5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
8 Access to deprecated functions can be re-enabled by running config with
9 "enable-deprecated". In addition applications wishing to use deprecated
10 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
11 will, by default, disable some transitive includes that previously existed
12 in the header files (e.g. ec.h will no longer, by default, include bn.h)
15 *) Added support for OCB mode. OpenSSL has been granted a patent license
16 compatible with the OpenSSL license for use of OCB. Details are available
17 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
18 for OCB can be removed by calling config with no-ocb.
21 *) SSLv2 support has been removed. It still supports receiving a SSLv2
22 compatible client hello.
25 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
26 done while fixing the error code for the key-too-small case.
27 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
29 *) Remove various unsupported platforms:
35 Sinix/ReliantUNIX RM400
42 *) Experimental support for a new, fast, unbiased prime candidate generator,
43 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
44 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
46 *) New output format NSS in the sess_id command line tool. This allows
47 exporting the session id and the master key in NSS keylog format.
48 [Martin Kaiser <martin@kaiser.cx>]
50 *) Harmonize version and its documentation. -f flag is used to display
52 [mancha <mancha1@zoho.com>]
54 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
56 [mancha <mancha1@zoho.com>]
58 *) Fix some double frees. These are not thought to be exploitable.
59 [mancha <mancha1@zoho.com>]
61 *) A missing bounds check in the handling of the TLS heartbeat extension
62 can be used to reveal up to 64k of memory to a connected client or
65 Thanks for Neel Mehta of Google Security for discovering this bug and to
66 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
67 preparing the fix (CVE-2014-0160)
68 [Adam Langley, Bodo Moeller]
70 *) Fix for the attack described in the paper "Recovering OpenSSL
71 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
72 by Yuval Yarom and Naomi Benger. Details can be obtained from:
73 http://eprint.iacr.org/2014/140
75 Thanks to Yuval Yarom and Naomi Benger for discovering this
76 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
77 [Yuval Yarom and Naomi Benger]
79 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
80 this fixes a limitation in previous versions of OpenSSL.
83 *) Experimental encrypt-then-mac support.
85 Experimental support for encrypt then mac from
86 draft-gutmann-tls-encrypt-then-mac-02.txt
88 To enable it set the appropriate extension number (0x42 for the test
89 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
91 For non-compliant peers (i.e. just about everything) this should have no
94 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
98 *) Add EVP support for key wrapping algorithms, to avoid problems with
99 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
100 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
101 algorithms and include tests cases.
104 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
108 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
109 MGF1 digest and OAEP label.
112 *) Make openssl verify return errors.
113 [Chris Palmer <palmer@google.com> and Ben Laurie]
115 *) New function ASN1_TIME_diff to calculate the difference between two
116 ASN1_TIME structures or one structure and the current time.
119 *) Update fips_test_suite to support multiple command line options. New
120 test to induce all self test errors in sequence and check expected
124 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
125 sign or verify all in one operation.
128 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
129 test programs and fips_test_suite. Includes functionality to parse
130 the minimal script output of fipsalgest.pl directly.
133 *) Add authorisation parameter to FIPS_module_mode_set().
136 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
139 *) Use separate DRBG fields for internal and external flags. New function
140 FIPS_drbg_health_check() to perform on demand health checking. Add
141 generation tests to fips_test_suite with reduced health check interval to
142 demonstrate periodic health checking. Add "nodh" option to
143 fips_test_suite to skip very slow DH test.
146 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
150 *) More extensive health check for DRBG checking many more failure modes.
151 New function FIPS_selftest_drbg_all() to handle every possible DRBG
152 combination: call this in fips_test_suite.
155 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
156 and POST to handle Dual EC cases.
159 *) Add support for canonical generation of DSA parameter 'g'. See
162 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
163 POST to handle HMAC cases.
166 *) Add functions FIPS_module_version() and FIPS_module_version_text()
167 to return numerical and string versions of the FIPS module number.
170 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
171 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
172 outside the validated module in the FIPS capable OpenSSL.
175 *) Minor change to DRBG entropy callback semantics. In some cases
176 there is no multiple of the block length between min_len and
177 max_len. Allow the callback to return more than max_len bytes
178 of entropy but discard any extra: it is the callback's responsibility
179 to ensure that the extra data discarded does not impact the
180 requested amount of entropy.
183 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
184 information in FIPS186-3, SP800-57 and SP800-131A.
187 *) CCM support via EVP. Interface is very similar to GCM case except we
188 must supply all data in one chunk (i.e. no update, final) and the
189 message length must be supplied if AAD is used. Add algorithm test
193 *) Initial version of POST overhaul. Add POST callback to allow the status
194 of POST to be monitored and/or failures induced. Modify fips_test_suite
195 to use callback. Always run all selftests even if one fails.
198 *) XTS support including algorithm test driver in the fips_gcmtest program.
199 Note: this does increase the maximum key length from 32 to 64 bytes but
200 there should be no binary compatibility issues as existing applications
201 will never use XTS mode.
204 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
205 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
206 performs algorithm blocking for unapproved PRNG types. Also do not
207 set PRNG type in FIPS_mode_set(): leave this to the application.
208 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
209 the standard OpenSSL PRNG: set additional data to a date time vector.
212 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
213 This shouldn't present any incompatibility problems because applications
214 shouldn't be using these directly and any that are will need to rethink
215 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
218 *) Extensive self tests and health checking required by SP800-90 DRBG.
219 Remove strength parameter from FIPS_drbg_instantiate and always
220 instantiate at maximum supported strength.
223 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
226 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
229 *) New function DH_compute_key_padded() to compute a DH key and pad with
230 leading zeroes if needed: this complies with SP800-56A et al.
233 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
234 anything, incomplete, subject to change and largely untested at present.
237 *) Modify fipscanisteronly build option to only build the necessary object
238 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
241 *) Add experimental option FIPSSYMS to give all symbols in
242 fipscanister.o and FIPS or fips prefix. This will avoid
243 conflicts with future versions of OpenSSL. Add perl script
244 util/fipsas.pl to preprocess assembly language source files
245 and rename any affected symbols.
248 *) Add selftest checks and algorithm block of non-fips algorithms in
249 FIPS mode. Remove DES2 from selftests.
252 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
253 return internal method without any ENGINE dependencies. Add new
254 tiny fips sign and verify functions.
257 *) New build option no-ec2m to disable characteristic 2 code.
260 *) New build option "fipscanisteronly". This only builds fipscanister.o
261 and (currently) associated fips utilities. Uses the file Makefile.fips
262 instead of Makefile.org as the prototype.
265 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
266 Update fips_gcmtest to use IV generator.
269 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
270 setting output buffer to NULL. The *Final function must be
271 called although it will not retrieve any additional data. The tag
272 can be set or retrieved with a ctrl. The IV length is by default 12
273 bytes (96 bits) but can be set to an alternative value. If the IV
274 length exceeds the maximum IV length (currently 16 bytes) it cannot be
278 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
279 underlying do_cipher function handles all cipher semantics itself
280 including padding and finalisation. This is useful if (for example)
281 an ENGINE cipher handles block padding itself. The behaviour of
282 do_cipher is subtly changed if this flag is set: the return value
283 is the number of characters written to the output buffer (zero is
284 no longer an error code) or a negative error code. Also if the
285 input buffer is NULL and length 0 finalisation should be performed.
288 *) If a candidate issuer certificate is already part of the constructed
289 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
292 *) Improve forward-security support: add functions
294 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
295 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
297 for use by SSL/TLS servers; the callback function will be called whenever a
298 new session is created, and gets to decide whether the session may be
299 cached to make it resumable (return 0) or not (return 1). (As by the
300 SSL/TLS protocol specifications, the session_id sent by the server will be
301 empty to indicate that the session is not resumable; also, the server will
302 not generate RFC 4507 (RFC 5077) session tickets.)
304 A simple reasonable callback implementation is to return is_forward_secure.
305 This parameter will be set to 1 or 0 depending on the ciphersuite selected
306 by the SSL/TLS server library, indicating whether it can provide forward
308 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
310 *) New -verify_name option in command line utilities to set verification
314 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
315 Add CMAC pkey methods.
318 *) Experimental renegotiation in s_server -www mode. If the client
319 browses /reneg connection is renegotiated. If /renegcert it is
320 renegotiated requesting a certificate.
323 *) Add an "external" session cache for debugging purposes to s_server. This
324 should help trace issues which normally are only apparent in deployed
325 multi-process servers.
328 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
329 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
330 BIO_set_cipher() and some obscure PEM functions were changed so they
331 can now return an error. The RAND changes required a change to the
332 RAND_METHOD structure.
335 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
336 a gcc attribute to warn if the result of a function is ignored. This
337 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
338 whose return value is often ignored.
341 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
343 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
344 ARMv5 through ARMv8, as opposite to "locking" it to single one.
345 So far those who have to target multiple plaforms would compromise
346 and argue that binary targeting say ARMv5 would still execute on
347 ARMv8. "Universal" build resolves this compromise by providing
348 near-optimal performance even on newer platforms.
351 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
352 (other platforms pending).
353 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
355 *) Add support for the SignedCertificateTimestampList certificate and
356 OCSP response extensions from RFC6962.
359 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
360 for corner cases. (Certain input points at infinity could lead to
361 bogus results, with non-infinity inputs mapped to infinity too.)
364 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
365 This covers AES, SHA256/512 and GHASH. "Initial" means that most
366 common cases are optimized and there still is room for further
367 improvements. Vector Permutation AES for Altivec is also added.
370 *) Add support for little-endian ppc64 Linux target.
371 [Marcelo Cerri (IBM)]
373 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
374 SHA1, SHA256 and GHASH. "Initial" means that most common cases
375 are optimized and there still is room for further improvements.
376 Both 32- and 64-bit modes are supported.
377 [Andy Polyakov, Ard Biesheuvel (Linaro)]
379 *) Improved ARMv7 NEON support.
382 *) Support for SPARC Architecture 2011 crypto extensions, first
383 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
384 SHA256/512, MD5, GHASH and modular exponentiation.
385 [Andy Polyakov, David Miller]
387 *) Accelerated modular exponentiation for Intel processors, a.k.a.
389 [Shay Gueron & Vlad Krasnov (Intel Corp)]
391 *) Support for new and upcoming Intel processors, including AVX2,
392 BMI and SHA ISA extensions. This includes additional "stitched"
393 implementations, AESNI-SHA256 and GCM, and multi-buffer support
396 This work was sponsored by Intel Corp.
399 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
400 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
401 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
404 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
405 this fixes a limiation in previous versions of OpenSSL.
408 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
409 MGF1 digest and OAEP label.
412 *) Add EVP support for key wrapping algorithms, to avoid problems with
413 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
414 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
415 algorithms and include tests cases.
418 *) Add functions to allocate and set the fields of an ECDSA_METHOD
420 [Douglas E. Engert, Steve Henson]
422 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
423 difference in days and seconds between two tm or ASN1_TIME structures.
426 *) Add -rev test option to s_server to just reverse order of characters
427 received by client and send back to server. Also prints an abbreviated
428 summary of the connection parameters.
431 *) New option -brief for s_client and s_server to print out a brief summary
432 of connection parameters.
435 *) Add callbacks for arbitrary TLS extensions.
436 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
438 *) New option -crl_download in several openssl utilities to download CRLs
439 from CRLDP extension in certificates.
442 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
445 *) New function X509_CRL_diff to generate a delta CRL from the difference
446 of two full CRLs. Add support to "crl" utility.
449 *) New functions to set lookup_crls function and to retrieve
450 X509_STORE from X509_STORE_CTX.
453 *) Print out deprecated issuer and subject unique ID fields in
457 *) Extend OCSP I/O functions so they can be used for simple general purpose
458 HTTP as well as OCSP. New wrapper function which can be used to download
459 CRLs using the OCSP API.
462 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
465 *) SSL_CONF* functions. These provide a common framework for application
466 configuration using configuration files or command lines.
469 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
470 message callback and prints the results. Needs compile time option
471 "enable-ssl-trace". New options to s_client and s_server to enable
475 *) New ctrl and macro to retrieve supported points extensions.
476 Print out extension in s_server and s_client.
479 *) New functions to retrieve certificate signature and signature
483 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
487 *) New Suite B modes for TLS code. These use and enforce the requirements
488 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
489 only use Suite B curves. The Suite B modes can be set by using the
490 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
493 *) New chain verification flags for Suite B levels of security. Check
494 algorithms are acceptable when flags are set in X509_verify_cert.
497 *) Make tls1_check_chain return a set of flags indicating checks passed
498 by a certificate chain. Add additional tests to handle client
499 certificates: checks for matching certificate type and issuer name
503 *) If an attempt is made to use a signature algorithm not in the peer
504 preference list abort the handshake. If client has no suitable
505 signature algorithms in response to a certificate request do not
509 *) If server EC tmp key is not in client preference list abort handshake.
512 *) Add support for certificate stores in CERT structure. This makes it
513 possible to have different stores per SSL structure or one store in
514 the parent SSL_CTX. Include distint stores for certificate chain
515 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
516 to build and store a certificate chain in CERT structure: returing
517 an error if the chain cannot be built: this will allow applications
518 to test if a chain is correctly configured.
520 Note: if the CERT based stores are not set then the parent SSL_CTX
521 store is used to retain compatibility with existing behaviour.
525 *) New function ssl_set_client_disabled to set a ciphersuite disabled
526 mask based on the current session, check mask when sending client
527 hello and checking the requested ciphersuite.
530 *) New ctrls to retrieve and set certificate types in a certificate
531 request message. Print out received values in s_client. If certificate
532 types is not set with custom values set sensible values based on
533 supported signature algorithms.
536 *) Support for distinct client and server supported signature algorithms.
539 *) Add certificate callback. If set this is called whenever a certificate
540 is required by client or server. An application can decide which
541 certificate chain to present based on arbitrary criteria: for example
542 supported signature algorithms. Add very simple example to s_server.
543 This fixes many of the problems and restrictions of the existing client
544 certificate callback: for example you can now clear an existing
545 certificate and specify the whole chain.
548 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
549 the certificate can be used for (if anything). Set valid_flags field
550 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
551 to have similar checks in it.
553 Add new "cert_flags" field to CERT structure and include a "strict mode".
554 This enforces some TLS certificate requirements (such as only permitting
555 certificate signature algorithms contained in the supported algorithms
556 extension) which some implementations ignore: this option should be used
557 with caution as it could cause interoperability issues.
560 *) Update and tidy signature algorithm extension processing. Work out
561 shared signature algorithms based on preferences and peer algorithms
562 and print them out in s_client and s_server. Abort handshake if no
563 shared signature algorithms.
566 *) Add new functions to allow customised supported signature algorithms
567 for SSL and SSL_CTX structures. Add options to s_client and s_server
571 *) New function SSL_certs_clear() to delete all references to certificates
572 from an SSL structure. Before this once a certificate had been added
573 it couldn't be removed.
576 *) Integrate hostname, email address and IP address checking with certificate
577 verification. New verify options supporting checking in opensl utility.
580 *) Fixes and wildcard matching support to hostname and email checking
581 functions. Add manual page.
582 [Florian Weimer (Red Hat Product Security Team)]
584 *) New functions to check a hostname email or IP address against a
585 certificate. Add options x509 utility to print results of checks against
589 *) Fix OCSP checking.
590 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
592 *) Initial experimental support for explicitly trusted non-root CAs.
593 OpenSSL still tries to build a complete chain to a root but if an
594 intermediate CA has a trust setting included that is used. The first
595 setting is used: whether to trust (e.g., -addtrust option to the x509
599 *) Add -trusted_first option which attempts to find certificates in the
600 trusted store even if an untrusted chain is also supplied.
603 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
604 platform support for Linux and Android.
607 *) Support for linux-x32, ILP32 environment in x86_64 framework.
610 *) Experimental multi-implementation support for FIPS capable OpenSSL.
611 When in FIPS mode the approved implementations are used as normal,
612 when not in FIPS mode the internal unapproved versions are used instead.
613 This means that the FIPS capable OpenSSL isn't forced to use the
614 (often lower performance) FIPS implementations outside FIPS mode.
617 *) Transparently support X9.42 DH parameters when calling
618 PEM_read_bio_DHparameters. This means existing applications can handle
619 the new parameter format automatically.
622 *) Initial experimental support for X9.42 DH parameter format: mainly
623 to support use of 'q' parameter for RFC5114 parameters.
626 *) Add DH parameters from RFC5114 including test data to dhtest.
629 *) Support for automatic EC temporary key parameter selection. If enabled
630 the most preferred EC parameters are automatically used instead of
631 hardcoded fixed parameters. Now a server just has to call:
632 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
633 support ECDH and use the most appropriate parameters.
636 *) Enhance and tidy EC curve and point format TLS extension code. Use
637 static structures instead of allocation if default values are used.
638 New ctrls to set curves we wish to support and to retrieve shared curves.
639 Print out shared curves in s_server. New options to s_server and s_client
640 to set list of supported curves.
643 *) New ctrls to retrieve supported signature algorithms and
644 supported curve values as an array of NIDs. Extend openssl utility
645 to print out received values.
648 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
649 between NIDs and the more common NIST names such as "P-256". Enhance
650 ecparam utility and ECC method to recognise the NIST names for curves.
653 *) Enhance SSL/TLS certificate chain handling to support different
654 chains for each certificate instead of one chain in the parent SSL_CTX.
657 *) Support for fixed DH ciphersuite client authentication: where both
658 server and client use DH certificates with common parameters.
661 *) Support for fixed DH ciphersuites: those requiring DH server
665 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
667 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
668 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
669 X509_CINF_get_signature were reverted post internal team review.
671 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
673 *) Abort handshake if server key exchange message is omitted for ephemeral
676 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
677 reporting this issue.
681 *) Remove non-export ephemeral RSA code on client and server. This code
682 violated the TLS standard by allowing the use of temporary RSA keys in
683 non-export ciphersuites and could be used by a server to effectively
684 downgrade the RSA key length used to a value smaller than the server
685 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
686 INRIA or reporting this issue.
690 *) Ensure that the session ID context of an SSL is updated when its
691 SSL_CTX is updated via SSL_set_SSL_CTX.
693 The session ID context is typically set from the parent SSL_CTX,
694 and can vary with the CTX.
697 *) Fix various certificate fingerprint issues.
699 By using non-DER or invalid encodings outside the signed portion of a
700 certificate the fingerprint can be changed without breaking the signature.
701 Although no details of the signed portion of the certificate can be changed
702 this can cause problems with some applications: e.g. those using the
703 certificate fingerprint for blacklists.
705 1. Reject signatures with non zero unused bits.
707 If the BIT STRING containing the signature has non zero unused bits reject
708 the signature. All current signature algorithms require zero unused bits.
710 2. Check certificate algorithm consistency.
712 Check the AlgorithmIdentifier inside TBS matches the one in the
713 certificate signature. NB: this will result in signature failure
714 errors for some broken certificates.
716 Thanks to Konrad Kraszewski from Google for reporting this issue.
718 3. Check DSA/ECDSA signatures use DER.
720 Reencode DSA/ECDSA signatures and compare with the original received
721 signature. Return an error if there is a mismatch.
723 This will reject various cases including garbage after signature
724 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
725 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
726 (negative or with leading zeroes).
728 Further analysis was conducted and fixes were developed by Stephen Henson
729 of the OpenSSL core team.
734 *) Do not resume sessions on the server if the negotiated protocol
735 version does not match the session's version. Resuming with a different
736 version, while not strictly forbidden by the RFC, is of questionable
737 sanity and breaks all known clients.
738 [David Benjamin, Emilia Käsper]
740 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
741 early CCS messages during renegotiation. (Note that because
742 renegotiation is encrypted, this early CCS was not exploitable.)
