially intended to pass the tests outlined in "Conformance Testing
3732 of Relying Party Client Certificate Path Processing Logic" v1.07.
3735 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3736 remained unused and not that useful. A variety of other little bignum
3737 tweaks and fixes have also been made continuing on from the audit (see
3741 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3742 associated ASN1, EVP and SSL functions and old ASN1 macros.
3745 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3746 and this should never fail. So the return value from the use of
3747 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3748 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3751 *) BN_CTX_get() should return zero-valued bignums, providing the same
3752 initialised value as BN_new().
3753 [Geoff Thorpe, suggested by Ulf Möller]
3755 *) Support for inhibitAnyPolicy certificate extension.
3758 *) An audit of the BIGNUM code is underway, for which debugging code is
3759 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3760 is considered valid when processing BIGNUMs, and causes execution to
3761 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3762 further steps are taken to deliberately pollute unused data in BIGNUM
3763 structures to try and expose faulty code further on. For now, openssl will
3764 (in its default mode of operation) continue to tolerate the inconsistent
3765 forms that it has tolerated in the past, but authors and packagers should
3766 consider trying openssl and their own applications when compiled with
3767 these debugging symbols defined. It will help highlight potential bugs in
3768 their own code, and will improve the test coverage for OpenSSL itself. At
3769 some point, these tighter rules will become openssl's default to improve
3770 maintainability, though the assert()s and other overheads will remain only
3771 in debugging configurations. See bn.h for more details.
3772 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3774 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3775 that can only be obtained through BN_CTX_new() (which implicitly
3776 initialises it). The presence of this function only made it possible
3777 to overwrite an existing structure (and cause memory leaks).
3780 *) Because of the callback-based approach for implementing LHASH as a
3781 template type, lh_insert() adds opaque objects to hash-tables and
3782 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3783 to clean up those corresponding objects before destroying the hash table
3784 (and losing the object pointers). So some over-zealous constifications in
3785 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3786 objects as "const" and the lh_doall[_arg] callback wrappers are not
3787 prototyped to have "const" restrictions on the object pointers they are
3788 given (and so aren't required to cast them away any more).
3791 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3792 (speed) prefers to use its own implementation. The two implementations
3793 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3794 its object type properly exposed (MS_TM) instead of casting to/from "char
3795 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3796 aren't necessarily the greatest nomenclatures - but this is what was used
3797 internally to the implementation so I've used that for now.
3800 *) Ensure that deprecated functions do not get compiled when
3801 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3802 the self-tests were still using deprecated key-generation functions so
3803 these have been updated also.
3806 *) Reorganise PKCS#7 code to separate the digest location functionality
3807 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3808 New function PKCS7_set_digest() to set the digest type for PKCS#7
3809 digestedData type. Add additional code to correctly generate the
3810 digestedData type and add support for this type in PKCS7 initialization
3814 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3815 structure of type "other".
3818 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3819 sure the loop does correctly stop and breaking ("division by zero")
3820 modulus operations are not performed. The (pre-generated) prime
3821 table crypto/bn/bn_prime.h was already correct, but it could not be
3822 re-generated on some platforms because of the "division by zero"
3823 situation in the script.
3824 [Ralf S. Engelschall]
3826 *) Update support for ECC-based TLS ciphersuites according to
3827 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3828 SHA-1 now is only used for "small" curves (where the
3829 representation of a field element takes up to 24 bytes); for
3830 larger curves, the field element resulting from ECDH is directly
3831 used as premaster secret.
3832 [Douglas Stebila (Sun Microsystems Laboratories)]
3834 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3835 curve secp160r1 to the tests.
3836 [Douglas Stebila (Sun Microsystems Laboratories)]
3838 *) Add the possibility to load symbols globally with DSO.
3839 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3841 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3842 control of the error stack.
3845 *) Add support for STORE in ENGINE.
3848 *) Add the STORE type. The intention is to provide a common interface
3849 to certificate and key stores, be they simple file-based stores, or
3850 HSM-type store, or LDAP stores, or...
3851 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3854 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3855 pass a list of arguments to any function as well as provide a way
3856 for a function to pass data back to the caller.
3859 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3860 works like BUF_strdup() but can be used to duplicate a portion of
3861 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3865 *) Add the function sk_find_ex() which works like sk_find(), but will
3866 return an index to an element even if an exact match couldn't be
3867 found. The index is guaranteed to point at the element where the
3868 searched-for key would be inserted to preserve sorting order.
3871 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3872 takes an extra flags argument for optional functionality. Currently,
3873 the following flags are defined:
3875 OBJ_BSEARCH_VALUE_ON_NOMATCH
3876 This one gets OBJ_bsearch_ex() to return a pointer to the first
3877 element where the comparing function returns a negative or zero
3880 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3881 This one gets OBJ_bsearch_ex() to return a pointer to the first
3882 element where the comparing function returns zero. This is useful
3883 if there are more than one element where the comparing function
3887 *) Make it possible to create self-signed certificates with 'openssl ca'
3888 in such a way that the self-signed certificate becomes part of the
3889 CA database and uses the same mechanisms for serial number generation
3890 as all other certificate signing. The new flag '-selfsign' enables
3891 this functionality. Adapt CA.sh and CA.pl.in.
3894 *) Add functionality to check the public key of a certificate request
3895 against a given private. This is useful to check that a certificate
3896 request can be signed by that key (self-signing).
3899 *) Make it possible to have multiple active certificates with the same
3900 subject in the CA index file. This is done only if the keyword
3901 'unique_subject' is set to 'no' in the main CA section (default
3902 if 'CA_default') of the configuration file. The value is saved
3903 with the database itself in a separate index attribute file,
3904 named like the index file with '.attr' appended to the name.
3907 *) Generate muti valued AVAs using '+' notation in config files for
3911 *) Support for nameConstraints certificate extension.
3914 *) Support for policyConstraints certificate extension.
3917 *) Support for policyMappings certificate extension.
3920 *) Make sure the default DSA_METHOD implementation only uses its
3921 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3922 and change its own handlers to be NULL so as to remove unnecessary
3923 indirection. This lets alternative implementations fallback to the
3924 default implementation more easily.
3927 *) Support for directoryName in GeneralName related extensions
3931 *) Make it possible to link applications using Makefile.shared.
3932 Make that possible even when linking against static libraries!
3935 *) Support for single pass processing for S/MIME signing. This now
3936 means that S/MIME signing can be done from a pipe, in addition
3937 cleartext signing (multipart/signed type) is effectively streaming
3938 and the signed data does not need to be all held in memory.
3940 This is done with a new flag PKCS7_STREAM. When this flag is set
3941 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3942 is done after the data is output (and digests calculated) in
3943 SMIME_write_PKCS7().
3946 *) Add full support for -rpath/-R, both in shared libraries and
3947 applications, at least on the platforms where it's known how
3951 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3952 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3953 will now compute a table of multiples of the generator that
3954 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3955 faster (notably in the case of a single point multiplication,
3956 scalar * generator).
3957 [Nils Larsch, Bodo Moeller]
3959 *) IPv6 support for certificate extensions. The various extensions
3960 which use the IP:a.b.c.d can now take IPv6 addresses using the
3961 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3965 *) Added an ENGINE that implements RSA by performing private key
3966 exponentiations with the GMP library. The conversions to and from
3967 GMP's mpz_t format aren't optimised nor are any montgomery forms
3968 cached, and on x86 it appears OpenSSL's own performance has caught up.
3969 However there are likely to be other architectures where GMP could
3970 provide a boost. This ENGINE is not built in by default, but it can be
3971 specified at Configure time and should be accompanied by the necessary
3972 linker additions, eg;
3973 ./config -DOPENSSL_USE_GMP -lgmp
3976 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3977 testing availability of engines with "-t" - the old behaviour is
3978 produced by increasing the feature's verbosity with "-tt".
3981 *) ECDSA routines: under certain error conditions uninitialized BN objects
3982 could be freed. Solution: make sure initialization is performed early
3983 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3987 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3988 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3989 software implementations. For DSA and DH, parameter generation can
3990 also be overriden by providing the appropriate method callbacks.
3993 *) Change the "progress" mechanism used in key-generation and
3994 primality testing to functions that take a new BN_GENCB pointer in
3995 place of callback/argument pairs. The new API functions have "_ex"
3996 postfixes and the older functions are reimplemented as wrappers for
3997 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3998 declarations of the old functions to help (graceful) attempts to
3999 migrate to the new functions. Also, the new key-generation API
4000 functions operate on a caller-supplied key-structure and return
4001 success/failure rather than returning a key or NULL - this is to
4002 help make "keygen" another member function of RSA_METHOD etc.
4004 Example for using the new callback interface:
4006 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4010 BN_GENCB_set(&my_cb, my_callback, my_arg);
4012 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4013 /* For the meaning of a, b in calls to my_callback(), see the
4014 * documentation of the function that calls the callback.
4015 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4016 * my_callback should return 1 if it wants BN_is_prime_ex()
4017 * to continue, or 0 to stop.
4022 *) Change the ZLIB compression method to be stateful, and make it
4023 available to TLS with the number defined in
4024 draft-ietf-tls-compression-04.txt.
4027 *) Add the ASN.1 structures and functions for CertificatePair, which
4028 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4030 CertificatePair ::= SEQUENCE {
4031 forward [0] Certificate OPTIONAL,
4032 reverse [1] Certificate OPTIONAL,
4033 -- at least one of the pair shall be present -- }
4035 Also implement the PEM functions to read and write certificate
4036 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4038 This needed to be defined, mostly for the sake of the LDAP
4039 attribute crossCertificatePair, but may prove useful elsewhere as
4043 *) Make it possible to inhibit symlinking of shared libraries in
4044 Makefile.shared, for Cygwin's sake.
4047 *) Extend the BIGNUM API by creating a function
4048 void BN_set_negative(BIGNUM *a, int neg);
4049 and a macro that behave like
4050 int BN_is_negative(const BIGNUM *a);
4052 to avoid the need to access 'a->neg' directly in applications.
4055 *) Implement fast modular reduction for pseudo-Mersenne primes
4056 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4057 EC_GROUP_new_curve_GFp() will now automatically use this
4059 [Nils Larsch <nla@trustcenter.de>]
4061 *) Add new lock type (CRYPTO_LOCK_BN).
4064 *) Change the ENGINE framework to automatically load engines
4065 dynamically from specific directories unless they could be
4066 found to already be built in or loaded. Move all the
4067 current engines except for the cryptodev one to a new
4069 The engines in engines/ are built as shared libraries if
4070 the "shared" options was given to ./Configure or ./config.
4071 Otherwise, they are inserted in libcrypto.a.
4072 /usr/local/ssl/engines is the default directory for dynamic
4073 engines, but that can be overriden at configure time through
4074 the usual use of --prefix and/or --openssldir, and at run
4075 time with the environment variable OPENSSL_ENGINES.
4076 [Geoff Thorpe and Richard Levitte]
4078 *) Add Makefile.shared, a helper makefile to build shared
4079 libraries. Adapt Makefile.org.
4082 *) Add version info to Win32 DLLs.
4083 [Peter 'Luna' Runestig" <peter@runestig.com>]
4085 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4086 can be added using this API to created arbitrary PKCS#12
4087 files while avoiding the low level API.
4089 New options to PKCS12_create(), key or cert can be NULL and
4090 will then be omitted from the output file. The encryption
4091 algorithm NIDs can be set to -1 for no encryption, the mac
4092 iteration count can be set to 0 to omit the mac.
4094 Enhance pkcs12 utility by making the -nokeys and -nocerts
4095 options work when creating a PKCS#12 file. New option -nomac
4096 to omit the mac, NONE can be set for an encryption algorithm.
4097 New code is modified to use the enhanced PKCS12_create()
4098 instead of the low level API.
4101 *) Extend ASN1 encoder to support indefinite length constructed
4102 encoding. This can output sequences tags and octet strings in
4103 this form. Modify pk7_asn1.c to support indefinite length
4104 encoding. This is experimental and needs additional code to
4105 be useful, such as an ASN1 bio and some enhanced streaming
4108 Extend template encode functionality so that tagging is passed
4109 down to the template encoder.
4112 *) Let 'openssl req' fail if an argument to '-newkey' is not
4113 recognized instead of using RSA as a default.
4116 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4117 As these are not official, they are not included in "ALL";
4118 the "ECCdraft" ciphersuite group alias can be used to select them.
4119 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4121 *) Add ECDH engine support.
4122 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4124 *) Add ECDH in new directory crypto/ecdh/.
4125 [Douglas Stebila (Sun Microsystems Laboratories)]
4127 *) Let BN_rand_range() abort with an error after 100 iterations
4128 without success (which indicates a broken PRNG).
4131 *) Change BN_mod_sqrt() so that it verifies that the input value
4132 is really the square of the return value. (Previously,
4133 BN_mod_sqrt would show GIGO behaviour.)
4136 *) Add named elliptic curves over binary fields from X9.62, SECG,
4137 and WAP/WTLS; add OIDs that were still missing.
4139 [Sheueling Chang Shantz and Douglas Stebila
4140 (Sun Microsystems Laboratories)]
4142 *) Extend the EC library for elliptic curves over binary fields
4143 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4146 EC_GF2m_simple_method
4150 EC_GROUP_new_curve_GF2m
4151 EC_GROUP_set_curve_GF2m
4152 EC_GROUP_get_curve_GF2m
4153 EC_POINT_set_affine_coordinates_GF2m
4154 EC_POINT_get_affine_coordinates_GF2m
4155 EC_POINT_set_compressed_coordinates_GF2m
4157 Point compression for binary fields is disabled by default for
4158 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4161 As binary polynomials are represented as BIGNUMs, various members
4162 of the EC_GROUP and EC_POINT data structures can be shared
4163 between the implementations for prime fields and binary fields;
4164 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4165 are essentially identical to their ..._GFp counterparts.
4166 (For simplicity, the '..._GFp' prefix has been dropped from
4167 various internal method names.)
4169 An internal 'field_div' method (similar to 'field_mul' and
4170 'field_sqr') has been added; this is used only for binary fields.
4172 [Sheueling Chang Shantz and Douglas Stebila
4173 (Sun Microsystems Laboratories)]
4175 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4176 through methods ('mul', 'precompute_mult').
4178 The generic implementations (now internally called 'ec_wNAF_mul'
4179 and 'ec_wNAF_precomputed_mult') remain the default if these
4180 methods are undefined.
4182 [Sheueling Chang Shantz and Douglas Stebila
4183 (Sun Microsystems Laboratories)]
4185 *) New function EC_GROUP_get_degree, which is defined through
4186 EC_METHOD. For curves over prime fields, this returns the bit
4187 length of the modulus.
4189 [Sheueling Chang Shantz and Douglas Stebila
4190 (Sun Microsystems Laboratories)]
4192 *) New functions EC_GROUP_dup, EC_POINT_dup.
4193 (These simply call ..._new and ..._copy).
4195 [Sheueling Chang Shantz and Douglas Stebila
4196 (Sun Microsystems Laboratories)]
4198 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4199 Polynomials are represented as BIGNUMs (where the sign bit is not
4200 used) in the following functions [macros]:
4203 BN_GF2m_sub [= BN_GF2m_add]
4204 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4205 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4206 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4208 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4209 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4210 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4211 BN_GF2m_cmp [= BN_ucmp]
4213 (Note that only the 'mod' functions are actually for fields GF(2^m).
4214 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4216 For some functions, an the irreducible polynomial defining a
4217 field can be given as an 'unsigned int[]' with strictly
4218 decreasing elements giving the indices of those bits that are set;
4219 i.e., p[] represents the polynomial
4220 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4222 p[0] > p[1] > ... > p[k] = 0.
4223 This applies to the following functions:
4228 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4229 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4231 BN_GF2m_mod_sqrt_arr
4232 BN_GF2m_mod_solve_quad_arr
4236 Conversion can be performed by the following functions:
4241 bntest.c has additional tests for binary polynomial arithmetic.
4243 Two implementations for BN_GF2m_mod_div() are available.
4244 The default algorithm simply uses BN_GF2m_mod_inv() and
4245 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4246 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4247 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4249 [Sheueling Chang Shantz and Douglas Stebila
4250 (Sun Microsystems Laboratories)]
4252 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4253 functionality is disabled at compile-time.
4254 [Douglas Stebila <douglas.stebila@sun.com>]
4256 *) Change default behaviour of 'openssl asn1parse' so that more
4257 information is visible when viewing, e.g., a certificate:
4259 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4260 mode the content of non-printable OCTET STRINGs is output in a
4261 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4262 avoid the appearance of a printable string.
4263 [Nils Larsch <nla@trustcenter.de>]
4265 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4267 EC_GROUP_set_asn1_flag()
4268 EC_GROUP_get_asn1_flag()
4269 EC_GROUP_set_point_conversion_form()
4270 EC_GROUP_get_point_conversion_form()
4271 These control ASN1 encoding details:
4272 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4273 has been set to OPENSSL_EC_NAMED_CURVE.
4274 - Points are encoded in uncompressed form by default; options for
4275 asn1_for are as for point2oct, namely
4276 POINT_CONVERSION_COMPRESSED
4277 POINT_CONVERSION_UNCOMPRESSED
4278 POINT_CONVERSION_HYBRID
4280 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4283 EC_GROUP_get0_seed()
4284 EC_GROUP_get_seed_len()
4285 This is used only for ASN1 purposes (so far).
4286 [Nils Larsch <nla@trustcenter.de>]
4288 *) Add 'field_type' member to EC_METHOD, which holds the NID
4289 of the appropriate field type OID. The new function
4290 EC_METHOD_get_field_type() returns this value.
4291 [Nils Larsch <nla@trustcenter.de>]
4296 EC_POINT_point2hex()
4297 EC_POINT_hex2point()
4298 providing useful interfaces to EC_POINT_point2oct() and
4299 EC_POINT_oct2point().
4300 [Nils Larsch <nla@trustcenter.de>]
4302 *) Change internals of the EC library so that the functions
4303 EC_GROUP_set_generator()
4304 EC_GROUP_get_generator()
4305 EC_GROUP_get_order()
4306 EC_GROUP_get_cofactor()
4307 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4308 to methods, which would lead to unnecessary code duplication when
4309 adding different types of curves.
4310 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4312 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4313 arithmetic, and such that modified wNAFs are generated
4314 (which avoid length expansion in many cases).
4317 *) Add a function EC_GROUP_check_discriminant() (defined via
4318 EC_METHOD) that verifies that the curve discriminant is non-zero.
4320 Add a function EC_GROUP_check() that makes some sanity tests
4321 on a EC_GROUP, its generator and order. This includes
4322 EC_GROUP_check_discriminant().
4323 [Nils Larsch <nla@trustcenter.de>]
4325 *) Add ECDSA in new directory crypto/ecdsa/.
4327 Add applications 'openssl ecparam' and 'openssl ecdsa'
4328 (these are based on 'openssl dsaparam' and 'openssl dsa').
4330 ECDSA support is also included in various other files across the
4331 library. Most notably,
4332 - 'openssl req' now has a '-newkey ecdsa:file' option;
4333 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4334 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4335 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4336 them suitable for ECDSA where domain parameters must be
4337 extracted before the specific public key;
4338 - ECDSA engine support has been added.
4339 [Nils Larsch <nla@trustcenter.de>]
4341 *) Include some named elliptic curves, and add OIDs from X9.62,
4342 SECG, and WAP/WTLS. Each curve can be obtained from the new
4344 EC_GROUP_new_by_curve_name(),
4345 and the list of available named curves can be obtained with
4346 EC_get_builtin_curves().
4347 Also add a 'curve_name' member to EC_GROUP objects, which can be
4349 EC_GROUP_set_curve_name()
4350 EC_GROUP_get_curve_name()
4351 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4353 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4354 was actually never needed) and in BN_mul(). The removal in BN_mul()
4355 required a small change in bn_mul_part_recursive() and the addition
4356 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4357 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4358 bn_sub_words() and bn_add_words() except they take arrays with
4362 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4364 *) Cleanse PEM buffers before freeing them since they may contain
4366 [Benjamin Bennett <ben@psc.edu>]
4368 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4369 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4370 authentication-only ciphersuites.
4373 *) Since AES128 and AES256 share a single mask bit in the logic of
4374 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4375 kludge to work properly if AES128 is available and AES256 isn't.
4378 *) Expand security boundary to match 1.1.1 module.
4381 *) Remove redundant features: hash file source, editing of test vectors
4382 modify fipsld to use external fips_premain.c signature.
4385 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4386 run algorithm test programs.
4389 *) Make algorithm test programs more tolerant of whitespace.
4392 *) Have SSL/TLS server implementation tolerate "mismatched" record
4393 protocol version while receiving ClientHello even if the
4394 ClientHello is fragmented. (The server can't insist on the
4395 particular protocol version it has chosen before the ServerHello
4396 message has informed the client about his choice.)
4399 *) Load error codes if they are not already present instead of using a
4400 static variable. This allows them to be cleanly unloaded and reloaded.
4403 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4405 *) Introduce limits to prevent malicious keys being able to
4406 cause a denial of service. (CVE-2006-2940)
4407 [Steve Henson, Bodo Moeller]
4409 *) Fix ASN.1 parsing of certain invalid structures that can result
4410 in a denial of service. (CVE-2006-2937) [Steve Henson]
4412 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4413 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4415 *) Fix SSL client code which could crash if connecting to a
4416 malicious SSLv2 server. (CVE-2006-4343)
4417 [Tavis Ormandy and Will Drewry, Google Security Team]
4419 *) Change ciphersuite string processing so that an explicit
4420 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4421 will no longer include "AES128-SHA"), and any other similar
4422 ciphersuite (same bitmap) from *other* protocol versions (so that
4423 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4424 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4425 changes from 0.9.8b and 0.9.8d.
4428 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4430 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4431 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4433 *) Change the Unix randomness entropy gathering to use poll() when
4434 possible instead of select(), since the latter has some
4435 undesirable limitations.
4436 [Darryl Miles via Richard Levitte and Bodo Moeller]
4438 *) Disable rogue ciphersuites:
4440 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4441 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4442 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4444 The latter two were purportedly from
4445 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4448 Also deactive the remaining ciphersuites from
4449 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4450 unofficial, and the ID has long expired.
4453 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4454 dual-core machines) and other potential thread-safety issues.
4457 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4459 *) Adapt fipsld and the build system to link against the validated FIPS
4460 module in FIPS mode.
4463 *) Fixes for VC++ 2005 build under Windows.
4466 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4467 from a Windows bash shell such as MSYS. It is autodetected from the
4468 "config" script when run from a VC++ environment. Modify standard VC++
4469 build to use fipscanister.o from the GNU make build.
4472 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4474 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4475 The value now differs depending on if you build for FIPS or not.
4476 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4477 safely run with a non-FIPSed libcrypto, as it may crash because of
4478 the difference induced by this change.
4481 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4483 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4484 (part of SSL_OP_ALL). This option used to disable the
4485 countermeasure against man-in-the-middle protocol-version
4486 rollback in the SSL 2.0 server implementation, which is a bad
4487 idea. (CVE-2005-2969)
4489 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4490 for Information Security, National Institute of Advanced Industrial
4491 Science and Technology [AIST], Japan)]
4493 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4494 mainly for FIPS compliance and not fully integrated at this stage.
4497 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4498 the exponentiation using a fixed-length exponent. (Otherwise,
4499 the information leaked through timing could expose the secret key
4500 after many signatures; cf. Bleichenbacher's attack on DSA with
4504 *) Make a new fixed-window mod_exp implementation the default for
4505 RSA, DSA, and DH private-key operations so that the sequence of
4506 squares and multiplies and the memory access pattern are
4507 independent of the particular secret key. This will mitigate
4508 cache-timing and potential related attacks.
4510 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4511 and this is automatically used by BN_mod_exp_mont() if the new flag
4512 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4513 will use this BN flag for private exponents unless the flag
4514 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4515 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4517 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4519 *) Change the client implementation for SSLv23_method() and
4520 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4521 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4522 (Previously, the SSL 2.0 backwards compatible Client Hello
4523 message format would be used even with SSL_OP_NO_SSLv2.)
4526 *) Add support for smime-type MIME parameter in S/MIME messages which some
4530 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4531 a threadsafe manner. Modify rsa code to use new function and add calls
4532 to dsa and dh code (which had race conditions before).
4535 *) Include the fixed error library code in the C error file definitions
4536 instead of fixing them up at runtime. This keeps the error code
4537 structures constant.
4540 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4542 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4545 *) Fixes for newer kerberos headers. NB: the casts are needed because
4546 the 'length' field is signed on one version and unsigned on another
4547 with no (?) obvious way to tell the difference, without these VC++
4548 complains. Also the "definition" of FAR (blank) is no longer included
4549 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4550 some needed definitions.
4553 *) Undo Cygwin change.
4556 *) Added support for proxy certificates according to RFC 3820.
4557 Because they may be a security thread to unaware applications,
4558 they must be explicitely allowed in run-time. See
4559 docs/HOWTO/proxy_certificates.txt for further information.
4562 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4564 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4565 server and client random values. Previously
4566 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4567 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4569 This change has negligible security impact because:
4571 1. Server and client random values still have 24 bytes of pseudo random
4574 2. Server and client random values are sent in the clear in the initial
4577 3. The master secret is derived using the premaster secret (48 bytes in
4578 size for static RSA ciphersuites) as well as client server and random
4581 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4584 [Stephen Henson, reported by UK NISCC]
4586 *) Use Windows randomness collection on Cygwin.
4589 *) Fix hang in EGD/PRNGD query when communication socket is closed
4590 prematurely by EGD/PRNGD.
4591 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4593 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4596 *) Back-port of selected performance improvements from development
4597 branch, as well as improved support for PowerPC platforms.
4600 *) Add lots of checks for memory allocation failure, error codes to indicate
4601 failure and freeing up memory if a failure occurs.
4602 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4604 *) Add new -passin argument to dgst.
4607 *) Perform some character comparisons of different types in X509_NAME_cmp:
4608 this is needed for some certificates that reencode DNs into UTF8Strings
4609 (in violation of RFC3280) and can't or wont issue name rollover
4613 *) Make an explicit check during certificate validation to see that
4614 the CA setting in each certificate on the chain is correct. As a
4615 side effect always do the following basic checks on extensions,
4616 not just when there's an associated purpose to the check:
4618 - if there is an unhandled critical extension (unless the user
4619 has chosen to ignore this fault)
4620 - if the path length has been exceeded (if one is set at all)
4621 - that certain extensions fit the associated purpose (if one has
4625 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4627 *) Avoid a race condition when CRLs are checked in a multi threaded
4628 environment. This would happen due to the reordering of the revoked
4629 entries during signature checking and serial number lookup. Now the
4630 encoding is cached and the serial number sort performed under a lock.