745 *) Tighten client-side session ticket handling during renegotiation:
746 ensure that the client only accepts a session ticket if the server sends
747 the extension anew in the ServerHello. Previously, a TLS client would
748 reuse the old extension state and thus accept a session ticket if one was
749 announced in the initial ServerHello.
751 Similarly, ensure that the client requires a session ticket if one
752 was advertised in the ServerHello. Previously, a TLS client would
753 ignore a missing NewSessionTicket message.
756 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
760 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
761 sends a carefully crafted handshake message, to cause OpenSSL to fail
762 to free up to 64k of memory causing a memory leak. This could be
763 exploited in a Denial Of Service attack. This issue affects OpenSSL
764 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
765 whether SRTP is used or configured. Implementations of OpenSSL that
766 have been compiled with OPENSSL_NO_SRTP defined are not affected.
768 The fix was developed by the OpenSSL team.
772 *) Session Ticket Memory Leak.
774 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
775 integrity of that ticket is first verified. In the event of a session
776 ticket integrity check failing, OpenSSL will fail to free memory
777 causing a memory leak. By sending a large number of invalid session
778 tickets an attacker could exploit this issue in a Denial Of Service
783 *) Build option no-ssl3 is incomplete.
785 When OpenSSL is configured with "no-ssl3" as a build option, servers
786 could accept and complete a SSL 3.0 handshake, and clients could be
787 configured to send them.
789 [Akamai and the OpenSSL team]
791 *) Add support for TLS_FALLBACK_SCSV.
792 Client applications doing fallback retries should call
793 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
795 [Adam Langley, Bodo Moeller]
797 *) Add additional DigestInfo checks.
799 Reencode DigestInto in DER and check against the original when
800 verifying RSA signature: this will reject any improperly encoded
801 DigestInfo structures.
803 Note: this is a precautionary measure and no attacks are currently known.
807 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
809 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
810 SRP code can be overrun an internal buffer. Add sanity check that
811 g, A, B < N to SRP code.
813 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
814 Group for discovering this issue.
818 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
819 TLS 1.0 instead of higher protocol versions when the ClientHello message
820 is badly fragmented. This allows a man-in-the-middle attacker to force a
821 downgrade to TLS 1.0 even if both the server and the client support a
822 higher protocol version, by modifying the client's TLS records.
824 Thanks to David Benjamin and Adam Langley (Google) for discovering and
825 researching this issue.
829 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
830 to a denial of service attack. A malicious server can crash the client
831 with a null pointer dereference (read) by specifying an anonymous (EC)DH
832 ciphersuite and sending carefully crafted handshake messages.
834 Thanks to Felix Gröbert (Google) for discovering and researching this
839 *) By sending carefully crafted DTLS packets an attacker could cause openssl
840 to leak memory. This can be exploited through a Denial of Service attack.
841 Thanks to Adam Langley for discovering and researching this issue.
845 *) An attacker can force openssl to consume large amounts of memory whilst
846 processing DTLS handshake messages. This can be exploited through a
847 Denial of Service attack.
848 Thanks to Adam Langley for discovering and researching this issue.
852 *) An attacker can force an error condition which causes openssl to crash
853 whilst processing DTLS packets due to memory being freed twice. This
854 can be exploited through a Denial of Service attack.
855 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
860 *) If a multithreaded client connects to a malicious server using a resumed
861 session and the server sends an ec point format extension it could write
862 up to 255 bytes to freed memory.
864 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
869 *) A malicious server can crash an OpenSSL client with a null pointer
870 dereference (read) by specifying an SRP ciphersuite even though it was not
871 properly negotiated with the client. This can be exploited through a
872 Denial of Service attack.
874 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
875 discovering and researching this issue.
879 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
880 X509_name_oneline, X509_name_print_ex et al. to leak some information
881 from the stack. Applications may be affected if they echo pretty printing
882 output to the attacker.
884 Thanks to Ivan Fratric (Google) for discovering this issue.
886 [Emilia Käsper, and Steve Henson]
888 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
889 for corner cases. (Certain input points at infinity could lead to
890 bogus results, with non-infinity inputs mapped to infinity too.)
893 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
895 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
896 handshake can force the use of weak keying material in OpenSSL
897 SSL/TLS clients and servers.
899 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
900 researching this issue. (CVE-2014-0224)
901 [KIKUCHI Masashi, Steve Henson]
903 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
904 OpenSSL DTLS client the code can be made to recurse eventually crashing
907 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
909 [Imre Rad, Steve Henson]
911 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
912 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
913 client or server. This is potentially exploitable to run arbitrary
914 code on a vulnerable client or server.
916 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
917 [Jüri Aedla, Steve Henson]
919 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
920 are subject to a denial of service attack.
922 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
923 this issue. (CVE-2014-3470)
924 [Felix Gröbert, Ivan Fratric, Steve Henson]
926 *) Harmonize version and its documentation. -f flag is used to display
928 [mancha <mancha1@zoho.com>]
930 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
932 [mancha <mancha1@zoho.com>]
934 *) Fix some double frees. These are not thought to be exploitable.
935 [mancha <mancha1@zoho.com>]
937 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
939 *) A missing bounds check in the handling of the TLS heartbeat extension
940 can be used to reveal up to 64k of memory to a connected client or
943 Thanks for Neel Mehta of Google Security for discovering this bug and to
944 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
945 preparing the fix (CVE-2014-0160)
946 [Adam Langley, Bodo Moeller]
948 *) Fix for the attack described in the paper "Recovering OpenSSL
949 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
950 by Yuval Yarom and Naomi Benger. Details can be obtained from:
951 http://eprint.iacr.org/2014/140
953 Thanks to Yuval Yarom and Naomi Benger for discovering this
954 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
955 [Yuval Yarom and Naomi Benger]
957 *) TLS pad extension: draft-agl-tls-padding-03
959 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
960 TLS client Hello record length value would otherwise be > 255 and
961 less that 512 pad with a dummy extension containing zeroes so it
962 is at least 512 bytes long.
964 [Adam Langley, Steve Henson]
966 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
968 *) Fix for TLS record tampering bug. A carefully crafted invalid
969 handshake could crash OpenSSL with a NULL pointer exception.
970 Thanks to Anton Johansson for reporting this issues.
973 *) Keep original DTLS digest and encryption contexts in retransmission
974 structures so we can use the previous session parameters if they need
975 to be resent. (CVE-2013-6450)
978 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
979 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
980 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
981 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
982 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
983 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
984 [Rob Stradling, Adam Langley]
986 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
988 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
989 supporting platforms or when small records were transferred.
990 [Andy Polyakov, Steve Henson]
992 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
994 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
996 This addresses the flaw in CBC record processing discovered by
997 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
998 at: http://www.isg.rhul.ac.uk/tls/
1000 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1001 Security Group at Royal Holloway, University of London
1002 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1003 Emilia Käsper for the initial patch.
1005 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1007 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1008 ciphersuites which can be exploited in a denial of service attack.
1009 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1010 and detecting this bug and to Wolfgang Ettlinger
1011 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1015 *) Return an error when checking OCSP signatures when key is NULL.
1016 This fixes a DoS attack. (CVE-2013-0166)
1019 *) Make openssl verify return errors.
1020 [Chris Palmer <palmer@google.com> and Ben Laurie]
1022 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1023 the right response is stapled. Also change SSL_get_certificate()
1024 so it returns the certificate actually sent.
1025 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1026 [Rob Stradling <rob.stradling@comodo.com>]
1028 *) Fix possible deadlock when decoding public keys.
1031 *) Don't use TLS 1.0 record version number in initial client hello
1035 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1037 *) Sanity check record length before skipping explicit IV in TLS
1038 1.2, 1.1 and DTLS to fix DoS attack.
1040 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1041 fuzzing as a service testing platform.
1045 *) Initialise tkeylen properly when encrypting CMS messages.
1046 Thanks to Solar Designer of Openwall for reporting this issue.
1049 *) In FIPS mode don't try to use composite ciphers as they are not
1053 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1055 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1056 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1057 mean any application compiled against OpenSSL 1.0.0 headers setting
1058 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1059 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1060 0x10000000L Any application which was previously compiled against
1061 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1062 will need to be recompiled as a result. Letting be results in
1063 inability to disable specifically TLS 1.1 and in client context,
1064 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1067 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1068 disable just protocol X, but all protocols above X *if* there are
1069 protocols *below* X still enabled. In more practical terms it means
1070 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1071 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1072 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1076 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1078 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1079 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1080 in CRYPTO_realloc_clean.
1082 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1083 issue and to Adam Langley <agl@chromium.org> for fixing it.
1085 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1087 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1090 *) Workarounds for some broken servers that "hang" if a client hello
1091 record length exceeds 255 bytes.
1093 1. Do not use record version number > TLS 1.0 in initial client
1094 hello: some (but not all) hanging servers will now work.
1095 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1096 the number of ciphers sent in the client hello. This should be
1097 set to an even number, such as 50, for example by passing:
1098 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1099 Most broken servers should now work.
1100 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1101 TLS 1.2 client support entirely.
1104 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1107 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1109 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1110 STRING form instead of a DigestInfo.
1113 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1114 and the RSA_sign/RSA_verify functions. This was made more apparent when
1115 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1116 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1117 the correct format in RSA_verify so both forms transparently work.
1120 *) Some servers which support TLS 1.0 can choke if we initially indicate
1121 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1122 encrypted premaster secret. As a workaround use the maximum permitted
1123 client version in client hello, this should keep such servers happy
1124 and still work with previous versions of OpenSSL.
1127 *) Add support for TLS/DTLS heartbeats.
1128 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1130 *) Add support for SCTP.
1131 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1133 *) Improved PRNG seeding for VOS.
1134 [Paul Green <Paul.Green@stratus.com>]
1136 *) Extensive assembler packs updates, most notably:
1138 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1139 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1140 - x86_64: bit-sliced AES implementation;
1141 - ARM: NEON support, contemporary platforms optimizations;
1142 - s390x: z196 support;
1143 - *: GHASH and GF(2^m) multiplication implementations;
1147 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1148 (removal of unnecessary code)
1149 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1151 *) Add TLS key material exporter from RFC 5705.
1154 *) Add DTLS-SRTP negotiation from RFC 5764.
1157 *) Add Next Protocol Negotiation,
1158 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1159 disabled with a no-npn flag to config or Configure. Code donated
1161 [Adam Langley <agl@google.com> and Ben Laurie]
1163 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1164 NIST-P256, NIST-P521, with constant-time single point multiplication on
1165 typical inputs. Compiler support for the nonstandard type __uint128_t is
1166 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1167 Code made available under Apache License version 2.0.
1169 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1170 line to include this in your build of OpenSSL, and run "make depend" (or
1171 "make update"). This enables the following EC_METHODs:
1173 EC_GFp_nistp224_method()
1174 EC_GFp_nistp256_method()
1175 EC_GFp_nistp521_method()
1177 EC_GROUP_new_by_curve_name() will automatically use these (while
1178 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1180 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1182 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1183 all platforms. Move ssize_t definition from e_os.h to the public
1184 header file e_os2.h as it now appears in public header file cms.h
1187 *) New -sigopt option to the ca, req and x509 utilities. Additional
1188 signature parameters can be passed using this option and in
1192 *) Add RSA PSS signing function. This will generate and set the
1193 appropriate AlgorithmIdentifiers for PSS based on those in the
1194 corresponding EVP_MD_CTX structure. No application support yet.
1197 *) Support for companion algorithm specific ASN1 signing routines.
1198 New function ASN1_item_sign_ctx() signs a pre-initialised
1199 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1200 the appropriate parameters.
1203 *) Add new algorithm specific ASN1 verification initialisation function
1204 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1205 handling will be the same no matter what EVP_PKEY_METHOD is used.
1206 Add a PSS handler to support verification of PSS signatures: checked
1207 against a number of sample certificates.
1210 *) Add signature printing for PSS. Add PSS OIDs.
1211 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1213 *) Add algorithm specific signature printing. An individual ASN1 method
1214 can now print out signatures instead of the standard hex dump.
1216 More complex signatures (e.g. PSS) can print out more meaningful
1217 information. Include DSA version that prints out the signature
1221 *) Password based recipient info support for CMS library: implementing
1225 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1226 neatly separates the code into cipher and PBE sections and is required
1227 for some algorithms that split PBES2 into separate pieces (such as
1228 password based CMS).
1231 *) Session-handling fixes:
1232 - Fix handling of connections that are resuming with a session ID,
1233 but also support Session Tickets.
1234 - Fix a bug that suppressed issuing of a new ticket if the client
1235 presented a ticket with an expired session.
1236 - Try to set the ticket lifetime hint to something reasonable.
1237 - Make tickets shorter by excluding irrelevant information.
1238 - On the client side, don't ignore renewed tickets.
1239 [Adam Langley, Bodo Moeller (Google)]
1241 *) Fix PSK session representation.
1244 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1246 This work was sponsored by Intel.
1249 *) Add GCM support to TLS library. Some custom code is needed to split
1250 the IV between the fixed (from PRF) and explicit (from TLS record)
1251 portions. This adds all GCM ciphersuites supported by RFC5288 and
1252 RFC5289. Generalise some AES* cipherstrings to include GCM and
1253 add a special AESGCM string for GCM only.
1256 *) Expand range of ctrls for AES GCM. Permit setting invocation
1257 field on decrypt and retrieval of invocation field only on encrypt.
1260 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1261 As required by RFC5289 these ciphersuites cannot be used if for
1262 versions of TLS earlier than 1.2.
1265 *) For FIPS capable OpenSSL interpret a NULL default public key method
1266 as unset and return the appropriate default but do *not* set the default.
1267 This means we can return the appropriate method in applications that
1268 switch between FIPS and non-FIPS modes.
1271 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1272 ENGINE is used then we cannot handle that in the FIPS module so we
1273 keep original code iff non-FIPS operations are allowed.
1276 *) Add -attime option to openssl utilities.
1277 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1279 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1282 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1283 FIPS EC methods unconditionally for now.
1286 *) New build option no-ec2m to disable characteristic 2 code.
1289 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1290 all cases can be covered as some introduce binary incompatibilities.
1293 *) Redirect RSA operations to FIPS module including keygen,
1294 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1297 *) Add similar low level API blocking to ciphers.
1300 *) Low level digest APIs are not approved in FIPS mode: any attempt
1301 to use these will cause a fatal error. Applications that *really* want
1302 to use them can use the private_* version instead.
1305 *) Redirect cipher operations to FIPS module for FIPS builds.
1308 *) Redirect digest operations to FIPS module for FIPS builds.
1311 *) Update build system to add "fips" flag which will link in fipscanister.o
1312 for static and shared library builds embedding a signature if needed.
1315 *) Output TLS supported curves in preference order instead of numerical
1316 order. This is currently hardcoded for the highest order curves first.
1317 This should be configurable so applications can judge speed vs strength.
1320 *) Add TLS v1.2 server support for client authentication.
1323 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1327 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1328 FIPS modules versions.
1331 *) Add TLS v1.2 client side support for client authentication. Keep cache
1332 of handshake records longer as we don't know the hash algorithm to use
1333 until after the certificate request message is received.
1336 *) Initial TLS v1.2 client support. Add a default signature algorithms
1337 extension including all the algorithms we support. Parse new signature
1338 format in client key exchange. Relax some ECC signing restrictions for
1339 TLS v1.2 as indicated in RFC5246.
1342 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1343 to new signature format when needed using client digest preference.
1344 All server ciphersuites should now work correctly in TLS v1.2. No client
1345 support yet and no support for client certificates.
1348 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1349 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1350 ciphersuites. At present only RSA key exchange ciphersuites work with
1351 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1352 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1353 and version checking.
1356 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1357 with this defined it will not be affected by any changes to ssl internal
1358 structures. Add several utility functions to allow openssl application
1359 to work with OPENSSL_NO_SSL_INTERN defined.
1363 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1365 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1368 *) Permit abbreviated handshakes when renegotiating using the function
1369 SSL_renegotiate_abbreviated().
1370 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1372 *) Add call to ENGINE_register_all_complete() to
1373 ENGINE_load_builtin_engines(), so some implementations get used
1374 automatically instead of needing explicit application support.
1377 *) Add support for TLS key exporter as described in RFC5705.
1378 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1380 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1381 a few changes are required:
1383 Add SSL_OP_NO_TLSv1_1 flag.
1384 Add TLSv1_1 methods.
1385 Update version checking logic to handle version 1.1.
1386 Add explicit IV handling (ported from DTLS code).
1387 Add command line options to s_client/s_server.
1390 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1392 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1394 This addresses the flaw in CBC record processing discovered by
1395 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1396 at: http://www.isg.rhul.ac.uk/tls/
1398 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1399 Security Group at Royal Holloway, University of London
1400 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1401 Emilia Käsper for the initial patch.
1403 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1405 *) Return an error when checking OCSP signatures when key is NULL.
1406 This fixes a DoS attack. (CVE-2013-0166)
1409 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1410 the right response is stapled. Also change SSL_get_certificate()
1411 so it returns the certificate actually sent.
1412 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1413 (This is a backport)
1414 [Rob Stradling <rob.stradling@comodo.com>]
1416 *) Fix possible deadlock when decoding public keys.
1419 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1421 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1424 *) Sanity check record length before skipping explicit IV in DTLS
1427 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1428 fuzzing as a service testing platform.
1432 *) Initialise tkeylen properly when encrypting CMS messages.
1433 Thanks to Solar Designer of Openwall for reporting this issue.
1436 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1438 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1439 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1440 in CRYPTO_realloc_clean.
1442 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1443 issue and to Adam Langley <agl@chromium.org> for fixing it.
1445 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1447 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1449 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1450 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1451 content decryption and always return the same error. Note: this attack
1452 needs on average 2^20 messages so it only affects automated senders. The
1453 old behaviour can be reenabled in the CMS code by setting the
1454 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1455 an MMA defence is not necessary.
1456 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1457 this issue. (CVE-2012-0884)
1460 *) Fix CVE-2011-4619: make sure we really are receiving a
1461 client hello before rejecting multiple SGC restarts. Thanks to
1462 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1465 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1467 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1468 Thanks to Antonio Martin, Enterprise Secure Access Research and
1469 Development, Cisco Systems, Inc. for discovering this bug and
1470 preparing a fix. (CVE-2012-0050)
1473 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1475 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1476 of the Vaudenay padding oracle attack on CBC mode encryption
1477 which enables an efficient plaintext recovery attack against
1478 the OpenSSL implementation of DTLS. Their attack exploits timing
1479 differences arising during decryption processing. A research
1480 paper describing this attack can be found at:
1481 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1482 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1483 Security Group at Royal Holloway, University of London
1484 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1485 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1486 for preparing the fix. (CVE-2011-4108)
1487 [Robin Seggelmann, Michael Tuexen]
1489 *) Clear bytes used for block padding of SSL 3.0 records.
1491 [Adam Langley (Google)]
1493 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1494 Kadianakis <desnacked@gmail.com> for discovering this issue and
1495 Adam Langley for preparing the fix. (CVE-2011-4619)
1496 [Adam Langley (Google)]
1498 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1499 [Andrey Kulikov <amdeich@gmail.com>]
1501 *) Prevent malformed RFC3779 data triggering an assertion failure.
1502 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1503 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1504 [Rob Austein <sra@hactrn.net>]
1506 *) Improved PRNG seeding for VOS.
1507 [Paul Green <Paul.Green@stratus.com>]
1509 *) Fix ssl_ciph.c set-up race.
1510 [Adam Langley (Google)]
1512 *) Fix spurious failures in ecdsatest.c.
1513 [Emilia Käsper (Google)]
1515 *) Fix the BIO_f_buffer() implementation (which was mixing different
1516 interpretations of the '..._len' fields).
1517 [Adam Langley (Google)]
1519 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1520 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1521 threads won't reuse the same blinding coefficients.
1523 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1524 lock to call BN_BLINDING_invert_ex, and avoids one use of
1525 BN_BLINDING_update for each BN_BLINDING structure (previously,
1526 the last update always remained unused).
1527 [Emilia Käsper (Google)]
1529 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1530 [Bob Buckholz (Google)]
1532 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1534 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1535 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1536 [Kaspar Brand <ossl@velox.ch>]
1538 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1539 for multi-threaded use of ECDH. (CVE-2011-3210)
1540 [Adam Langley (Google)]
1542 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1545 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1546 signature public key algorithm by using OID xref utilities instead.
1547 Before this you could only use some ECC ciphersuites with SHA1 only.
1550 *) Add protection against ECDSA timing attacks as mentioned in the paper
1551 by Billy Bob Brumley and Nicola Tuveri, see:
1553 http://eprint.iacr.org/2011/232.pdf
1555 [Billy Bob Brumley and Nicola Tuveri]
1557 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1559 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1560 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1562 *) Fix bug in string printing code: if *any* escaping is enabled we must
1563 escape the escape character (backslash) or the resulting string is
1567 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1569 *) Disable code workaround for ancient and obsolete Netscape browsers
1570 and servers: an attacker can use it in a ciphersuite downgrade attack.
1571 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1574 *) Fixed J-PAKE implementation error, originally discovered by
1575 Sebastien Martini, further info and confirmation from Stefan
1576 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1579 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1581 *) Fix extension code to avoid race conditions which can result in a buffer
1582 overrun vulnerability: resumed sessions must not be modified as they can
1583 be shared by multiple threads. CVE-2010-3864
1586 *) Fix WIN32 build system to correctly link an ENGINE directory into
1590 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1592 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1594 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1596 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1598 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1599 context. The operation can be customised via the ctrl mechanism in
1600 case ENGINEs want to include additional functionality.
1603 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1606 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1607 output hashes compatible with older versions of OpenSSL.
1608 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1610 *) Fix compression algorithm handling: if resuming a session use the
1611 compression algorithm of the resumed session instead of determining
1612 it from client hello again. Don't allow server to change algorithm.
1615 *) Add load_crls() function to apps tidying load_certs() too. Add option
1616 to verify utility to allow additional CRLs to be included.
1619 *) Update OCSP request code to permit adding custom headers to the request:
1620 some responders need this.
1623 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1625 [Julia Lawall <julia@diku.dk>]
1627 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1628 needlessly dereferenced structures, used obsolete functions and
1629 didn't handle all updated verify codes correctly.
1632 *) Disable MD2 in the default configuration.
1635 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1636 indicate the initial BIO being pushed or popped. This makes it possible
1637 to determine whether the BIO is the one explicitly called or as a result
1638 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1639 it handles reference counts correctly and doesn't zero out the I/O bio
1640 when it is not being explicitly popped. WARNING: applications which
1641 included workarounds for the old buggy behaviour will need to be modified
1642 or they could free up already freed BIOs.
1645 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1646 renaming to all platforms (within the 0.9.8 branch, this was
1647 done conditionally on Netware platforms to avoid a name clash).
1648 [Guenter <lists@gknw.net>]
1650 *) Add ECDHE and PSK support to DTLS.
1651 [Michael Tuexen <tuexen@fh-muenster.de>]
1653 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1657 *) Add "missing" function EVP_MD_flags() (without this the only way to
1658 retrieve a digest flags is by accessing the structure directly. Update
1659 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1660 or cipher is registered as in the "from" argument. Print out all
1661 registered digests in the dgst usage message instead of manually
1662 attempting to work them out.
1665 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1666 this allows the use of compression and extensions. Change default cipher
1667 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1668 by default unless an application cipher string requests it.
1671 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1672 key ids to find matching certificates and keys but some PKCS#12 files
1673 don't follow the (somewhat unwritten) rules and this strategy fails.
1674 Now just gather all certificates together and the first private key
1675 then look for the first certificate that matches the key.
1678 *) Support use of registered digest and cipher names for dgst and cipher
1679 commands instead of having to add each one as a special case. So now
1686 openssl dgst -sha256 foo
1688 and this works for ENGINE based algorithms too.
1692 *) Update Gost ENGINE to support parameter files.
1693 [Victor B. Wagner <vitus@cryptocom.ru>]
1695 *) Support GeneralizedTime in ca utility.
1696 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1698 *) Enhance the hash format used for certificate directory links. The new
1699 form uses the canonical encoding (meaning equivalent names will work
1700 even if they aren't identical) and uses SHA1 instead of MD5. This form
1701 is incompatible with the older format and as a result c_rehash should
1702 be used to rebuild symbolic links.
1705 *) Make PKCS#8 the default write format for private keys, replacing the
1706 traditional format. This form is standardised, more secure and doesn't
1707 include an implicit MD5 dependency.
1710 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1711 committed to OpenSSL should pass this lot as a minimum.
1714 *) Add session ticket override functionality for use by EAP-FAST.
1715 [Jouni Malinen <j@w1.fi>]
1717 *) Modify HMAC functions to return a value. Since these can be implemented
1718 in an ENGINE errors can occur.
1721 *) Type-checked OBJ_bsearch_ex.
1724 *) Type-checked OBJ_bsearch. Also some constification necessitated
1725 by type-checking. Still to come: TXT_DB, bsearch(?),
1726 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1730 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1731 seconds to a tm structure directly, instead of going through OS
1732 specific date routines. This avoids any issues with OS routines such
1733 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1734 and X509_time_adj_ex() to cover the extended range. The existing
1735 X509_time_adj() is still usable and will no longer have any date issues.
1738 *) Delta CRL support. New use deltas option which will attempt to locate
1739 and search any appropriate delta CRLs available.
1741 This work was sponsored by Google.
1744 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1745 code and add additional score elements. Validate alternate CRL paths
1746 as part of the CRL checking and indicate a new error "CRL path validation
1747 error" in this case. Applications wanting additional details can use
1748 the verify callback and check the new "parent" field. If this is not
1749 NULL CRL path validation is taking place. Existing applications wont
1750 see this because it requires extended CRL support which is off by
1753 This work was sponsored by Google.
1756 *) Support for freshest CRL extension.
1758 This work was sponsored by Google.
1761 *) Initial indirect CRL support. Currently only supported in the CRLs
1762 passed directly and not via lookup. Process certificate issuer
1763 CRL entry extension and lookup CRL entries by bother issuer name
1764 and serial number. Check and process CRL issuer entry in IDP extension.
1766 This work was sponsored by Google.
1769 *) Add support for distinct certificate and CRL paths. The CRL issuer
1770 certificate is validated separately in this case. Only enabled if
1771 an extended CRL support flag is set: this flag will enable additional
1772 CRL functionality in future.
1774 This work was sponsored by Google.
1777 *) Add support for policy mappings extension.
1779 This work was sponsored by Google.
1782 *) Fixes to pathlength constraint, self issued certificate handling,
1783 policy processing to align with RFC3280 and PKITS tests.
1785 This work was sponsored by Google.
1788 *) Support for name constraints certificate extension. DN, email, DNS
1789 and URI types are currently supported.
1791 This work was sponsored by Google.
1794 *) To cater for systems that provide a pointer-based thread ID rather
1795 than numeric, deprecate the current numeric thread ID mechanism and
1796 replace it with a structure and associated callback type. This
1797 mechanism allows a numeric "hash" to be extracted from a thread ID in
1798 either case, and on platforms where pointers are larger than 'long',
1799 mixing is done to help ensure the numeric 'hash' is usable even if it
1800 can't be guaranteed unique. The default mechanism is to use "&errno"
1801 as a pointer-based thread ID to distinguish between threads.
1803 Applications that want to provide their own thread IDs should now use
1804 CRYPTO_THREADID_set_callback() to register a callback that will call
1805 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1807 Note that ERR_remove_state() is now deprecated, because it is tied
1808 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1809 to free the current thread's error state should be replaced by
1810 ERR_remove_thread_state(NULL).
1812 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1813 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1814 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1815 application was previously providing a numeric thread callback that
1816 was inappropriate for distinguishing threads, then uniqueness might
1817 have been obtained with &errno that happened immediately in the
1818 intermediate development versions of OpenSSL; this is no longer the
1819 case, the numeric thread callback will now override the automatic use
1821 [Geoff Thorpe, with help from Bodo Moeller]
1823 *) Initial support for different CRL issuing certificates. This covers a
1824 simple case where the self issued certificates in the chain exist and
1825 the real CRL issuer is higher in the existing chain.
1827 This work was sponsored by Google.
1830 *) Removed effectively defunct crypto/store from the build.
1833 *) Revamp of STACK to provide stronger type-checking. Still to come:
1834 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1835 ASN1_STRING, CONF_VALUE.
1838 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1839 RAM on SSL connections. This option can save about 34k per idle SSL.
1842 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1843 STACK, TXT_DB, bsearch, qsort.
1846 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1847 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1848 support for data, signedData, compressedData, digestedData and
1849 encryptedData, envelopedData types included. Scripts to check against
1850 RFC4134 examples draft and interop and consistency checks of many
1851 content types and variants.
1854 *) Add options to enc utility to support use of zlib compression BIO.
1857 *) Extend mk1mf to support importing of options and assembly language
1858 files from Configure script, currently only included in VC-WIN32.
1859 The assembly language rules can now optionally generate the source
1860 files from the associated perl scripts.
1863 *) Implement remaining functionality needed to support GOST ciphersuites.
1864 Interop testing has been performed using CryptoPro implementations.
1865 [Victor B. Wagner <vitus@cryptocom.ru>]
1867 *) s390x assembler pack.
1870 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1874 *) Implement Opaque PRF Input TLS extension as specified in
1875 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1876 official specification yet and no extension type assignment by
1877 IANA exists, this extension (for now) will have to be explicitly
1878 enabled when building OpenSSL by providing the extension number
1879 to use. For example, specify an option
1881 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1883 to the "config" or "Configure" script to enable the extension,
1884 assuming extension number 0x9527 (which is a completely arbitrary
1885 and unofficial assignment based on the MD5 hash of the Internet
1886 Draft). Note that by doing so, you potentially lose
1887 interoperability with other TLS implementations since these might
1888 be using the same extension number for other purposes.
1890 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1891 opaque PRF input value to use in the handshake. This will create
1892 an interal copy of the length-'len' string at 'src', and will
1893 return non-zero for success.
1895 To get more control and flexibility, provide a callback function
1898 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1899 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1903 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1906 Callback function 'cb' will be called in handshakes, and is
1907 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1908 Argument 'arg' is for application purposes (the value as given to
1909 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1910 be provided to the callback function). The callback function
1911 has to return non-zero to report success: usually 1 to use opaque
1912 PRF input just if possible, or 2 to enforce use of the opaque PRF
1913 input. In the latter case, the library will abort the handshake
1914 if opaque PRF input is not successfully negotiated.
1916 Arguments 'peerinput' and 'len' given to the callback function
1917 will always be NULL and 0 in the case of a client. A server will
1918 see the client's opaque PRF input through these variables if
1919 available (NULL and 0 otherwise). Note that if the server
1920 provides an opaque PRF input, the length must be the same as the
1921 length of the client's opaque PRF input.
1923 Note that the callback function will only be called when creating
1924 a new session (session resumption can resume whatever was
1925 previously negotiated), and will not be called in SSL 2.0
1926 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1927 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1928 for applications that need to enforce opaque PRF input.
1932 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1935 [Victor B. Wagner <vitus@cryptocom.ru>]
1937 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1938 RFC4507bis. The encrypted ticket format is an encrypted encoded
1939 SSL_SESSION structure, that way new session features are automatically
1942 If a client application caches session in an SSL_SESSION structure
1943 support is transparent because tickets are now stored in the encoded
1946 The SSL_CTX structure automatically generates keys for ticket
1947 protection in servers so again support should be possible
1948 with no application modification.
1950 If a client or server wishes to disable RFC4507 support then the option
1951 SSL_OP_NO_TICKET can be set.
1953 Add a TLS extension debugging callback to allow the contents of any client
1954 or server extensions to be examined.
1956 This work was sponsored by Google.
1959 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1960 OpenSSL should now compile cleanly on gcc 4.2
1961 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1963 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1964 support including streaming MAC support: this is required for GOST
1965 ciphersuite support.
1966 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1968 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1969 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1970 to output in BER and PEM format.
1973 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1974 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1975 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1976 ENGINE support for HMAC keys which are unextractable. New -mac and
1977 -macopt options to dgst utility.
1980 *) New option -sigopt to dgst utility. Update dgst to use
1981 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1982 alternative signing paramaters such as X9.31 or PSS in the dgst
1986 *) Change ssl_cipher_apply_rule(), the internal function that does
1987 the work each time a ciphersuite string requests enabling
1988 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1989 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1990 the order of disabled ciphersuites such that those ciphersuites
1991 that most recently went from enabled to disabled not only stay
1992 in order with respect to each other, but also have higher priority
1993 than other disabled ciphersuites the next time ciphersuites are
1996 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1997 the same ciphersuites as with "HIGH" alone, but in a specific
1998 order where the PSK ciphersuites come first (since they are the
1999 most recently disabled ciphersuites when "HIGH" is parsed).
2001 Also, change ssl_create_cipher_list() (using this new
2002 funcionality) such that between otherwise identical
2003 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2007 *) Change ssl_create_cipher_list() so that it automatically
2008 arranges the ciphersuites in reasonable order before starting
2009 to process the rule string. Thus, the definition for "DEFAULT"
2010 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2011 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2012 This makes it much easier to arrive at a reasonable default order
2013 in applications for which anonymous ciphers are OK (meaning
2014 that you can't actually use DEFAULT).
2015 [Bodo Moeller; suggested by Victor Duchovni]
2017 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2018 processing) into multiple integers instead of setting
2019 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2020 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2021 (These masks as well as the individual bit definitions are hidden
2022 away into the non-exported interface ssl/ssl_locl.h, so this
2023 change to the definition of the SSL_CIPHER structure shouldn't
2024 affect applications.) This give us more bits for each of these
2025 categories, so there is no longer a need to coagulate AES128 and
2026 AES256 into a single algorithm bit, and to coagulate Camellia128
2027 and Camellia256 into a single algorithm bit, which has led to all
2030 Thus, among other things, the kludge introduced in 0.9.7m and
2031 0.9.8e for masking out AES256 independently of AES128 or masking
2032 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2034 With the change, we also introduce new ciphersuite aliases that
2035 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2039 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2040 Use the leftmost N bytes of the signature input if the input is
2041 larger than the prime q (with N being the size in bytes of q).
2044 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2045 it yet and it is largely untested.
2048 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2051 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2052 some compilers (gcc 4.2 and later) reject their use. Safestack is
2053 reimplemented. Update ASN1 to avoid use of legacy functions.
2056 *) Win32/64 targets are linked with Winsock2.
2059 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2060 to external functions. This can be used to increase CRL handling
2061 efficiency especially when CRLs are very large by (for example) storing
2062 the CRL revoked certificates in a database.
2065 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2066 new CRLs added to a directory can be used. New command line option
2067 -verify_return_error to s_client and s_server. This causes real errors
2068 to be returned by the verify callback instead of carrying on no matter
2069 what. This reflects the way a "real world" verify callback would behave.
2072 *) GOST engine, supporting several GOST algorithms and public key formats.
2073 Kindly donated by Cryptocom.
2076 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2077 partitioned by DP are handled but no indirect CRL or reason partitioning
2078 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2079 selected via a scoring technique which handles IDP and AKID in CRLs.
2082 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2083 will ultimately be used for all verify operations: this will remove the
2084 X509_STORE dependency on certificate verification and allow alternative
2085 lookup methods. X509_STORE based implementations of these two callbacks.
2088 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2089 Modify get_crl() to find a valid (unexpired) CRL if possible.
2092 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2093 this would be called X509_CRL_cmp() but that name is already used by
2094 a function that just compares CRL issuer names. Cache several CRL
2095 extensions in X509_CRL structure and cache CRLDP in X509.
2098 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2099 this maps equivalent X509_NAME structures into a consistent structure.
2100 Name comparison can then be performed rapidly using memcmp().
2103 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2107 *) Allow digests to supply their own micalg string for S/MIME type using
2108 the ctrl EVP_MD_CTRL_MICALG.
2111 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2112 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2113 ctrl. It can then customise the structure before and/or after signing
2117 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2118 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2119 to free up any added signature OIDs.
2122 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2123 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2124 digest and cipher tables. New options added to openssl utility:
2125 list-message-digest-algorithms and list-cipher-algorithms.
2128 *) Change the array representation of binary polynomials: the list
2129 of degrees of non-zero coefficients is now terminated with -1.
2130 Previously it was terminated with 0, which was also part of the
2131 value; thus, the array representation was not applicable to
2132 polynomials where t^0 has coefficient zero. This change makes
2133 the array representation useful in a more general context.
2136 *) Various modifications and fixes to SSL/TLS cipher string
2137 handling. For ECC, the code now distinguishes between fixed ECDH
2138 with RSA certificates on the one hand and with ECDSA certificates
2139 on the other hand, since these are separate ciphersuites. The
2140 unused code for Fortezza ciphersuites has been removed.
2142 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2143 (not "ECDHE"). For consistency with the code for DH
2144 certificates, use of ECDH certificates is now considered ECDH
2145 authentication, not RSA or ECDSA authentication (the latter is
2146 merely the CA's signing algorithm and not actively used in the
2149 The temporary ciphersuite alias "ECCdraft" is no longer
2150 available, and ECC ciphersuites are no longer excluded from "ALL"
2151 and "DEFAULT". The following aliases now exist for RFC 4492
2152 ciphersuites, most of these by analogy with the DH case:
2154 kECDHr - ECDH cert, signed with RSA
2155 kECDHe - ECDH cert, signed with ECDSA
2156 kECDH - ECDH cert (signed with either RSA or ECDSA)
2157 kEECDH - ephemeral ECDH
2158 ECDH - ECDH cert or ephemeral ECDH
2164 AECDH - anonymous ECDH
2165 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2169 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2170 Use correct micalg parameters depending on digest(s) in signed message.
2173 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2174 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2177 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2178 an engine to register a method. Add ENGINE lookups for methods and
2179 functional reference processing.
2182 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2183 EVP_{Sign,Verify}* which allow an application to customise the signature
2187 *) New -resign option to smime utility. This adds one or more signers
2188 to an existing PKCS#7 signedData structure. Also -md option to use an
2189 alternative message digest algorithm for signing.
2192 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2193 create PKCS7 structures containing multiple signers. Update smime
2194 application to support multiple signers.
2197 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2201 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2202 Reorganize PBE internals to lookup from a static table using NIDs,
2203 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2204 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2205 PRF which will be automatically used with PBES2.
2208 *) Replace the algorithm specific calls to generate keys in "req" with the
2212 *) Update PKCS#7 enveloped data routines to use new API. This is now
2213 supported by any public key method supporting the encrypt operation. A
2214 ctrl is added to allow the public key algorithm to examine or modify
2215 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2219 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2220 a default digest type to use. In most cases this will be SHA1 but some
2221 algorithms (such as GOST) need to specify an alternative digest. The
2222 return value indicates how strong the preference is 1 means optional and
2223 2 is mandatory (that is it is the only supported type). Modify
2224 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2225 use the default md. Update openssl utilities to use the default digest
2226 type for signing if it is not explicitly indicated.