4631 Add new STACK function sk_is_sorted().
4634 *) Add Delta CRL to the extension code.
4637 *) Various fixes to s3_pkt.c so alerts are sent properly.
4638 [David Holmes <d.holmes@f5.com>]
4640 *) Reduce the chances of duplicate issuer name and serial numbers (in
4641 violation of RFC3280) using the OpenSSL certificate creation utilities.
4642 This is done by creating a random 64 bit value for the initial serial
4643 number when a serial number file is created or when a self signed
4644 certificate is created using 'openssl req -x509'. The initial serial
4645 number file is created using 'openssl x509 -next_serial' in CA.pl
4646 rather than being initialized to 1.
4649 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4651 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4652 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4653 [Joe Orton, Steve Henson]
4655 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4657 [Joe Orton, Steve Henson]
4659 *) Make it possible to have multiple active certificates with the same
4660 subject in the CA index file. This is done only if the keyword
4661 'unique_subject' is set to 'no' in the main CA section (default
4662 if 'CA_default') of the configuration file. The value is saved
4663 with the database itself in a separate index attribute file,
4664 named like the index file with '.attr' appended to the name.
4667 *) X509 verify fixes. Disable broken certificate workarounds when
4668 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4669 keyUsage extension present. Don't accept CRLs with unhandled critical
4670 extensions: since verify currently doesn't process CRL extensions this
4671 rejects a CRL with *any* critical extensions. Add new verify error codes
4675 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4676 A clarification of RFC2560 will require the use of OCTET STRINGs and
4677 some implementations cannot handle the current raw format. Since OpenSSL
4678 copies and compares OCSP nonces as opaque blobs without any attempt at
4679 parsing them this should not create any compatibility issues.
4682 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4683 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4684 this HMAC (and other) operations are several times slower than OpenSSL
4688 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4689 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4691 *) Use the correct content when signing type "other".
4694 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4696 *) Fix various bugs revealed by running the NISCC test suite:
4698 Stop out of bounds reads in the ASN1 code when presented with
4699 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4701 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4703 If verify callback ignores invalid public key errors don't try to check
4704 certificate signature with the NULL public key.
4708 *) New -ignore_err option in ocsp application to stop the server
4709 exiting on the first error in a request.
4712 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4713 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4717 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4718 extra data after the compression methods not only for TLS 1.0
4719 but also for SSL 3.0 (as required by the specification).
4720 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4722 *) Change X509_certificate_type() to mark the key as exported/exportable
4723 when it's 512 *bits* long, not 512 bytes.
4726 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4727 blocks during encryption.
4730 *) Various fixes to base64 BIO and non blocking I/O. On write
4731 flushes were not handled properly if the BIO retried. On read
4732 data was not being buffered properly and had various logic bugs.
4733 This also affects blocking I/O when the data being decoded is a
4737 *) Various S/MIME bugfixes and compatibility changes:
4738 output correct application/pkcs7 MIME type if
4739 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4740 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4741 of files as .eml work). Correctly handle very long lines in MIME
4745 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4747 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4748 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4749 a protocol version number mismatch like a decryption error
4750 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4753 *) Turn on RSA blinding by default in the default implementation
4754 to avoid a timing attack. Applications that don't want it can call
4755 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4756 They would be ill-advised to do so in most cases.
4757 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4759 *) Change RSA blinding code so that it works when the PRNG is not
4760 seeded (in this case, the secret RSA exponent is abused as
4761 an unpredictable seed -- if it is not unpredictable, there
4762 is no point in blinding anyway). Make RSA blinding thread-safe
4763 by remembering the creator's thread ID in rsa->blinding and
4764 having all other threads use local one-time blinding factors
4765 (this requires more computation than sharing rsa->blinding, but
4766 avoids excessive locking; and if an RSA object is not shared
4767 between threads, blinding will still be very fast).
4770 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4771 ENGINE as defaults for all supported algorithms irrespective of
4772 the 'flags' parameter. 'flags' is now honoured, so applications
4773 should make sure they are passing it correctly.
4776 *) Target "mingw" now allows native Windows code to be generated in
4777 the Cygwin environment as well as with the MinGW compiler.
4780 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4782 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4783 via timing by performing a MAC computation even if incorrect
4784 block cipher padding has been found. This is a countermeasure
4785 against active attacks where the attacker has to distinguish
4786 between bad padding and a MAC verification error. (CVE-2003-0078)
4788 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4789 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4790 Martin Vuagnoux (EPFL, Ilion)]
4792 *) Make the no-err option work as intended. The intention with no-err
4793 is not to have the whole error stack handling routines removed from
4794 libcrypto, it's only intended to remove all the function name and
4795 reason texts, thereby removing some of the footprint that may not
4796 be interesting if those errors aren't displayed anyway.
4798 NOTE: it's still possible for any application or module to have it's
4799 own set of error texts inserted. The routines are there, just not
4800 used by default when no-err is given.
4803 *) Add support for FreeBSD on IA64.
4804 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4806 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4807 Kerberos function mit_des_cbc_cksum(). Before this change,
4808 the value returned by DES_cbc_cksum() was like the one from
4809 mit_des_cbc_cksum(), except the bytes were swapped.
4810 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4812 *) Allow an application to disable the automatic SSL chain building.
4813 Before this a rather primitive chain build was always performed in
4814 ssl3_output_cert_chain(): an application had no way to send the
4815 correct chain if the automatic operation produced an incorrect result.
4817 Now the chain builder is disabled if either:
4819 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4821 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4823 The reasoning behind this is that an application would not want the
4824 auto chain building to take place if extra chain certificates are
4825 present and it might also want a means of sending no additional
4826 certificates (for example the chain has two certificates and the
4830 *) Add the possibility to build without the ENGINE framework.
4831 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4833 *) Under Win32 gmtime() can return NULL: check return value in
4834 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4837 *) DSA routines: under certain error conditions uninitialized BN objects
4838 could be freed. Solution: make sure initialization is performed early
4839 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4840 Nils Larsch <nla@trustcenter.de> via PR#459)
4843 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4844 checked on reconnect on the client side, therefore session resumption
4845 could still fail with a "ssl session id is different" error. This
4846 behaviour is masked when SSL_OP_ALL is used due to
4847 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4848 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4849 followup to PR #377.
4852 *) IA-32 assembler support enhancements: unified ELF targets, support
4853 for SCO/Caldera platforms, fix for Cygwin shared build.
4856 *) Add support for FreeBSD on sparc64. As a consequence, support for
4857 FreeBSD on non-x86 processors is separate from x86 processors on
4858 the config script, much like the NetBSD support.
4859 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4861 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4863 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4866 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4867 code (06) was taken as the first octet of the session ID and the last
4868 octet was ignored consequently. As a result SSLv2 client side session
4869 caching could not have worked due to the session ID mismatch between
4871 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4875 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4876 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4880 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4881 seems that in spite of existing for more than a year, many application
4882 author have done nothing to provide the necessary callbacks, which
4883 means that this particular engine will not work properly anywhere.
4884 This is a very unfortunate situation which forces us, in the name
4885 of usability, to give the hw_ncipher.c a static lock, which is part
4887 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4888 appear in 0.9.8 or later. We EXPECT application authors to have
4889 dealt properly with this when 0.9.8 is released (unless we actually
4890 make such changes in the libcrypto locking code that changes will
4891 have to be made anyway).
4894 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4895 octets have been read, EOF or an error occurs. Without this change
4896 some truncated ASN1 structures will not produce an error.
4899 *) Disable Heimdal support, since it hasn't been fully implemented.
4900 Still give the possibility to force the use of Heimdal, but with
4901 warnings and a request that patches get sent to openssl-dev.
4904 *) Add the VC-CE target, introduce the WINCE sysname, and add
4905 INSTALL.WCE and appropriate conditionals to make it build.
4906 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4908 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4909 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4910 edit numbers of the version.
4911 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4913 *) Introduce safe string copy and catenation functions
4914 (BUF_strlcpy() and BUF_strlcat()).
4915 [Ben Laurie (CHATS) and Richard Levitte]
4917 *) Avoid using fixed-size buffers for one-line DNs.
4918 [Ben Laurie (CHATS)]
4920 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4921 resizing buffers containing secrets, and use where appropriate.
4922 [Ben Laurie (CHATS)]
4924 *) Avoid using fixed size buffers for configuration file location.
4925 [Ben Laurie (CHATS)]
4927 *) Avoid filename truncation for various CA files.
4928 [Ben Laurie (CHATS)]
4930 *) Use sizeof in preference to magic numbers.
4931 [Ben Laurie (CHATS)]
4933 *) Avoid filename truncation in cert requests.
4934 [Ben Laurie (CHATS)]
4936 *) Add assertions to check for (supposedly impossible) buffer
4938 [Ben Laurie (CHATS)]
4940 *) Don't cache truncated DNS entries in the local cache (this could
4941 potentially lead to a spoofing attack).
4942 [Ben Laurie (CHATS)]
4944 *) Fix various buffers to be large enough for hex/decimal
4945 representations in a platform independent manner.
4946 [Ben Laurie (CHATS)]
4948 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4949 resizing buffers containing secrets, and use where appropriate.
4950 [Ben Laurie (CHATS)]
4952 *) Add BIO_indent() to avoid much slightly worrying code to do
4954 [Ben Laurie (CHATS)]
4956 *) Convert sprintf()/BIO_puts() to BIO_printf().
4957 [Ben Laurie (CHATS)]
4959 *) buffer_gets() could terminate with the buffer only half
4961 [Ben Laurie (CHATS)]
4963 *) Add assertions to prevent user-supplied crypto functions from
4964 overflowing internal buffers by having large block sizes, etc.
4965 [Ben Laurie (CHATS)]
4967 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4969 [Ben Laurie (CHATS)]
4971 *) Eliminate unused copy of key in RC4.
4972 [Ben Laurie (CHATS)]
4974 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4975 [Ben Laurie (CHATS)]
4977 *) Fix off-by-one error in EGD path.
4978 [Ben Laurie (CHATS)]
4980 *) If RANDFILE path is too long, ignore instead of truncating.
4981 [Ben Laurie (CHATS)]
4983 *) Eliminate unused and incorrectly sized X.509 structure
4985 [Ben Laurie (CHATS)]
4987 *) Eliminate unused and dangerous function knumber().
4988 [Ben Laurie (CHATS)]
4990 *) Eliminate unused and dangerous structure, KSSL_ERR.
4991 [Ben Laurie (CHATS)]
4993 *) Protect against overlong session ID context length in an encoded
4994 session object. Since these are local, this does not appear to be
4996 [Ben Laurie (CHATS)]
4998 *) Change from security patch (see 0.9.6e below) that did not affect
4999 the 0.9.6 release series:
5001 Remote buffer overflow in SSL3 protocol - an attacker could
5002 supply an oversized master key in Kerberos-enabled versions.
5004 [Ben Laurie (CHATS)]
5006 *) Change the SSL kerb5 codes to match RFC 2712.
5009 *) Make -nameopt work fully for req and add -reqopt switch.
5010 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5012 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5013 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5015 *) Make sure tests can be performed even if the corresponding algorithms
5016 have been removed entirely. This was also the last step to make
5017 OpenSSL compilable with DJGPP under all reasonable conditions.
5018 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5020 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5021 to allow version independent disabling of normally unselected ciphers,
5022 which may be activated as a side-effect of selecting a single cipher.
5024 (E.g., cipher list string "RSA" enables ciphersuites that are left
5025 out of "ALL" because they do not provide symmetric encryption.
5026 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5027 [Lutz Jaenicke, Bodo Moeller]
5029 *) Add appropriate support for separate platform-dependent build
5030 directories. The recommended way to make a platform-dependent
5031 build directory is the following (tested on Linux), maybe with
5034 # Place yourself outside of the OpenSSL source tree. In
5035 # this example, the environment variable OPENSSL_SOURCE
5036 # is assumed to contain the absolute OpenSSL source directory.
5037 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5038 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5039 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5040 mkdir -p `dirname $F`
5041 ln -s $OPENSSL_SOURCE/$F $F
5044 To be absolutely sure not to disturb the source tree, a "make clean"
5045 is a good thing. If it isn't successfull, don't worry about it,
5046 it probably means the source directory is very clean.
5049 *) Make sure any ENGINE control commands make local copies of string
5050 pointers passed to them whenever necessary. Otherwise it is possible
5051 the caller may have overwritten (or deallocated) the original string
5052 data when a later ENGINE operation tries to use the stored values.
5053 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5055 *) Improve diagnostics in file reading and command-line digests.
5056 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5058 *) Add AES modes CFB and OFB to the object database. Correct an
5059 error in AES-CFB decryption.
5062 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5063 allows existing EVP_CIPHER_CTX structures to be reused after
5064 calling EVP_*Final(). This behaviour is used by encryption
5065 BIOs and some applications. This has the side effect that
5066 applications must explicitly clean up cipher contexts with
5067 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5070 *) Check the values of dna and dnb in bn_mul_recursive before calling
5071 bn_mul_comba (a non zero value means the a or b arrays do not contain
5072 n2 elements) and fallback to bn_mul_normal if either is not zero.
5075 *) Fix escaping of non-ASCII characters when using the -subj option
5076 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5079 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5080 form for "surname", serialNumber has no short form.
5081 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5082 therefore remove "mail" short name for "internet 7".
5083 The OID for unique identifiers in X509 certificates is
5084 x500UniqueIdentifier, not uniqueIdentifier.
5085 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5088 *) Add an "init" command to the ENGINE config module and auto initialize
5089 ENGINEs. Without any "init" command the ENGINE will be initialized
5090 after all ctrl commands have been executed on it. If init=1 the
5091 ENGINE is initailized at that point (ctrls before that point are run
5092 on the uninitialized ENGINE and after on the initialized one). If
5093 init=0 then the ENGINE will not be iniatialized at all.
5096 *) Fix the 'app_verify_callback' interface so that the user-defined
5097 argument is actually passed to the callback: In the
5098 SSL_CTX_set_cert_verify_callback() prototype, the callback
5099 declaration has been changed from
5102 int (*cb)(X509_STORE_CTX *,void *);
5103 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5104 i=s->ctx->app_verify_callback(&ctx)
5105 has been changed into
5106 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5108 To update applications using SSL_CTX_set_cert_verify_callback(),
5109 a dummy argument can be added to their callback functions.
5110 [D. K. Smetters <smetters@parc.xerox.com>]
5112 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5113 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5115 *) Add and OPENSSL_LOAD_CONF define which will cause
5116 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5117 This allows older applications to transparently support certain
5118 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5119 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5120 load the config file and OPENSSL_add_all_algorithms_conf() which will
5121 always load it have also been added.
5124 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5125 Adjust NIDs and EVP layer.
5126 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5128 *) Config modules support in openssl utility.
5130 Most commands now load modules from the config file,
5131 though in a few (such as version) this isn't done
5132 because it couldn't be used for anything.
5134 In the case of ca and req the config file used is
5135 the same as the utility itself: that is the -config
5136 command line option can be used to specify an
5140 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5141 use "openssl_conf" if filename is NULL use default openssl config file.
5144 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5145 config section name. Add a new flag to tolerate a missing config file
5146 and move code to CONF_modules_load_file().
5149 *) Support for crypto accelerator cards from Accelerated Encryption
5150 Processing, www.aep.ie. (Use engine 'aep')
5151 The support was copied from 0.9.6c [engine] and adapted/corrected
5152 to work with the new engine framework.
5153 [AEP Inc. and Richard Levitte]
5155 *) Support for SureWare crypto accelerator cards from Baltimore
5156 Technologies. (Use engine 'sureware')
5157 The support was copied from 0.9.6c [engine] and adapted
5158 to work with the new engine framework.
5161 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5162 make the newer ENGINE framework commands for the CHIL engine work.
5163 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5165 *) Make it possible to produce shared libraries on ReliantUNIX.
5166 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5168 *) Add the configuration target debug-linux-ppro.
5169 Make 'openssl rsa' use the general key loading routines
5170 implemented in apps.c, and make those routines able to
5171 handle the key format FORMAT_NETSCAPE and the variant
5173 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5175 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5176 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5178 *) Add -keyform to rsautl, and document -engine.
5179 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5181 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5182 BIO_R_NO_SUCH_FILE error code rather than the generic
5183 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5186 *) Add new functions
5188 ERR_peek_last_error_line
5189 ERR_peek_last_error_line_data.
5190 These are similar to
5193 ERR_peek_error_line_data,
5194 but report on the latest error recorded rather than the first one
5195 still in the error queue.
5196 [Ben Laurie, Bodo Moeller]
5198 *) default_algorithms option in ENGINE config module. This allows things
5200 default_algorithms = ALL
5201 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5204 *) Preliminary ENGINE config module.
5207 *) New experimental application configuration code.
5210 *) Change the AES code to follow the same name structure as all other
5211 symmetric ciphers, and behave the same way. Move everything to
5212 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5213 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5215 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5216 [Ben Laurie and Theo de Raadt]
5218 *) Add option to output public keys in req command.
5219 [Massimiliano Pala madwolf@openca.org]
5221 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5222 (up to about 10% better than before for P-192 and P-224).
5225 *) New functions/macros
5227 SSL_CTX_set_msg_callback(ctx, cb)
5228 SSL_CTX_set_msg_callback_arg(ctx, arg)
5229 SSL_set_msg_callback(ssl, cb)
5230 SSL_set_msg_callback_arg(ssl, arg)
5232 to request calling a callback function
5234 void cb(int write_p, int version, int content_type,
5235 const void *buf, size_t len, SSL *ssl, void *arg)
5237 whenever a protocol message has been completely received
5238 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5239 protocol version according to which the SSL library interprets
5240 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5241 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5242 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5243 specification (change_cipher_spec(20), alert(21), handshake(22)).
5244 'buf' and 'len' point to the actual message, 'ssl' to the
5245 SSL object, and 'arg' is the application-defined value set by
5246 SSL[_CTX]_set_msg_callback_arg().
5248 'openssl s_client' and 'openssl s_server' have new '-msg' options
5249 to enable a callback that displays all protocol messages.
5252 *) Change the shared library support so shared libraries are built as
5253 soon as the corresponding static library is finished, and thereby get
5254 openssl and the test programs linked against the shared library.
5255 This still only happens when the keyword "shard" has been given to
5256 the configuration scripts.
5258 NOTE: shared library support is still an experimental thing, and
5259 backward binary compatibility is still not guaranteed.
5260 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5262 *) Add support for Subject Information Access extension.
5263 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5265 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5266 additional bytes when new memory had to be allocated, not just
5267 when reusing an existing buffer.
5270 *) New command line and configuration option 'utf8' for the req command.
5271 This allows field values to be specified as UTF8 strings.
5274 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5275 runs for the former and machine-readable output for the latter.
5278 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5279 of the e-mail address in the DN (i.e., it will go into a certificate
5280 extension only). The new configuration file option 'email_in_dn = no'
5281 has the same effect.
5282 [Massimiliano Pala madwolf@openca.org]
5284 *) Change all functions with names starting with des_ to be starting
5285 with DES_ instead. Add wrappers that are compatible with libdes,
5286 but are named _ossl_old_des_*. Finally, add macros that map the
5287 des_* symbols to the corresponding _ossl_old_des_* if libdes
5288 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5289 desired, the des_* symbols will be mapped to DES_*, with one
5292 Since we provide two compatibility mappings, the user needs to
5293 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5294 compatibility is desired. The default (i.e., when that macro
5295 isn't defined) is OpenSSL 0.9.6c compatibility.
5297 There are also macros that enable and disable the support of old
5298 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5299 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5300 are defined, the default will apply: to support the old des routines.
5302 In either case, one must include openssl/des.h to get the correct
5303 definitions. Do not try to just include openssl/des_old.h, that
5306 NOTE: This is a major break of an old API into a new one. Software
5307 authors are encouraged to switch to the DES_ style functions. Some
5308 time in the future, des_old.h and the libdes compatibility functions
5309 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5310 default), and then completely removed.
5313 *) Test for certificates which contain unsupported critical extensions.
5314 If such a certificate is found during a verify operation it is
5315 rejected by default: this behaviour can be overridden by either
5316 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5317 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5318 X509_supported_extension() has also been added which returns 1 if a
5319 particular extension is supported.
5322 *) Modify the behaviour of EVP cipher functions in similar way to digests
5323 to retain compatibility with existing code.
5326 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5327 compatibility with existing code. In particular the 'ctx' parameter does
5328 not have to be to be initialized before the call to EVP_DigestInit() and
5329 it is tidied up after a call to EVP_DigestFinal(). New function
5330 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5331 EVP_MD_CTX_copy() changed to not require the destination to be
5332 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5333 requires the destination to be valid.
5335 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5336 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5339 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5340 so that complete 'Handshake' protocol structures are kept in memory
5341 instead of overwriting 'msg_type' and 'length' with 'body' data.
5344 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5345 [Massimo Santin via Richard Levitte]
5347 *) Major restructuring to the underlying ENGINE code. This includes
5348 reduction of linker bloat, separation of pure "ENGINE" manipulation
5349 (initialisation, etc) from functionality dealing with implementations
5350 of specific crypto iterfaces. This change also introduces integrated
5351 support for symmetric ciphers and digest implementations - so ENGINEs
5352 can now accelerate these by providing EVP_CIPHER and EVP_MD
5353 implementations of their own. This is detailed in crypto/engine/README
5354 as it couldn't be adequately described here. However, there are a few
5355 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5356 were changed in the original introduction of ENGINE code have now
5357 reverted back - the hooking from this code to ENGINE is now a good
5358 deal more passive and at run-time, operations deal directly with
5359 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5360 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5361 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5362 they were not being used by the framework as there is no concept of a
5363 BIGNUM_METHOD and they could not be generalised to the new
5364 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5365 ENGINE_cpy() has been removed as it cannot be consistently defined in
5369 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5372 *) Change mkdef.pl to sort symbols that get the same entry number,
5373 and make sure the automatically generated functions ERR_load_*
5374 become part of libeay.num as well.
5377 *) New function SSL_renegotiate_pending(). This returns true once
5378 renegotiation has been requested (either SSL_renegotiate() call
5379 or HelloRequest/ClientHello received from the peer) and becomes
5380 false once a handshake has been completed.
5381 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5382 sends a HelloRequest, but does not ensure that a handshake takes
5383 place. SSL_renegotiate_pending() is useful for checking if the
5384 client has followed the request.)
5387 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5388 By default, clients may request session resumption even during
5389 renegotiation (if session ID contexts permit); with this option,
5390 session resumption is possible only in the first handshake.
5392 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5393 more bits available for options that should not be part of
5394 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5397 *) Add some demos for certificate and certificate request creation.
5400 *) Make maximum certificate chain size accepted from the peer application
5401 settable (SSL*_get/set_max_cert_list()), as proposed by
5402 "Douglas E. Engert" <deengert@anl.gov>.
5405 *) Add support for shared libraries for Unixware-7
5406 (Boyd Lynn Gerber <gerberb@zenez.com>).
5409 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5410 be done prior to destruction. Use this to unload error strings from
5411 ENGINEs that load their own error strings. NB: This adds two new API
5412 functions to "get" and "set" this destroy handler in an ENGINE.
5415 *) Alter all existing ENGINE implementations (except "openssl" and
5416 "openbsd") to dynamically instantiate their own error strings. This
5417 makes them more flexible to be built both as statically-linked ENGINEs
5418 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5419 Also, add stub code to each that makes building them as self-contained
5420 shared-libraries easier (see README.ENGINE).
5423 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5424 implementations into applications that are completely implemented in
5425 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5426 commands that can be used to configure what shared-library to load and
5427 to control aspects of the way it is handled. Also, made an update to
5428 the README.ENGINE file that brings its information up-to-date and
5429 provides some information and instructions on the "dynamic" ENGINE
5430 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5433 *) Make it possible to unload ranges of ERR strings with a new
5434 "ERR_unload_strings" function.
5437 *) Add a copy() function to EVP_MD.
5440 *) Make EVP_MD routines take a context pointer instead of just the
5441 md_data void pointer.
5444 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5445 that the digest can only process a single chunk of data
5446 (typically because it is provided by a piece of
5447 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5448 is only going to provide a single chunk of data, and hence the
5449 framework needn't accumulate the data for oneshot drivers.
5452 *) As with "ERR", make it possible to replace the underlying "ex_data"
5453 functions. This change also alters the storage and management of global
5454 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5455 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5456 index counters. The API functions that use this state have been changed
5457 to take a "class_index" rather than pointers to the class's local STACK
5458 and counter, and there is now an API function to dynamically create new
5459 classes. This centralisation allows us to (a) plug a lot of the
5460 thread-safety problems that existed, and (b) makes it possible to clean
5461 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5462 such data would previously have always leaked in application code and
5463 workarounds were in place to make the memory debugging turn a blind eye
5464 to it. Application code that doesn't use this new function will still
5465 leak as before, but their memory debugging output will announce it now
5466 rather than letting it slide.
5468 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5469 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5470 has a return value to indicate success or failure.
5473 *) Make it possible to replace the underlying "ERR" functions such that the
5474 global state (2 LHASH tables and 2 locks) is only used by the "default"
5475 implementation. This change also adds two functions to "get" and "set"
5476 the implementation prior to it being automatically set the first time
5477 any other ERR function takes place. Ie. an application can call "get",
5478 pass the return value to a module it has just loaded, and that module
5479 can call its own "set" function using that value. This means the
5480 module's "ERR" operations will use (and modify) the error state in the
5481 application and not in its own statically linked copy of OpenSSL code.
5484 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5485 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5486 the operation, and provides a more encapsulated way for external code
5487 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5488 to use these functions rather than manually incrementing the counts.
5490 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5493 *) Add EVP test program.
5496 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5499 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5500 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5501 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5502 These allow a CRL to be built without having to access X509_CRL fields
5503 directly. Modify 'ca' application to use new functions.
5506 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5507 bug workarounds. Rollback attack detection is a security feature.
5508 The problem will only arise on OpenSSL servers when TLSv1 is not
5509 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5510 Software authors not wanting to support TLSv1 will have special reasons
5511 for their choice and can explicitly enable this option.
5512 [Bodo Moeller, Lutz Jaenicke]
5514 *) Rationalise EVP so it can be extended: don't include a union of
5515 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5516 (similar to those existing for EVP_CIPHER_CTX).