2229 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2230 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2231 signing method from the key type. This effectively removes the link
2232 between digests and public key types.
2235 *) Add an OID cross reference table and utility functions. Its purpose is to
2236 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2237 rsaEncryption. This will allow some of the algorithm specific hackery
2238 needed to use the correct OID to be removed.
2241 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2242 structures for PKCS7_sign(). They are now set up by the relevant public
2246 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2249 *) Add support for key derivation (agreement) in the API, DH method and
2253 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2254 public and private key formats. As a side effect these add additional
2255 command line functionality not previously available: DSA signatures can be
2256 generated and verified using pkeyutl and DH key support and generation in
2261 [Oliver Tappe <zooey@hirschkaefer.de>]
2263 *) New make target "install_html_docs" installs HTML renditions of the
2265 [Oliver Tappe <zooey@hirschkaefer.de>]
2267 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2268 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2269 support key and parameter generation and add initial key generation
2270 functionality for RSA.
2273 *) Add functions for main EVP_PKEY_method operations. The undocumented
2274 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2275 EVP_PKEY_{encrypt,decrypt}_old.
2278 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2279 key API, doesn't do much yet.
2282 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2283 public key algorithms. New option to openssl utility:
2284 "list-public-key-algorithms" to print out info.
2287 *) Implement the Supported Elliptic Curves Extension for
2288 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2291 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2292 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2295 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2296 utilities such as rsa, dsa, dsaparam etc except they process any key
2300 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2301 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2302 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2306 *) Initial support for pluggable public key ASN1.
2307 De-spaghettify the public key ASN1 handling. Move public and private
2308 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2309 algorithm specific handling to a single module within the relevant
2310 algorithm directory. Add functions to allow (near) opaque processing
2311 of public and private key structures.
2314 *) Implement the Supported Point Formats Extension for
2315 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2318 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2319 for the psk identity [hint] and the psk callback functions to the
2320 SSL_SESSION, SSL and SSL_CTX structure.
2323 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2327 SSL_CTX_use_psk_identity_hint
2328 SSL_get_psk_identity_hint
2329 SSL_get_psk_identity
2330 SSL_use_psk_identity_hint
2332 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2334 *) Add RFC 3161 compliant time stamp request creation, response generation
2335 and response verification functionality.
2336 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2338 *) Add initial support for TLS extensions, specifically for the server_name
2339 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2340 have new members for a host name. The SSL data structure has an
2341 additional member SSL_CTX *initial_ctx so that new sessions can be
2342 stored in that context to allow for session resumption, even after the
2343 SSL has been switched to a new SSL_CTX in reaction to a client's
2344 server_name extension.
2346 New functions (subject to change):
2348 SSL_get_servername()
2349 SSL_get_servername_type()
2352 New CTRL codes and macros (subject to change):
2354 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2355 - SSL_CTX_set_tlsext_servername_callback()
2356 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2357 - SSL_CTX_set_tlsext_servername_arg()
2358 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2360 openssl s_client has a new '-servername ...' option.
2362 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2363 '-key2 ...', '-servername_fatal' (subject to change). This allows
2364 testing the HostName extension for a specific single host name ('-cert'
2365 and '-key' remain fallbacks for handshakes without HostName
2366 negotiation). If the unrecognized_name alert has to be sent, this by
2367 default is a warning; it becomes fatal with the '-servername_fatal'
2370 [Peter Sylvester, Remy Allais, Christophe Renou]
2372 *) Whirlpool hash implementation is added.
2375 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2376 bn(64,32). Because of instruction set limitations it doesn't have
2377 any negative impact on performance. This was done mostly in order
2378 to make it possible to share assembler modules, such as bn_mul_mont
2379 implementations, between 32- and 64-bit builds without hassle.
2382 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2383 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2387 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2388 dedicated Montgomery multiplication procedure, is introduced.
2389 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2390 "64-bit" performance on certain 32-bit targets.
2393 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2394 in SSL structures. New SSL ctrl to set maximum send fragment size.
2395 Save memory by seeting the I/O buffer sizes dynamically instead of
2396 using the maximum available value.
2399 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2400 in addition to the text details.
2403 *) Very, very preliminary EXPERIMENTAL support for printing of general
2404 ASN1 structures. This currently produces rather ugly output and doesn't
2405 handle several customised structures at all.
2408 *) Integrated support for PVK file format and some related formats such
2409 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2410 these in the 'rsa' and 'dsa' utilities.
2413 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2416 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2417 place for the (very old) "NETSCAPE" format certificates which are now
2418 handled using new ASN1 code equivalents.
2421 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2422 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2423 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2426 *) Modify CRL distribution points extension code to print out previously
2427 unsupported fields. Enhance extension setting code to allow setting of
2431 *) Add print and set support for Issuing Distribution Point CRL extension.
2434 *) Change 'Configure' script to enable Camellia by default.
2437 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2439 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2441 This addresses the flaw in CBC record processing discovered by
2442 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2443 at: http://www.isg.rhul.ac.uk/tls/
2445 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2446 Security Group at Royal Holloway, University of London
2447 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2448 Emilia Käsper for the initial patch.
2450 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2452 *) Return an error when checking OCSP signatures when key is NULL.
2453 This fixes a DoS attack. (CVE-2013-0166)
2456 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2457 the right response is stapled. Also change SSL_get_certificate()
2458 so it returns the certificate actually sent.
2459 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2460 (This is a backport)
2461 [Rob Stradling <rob.stradling@comodo.com>]
2463 *) Fix possible deadlock when decoding public keys.
2466 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2468 *) Sanity check record length before skipping explicit IV in DTLS
2471 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2472 fuzzing as a service testing platform.
2476 *) Initialise tkeylen properly when encrypting CMS messages.
2477 Thanks to Solar Designer of Openwall for reporting this issue.
2480 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2482 *) The fix for CVE-2012-2110 did not take into account that the
2483 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2484 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2485 rejecting negative len parameter. (CVE-2012-2131)
2486 [Tomas Hoger <thoger@redhat.com>]
2488 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2490 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2491 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2492 in CRYPTO_realloc_clean.
2494 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2495 issue and to Adam Langley <agl@chromium.org> for fixing it.
2497 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2499 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2501 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2502 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2503 content decryption and always return the same error. Note: this attack
2504 needs on average 2^20 messages so it only affects automated senders. The
2505 old behaviour can be reenabled in the CMS code by setting the
2506 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2507 an MMA defence is not necessary.
2508 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2509 this issue. (CVE-2012-0884)
2512 *) Fix CVE-2011-4619: make sure we really are receiving a
2513 client hello before rejecting multiple SGC restarts. Thanks to
2514 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2517 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2519 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2520 Thanks to Antonio Martin, Enterprise Secure Access Research and
2521 Development, Cisco Systems, Inc. for discovering this bug and
2522 preparing a fix. (CVE-2012-0050)
2525 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2527 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2528 of the Vaudenay padding oracle attack on CBC mode encryption
2529 which enables an efficient plaintext recovery attack against
2530 the OpenSSL implementation of DTLS. Their attack exploits timing
2531 differences arising during decryption processing. A research
2532 paper describing this attack can be found at:
2533 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2534 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2535 Security Group at Royal Holloway, University of London
2536 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2537 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2538 for preparing the fix. (CVE-2011-4108)
2539 [Robin Seggelmann, Michael Tuexen]
2541 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2542 [Ben Laurie, Kasper <ekasper@google.com>]
2544 *) Clear bytes used for block padding of SSL 3.0 records.
2546 [Adam Langley (Google)]
2548 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2549 Kadianakis <desnacked@gmail.com> for discovering this issue and
2550 Adam Langley for preparing the fix. (CVE-2011-4619)
2551 [Adam Langley (Google)]
2553 *) Prevent malformed RFC3779 data triggering an assertion failure.
2554 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2555 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2556 [Rob Austein <sra@hactrn.net>]
2558 *) Fix ssl_ciph.c set-up race.
2559 [Adam Langley (Google)]
2561 *) Fix spurious failures in ecdsatest.c.
2562 [Emilia Käsper (Google)]
2564 *) Fix the BIO_f_buffer() implementation (which was mixing different
2565 interpretations of the '..._len' fields).
2566 [Adam Langley (Google)]
2568 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2569 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2570 threads won't reuse the same blinding coefficients.
2572 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2573 lock to call BN_BLINDING_invert_ex, and avoids one use of
2574 BN_BLINDING_update for each BN_BLINDING structure (previously,
2575 the last update always remained unused).
2576 [Emilia Käsper (Google)]
2578 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2579 for multi-threaded use of ECDH.
2580 [Adam Langley (Google)]
2582 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2585 *) Add protection against ECDSA timing attacks as mentioned in the paper
2586 by Billy Bob Brumley and Nicola Tuveri, see:
2588 http://eprint.iacr.org/2011/232.pdf
2590 [Billy Bob Brumley and Nicola Tuveri]
2592 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2594 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2595 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2597 *) Fix bug in string printing code: if *any* escaping is enabled we must
2598 escape the escape character (backslash) or the resulting string is
2602 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2604 *) Disable code workaround for ancient and obsolete Netscape browsers
2605 and servers: an attacker can use it in a ciphersuite downgrade attack.
2606 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2609 *) Fixed J-PAKE implementation error, originally discovered by
2610 Sebastien Martini, further info and confirmation from Stefan
2611 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2614 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2616 *) Fix extension code to avoid race conditions which can result in a buffer
2617 overrun vulnerability: resumed sessions must not be modified as they can
2618 be shared by multiple threads. CVE-2010-3864
2621 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2624 *) Don't reencode certificate when calculating signature: cache and use
2625 the original encoding instead. This makes signature verification of
2626 some broken encodings work correctly.
2629 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2630 is also one of the inputs.
2631 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2633 *) Don't repeatedly append PBE algorithms to table if they already exist.
2634 Sort table on each new add. This effectively makes the table read only
2635 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2639 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2641 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2644 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2645 access or freeing data twice (CVE-2010-0742)
2646 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2648 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2649 common in certificates and some applications which only call
2650 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2654 Reduce copying into .apps and .test in makevms.com
2655 Don't try to use blank CA certificate in CA.com
2656 Allow use of C files from original directories in maketests.com
2657 [Steven M. Schweda" <sms@antinode.info>]
2659 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2661 *) When rejecting SSL/TLS records due to an incorrect version number, never
2662 update s->server with a new major version number. As of
2663 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2664 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2665 the previous behavior could result in a read attempt at NULL when
2666 receiving specific incorrect SSL/TLS records once record payload
2667 protection is active. (CVE-2010-0740)
2668 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2670 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2671 could be crashed if the relevant tables were not present (e.g. chrooted).
2672 [Tomas Hoger <thoger@redhat.com>]
2674 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2676 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2677 [Martin Olsson, Neel Mehta]
2679 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2680 accommodate for stack sorting, always a write lock!).
2683 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2684 excessive delays in the RAND_poll(): over a minute. As a workaround
2685 include a time check in the inner Heap32Next loop too.
2688 *) The code that handled flushing of data in SSL/TLS originally used the
2689 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2690 the problem outlined in PR#1949. The fix suggested there however can
2691 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2692 of Apache). So instead simplify the code to flush unconditionally.
2693 This should be fine since flushing with no data to flush is a no op.
2696 *) Handle TLS versions 2.0 and later properly and correctly use the
2697 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2698 off ancient servers have a habit of sticking around for a while...
2701 *) Modify compression code so it frees up structures without using the
2702 ex_data callbacks. This works around a problem where some applications
2703 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2704 restarting) then use compression (e.g. SSL with compression) later.
2705 This results in significant per-connection memory leaks and
2706 has caused some security issues including CVE-2008-1678 and
2710 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2711 change when encrypting or decrypting.
2714 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2715 connect and renegotiate with servers which do not support RI.
2716 Until RI is more widely deployed this option is enabled by default.
2719 *) Add "missing" ssl ctrls to clear options and mode.
2722 *) If client attempts to renegotiate and doesn't support RI respond with
2723 a no_renegotiation alert as required by RFC5746. Some renegotiating
2724 TLS clients will continue a connection gracefully when they receive
2725 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2726 waiting for a server hello which it will never receive. Now we treat a
2727 received no_renegotiation alert as a fatal error. This is because
2728 applications requesting a renegotiation might well expect it to succeed
2729 and would have no code in place to handle the server denying it so the
2730 only safe thing to do is to terminate the connection.
2733 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2734 peer supports secure renegotiation and 0 otherwise. Print out peer
2735 renegotiation support in s_client/s_server.
2738 *) Replace the highly broken and deprecated SPKAC certification method with
2739 the updated NID creation version. This should correctly handle UTF8.
2742 *) Implement RFC5746. Re-enable renegotiation but require the extension
2743 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2744 turns out to be a bad idea. It has been replaced by
2745 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2746 SSL_CTX_set_options(). This is really not recommended unless you
2747 know what you are doing.
2748 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2750 *) Fixes to stateless session resumption handling. Use initial_ctx when
2751 issuing and attempting to decrypt tickets in case it has changed during
2752 servername handling. Use a non-zero length session ID when attempting
2753 stateless session resumption: this makes it possible to determine if
2754 a resumption has occurred immediately after receiving server hello
2755 (several places in OpenSSL subtly assume this) instead of later in
2759 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2760 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2761 fixes for a few places where the return code is not checked
2763 [Julia Lawall <julia@diku.dk>]
2765 *) Add --strict-warnings option to Configure script to include devteam
2766 warnings in other configurations.
2769 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2770 makes it possible to install openssl libraries in locations which
2771 have names other than "lib", for example "/usr/lib64" which some
2773 [Steve Henson, based on patch from Jeremy Utley]
2775 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2776 X690 8.9.12 and can produce some misleading textual output of OIDs.
2777 [Steve Henson, reported by Dan Kaminsky]
2779 *) Delete MD2 from algorithm tables. This follows the recommendation in
2780 several standards that it is not used in new applications due to
2781 several cryptographic weaknesses. For binary compatibility reasons
2782 the MD2 API is still compiled in by default.
2785 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2789 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2790 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2792 [Guenter <lists@gknw.net>]
2794 *) Fix the server certificate chain building code to use X509_verify_cert(),
2795 it used to have an ad-hoc builder which was unable to cope with anything
2796 other than a simple chain.
2797 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2799 *) Don't check self signed certificate signatures in X509_verify_cert()
2800 by default (a flag can override this): it just wastes time without
2801 adding any security. As a useful side effect self signed root CAs
2802 with non-FIPS digests are now usable in FIPS mode.
2805 *) In dtls1_process_out_of_seq_message() the check if the current message
2806 is already buffered was missing. For every new message was memory
2807 allocated, allowing an attacker to perform an denial of service attack
2808 with sending out of seq handshake messages until there is no memory
2809 left. Additionally every future messege was buffered, even if the
2810 sequence number made no sense and would be part of another handshake.
2811 So only messages with sequence numbers less than 10 in advance will be
2812 buffered. (CVE-2009-1378)
2813 [Robin Seggelmann, discovered by Daniel Mentz]
2815 *) Records are buffered if they arrive with a future epoch to be
2816 processed after finishing the corresponding handshake. There is
2817 currently no limitation to this buffer allowing an attacker to perform
2818 a DOS attack with sending records with future epochs until there is no
2819 memory left. This patch adds the pqueue_size() function to determine
2820 the size of a buffer and limits the record buffer to 100 entries.
2822 [Robin Seggelmann, discovered by Daniel Mentz]
2824 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2825 parent structure is freed. (CVE-2009-1379)
2828 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2829 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2832 [Ilya O. <vrghost@gmail.com>]
2834 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2836 *) Disable renegotiation completely - this fixes a severe security
2837 problem (CVE-2009-3555) at the cost of breaking all
2838 renegotiation. Renegotiation can be re-enabled by setting
2839 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2840 run-time. This is really not recommended unless you know what
2844 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2846 *) Don't set val to NULL when freeing up structures, it is freed up by
2847 underlying code. If sizeof(void *) > sizeof(long) this can result in
2848 zeroing past the valid field. (CVE-2009-0789)
2849 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2851 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2852 checked correctly. This would allow some invalid signed attributes to
2853 appear to verify correctly. (CVE-2009-0591)
2854 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2856 *) Reject UniversalString and BMPString types with invalid lengths. This
2857 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2858 a legal length. (CVE-2009-0590)
2861 *) Set S/MIME signing as the default purpose rather than setting it
2862 unconditionally. This allows applications to override it at the store
2866 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2867 to handle some structures.
2870 *) Improve efficiency of mem_gets: don't search whole buffer each time
2872 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2874 *) New -hex option for openssl rand.
2877 *) Print out UTF8String and NumericString when parsing ASN1.
2880 *) Support NumericString type for name components.
2883 *) Allow CC in the environment to override the automatically chosen
2884 compiler. Note that nothing is done to ensure flags work with the
2888 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2890 *) Properly check EVP_VerifyFinal() and similar return values
2892 [Ben Laurie, Bodo Moeller, Google Security Team]
2894 *) Enable TLS extensions by default.
2897 *) Allow the CHIL engine to be loaded, whether the application is
2898 multithreaded or not. (This does not release the developer from the
2899 obligation to set up the dynamic locking callbacks.)
2900 [Sander Temme <sander@temme.net>]
2902 *) Use correct exit code if there is an error in dgst command.
2903 [Steve Henson; problem pointed out by Roland Dirlewanger]
2905 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2906 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2909 *) Add experimental JPAKE support, including demo authentication in
2910 s_client and s_server.
2913 *) Set the comparison function in v3_addr_canonize().
2914 [Rob Austein <sra@hactrn.net>]
2916 *) Add support for XMPP STARTTLS in s_client.
2917 [Philip Paeps <philip@freebsd.org>]
2919 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2920 to ensure that even with this option, only ciphersuites in the
2921 server's preference list will be accepted. (Note that the option
2922 applies only when resuming a session, so the earlier behavior was
2923 just about the algorithm choice for symmetric cryptography.)
2926 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2928 *) Fix NULL pointer dereference if a DTLS server received
2929 ChangeCipherSpec as first record (CVE-2009-1386).
2932 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2933 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2936 *) The fix in 0.9.8c that supposedly got rid of unsafe
2937 double-checked locking was incomplete for RSA blinding,
2938 addressing just one layer of what turns out to have been
2939 doubly unsafe triple-checked locking.
2941 So now fix this for real by retiring the MONT_HELPER macro
2942 in crypto/rsa/rsa_eay.c.
2944 [Bodo Moeller; problem pointed out by Marius Schilder]
2946 *) Various precautionary measures:
2948 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2950 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2951 (NB: This would require knowledge of the secret session ticket key
2952 to exploit, in which case you'd be SOL either way.)
2954 - Change bn_nist.c so that it will properly handle input BIGNUMs
2955 outside the expected range.
2957 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2960 [Neel Mehta, Bodo Moeller]
2962 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2963 the load fails. Useful for distros.
2964 [Ben Laurie and the FreeBSD team]
2966 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2969 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2972 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2974 This work was sponsored by Logica.
2977 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2978 keystores. Support for SSL/TLS client authentication too.
2979 Not compiled unless enable-capieng specified to Configure.
2981 This work was sponsored by Logica.
2984 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2985 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2986 attribute creation routines such as certifcate requests and PKCS#12
2990 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2992 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2993 handshake which could lead to a cilent crash as found using the
2994 Codenomicon TLS test suite (CVE-2008-1672)
2995 [Steve Henson, Mark Cox]
2997 *) Fix double free in TLS server name extensions which could lead to
2998 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3001 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3003 Clear the error queue to ensure that error entries left from
3004 older function calls do not interfere with the correct operation.
3005 [Lutz Jaenicke, Erik de Castro Lopo]
3007 *) Remove root CA certificates of commercial CAs:
3009 The OpenSSL project does not recommend any specific CA and does not
3010 have any policy with respect to including or excluding any CA.
3011 Therefore it does not make any sense to ship an arbitrary selection
3012 of root CA certificates with the OpenSSL software.
3015 *) RSA OAEP patches to fix two separate invalid memory reads.
3016 The first one involves inputs when 'lzero' is greater than
3017 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3018 before the beginning of from). The second one involves inputs where
3019 the 'db' section contains nothing but zeroes (there is a one-byte
3020 invalid read after the end of 'db').
3021 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3023 *) Partial backport from 0.9.9-dev:
3025 Introduce bn_mul_mont (dedicated Montgomery multiplication
3026 procedure) as a candidate for BIGNUM assembler implementation.
3027 While 0.9.9-dev uses assembler for various architectures, only
3028 x86_64 is available by default here in the 0.9.8 branch, and
3029 32-bit x86 is available through a compile-time setting.
3031 To try the 32-bit x86 assembler implementation, use Configure
3032 option "enable-montasm" (which exists only for this backport).
3034 As "enable-montasm" for 32-bit x86 disclaims code stability
3035 anyway, in this constellation we activate additional code
3036 backported from 0.9.9-dev for further performance improvements,
3037 namely BN_from_montgomery_word. (To enable this otherwise,
3038 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3040 [Andy Polyakov (backport partially by Bodo Moeller)]
3042 *) Add TLS session ticket callback. This allows an application to set
3043 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3044 values. This is useful for key rollover for example where several key
3045 sets may exist with different names.
3048 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3049 This was broken until now in 0.9.8 releases, such that the only way
3050 a registered ENGINE could be used (assuming it initialises
3051 successfully on the host) was to explicitly set it as the default
3052 for the relevant algorithms. This is in contradiction with 0.9.7
3053 behaviour and the documentation. With this fix, when an ENGINE is
3054 registered into a given algorithm's table of implementations, the
3055 'uptodate' flag is reset so that auto-discovery will be used next
3056 time a new context for that algorithm attempts to select an
3058 [Ian Lister (tweaked by Geoff Thorpe)]
3060 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3061 implemention in the following ways:
3063 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3066 Lack of BER streaming support means one pass streaming processing is
3067 only supported if data is detached: setting the streaming flag is
3068 ignored for embedded content.
3070 CMS support is disabled by default and must be explicitly enabled
3071 with the enable-cms configuration option.
3074 *) Update the GMP engine glue to do direct copies between BIGNUM and
3075 mpz_t when openssl and GMP use the same limb size. Otherwise the
3076 existing "conversion via a text string export" trick is still used.
3077 [Paul Sheer <paulsheer@gmail.com>]
3079 *) Zlib compression BIO. This is a filter BIO which compressed and
3080 uncompresses any data passed through it.
3083 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3084 RFC3394 compatible AES key wrapping.
3087 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3088 sets string data without copying. X509_ALGOR_set0() and
3089 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3090 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3091 from an X509_ATTRIBUTE structure optionally checking it occurs only
3092 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3096 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3097 to get the expected BN_FLG_CONSTTIME behavior.
3098 [Bodo Moeller (Google)]
3102 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3103 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3104 - added some more tests to do_tests.pl
3105 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3106 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3107 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3108 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3109 - various changes to netware.pl to enable gcc-cross builds on Win32
3111 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3112 - various changes to fix missing prototype warnings
3113 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3114 - added AES, WHIRLPOOL and CPUID assembler code to build files
3115 - added missing AES assembler make rules to mk1mf.pl
3116 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3117 [Guenter Knauf <eflash@gmx.net>]
3119 *) Implement certificate status request TLS extension defined in RFC3546.
3120 A client can set the appropriate parameters and receive the encoded
3121 OCSP response via a callback. A server can query the supplied parameters
3122 and set the encoded OCSP response in the callback. Add simplified examples
3123 to s_client and s_server.
3126 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3128 *) Fix various bugs:
3129 + Binary incompatibility of ssl_ctx_st structure
3130 + DTLS interoperation with non-compliant servers
3131 + Don't call get_session_cb() without proposed session
3132 + Fix ia64 assembler code
3133 [Andy Polyakov, Steve Henson]
3135 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3137 *) DTLS Handshake overhaul. There were longstanding issues with
3138 OpenSSL DTLS implementation, which were making it impossible for
3139 RFC 4347 compliant client to communicate with OpenSSL server.
3140 Unfortunately just fixing these incompatibilities would "cut off"
3141 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3142 server keeps tolerating non RFC compliant syntax. The opposite is
3143 not true, 0.9.8f client can not communicate with earlier server.
3144 This update even addresses CVE-2007-4995.
3147 *) Changes to avoid need for function casts in OpenSSL: some compilers
3148 (gcc 4.2 and later) reject their use.
3149 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3152 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3153 RFC4507bis. The encrypted ticket format is an encrypted encoded
3154 SSL_SESSION structure, that way new session features are automatically
3157 If a client application caches session in an SSL_SESSION structure
3158 support is transparent because tickets are now stored in the encoded
3161 The SSL_CTX structure automatically generates keys for ticket
3162 protection in servers so again support should be possible
3163 with no application modification.
3165 If a client or server wishes to disable RFC4507 support then the option
3166 SSL_OP_NO_TICKET can be set.
3168 Add a TLS extension debugging callback to allow the contents of any client
3169 or server extensions to be examined.
3171 This work was sponsored by Google.
3174 *) Add initial support for TLS extensions, specifically for the server_name
3175 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3176 have new members for a host name. The SSL data structure has an
3177 additional member SSL_CTX *initial_ctx so that new sessions can be
3178 stored in that context to allow for session resumption, even after the
3179 SSL has been switched to a new SSL_CTX in reaction to a client's
3180 server_name extension.
3182 New functions (subject to change):
3184 SSL_get_servername()
3185 SSL_get_servername_type()
3188 New CTRL codes and macros (subject to change):
3190 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3191 - SSL_CTX_set_tlsext_servername_callback()
3192 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3193 - SSL_CTX_set_tlsext_servername_arg()
3194 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3196 openssl s_client has a new '-servername ...' option.
3198 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3199 '-key2 ...', '-servername_fatal' (subject to change). This allows
3200 testing the HostName extension for a specific single host name ('-cert'
3201 and '-key' remain fallbacks for handshakes without HostName
3202 negotiation). If the unrecognized_name alert has to be sent, this by
3203 default is a warning; it becomes fatal with the '-servername_fatal'
3206 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3208 *) Add AES and SSE2 assembly language support to VC++ build.
3211 *) Mitigate attack on final subtraction in Montgomery reduction.
3214 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3215 (which previously caused an internal error).
3218 *) Squeeze another 10% out of IGE mode when in != out.
3221 *) AES IGE mode speedup.
3222 [Dean Gaudet (Google)]
3224 *) Add the Korean symmetric 128-bit cipher SEED (see
3225 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3226 add SEED ciphersuites from RFC 4162:
3228 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3229 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3230 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3231 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3233 To minimize changes between patchlevels in the OpenSSL 0.9.8
3234 series, SEED remains excluded from compilation unless OpenSSL
3235 is configured with 'enable-seed'.
3236 [KISA, Bodo Moeller]
3238 *) Mitigate branch prediction attacks, which can be practical if a
3239 single processor is shared, allowing a spy process to extract
3240 information. For detailed background information, see
3241 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3242 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3243 and Necessary Software Countermeasures"). The core of the change
3244 are new versions BN_div_no_branch() and
3245 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3246 respectively, which are slower, but avoid the security-relevant
3247 conditional branches. These are automatically called by BN_div()
3248 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3249 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3250 remove a conditional branch.
3252 BN_FLG_CONSTTIME is the new name for the previous
3253 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3254 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3255 in the exponent causes BN_mod_exp_mont() to use the alternative
3256 implementation in BN_mod_exp_mont_consttime().) The old name
3257 remains as a deprecated alias.
3259 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3260 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3261 constant-time implementations for more than just exponentiation.
3262 Here too the old name is kept as a deprecated alias.
3264 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3265 the BN_BLINDING structure gets an independent copy of the
3266 modulus. This means that the previous "BIGNUM *m" argument to
3267 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3268 essentially becomes "const BIGNUM *m", although we can't actually
3269 change this in the header file before 0.9.9. It allows
3270 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3271 enable BN_FLG_CONSTTIME.
3273 [Matthew D Wood (Intel Corp)]
3275 *) In the SSL/TLS server implementation, be strict about session ID
3276 context matching (which matters if an application uses a single
3277 external cache for different purposes). Previously,
3278 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3279 set. This did ensure strict client verification, but meant that,
3280 with applications using a single external cache for quite
3281 different requirements, clients could circumvent ciphersuite
3282 restrictions for a given session ID context by starting a session
3283 in a different context.
3286 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3287 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3288 authentication-only ciphersuites.
3291 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3292 not complete and could lead to a possible single byte overflow
3293 (CVE-2007-5135) [Ben Laurie]
3295 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3297 *) Since AES128 and AES256 (and similarly Camellia128 and
3298 Camellia256) share a single mask bit in the logic of
3299 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3300 kludge to work properly if AES128 is available and AES256 isn't
3301 (or if Camellia128 is available and Camellia256 isn't).
3304 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3305 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3306 When a point or a seed is encoded in a BIT STRING, we need to
3307 prevent the removal of trailing zero bits to get the proper DER
3308 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3309 of a NamedBitList, for which trailing 0 bits need to be removed.)
3312 *) Have SSL/TLS server implementation tolerate "mismatched" record
3313 protocol version while receiving ClientHello even if the
3314 ClientHello is fragmented. (The server can't insist on the
3315 particular protocol version it has chosen before the ServerHello
3316 message has informed the client about his choice.)
3319 *) Add RFC 3779 support.
3320 [Rob Austein for ARIN, Ben Laurie]
3322 *) Load error codes if they are not already present instead of using a
3323 static variable. This allows them to be cleanly unloaded and reloaded.
3324 Improve header file function name parsing.
3327 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3328 or CAPABILITY handshake as required by RFCs.
3331 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3333 *) Introduce limits to prevent malicious keys being able to
3334 cause a denial of service. (CVE-2006-2940)
3335 [Steve Henson, Bodo Moeller]
3337 *) Fix ASN.1 parsing of certain invalid structures that can result
3338 in a denial of service. (CVE-2006-2937) [Steve Henson]
3340 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3341 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3343 *) Fix SSL client code which could crash if connecting to a
3344 malicious SSLv2 server. (CVE-2006-4343)
3345 [Tavis Ormandy and Will Drewry, Google Security Team]
3347 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3348 match only those. Before that, "AES256-SHA" would be interpreted
3349 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3350 the same strength classification in 0.9.7h) as we currently only
3351 have a single AES bit in the ciphersuite description bitmap.
3352 That change, however, also applied to ciphersuite strings such as
3353 "RC4-MD5" that intentionally matched multiple ciphersuites --
3354 namely, SSL 2.0 ciphersuites in addition to the more common ones
3355 from SSL 3.0/TLS 1.0.
3357 So we change the selection algorithm again: Naming an explicit
3358 ciphersuite selects this one ciphersuite, and any other similar
3359 ciphersuite (same bitmap) from *other* protocol versions.
3360 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3361 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3363 Since SSL 2.0 does not have any ciphersuites for which the
3364 128/256 bit distinction would be relevant, this works for now.
3365 The proper fix will be to use different bits for AES128 and
3366 AES256, which would have avoided the problems from the beginning;
3367 however, bits are scarce, so we can only do this in a new release
3368 (not just a patchlevel) when we can change the SSL_CIPHER
3369 definition to split the single 'unsigned long mask' bitmap into
3370 multiple values to extend the available space.
3374 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3376 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3377 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3379 *) Add AES IGE and biIGE modes.
3382 *) Change the Unix randomness entropy gathering to use poll() when
3383 possible instead of select(), since the latter has some
3384 undesirable limitations.
3385 [Darryl Miles via Richard Levitte and Bodo Moeller]
3387 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3388 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3389 cannot be implicitly activated as part of, e.g., the "AES" alias.
3390 However, please upgrade to OpenSSL 0.9.9[-dev] for
3391 non-experimental use of the ECC ciphersuites to get TLS extension
3392 support, which is required for curve and point format negotiation
3393 to avoid potential handshake problems.
3396 *) Disable rogue ciphersuites:
3398 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3399 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3400 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3402 The latter two were purportedly from
3403 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3406 Also deactivate the remaining ciphersuites from
3407 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3408 unofficial, and the ID has long expired.
3411 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3412 dual-core machines) and other potential thread-safety issues.
3415 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3416 versions), which is now available for royalty-free use
3417 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3418 Also, add Camellia TLS ciphersuites from RFC 4132.
3420 To minimize changes between patchlevels in the OpenSSL 0.9.8
3421 series, Camellia remains excluded from compilation unless OpenSSL
3422 is configured with 'enable-camellia'.
3425 *) Disable the padding bug check when compression is in use. The padding
3426 bug check assumes the first packet is of even length, this is not
3427 necessarily true if compresssion is enabled and can result in false
3428 positives causing handshake failure. The actual bug test is ancient
3429 code so it is hoped that implementations will either have fixed it by
3430 now or any which still have the bug do not support compression.
3433 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3435 *) When applying a cipher rule check to see if string match is an explicit
3436 cipher suite and only match that one cipher suite if it is.
3439 *) Link in manifests for VC++ if needed.
3440 [Austin Ziegler <halostatue@gmail.com>]
3442 *) Update support for ECC-based TLS ciphersuites according to
3443 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3444 TLS extensions, which are supported starting with the 0.9.9
3445 branch, not in the OpenSSL 0.9.8 branch).
3448 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3449 opaque EVP_CIPHER_CTX handling.
3452 *) Fixes and enhancements to zlib compression code. We now only use
3453 "zlib1.dll" and use the default __cdecl calling convention on Win32
3454 to conform with the standards mentioned here:
3455 http://www.zlib.net/DLL_FAQ.txt
3456 Static zlib linking now works on Windows and the new --with-zlib-include
3457 --with-zlib-lib options to Configure can be used to supply the location
3458 of the headers and library. Gracefully handle case where zlib library
3462 *) Several fixes and enhancements to the OID generation code. The old code
3463 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3464 handle numbers larger than ULONG_MAX, truncated printing and had a
3465 non standard OBJ_obj2txt() behaviour.
3468 *) Add support for building of engines under engine/ as shared libraries
3469 under VC++ build system.
3472 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3473 Hopefully, we will not see any false combination of paths any more.
3476 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3478 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3479 (part of SSL_OP_ALL). This option used to disable the
3480 countermeasure against man-in-the-middle protocol-version
3481 rollback in the SSL 2.0 server implementation, which is a bad
3482 idea. (CVE-2005-2969)
3484 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3485 for Information Security, National Institute of Advanced Industrial
3486 Science and Technology [AIST], Japan)]
3488 *) Add two function to clear and return the verify parameter flags.
3491 *) Keep cipherlists sorted in the source instead of sorting them at
3492 runtime, thus removing the need for a lock.
3495 *) Avoid some small subgroup attacks in Diffie-Hellman.
3496 [Nick Mathewson and Ben Laurie]
3498 *) Add functions for well-known primes.
3501 *) Extended Windows CE support.
3502 [Satoshi Nakamura and Andy Polyakov]
3504 *) Initialize SSL_METHOD structures at compile time instead of during
3505 runtime, thus removing the need for a lock.
3508 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3509 attempting to decrypt each encrypted key in turn. Add support to
3513 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3515 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3518 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3521 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3522 key into the same file any more.
3525 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3528 *) Add -utf8 command line and config file option to 'ca'.
3529 [Stefan <stf@udoma.org]
3531 *) Removed the macro des_crypt(), as it seems to conflict with some
3532 libraries. Use DES_crypt().
3535 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3536 involves renaming the source and generated shared-libs for
3537 both. The engines will accept the corrected or legacy ids
3538 ('ncipher' and '4758_cca' respectively) when binding. NB,
3539 this only applies when building 'shared'.
3540 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3542 *) Add attribute functions to EVP_PKEY structure. Modify
3543 PKCS12_create() to recognize a CSP name attribute and
3544 use it. Make -CSP option work again in pkcs12 utility.
3547 *) Add new functionality to the bn blinding code:
3548 - automatic re-creation of the BN_BLINDING parameters after
3549 a fixed number of uses (currently 32)
3550 - add new function for parameter creation
3551 - introduce flags to control the update behaviour of the
3552 BN_BLINDING parameters
3553 - hide BN_BLINDING structure
3554 Add a second BN_BLINDING slot to the RSA structure to improve
3555 performance when a single RSA object is shared among several
3559 *) Add support for DTLS.
3560 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3562 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3563 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3566 *) Remove buggy and incomplete DH cert support from
3567 ssl/ssl_rsa.c and ssl/s3_both.c
3570 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3571 the apps/openssl applications.
3574 *) Compile clean with "-Wall -Wmissing-prototypes
3575 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3576 DEBUG_SAFESTACK must also be set.
3579 *) Change ./Configure so that certain algorithms can be disabled by default.
3580 The new counterpiece to "no-xxx" is "enable-xxx".
3582 The patented RC5 and MDC2 algorithms will now be disabled unless
3583 "enable-rc5" and "enable-mdc2", respectively, are specified.
3585 (IDEA remains enabled despite being patented. This is because IDEA
3586 is frequently required for interoperability, and there is no license
3587 fee for non-commercial use. As before, "no-idea" can be used to
3588 avoid this algorithm.)
3592 *) Add processing of proxy certificates (see RFC 3820). This work was
3593 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3594 EGEE (Enabling Grids for E-science in Europe).
3597 *) RC4 performance overhaul on modern architectures/implementations, such
3598 as Intel P4, IA-64 and AMD64.
3601 *) New utility extract-section.pl. This can be used specify an alternative
3602 section number in a pod file instead of having to treat each file as
3603 a separate case in Makefile. This can be done by adding two lines to the
3606 =for comment openssl_section:XXX
3608 The blank line is mandatory.
3612 *) New arguments -certform, -keyform and -pass for s_client and s_server
3613 to allow alternative format key and certificate files and passphrase
3617 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3618 update associated structures and add various utility functions.
3620 Add new policy related verify parameters, include policy checking in
3621 standard verify code. Enhance 'smime' application with extra parameters
3622 to support policy checking and print out.
3625 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3626 Nehemiah processors. These extensions support AES encryption in hardware
3627 as well as RNG (though RNG support is currently disabled).
3628 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3630 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3633 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3634 [Andy Polyakov and a number of other people]
3636 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3637 implementation contributed by IBM.
3638 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3640 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3641 exponent rather than 'unsigned long'. There is a corresponding change to
3642 the new 'rsa_keygen' element of the RSA_METHOD structure.
3643 [Jelte Jansen, Geoff Thorpe]
3645 *) Functionality for creating the initial serial number file is now
3646 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3648 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3649 number file to 1, which is bound to cause problems. To avoid
3650 the problems while respecting compatibility between different 0.9.7
3651 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3652 CA.pl for serial number initialization. With the new release 0.9.8,
3653 we can fix the problem directly in the 'ca' utility.)
3656 *) Reduced header interdepencies by declaring more opaque objects in
3657 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3658 give fewer recursive includes, which could break lazy source code - so
3659 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3660 developers should define this symbol when building and using openssl to
3661 ensure they track the recommended behaviour, interfaces, [etc], but
3662 backwards-compatible behaviour prevails when this isn't defined.