5521 EVP_MD_CTX_init(&md); /* new function call */
5522 EVP_DigestInit(&md, EVP_sha1());
5523 EVP_DigestUpdate(&md, in, len);
5524 EVP_DigestFinal(&md, out, NULL);
5525 EVP_MD_CTX_cleanup(&md); /* new function call */
5529 *) Make DES key schedule conform to the usual scheme, as well as
5530 correcting its structure. This means that calls to DES functions
5531 now have to pass a pointer to a des_key_schedule instead of a
5532 plain des_key_schedule (which was actually always a pointer
5535 des_key_schedule ks;
5537 des_set_key_checked(..., &ks);
5538 des_ncbc_encrypt(..., &ks, ...);
5540 (Note that a later change renames 'des_...' into 'DES_...'.)
5543 *) Initial reduction of linker bloat: the use of some functions, such as
5544 PEM causes large amounts of unused functions to be linked in due to
5545 poor organisation. For example pem_all.c contains every PEM function
5546 which has a knock on effect of linking in large amounts of (unused)
5547 ASN1 code. Grouping together similar functions and splitting unrelated
5548 functions prevents this.
5551 *) Cleanup of EVP macros.
5554 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5555 correct _ecb suffix.
5558 *) Add initial OCSP responder support to ocsp application. The
5559 revocation information is handled using the text based index
5560 use by the ca application. The responder can either handle
5561 requests generated internally, supplied in files (for example
5562 via a CGI script) or using an internal minimal server.
5565 *) Add configuration choices to get zlib compression for TLS.
5568 *) Changes to Kerberos SSL for RFC 2712 compliance:
5569 1. Implemented real KerberosWrapper, instead of just using
5570 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5571 2. Implemented optional authenticator field of KerberosWrapper.
5573 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5574 and authenticator structs; see crypto/krb5/.
5576 Generalized Kerberos calls to support multiple Kerberos libraries.
5577 [Vern Staats <staatsvr@asc.hpc.mil>,
5578 Jeffrey Altman <jaltman@columbia.edu>
5579 via Richard Levitte]
5581 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5582 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5583 values for each of the key sizes rather than having just
5584 parameters (and 'speed' generating keys each time).
5587 *) Speed up EVP routines.
5590 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5591 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5592 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5593 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5595 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5596 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5597 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5600 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5602 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5605 *) Added the OS2-EMX target.
5606 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5608 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5609 to support NCONF routines in extension code. New function CONF_set_nconf()
5610 to allow functions which take an NCONF to also handle the old LHASH
5611 structure: this means that the old CONF compatible routines can be
5612 retained (in particular wrt extensions) without having to duplicate the
5613 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5616 *) Enhance the general user interface with mechanisms for inner control
5617 and with possibilities to have yes/no kind of prompts.
5620 *) Change all calls to low level digest routines in the library and
5621 applications to use EVP. Add missing calls to HMAC_cleanup() and
5622 don't assume HMAC_CTX can be copied using memcpy().
5623 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5625 *) Add the possibility to control engines through control names but with
5626 arbitrary arguments instead of just a string.
5627 Change the key loaders to take a UI_METHOD instead of a callback
5628 function pointer. NOTE: this breaks binary compatibility with earlier
5629 versions of OpenSSL [engine].
5630 Adapt the nCipher code for these new conditions and add a card insertion
5634 *) Enhance the general user interface with mechanisms to better support
5635 dialog box interfaces, application-defined prompts, the possibility
5636 to use defaults (for example default passwords from somewhere else)
5637 and interrupts/cancellations.
5640 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5641 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5644 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5645 tidy up some unnecessarily weird code in 'sk_new()').
5646 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5648 *) Change the key loading routines for ENGINEs to use the same kind
5649 callback (pem_password_cb) as all other routines that need this
5653 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5654 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5655 than this minimum value is recommended.
5658 *) New random seeder for OpenVMS, using the system process statistics
5659 that are easily reachable.
5662 *) Windows apparently can't transparently handle global
5663 variables defined in DLLs. Initialisations such as:
5665 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5667 wont compile. This is used by the any applications that need to
5668 declare their own ASN1 modules. This was fixed by adding the option
5669 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5670 needed for static libraries under Win32.
5673 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5674 setting of purpose and trust fields. New X509_STORE trust and
5675 purpose functions and tidy up setting in other SSL functions.
5678 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5679 structure. These are inherited by X509_STORE_CTX when it is
5680 initialised. This allows various defaults to be set in the
5681 X509_STORE structure (such as flags for CRL checking and custom
5682 purpose or trust settings) for functions which only use X509_STORE_CTX
5683 internally such as S/MIME.
5685 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5686 trust settings if they are not set in X509_STORE. This allows X509_STORE
5687 purposes and trust (in S/MIME for example) to override any set by default.
5689 Add command line options for CRL checking to smime, s_client and s_server
5693 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5694 are set then the CRL is looked up in the X509_STORE structure and
5695 its validity and signature checked, then if the certificate is found
5696 in the CRL the verify fails with a revoked error.
5698 Various new CRL related callbacks added to X509_STORE_CTX structure.
5700 Command line options added to 'verify' application to support this.
5702 This needs some additional work, such as being able to handle multiple
5703 CRLs with different times, extension based lookup (rather than just
5704 by subject name) and ultimately more complete V2 CRL extension
5708 *) Add a general user interface API (crypto/ui/). This is designed
5709 to replace things like des_read_password and friends (backward
5710 compatibility functions using this new API are provided).
5711 The purpose is to remove prompting functions from the DES code
5712 section as well as provide for prompting through dialog boxes in
5713 a window system and the like.
5716 *) Add "ex_data" support to ENGINE so implementations can add state at a
5717 per-structure level rather than having to store it globally.
5720 *) Make it possible for ENGINE structures to be copied when retrieved by
5721 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5722 This causes the "original" ENGINE structure to act like a template,
5723 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5724 operational state can be localised to each ENGINE structure, despite the
5725 fact they all share the same "methods". New ENGINE structures returned in
5726 this case have no functional references and the return value is the single
5727 structural reference. This matches the single structural reference returned
5728 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5732 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5733 needs to match any other type at all we need to manually clear the
5737 *) Changes to the "openssl engine" utility to include;
5738 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5739 about an ENGINE's available control commands.
5740 - executing control commands from command line arguments using the
5741 '-pre' and '-post' switches. '-post' is only used if '-t' is
5742 specified and the ENGINE is successfully initialised. The syntax for
5743 the individual commands are colon-separated, for example;
5744 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5747 *) New dynamic control command support for ENGINEs. ENGINEs can now
5748 declare their own commands (numbers), names (strings), descriptions,
5749 and input types for run-time discovery by calling applications. A
5750 subset of these commands are implicitly classed as "executable"
5751 depending on their input type, and only these can be invoked through
5752 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5753 can be based on user input, config files, etc). The distinction is
5754 that "executable" commands cannot return anything other than a boolean
5755 result and can only support numeric or string input, whereas some
5756 discoverable commands may only be for direct use through
5757 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5758 pointers, or other custom uses. The "executable" commands are to
5759 support parameterisations of ENGINE behaviour that can be
5760 unambiguously defined by ENGINEs and used consistently across any
5761 OpenSSL-based application. Commands have been added to all the
5762 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5763 control over shared-library paths without source code alterations.
5766 *) Changed all ENGINE implementations to dynamically allocate their
5767 ENGINEs rather than declaring them statically. Apart from this being
5768 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5769 this also allows the implementations to compile without using the
5770 internal engine_int.h header.
5773 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5774 'const' value. Any code that should be able to modify a RAND_METHOD
5775 should already have non-const pointers to it (ie. they should only
5776 modify their own ones).
5779 *) Made a variety of little tweaks to the ENGINE code.
5780 - "atalla" and "ubsec" string definitions were moved from header files
5781 to C code. "nuron" string definitions were placed in variables
5782 rather than hard-coded - allowing parameterisation of these values
5783 later on via ctrl() commands.
5784 - Removed unused "#if 0"'d code.
5785 - Fixed engine list iteration code so it uses ENGINE_free() to release
5786 structural references.
5787 - Constified the RAND_METHOD element of ENGINE structures.
5788 - Constified various get/set functions as appropriate and added
5789 missing functions (including a catch-all ENGINE_cpy that duplicates
5790 all ENGINE values onto a new ENGINE except reference counts/state).
5791 - Removed NULL parameter checks in get/set functions. Setting a method
5792 or function to NULL is a way of cancelling out a previously set
5793 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5794 and doesn't justify the extra error symbols and code.
5795 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5796 flags from engine_int.h to engine.h.
5797 - Changed prototypes for ENGINE handler functions (init(), finish(),
5798 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5801 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5802 to the algorithm using long division. The binary algorithm can be
5803 used only if the modulus is odd. On 32-bit systems, it is faster
5804 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5805 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5806 up to 450 bits. In 64-bit environments, the binary algorithm
5807 appears to be advantageous for much longer moduli; here we use it
5808 for moduli up to 2048 bits.
5811 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5812 could not support the combine flag in choice fields.
5815 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5816 extensions from a certificate request to the certificate.
5819 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5820 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5821 file: this allows the display of the certificate about to be
5822 signed to be customised, to allow certain fields to be included
5823 or excluded and extension details. The old system didn't display
5824 multicharacter strings properly, omitted fields not in the policy
5825 and couldn't display additional details such as extensions.
5828 *) Function EC_POINTs_mul for multiple scalar multiplication
5829 of an arbitrary number of elliptic curve points
5830 \sum scalars[i]*points[i],
5831 optionally including the generator defined for the EC_GROUP:
5832 scalar*generator + \sum scalars[i]*points[i].
5834 EC_POINT_mul is a simple wrapper function for the typical case
5835 that the point list has just one item (besides the optional
5839 *) First EC_METHODs for curves over GF(p):
5841 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5842 operations and provides various method functions that can also
5843 operate with faster implementations of modular arithmetic.
5845 EC_GFp_mont_method() reuses most functions that are part of
5846 EC_GFp_simple_method, but uses Montgomery arithmetic.
5848 [Bodo Moeller; point addition and point doubling
5849 implementation directly derived from source code provided by
5850 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5852 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5853 crypto/ec/ec_lib.c):
5855 Curves are EC_GROUP objects (with an optional group generator)
5856 based on EC_METHODs that are built into the library.
5858 Points are EC_POINT objects based on EC_GROUP objects.
5860 Most of the framework would be able to handle curves over arbitrary
5861 finite fields, but as there are no obvious types for fields other
5862 than GF(p), some functions are limited to that for now.
5865 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5866 that the file contains a complete HTTP response.
5869 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5870 change the def and num file printf format specifier from "%-40sXXX"
5871 to "%-39s XXX". The latter will always guarantee a space after the
5872 field while the former will cause them to run together if the field
5873 is 40 of more characters long.
5876 *) Constify the cipher and digest 'method' functions and structures
5877 and modify related functions to take constant EVP_MD and EVP_CIPHER
5881 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5882 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5885 *) Modify EVP_Digest*() routines so they now return values. Although the
5886 internal software routines can never fail additional hardware versions
5890 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5892 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5893 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5896 ERR_R_NESTED_ASN1_ERROR
5898 ERR_R_MISSING_ASN1_EOS
5899 were 4 .. 9, conflicting with
5900 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5902 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5903 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5905 Add new error code 'ERR_R_INTERNAL_ERROR'.
5908 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5912 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5913 sets the subject name for a new request or supersedes the
5914 subject name in a given request. Formats that can be parsed are
5915 'CN=Some Name, OU=myOU, C=IT'
5917 'CN=Some Name/OU=myOU/C=IT'.
5919 Add options '-batch' and '-verbose' to 'openssl req'.
5920 [Massimiliano Pala <madwolf@hackmasters.net>]
5922 *) Introduce the possibility to access global variables through
5923 functions on platform were that's the best way to handle exporting
5924 global variables in shared libraries. To enable this functionality,
5925 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5926 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5927 is normally done by Configure or something similar).
5929 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5930 in the source file (foo.c) like this:
5932 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5933 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5935 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5936 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5938 OPENSSL_DECLARE_GLOBAL(int,foo);
5939 #define foo OPENSSL_GLOBAL_REF(foo)
5940 OPENSSL_DECLARE_GLOBAL(double,bar);
5941 #define bar OPENSSL_GLOBAL_REF(bar)
5943 The #defines are very important, and therefore so is including the
5944 header file everywhere where the defined globals are used.
5946 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5947 of ASN.1 items, but that structure is a bit different.
5949 The largest change is in util/mkdef.pl which has been enhanced with
5950 better and easier to understand logic to choose which symbols should
5951 go into the Windows .def files as well as a number of fixes and code
5952 cleanup (among others, algorithm keywords are now sorted
5953 lexicographically to avoid constant rewrites).
5956 *) In BN_div() keep a copy of the sign of 'num' before writing the
5957 result to 'rm' because if rm==num the value will be overwritten
5958 and produce the wrong result if 'num' is negative: this caused
5959 problems with BN_mod() and BN_nnmod().
5962 *) Function OCSP_request_verify(). This checks the signature on an
5963 OCSP request and verifies the signer certificate. The signer
5964 certificate is just checked for a generic purpose and OCSP request
5968 *) Add OCSP_check_validity() function to check the validity of OCSP
5969 responses. OCSP responses are prepared in real time and may only
5970 be a few seconds old. Simply checking that the current time lies
5971 between thisUpdate and nextUpdate max reject otherwise valid responses
5972 caused by either OCSP responder or client clock inaccuracy. Instead
5973 we allow thisUpdate and nextUpdate to fall within a certain period of
5974 the current time. The age of the response can also optionally be
5975 checked. Two new options -validity_period and -status_age added to
5979 *) If signature or public key algorithm is unrecognized print out its
5980 OID rather that just UNKNOWN.
5983 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5984 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5985 ID to be generated from the issuer certificate alone which can then be
5986 passed to OCSP_id_issuer_cmp().
5989 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5990 ASN1 modules to export functions returning ASN1_ITEM pointers
5991 instead of the ASN1_ITEM structures themselves. This adds several
5992 new macros which allow the underlying ASN1 function/structure to
5993 be accessed transparently. As a result code should not use ASN1_ITEM
5994 references directly (such as &X509_it) but instead use the relevant
5995 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5996 use of the new ASN1 code on platforms where exporting structures
5997 is problematical (for example in shared libraries) but exporting
5998 functions returning pointers to structures is not.
6001 *) Add support for overriding the generation of SSL/TLS session IDs.
6002 These callbacks can be registered either in an SSL_CTX or per SSL.
6003 The purpose of this is to allow applications to control, if they wish,
6004 the arbitrary values chosen for use as session IDs, particularly as it
6005 can be useful for session caching in multiple-server environments. A
6006 command-line switch for testing this (and any client code that wishes
6007 to use such a feature) has been added to "s_server".
6008 [Geoff Thorpe, Lutz Jaenicke]
6010 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6011 of the form '#if defined(...) || defined(...) || ...' and
6012 '#if !defined(...) && !defined(...) && ...'. This also avoids
6013 the growing number of special cases it was previously handling.
6016 *) Make all configuration macros available for application by making
6017 sure they are available in opensslconf.h, by giving them names starting
6018 with "OPENSSL_" to avoid conflicts with other packages and by making
6019 sure e_os2.h will cover all platform-specific cases together with
6021 Additionally, it is now possible to define configuration/platform-
6022 specific names (called "system identities"). In the C code, these
6023 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6024 macro with the name beginning with "OPENSSL_SYS_", which is determined
6025 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6029 *) New option -set_serial to 'req' and 'x509' this allows the serial
6030 number to use to be specified on the command line. Previously self
6031 signed certificates were hard coded with serial number 0 and the
6032 CA options of 'x509' had to use a serial number in a file which was
6036 *) New options to 'ca' utility to support V2 CRL entry extensions.
6037 Currently CRL reason, invalidity date and hold instruction are
6038 supported. Add new CRL extensions to V3 code and some new objects.
6041 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6042 disable standard block padding (aka PKCS#5 padding) in the EVP
6043 API, which was previously mandatory. This means that the data is
6044 not padded in any way and so the total length much be a multiple
6045 of the block size, otherwise an error occurs.
6048 *) Initial (incomplete) OCSP SSL support.
6051 *) New function OCSP_parse_url(). This splits up a URL into its host,
6052 port and path components: primarily to parse OCSP URLs. New -url
6053 option to ocsp utility.
6056 *) New nonce behavior. The return value of OCSP_check_nonce() now
6057 reflects the various checks performed. Applications can decide
6058 whether to tolerate certain situations such as an absent nonce
6059 in a response when one was present in a request: the ocsp application
6060 just prints out a warning. New function OCSP_add1_basic_nonce()
6061 this is to allow responders to include a nonce in a response even if
6062 the request is nonce-less.
6065 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6066 skipped when using openssl x509 multiple times on a single input file,
6067 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6070 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6071 set string type: to handle setting ASN1_TIME structures. Fix ca
6072 utility to correctly initialize revocation date of CRLs.
6075 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6076 the clients preferred ciphersuites and rather use its own preferences.
6077 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6078 Internet Explorer by ensuring unchanged hash method during stepup.
6079 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6082 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6083 to aes and add a new 'exist' option to print out symbols that don't
6087 *) Additional options to ocsp utility to allow flags to be set and
6088 additional certificates supplied.
6091 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6092 OCSP client a number of certificate to only verify the response
6096 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6097 handle the new API. Currently only ECB, CBC modes supported. Add new
6100 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6101 Encryption Standard (AES) Ciphersuites for Transport Layer
6102 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6103 not enabled by default and were not part of the "ALL" ciphersuite
6104 alias because they were not yet official; they could be
6105 explicitly requested by specifying the "AESdraft" ciphersuite
6106 group alias. In the final release of OpenSSL 0.9.7, the group
6107 alias is called "AES" and is part of "ALL".)
6108 [Ben Laurie, Steve Henson, Bodo Moeller]
6110 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6111 request to response.
6114 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6115 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6116 extract information from a certificate request. OCSP_response_create()
6117 creates a response and optionally adds a basic response structure.
6118 OCSP_basic_add1_status() adds a complete single response to a basic
6119 response and returns the OCSP_SINGLERESP structure just added (to allow
6120 extensions to be included for example). OCSP_basic_add1_cert() adds a
6121 certificate to a basic response and OCSP_basic_sign() signs a basic
6122 response with various flags. New helper functions ASN1_TIME_check()
6123 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6124 (converts ASN1_TIME to GeneralizedTime).
6127 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6128 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6129 structure from a certificate. X509_pubkey_digest() digests the public_key
6130 contents: this is used in various key identifiers.
6133 *) Make sk_sort() tolerate a NULL argument.
6134 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6136 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6137 passed by the function are trusted implicitly. If any of them signed the
6138 response then it is assumed to be valid and is not verified.
6141 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6142 to data. This was previously part of the PKCS7 ASN1 code. This
6143 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6144 [Steve Henson, reported by Kenneth R. Robinette
6145 <support@securenetterm.com>]
6147 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6148 routines: without these tracing memory leaks is very painful.
6149 Fix leaks in PKCS12 and PKCS7 routines.
6152 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6153 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6154 effectively meant GeneralizedTime would never be used. Now it
6155 is initialised to -1 but X509_time_adj() now has to check the value
6156 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6157 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6158 [Steve Henson, reported by Kenneth R. Robinette
6159 <support@securenetterm.com>]
6161 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6162 result in a zero length in the ASN1_INTEGER structure which was
6163 not consistent with the structure when d2i_ASN1_INTEGER() was used
6164 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6165 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6166 where it did not print out a minus for negative ASN1_INTEGER.
6169 *) Add summary printout to ocsp utility. The various functions which
6170 convert status values to strings have been renamed to:
6171 OCSP_response_status_str(), OCSP_cert_status_str() and
6172 OCSP_crl_reason_str() and are no longer static. New options
6173 to verify nonce values and to disable verification. OCSP response
6174 printout format cleaned up.
6177 *) Add additional OCSP certificate checks. These are those specified
6178 in RFC2560. This consists of two separate checks: the CA of the
6179 certificate being checked must either be the OCSP signer certificate
6180 or the issuer of the OCSP signer certificate. In the latter case the
6181 OCSP signer certificate must contain the OCSP signing extended key
6182 usage. This check is performed by attempting to match the OCSP
6183 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6184 in the OCSP_CERTID structures of the response.
6187 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6188 and related routines. This uses the standard OpenSSL certificate
6189 verify routines to perform initial checks (just CA validity) and
6190 to obtain the certificate chain. Then additional checks will be
6191 performed on the chain. Currently the root CA is checked to see
6192 if it is explicitly trusted for OCSP signing. This is used to set
6193 a root CA as a global signing root: that is any certificate that
6194 chains to that CA is an acceptable OCSP signing certificate.
6197 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6198 extensions from a separate configuration file.
6199 As when reading extensions from the main configuration file,
6200 the '-extensions ...' option may be used for specifying the
6202 [Massimiliano Pala <madwolf@comune.modena.it>]
6204 *) New OCSP utility. Allows OCSP requests to be generated or
6205 read. The request can be sent to a responder and the output
6206 parsed, outputed or printed in text form. Not complete yet:
6207 still needs to check the OCSP response validity.
6210 *) New subcommands for 'openssl ca':
6211 'openssl ca -status <serial>' prints the status of the cert with
6212 the given serial number (according to the index file).
6213 'openssl ca -updatedb' updates the expiry status of certificates
6215 [Massimiliano Pala <madwolf@comune.modena.it>]
6217 *) New '-newreq-nodes' command option to CA.pl. This is like
6218 '-newreq', but calls 'openssl req' with the '-nodes' option
6219 so that the resulting key is not encrypted.
6220 [Damien Miller <djm@mindrot.org>]
6222 *) New configuration for the GNU Hurd.
6223 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6225 *) Initial code to implement OCSP basic response verify. This
6226 is currently incomplete. Currently just finds the signer's
6227 certificate and verifies the signature on the response.
6230 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6231 value of OPENSSLDIR. This is available via the new '-d' option
6232 to 'openssl version', and is also included in 'openssl version -a'.
6235 *) Allowing defining memory allocation callbacks that will be given
6236 file name and line number information in additional arguments
6237 (a const char* and an int). The basic functionality remains, as
6238 well as the original possibility to just replace malloc(),
6239 realloc() and free() by functions that do not know about these
6240 additional arguments. To register and find out the current
6241 settings for extended allocation functions, the following
6242 functions are provided:
6244 CRYPTO_set_mem_ex_functions
6245 CRYPTO_set_locked_mem_ex_functions
6246 CRYPTO_get_mem_ex_functions
6247 CRYPTO_get_locked_mem_ex_functions
6249 These work the same way as CRYPTO_set_mem_functions and friends.
6250 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6251 extended allocation function is enabled.
6252 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6253 a conventional allocation function is enabled.
6254 [Richard Levitte, Bodo Moeller]
6256 *) Finish off removing the remaining LHASH function pointer casts.
6257 There should no longer be any prototype-casting required when using
6258 the LHASH abstraction, and any casts that remain are "bugs". See
6259 the callback types and macros at the head of lhash.h for details
6260 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6263 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6264 If /dev/[u]random devices are not available or do not return enough
6265 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6267 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6268 /etc/entropy will be queried once each in this sequence, quering stops
6269 when enough entropy was collected without querying more sockets.
6272 *) Change the Unix RAND_poll() variant to be able to poll several
6273 random devices, as specified by DEVRANDOM, until a sufficient amount
6274 of data has been collected. We spend at most 10 ms on each file
6275 (select timeout) and read in non-blocking mode. DEVRANDOM now
6276 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6277 (previously it was just the string "/dev/urandom"), so on typical
6278 platforms the 10 ms delay will never occur.
6279 Also separate out the Unix variant to its own file, rand_unix.c.
6280 For VMS, there's a currently-empty rand_vms.c.
6283 *) Move OCSP client related routines to ocsp_cl.c. These
6284 provide utility functions which an application needing
6285 to issue a request to an OCSP responder and analyse the
6286 response will typically need: as opposed to those which an
6287 OCSP responder itself would need which will be added later.
6289 OCSP_request_sign() signs an OCSP request with an API similar
6290 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6291 response. OCSP_response_get1_basic() extracts basic response
6292 from response. OCSP_resp_find_status(): finds and extracts status
6293 information from an OCSP_CERTID structure (which will be created
6294 when the request structure is built). These are built from lower
6295 level functions which work on OCSP_SINGLERESP structures but
6296 wont normally be used unless the application wishes to examine
6297 extensions in the OCSP response for example.
6299 Replace nonce routines with a pair of functions.
6300 OCSP_request_add1_nonce() adds a nonce value and optionally
6301 generates a random value. OCSP_check_nonce() checks the
6302 validity of the nonce in an OCSP response.
6305 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6306 This doesn't copy the supplied OCSP_CERTID and avoids the
6307 need to free up the newly created id. Change return type
6308 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6309 This can then be used to add extensions to the request.
6310 Deleted OCSP_request_new(), since most of its functionality
6311 is now in OCSP_REQUEST_new() (and the case insensitive name
6312 clash) apart from the ability to set the request name which
6313 will be added elsewhere.
6316 *) Update OCSP API. Remove obsolete extensions argument from
6317 various functions. Extensions are now handled using the new
6318 OCSP extension code. New simple OCSP HTTP function which
6319 can be used to send requests and parse the response.
6322 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6323 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6324 uses the special reorder version of SET OF to sort the attributes
6325 and reorder them to match the encoded order. This resolves a long
6326 standing problem: a verify on a PKCS7 structure just after signing
6327 it used to fail because the attribute order did not match the
6328 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6329 it uses the received order. This is necessary to tolerate some broken
6330 software that does not order SET OF. This is handled by encoding
6331 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6332 to produce the required SET OF.
6335 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6336 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6337 files to get correct declarations of the ASN.1 item variables.
6340 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6341 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6342 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6343 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6344 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6345 ASN1_ITEM and no wrapper functions.
6348 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6349 replace the old function pointer based I/O routines. Change most of
6350 the *_d2i_bio() and *_d2i_fp() functions to use these.
6353 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6354 lines, recognice more "algorithms" that can be deselected, and make
6355 it complain about algorithm deselection that isn't recognised.
6358 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6359 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6360 to use new functions. Add NO_ASN1_OLD which can be set to remove
6361 some old style ASN1 functions: this can be used to determine if old
6362 code will still work when these eventually go away.
6365 *) New extension functions for OCSP structures, these follow the
6366 same conventions as certificates and CRLs.
6369 *) New function X509V3_add1_i2d(). This automatically encodes and
6370 adds an extension. Its behaviour can be customised with various
6371 flags to append, replace or delete. Various wrappers added for
6372 certifcates and CRLs.
6375 *) Fix to avoid calling the underlying ASN1 print routine when
6376 an extension cannot be parsed. Correct a typo in the
6377 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6380 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6381 entries for variables.
6384 *) Add functionality to apps/openssl.c for detecting locking
6385 problems: As the program is single-threaded, all we have
6386 to do is register a locking callback using an array for
6387 storing which locks are currently held by the program.
6390 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6391 SSL_get_ex_data_X509_STORE_idx(), which is used in
6392 ssl_verify_cert_chain() and thus can be called at any time
6393 during TLS/SSL handshakes so that thread-safety is essential.
6394 Unfortunately, the ex_data design is not at all suited
6395 for multi-threaded use, so it probably should be abolished.
6398 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6399 [Broadcom, tweaked and integrated by Geoff Thorpe]
6401 *) Move common extension printing code to new function
6402 X509V3_print_extensions(). Reorganise OCSP print routines and
6403 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6406 *) New function X509_signature_print() to remove duplication in some
6410 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6411 set (this was treated exactly the same as SET OF previously). This
6412 is used to reorder the STACK representing the structure to match the
6413 encoding. This will be used to get round a problem where a PKCS7
6414 structure which was signed could not be verified because the STACK
6415 order did not reflect the encoded order.
6418 *) Reimplement the OCSP ASN1 module using the new code.
6421 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6422 for its ASN1 operations. The old style function pointers still exist
6423 for now but they will eventually go away.
6426 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6427 completely replaces the old ASN1 functionality with a table driven
6428 encoder and decoder which interprets an ASN1_ITEM structure describing
6429 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6430 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6431 has also been converted to the new form.
6434 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6435 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6436 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6437 for negative moduli.
6440 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6441 of not touching the result's sign bit.
6444 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6448 *) Changed the LHASH code to use prototypes for callbacks, and created
6449 macros to declare and implement thin (optionally static) functions
6450 that provide type-safety and avoid function pointer casting for the
6451 type-specific callbacks.
6454 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6456 [Veers Staats <staatsvr@asc.hpc.mil>,
6457 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6459 *) Reformat the FAQ so the different questions and answers can be divided
6460 in sections depending on the subject.
6463 *) Have the zlib compression code load ZLIB.DLL dynamically under
6467 *) New function BN_mod_sqrt for computing square roots modulo a prime
6468 (using the probabilistic Tonelli-Shanks algorithm unless
6469 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6470 be handled deterministically).
6471 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6473 *) Make BN_mod_inverse faster by explicitly handling small quotients
6474 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6475 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6478 *) New function BN_kronecker.
6481 *) Fix BN_gcd so that it works on negative inputs; the result is
6482 positive unless both parameters are zero.
6483 Previously something reasonably close to an infinite loop was
6484 possible because numbers could be growing instead of shrinking
6485 in the implementation of Euclid's algorithm.
6488 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6489 sign of the number in question.
6491 Fix BN_is_word(a,w) to work correctly for w == 0.
6493 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6494 because its test if the absolute value of 'a' equals 'w'.
6495 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6496 it exists mostly for use in the implementations of BN_is_zero(),
6497 BN_is_one(), and BN_is_word().
6500 *) New function BN_swap.
6503 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6504 the exponentiation functions are more likely to produce reasonable
6505 results on negative inputs.
6508 *) Change BN_mod_mul so that the result is always non-negative.
6509 Previously, it could be negative if one of the factors was negative;
6510 I don't think anyone really wanted that behaviour.
6513 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6514 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6515 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6516 and add new functions:
6525 BN_mod_lshift1_quick
6529 These functions always generate non-negative results.
6531 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6532 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6534 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6535 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6536 be reduced modulo m.
6537 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6540 The following entry accidentally appeared in the CHANGES file
6541 distributed with OpenSSL 0.9.7. The modifications described in
6542 it do *not* apply to OpenSSL 0.9.7.
6544 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6545 was actually never needed) and in BN_mul(). The removal in BN_mul()
6546 required a small change in bn_mul_part_recursive() and the addition
6547 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6548 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6549 bn_sub_words() and bn_add_words() except they take arrays with
6554 *) In 'openssl passwd', verify passwords read from the terminal
6555 unless the '-salt' option is used (which usually means that
6556 verification would just waste user's time since the resulting
6557 hash is going to be compared with some given password hash)
6558 or the new '-noverify' option is used.
6560 This is an incompatible change, but it does not affect
6561 non-interactive use of 'openssl passwd' (passwords on the command
6562 line, '-stdin' option, '-in ...' option) and thus should not
6566 *) Remove all references to RSAref, since there's no more need for it.
6569 *) Make DSO load along a path given through an environment variable
6570 (SHLIB_PATH) with shl_load().
6573 *) Constify the ENGINE code as a result of BIGNUM constification.
6574 Also constify the RSA code and most things related to it. In a
6575 few places, most notable in the depth of the ASN.1 code, ugly
6576 casts back to non-const were required (to be solved at a later
6580 *) Make it so the openssl application has all engines loaded by default.
6583 *) Constify the BIGNUM routines a little more.
6586 *) Add the following functions:
6588 ENGINE_load_cswift()
6590 ENGINE_load_atalla()
6592 ENGINE_load_builtin_engines()
6594 That way, an application can itself choose if external engines that
6595 are built-in in OpenSSL shall ever be used or not. The benefit is
6596 that applications won't have to be linked with libdl or other dso
6597 libraries unless it's really needed.
6599 Changed 'openssl engine' to load all engines on demand.
6600 Changed the engine header files to avoid the duplication of some
6601 declarations (they differed!).
6604 *) 'openssl engine' can now list capabilities.
6607 *) Better error reporting in 'openssl engine'.
6610 *) Never call load_dh_param(NULL) in s_server.
6613 *) Add engine application. It can currently list engines by name and
6614 identity, and test if they are actually available.
6617 *) Improve RPM specification file by forcing symbolic linking and making
6618 sure the installed documentation is also owned by root.root.
6619 [Damien Miller <djm@mindrot.org>]
6621 *) Give the OpenSSL applications more possibilities to make use of
6622 keys (public as well as private) handled by engines.
6625 *) Add OCSP code that comes from CertCo.
6628 *) Add VMS support for the Rijndael code.
6631 *) Added untested support for Nuron crypto accelerator.
6634 *) Add support for external cryptographic devices. This code was
6635 previously distributed separately as the "engine" branch.
6636 [Geoff Thorpe, Richard Levitte]
6638 *) Rework the filename-translation in the DSO code. It is now possible to
6639 have far greater control over how a "name" is turned into a filename
6640 depending on the operating environment and any oddities about the
6641 different shared library filenames on each system.
6644 *) Support threads on FreeBSD-elf in Configure.
6647 *) Fix for SHA1 assembly problem with MASM: it produces
6648 warnings about corrupt line number information when assembling
6649 with debugging information. This is caused by the overlapping
6651 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6654 NCONF_get_number() has no error checking at all. As a replacement,
6655 NCONF_get_number_e() is defined (_e for "error checking") and is
6656 promoted strongly. The old NCONF_get_number is kept around for
6657 binary backward compatibility.
6658 Make it possible for methods to load from something other than a BIO,
6659 by providing a function pointer that is given a name instead of a BIO.
6660 For example, this could be used to load configuration data from an
6664 *) Fix for non blocking accept BIOs. Added new I/O special reason
6665 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6666 with non blocking I/O was not possible because no retry code was
6667 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6671 *) Added the beginnings of Rijndael support.
6674 *) Fix for bug in DirectoryString mask setting. Add support for
6675 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6676 to allow certificate printing to more controllable, additional
6677 'certopt' option to 'x509' to allow new printing options to be
6681 *) Clean old EAY MD5 hack from e_os.h.
6684 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6686 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6687 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6688 [Joe Orton, Steve Henson]
6690 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6692 *) Fix additional bug revealed by the NISCC test suite:
6694 Stop bug triggering large recursion when presented with
6695 certain ASN.1 tags (CVE-2003-0851)
6698 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6700 *) Fix various bugs revealed by running the NISCC test suite:
6702 Stop out of bounds reads in the ASN1 code when presented with
6703 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6705 If verify callback ignores invalid public key errors don't try to check
6706 certificate signature with the NULL public key.
6710 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6711 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6715 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6716 extra data after the compression methods not only for TLS 1.0
6717 but also for SSL 3.0 (as required by the specification).
6718 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6720 *) Change X509_certificate_type() to mark the key as exported/exportable
6721 when it's 512 *bits* long, not 512 bytes.
6724 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6726 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6727 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6728 a protocol version number mismatch like a decryption error
6729 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6732 *) Turn on RSA blinding by default in the default implementation
6733 to avoid a timing attack. Applications that don't want it can call
6734 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6735 They would be ill-advised to do so in most cases.
6736 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6738 *) Change RSA blinding code so that it works when the PRNG is not
6739 seeded (in this case, the secret RSA exponent is abused as
6740 an unpredictable seed -- if it is not unpredictable, there
6741 is no point in blinding anyway). Make RSA blinding thread-safe
6742 by remembering the creator's thread ID in rsa->blinding and
6743 having all other threads use local one-time blinding factors
6744 (this requires more computation than sharing rsa->blinding, but
6745 avoids excessive locking; and if an RSA object is not shared
6746 between threads, blinding will still be very fast).
6749 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6751 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6752 via timing by performing a MAC computation even if incorrrect
6753 block cipher padding has been found. This is a countermeasure
6754 against active attacks where the attacker has to distinguish
6755 between bad padding and a MAC verification error. (CVE-2003-0078)
6757 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6758 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6759 Martin Vuagnoux (EPFL, Ilion)]
6761 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6763 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6764 memory from it's contents. This is done with a counter that will
6765 place alternating values in each byte. This can be used to solve
6766 two issues: 1) the removal of calls to memset() by highly optimizing
6767 compilers, and 2) cleansing with other values than 0, since those can
6768 be read through on certain media, for example a swap space on disk.
6771 *) Bugfix: client side session caching did not work with external caching,
6772 because the session->cipher setting was not restored when reloading
6773 from the external cache. This problem was masked, when
6774 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6775 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6778 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6779 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6780 [Zeev Lieber <zeev-l@yahoo.com>]
6782 *) Undo an undocumented change introduced in 0.9.6e which caused
6783 repeated calls to OpenSSL_add_all_ciphers() and
6784 OpenSSL_add_all_digests() to be ignored, even after calling
6788 *) Change the default configuration reader to deal with last line not
6789 being properly terminated.
6792 *) Change X509_NAME_cmp() so it applies the special rules on handling
6793 DN values that are of type PrintableString, as well as RDNs of type
6794 emailAddress where the value has the type ia5String.
6795 [stefank@valicert.com via Richard Levitte]
6797 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6798 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6799 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6800 the bitwise-OR of the two for use by the majority of applications
6801 wanting this behaviour, and update the docs. The documented
6802 behaviour and actual behaviour were inconsistent and had been
6803 changing anyway, so this is more a bug-fix than a behavioural
6805 [Geoff Thorpe, diagnosed by Nadav Har'El]
6807 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6808 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6811 *) Fix initialization code race conditions in
6812 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6813 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6814 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6815 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6816 ssl2_get_cipher_by_char(),
6817 ssl3_get_cipher_by_char().
6818 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6820 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6821 the cached sessions are flushed, as the remove_cb() might use ex_data
6822 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6823 (see [openssl.org #212]).
6824 [Geoff Thorpe, Lutz Jaenicke]
6826 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6827 length, instead of the encoding length to d2i_ASN1_OBJECT.
6830 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6832 *) [In 0.9.6g-engine release:]
6833 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6834 [Lynn Gazis <lgazis@rainbow.com>]
6836 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6838 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6839 and get fix the header length calculation.
6840 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6841 Alon Kantor <alonk@checkpoint.com> (and others),
6844 *) Use proper error handling instead of 'assertions' in buffer
6845 overflow checks added in 0.9.6e. This prevents DoS (the
6846 assertions could call abort()).
6847 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6849 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6851 *) Add various sanity checks to asn1_get_length() to reject
6852 the ASN1 length bytes if they exceed sizeof(long), will appear
6853 negative or the content length exceeds the length of the
6855 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6857 *) Fix cipher selection routines: ciphers without encryption had no flags
6858 for the cipher strength set and where therefore not handled correctly
6859 by the selection routines (PR #130).
6862 *) Fix EVP_dsa_sha macro.
6866 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6867 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6868 that was added in OpenSSL 0.9.6d.
6870 As the countermeasure turned out to be incompatible with some
6871 broken SSL implementations, the new option is part of SSL_OP_ALL.
6872 SSL_OP_ALL is usually employed when compatibility with weird SSL
6873 implementations is desired (e.g. '-bugs' option to 's_client' and
6874 's_server'), so the new option is automatically set in many
6878 *) Changes in security patch:
6880 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6881 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6882 Air Force Materiel Command, USAF, under agreement number
6885 *) Add various sanity checks to asn1_get_length() to reject
6886 the ASN1 length bytes if they exceed sizeof(long), will appear
6887 negative or the content length exceeds the length of the
6888 supplied buffer. (CVE-2002-0659)
6889 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6891 *) Assertions for various potential buffer overflows, not known to
6893 [Ben Laurie (CHATS)]
6895 *) Various temporary buffers to hold ASCII versions of integers were
6896 too small for 64 bit platforms. (CVE-2002-0655)
6897 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6899 *) Remote buffer overflow in SSL3 protocol - an attacker could
6900 supply an oversized session ID to a client. (CVE-2002-0656)
6901 [Ben Laurie (CHATS)]
6903 *) Remote buffer overflow in SSL2 protocol - an attacker could
6904 supply an oversized client master key. (CVE-2002-0656)
6905 [Ben Laurie (CHATS)]
6907 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6909 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6910 encoded as NULL) with id-dsa-with-sha1.
6911 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6913 *) Check various X509_...() return values in apps/req.c.
6914 [Nils Larsch <nla@trustcenter.de>]
6916 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6917 an end-of-file condition would erronously be flagged, when the CRLF
6918 was just at the end of a processed block. The bug was discovered when
6919 processing data through a buffering memory BIO handing the data to a
6920 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6921 <ptsekov@syntrex.com> and Nedelcho Stanev.
6924 *) Implement a countermeasure against a vulnerability recently found
6925 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6926 before application data chunks to avoid the use of known IVs
6927 with data potentially chosen by the attacker.
6930 *) Fix length checks in ssl3_get_client_hello().
6933 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6934 to prevent ssl3_read_internal() from incorrectly assuming that
6935 ssl3_read_bytes() found application data while handshake
6936 processing was enabled when in fact s->s3->in_read_app_data was
6937 merely automatically cleared during the initial handshake.
6938 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6940 *) Fix object definitions for Private and Enterprise: they were not
6941 recognized in their shortname (=lowercase) representation. Extend
6942 obj_dat.pl to issue an error when using undefined keywords instead
6943 of silently ignoring the problem (Svenning Sorensen
6944 <sss@sss.dnsalias.net>).
6947 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6948 generators, i.e. generators other than 2 and 5. (Previously, the
6949 code did not properly initialise the 'add' and 'rem' values to
6950 BN_generate_prime().)
6952 In the new general case, we do not insist that 'generator' is
6953 actually a primitive root: This requirement is rather pointless;
6954 a generator of the order-q subgroup is just as good, if not
6958 *) Map new X509 verification errors to alerts. Discovered and submitted by
6959 Tom Wu <tom@arcot.com>.
6962 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6963 returning non-zero before the data has been completely received
6964 when using non-blocking I/O.
6965 [Bodo Moeller; problem pointed out by John Hughes]
6967 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6968 [Ben Laurie, Lutz Jaenicke]
6970 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6971 Yoram Zahavi <YoramZ@gilian.com>).
6974 *) Add information about CygWin 1.3 and on, and preserve proper
6975 configuration for the versions before that.
6976 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6978 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6979 check whether we deal with a copy of a session and do not delete from
6980 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6981 <izhar@checkpoint.com>.
6984 *) Do not store session data into the internal session cache, if it
6985 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6986 flag is set). Proposed by Aslam <aslam@funk.com>.
6989 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6993 *) [In 0.9.6d-engine release:]
6994 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6995 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6997 *) Add the configuration target linux-s390x.
6998 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7000 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7001 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7002 variable as an indication that a ClientHello message has been
7003 received. As the flag value will be lost between multiple
7004 invocations of ssl3_accept when using non-blocking I/O, the
7005 function may not be aware that a handshake has actually taken
7006 place, thus preventing a new session from being added to the
7009 To avoid this problem, we now set s->new_session to 2 instead of
7010 using a local variable.
7011 [Lutz Jaenicke, Bodo Moeller]
7013 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7014 if the SSL_R_LENGTH_MISMATCH error is detected.
7015 [Geoff Thorpe, Bodo Moeller]
7017 *) New 'shared_ldflag' column in Configure platform table.
7020 *) Fix EVP_CIPHER_mode macro.
7021 ["Dan S. Camper" <dan@bti.net>]
7023 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7024 type, we must throw them away by setting rr->length to 0.
7025 [D P Chang <dpc@qualys.com>]
7027 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7029 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7030 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7031 worked incorrectly for those cases where range = 10..._2 and
7032 3*range is two bits longer than range.)
7035 *) Only add signing time to PKCS7 structures if it is not already
7039 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7040 OBJ_ld_ce should be OBJ_id_ce.
7041 Also some ip-pda OIDs in crypto/objects/objects.txt were
7042 incorrect (cf. RFC 3039).
7043 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7045 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7046 returns early because it has nothing to do.
7047 [Andy Schneider <andy.schneider@bjss.co.uk>]
7049 *) [In 0.9.6c-engine release:]
7050 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7051 [Andy Schneider <andy.schneider@bjss.co.uk>]
7053 *) [In 0.9.6c-engine release:]
7054 Add support for Cryptographic Appliance's keyserver technology.
7055 (Use engine 'keyclient')
7056 [Cryptographic Appliances and Geoff Thorpe]
7058 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7059 is called via tools/c89.sh because arguments have to be
7060 rearranged (all '-L' options must appear before the first object
7062 [Richard Shapiro <rshapiro@abinitio.com>]
7064 *) [In 0.9.6c-engine release:]
7065 Add support for Broadcom crypto accelerator cards, backported
7067 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7069 *) [In 0.9.6c-engine release:]
7070 Add support for SureWare crypto accelerator cards from
7071 Baltimore Technologies. (Use engine 'sureware')
7072 [Baltimore Technologies and Mark Cox]
7074 *) [In 0.9.6c-engine release:]
7075 Add support for crypto accelerator cards from Accelerated
7076 Encryption Processing, www.aep.ie. (Use engine 'aep')
7077 [AEP Inc. and Mark Cox]
7079 *) Add a configuration entry for gcc on UnixWare.
7080 [Gary Benson <gbenson@redhat.com>]
7082 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7083 messages are stored in a single piece (fixed-length part and
7084 variable-length part combined) and fix various bugs found on the way.
7087 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7088 instead. BIO_gethostbyname() does not know what timeouts are
7089 appropriate, so entries would stay in cache even when they have
7091 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7093 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7094 faced with a pathologically small ClientHello fragment that does
7095 not contain client_version: Instead of aborting with an error,
7096 simply choose the highest available protocol version (i.e.,
7097 TLS 1.0 unless it is disabled). In practice, ClientHello
7098 messages are never sent like this, but this change gives us
7099 strictly correct behaviour at least for TLS.
7102 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7103 never resets s->method to s->ctx->method when called from within
7104 one of the SSL handshake functions.
7105 [Bodo Moeller; problem pointed out by Niko Baric]
7107 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7108 (sent using the client's version number) if client_version is
7109 smaller than the protocol version in use. Also change
7110 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7111 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7112 the client will at least see that alert.
7115 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7119 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7120 client receives HelloRequest while in a handshake.
7121 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7123 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7124 should end in 'break', not 'goto end' which circumvents various
7125 cleanups done in state SSL_ST_OK. But session related stuff
7126 must be disabled for SSL_ST_OK in the case that we just sent a
7129 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7130 before just sending a HelloRequest.
7131 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7133 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7134 reveal whether illegal block cipher padding was found or a MAC
7135 verification error occured. (Neither SSLerr() codes nor alerts
7136 are directly visible to potential attackers, but the information
7137 may leak via logfiles.)
7139 Similar changes are not required for the SSL 2.0 implementation
7140 because the number of padding bytes is sent in clear for SSL 2.0,
7141 and the extra bytes are just ignored. However ssl/s2_pkt.c
7142 failed to verify that the purported number of padding bytes is in
7146 *) Add OpenUNIX-8 support including shared libraries
7147 (Boyd Lynn Gerber <gerberb@zenez.com>).
7150 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7151 'wristwatch attack' using huge encoding parameters (cf.
7152 James H. Manger's CRYPTO 2001 paper). Note that the
7153 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7154 encoding parameters and hence was not vulnerable.
7157 *) BN_sqr() bug fix.
7158 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7160 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7161 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7162 followed by modular reduction.
7163 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7165 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7166 equivalent based on BN_pseudo_rand() instead of BN_rand().
7169 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7170 This function was broken, as the check for a new client hello message
7171 to handle SGC did not allow these large messages.
7172 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7175 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7178 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7179 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7182 *) Rework the configuration and shared library support for Tru64 Unix.
7183 The configuration part makes use of modern compiler features and
7184 still retains old compiler behavior for those that run older versions
7185 of the OS. The shared library support part includes a variant that
7186 uses the RPATH feature, and is available through the special
7187 configuration target "alpha-cc-rpath", which will never be selected
7189 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7191 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7192 with the same message size as in ssl3_get_certificate_request().
7193 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7194 messages might inadvertently be reject as too long.
7195 [Petr Lampa <lampa@fee.vutbr.cz>]
7197 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7200 *) Modified SSL library such that the verify_callback that has been set
7201 specificly for an SSL object with SSL_set_verify() is actually being
7202 used. Before the change, a verify_callback set with this function was
7203 ignored and the verify_callback() set in the SSL_CTX at the time of
7204 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7205 to allow the necessary settings.
7208 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7209 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7210 done automatically (in contradiction to the requirements of the C
7211 standard). This made problems when used from OpenSSH.
7214 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7215 dh->length and always used
7217 BN_rand_range(priv_key, dh->p).
7219 BN_rand_range() is not necessary for Diffie-Hellman, and this
7220 specific range makes Diffie-Hellman unnecessarily inefficient if
7221 dh->length (recommended exponent length) is much smaller than the
7222 length of dh->p. We could use BN_rand_range() if the order of
7223 the subgroup was stored in the DH structure, but we only have
7228 BN_rand(priv_key, l, ...)
7230 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7236 RSA_eay_public_encrypt
7237 RSA_eay_private_decrypt
7238 RSA_eay_private_encrypt (signing)
7239 RSA_eay_public_decrypt (signature verification)
7241 (default implementations for RSA_public_encrypt,
7242 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7243 always reject numbers >= n.
7246 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7247 to synchronize access to 'locking_thread'. This is necessary on
7248 systems where access to 'locking_thread' (an 'unsigned long'
7249 variable) is not atomic.
7252 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7253 *before* setting the 'crypto_lock_rand' flag. The previous code had
7254 a race condition if 0 is a valid thread ID.
7255 [Travis Vitek <vitek@roguewave.com>]
7257 *) Add support for shared libraries under Irix.
7258 [Albert Chin-A-Young <china@thewrittenword.com>]
7260 *) Add configuration option to build on Linux on both big-endian and
7262 [Ralf Baechle <ralf@uni-koblenz.de>]
7264 *) Add the possibility to create shared libraries on HP-UX.
7267 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7269 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7270 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7271 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7272 PRNG state recovery was possible based on the output of
7273 one PRNG request appropriately sized to gain knowledge on
7274 'md' followed by enough consecutive 1-byte PRNG requests
7275 to traverse all of 'state'.
7277 1. When updating 'md_local' (the current thread's copy of 'md')
7278 during PRNG output generation, hash all of the previous
7279 'md_local' value, not just the half used for PRNG output.
7281 2. Make the number of bytes from 'state' included into the hash
7282 independent from the number of PRNG bytes requested.
7284 The first measure alone would be sufficient to avoid
7285 Markku-Juhani's attack. (Actually it had never occurred
7286 to me that the half of 'md_local' used for chaining was the
7287 half from which PRNG output bytes were taken -- I had always
7288 assumed that the secret half would be used.) The second
7289 measure makes sure that additional data from 'state' is never
7290 mixed into 'md_local' in small portions; this heuristically
7291 further strengthens the PRNG.
7294 *) Fix crypto/bn/asm/mips3.s.
7297 *) When only the key is given to "enc", the IV is undefined. Print out
7298 an error message in this case.
7301 *) Handle special case when X509_NAME is empty in X509 printing routines.
7304 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7305 positive and less than q.
7308 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7309 used: it isn't thread safe and the add_lock_callback should handle
7311 [Paul Rose <Paul.Rose@bridge.com>]
7313 *) Verify that incoming data obeys the block size in
7314 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7318 [Ulf Möller, Bodo Möller]
7320 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7321 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7322 when fixing the server behaviour for backwards-compatible 'client
7323 hello' messages. (Note that the attack is impractical against
7324 SSL 3.0 and TLS 1.0 anyway because length and version checking
7325 means that the probability of guessing a valid ciphertext is
7326 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7329 Before 0.9.5, the countermeasure (hide the error by generating a
7330 random 'decryption result') did not work properly because
7331 ERR_clear_error() was missing, meaning that SSL_get_error() would
7332 detect the supposedly ignored error.
7334 Both problems are now fixed.
7337 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7338 (previously it was 1024).
7341 *) Fix for compatibility mode trust settings: ignore trust settings
7342 unless some valid trust or reject settings are present.
7345 *) Fix for blowfish EVP: its a variable length cipher.
7348 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7349 parameters in DSA public key structures and return an error in the
7350 DSA routines if parameters are absent.
7353 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7354 in the current directory if neither $RANDFILE nor $HOME was set.
7355 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7356 caused some confusion to Windows users who haven't defined $HOME.
7357 Thus RAND_file_name() is changed again: e_os.h can define a
7358 DEFAULT_HOME, which will be used if $HOME is not set.
7359 For Windows, we use "C:"; on other platforms, we still require
7360 environment variables.
7362 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7363 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7364 having multiple threads call RAND_poll() concurrently.
7367 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7368 combination of a flag and a thread ID variable.