3665 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3668 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3669 This will generate a random key of the appropriate length based on the
3670 cipher context. The EVP_CIPHER can provide its own random key generation
3671 routine to support keys of a specific form. This is used in the des and
3672 3des routines to generate a key of the correct parity. Update S/MIME
3673 code to use new functions and hence generate correct parity DES keys.
3674 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3675 valid (weak or incorrect parity).
3678 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3679 as looking them up. This is useful when the verified structure may contain
3680 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3681 present unless the new PKCS7_NO_CRL flag is asserted.
3684 *) Extend ASN1 oid configuration module. It now additionally accepts the
3687 shortName = some long name, 1.2.3.4
3690 *) Reimplemented the BN_CTX implementation. There is now no more static
3691 limitation on the number of variables it can handle nor the depth of the
3692 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3693 information can now expand as required, and rather than having a single
3694 static array of bignums, BN_CTX now uses a linked-list of such arrays
3695 allowing it to expand on demand whilst maintaining the usefulness of
3696 BN_CTX's "bundling".
3699 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3700 to allow all RSA operations to function using a single BN_CTX.
3703 *) Preliminary support for certificate policy evaluation and checking. This
3704 is initially intended to pass the tests outlined in "Conformance Testing
3705 of Relying Party Client Certificate Path Processing Logic" v1.07.
3708 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3709 remained unused and not that useful. A variety of other little bignum
3710 tweaks and fixes have also been made continuing on from the audit (see
3714 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3715 associated ASN1, EVP and SSL functions and old ASN1 macros.
3718 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3719 and this should never fail. So the return value from the use of
3720 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3721 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3724 *) BN_CTX_get() should return zero-valued bignums, providing the same
3725 initialised value as BN_new().
3726 [Geoff Thorpe, suggested by Ulf Möller]
3728 *) Support for inhibitAnyPolicy certificate extension.
3731 *) An audit of the BIGNUM code is underway, for which debugging code is
3732 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3733 is considered valid when processing BIGNUMs, and causes execution to
3734 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3735 further steps are taken to deliberately pollute unused data in BIGNUM
3736 structures to try and expose faulty code further on. For now, openssl will
3737 (in its default mode of operation) continue to tolerate the inconsistent
3738 forms that it has tolerated in the past, but authors and packagers should
3739 consider trying openssl and their own applications when compiled with
3740 these debugging symbols defined. It will help highlight potential bugs in
3741 their own code, and will improve the test coverage for OpenSSL itself. At
3742 some point, these tighter rules will become openssl's default to improve
3743 maintainability, though the assert()s and other overheads will remain only
3744 in debugging configurations. See bn.h for more details.
3745 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3747 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3748 that can only be obtained through BN_CTX_new() (which implicitly
3749 initialises it). The presence of this function only made it possible
3750 to overwrite an existing structure (and cause memory leaks).
3753 *) Because of the callback-based approach for implementing LHASH as a
3754 template type, lh_insert() adds opaque objects to hash-tables and
3755 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3756 to clean up those corresponding objects before destroying the hash table
3757 (and losing the object pointers). So some over-zealous constifications in
3758 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3759 objects as "const" and the lh_doall[_arg] callback wrappers are not
3760 prototyped to have "const" restrictions on the object pointers they are
3761 given (and so aren't required to cast them away any more).
3764 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3765 (speed) prefers to use its own implementation. The two implementations
3766 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3767 its object type properly exposed (MS_TM) instead of casting to/from "char
3768 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3769 aren't necessarily the greatest nomenclatures - but this is what was used
3770 internally to the implementation so I've used that for now.
3773 *) Ensure that deprecated functions do not get compiled when
3774 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3775 the self-tests were still using deprecated key-generation functions so
3776 these have been updated also.
3779 *) Reorganise PKCS#7 code to separate the digest location functionality
3780 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3781 New function PKCS7_set_digest() to set the digest type for PKCS#7
3782 digestedData type. Add additional code to correctly generate the
3783 digestedData type and add support for this type in PKCS7 initialization
3787 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3788 structure of type "other".
3791 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3792 sure the loop does correctly stop and breaking ("division by zero")
3793 modulus operations are not performed. The (pre-generated) prime
3794 table crypto/bn/bn_prime.h was already correct, but it could not be
3795 re-generated on some platforms because of the "division by zero"
3796 situation in the script.
3797 [Ralf S. Engelschall]
3799 *) Update support for ECC-based TLS ciphersuites according to
3800 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3801 SHA-1 now is only used for "small" curves (where the
3802 representation of a field element takes up to 24 bytes); for
3803 larger curves, the field element resulting from ECDH is directly
3804 used as premaster secret.
3805 [Douglas Stebila (Sun Microsystems Laboratories)]
3807 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3808 curve secp160r1 to the tests.
3809 [Douglas Stebila (Sun Microsystems Laboratories)]
3811 *) Add the possibility to load symbols globally with DSO.
3812 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3814 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3815 control of the error stack.
3818 *) Add support for STORE in ENGINE.
3821 *) Add the STORE type. The intention is to provide a common interface
3822 to certificate and key stores, be they simple file-based stores, or
3823 HSM-type store, or LDAP stores, or...
3824 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3827 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3828 pass a list of arguments to any function as well as provide a way
3829 for a function to pass data back to the caller.
3832 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3833 works like BUF_strdup() but can be used to duplicate a portion of
3834 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3838 *) Add the function sk_find_ex() which works like sk_find(), but will
3839 return an index to an element even if an exact match couldn't be
3840 found. The index is guaranteed to point at the element where the
3841 searched-for key would be inserted to preserve sorting order.
3844 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3845 takes an extra flags argument for optional functionality. Currently,
3846 the following flags are defined:
3848 OBJ_BSEARCH_VALUE_ON_NOMATCH
3849 This one gets OBJ_bsearch_ex() to return a pointer to the first
3850 element where the comparing function returns a negative or zero
3853 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3854 This one gets OBJ_bsearch_ex() to return a pointer to the first
3855 element where the comparing function returns zero. This is useful
3856 if there are more than one element where the comparing function
3860 *) Make it possible to create self-signed certificates with 'openssl ca'
3861 in such a way that the self-signed certificate becomes part of the
3862 CA database and uses the same mechanisms for serial number generation
3863 as all other certificate signing. The new flag '-selfsign' enables
3864 this functionality. Adapt CA.sh and CA.pl.in.
3867 *) Add functionality to check the public key of a certificate request
3868 against a given private. This is useful to check that a certificate
3869 request can be signed by that key (self-signing).
3872 *) Make it possible to have multiple active certificates with the same
3873 subject in the CA index file. This is done only if the keyword
3874 'unique_subject' is set to 'no' in the main CA section (default
3875 if 'CA_default') of the configuration file. The value is saved
3876 with the database itself in a separate index attribute file,
3877 named like the index file with '.attr' appended to the name.
3880 *) Generate muti valued AVAs using '+' notation in config files for
3884 *) Support for nameConstraints certificate extension.
3887 *) Support for policyConstraints certificate extension.
3890 *) Support for policyMappings certificate extension.
3893 *) Make sure the default DSA_METHOD implementation only uses its
3894 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3895 and change its own handlers to be NULL so as to remove unnecessary
3896 indirection. This lets alternative implementations fallback to the
3897 default implementation more easily.
3900 *) Support for directoryName in GeneralName related extensions
3904 *) Make it possible to link applications using Makefile.shared.
3905 Make that possible even when linking against static libraries!
3908 *) Support for single pass processing for S/MIME signing. This now
3909 means that S/MIME signing can be done from a pipe, in addition
3910 cleartext signing (multipart/signed type) is effectively streaming
3911 and the signed data does not need to be all held in memory.
3913 This is done with a new flag PKCS7_STREAM. When this flag is set
3914 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3915 is done after the data is output (and digests calculated) in
3916 SMIME_write_PKCS7().
3919 *) Add full support for -rpath/-R, both in shared libraries and
3920 applications, at least on the platforms where it's known how
3924 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3925 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3926 will now compute a table of multiples of the generator that
3927 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3928 faster (notably in the case of a single point multiplication,
3929 scalar * generator).
3930 [Nils Larsch, Bodo Moeller]
3932 *) IPv6 support for certificate extensions. The various extensions
3933 which use the IP:a.b.c.d can now take IPv6 addresses using the
3934 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3938 *) Added an ENGINE that implements RSA by performing private key
3939 exponentiations with the GMP library. The conversions to and from
3940 GMP's mpz_t format aren't optimised nor are any montgomery forms
3941 cached, and on x86 it appears OpenSSL's own performance has caught up.
3942 However there are likely to be other architectures where GMP could
3943 provide a boost. This ENGINE is not built in by default, but it can be
3944 specified at Configure time and should be accompanied by the necessary
3945 linker additions, eg;
3946 ./config -DOPENSSL_USE_GMP -lgmp
3949 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3950 testing availability of engines with "-t" - the old behaviour is
3951 produced by increasing the feature's verbosity with "-tt".
3954 *) ECDSA routines: under certain error conditions uninitialized BN objects
3955 could be freed. Solution: make sure initialization is performed early
3956 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3960 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3961 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3962 software implementations. For DSA and DH, parameter generation can
3963 also be overriden by providing the appropriate method callbacks.
3966 *) Change the "progress" mechanism used in key-generation and
3967 primality testing to functions that take a new BN_GENCB pointer in
3968 place of callback/argument pairs. The new API functions have "_ex"
3969 postfixes and the older functions are reimplemented as wrappers for
3970 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3971 declarations of the old functions to help (graceful) attempts to
3972 migrate to the new functions. Also, the new key-generation API
3973 functions operate on a caller-supplied key-structure and return
3974 success/failure rather than returning a key or NULL - this is to
3975 help make "keygen" another member function of RSA_METHOD etc.
3977 Example for using the new callback interface:
3979 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3983 BN_GENCB_set(&my_cb, my_callback, my_arg);
3985 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3986 /* For the meaning of a, b in calls to my_callback(), see the
3987 * documentation of the function that calls the callback.
3988 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3989 * my_callback should return 1 if it wants BN_is_prime_ex()
3990 * to continue, or 0 to stop.
3995 *) Change the ZLIB compression method to be stateful, and make it
3996 available to TLS with the number defined in
3997 draft-ietf-tls-compression-04.txt.
4000 *) Add the ASN.1 structures and functions for CertificatePair, which
4001 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4003 CertificatePair ::= SEQUENCE {
4004 forward [0] Certificate OPTIONAL,
4005 reverse [1] Certificate OPTIONAL,
4006 -- at least one of the pair shall be present -- }
4008 Also implement the PEM functions to read and write certificate
4009 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4011 This needed to be defined, mostly for the sake of the LDAP
4012 attribute crossCertificatePair, but may prove useful elsewhere as
4016 *) Make it possible to inhibit symlinking of shared libraries in
4017 Makefile.shared, for Cygwin's sake.
4020 *) Extend the BIGNUM API by creating a function
4021 void BN_set_negative(BIGNUM *a, int neg);
4022 and a macro that behave like
4023 int BN_is_negative(const BIGNUM *a);
4025 to avoid the need to access 'a->neg' directly in applications.
4028 *) Implement fast modular reduction for pseudo-Mersenne primes
4029 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4030 EC_GROUP_new_curve_GFp() will now automatically use this
4032 [Nils Larsch <nla@trustcenter.de>]
4034 *) Add new lock type (CRYPTO_LOCK_BN).
4037 *) Change the ENGINE framework to automatically load engines
4038 dynamically from specific directories unless they could be
4039 found to already be built in or loaded. Move all the
4040 current engines except for the cryptodev one to a new
4042 The engines in engines/ are built as shared libraries if
4043 the "shared" options was given to ./Configure or ./config.
4044 Otherwise, they are inserted in libcrypto.a.
4045 /usr/local/ssl/engines is the default directory for dynamic
4046 engines, but that can be overriden at configure time through
4047 the usual use of --prefix and/or --openssldir, and at run
4048 time with the environment variable OPENSSL_ENGINES.
4049 [Geoff Thorpe and Richard Levitte]
4051 *) Add Makefile.shared, a helper makefile to build shared
4052 libraries. Adapt Makefile.org.
4055 *) Add version info to Win32 DLLs.
4056 [Peter 'Luna' Runestig" <peter@runestig.com>]
4058 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4059 can be added using this API to created arbitrary PKCS#12
4060 files while avoiding the low level API.
4062 New options to PKCS12_create(), key or cert can be NULL and
4063 will then be omitted from the output file. The encryption
4064 algorithm NIDs can be set to -1 for no encryption, the mac
4065 iteration count can be set to 0 to omit the mac.
4067 Enhance pkcs12 utility by making the -nokeys and -nocerts
4068 options work when creating a PKCS#12 file. New option -nomac
4069 to omit the mac, NONE can be set for an encryption algorithm.
4070 New code is modified to use the enhanced PKCS12_create()
4071 instead of the low level API.
4074 *) Extend ASN1 encoder to support indefinite length constructed
4075 encoding. This can output sequences tags and octet strings in
4076 this form. Modify pk7_asn1.c to support indefinite length
4077 encoding. This is experimental and needs additional code to
4078 be useful, such as an ASN1 bio and some enhanced streaming
4081 Extend template encode functionality so that tagging is passed
4082 down to the template encoder.
4085 *) Let 'openssl req' fail if an argument to '-newkey' is not
4086 recognized instead of using RSA as a default.
4089 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4090 As these are not official, they are not included in "ALL";
4091 the "ECCdraft" ciphersuite group alias can be used to select them.
4092 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4094 *) Add ECDH engine support.
4095 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4097 *) Add ECDH in new directory crypto/ecdh/.
4098 [Douglas Stebila (Sun Microsystems Laboratories)]
4100 *) Let BN_rand_range() abort with an error after 100 iterations
4101 without success (which indicates a broken PRNG).
4104 *) Change BN_mod_sqrt() so that it verifies that the input value
4105 is really the square of the return value. (Previously,
4106 BN_mod_sqrt would show GIGO behaviour.)
4109 *) Add named elliptic curves over binary fields from X9.62, SECG,
4110 and WAP/WTLS; add OIDs that were still missing.
4112 [Sheueling Chang Shantz and Douglas Stebila
4113 (Sun Microsystems Laboratories)]
4115 *) Extend the EC library for elliptic curves over binary fields
4116 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4119 EC_GF2m_simple_method
4123 EC_GROUP_new_curve_GF2m
4124 EC_GROUP_set_curve_GF2m
4125 EC_GROUP_get_curve_GF2m
4126 EC_POINT_set_affine_coordinates_GF2m
4127 EC_POINT_get_affine_coordinates_GF2m
4128 EC_POINT_set_compressed_coordinates_GF2m
4130 Point compression for binary fields is disabled by default for
4131 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4134 As binary polynomials are represented as BIGNUMs, various members
4135 of the EC_GROUP and EC_POINT data structures can be shared
4136 between the implementations for prime fields and binary fields;
4137 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4138 are essentially identical to their ..._GFp counterparts.
4139 (For simplicity, the '..._GFp' prefix has been dropped from
4140 various internal method names.)
4142 An internal 'field_div' method (similar to 'field_mul' and
4143 'field_sqr') has been added; this is used only for binary fields.
4145 [Sheueling Chang Shantz and Douglas Stebila
4146 (Sun Microsystems Laboratories)]
4148 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4149 through methods ('mul', 'precompute_mult').
4151 The generic implementations (now internally called 'ec_wNAF_mul'
4152 and 'ec_wNAF_precomputed_mult') remain the default if these
4153 methods are undefined.
4155 [Sheueling Chang Shantz and Douglas Stebila
4156 (Sun Microsystems Laboratories)]
4158 *) New function EC_GROUP_get_degree, which is defined through
4159 EC_METHOD. For curves over prime fields, this returns the bit
4160 length of the modulus.
4162 [Sheueling Chang Shantz and Douglas Stebila
4163 (Sun Microsystems Laboratories)]
4165 *) New functions EC_GROUP_dup, EC_POINT_dup.
4166 (These simply call ..._new and ..._copy).
4168 [Sheueling Chang Shantz and Douglas Stebila
4169 (Sun Microsystems Laboratories)]
4171 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4172 Polynomials are represented as BIGNUMs (where the sign bit is not
4173 used) in the following functions [macros]:
4176 BN_GF2m_sub [= BN_GF2m_add]
4177 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4178 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4179 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4181 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4182 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4183 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4184 BN_GF2m_cmp [= BN_ucmp]
4186 (Note that only the 'mod' functions are actually for fields GF(2^m).
4187 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4189 For some functions, an the irreducible polynomial defining a
4190 field can be given as an 'unsigned int[]' with strictly
4191 decreasing elements giving the indices of those bits that are set;
4192 i.e., p[] represents the polynomial
4193 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4195 p[0] > p[1] > ... > p[k] = 0.
4196 This applies to the following functions:
4201 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4202 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4204 BN_GF2m_mod_sqrt_arr
4205 BN_GF2m_mod_solve_quad_arr
4209 Conversion can be performed by the following functions:
4214 bntest.c has additional tests for binary polynomial arithmetic.
4216 Two implementations for BN_GF2m_mod_div() are available.
4217 The default algorithm simply uses BN_GF2m_mod_inv() and
4218 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4219 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4220 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4222 [Sheueling Chang Shantz and Douglas Stebila
4223 (Sun Microsystems Laboratories)]
4225 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4226 functionality is disabled at compile-time.
4227 [Douglas Stebila <douglas.stebila@sun.com>]
4229 *) Change default behaviour of 'openssl asn1parse' so that more
4230 information is visible when viewing, e.g., a certificate:
4232 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4233 mode the content of non-printable OCTET STRINGs is output in a
4234 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4235 avoid the appearance of a printable string.
4236 [Nils Larsch <nla@trustcenter.de>]
4238 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4240 EC_GROUP_set_asn1_flag()
4241 EC_GROUP_get_asn1_flag()
4242 EC_GROUP_set_point_conversion_form()
4243 EC_GROUP_get_point_conversion_form()
4244 These control ASN1 encoding details:
4245 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4246 has been set to OPENSSL_EC_NAMED_CURVE.
4247 - Points are encoded in uncompressed form by default; options for
4248 asn1_for are as for point2oct, namely
4249 POINT_CONVERSION_COMPRESSED
4250 POINT_CONVERSION_UNCOMPRESSED
4251 POINT_CONVERSION_HYBRID
4253 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4256 EC_GROUP_get0_seed()
4257 EC_GROUP_get_seed_len()
4258 This is used only for ASN1 purposes (so far).
4259 [Nils Larsch <nla@trustcenter.de>]
4261 *) Add 'field_type' member to EC_METHOD, which holds the NID
4262 of the appropriate field type OID. The new function
4263 EC_METHOD_get_field_type() returns this value.
4264 [Nils Larsch <nla@trustcenter.de>]
4269 EC_POINT_point2hex()
4270 EC_POINT_hex2point()
4271 providing useful interfaces to EC_POINT_point2oct() and
4272 EC_POINT_oct2point().
4273 [Nils Larsch <nla@trustcenter.de>]
4275 *) Change internals of the EC library so that the functions
4276 EC_GROUP_set_generator()
4277 EC_GROUP_get_generator()
4278 EC_GROUP_get_order()
4279 EC_GROUP_get_cofactor()
4280 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4281 to methods, which would lead to unnecessary code duplication when
4282 adding different types of curves.
4283 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4285 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4286 arithmetic, and such that modified wNAFs are generated
4287 (which avoid length expansion in many cases).
4290 *) Add a function EC_GROUP_check_discriminant() (defined via
4291 EC_METHOD) that verifies that the curve discriminant is non-zero.
4293 Add a function EC_GROUP_check() that makes some sanity tests
4294 on a EC_GROUP, its generator and order. This includes
4295 EC_GROUP_check_discriminant().
4296 [Nils Larsch <nla@trustcenter.de>]
4298 *) Add ECDSA in new directory crypto/ecdsa/.