7369 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7370 flag), *other* threads can enter ssleay_add_bytes without obeying
7371 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7372 that they do not hold after the first thread unsets add_do_not_lock).
7375 *) Change bctest again: '-x' expressions are not available in all
7379 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7381 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7382 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7384 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7385 the default extension for executables, if any. Also, make the perl
7386 scripts that use symlink() to test if it really exists and use "cp"
7387 if it doesn't. All this made OpenSSL compilable and installable in
7391 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7392 If SEQUENCE is length is indefinite just set c->slen to the total
7393 amount of data available.
7394 [Steve Henson, reported by shige@FreeBSD.org]
7395 [This change does not apply to 0.9.7.]
7397 *) Change bctest to avoid here-documents inside command substitution
7398 (workaround for FreeBSD /bin/sh bug).
7399 For compatibility with Ultrix, avoid shell functions (introduced
7400 in the bctest version that searches along $PATH).
7403 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7404 with des_encrypt() defined on some operating systems, like Solaris
7408 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7409 On the Importance of Eliminating Errors in Cryptographic
7410 Computations, J. Cryptology 14 (2001) 2, 101-119,
7411 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7414 *) MIPS assembler BIGNUM division bug fix.
7417 *) Disabled incorrect Alpha assembler code.
7420 *) Fix PKCS#7 decode routines so they correctly update the length
7421 after reading an EOC for the EXPLICIT tag.
7423 [This change does not apply to 0.9.7.]
7425 *) Fix bug in PKCS#12 key generation routines. This was triggered
7426 if a 3DES key was generated with a 0 initial byte. Include
7427 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7428 (but broken) behaviour.
7431 *) Enhance bctest to search for a working bc along $PATH and print
7433 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7435 *) Fix memory leaks in err.c: free err_data string if necessary;
7436 don't write to the wrong index in ERR_set_error_data.
7439 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7443 *) Replace rdtsc with _emit statements for VC++ version 5.
7444 [Jeremy Cooper <jeremy@baymoo.org>]
7446 *) Make it possible to reuse SSLv2 sessions.
7449 *) In copy_email() check for >= 0 as a return value for
7450 X509_NAME_get_index_by_NID() since 0 is a valid index.
7451 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7453 *) Avoid coredump with unsupported or invalid public keys by checking if
7454 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7455 PKCS7_verify() fails with non detached data.
7458 *) Don't use getenv in library functions when run as setuid/setgid.
7459 New function OPENSSL_issetugid().
7462 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7463 due to incorrect handling of multi-threading:
7465 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7467 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7469 3. Count how many times MemCheck_off() has been called so that
7470 nested use can be treated correctly. This also avoids
7471 inband-signalling in the previous code (which relied on the
7472 assumption that thread ID 0 is impossible).
7475 *) Add "-rand" option also to s_client and s_server.
7478 *) Fix CPU detection on Irix 6.x.
7479 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7480 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7482 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7485 [This change does not apply to 0.9.7.]
7487 *) Use the cached encoding of an X509_NAME structure rather than
7488 copying it. This is apparently the reason for the libsafe "errors"
7489 but the code is actually correct.
7492 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7493 Bleichenbacher's DSA attack.
7494 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7495 to be set and top=0 forces the highest bit to be set; top=-1 is new
7496 and leaves the highest bit random.
7497 [Ulf Moeller, Bodo Moeller]
7499 *) In the NCONF_...-based implementations for CONF_... queries
7500 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7501 a temporary CONF structure with the data component set to NULL
7502 (which gives segmentation faults in lh_retrieve).
7503 Instead, use NULL for the CONF pointer in CONF_get_string and
7504 CONF_get_number (which may use environment variables) and directly
7505 return NULL from CONF_get_section.
7508 *) Fix potential buffer overrun for EBCDIC.
7511 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7512 keyUsage if basicConstraints absent for a CA.
7515 *) Make SMIME_write_PKCS7() write mail header values with a format that
7516 is more generally accepted (no spaces before the semicolon), since
7517 some programs can't parse those values properly otherwise. Also make
7518 sure BIO's that break lines after each write do not create invalid
7522 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7523 macros previously used would not encode an empty SEQUENCE OF
7524 and break the signature.
7526 [This change does not apply to 0.9.7.]
7528 *) Zero the premaster secret after deriving the master secret in
7532 *) Add some EVP_add_digest_alias registrations (as found in
7533 OpenSSL_add_all_digests()) to SSL_library_init()
7534 aka OpenSSL_add_ssl_algorithms(). This provides improved
7535 compatibility with peers using X.509 certificates
7536 with unconventional AlgorithmIdentifier OIDs.
7539 *) Fix for Irix with NO_ASM.
7540 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7542 *) ./config script fixes.
7543 [Ulf Moeller, Richard Levitte]
7545 *) Fix 'openssl passwd -1'.
7548 *) Change PKCS12_key_gen_asc() so it can cope with non null
7549 terminated strings whose length is passed in the passlen
7550 parameter, for example from PEM callbacks. This was done
7551 by adding an extra length parameter to asc2uni().
7552 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7554 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7555 call failed, free the DSA structure.
7558 *) Fix to uni2asc() to cope with zero length Unicode strings.
7559 These are present in some PKCS#12 files.
7562 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7563 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7564 when writing a 32767 byte record.
7565 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7567 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7568 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7570 (RSA objects have a reference count access to which is protected
7571 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7572 so they are meant to be shared between threads.)
7573 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7574 "Reddie, Steven" <Steven.Reddie@ca.com>]
7576 *) Fix a deadlock in CRYPTO_mem_leaks().
7579 *) Use better test patterns in bntest.
7582 *) rand_win.c fix for Borland C.
7585 *) BN_rshift bugfix for n == 0.
7588 *) Add a 'bctest' script that checks for some known 'bc' bugs
7589 so that 'make test' does not abort just because 'bc' is broken.
7592 *) Store verify_result within SSL_SESSION also for client side to
7593 avoid potential security hole. (Re-used sessions on the client side
7594 always resulted in verify_result==X509_V_OK, not using the original
7595 result of the server certificate verification.)
7598 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7599 SSL3_RT_APPLICATION_DATA, return 0.
7600 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7604 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7605 releases, have been re-implemented by renaming the previous
7606 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7607 and ssl3_read_internal, respectively, and adding 'peek' parameters
7608 to them. The new ssl[23]_{read,peek} functions are calls to
7609 ssl[23]_read_internal with the 'peek' flag set appropriately.
7610 A 'peek' parameter has also been added to ssl3_read_bytes, which
7611 does the actual work for ssl3_read_internal.
7614 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7615 the method-specific "init()" handler. Also clean up ex_data after
7616 calling the method-specific "finish()" handler. Previously, this was
7617 happening the other way round.
7620 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7621 The previous value, 12, was not always sufficient for BN_mod_exp().
7624 *) Make sure that shared libraries get the internal name engine with
7625 the full version number and not just 0. This should mark the
7626 shared libraries as not backward compatible. Of course, this should
7627 be changed again when we can guarantee backward binary compatibility.
7630 *) Fix typo in get_cert_by_subject() in by_dir.c
7631 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7633 *) Rework the system to generate shared libraries:
7635 - Make note of the expected extension for the shared libraries and
7636 if there is a need for symbolic links from for example libcrypto.so.0
7637 to libcrypto.so.0.9.7. There is extended info in Configure for
7640 - Make as few rebuilds of the shared libraries as possible.
7642 - Still avoid linking the OpenSSL programs with the shared libraries.
7644 - When installing, install the shared libraries separately from the
7648 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7650 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7651 and not in SSL_clear because the latter is also used by the
7652 accept/connect functions; previously, the settings made by
7653 SSL_set_read_ahead would be lost during the handshake.
7654 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7656 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7657 Previously, it would create entries for disableed algorithms no
7661 *) Added several new manual pages for SSL_* function.
7664 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7666 *) In ssl23_get_client_hello, generate an error message when faced
7667 with an initial SSL 3.0/TLS record that is too small to contain the
7668 first two bytes of the ClientHello message, i.e. client_version.
7669 (Note that this is a pathologic case that probably has never happened
7670 in real life.) The previous approach was to use the version number
7671 from the record header as a substitute; but our protocol choice
7672 should not depend on that one because it is not authenticated
7673 by the Finished messages.
7676 *) More robust randomness gathering functions for Windows.
7677 [Jeffrey Altman <jaltman@columbia.edu>]
7679 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7680 not set then we don't setup the error code for issuer check errors
7681 to avoid possibly overwriting other errors which the callback does
7682 handle. If an application does set the flag then we assume it knows
7683 what it is doing and can handle the new informational codes
7687 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7688 a general "ANY" type, as such it should be able to decode anything
7689 including tagged types. However it didn't check the class so it would
7690 wrongly interpret tagged types in the same way as their universal
7691 counterpart and unknown types were just rejected. Changed so that the
7692 tagged and unknown types are handled in the same way as a SEQUENCE:
7693 that is the encoding is stored intact. There is also a new type
7694 "V_ASN1_OTHER" which is used when the class is not universal, in this
7695 case we have no idea what the actual type is so we just lump them all
7699 *) On VMS, stdout may very well lead to a file that is written to
7700 in a record-oriented fashion. That means that every write() will
7701 write a separate record, which will be read separately by the
7702 programs trying to read from it. This can be very confusing.
7704 The solution is to put a BIO filter in the way that will buffer
7705 text until a linefeed is reached, and then write everything a
7706 line at a time, so every record written will be an actual line,
7707 not chunks of lines and not (usually doesn't happen, but I've
7708 seen it once) several lines in one record. BIO_f_linebuffer() is
7711 Currently, it's a VMS-only method, because that's where it has
7712 been tested well enough.
7715 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7716 it can return incorrect results.
7717 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7718 but it was in 0.9.6-beta[12].)
7721 *) Disable the check for content being present when verifying detached
7722 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7723 include zero length content when signing messages.
7726 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7727 BIO_ctrl (for BIO pairs).
7730 *) Add DSO method for VMS.
7733 *) Bug fix: Montgomery multiplication could produce results with the
7737 *) Add RPM specification openssl.spec and modify it to build three
7738 packages. The default package contains applications, application
7739 documentation and run-time libraries. The devel package contains
7740 include files, static libraries and function documentation. The
7741 doc package contains the contents of the doc directory. The original
7742 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7745 *) Add a large number of documentation files for many SSL routines.
7746 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7748 *) Add a configuration entry for Sony News 4.
7749 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7751 *) Don't set the two most significant bits to one when generating a
7752 random number < q in the DSA library.
7755 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7756 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7757 the underlying transport is blocking) if a handshake took place.
7758 (The default behaviour is needed by applications such as s_client
7759 and s_server that use select() to determine when to use SSL_read;
7760 but for applications that know in advance when to expect data, it
7761 just makes things more complicated.)
7764 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7768 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7769 work better on such systems.
7770 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7772 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7773 Update PKCS12_parse() so it copies the friendlyName and the
7774 keyid to the certificates aux info.
7777 *) Fix bug in PKCS7_verify() which caused an infinite loop
7778 if there was more than one signature.
7779 [Sven Uszpelkat <su@celocom.de>]
7781 *) Major change in util/mkdef.pl to include extra information
7782 about each symbol, as well as presenting variables as well
7783 as functions. This change means that there's n more need
7784 to rebuild the .num files when some algorithms are excluded.
7787 *) Allow the verify time to be set by an application,
7788 rather than always using the current time.
7791 *) Phase 2 verify code reorganisation. The certificate
7792 verify code now looks up an issuer certificate by a
7793 number of criteria: subject name, authority key id
7794 and key usage. It also verifies self signed certificates
7795 by the same criteria. The main comparison function is
7796 X509_check_issued() which performs these checks.
7798 Lot of changes were necessary in order to support this
7799 without completely rewriting the lookup code.
7801 Authority and subject key identifier are now cached.
7803 The LHASH 'certs' is X509_STORE has now been replaced
7804 by a STACK_OF(X509_OBJECT). This is mainly because an
7805 LHASH can't store or retrieve multiple objects with
7806 the same hash value.
7808 As a result various functions (which were all internal
7809 use only) have changed to handle the new X509_STORE
7810 structure. This will break anything that messed round
7811 with X509_STORE internally.
7813 The functions X509_STORE_add_cert() now checks for an
7814 exact match, rather than just subject name.
7816 The X509_STORE API doesn't directly support the retrieval
7817 of multiple certificates matching a given criteria, however
7818 this can be worked round by performing a lookup first
7819 (which will fill the cache with candidate certificates)
7820 and then examining the cache for matches. This is probably
7821 the best we can do without throwing out X509_LOOKUP
7822 entirely (maybe later...).
7824 The X509_VERIFY_CTX structure has been enhanced considerably.
7826 All certificate lookup operations now go via a get_issuer()
7827 callback. Although this currently uses an X509_STORE it
7828 can be replaced by custom lookups. This is a simple way
7829 to bypass the X509_STORE hackery necessary to make this
7830 work and makes it possible to use more efficient techniques
7831 in future. A very simple version which uses a simple
7832 STACK for its trusted certificate store is also provided
7833 using X509_STORE_CTX_trusted_stack().
7835 The verify_cb() and verify() callbacks now have equivalents
7836 in the X509_STORE_CTX structure.
7838 X509_STORE_CTX also has a 'flags' field which can be used
7839 to customise the verify behaviour.
7842 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7843 excludes S/MIME capabilities.
7846 *) When a certificate request is read in keep a copy of the
7847 original encoding of the signed data and use it when outputing
7848 again. Signatures then use the original encoding rather than
7849 a decoded, encoded version which may cause problems if the
7850 request is improperly encoded.
7853 *) For consistency with other BIO_puts implementations, call
7854 buffer_write(b, ...) directly in buffer_puts instead of calling
7857 In BIO_puts, increment b->num_write as in BIO_write.
7858 [Peter.Sylvester@EdelWeb.fr]
7860 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7861 BN_zero, we may not return a BIGNUM with an array consisting of
7865 *) Avoid calling abort() from within the library when problems are
7866 detected, except if preprocessor symbols have been defined
7867 (such as REF_CHECK, BN_DEBUG etc.).
7870 *) New openssl application 'rsautl'. This utility can be
7871 used for low level RSA operations. DER public key
7872 BIO/fp routines also added.
7875 *) New Configure entry and patches for compiling on QNX 4.
7876 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7878 *) A demo state-machine implementation was sponsored by
7879 Nuron (http://www.nuron.com/) and is now available in
7880 demos/state_machine.
7883 *) New options added to the 'dgst' utility for signature
7884 generation and verification.
7887 *) Unrecognized PKCS#7 content types are now handled via a
7888 catch all ASN1_TYPE structure. This allows unsupported
7889 types to be stored as a "blob" and an application can
7890 encode and decode it manually.
7893 *) Fix various signed/unsigned issues to make a_strex.c
7895 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7897 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7898 length if passed a buffer. ASN1_INTEGER_to_BN failed
7899 if passed a NULL BN and its argument was negative.
7900 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7902 *) Modification to PKCS#7 encoding routines to output definite
7903 length encoding. Since currently the whole structures are in
7904 memory there's not real point in using indefinite length
7905 constructed encoding. However if OpenSSL is compiled with
7906 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7909 *) Added BIO_vprintf() and BIO_vsnprintf().
7912 *) Added more prefixes to parse for in the the strings written
7913 through a logging bio, to cover all the levels that are available
7914 through syslog. The prefixes are now:
7916 PANIC, EMERG, EMR => LOG_EMERG
7917 ALERT, ALR => LOG_ALERT
7918 CRIT, CRI => LOG_CRIT
7919 ERROR, ERR => LOG_ERR
7920 WARNING, WARN, WAR => LOG_WARNING
7921 NOTICE, NOTE, NOT => LOG_NOTICE
7922 INFO, INF => LOG_INFO
7923 DEBUG, DBG => LOG_DEBUG
7925 and as before, if none of those prefixes are present at the
7926 beginning of the string, LOG_ERR is chosen.
7928 On Win32, the LOG_* levels are mapped according to this:
7930 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7931 LOG_WARNING => EVENTLOG_WARNING_TYPE
7932 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7936 *) Made it possible to reconfigure with just the configuration
7937 argument "reconf" or "reconfigure". The command line arguments
7938 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7939 and are retrieved from there when reconfiguring.
7943 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7945 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7948 *) The obj_dat.pl script was messing up the sorting of object
7949 names. The reason was that it compared the quoted version
7950 of strings as a result "OCSP" > "OCSP Signing" because
7951 " > SPACE. Changed script to store unquoted versions of
7952 names and add quotes on output. It was also omitting some
7953 names from the lookup table if they were given a default
7954 value (that is if SN is missing it is given the same
7955 value as LN and vice versa), these are now added on the
7956 grounds that if an object has a name we should be able to
7957 look it up. Finally added warning output when duplicate
7958 short or long names are found.
7961 *) Changes needed for Tandem NSK.
7962 [Scott Uroff <scott@xypro.com>]
7964 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7965 RSA_padding_check_SSLv23(), special padding was never detected
7966 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7967 version rollback attacks was not effective.
7969 In s23_clnt.c, don't use special rollback-attack detection padding
7970 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7971 client; similarly, in s23_srvr.c, don't do the rollback check if
7972 SSL 2.0 is the only protocol enabled in the server.
7975 *) Make it possible to get hexdumps of unprintable data with 'openssl
7976 asn1parse'. By implication, the functions ASN1_parse_dump() and
7977 BIO_dump_indent() are added.
7980 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7981 these print out strings and name structures based on various
7982 flags including RFC2253 support and proper handling of
7983 multibyte characters. Added options to the 'x509' utility
7984 to allow the various flags to be set.
7987 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7988 Also change the functions X509_cmp_current_time() and
7989 X509_gmtime_adj() work with an ASN1_TIME structure,
7990 this will enable certificates using GeneralizedTime in validity
7991 dates to be checked.
7994 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7995 negative public key encodings) on by default,
7996 NO_NEG_PUBKEY_BUG can be set to disable it.
7999 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8000 content octets. An i2c_ASN1_OBJECT is unnecessary because
8001 the encoding can be trivially obtained from the structure.
8004 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8005 not read locks (CRYPTO_r_[un]lock).
8008 *) A first attempt at creating official support for shared
8009 libraries through configuration. I've kept it so the
8010 default is static libraries only, and the OpenSSL programs
8011 are always statically linked for now, but there are
8012 preparations for dynamic linking in place.
8013 This has been tested on Linux and Tru64.
8016 *) Randomness polling function for Win9x, as described in:
8017 Peter Gutmann, Software Generation of Practically Strong
8021 *) Fix so PRNG is seeded in req if using an already existing
8025 *) New options to smime application. -inform and -outform
8026 allow alternative formats for the S/MIME message including
8027 PEM and DER. The -content option allows the content to be
8028 specified separately. This should allow things like Netscape
8029 form signing output easier to verify.
8032 *) Fix the ASN1 encoding of tags using the 'long form'.
8035 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8036 STRING types. These convert content octets to and from the
8037 underlying type. The actual tag and length octets are
8038 already assumed to have been read in and checked. These
8039 are needed because all other string types have virtually
8040 identical handling apart from the tag. By having versions
8041 of the ASN1 functions that just operate on content octets
8042 IMPLICIT tagging can be handled properly. It also allows
8043 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8044 and ASN1_INTEGER are identical apart from the tag.
8047 *) Change the handling of OID objects as follows:
8049 - New object identifiers are inserted in objects.txt, following
8050 the syntax given in objects.README.
8051 - objects.pl is used to process obj_mac.num and create a new
8053 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8056 This is currently kind of a hack, and the perl code in objects.pl
8057 isn't very elegant, but it works as I intended. The simplest way
8058 to check that it worked correctly is to look in obj_dat.h and
8059 check the array nid_objs and make sure the objects haven't moved
8060 around (this is important!). Additions are OK, as well as
8061 consistent name changes.
8064 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8067 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8068 The given file adds to whatever has already been seeded into the
8069 random pool through the RANDFILE configuration file option or
8070 environment variable, or the default random state file.
8073 *) mkstack.pl now sorts each macro group into lexical order.
8074 Previously the output order depended on the order the files
8075 appeared in the directory, resulting in needless rewriting
8079 *) Patches to make OpenSSL compile under Win32 again. Mostly
8080 work arounds for the VC++ problem that it treats func() as
8081 func(void). Also stripped out the parts of mkdef.pl that
8082 added extra typesafe functions: these no longer exist.
8085 *) Reorganisation of the stack code. The macros are now all
8086 collected in safestack.h . Each macro is defined in terms of
8087 a "stack macro" of the form SKM_<name>(type, a, b). The
8088 DEBUG_SAFESTACK is now handled in terms of function casts,
8089 this has the advantage of retaining type safety without the
8090 use of additional functions. If DEBUG_SAFESTACK is not defined
8091 then the non typesafe macros are used instead. Also modified the
8092 mkstack.pl script to handle the new form. Needs testing to see
8093 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8094 the default if no major problems. Similar behaviour for ASN1_SET_OF
8095 and PKCS12_STACK_OF.
8098 *) When some versions of IIS use the 'NET' form of private key the
8099 key derivation algorithm is different. Normally MD5(password) is
8100 used as a 128 bit RC4 key. In the modified case
8101 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8102 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8103 as the old Netscape_RSA functions except they have an additional
8104 'sgckey' parameter which uses the modified algorithm. Also added
8105 an -sgckey command line option to the rsa utility. Thanks to
8106 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8107 algorithm to openssl-dev.
8110 *) The evp_local.h macros were using 'c.##kname' which resulted in
8111 invalid expansion on some systems (SCO 5.0.5 for example).
8112 Corrected to 'c.kname'.
8113 [Phillip Porch <root@theporch.com>]
8115 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8116 a STACK of email addresses from a certificate or request, these look
8117 in the subject name and the subject alternative name extensions and
8118 omit any duplicate addresses.
8121 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8122 This makes DSA verification about 2 % faster.
8125 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8126 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8127 plus overhead for 1024 bit moduli).
8128 This makes exponentiations about 0.5 % faster for 1024 bit
8129 exponents (as measured by "openssl speed rsa2048").
8132 *) Rename memory handling macros to avoid conflicts with other
8134 Malloc => OPENSSL_malloc
8135 Malloc_locked => OPENSSL_malloc_locked
8136 Realloc => OPENSSL_realloc
8137 Free => OPENSSL_free
8140 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8141 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8144 *) CygWin32 support.
8145 [John Jarvie <jjarvie@newsguy.com>]
8147 *) The type-safe stack code has been rejigged. It is now only compiled
8148 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8149 by default all type-specific stack functions are "#define"d back to
8150 standard stack functions. This results in more streamlined output
8151 but retains the type-safety checking possibilities of the original
8155 *) The STACK code has been cleaned up, and certain type declarations
8156 that didn't make a lot of sense have been brought in line. This has
8157 also involved a cleanup of sorts in safestack.h to more correctly
8158 map type-safe stack functions onto their plain stack counterparts.
8159 This work has also resulted in a variety of "const"ifications of
8160 lots of the code, especially "_cmp" operations which should normally
8161 be prototyped with "const" parameters anyway.
8164 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8165 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8166 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8167 where all of 'md' is used each time the PRNG is used, but 'state'
8168 is used only indexed by a cyclic counter. As entropy may not be
8169 well distributed from the beginning, 'md' is important as a
8170 chaining variable. However, the output function chains only half
8171 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8172 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8173 in all of 'state' being rewritten, with the new values depending
8174 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8177 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8178 the handshake is continued after ssl_verify_cert_chain();
8179 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8180 can lead to 'unexplainable' connection aborts later.
8181 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8183 *) Major EVP API cipher revision.
8184 Add hooks for extra EVP features. This allows various cipher
8185 parameters to be set in the EVP interface. Support added for variable
8186 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8187 setting of RC2 and RC5 parameters.
8189 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8192 Remove lots of duplicated code from the EVP library. For example *every*
8193 cipher init() function handles the 'iv' in the same way according to the
8194 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8195 for CFB and OFB modes they zero ctx->num.
8197 New functionality allows removal of S/MIME code RC2 hack.
8199 Most of the routines have the same form and so can be declared in terms
8202 By shifting this to the top level EVP_CipherInit() it can be removed from
8203 all individual ciphers. If the cipher wants to handle IVs or keys
8204 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8207 Change lots of functions like EVP_EncryptUpdate() to now return a
8208 value: although software versions of the algorithms cannot fail
8209 any installed hardware versions can.
8212 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8213 this option is set, tolerate broken clients that send the negotiated
8214 protocol version number instead of the requested protocol version
8218 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8219 i.e. non-zero for export ciphersuites, zero otherwise.
8220 Previous versions had this flag inverted, inconsistent with
8221 rsa_tmp_cb (..._TMP_RSA_CB).
8222 [Bodo Moeller; problem reported by Amit Chopra]
8224 *) Add missing DSA library text string. Work around for some IIS
8225 key files with invalid SEQUENCE encoding.
8228 *) Add a document (doc/standards.txt) that list all kinds of standards
8229 and so on that are implemented in OpenSSL.
8232 *) Enhance c_rehash script. Old version would mishandle certificates
8233 with the same subject name hash and wouldn't handle CRLs at all.
8234 Added -fingerprint option to crl utility, to support new c_rehash
8238 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8241 *) Fix for SSL server purpose checking. Server checking was
8242 rejecting certificates which had extended key usage present
8243 but no ssl client purpose.
8244 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8246 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8247 is a little unclear about how a blank password is handled.
8248 Since the password in encoded as a BMPString with terminating
8249 double NULL a zero length password would end up as just the
8250 double NULL. However no password at all is different and is
8251 handled differently in the PKCS#12 key generation code. NS
8252 treats a blank password as zero length. MSIE treats it as no
8253 password on export: but it will try both on import. We now do
8254 the same: PKCS12_parse() tries zero length and no password if
8255 the password is set to "" or NULL (NULL is now a valid password:
8256 it wasn't before) as does the pkcs12 application.
8259 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8260 perror when PEM_read_bio_X509_REQ fails, the error message must
8261 be obtained from the error queue.
8264 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8265 it in ERR_remove_state if appropriate, and change ERR_get_state
8266 accordingly to avoid race conditions (this is necessary because
8267 thread_hash is no longer constant once set).
8270 *) Bugfix for linux-elf makefile.one.
8273 *) RSA_get_default_method() will now cause a default
8274 RSA_METHOD to be chosen if one doesn't exist already.
8275 Previously this was only set during a call to RSA_new()
8276 or RSA_new_method(NULL) meaning it was possible for
8277 RSA_get_default_method() to return NULL.
8280 *) Added native name translation to the existing DSO code
8281 that will convert (if the flag to do so is set) filenames
8282 that are sufficiently small and have no path information
8283 into a canonical native form. Eg. "blah" converted to
8284 "libblah.so" or "blah.dll" etc.
8287 *) New function ERR_error_string_n(e, buf, len) which is like
8288 ERR_error_string(e, buf), but writes at most 'len' bytes
8289 including the 0 terminator. For ERR_error_string_n, 'buf'
8291 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8293 *) CONF library reworked to become more general. A new CONF
8294 configuration file reader "class" is implemented as well as a
8295 new functions (NCONF_*, for "New CONF") to handle it. The now
8296 old CONF_* functions are still there, but are reimplemented to
8297 work in terms of the new functions. Also, a set of functions
8298 to handle the internal storage of the configuration data is
8299 provided to make it easier to write new configuration file
8300 reader "classes" (I can definitely see something reading a
8301 configuration file in XML format, for example), called _CONF_*,
8302 or "the configuration storage API"...
8304 The new configuration file reading functions are:
8306 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8307 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8309 NCONF_default, NCONF_WIN32
8311 NCONF_dump_fp, NCONF_dump_bio
8313 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8314 NCONF_new creates a new CONF object. This works in the same way
8315 as other interfaces in OpenSSL, like the BIO interface.
8316 NCONF_dump_* dump the internal storage of the configuration file,
8317 which is useful for debugging. All other functions take the same
8318 arguments as the old CONF_* functions wth the exception of the
8319 first that must be a `CONF *' instead of a `LHASH *'.
8321 To make it easer to use the new classes with the old CONF_* functions,
8322 the function CONF_set_default_method is provided.
8325 *) Add '-tls1' option to 'openssl ciphers', which was already
8326 mentioned in the documentation but had not been implemented.
8327 (This option is not yet really useful because even the additional
8328 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8331 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8332 OpenSSL-based applications) load shared libraries and bind to
8333 them in a portable way.
8334 [Geoff Thorpe, with contributions from Richard Levitte]
8336 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8338 *) Make sure _lrotl and _lrotr are only used with MSVC.
8340 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8341 (the default implementation of RAND_status).
8343 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8344 to '-clrext' (= clear extensions), as intended and documented.
8345 [Bodo Moeller; inconsistency pointed out by Michael Attili
8348 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8349 was larger than the MD block size.
8350 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8352 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8353 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8354 using the passed key: if the passed key was a private key the result
8355 of X509_print(), for example, would be to print out all the private key
8359 *) des_quad_cksum() byte order bug fix.
8360 [Ulf Möller, using the problem description in krb4-0.9.7, where
8361 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8363 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8365 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8367 *) For easily testing in shell scripts whether some command
8368 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8369 returns with exit code 0 iff no command of the given name is available.
8370 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8371 the output goes to stdout and nothing is printed to stderr.
8372 Additional arguments are always ignored.
8374 Since for each cipher there is a command of the same name,
8375 the 'no-cipher' compilation switches can be tested this way.
8377 ('openssl no-XXX' is not able to detect pseudo-commands such
8378 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8381 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8384 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8385 is set; it will be thrown away anyway because each handshake creates
8387 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8388 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8389 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8390 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8393 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8394 'Q' and 'R' lose their special meanings (quit/renegotiate).
8395 This is part of what -quiet does; unlike -quiet, -ign_eof
8396 does not suppress any output.
8399 *) Add compatibility options to the purpose and trust code. The
8400 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8401 accepts a certificate or CA, this was the previous behaviour,
8402 with all the associated security issues.
8404 X509_TRUST_COMPAT is the old trust behaviour: only and
8405 automatically trust self signed roots in certificate store. A
8406 new trust setting X509_TRUST_DEFAULT is used to specify that
8407 a purpose has no associated trust setting and it should instead
8408 use the value in the default purpose.
8411 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8412 and fix a memory leak.
8415 *) In util/mkerr.pl (which implements 'make errors'), preserve
8416 reason strings from the previous version of the .c file, as
8417 the default to have only downcase letters (and digits) in
8418 automatically generated reasons codes is not always appropriate.
8421 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8422 using strerror. Previously, ERR_reason_error_string() returned
8423 library names as reason strings for SYSerr; but SYSerr is a special
8424 case where small numbers are errno values, not library numbers.
8427 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8428 converts DSA parameters into DH parameters. (When creating parameters,
8429 DSA_generate_parameters is used.)
8432 *) Include 'length' (recommended exponent length) in C code generated
8433 by 'openssl dhparam -C'.
8436 *) The second argument to set_label in perlasm was already being used
8437 so couldn't be used as a "file scope" flag. Moved to third argument
8441 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8442 instead of RAND_bytes for encryption IVs and salts.
8445 *) Include RAND_status() into RAND_METHOD instead of implementing
8446 it only for md_rand.c Otherwise replacing the PRNG by calling
8447 RAND_set_rand_method would be impossible.
8450 *) Don't let DSA_generate_key() enter an infinite loop if the random
8451 number generation fails.
8454 *) New 'rand' application for creating pseudo-random output.
8457 *) Added configuration support for Linux/IA64
8458 [Rolf Haberrecker <rolf@suse.de>]
8460 *) Assembler module support for Mingw32.
8463 *) Shared library support for HPUX (in shlib/).
8464 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8466 *) Shared library support for Solaris gcc.
8467 [Lutz Behnke <behnke@trustcenter.de>]
8469 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8471 *) PKCS7_encrypt() was adding text MIME headers twice because they
8472 were added manually and by SMIME_crlf_copy().
8475 *) In bntest.c don't call BN_rand with zero bits argument.
8476 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8478 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8479 case was implemented. This caused BN_div_recp() to fail occasionally.
8482 *) Add an optional second argument to the set_label() in the perl
8483 assembly language builder. If this argument exists and is set
8484 to 1 it signals that the assembler should use a symbol whose
8485 scope is the entire file, not just the current function. This
8486 is needed with MASM which uses the format label:: for this scope.
8487 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8489 *) Change the ASN1 types so they are typedefs by default. Before
8490 almost all types were #define'd to ASN1_STRING which was causing
8491 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8495 *) Change names of new functions to the new get1/get0 naming
8496 convention: After 'get1', the caller owns a reference count
8497 and has to call ..._free; 'get0' returns a pointer to some
8498 data structure without incrementing reference counters.
8499 (Some of the existing 'get' functions increment a reference
8500 counter, some don't.)
8501 Similarly, 'set1' and 'add1' functions increase reference
8502 counters or duplicate objects.
8505 *) Allow for the possibility of temp RSA key generation failure:
8506 the code used to assume it always worked and crashed on failure.
8509 *) Fix potential buffer overrun problem in BIO_printf().
8510 [Ulf Möller, using public domain code by Patrick Powell; problem
8511 pointed out by David Sacerdote <das33@cornell.edu>]
8513 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8514 RAND_egd() and RAND_status(). In the command line application,
8515 the EGD socket can be specified like a seed file using RANDFILE
8519 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8520 Some CAs (e.g. Verisign) distribute certificates in this form.
8523 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8524 list to exclude them. This means that no special compilation option
8525 is needed to use anonymous DH: it just needs to be included in the
8529 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8530 EVP_MD_type. The old functionality is available in a new macro called
8531 EVP_MD_md(). Change code that uses it and update docs.
8534 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8535 where the 'void *' argument is replaced by a function pointer argument.
8536 Previously 'void *' was abused to point to functions, which works on
8537 many platforms, but is not correct. As these functions are usually
8538 called by macros defined in OpenSSL header files, most source code
8539 should work without changes.
8542 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8543 sections with information on -D... compiler switches used for
8544 compiling the library so that applications can see them. To enable
8545 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8546 must be defined. E.g.,
8547 #define OPENSSL_ALGORITHM_DEFINES
8548 #include <openssl/opensslconf.h>
8549 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8550 [Richard Levitte, Ulf and Bodo Möller]
8552 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8556 *) Change the 'other' type in certificate aux info to a STACK_OF
8557 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8558 the required ASN1 format: arbitrary types determined by an OID.
8561 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8562 argument to 'req'. This is not because the function is newer or
8563 better than others it just uses the work 'NEW' in the certificate
8564 request header lines. Some software needs this.
8567 *) Reorganise password command line arguments: now passwords can be
8568 obtained from various sources. Delete the PEM_cb function and make
8569 it the default behaviour: i.e. if the callback is NULL and the
8570 usrdata argument is not NULL interpret it as a null terminated pass
8571 phrase. If usrdata and the callback are NULL then the pass phrase
8572 is prompted for as usual.
8575 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8576 the support is automatically enabled. The resulting binaries will
8577 autodetect the card and use it if present.
8578 [Ben Laurie and Compaq Inc.]
8580 *) Work around for Netscape hang bug. This sends certificate request
8581 and server done in one record. Since this is perfectly legal in the
8582 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8583 the bugs/SSLv3 entry for more info.
8586 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8589 *) Add -rand argument to smime and pkcs12 applications and read/write
8593 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8596 *) Add command line password options to the remaining applications.
8599 *) Bug fix for BN_div_recp() for numerators with an even number of
8603 *) More tests in bntest.c, and changed test_bn output.
8606 *) ./config recognizes MacOS X now.
8609 *) Bug fix for BN_div() when the first words of num and divsor are
8610 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8613 *) Add support for various broken PKCS#8 formats, and command line
8614 options to produce them.
8617 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8618 get temporary BIGNUMs from a BN_CTX.
8621 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8625 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8626 include a #define from the old name to the new. The original intent
8627 was that statically linked binaries could for example just call
8628 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8629 link with digests. This never worked because SSLeay_add_all_digests()
8630 and SSLeay_add_all_ciphers() were in the same source file so calling
8631 one would link with the other. They are now in separate source files.
8634 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8637 *) Use a less unusual form of the Miller-Rabin primality test (it used
8638 a binary algorithm for exponentiation integrated into the Miller-Rabin
8639 loop, our standard modexp algorithms are faster).
8642 *) Support for the EBCDIC character set completed.
8643 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8645 *) Source code cleanups: use const where appropriate, eliminate casts,
8646 use void * instead of char * in lhash.
8649 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8650 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8651 this the server could overwrite ephemeral keys that the client
8655 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8656 using 50 iterations of the Rabin-Miller test.
8658 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8659 iterations of the Rabin-Miller test as required by the appendix
8660 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8661 As BN_is_prime_fasttest includes trial division, DSA parameter
8662 generation becomes much faster.
8664 This implies a change for the callback functions in DSA_is_prime
8665 and DSA_generate_parameters: The callback function is called once
8666 for each positive witness in the Rabin-Miller test, not just
8667 occasionally in the inner loop; and the parameters to the
8668 callback function now provide an iteration count for the outer
8669 loop rather than for the current invocation of the inner loop.
8670 DSA_generate_parameters additionally can call the callback
8671 function with an 'iteration count' of -1, meaning that a
8672 candidate has passed the trial division test (when q is generated
8673 from an application-provided seed, trial division is skipped).
8676 *) New function BN_is_prime_fasttest that optionally does trial
8677 division before starting the Rabin-Miller test and has
8678 an additional BN_CTX * argument (whereas BN_is_prime always
8679 has to allocate at least one BN_CTX).
8680 'callback(1, -1, cb_arg)' is called when a number has passed the
8681 trial division stage.
8684 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8688 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8691 *) New function BN_pseudo_rand().
8694 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8695 bignum version of BN_from_montgomery() with the working code from
8696 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8700 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8701 made it impossible to use the same SSL_SESSION data structure in
8702 SSL2 clients in multiple threads.
8705 *) The return value of RAND_load_file() no longer counts bytes obtained
8706 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8707 to seed the PRNG (previously an explicit byte count was required).
8708 [Ulf Möller, Bodo Möller]
8710 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8711 used (char *) instead of (void *) and had casts all over the place.
8714 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8717 *) Retain source code compatibility for BN_prime_checks macro:
8718 BN_is_prime(..., BN_prime_checks, ...) now uses
8719 BN_prime_checks_for_size to determine the appropriate number of
8720 Rabin-Miller iterations.
8723 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8724 DH_CHECK_P_NOT_SAFE_PRIME.
8725 (Check if this is true? OpenPGP calls them "strong".)
8728 *) Merge the functionality of "dh" and "gendh" programs into a new program
8729 "dhparam". The old programs are retained for now but will handle DH keys
8730 (instead of parameters) in future.
8733 *) Make the ciphers, s_server and s_client programs check the return values
8734 when a new cipher list is set.
8737 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8738 ciphers. Before when the 56bit ciphers were enabled the sorting was
8741 The syntax for the cipher sorting has been extended to support sorting by
8742 cipher-strength (using the strength_bits hard coded in the tables).
8743 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8745 Fix a bug in the cipher-command parser: when supplying a cipher command
8746 string with an "undefined" symbol (neither command nor alphanumeric
8747 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8748 an error is flagged.
8750 Due to the strength-sorting extension, the code of the
8751 ssl_create_cipher_list() function was completely rearranged. I hope that
8752 the readability was also increased :-)
8753 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8755 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8756 for the first serial number and places 2 in the serial number file. This
8757 avoids problems when the root CA is created with serial number zero and
8758 the first user certificate has the same issuer name and serial number
8762 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8763 the new code. Add documentation for this stuff.
8766 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8767 X509_*() to X509at_*() on the grounds that they don't handle X509
8768 structures and behave in an analogous way to the X509v3 functions:
8769 they shouldn't be called directly but wrapper functions should be used
8772 So we also now have some wrapper functions that call the X509at functions
8773 when passed certificate requests. (TO DO: similar things can be done with
8774 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8775 things. Some of these need some d2i or i2d and print functionality
8776 because they handle more complex structures.)
8779 *) Add missing #ifndefs that caused missing symbols when building libssl
8780 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8781 NO_RSA in ssl/s2*.c.
8782 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8784 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8785 has a return value which indicates the quality of the random data
8786 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8787 error queue. New function RAND_pseudo_bytes() generates output that is
8788 guaranteed to be unique but not unpredictable. RAND_add is like
8789 RAND_seed, but takes an extra argument for an entropy estimate
8790 (RAND_seed always assumes full entropy).
8793 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8794 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8795 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8796 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8797 false-positive rate of at most 2^-80 for random input.
8800 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8803 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8804 in the 0.9.5 release), this returns the chain
8805 from an X509_CTX structure with a dup of the stack and all
8806 the X509 reference counts upped: so the stack will exist
8807 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8810 Also make SSL_SESSION_print() print out the verify return
8814 *) Add manpage for the pkcs12 command. Also change the default
8815 behaviour so MAC iteration counts are used unless the new
8816 -nomaciter option is used. This improves file security and
8817 only older versions of MSIE (4.0 for example) need it.
8820 *) Honor the no-xxx Configure options when creating .DEF files.
8823 *) Add PKCS#10 attributes to field table: challengePassword,
8824 unstructuredName and unstructuredAddress. These are taken from
8825 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8826 international characters are used.
8828 More changes to X509_ATTRIBUTE code: allow the setting of types
8829 based on strings. Remove the 'loc' parameter when adding
8830 attributes because these will be a SET OF encoding which is sorted
8834 *) Initial changes to the 'req' utility to allow request generation
8835 automation. This will allow an application to just generate a template
8836 file containing all the field values and have req construct the
8839 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8840 used all over the place including certificate requests and PKCS#7
8841 structures. They are currently handled manually where necessary with
8842 some primitive wrappers for PKCS#7. The new functions behave in a
8843 manner analogous to the X509 extension functions: they allow
8844 attributes to be looked up by NID and added.
8846 Later something similar to the X509V3 code would be desirable to
8847 automatically handle the encoding, decoding and printing of the
8848 more complex types. The string types like challengePassword can
8849 be handled by the string table functions.
8851 Also modified the multi byte string table handling. Now there is
8852 a 'global mask' which masks out certain types. The table itself
8853 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8854 is useful when for example there is only one permissible type
8855 (as in countryName) and using the mask might result in no valid
8859 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8860 SSL_get_peer_finished to allow applications to obtain the latest
8861 Finished messages sent to the peer or expected from the peer,
8862 respectively. (SSL_get_peer_finished is usually the Finished message
8863 actually received from the peer, otherwise the protocol will be aborted.)
8865 As the Finished message are message digests of the complete handshake
8866 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8867 be used for external authentication procedures when the authentication
8868 provided by SSL/TLS is not desired or is not enough.
8871 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8872 the host supports BWX extension and if Compaq C is present on the
8873 $PATH. Just exploiting of the BWX extension results in 20-30%
8874 performance kick for some algorithms, e.g. DES and RC4 to mention
8875 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8879 *) Add support for MS "fast SGC". This is arguably a violation of the
8880 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8881 weak crypto and after checking the certificate is SGC a second one
8882 with strong crypto. MS SGC stops the first handshake after receiving
8883 the server certificate message and sends a second client hello. Since
8884 a server will typically do all the time consuming operations before
8885 expecting any further messages from the client (server key exchange
8886 is the most expensive) there is little difference between the two.
8888 To get OpenSSL to support MS SGC we have to permit a second client
8889 hello message after we have sent server done. In addition we have to
8890 reset the MAC if we do get this second client hello.
8893 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8894 if a DER encoded private key is RSA or DSA traditional format. Changed
8895 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8896 format DER encoded private key. Newer code should use PKCS#8 format which
8897 has the key type encoded in the ASN1 structure. Added DER private key
8898 support to pkcs8 application.
8901 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8902 ciphersuites has been selected (as required by the SSL 3/TLS 1
8903 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8904 is set, we interpret this as a request to violate the specification
8905 (the worst that can happen is a handshake failure, and 'correct'
8906 behaviour would result in a handshake failure anyway).
8909 *) In SSL_CTX_add_session, take into account that there might be multiple
8910 SSL_SESSION structures with the same session ID (e.g. when two threads
8911 concurrently obtain them from an external cache).
8912 The internal cache can handle only one SSL_SESSION with a given ID,
8913 so if there's a conflict, we now throw out the old one to achieve
8917 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8918 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8919 some routines that use cipher OIDs: some ciphers do not have OIDs
8920 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8924 *) Simplify the trust setting structure and code. Now we just have
8925 two sequences of OIDs for trusted and rejected settings. These will
8926 typically have values the same as the extended key usage extension
8927 and any application specific purposes.
8929 The trust checking code now has a default behaviour: it will just
8930 check for an object with the same NID as the passed id. Functions can
8931 be provided to override either the default behaviour or the behaviour
8932 for a given id. SSL client, server and email already have functions
8933 in place for compatibility: they check the NID and also return "trusted"
8934 if the certificate is self signed.
8937 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8938 traditional format into an EVP_PKEY structure.
8941 *) Add a password callback function PEM_cb() which either prompts for
8942 a password if usr_data is NULL or otherwise assumes it is a null
8943 terminated password. Allow passwords to be passed on command line
8944 environment or config files in a few more utilities.
8947 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8948 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8949 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8950 Update documentation.
8953 *) Support for ASN1 "NULL" type. This could be handled before by using
8954 ASN1_TYPE but there wasn't any function that would try to read a NULL
8955 and produce an error if it couldn't. For compatibility we also have
8956 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8957 don't allocate anything because they don't need to.
8960 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8962 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8964 *) Rebuild of the memory allocation routines used by OpenSSL code and
8965 possibly others as well. The purpose is to make an interface that
8966 provide hooks so anyone can build a separate set of allocation and
8967 deallocation routines to be used by OpenSSL, for example memory
8968 pool implementations, or something else, which was previously hard
8969 since Malloc(), Realloc() and Free() were defined as macros having
8970 the values malloc, realloc and free, respectively (except for Win32
8971 compilations). The same is provided for memory debugging code.
8972 OpenSSL already comes with functionality to find memory leaks, but
8973 this gives people a chance to debug other memory problems.
8975 With these changes, a new set of functions and macros have appeared:
8977 CRYPTO_set_mem_debug_functions() [F]
8978 CRYPTO_get_mem_debug_functions() [F]
8979 CRYPTO_dbg_set_options() [F]
8980 CRYPTO_dbg_get_options() [F]
8981 CRYPTO_malloc_debug_init() [M]
8983 The memory debug functions are NULL by default, unless the library
8984 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8985 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8986 gives the standard debugging functions that come with OpenSSL) or
8987 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8988 provided by the library user) must be used. When the standard
8989 debugging functions are used, CRYPTO_dbg_set_options can be used to
8990 request additional information:
8991 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8992 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8994 Also, things like CRYPTO_set_mem_functions will always give the
8995 expected result (the new set of functions is used for allocation
8996 and deallocation) at all times, regardless of platform and compiler
8999 To finish it up, some functions that were never use in any other
9000 way than through macros have a new API and new semantic:
9003 CRYPTO_dbg_realloc()
9006 All macros of value have retained their old syntax.
9007 [Richard Levitte and Bodo Moeller]
9009 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9010 ordering of SMIMECapabilities wasn't in "strength order" and there
9011 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9015 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9016 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9017 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9019 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9020 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9021 functionality to handle multipart/signed properly) and a utility
9022 called 'smime' to call all this stuff. This is based on code I
9023 originally wrote for Celo who have kindly allowed it to be
9024 included in OpenSSL.
9027 *) Add variants des_set_key_checked and des_set_key_unchecked of
9028 des_set_key (aka des_key_sched). Global variable des_check_key
9029 decides which of these is called by des_set_key; this way
9030 des_check_key behaves as it always did, but applications and
9031 the library itself, which was buggy for des_check_key == 1,
9032 have a cleaner way to pick the version they need.
9035 *) New function PKCS12_newpass() which changes the password of a
9039 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9040 dynamic mix. In both cases the ids can be used as an index into the
9041 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9042 functions so they accept a list of the field values and the
9043 application doesn't need to directly manipulate the X509_TRUST
9047 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9051 *) Modify the way the V3 extension code looks up extensions. This now
9052 works in a similar way to the object code: we have some "standard"
9053 extensions in a static table which is searched with OBJ_bsearch()
9054 and the application can add dynamic ones if needed. The file
9055 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9056 updated whenever a new extension is added to the core code and kept
9057 in ext_nid order. There is a simple program 'tabtest.c' which checks
9058 this. New extensions are not added too often so this file can readily
9059 be maintained manually.
9061 There are two big advantages in doing things this way. The extensions
9062 can be looked up immediately and no longer need to be "added" using
9063 X509V3_add_standard_extensions(): this function now does nothing.
9064 [Side note: I get *lots* of email saying the extension code doesn't
9065 work because people forget to call this function]
9066 Also no dynamic allocation is done unless new extensions are added:
9067 so if we don't add custom extensions there is no need to call
9068 X509V3_EXT_cleanup().
9071 *) Modify enc utility's salting as follows: make salting the default. Add a
9072 magic header, so unsalted files fail gracefully instead of just decrypting
9073 to garbage. This is because not salting is a big security hole, so people
9074 should be discouraged from doing it.
9077 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9078 digest to be passed on the command line but it only used this
9079 parameter when signing a certificate. Modified so all relevant
9080 operations are affected by the digest parameter including the
9081 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9082 DSA key was used because it didn't fix the digest.
9085 *) Initial certificate chain verify code. Currently tests the untrusted
9086 certificates for consistency with the verify purpose (which is set
9087 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9089 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9090 this is because it will reject chains with invalid extensions whereas
9091 every previous version of OpenSSL and SSLeay made no checks at all.
9093 Trust code: checks the root CA for the relevant trust settings. Trust
9094 settings have an initial value consistent with the verify purpose: e.g.
9095 if the verify purpose is for SSL client use it expects the CA to be
9096 trusted for SSL client use. However the default value can be changed to
9097 permit custom trust settings: one example of this would be to only trust
9098 certificates from a specific "secure" set of CAs.
9100 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9101 which should be used for version portability: especially since the
9102 verify structure is likely to change more often now.
9104 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9105 to set them. If not set then assume SSL clients will verify SSL servers
9108 Two new options to the verify program: -untrusted allows a set of
9109 untrusted certificates to be passed in and -purpose which sets the
9110 intended purpose of the certificate. If a purpose is set then the
9111 new chain verify code is used to check extension consistency.
9114 *) Support for the authority information access extension.
9117 *) Modify RSA and DSA PEM read routines to transparently handle
9118 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9119 public keys in a format compatible with certificate
9120 SubjectPublicKeyInfo structures. Unfortunately there were already
9121 functions called *_PublicKey_* which used various odd formats so
9122 these are retained for compatibility: however the DSA variants were
9123 never in a public release so they have been deleted. Changed dsa/rsa
9124 utilities to handle the new format: note no releases ever handled public
9125 keys so we should be OK.
9127 The primary motivation for this change is to avoid the same fiasco
9128 that dogs private keys: there are several incompatible private key
9129 formats some of which are standard and some OpenSSL specific and
9130 require various evil hacks to allow partial transparent handling and
9131 even then it doesn't work with DER formats. Given the option anything
9132 other than PKCS#8 should be dumped: but the other formats have to
9133 stay in the name of compatibility.
9135 With public keys and the benefit of hindsight one standard format
9136 is used which works with EVP_PKEY, RSA or DSA structures: though
9137 it clearly returns an error if you try to read the wrong kind of key.
9139 Added a -pubkey option to the 'x509' utility to output the public key.
9140 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9141 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9142 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9143 that do the same as the EVP_PKEY_assign_*() except they up the
9144 reference count of the added key (they don't "swallow" the
9148 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9149 CRLs would fail if the file contained no certificates or no CRLs:
9150 added a new function to read in both types and return the number
9151 read: this means that if none are read it will be an error. The
9152 DER versions of the certificate and CRL reader would always fail
9153 because it isn't possible to mix certificates and CRLs in DER format
9154 without choking one or the other routine. Changed this to just read
9155 a certificate: this is the best we can do. Also modified the code
9156 in apps/verify.c to take notice of return codes: it was previously
9157 attempting to read in certificates from NULL pointers and ignoring
9158 any errors: this is one reason why the cert and CRL reader seemed
9159 to work. It doesn't check return codes from the default certificate
9160 routines: these may well fail if the certificates aren't installed.
9163 *) Code to support otherName option in GeneralName.
9166 *) First update to verify code. Change the verify utility
9167 so it warns if it is passed a self signed certificate:
9168 for consistency with the normal behaviour. X509_verify
9169 has been modified to it will now verify a self signed
9170 certificate if *exactly* the same certificate appears
9171 in the store: it was previously impossible to trust a
9172 single self signed certificate. This means that:
9173 openssl verify ss.pem
9174 now gives a warning about a self signed certificate but
9175 openssl verify -CAfile ss.pem ss.pem
9179 *) For servers, store verify_result in SSL_SESSION data structure
9180 (and add it to external session representation).