4300 Add applications 'openssl ecparam' and 'openssl ecdsa'
4301 (these are based on 'openssl dsaparam' and 'openssl dsa').
4303 ECDSA support is also included in various other files across the
4304 library. Most notably,
4305 - 'openssl req' now has a '-newkey ecdsa:file' option;
4306 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4307 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4308 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4309 them suitable for ECDSA where domain parameters must be
4310 extracted before the specific public key;
4311 - ECDSA engine support has been added.
4312 [Nils Larsch <nla@trustcenter.de>]
4314 *) Include some named elliptic curves, and add OIDs from X9.62,
4315 SECG, and WAP/WTLS. Each curve can be obtained from the new
4317 EC_GROUP_new_by_curve_name(),
4318 and the list of available named curves can be obtained with
4319 EC_get_builtin_curves().
4320 Also add a 'curve_name' member to EC_GROUP objects, which can be
4322 EC_GROUP_set_curve_name()
4323 EC_GROUP_get_curve_name()
4324 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4326 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4327 was actually never needed) and in BN_mul(). The removal in BN_mul()
4328 required a small change in bn_mul_part_recursive() and the addition
4329 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4330 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4331 bn_sub_words() and bn_add_words() except they take arrays with
4335 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4337 *) Cleanse PEM buffers before freeing them since they may contain
4339 [Benjamin Bennett <ben@psc.edu>]
4341 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4342 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4343 authentication-only ciphersuites.
4346 *) Since AES128 and AES256 share a single mask bit in the logic of
4347 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4348 kludge to work properly if AES128 is available and AES256 isn't.
4351 *) Expand security boundary to match 1.1.1 module.
4354 *) Remove redundant features: hash file source, editing of test vectors
4355 modify fipsld to use external fips_premain.c signature.
4358 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4359 run algorithm test programs.
4362 *) Make algorithm test programs more tolerant of whitespace.
4365 *) Have SSL/TLS server implementation tolerate "mismatched" record
4366 protocol version while receiving ClientHello even if the
4367 ClientHello is fragmented. (The server can't insist on the
4368 particular protocol version it has chosen before the ServerHello
4369 message has informed the client about his choice.)
4372 *) Load error codes if they are not already present instead of using a
4373 static variable. This allows them to be cleanly unloaded and reloaded.
4376 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4378 *) Introduce limits to prevent malicious keys being able to
4379 cause a denial of service. (CVE-2006-2940)
4380 [Steve Henson, Bodo Moeller]
4382 *) Fix ASN.1 parsing of certain invalid structures that can result
4383 in a denial of service. (CVE-2006-2937) [Steve Henson]
4385 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4386 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4388 *) Fix SSL client code which could crash if connecting to a
4389 malicious SSLv2 server. (CVE-2006-4343)
4390 [Tavis Ormandy and Will Drewry, Google Security Team]
4392 *) Change ciphersuite string processing so that an explicit
4393 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4394 will no longer include "AES128-SHA"), and any other similar
4395 ciphersuite (same bitmap) from *other* protocol versions (so that
4396 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4397 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4398 changes from 0.9.8b and 0.9.8d.
4401 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4403 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4404 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4406 *) Change the Unix randomness entropy gathering to use poll() when
4407 possible instead of select(), since the latter has some
4408 undesirable limitations.
4409 [Darryl Miles via Richard Levitte and Bodo Moeller]
4411 *) Disable rogue ciphersuites:
4413 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4414 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4415 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4417 The latter two were purportedly from
4418 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4421 Also deactive the remaining ciphersuites from
4422 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4423 unofficial, and the ID has long expired.
4426 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4427 dual-core machines) and other potential thread-safety issues.
4430 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4432 *) Adapt fipsld and the build system to link against the validated FIPS
4433 module in FIPS mode.
4436 *) Fixes for VC++ 2005 build under Windows.
4439 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4440 from a Windows bash shell such as MSYS. It is autodetected from the
4441 "config" script when run from a VC++ environment. Modify standard VC++
4442 build to use fipscanister.o from the GNU make build.
4445 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4447 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4448 The value now differs depending on if you build for FIPS or not.
4449 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4450 safely run with a non-FIPSed libcrypto, as it may crash because of
4451 the difference induced by this change.
4454 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4456 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4457 (part of SSL_OP_ALL). This option used to disable the
4458 countermeasure against man-in-the-middle protocol-version
4459 rollback in the SSL 2.0 server implementation, which is a bad
4460 idea. (CVE-2005-2969)
4462 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4463 for Information Security, National Institute of Advanced Industrial
4464 Science and Technology [AIST], Japan)]
4466 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4467 mainly for FIPS compliance and not fully integrated at this stage.
4470 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4471 the exponentiation using a fixed-length exponent. (Otherwise,
4472 the information leaked through timing could expose the secret key
4473 after many signatures; cf. Bleichenbacher's attack on DSA with
4477 *) Make a new fixed-window mod_exp implementation the default for
4478 RSA, DSA, and DH private-key operations so that the sequence of
4479 squares and multiplies and the memory access pattern are
4480 independent of the particular secret key. This will mitigate
4481 cache-timing and potential related attacks.
4483 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4484 and this is automatically used by BN_mod_exp_mont() if the new flag
4485 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4486 will use this BN flag for private exponents unless the flag
4487 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4488 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4490 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4492 *) Change the client implementation for SSLv23_method() and
4493 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4494 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4495 (Previously, the SSL 2.0 backwards compatible Client Hello
4496 message format would be used even with SSL_OP_NO_SSLv2.)
4499 *) Add support for smime-type MIME parameter in S/MIME messages which some
4503 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4504 a threadsafe manner. Modify rsa code to use new function and add calls
4505 to dsa and dh code (which had race conditions before).
4508 *) Include the fixed error library code in the C error file definitions
4509 instead of fixing them up at runtime. This keeps the error code
4510 structures constant.
4513 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4515 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4518 *) Fixes for newer kerberos headers. NB: the casts are needed because
4519 the 'length' field is signed on one version and unsigned on another
4520 with no (?) obvious way to tell the difference, without these VC++
4521 complains. Also the "definition" of FAR (blank) is no longer included
4522 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4523 some needed definitions.
4526 *) Undo Cygwin change.
4529 *) Added support for proxy certificates according to RFC 3820.
4530 Because they may be a security thread to unaware applications,
4531 they must be explicitely allowed in run-time. See
4532 docs/HOWTO/proxy_certificates.txt for further information.
4535 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4537 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4538 server and client random values. Previously
4539 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4540 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4542 This change has negligible security impact because:
4544 1. Server and client random values still have 24 bytes of pseudo random
4547 2. Server and client random values are sent in the clear in the initial
4550 3. The master secret is derived using the premaster secret (48 bytes in
4551 size for static RSA ciphersuites) as well as client server and random
4554 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4557 [Stephen Henson, reported by UK NISCC]
4559 *) Use Windows randomness collection on Cygwin.
4562 *) Fix hang in EGD/PRNGD query when communication socket is closed
4563 prematurely by EGD/PRNGD.
4564 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4566 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4569 *) Back-port of selected performance improvements from development
4570 branch, as well as improved support for PowerPC platforms.
4573 *) Add lots of checks for memory allocation failure, error codes to indicate
4574 failure and freeing up memory if a failure occurs.
4575 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4577 *) Add new -passin argument to dgst.
4580 *) Perform some character comparisons of different types in X509_NAME_cmp:
4581 this is needed for some certificates that reencode DNs into UTF8Strings
4582 (in violation of RFC3280) and can't or wont issue name rollover
4586 *) Make an explicit check during certificate validation to see that
4587 the CA setting in each certificate on the chain is correct. As a
4588 side effect always do the following basic checks on extensions,
4589 not just when there's an associated purpose to the check:
4591 - if there is an unhandled critical extension (unless the user
4592 has chosen to ignore this fault)
4593 - if the path length has been exceeded (if one is set at all)
4594 - that certain extensions fit the associated purpose (if one has
4598 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4600 *) Avoid a race condition when CRLs are checked in a multi threaded
4601 environment. This would happen due to the reordering of the revoked
4602 entries during signature checking and serial number lookup. Now the
4603 encoding is cached and the serial number sort performed under a lock.
4604 Add new STACK function sk_is_sorted().
4607 *) Add Delta CRL to the extension code.
4610 *) Various fixes to s3_pkt.c so alerts are sent properly.
4611 [David Holmes <d.holmes@f5.com>]
4613 *) Reduce the chances of duplicate issuer name and serial numbers (in
4614 violation of RFC3280) using the OpenSSL certificate creation utilities.
4615 This is done by creating a random 64 bit value for the initial serial
4616 number when a serial number file is created or when a self signed
4617 certificate is created using 'openssl req -x509'. The initial serial
4618 number file is created using 'openssl x509 -next_serial' in CA.pl
4619 rather than being initialized to 1.
4622 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4624 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4625 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4626 [Joe Orton, Steve Henson]
4628 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4630 [Joe Orton, Steve Henson]
4632 *) Make it possible to have multiple active certificates with the same
4633 subject in the CA index file. This is done only if the keyword
4634 'unique_subject' is set to 'no' in the main CA section (default
4635 if 'CA_default') of the configuration file. The value is saved
4636 with the database itself in a separate index attribute file,
4637 named like the index file with '.attr' appended to the name.
4640 *) X509 verify fixes. Disable broken certificate workarounds when
4641 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4642 keyUsage extension present. Don't accept CRLs with unhandled critical
4643 extensions: since verify currently doesn't process CRL extensions this
4644 rejects a CRL with *any* critical extensions. Add new verify error codes
4648 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4649 A clarification of RFC2560 will require the use of OCTET STRINGs and
4650 some implementations cannot handle the current raw format. Since OpenSSL
4651 copies and compares OCSP nonces as opaque blobs without any attempt at
4652 parsing them this should not create any compatibility issues.
4655 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4656 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4657 this HMAC (and other) operations are several times slower than OpenSSL
4661 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4662 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4664 *) Use the correct content when signing type "other".
4667 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4669 *) Fix various bugs revealed by running the NISCC test suite:
4671 Stop out of bounds reads in the ASN1 code when presented with
4672 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4674 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4676 If verify callback ignores invalid public key errors don't try to check
4677 certificate signature with the NULL public key.
4681 *) New -ignore_err option in ocsp application to stop the server
4682 exiting on the first error in a request.
4685 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4686 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4690 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4691 extra data after the compression methods not only for TLS 1.0
4692 but also for SSL 3.0 (as required by the specification).
4693 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4695 *) Change X509_certificate_type() to mark the key as exported/exportable
4696 when it's 512 *bits* long, not 512 bytes.
4699 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4700 blocks during encryption.
4703 *) Various fixes to base64 BIO and non blocking I/O. On write
4704 flushes were not handled properly if the BIO retried. On read
4705 data was not being buffered properly and had various logic bugs.
4706 This also affects blocking I/O when the data being decoded is a
4710 *) Various S/MIME bugfixes and compatibility changes:
4711 output correct application/pkcs7 MIME type if
4712 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4713 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4714 of files as .eml work). Correctly handle very long lines in MIME
4718 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4720 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4721 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4722 a protocol version number mismatch like a decryption error
4723 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4726 *) Turn on RSA blinding by default in the default implementation
4727 to avoid a timing attack. Applications that don't want it can call
4728 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4729 They would be ill-advised to do so in most cases.
4730 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4732 *) Change RSA blinding code so that it works when the PRNG is not
4733 seeded (in this case, the secret RSA exponent is abused as
4734 an unpredictable seed -- if it is not unpredictable, there
4735 is no point in blinding anyway). Make RSA blinding thread-safe
4736 by remembering the creator's thread ID in rsa->blinding and
4737 having all other threads use local one-time blinding factors
4738 (this requires more computation than sharing rsa->blinding, but
4739 avoids excessive locking; and if an RSA object is not shared
4740 between threads, blinding will still be very fast).
4743 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4744 ENGINE as defaults for all supported algorithms irrespective of
4745 the 'flags' parameter. 'flags' is now honoured, so applications
4746 should make sure they are passing it correctly.
4749 *) Target "mingw" now allows native Windows code to be generated in
4750 the Cygwin environment as well as with the MinGW compiler.
4753 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4755 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4756 via timing by performing a MAC computation even if incorrect
4757 block cipher padding has been found. This is a countermeasure
4758 against active attacks where the attacker has to distinguish
4759 between bad padding and a MAC verification error. (CVE-2003-0078)
4761 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4762 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4763 Martin Vuagnoux (EPFL, Ilion)]
4765 *) Make the no-err option work as intended. The intention with no-err
4766 is not to have the whole error stack handling routines removed from
4767 libcrypto, it's only intended to remove all the function name and
4768 reason texts, thereby removing some of the footprint that may not
4769 be interesting if those errors aren't displayed anyway.
4771 NOTE: it's still possible for any application or module to have it's
4772 own set of error texts inserted. The routines are there, just not
4773 used by default when no-err is given.
4776 *) Add support for FreeBSD on IA64.
4777 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4779 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4780 Kerberos function mit_des_cbc_cksum(). Before this change,
4781 the value returned by DES_cbc_cksum() was like the one from
4782 mit_des_cbc_cksum(), except the bytes were swapped.
4783 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4785 *) Allow an application to disable the automatic SSL chain building.
4786 Before this a rather primitive chain build was always performed in
4787 ssl3_output_cert_chain(): an application had no way to send the
4788 correct chain if the automatic operation produced an incorrect result.
4790 Now the chain builder is disabled if either:
4792 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4794 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4796 The reasoning behind this is that an application would not want the
4797 auto chain building to take place if extra chain certificates are
4798 present and it might also want a means of sending no additional
4799 certificates (for example the chain has two certificates and the
4803 *) Add the possibility to build without the ENGINE framework.
4804 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4806 *) Under Win32 gmtime() can return NULL: check return value in
4807 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4810 *) DSA routines: under certain error conditions uninitialized BN objects
4811 could be freed. Solution: make sure initialization is performed early
4812 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4813 Nils Larsch <nla@trustcenter.de> via PR#459)
4816 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4817 checked on reconnect on the client side, therefore session resumption
4818 could still fail with a "ssl session id is different" error. This
4819 behaviour is masked when SSL_OP_ALL is used due to
4820 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4821 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4822 followup to PR #377.
4825 *) IA-32 assembler support enhancements: unified ELF targets, support
4826 for SCO/Caldera platforms, fix for Cygwin shared build.
4829 *) Add support for FreeBSD on sparc64. As a consequence, support for
4830 FreeBSD on non-x86 processors is separate from x86 processors on
4831 the config script, much like the NetBSD support.
4832 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4834 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4836 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4839 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4840 code (06) was taken as the first octet of the session ID and the last
4841 octet was ignored consequently. As a result SSLv2 client side session
4842 caching could not have worked due to the session ID mismatch between
4844 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4848 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4849 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4853 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4854 seems that in spite of existing for more than a year, many application
4855 author have done nothing to provide the necessary callbacks, which
4856 means that this particular engine will not work properly anywhere.
4857 This is a very unfortunate situation which forces us, in the name
4858 of usability, to give the hw_ncipher.c a static lock, which is part
4860 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4861 appear in 0.9.8 or later. We EXPECT application authors to have
4862 dealt properly with this when 0.9.8 is released (unless we actually
4863 make such changes in the libcrypto locking code that changes will
4864 have to be made anyway).
4867 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4868 octets have been read, EOF or an error occurs. Without this change
4869 some truncated ASN1 structures will not produce an error.
4872 *) Disable Heimdal support, since it hasn't been fully implemented.
4873 Still give the possibility to force the use of Heimdal, but with
4874 warnings and a request that patches get sent to openssl-dev.
4877 *) Add the VC-CE target, introduce the WINCE sysname, and add
4878 INSTALL.WCE and appropriate conditionals to make it build.
4879 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4881 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4882 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4883 edit numbers of the version.
4884 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4886 *) Introduce safe string copy and catenation functions
4887 (BUF_strlcpy() and BUF_strlcat()).
4888 [Ben Laurie (CHATS) and Richard Levitte]
4890 *) Avoid using fixed-size buffers for one-line DNs.
4891 [Ben Laurie (CHATS)]
4893 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4894 resizing buffers containing secrets, and use where appropriate.
4895 [Ben Laurie (CHATS)]
4897 *) Avoid using fixed size buffers for configuration file location.
4898 [Ben Laurie (CHATS)]
4900 *) Avoid filename truncation for various CA files.
4901 [Ben Laurie (CHATS)]
4903 *) Use sizeof in preference to magic numbers.
4904 [Ben Laurie (CHATS)]
4906 *) Avoid filename truncation in cert requests.
4907 [Ben Laurie (CHATS)]
4909 *) Add assertions to check for (supposedly impossible) buffer
4911 [Ben Laurie (CHATS)]
4913 *) Don't cache truncated DNS entries in the local cache (this could
4914 potentially lead to a spoofing attack).
4915 [Ben Laurie (CHATS)]
4917 *) Fix various buffers to be large enough for hex/decimal
4918 representations in a platform independent manner.
4919 [Ben Laurie (CHATS)]
4921 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4922 resizing buffers containing secrets, and use where appropriate.
4923 [Ben Laurie (CHATS)]
4925 *) Add BIO_indent() to avoid much slightly worrying code to do
4927 [Ben Laurie (CHATS)]
4929 *) Convert sprintf()/BIO_puts() to BIO_printf().
4930 [Ben Laurie (CHATS)]
4932 *) buffer_gets() could terminate with the buffer only half
4934 [Ben Laurie (CHATS)]
4936 *) Add assertions to prevent user-supplied crypto functions from
4937 overflowing internal buffers by having large block sizes, etc.
4938 [Ben Laurie (CHATS)]
4940 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4942 [Ben Laurie (CHATS)]
4944 *) Eliminate unused copy of key in RC4.
4945 [Ben Laurie (CHATS)]
4947 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4948 [Ben Laurie (CHATS)]
4950 *) Fix off-by-one error in EGD path.
4951 [Ben Laurie (CHATS)]
4953 *) If RANDFILE path is too long, ignore instead of truncating.
4954 [Ben Laurie (CHATS)]
4956 *) Eliminate unused and incorrectly sized X.509 structure
4958 [Ben Laurie (CHATS)]
4960 *) Eliminate unused and dangerous function knumber().
4961 [Ben Laurie (CHATS)]
4963 *) Eliminate unused and dangerous structure, KSSL_ERR.
4964 [Ben Laurie (CHATS)]
4966 *) Protect against overlong session ID context length in an encoded
4967 session object. Since these are local, this does not appear to be
4969 [Ben Laurie (CHATS)]
4971 *) Change from security patch (see 0.9.6e below) that did not affect
4972 the 0.9.6 release series:
4974 Remote buffer overflow in SSL3 protocol - an attacker could
4975 supply an oversized master key in Kerberos-enabled versions.
4977 [Ben Laurie (CHATS)]
4979 *) Change the SSL kerb5 codes to match RFC 2712.
4982 *) Make -nameopt work fully for req and add -reqopt switch.
4983 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4985 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4986 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4988 *) Make sure tests can be performed even if the corresponding algorithms
4989 have been removed entirely. This was also the last step to make
4990 OpenSSL compilable with DJGPP under all reasonable conditions.
4991 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4993 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4994 to allow version independent disabling of normally unselected ciphers,
4995 which may be activated as a side-effect of selecting a single cipher.
4997 (E.g., cipher list string "RSA" enables ciphersuites that are left
4998 out of "ALL" because they do not provide symmetric encryption.