9181 This is needed when client certificate verifications fails,
9182 but an application-provided verification callback (set by
9183 SSL_CTX_set_cert_verify_callback) allows accepting the session
9184 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9185 but returns 1): When the session is reused, we have to set
9186 ssl->verify_result to the appropriate error code to avoid
9188 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9190 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9191 case in PKCS7_dataInit() where the signed PKCS7 structure
9192 didn't contain any existing data because it was being created.
9193 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9195 *) Add a salt to the key derivation routines in enc.c. This
9196 forms the first 8 bytes of the encrypted file. Also add a
9197 -S option to allow a salt to be input on the command line.
9200 *) New function X509_cmp(). Oddly enough there wasn't a function
9201 to compare two certificates. We do this by working out the SHA1
9202 hash and comparing that. X509_cmp() will be needed by the trust
9206 *) SSL_get1_session() is like SSL_get_session(), but increments
9207 the reference count in the SSL_SESSION returned.
9208 [Geoff Thorpe <geoff@eu.c2.net>]
9210 *) Fix for 'req': it was adding a null to request attributes.
9211 Also change the X509_LOOKUP and X509_INFO code to handle
9212 certificate auxiliary information.
9215 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9219 *) Add the possibility to add extra information to the memory leak
9220 detecting output, to form tracebacks, showing from where each
9221 allocation was originated: CRYPTO_push_info("constant string") adds
9222 the string plus current file name and line number to a per-thread
9223 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9224 is like calling CYRPTO_pop_info() until the stack is empty.
9225 Also updated memory leak detection code to be multi-thread-safe.
9228 *) Add options -text and -noout to pkcs7 utility and delete the
9229 encryption options which never did anything. Update docs.
9232 *) Add options to some of the utilities to allow the pass phrase
9233 to be included on either the command line (not recommended on
9234 OSes like Unix) or read from the environment. Update the
9235 manpages and fix a few bugs.
9238 *) Add a few manpages for some of the openssl commands.
9241 *) Fix the -revoke option in ca. It was freeing up memory twice,
9242 leaking and not finding already revoked certificates.
9245 *) Extensive changes to support certificate auxiliary information.
9246 This involves the use of X509_CERT_AUX structure and X509_AUX
9247 functions. An X509_AUX function such as PEM_read_X509_AUX()
9248 can still read in a certificate file in the usual way but it
9249 will also read in any additional "auxiliary information". By
9250 doing things this way a fair degree of compatibility can be
9251 retained: existing certificates can have this information added
9252 using the new 'x509' options.
9254 Current auxiliary information includes an "alias" and some trust
9255 settings. The trust settings will ultimately be used in enhanced
9256 certificate chain verification routines: currently a certificate
9257 can only be trusted if it is self signed and then it is trusted
9261 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9262 The problem was that one of the replacement routines had not been working
9263 since SSLeay releases. For now the offending routine has been replaced
9264 with non-optimised assembler. Even so, this now gives around 95%
9265 performance improvement for 1024 bit RSA signs.
9268 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9269 handling. Most clients have the effective key size in bits equal to
9270 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9271 A few however don't do this and instead use the size of the decrypted key
9272 to determine the RC2 key length and the AlgorithmIdentifier to determine
9273 the effective key length. In this case the effective key length can still
9274 be 40 bits but the key length can be 168 bits for example. This is fixed
9275 by manually forcing an RC2 key into the EVP_PKEY structure because the
9276 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9277 the key length and effective key length are equal.
9280 *) Add a bunch of functions that should simplify the creation of
9281 X509_NAME structures. Now you should be able to do:
9282 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9283 and have it automatically work out the correct field type and fill in
9284 the structures. The more adventurous can try:
9285 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9286 and it will (hopefully) work out the correct multibyte encoding.
9289 *) Change the 'req' utility to use the new field handling and multibyte
9290 copy routines. Before the DN field creation was handled in an ad hoc
9291 way in req, ca, and x509 which was rather broken and didn't support
9292 BMPStrings or UTF8Strings. Since some software doesn't implement
9293 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9294 using the dirstring_type option. See the new comment in the default
9295 openssl.cnf for more info.
9298 *) Make crypto/rand/md_rand.c more robust:
9299 - Assure unique random numbers after fork().
9300 - Make sure that concurrent threads access the global counter and
9301 md serializably so that we never lose entropy in them
9302 or use exactly the same state in multiple threads.
9303 Access to the large state is not always serializable because
9304 the additional locking could be a performance killer, and
9305 md should be large enough anyway.
9308 *) New file apps/app_rand.c with commonly needed functionality
9309 for handling the random seed file.
9311 Use the random seed file in some applications that previously did not:
9313 dsaparam -genkey (which also ignored its '-rand' option),
9316 x509 (when signing).
9317 Except on systems with /dev/urandom, it is crucial to have a random
9318 seed file at least for key creation, DSA signing, and for DH exchanges;
9319 for RSA signatures we could do without one.
9321 gendh and gendsa (unlike genrsa) used to read only the first byte
9322 of each file listed in the '-rand' option. The function as previously
9323 found in genrsa is now in app_rand.c and is used by all programs
9324 that support '-rand'.
9327 *) In RAND_write_file, use mode 0600 for creating files;
9328 don't just chmod when it may be too late.
9331 *) Report an error from X509_STORE_load_locations
9332 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9335 *) New function ASN1_mbstring_copy() this copies a string in either
9336 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9337 into an ASN1_STRING type. A mask of permissible types is passed
9338 and it chooses the "minimal" type to use or an error if not type
9342 *) Add function equivalents to the various macros in asn1.h. The old
9343 macros are retained with an M_ prefix. Code inside the library can
9344 use the M_ macros. External code (including the openssl utility)
9345 should *NOT* in order to be "shared library friendly".
9348 *) Add various functions that can check a certificate's extensions
9349 to see if it usable for various purposes such as SSL client,
9350 server or S/MIME and CAs of these types. This is currently
9351 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9352 verification. Also added a -purpose flag to x509 utility to
9353 print out all the purposes.
9356 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9360 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9361 for, obtain and decode and extension and obtain its critical flag.
9362 This allows all the necessary extension code to be handled in a
9363 single function call.
9366 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9367 platforms. See crypto/rc4/rc4_enc.c for further details.
9370 *) New -noout option to asn1parse. This causes no output to be produced
9371 its main use is when combined with -strparse and -out to extract data
9372 from a file (which may not be in ASN.1 format).
9375 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9376 when producing the local key id.
9377 [Richard Levitte <levitte@stacken.kth.se>]
9379 *) New option -dhparam in s_server. This allows a DH parameter file to be
9380 stated explicitly. If it is not stated then it tries the first server
9381 certificate file. The previous behaviour hard coded the filename
9385 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9386 a public key to be input or output. For example:
9387 openssl rsa -in key.pem -pubout -out pubkey.pem
9388 Also added necessary DSA public key functions to handle this.
9391 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9392 in the message. This was handled by allowing
9393 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9394 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9396 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9397 to the end of the strings whereas this didn't. This would cause problems
9398 if strings read with d2i_ASN1_bytes() were later modified.
9399 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9401 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9402 data and it contains EOF it will end up returning an error. This is
9403 caused by input 46 bytes long. The cause is due to the way base64
9404 BIOs find the start of base64 encoded data. They do this by trying a
9405 trial decode on each line until they find one that works. When they
9406 do a flag is set and it starts again knowing it can pass all the
9407 data directly through the decoder. Unfortunately it doesn't reset
9408 the context it uses. This means that if EOF is reached an attempt
9409 is made to pass two EOFs through the context and this causes the
9410 resulting error. This can also cause other problems as well. As is
9411 usual with these problems it takes *ages* to find and the fix is
9412 trivial: move one line.
9413 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9415 *) Ugly workaround to get s_client and s_server working under Windows. The
9416 old code wouldn't work because it needed to select() on sockets and the
9417 tty (for keypresses and to see if data could be written). Win32 only
9418 supports select() on sockets so we select() with a 1s timeout on the
9419 sockets and then see if any characters are waiting to be read, if none
9420 are present then we retry, we also assume we can always write data to
9421 the tty. This isn't nice because the code then blocks until we've
9422 received a complete line of data and it is effectively polling the
9423 keyboard at 1s intervals: however it's quite a bit better than not
9424 working at all :-) A dedicated Windows application might handle this
9425 with an event loop for example.
9428 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9429 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9430 will be called when RSA_sign() and RSA_verify() are used. This is useful
9431 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9432 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9433 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9434 This necessitated the support of an extra signature type NID_md5_sha1
9435 for SSL signatures and modifications to the SSL library to use it instead
9436 of calling RSA_public_decrypt() and RSA_private_encrypt().
9439 *) Add new -verify -CAfile and -CApath options to the crl program, these
9440 will lookup a CRL issuers certificate and verify the signature in a
9441 similar way to the verify program. Tidy up the crl program so it
9442 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9443 less strict. It will now permit CRL extensions even if it is not
9444 a V2 CRL: this will allow it to tolerate some broken CRLs.
9447 *) Initialize all non-automatic variables each time one of the openssl
9448 sub-programs is started (this is necessary as they may be started
9449 multiple times from the "OpenSSL>" prompt).
9450 [Lennart Bang, Bodo Moeller]
9452 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9453 removing all other RSA functionality (this is what NO_RSA does). This
9454 is so (for example) those in the US can disable those operations covered
9455 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9459 *) Non-copying interface to BIO pairs.
9460 (still largely untested)
9463 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9464 ASCII string. This was handled independently in various places before.
9467 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9468 UTF8 strings a character at a time.
9471 *) Use client_version from client hello to select the protocol
9472 (s23_srvr.c) and for RSA client key exchange verification
9473 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9476 *) Add various utility functions to handle SPKACs, these were previously
9477 handled by poking round in the structure internals. Added new function
9478 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9479 print, verify and generate SPKACs. Based on an original idea from
9480 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9483 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9486 *) Allow the config file extension section to be overwritten on the
9487 command line. Based on an original idea from Massimiliano Pala
9488 <madwolf@comune.modena.it>. The new option is called -extensions
9489 and can be applied to ca, req and x509. Also -reqexts to override
9490 the request extensions in req and -crlexts to override the crl extensions
9494 *) Add new feature to the SPKAC handling in ca. Now you can include
9495 the same field multiple times by preceding it by "XXXX." for example:
9498 this is the same syntax as used in the req config file.
9501 *) Allow certificate extensions to be added to certificate requests. These
9502 are specified in a 'req_extensions' option of the req section of the
9503 config file. They can be printed out with the -text option to req but
9504 are otherwise ignored at present.
9507 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9508 data read consists of only the final block it would not decrypted because
9509 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9510 A misplaced 'break' also meant the decrypted final block might not be
9511 copied until the next read.
9514 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9515 a few extra parameters to the DH structure: these will be useful if
9516 for example we want the value of 'q' or implement X9.42 DH.
9519 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9520 provides hooks that allow the default DSA functions or functions on a
9521 "per key" basis to be replaced. This allows hardware acceleration and
9522 hardware key storage to be handled without major modification to the
9523 library. Also added low level modexp hooks and CRYPTO_EX structure and
9524 associated functions.
9527 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9528 as "read only": it can't be written to and the buffer it points to will
9529 not be freed. Reading from a read only BIO is much more efficient than
9530 a normal memory BIO. This was added because there are several times when
9531 an area of memory needs to be read from a BIO. The previous method was
9532 to create a memory BIO and write the data to it, this results in two
9533 copies of the data and an O(n^2) reading algorithm. There is a new
9534 function BIO_new_mem_buf() which creates a read only memory BIO from
9535 an area of memory. Also modified the PKCS#7 routines to use read only
9539 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9540 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9541 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9542 but a retry condition occured while trying to read the rest.
9545 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9546 NID_pkcs7_encrypted by default: this was wrong since this should almost
9547 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9548 the encrypted data type: this is a more sensible place to put it and it
9549 allows the PKCS#12 code to be tidied up that duplicated this
9553 *) Changed obj_dat.pl script so it takes its input and output files on
9554 the command line. This should avoid shell escape redirection problems
9558 *) Initial support for certificate extension requests, these are included
9559 in things like Xenroll certificate requests. Included functions to allow
9560 extensions to be obtained and added.
9563 *) -crlf option to s_client and s_server for sending newlines as
9564 CRLF (as required by many protocols).
9567 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9569 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9570 [Ralf S. Engelschall]
9572 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9573 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9575 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9579 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9580 DH parameters/keys (q is lost during that conversion, but the resulting
9581 DH parameters contain its length).
9583 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9584 much faster than DH_generate_parameters (which creates parameters
9585 where p = 2*q + 1), and also the smaller q makes DH computations
9586 much more efficient (160-bit exponentiation instead of 1024-bit
9587 exponentiation); so this provides a convenient way to support DHE
9588 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9589 utter importance to use
9590 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9592 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9593 when such DH parameters are used, because otherwise small subgroup
9594 attacks may become possible!
9597 *) Avoid memory leak in i2d_DHparams.
9600 *) Allow the -k option to be used more than once in the enc program:
9601 this allows the same encrypted message to be read by multiple recipients.
9604 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9605 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9606 it will always use the numerical form of the OID, even if it has a short
9610 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9611 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9612 otherwise bn_mod_exp was called. In the case of hardware keys for example
9613 no private key components need be present and it might store extra data
9614 in the RSA structure, which cannot be accessed from bn_mod_exp.
9615 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9616 private key operations.
9619 *) Added support for SPARC Linux.
9622 *) pem_password_cb function type incompatibly changed from
9623 typedef int pem_password_cb(char *buf, int size, int rwflag);
9625 ....(char *buf, int size, int rwflag, void *userdata);
9626 so that applications can pass data to their callbacks:
9627 The PEM[_ASN1]_{read,write}... functions and macros now take an
9628 additional void * argument, which is just handed through whenever
9629 the password callback is called.
9630 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9632 New function SSL_CTX_set_default_passwd_cb_userdata.
9634 Compatibility note: As many C implementations push function arguments
9635 onto the stack in reverse order, the new library version is likely to
9636 interoperate with programs that have been compiled with the old
9637 pem_password_cb definition (PEM_whatever takes some data that
9638 happens to be on the stack as its last argument, and the callback
9639 just ignores this garbage); but there is no guarantee whatsoever that
9642 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9643 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9644 problems not only on Windows, but also on some Unix platforms.
9645 To avoid problematic command lines, these definitions are now in an
9646 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9647 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9650 *) MIPS III/IV assembler module is reimplemented.
9653 *) More DES library cleanups: remove references to srand/rand and
9654 delete an unused file.
9657 *) Add support for the the free Netwide assembler (NASM) under Win32,
9658 since not many people have MASM (ml) and it can be hard to obtain.
9659 This is currently experimental but it seems to work OK and pass all
9660 the tests. Check out INSTALL.W32 for info.
9663 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9664 without temporary keys kept an extra copy of the server key,
9665 and connections with temporary keys did not free everything in case
9669 *) New function RSA_check_key and new openssl rsa option -check
9670 for verifying the consistency of RSA keys.
9671 [Ulf Moeller, Bodo Moeller]
9673 *) Various changes to make Win32 compile work:
9674 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9675 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9676 comparison" warnings.
9677 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9680 *) Add a debugging option to PKCS#5 v2 key generation function: when
9681 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9682 derived keys are printed to stderr.
9685 *) Copy the flags in ASN1_STRING_dup().
9686 [Roman E. Pavlov <pre@mo.msk.ru>]
9688 *) The x509 application mishandled signing requests containing DSA
9689 keys when the signing key was also DSA and the parameters didn't match.
9691 It was supposed to omit the parameters when they matched the signing key:
9692 the verifying software was then supposed to automatically use the CA's
9693 parameters if they were absent from the end user certificate.
9695 Omitting parameters is no longer recommended. The test was also
9696 the wrong way round! This was probably due to unusual behaviour in
9697 EVP_cmp_parameters() which returns 1 if the parameters match.
9698 This meant that parameters were omitted when they *didn't* match and
9699 the certificate was useless. Certificates signed with 'ca' didn't have
9701 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9703 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9704 The interface is as follows:
9705 Applications can use
9706 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9707 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9708 "off" is now the default.
9709 The library internally uses
9710 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9711 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9712 to disable memory-checking temporarily.
9714 Some inconsistent states that previously were possible (and were
9715 even the default) are now avoided.
9717 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9718 with each memory chunk allocated; this is occasionally more helpful
9719 than just having a counter.
9721 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9723 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9727 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9728 which largely parallels "options", but is for changing API behaviour,
9729 whereas "options" are about protocol behaviour.
9730 Initial "mode" flags are:
9732 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9733 a single record has been written.
9734 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9735 retries use the same buffer location.
9736 (But all of the contents must be
9740 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9743 *) Fix problems with no-hmac etc.
9744 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9746 *) New functions RSA_get_default_method(), RSA_set_method() and
9747 RSA_get_method(). These allows replacement of RSA_METHODs without having
9748 to mess around with the internals of an RSA structure.
9751 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9752 Also really enable memory leak checks in openssl.c and in some
9754 [Chad C. Mulligan, Bodo Moeller]
9756 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9757 up the length of negative integers. This has now been simplified to just
9758 store the length when it is first determined and use it later, rather
9759 than trying to keep track of where data is copied and updating it to
9761 [Steve Henson, reported by Brien Wheeler
9762 <bwheeler@authentica-security.com>]
9764 *) Add a new function PKCS7_signatureVerify. This allows the verification
9765 of a PKCS#7 signature but with the signing certificate passed to the
9766 function itself. This contrasts with PKCS7_dataVerify which assumes the
9767 certificate is present in the PKCS#7 structure. This isn't always the
9768 case: certificates can be omitted from a PKCS#7 structure and be
9769 distributed by "out of band" means (such as a certificate database).
9772 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9773 function prototypes in pem.h, also change util/mkdef.pl to add the
9774 necessary function names.
9777 *) mk1mf.pl (used by Windows builds) did not properly read the
9778 options set by Configure in the top level Makefile, and Configure
9779 was not even able to write more than one option correctly.
9780 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9783 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9784 file to be loaded from a BIO or FILE pointer. The BIO version will
9785 for example allow memory BIOs to contain config info.
9788 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9789 Whoever hopes to achieve shared-library compatibility across versions
9790 must use this, not the compile-time macro.
9791 (Exercise 0.9.4: Which is the minimum library version required by
9793 Note: All this applies only to multi-threaded programs, others don't
9797 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9798 through a BIO pair triggered the default case, i.e.
9799 SSLerr(...,SSL_R_UNKNOWN_STATE).
9802 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9803 can use the SSL library even if none of the specific BIOs is
9807 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9808 for the encoded length.
9809 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9811 *) Add initial documentation of the X509V3 functions.
9814 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9815 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9816 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9817 secure PKCS#8 private key format with a high iteration count.
9820 *) Fix determination of Perl interpreter: A perl or perl5
9821 _directory_ in $PATH was also accepted as the interpreter.
9822 [Ralf S. Engelschall]
9824 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9825 wrong with it but it was very old and did things like calling
9826 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9830 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9831 to use the new extension code.
9834 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9835 with macros. This should make it easier to change their form, add extra
9836 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9840 *) Add to configuration table a new entry that can specify an alternative
9841 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9842 according to Mark Crispin <MRC@Panda.COM>.
9846 *) DES CBC did not update the IV. Weird.
9849 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9850 Changing the behaviour of the former might break existing programs --
9851 where IV updating is needed, des_ncbc_encrypt can be used.
9854 *) When bntest is run from "make test" it drives bc to check its
9855 calculations, as well as internally checking them. If an internal check
9856 fails, it needs to cause bc to give a non-zero result or make test carries
9857 on without noticing the failure. Fixed.
9860 *) DES library cleanups.
9863 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9864 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9865 ciphers. NOTE: although the key derivation function has been verified
9866 against some published test vectors it has not been extensively tested
9867 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9871 *) Instead of "mkdir -p", which is not fully portable, use new
9872 Perl script "util/mkdir-p.pl".
9875 *) Rewrite the way password based encryption (PBE) is handled. It used to
9876 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9877 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9878 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9879 the 'parameter' field of the AlgorithmIdentifier is passed to the
9880 underlying key generation function so it must do its own ASN1 parsing.
9881 This has also changed the EVP_PBE_CipherInit() function which now has a
9882 'parameter' argument instead of literal salt and iteration count values
9883 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9886 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9887 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9888 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9889 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9890 value was just used as a "magic string" and not used directly its
9891 value doesn't matter.
9894 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9898 *) "linux-sparc64" configuration (ultrapenguin).
9899 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9900 "linux-sparc" configuration.
9901 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9903 *) config now generates no-xxx options for missing ciphers.
9906 *) Support the EBCDIC character set (work in progress).
9907 File ebcdic.c not yet included because it has a different license.
9908 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9910 *) Support BS2000/OSD-POSIX.
9911 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9913 *) Make callbacks for key generation use void * instead of char *.
9916 *) Make S/MIME samples compile (not yet tested).
9919 *) Additional typesafe stacks.
9922 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9926 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9928 *) New configuration variant "sco5-gcc".
9930 *) Updated some demos.
9931 [Sean O Riordain, Wade Scholine]
9933 *) Add missing BIO_free at exit of pkcs12 application.
9936 *) Fix memory leak in conf.c.
9939 *) Updates for Win32 to assembler version of MD5.
9942 *) Set #! path to perl in apps/der_chop to where we found it
9943 instead of using a fixed path.
9946 *) SHA library changes for irix64-mips4-cc.
9949 *) Improvements for VMS support.
9953 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9955 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9956 This also avoids the problems with SC4.2 and unpatched SC5.
9957 [Andy Polyakov <appro@fy.chalmers.se>]
9959 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9960 These are required because of the typesafe stack would otherwise break
9961 existing code. If old code used a structure member which used to be STACK
9962 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9963 sk_num or sk_value it would produce an error because the num, data members
9964 are not present in STACK_OF. Now it just produces a warning. sk_set
9965 replaces the old method of assigning a value to sk_value
9966 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9967 that does this will no longer work (and should use sk_set instead) but
9968 this could be regarded as a "questionable" behaviour anyway.
9971 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9972 correctly handle encrypted S/MIME data.
9975 *) Change type of various DES function arguments from des_cblock
9976 (which means, in function argument declarations, pointer to char)
9977 to des_cblock * (meaning pointer to array with 8 char elements),
9978 which allows the compiler to do more typechecking; it was like
9979 that back in SSLeay, but with lots of ugly casts.
9981 Introduce new type const_des_cblock.
9984 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9985 problems: find RecipientInfo structure that matches recipient certificate
9986 and initialise the ASN1 structures properly based on passed cipher.
9989 *) Belatedly make the BN tests actually check the results.
9992 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9993 to and from BNs: it was completely broken. New compilation option
9994 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9995 key elements as negative integers.
9998 *) Reorganize and speed up MD5.
9999 [Andy Polyakov <appro@fy.chalmers.se>]
10002 [Richard Levitte <richard@levitte.org>]
10004 *) New option -out to asn1parse to allow the parsed structure to be
10005 output to a file. This is most useful when combined with the -strparse
10006 option to examine the output of things like OCTET STRINGS.
10009 *) Make SSL library a little more fool-proof by not requiring any longer
10010 that SSL_set_{accept,connect}_state be called before
10011 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10012 in many applications because usually everything *appeared* to work as
10013 intended anyway -- now it really works as intended).
10016 *) Move openssl.cnf out of lib/.
10019 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10020 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10021 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10022 [Ralf S. Engelschall]
10024 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10025 handle PKCS#7 enveloped data properly.
10026 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10028 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10029 copying pointers. The cert_st handling is changed by this in
10030 various ways (and thus what used to be known as ctx->default_cert
10031 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10032 any longer when s->cert does not give us what we need).
10033 ssl_cert_instantiate becomes obsolete by this change.
10034 As soon as we've got the new code right (possibly it already is?),
10035 we have solved a couple of bugs of the earlier code where s->cert
10036 was used as if it could not have been shared with other SSL structures.
10038 Note that using the SSL API in certain dirty ways now will result
10039 in different behaviour than observed with earlier library versions:
10040 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10041 does not influence s as it used to.
10043 In order to clean up things more thoroughly, inside SSL_SESSION
10044 we don't use CERT any longer, but a new structure SESS_CERT
10045 that holds per-session data (if available); currently, this is
10046 the peer's certificate chain and, for clients, the server's certificate
10047 and temporary key. CERT holds only those values that can have
10048 meaningful defaults in an SSL_CTX.
10051 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10052 from the internal representation. Various PKCS#7 fixes: remove some
10053 evil casts and set the enc_dig_alg field properly based on the signing
10057 *) Allow PKCS#12 password to be set from the command line or the
10058 environment. Let 'ca' get its config file name from the environment
10059 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10063 *) Allow certificate policies extension to use an IA5STRING for the
10064 organization field. This is contrary to the PKIX definition but
10065 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10069 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10070 without disallowing inline assembler and the like for non-pedantic builds.
10073 *) Support Borland C++ builder.
10074 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10076 *) Support Mingw32.
10079 *) SHA-1 cleanups and performance enhancements.
10080 [Andy Polyakov <appro@fy.chalmers.se>]
10082 *) Sparc v8plus assembler for the bignum library.
10083 [Andy Polyakov <appro@fy.chalmers.se>]
10085 *) Accept any -xxx and +xxx compiler options in Configure.
10088 *) Update HPUX configuration.
10091 *) Add missing sk_<type>_unshift() function to safestack.h
10092 [Ralf S. Engelschall]
10094 *) New function SSL_CTX_use_certificate_chain_file that sets the
10095 "extra_cert"s in addition to the certificate. (This makes sense
10096 only for "PEM" format files, as chains as a whole are not
10100 *) Support verify_depth from the SSL API.
10101 x509_vfy.c had what can be considered an off-by-one-error:
10102 Its depth (which was not part of the external interface)
10103 was actually counting the number of certificates in a chain;
10104 now it really counts the depth.
10107 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10108 instead of X509err, which often resulted in confusing error
10109 messages since the error codes are not globally unique
10110 (e.g. an alleged error in ssl3_accept when a certificate
10111 didn't match the private key).
10113 *) New function SSL_CTX_set_session_id_context that allows to set a default
10114 value (so that you don't need SSL_set_session_id_context for each
10115 connection using the SSL_CTX).
10118 *) OAEP decoding bug fix.
10121 *) Support INSTALL_PREFIX for package builders, as proposed by
10125 *) New Configure options "threads" and "no-threads". For systems
10126 where the proper compiler options are known (currently Solaris
10127 and Linux), "threads" is the default.
10130 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10133 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10134 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10135 such as /usr/local/bin.
10138 *) "make linux-shared" to build shared libraries.
10139 [Niels Poppe <niels@netbox.org>]
10141 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10144 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10145 extension adding in x509 utility.
10148 *) Remove NOPROTO sections and error code comments.
10151 *) Partial rewrite of the DEF file generator to now parse the ANSI
10155 *) New Configure options --prefix=DIR and --openssldir=DIR.
10158 *) Complete rewrite of the error code script(s). It is all now handled
10159 by one script at the top level which handles error code gathering,
10160 header rewriting and C source file generation. It should be much better
10161 than the old method: it now uses a modified version of Ulf's parser to
10162 read the ANSI prototypes in all header files (thus the old K&R definitions
10163 aren't needed for error creation any more) and do a better job of
10164 translating function codes into names. The old 'ASN1 error code imbedded
10165 in a comment' is no longer necessary and it doesn't use .err files which
10166 have now been deleted. Also the error code call doesn't have to appear all
10167 on one line (which resulted in some large lines...).
10170 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10173 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10174 0 (which usually indicates a closed connection), but continue reading.
10177 *) Fix some race conditions.
10180 *) Add support for CRL distribution points extension. Add Certificate
10181 Policies and CRL distribution points documentation.
10184 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10187 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10188 8 of keying material. Merlin has also confirmed interop with this fix
10189 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10190 [Merlin Hughes <merlin@baltimore.ie>]
10192 *) Fix lots of warnings.
10193 [Richard Levitte <levitte@stacken.kth.se>]
10195 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10196 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10197 [Richard Levitte <levitte@stacken.kth.se>]
10199 *) Fix problems with sizeof(long) == 8.
10200 [Andy Polyakov <appro@fy.chalmers.se>]
10202 *) Change functions to ANSI C.
10205 *) Fix typos in error codes.
10206 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10208 *) Remove defunct assembler files from Configure.
10211 *) SPARC v8 assembler BIGNUM implementation.
10212 [Andy Polyakov <appro@fy.chalmers.se>]
10214 *) Support for Certificate Policies extension: both print and set.
10215 Various additions to support the r2i method this uses.
10218 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10219 return a const string when you are expecting an allocated buffer.
10222 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10223 types DirectoryString and DisplayText.
10226 *) Add code to allow r2i extensions to access the configuration database,
10227 add an LHASH database driver and add several ctx helper functions.
10230 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10231 fail when they extended the size of a BIGNUM.
10234 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10235 support typesafe stack.
10238 *) Fix typo in SSL_[gs]et_options().
10239 [Nils Frostberg <nils@medcom.se>]
10241 *) Delete various functions and files that belonged to the (now obsolete)
10242 old X509V3 handling code.
10245 *) New Configure option "rsaref".
10248 *) Don't auto-generate pem.h.
10251 *) Introduce type-safe ASN.1 SETs.
10254 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10255 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10257 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10258 that links with OpenSSL (well at least cause lots of warnings), but fear
10259 not: the conversion is trivial, and it eliminates loads of evil casts. A
10260 few STACKed things have been converted already. Feel free to convert more.
10261 In the fullness of time, I'll do away with the STACK type altogether.
10264 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10265 specified in <certfile> by updating the entry in the index.txt file.
10266 This way one no longer has to edit the index.txt file manually for
10267 revoking a certificate. The -revoke option does the gory details now.
10268 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10270 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10271 `-text' option at all and this way the `-noout -text' combination was
10272 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10273 [Ralf S. Engelschall]
10275 *) Make sure a corresponding plain text error message exists for the
10276 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10277 verify callback function determined that a certificate was revoked.
10278 [Ralf S. Engelschall]
10280 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10281 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10282 all available cipers including rc5, which was forgotten until now.
10283 In order to let the testing shell script know which algorithms
10284 are available, a new (up to now undocumented) command
10285 "openssl list-cipher-commands" is used.
10288 *) Bugfix: s_client occasionally would sleep in select() when
10289 it should have checked SSL_pending() first.
10292 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10293 the raw DSA values prior to ASN.1 encoding.
10296 *) Tweaks to Configure
10297 [Niels Poppe <niels@netbox.org>]
10299 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10303 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10306 *) New config option to avoid instructions that are illegal on the 80386.
10307 The default code is faster, but requires at least a 486.
10310 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10311 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10312 same as SSL2_VERSION anyway.
10315 *) New "-showcerts" option for s_client.
10318 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10319 application. Various cleanups and fixes.
10322 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10323 modify error routines to work internally. Add error codes and PBE init
10324 to library startup routines.
10327 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10328 packing functions to asn1 and evp. Changed function names and error
10329 codes along the way.
10332 *) PKCS12 integration: and so it begins... First of several patches to
10333 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10334 objects to objects.h
10337 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10338 and display support for Thawte strong extranet extension.
10341 *) Add LinuxPPC support.
10342 [Jeff Dubrule <igor@pobox.org>]
10344 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10345 bn_div_words in alpha.s.
10346 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10348 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10349 OAEP isn't supported when OpenSSL is built with RSAref.
10350 [Ulf Moeller <ulf@fitug.de>]
10352 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10353 so they no longer are missing under -DNOPROTO.
10354 [Soren S. Jorvang <soren@t.dk>]
10357 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10359 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10360 doesn't work when the session is reused. Coming soon!
10363 *) Fix a security hole, that allows sessions to be reused in the wrong
10364 context thus bypassing client cert protection! All software that uses
10365 client certs and session caches in multiple contexts NEEDS PATCHING to
10366 allow session reuse! A fuller solution is in the works.
10367 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10369 *) Some more source tree cleanups (removed obsolete files
10370 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10371 permission on "config" script to be executable) and a fix for the INSTALL
10373 [Ulf Moeller <ulf@fitug.de>]
10375 *) Remove some legacy and erroneous uses of malloc, free instead of
10377 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10379 *) Make rsa_oaep_test return non-zero on error.
10380 [Ulf Moeller <ulf@fitug.de>]
10382 *) Add support for native Solaris shared libraries. Configure
10383 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10384 if someone would make that last step automatic.
10385 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10387 *) ctx_size was not built with the right compiler during "make links". Fixed.
10390 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10391 except NULL ciphers". This means the default cipher list will no longer
10392 enable NULL ciphers. They need to be specifically enabled e.g. with
10393 the string "DEFAULT:eNULL".
10396 *) Fix to RSA private encryption routines: if p < q then it would
10397 occasionally produce an invalid result. This will only happen with
10398 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10401 *) Be less restrictive and allow also `perl util/perlpath.pl
10402 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10403 because this way one can also use an interpreter named `perl5' (which is
10404 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10405 installed as `perl').
10406 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10408 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10409 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10411 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10412 advapi32.lib to Win32 build and change the pem test comparison
10413 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10414 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10415 and crypto/des/ede_cbcm_enc.c.
10418 *) DES quad checksum was broken on big-endian architectures. Fixed.
10421 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10422 Win32 test batch file so it (might) work again. The Win32 test batch file
10423 is horrible: I feel ill....
10426 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10427 in e_os.h. Audit of header files to check ANSI and non ANSI
10428 sections: 10 functions were absent from non ANSI section and not exported
10429 from Windows DLLs. Fixed up libeay.num for new functions.
10432 *) Make `openssl version' output lines consistent.
10433 [Ralf S. Engelschall]
10435 *) Fix Win32 symbol export lists for BIO functions: Added
10436 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10437 to ms/libeay{16,32}.def.
10438 [Ralf S. Engelschall]
10440 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10441 fine under Unix and passes some trivial tests I've now added. But the
10442 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10443 added to make sure no one expects that this stuff really works in the
10444 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10445 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10447 [Ralf S. Engelschall]
10449 *) Fix the generation of two part addresses in perl.
10450 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10452 *) Add config entry for Linux on MIPS.
10453 [John Tobey <jtobey@channel1.com>]
10455 *) Make links whenever Configure is run, unless we are on Windoze.
10458 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10459 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10463 *) Add a useful kludge to allow package maintainers to specify compiler and
10464 other platforms details on the command line without having to patch the
10465 Configure script everytime: One now can use ``perl Configure
10466 <id>:<details>'', i.e. platform ids are allowed to have details appended
10467 to them (separated by colons). This is treated as there would be a static
10468 pre-configured entry in Configure's %table under key <id> with value
10469 <details> and ``perl Configure <id>'' is called. So, when you want to
10470 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10471 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10472 now, which overrides the FreeBSD-elf entry on-the-fly.
10473 [Ralf S. Engelschall]
10475 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10478 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10479 on the `perl Configure ...' command line. This way one can compile
10480 OpenSSL libraries with Position Independent Code (PIC) which is needed
10481 for linking it into DSOs.
10482 [Ralf S. Engelschall]
10484 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10488 *) Cleaned up the LICENSE document: The official contact for any license
10489 questions now is the OpenSSL core team under openssl-core@openssl.org.
10490 And add a paragraph about the dual-license situation to make sure people
10491 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10492 to the OpenSSL toolkit.
10493 [Ralf S. Engelschall]
10495 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10496 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10497 Additionally cleaned up the `make links' target: Remove unnecessary
10498 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10499 to speed processing and no longer clutter the display with confusing
10500 stuff. Instead only the actually done links are displayed.
10501 [Ralf S. Engelschall]
10503 *) Permit null encryption ciphersuites, used for authentication only. It used
10504 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10505 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10509 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10510 signed attributes when verifying signatures (this would break them),
10511 the detached data encoding was wrong and public keys obtained using
10512 X509_get_pubkey() weren't freed.
10515 *) Add text documentation for the BUFFER functions. Also added a work around
10516 to a Win95 console bug. This was triggered by the password read stuff: the
10517 last character typed gets carried over to the next fread(). If you were
10518 generating a new cert request using 'req' for example then the last
10519 character of the passphrase would be CR which would then enter the first
10523 *) Added the new `Includes OpenSSL Cryptography Software' button as
10524 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10525 button and can be used by applications based on OpenSSL to show the
10526 relationship to the OpenSSL project.
10527 [Ralf S. Engelschall]
10529 *) Remove confusing variables in function signatures in files
10530 ssl/ssl_lib.c and ssl/ssl.h.
10531 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10533 *) Don't install bss_file.c under PREFIX/include/
10534 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10536 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10537 functions that return function pointers and has support for NT specific
10538 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10539 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10540 unsigned to signed types: this was killing the Win32 compile.
10543 *) Add new certificate file to stack functions,
10544 SSL_add_dir_cert_subjects_to_stack() and
10545 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10546 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10547 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10548 This means that Apache-SSL and similar packages don't have to mess around
10549 to add as many CAs as they want to the preferred list.
10552 *) Experiment with doxygen documentation. Currently only partially applied to
10554 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10555 openssl.doxy as the configuration file.
10558 *) Get rid of remaining C++-style comments which strict C compilers hate.
10559 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10561 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10562 compiled in by default: it has problems with large keys.
10565 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10566 DH private keys and/or callback functions which directly correspond to
10567 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10568 is needed for applications which have to configure certificates on a
10569 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10571 For the RSA certificate situation is makes no difference, but
10572 for the DSA certificate situation this fixes the "no shared cipher"
10573 problem where the OpenSSL cipher selection procedure failed because the
10574 temporary keys were not overtaken from the context and the API provided
10575 no way to reconfigure them.
10576 The new functions now let applications reconfigure the stuff and they
10577 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10578 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10579 non-public-API function ssl_cert_instantiate() is used as a helper
10580 function and also to reduce code redundancy inside ssl_rsa.c.
10581 [Ralf S. Engelschall]
10583 *) Move s_server -dcert and -dkey options out of the undocumented feature
10584 area because they are useful for the DSA situation and should be
10585 recognized by the users.
10586 [Ralf S. Engelschall]
10588 *) Fix the cipher decision scheme for export ciphers: the export bits are
10589 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10590 SSL_EXP_MASK. So, the original variable has to be used instead of the
10591 already masked variable.
10592 [Richard Levitte <levitte@stacken.kth.se>]
10594 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10595 [Richard Levitte <levitte@stacken.kth.se>]
10597 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10598 from `int' to `unsigned int' because it's a length and initialized by
10599 EVP_DigestFinal() which expects an `unsigned int *'.
10600 [Richard Levitte <levitte@stacken.kth.se>]
10602 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10603 script. Instead use the usual Shell->Perl transition trick.
10604 [Ralf S. Engelschall]
10606 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10607 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10608 -noout -modulus' as it's already the case for `openssl rsa -noout
10609 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10610 currently the public key is printed (a decision which was already done by
10611 `openssl dsa -modulus' in the past) which serves a similar purpose.
10612 Additionally the NO_RSA no longer completely removes the whole -modulus
10613 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10615 [Ralf S. Engelschall]
10617 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10618 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10619 [Arne Ansper <arne@ats.cyber.ee>]
10621 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10622 to be added. Now both 'req' and 'ca' can use new objects defined in the
10626 *) Add cool BIO that does syslog (or event log on NT).
10627 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10629 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10630 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10631 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10632 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10635 *) Add preliminary config info for new extension code.
10638 *) Make RSA_NO_PADDING really use no padding.
10639 [Ulf Moeller <ulf@fitug.de>]
10641 *) Generate errors when private/public key check is done.
10644 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10645 for some CRL extensions and new objects added.
10648 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10649 key usage extension and fuller support for authority key id.
10652 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10653 padding method for RSA, which is recommended for new applications in PKCS
10654 #1 v2.0 (RFC 2437, October 1998).
10655 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10656 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10657 against Bleichbacher's attack on RSA.
10658 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10661 *) Updates to the new SSL compression code
10662 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10664 *) Fix so that the version number in the master secret, when passed
10665 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10666 (because the server will not accept higher), that the version number
10667 is 0x03,0x01, not 0x03,0x00
10668 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10670 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10671 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10672 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10675 *) Support for RAW extensions where an arbitrary extension can be
10676 created by including its DER encoding. See apps/openssl.cnf for
10680 *) Make sure latest Perl versions don't interpret some generated C array
10681 code as Perl array code in the crypto/err/err_genc.pl script.
10682 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10684 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10685 not many people have the assembler. Various Win32 compilation fixes and
10686 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10687 build instructions.
10690 *) Modify configure script 'Configure' to automatically create crypto/date.h
10691 file under Win32 and also build pem.h from pem.org. New script
10692 util/mkfiles.pl to create the MINFO file on environments that can't do a
10693 'make files': perl util/mkfiles.pl >MINFO should work.
10696 *) Major rework of DES function declarations, in the pursuit of correctness
10697 and purity. As a result, many evil casts evaporated, and some weirdness,
10698 too. You may find this causes warnings in your code. Zapping your evil
10699 casts will probably fix them. Mostly.
10702 *) Fix for a typo in asn1.h. Bug fix to object creation script
10703 obj_dat.pl. It considered a zero in an object definition to mean
10704 "end of object": none of the objects in objects.h have any zeros
10705 so it wasn't spotted.
10706 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10708 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10709 Masking (CBCM). In the absence of test vectors, the best I have been able
10710 to do is check that the decrypt undoes the encrypt, so far. Send me test
10711 vectors if you have them.
10714 *) Correct calculation of key length for export ciphers (too much space was
10715 allocated for null ciphers). This has not been tested!
10718 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10719 message is now correct (it understands "crypto" and "ssl" on its
10720 command line). There is also now an "update" option. This will update
10721 the util/ssleay.num and util/libeay.num files with any new functions.
10723 perl util/mkdef.pl crypto ssl update
10724 it will update them.
10727 *) Overhauled the Perl interface (perl/*):
10728 - ported BN stuff to OpenSSL's different BN library
10729 - made the perl/ source tree CVS-aware
10730 - renamed the package from SSLeay to OpenSSL (the files still contain
10731 their history because I've copied them in the repository)
10732 - removed obsolete files (the test scripts will be replaced
10733 by better Test::Harness variants in the future)
10734 [Ralf S. Engelschall]
10736 *) First cut for a very conservative source tree cleanup:
10737 1. merge various obsolete readme texts into doc/ssleay.txt
10738 where we collect the old documents and readme texts.
10739 2. remove the first part of files where I'm already sure that we no
10740 longer need them because of three reasons: either they are just temporary
10741 files which were left by Eric or they are preserved original files where
10742 I've verified that the diff is also available in the CVS via "cvs diff
10743 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10744 the crypto/md/ stuff).
10745 [Ralf S. Engelschall]
10747 *) More extension code. Incomplete support for subject and issuer alt
10748 name, issuer and authority key id. Change the i2v function parameters
10749 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10750 what that's for :-) Fix to ASN1 macro which messed up
10751 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10754 *) Preliminary support for ENUMERATED type. This is largely copied from the
10758 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10759 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10761 *) Make sure `make rehash' target really finds the `openssl' program.
10762 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10764 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10765 like to hear about it if this slows down other processors.
10768 *) Add CygWin32 platform information to Configure script.
10769 [Alan Batie <batie@aahz.jf.intel.com>]
10771 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10772 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10774 *) New program nseq to manipulate netscape certificate sequences
10777 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10781 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10782 but the BN code had some problems that would cause failures when
10783 doing certificate verification and some other functions.
10784 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10786 *) Add ASN1 and PEM code to support netscape certificate sequences.
10789 *) Add ASN1 and PEM code to support netscape certificate sequences.
10792 *) Add several PKIX and private extended key usage OIDs.
10795 *) Modify the 'ca' program to handle the new extension code. Modify
10796 openssl.cnf for new extension format, add comments.
10799 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10800 and add a sample to openssl.cnf so req -x509 now adds appropriate
10804 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10805 error code, add initial support to X509_print() and x509 application.
10808 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10809 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10810 stuff is currently isolated and isn't even compiled yet.
10813 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10814 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10815 Removed the versions check from X509 routines when loading extensions:
10816 this allows certain broken certificates that don't set the version
10817 properly to be processed.
10820 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10821 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10822 can still be regenerated with "make depend".
10825 *) Spelling mistake in C version of CAST-128.
10826 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10828 *) Changes to the error generation code. The perl script err-code.pl
10829 now reads in the old error codes and retains the old numbers, only
10830 adding new ones if necessary. It also only changes the .err files if new
10831 codes are added. The makefiles have been modified to only insert errors
10832 when needed (to avoid needlessly modifying header files). This is done
10833 by only inserting errors if the .err file is newer than the auto generated
10834 C file. To rebuild all the error codes from scratch (the old behaviour)
10835 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10836 or delete all the .err files.
10839 *) CAST-128 was incorrectly implemented for short keys. The C version has
10840 been fixed, but is untested. The assembler versions are also fixed, but
10841 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10842 to regenerate it if needed.
10843 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10844 Hagino <itojun@kame.net>]
10846 *) File was opened incorrectly in randfile.c.
10847 [Ulf Möller <ulf@fitug.de>]
10849 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10850 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10851 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10852 al: it's just almost always a UTCTime. Note this patch adds new error
10853 codes so do a "make errors" if there are problems.
10856 *) Correct Linux 1 recognition in config.
10857 [Ulf Möller <ulf@fitug.de>]
10859 *) Remove pointless MD5 hash when using DSA keys in ca.
10860 [Anonymous <nobody@replay.com>]
10862 *) Generate an error if given an empty string as a cert directory. Also
10863 generate an error if handed NULL (previously returned 0 to indicate an
10864 error, but didn't set one).
10865 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10867 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10870 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10871 parameters. This was causing a warning which killed off the Win32 compile.
10874 *) Remove C++ style comments from crypto/bn/bn_local.h.
10875 [Neil Costigan <neil.costigan@celocom.com>]
10877 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10878 based on a text string, looking up short and long names and finally
10879 "dot" format. The "dot" format stuff didn't work. Added new function
10880 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10881 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10882 OID is not part of the table.
10885 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10886 X509_LOOKUP_by_alias().
10889 *) Sort openssl functions by name.
10892 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10893 encryption from sample DSA keys (in case anyone is interested the password
10897 *) Make _all_ *_free functions accept a NULL pointer.
10898 [Frans Heymans <fheymans@isaserver.be>]
10900 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10902 [Anonymous <nobody@replay.com>]
10904 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10905 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10907 *) Don't blow it for numeric -newkey arguments to apps/req.
10908 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10910 *) Temp key "for export" tests were wrong in s3_srvr.c.
10911 [Anonymous <nobody@replay.com>]
10913 *) Add prototype for temp key callback functions
10914 SSL_CTX_set_tmp_{rsa,dh}_callback().
10917 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10918 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10921 *) X509_name_add_entry() freed the wrong thing after an error.
10922 [Arne Ansper <arne@ats.cyber.ee>]
10924 *) rsa_eay.c would attempt to free a NULL context.
10925 [Arne Ansper <arne@ats.cyber.ee>]
10927 *) BIO_s_socket() had a broken should_retry() on Windoze.
10928 [Arne Ansper <arne@ats.cyber.ee>]
10930 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10931 [Arne Ansper <arne@ats.cyber.ee>]
10933 *) Make sure the already existing X509_STORE->depth variable is initialized
10934 in X509_STORE_new(), but document the fact that this variable is still
10935 unused in the certificate verification process.
10936 [Ralf S. Engelschall]
10938 *) Fix the various library and apps files to free up pkeys obtained from
10939 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10942 *) Fix reference counting in X509_PUBKEY_get(). This makes
10943 demos/maurice/example2.c work, amongst others, probably.
10944 [Steve Henson and Ben Laurie]
10946 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10947 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10948 are no longer created. This way we have a single and consistent command
10949 line interface `openssl <command>', similar to `cvs <command>'.
10950 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10952 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10953 BIT STRING wrapper always have zero unused bits.
10956 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10959 *) Make the top-level INSTALL documentation easier to understand.
10962 *) Makefiles updated to exit if an error occurs in a sub-directory
10963 make (including if user presses ^C) [Paul Sutton]
10965 *) Make Montgomery context stuff explicit in RSA data structure.
10968 *) Fix build order of pem and err to allow for generated pem.h.
10971 *) Fix renumbering bug in X509_NAME_delete_entry().
10974 *) Enhanced the err-ins.pl script so it makes the error library number
10975 global and can add a library name. This is needed for external ASN1 and
10976 other error libraries.
10979 *) Fixed sk_insert which never worked properly.
10982 *) Fix ASN1 macros so they can handle indefinite length constructed
10983 EXPLICIT tags. Some non standard certificates use these: they can now
10987 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10988 into a single doc/ssleay.txt bundle. This way the information is still
10989 preserved but no longer messes up this directory. Now it's new room for
10990 the new set of documentation files.
10991 [Ralf S. Engelschall]
10993 *) SETs were incorrectly DER encoded. This was a major pain, because they
10994 shared code with SEQUENCEs, which aren't coded the same. This means that
10995 almost everything to do with SETs or SEQUENCEs has either changed name or
10996 number of arguments.
10997 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10999 *) Fix test data to work with the above.
11002 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11003 was already fixed by Eric for 0.9.1 it seems.
11004 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11006 *) Autodetect FreeBSD3.
11009 *) Fix various bugs in Configure. This affects the following platforms:
11013 unixware-2.0-pentium
11017 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11018 before they are needed.
11021 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11025 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11027 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11028 changed SSLeay to OpenSSL in version strings.
11029 [Ralf S. Engelschall]
11031 *) Some fixups to the top-level documents.
11034 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11035 because the symlink to include/ was missing.
11036 [Ralf S. Engelschall]
11038 *) Incorporated the popular no-RSA/DSA-only patches
11039 which allow to compile a RSA-free SSLeay.
11040 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11042 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11043 when "ssleay" is still not found.
11044 [Ralf S. Engelschall]
11046 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11047 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11049 *) Updated the README file.
11050 [Ralf S. Engelschall]
11052 *) Added various .cvsignore files in the CVS repository subdirs
11053 to make a "cvs update" really silent.
11054 [Ralf S. Engelschall]
11056 *) Recompiled the error-definition header files and added
11057 missing symbols to the Win32 linker tables.
11058 [Ralf S. Engelschall]
11060 *) Cleaned up the top-level documents;
11061 o new files: CHANGES and LICENSE
11062 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11063 o merged COPYRIGHT into LICENSE
11064 o removed obsolete TODO file
11065 o renamed MICROSOFT to INSTALL.W32
11066 [Ralf S. Engelschall]
11068 *) Removed dummy files from the 0.9.1b source tree:
11069 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11070 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11071 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11072 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11073 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11074 [Ralf S. Engelschall]
11076 *) Added various platform portability fixes.
11079 *) The Genesis of the OpenSSL rpject:
11080 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11081 Young and Tim J. Hudson created while they were working for C2Net until
11083 [The OpenSSL Project]
11086 Changes between 0.9.0b and 0.9.1b [not released]
11088 *) Updated a few CA certificates under certs/
11091 *) Changed some BIGNUM api stuff.
11094 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11095 DGUX x86, Linux Alpha, etc.
11098 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11099 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11103 *) Add -strparse option to asn1pars program which parses nested
11105 [Dr Stephen Henson <shenson@bigfoot.com>]
11107 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11110 *) DSA fix for "ca" program.
11113 *) Added "-genkey" option to "dsaparam" program.
11116 *) Added RIPE MD160 (rmd160) message digest.
11119 *) Added -a (all) option to "ssleay version" command.
11122 *) Added PLATFORM define which is the id given to Configure.
11125 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11128 *) Extended the ASN.1 parser routines.
11131 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11134 *) Added a BN_CTX to the BN library.
11137 *) Fixed the weak key values in DES library
11140 *) Changed API in EVP library for cipher aliases.
11143 *) Added support for RC2/64bit cipher.
11146 *) Converted the lhash library to the crypto/mem.c functions.
11149 *) Added more recognized ASN.1 object ids.
11152 *) Added more RSA padding checks for SSL/TLS.
11155 *) Added BIO proxy/filter functionality.
11158 *) Added extra_certs to SSL_CTX which can be used
11159 send extra CA certificates to the client in the CA cert chain sending
11160 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11163 *) Now Fortezza is denied in the authentication phase because
11164 this is key exchange mechanism is not supported by SSLeay at all.
11167 *) Additional PKCS1 checks.
11170 *) Support the string "TLSv1" for all TLS v1 ciphers.
11173 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11174 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11177 *) Fixed a few memory leaks.
11180 *) Fixed various code and comment typos.
11183 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11184 bytes sent in the client random.
11185 [Edward Bishop <ebishop@spyglass.com>]