4999 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5000 [Lutz Jaenicke, Bodo Moeller]
5002 *) Add appropriate support for separate platform-dependent build
5003 directories. The recommended way to make a platform-dependent
5004 build directory is the following (tested on Linux), maybe with
5007 # Place yourself outside of the OpenSSL source tree. In
5008 # this example, the environment variable OPENSSL_SOURCE
5009 # is assumed to contain the absolute OpenSSL source directory.
5010 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5011 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5012 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5013 mkdir -p `dirname $F`
5014 ln -s $OPENSSL_SOURCE/$F $F
5017 To be absolutely sure not to disturb the source tree, a "make clean"
5018 is a good thing. If it isn't successfull, don't worry about it,
5019 it probably means the source directory is very clean.
5022 *) Make sure any ENGINE control commands make local copies of string
5023 pointers passed to them whenever necessary. Otherwise it is possible
5024 the caller may have overwritten (or deallocated) the original string
5025 data when a later ENGINE operation tries to use the stored values.
5026 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5028 *) Improve diagnostics in file reading and command-line digests.
5029 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5031 *) Add AES modes CFB and OFB to the object database. Correct an
5032 error in AES-CFB decryption.
5035 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5036 allows existing EVP_CIPHER_CTX structures to be reused after
5037 calling EVP_*Final(). This behaviour is used by encryption
5038 BIOs and some applications. This has the side effect that
5039 applications must explicitly clean up cipher contexts with
5040 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5043 *) Check the values of dna and dnb in bn_mul_recursive before calling
5044 bn_mul_comba (a non zero value means the a or b arrays do not contain
5045 n2 elements) and fallback to bn_mul_normal if either is not zero.
5048 *) Fix escaping of non-ASCII characters when using the -subj option
5049 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5052 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5053 form for "surname", serialNumber has no short form.
5054 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5055 therefore remove "mail" short name for "internet 7".
5056 The OID for unique identifiers in X509 certificates is
5057 x500UniqueIdentifier, not uniqueIdentifier.
5058 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5061 *) Add an "init" command to the ENGINE config module and auto initialize
5062 ENGINEs. Without any "init" command the ENGINE will be initialized
5063 after all ctrl commands have been executed on it. If init=1 the
5064 ENGINE is initailized at that point (ctrls before that point are run
5065 on the uninitialized ENGINE and after on the initialized one). If
5066 init=0 then the ENGINE will not be iniatialized at all.
5069 *) Fix the 'app_verify_callback' interface so that the user-defined
5070 argument is actually passed to the callback: In the
5071 SSL_CTX_set_cert_verify_callback() prototype, the callback
5072 declaration has been changed from
5075 int (*cb)(X509_STORE_CTX *,void *);
5076 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5077 i=s->ctx->app_verify_callback(&ctx)
5078 has been changed into
5079 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5081 To update applications using SSL_CTX_set_cert_verify_callback(),
5082 a dummy argument can be added to their callback functions.
5083 [D. K. Smetters <smetters@parc.xerox.com>]
5085 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5086 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5088 *) Add and OPENSSL_LOAD_CONF define which will cause
5089 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5090 This allows older applications to transparently support certain
5091 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5092 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5093 load the config file and OPENSSL_add_all_algorithms_conf() which will
5094 always load it have also been added.
5097 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5098 Adjust NIDs and EVP layer.
5099 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5101 *) Config modules support in openssl utility.
5103 Most commands now load modules from the config file,
5104 though in a few (such as version) this isn't done
5105 because it couldn't be used for anything.
5107 In the case of ca and req the config file used is
5108 the same as the utility itself: that is the -config
5109 command line option can be used to specify an
5113 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5114 use "openssl_conf" if filename is NULL use default openssl config file.
5117 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5118 config section name. Add a new flag to tolerate a missing config file
5119 and move code to CONF_modules_load_file().
5122 *) Support for crypto accelerator cards from Accelerated Encryption
5123 Processing, www.aep.ie. (Use engine 'aep')
5124 The support was copied from 0.9.6c [engine] and adapted/corrected
5125 to work with the new engine framework.
5126 [AEP Inc. and Richard Levitte]
5128 *) Support for SureWare crypto accelerator cards from Baltimore
5129 Technologies. (Use engine 'sureware')
5130 The support was copied from 0.9.6c [engine] and adapted
5131 to work with the new engine framework.
5134 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5135 make the newer ENGINE framework commands for the CHIL engine work.
5136 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5138 *) Make it possible to produce shared libraries on ReliantUNIX.
5139 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5141 *) Add the configuration target debug-linux-ppro.
5142 Make 'openssl rsa' use the general key loading routines
5143 implemented in apps.c, and make those routines able to
5144 handle the key format FORMAT_NETSCAPE and the variant
5146 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5148 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5149 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5151 *) Add -keyform to rsautl, and document -engine.
5152 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5154 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5155 BIO_R_NO_SUCH_FILE error code rather than the generic
5156 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5159 *) Add new functions
5161 ERR_peek_last_error_line
5162 ERR_peek_last_error_line_data.
5163 These are similar to
5166 ERR_peek_error_line_data,
5167 but report on the latest error recorded rather than the first one
5168 still in the error queue.
5169 [Ben Laurie, Bodo Moeller]
5171 *) default_algorithms option in ENGINE config module. This allows things
5173 default_algorithms = ALL
5174 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5177 *) Preliminary ENGINE config module.
5180 *) New experimental application configuration code.
5183 *) Change the AES code to follow the same name structure as all other
5184 symmetric ciphers, and behave the same way. Move everything to
5185 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5186 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5188 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5189 [Ben Laurie and Theo de Raadt]
5191 *) Add option to output public keys in req command.
5192 [Massimiliano Pala madwolf@openca.org]
5194 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5195 (up to about 10% better than before for P-192 and P-224).
5198 *) New functions/macros
5200 SSL_CTX_set_msg_callback(ctx, cb)
5201 SSL_CTX_set_msg_callback_arg(ctx, arg)
5202 SSL_set_msg_callback(ssl, cb)
5203 SSL_set_msg_callback_arg(ssl, arg)
5205 to request calling a callback function
5207 void cb(int write_p, int version, int content_type,
5208 const void *buf, size_t len, SSL *ssl, void *arg)
5210 whenever a protocol message has been completely received
5211 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5212 protocol version according to which the SSL library interprets
5213 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5214 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5215 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5216 specification (change_cipher_spec(20), alert(21), handshake(22)).
5217 'buf' and 'len' point to the actual message, 'ssl' to the
5218 SSL object, and 'arg' is the application-defined value set by
5219 SSL[_CTX]_set_msg_callback_arg().
5221 'openssl s_client' and 'openssl s_server' have new '-msg' options
5222 to enable a callback that displays all protocol messages.
5225 *) Change the shared library support so shared libraries are built as
5226 soon as the corresponding static library is finished, and thereby get
5227 openssl and the test programs linked against the shared library.
5228 This still only happens when the keyword "shard" has been given to
5229 the configuration scripts.
5231 NOTE: shared library support is still an experimental thing, and
5232 backward binary compatibility is still not guaranteed.
5233 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5235 *) Add support for Subject Information Access extension.
5236 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5238 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5239 additional bytes when new memory had to be allocated, not just
5240 when reusing an existing buffer.
5243 *) New command line and configuration option 'utf8' for the req command.
5244 This allows field values to be specified as UTF8 strings.
5247 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5248 runs for the former and machine-readable output for the latter.
5251 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5252 of the e-mail address in the DN (i.e., it will go into a certificate
5253 extension only). The new configuration file option 'email_in_dn = no'
5254 has the same effect.
5255 [Massimiliano Pala madwolf@openca.org]
5257 *) Change all functions with names starting with des_ to be starting
5258 with DES_ instead. Add wrappers that are compatible with libdes,
5259 but are named _ossl_old_des_*. Finally, add macros that map the
5260 des_* symbols to the corresponding _ossl_old_des_* if libdes
5261 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5262 desired, the des_* symbols will be mapped to DES_*, with one
5265 Since we provide two compatibility mappings, the user needs to
5266 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5267 compatibility is desired. The default (i.e., when that macro
5268 isn't defined) is OpenSSL 0.9.6c compatibility.
5270 There are also macros that enable and disable the support of old
5271 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5272 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5273 are defined, the default will apply: to support the old des routines.
5275 In either case, one must include openssl/des.h to get the correct
5276 definitions. Do not try to just include openssl/des_old.h, that
5279 NOTE: This is a major break of an old API into a new one. Software
5280 authors are encouraged to switch to the DES_ style functions. Some
5281 time in the future, des_old.h and the libdes compatibility functions
5282 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5283 default), and then completely removed.
5286 *) Test for certificates which contain unsupported critical extensions.
5287 If such a certificate is found during a verify operation it is
5288 rejected by default: this behaviour can be overridden by either
5289 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5290 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5291 X509_supported_extension() has also been added which returns 1 if a
5292 particular extension is supported.
5295 *) Modify the behaviour of EVP cipher functions in similar way to digests
5296 to retain compatibility with existing code.
5299 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5300 compatibility with existing code. In particular the 'ctx' parameter does
5301 not have to be to be initialized before the call to EVP_DigestInit() and
5302 it is tidied up after a call to EVP_DigestFinal(). New function
5303 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5304 EVP_MD_CTX_copy() changed to not require the destination to be
5305 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5306 requires the destination to be valid.
5308 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5309 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5312 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5313 so that complete 'Handshake' protocol structures are kept in memory
5314 instead of overwriting 'msg_type' and 'length' with 'body' data.
5317 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5318 [Massimo Santin via Richard Levitte]
5320 *) Major restructuring to the underlying ENGINE code. This includes
5321 reduction of linker bloat, separation of pure "ENGINE" manipulation
5322 (initialisation, etc) from functionality dealing with implementations
5323 of specific crypto iterfaces. This change also introduces integrated
5324 support for symmetric ciphers and digest implementations - so ENGINEs
5325 can now accelerate these by providing EVP_CIPHER and EVP_MD
5326 implementations of their own. This is detailed in crypto/engine/README
5327 as it couldn't be adequately described here. However, there are a few
5328 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5329 were changed in the original introduction of ENGINE code have now
5330 reverted back - the hooking from this code to ENGINE is now a good
5331 deal more passive and at run-time, operations deal directly with
5332 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5333 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5334 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5335 they were not being used by the framework as there is no concept of a
5336 BIGNUM_METHOD and they could not be generalised to the new
5337 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5338 ENGINE_cpy() has been removed as it cannot be consistently defined in
5342 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5345 *) Change mkdef.pl to sort symbols that get the same entry number,
5346 and make sure the automatically generated functions ERR_load_*
5347 become part of libeay.num as well.
5350 *) New function SSL_renegotiate_pending(). This returns true once
5351 renegotiation has been requested (either SSL_renegotiate() call
5352 or HelloRequest/ClientHello received from the peer) and becomes
5353 false once a handshake has been completed.
5354 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5355 sends a HelloRequest, but does not ensure that a handshake takes
5356 place. SSL_renegotiate_pending() is useful for checking if the
5357 client has followed the request.)
5360 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5361 By default, clients may request session resumption even during
5362 renegotiation (if session ID contexts permit); with this option,
5363 session resumption is possible only in the first handshake.
5365 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5366 more bits available for options that should not be part of
5367 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5370 *) Add some demos for certificate and certificate request creation.
5373 *) Make maximum certificate chain size accepted from the peer application
5374 settable (SSL*_get/set_max_cert_list()), as proposed by
5375 "Douglas E. Engert" <deengert@anl.gov>.
5378 *) Add support for shared libraries for Unixware-7
5379 (Boyd Lynn Gerber <gerberb@zenez.com>).
5382 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5383 be done prior to destruction. Use this to unload error strings from
5384 ENGINEs that load their own error strings. NB: This adds two new API
5385 functions to "get" and "set" this destroy handler in an ENGINE.
5388 *) Alter all existing ENGINE implementations (except "openssl" and
5389 "openbsd") to dynamically instantiate their own error strings. This
5390 makes them more flexible to be built both as statically-linked ENGINEs
5391 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5392 Also, add stub code to each that makes building them as self-contained
5393 shared-libraries easier (see README.ENGINE).
5396 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5397 implementations into applications that are completely implemented in
5398 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5399 commands that can be used to configure what shared-library to load and
5400 to control aspects of the way it is handled. Also, made an update to
5401 the README.ENGINE file that brings its information up-to-date and
5402 provides some information and instructions on the "dynamic" ENGINE
5403 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5406 *) Make it possible to unload ranges of ERR strings with a new
5407 "ERR_unload_strings" function.
5410 *) Add a copy() function to EVP_MD.
5413 *) Make EVP_MD routines take a context pointer instead of just the
5414 md_data void pointer.
5417 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5418 that the digest can only process a single chunk of data
5419 (typically because it is provided by a piece of
5420 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5421 is only going to provide a single chunk of data, and hence the
5422 framework needn't accumulate the data for oneshot drivers.
5425 *) As with "ERR", make it possible to replace the underlying "ex_data"
5426 functions. This change also alters the storage and management of global
5427 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5428 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5429 index counters. The API functions that use this state have been changed
5430 to take a "class_index" rather than pointers to the class's local STACK
5431 and counter, and there is now an API function to dynamically create new
5432 classes. This centralisation allows us to (a) plug a lot of the
5433 thread-safety problems that existed, and (b) makes it possible to clean
5434 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5435 such data would previously have always leaked in application code and
5436 workarounds were in place to make the memory debugging turn a blind eye
5437 to it. Application code that doesn't use this new function will still
5438 leak as before, but their memory debugging output will announce it now
5439 rather than letting it slide.
5441 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5442 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5443 has a return value to indicate success or failure.
5446 *) Make it possible to replace the underlying "ERR" functions such that the
5447 global state (2 LHASH tables and 2 locks) is only used by the "default"
5448 implementation. This change also adds two functions to "get" and "set"
5449 the implementation prior to it being automatically set the first time
5450 any other ERR function takes place. Ie. an application can call "get",
5451 pass the return value to a module it has just loaded, and that module
5452 can call its own "set" function using that value. This means the
5453 module's "ERR" operations will use (and modify) the error state in the
5454 application and not in its own statically linked copy of OpenSSL code.
5457 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5458 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5459 the operation, and provides a more encapsulated way for external code
5460 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5461 to use these functions rather than manually incrementing the counts.
5463 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5466 *) Add EVP test program.
5469 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5472 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5473 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5474 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5475 These allow a CRL to be built without having to access X509_CRL fields
5476 directly. Modify 'ca' application to use new functions.
5479 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5480 bug workarounds. Rollback attack detection is a security feature.
5481 The problem will only arise on OpenSSL servers when TLSv1 is not
5482 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5483 Software authors not wanting to support TLSv1 will have special reasons
5484 for their choice and can explicitly enable this option.
5485 [Bodo Moeller, Lutz Jaenicke]
5487 *) Rationalise EVP so it can be extended: don't include a union of
5488 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5489 (similar to those existing for EVP_CIPHER_CTX).
5494 EVP_MD_CTX_init(&md); /* new function call */
5495 EVP_DigestInit(&md, EVP_sha1());
5496 EVP_DigestUpdate(&md, in, len);
5497 EVP_DigestFinal(&md, out, NULL);
5498 EVP_MD_CTX_cleanup(&md); /* new function call */
5502 *) Make DES key schedule conform to the usual scheme, as well as
5503 correcting its structure. This means that calls to DES functions
5504 now have to pass a pointer to a des_key_schedule instead of a
5505 plain des_key_schedule (which was actually always a pointer
5508 des_key_schedule ks;
5510 des_set_key_checked(..., &ks);
5511 des_ncbc_encrypt(..., &ks, ...);
5513 (Note that a later change renames 'des_...' into 'DES_...'.)
5516 *) Initial reduction of linker bloat: the use of some functions, such as
5517 PEM causes large amounts of unused functions to be linked in due to
5518 poor organisation. For example pem_all.c contains every PEM function
5519 which has a knock on effect of linking in large amounts of (unused)
5520 ASN1 code. Grouping together similar functions and splitting unrelated
5521 functions prevents this.
5524 *) Cleanup of EVP macros.
5527 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5528 correct _ecb suffix.
5531 *) Add initial OCSP responder support to ocsp application. The
5532 revocation information is handled using the text based index
5533 use by the ca application. The responder can either handle
5534 requests generated internally, supplied in files (for example
5535 via a CGI script) or using an internal minimal server.
5538 *) Add configuration choices to get zlib compression for TLS.
5541 *) Changes to Kerberos SSL for RFC 2712 compliance:
5542 1. Implemented real KerberosWrapper, instead of just using
5543 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5544 2. Implemented optional authenticator field of KerberosWrapper.
5546 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5547 and authenticator structs; see crypto/krb5/.
5549 Generalized Kerberos calls to support multiple Kerberos libraries.
5550 [Vern Staats <staatsvr@asc.hpc.mil>,
5551 Jeffrey Altman <jaltman@columbia.edu>
5552 via Richard Levitte]
5554 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5555 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5556 values for each of the key sizes rather than having just
5557 parameters (and 'speed' generating keys each time).
5560 *) Speed up EVP routines.
5563 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5564 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5565 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5566 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5568 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5569 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5570 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5573 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5575 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5578 *) Added the OS2-EMX target.
5579 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5581 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5582 to support NCONF routines in extension code. New function CONF_set_nconf()
5583 to allow functions which take an NCONF to also handle the old LHASH
5584 structure: this means that the old CONF compatible routines can be
5585 retained (in particular wrt extensions) without having to duplicate the
5586 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5589 *) Enhance the general user interface with mechanisms for inner control
5590 and with possibilities to have yes/no kind of prompts.
5593 *) Change all calls to low level digest routines in the library and
5594 applications to use EVP. Add missing calls to HMAC_cleanup() and
5595 don't assume HMAC_CTX can be copied using memcpy().
5596 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5598 *) Add the possibility to control engines through control names but with
5599 arbitrary arguments instead of just a string.
5600 Change the key loaders to take a UI_METHOD instead of a callback
5601 function pointer. NOTE: this breaks binary compatibility with earlier
5602 versions of OpenSSL [engine].
5603 Adapt the nCipher code for these new conditions and add a card insertion
5607 *) Enhance the general user interface with mechanisms to better support
5608 dialog box interfaces, application-defined prompts, the possibility
5609 to use defaults (for example default passwords from somewhere else)
5610 and interrupts/cancellations.
5613 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5614 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5617 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5618 tidy up some unnecessarily weird code in 'sk_new()').
5619 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5621 *) Change the key loading routines for ENGINEs to use the same kind
5622 callback (pem_password_cb) as all other routines that need this
5626 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5627 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5628 than this minimum value is recommended.
5631 *) New random seeder for OpenVMS, using the system process statistics
5632 that are easily reachable.
5635 *) Windows apparently can't transparently handle global
5636 variables defined in DLLs. Initialisations such as:
5638 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5640 wont compile. This is used by the any applications that need to
5641 declare their own ASN1 modules. This was fixed by adding the option
5642 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5643 needed for static libraries under Win32.
5646 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5647 setting of purpose and trust fields. New X509_STORE trust and
5648 purpose functions and tidy up setting in other SSL functions.
5651 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5652 structure. These are inherited by X509_STORE_CTX when it is
5653 initialised. This allows various defaults to be set in the
5654 X509_STORE structure (such as flags for CRL checking and custom
5655 purpose or trust settings) for functions which only use X509_STORE_CTX
5656 internally such as S/MIME.
5658 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5659 trust settings if they are not set in X509_STORE. This allows X509_STORE
5660 purposes and trust (in S/MIME for example) to override any set by default.
5662 Add command line options for CRL checking to smime, s_client and s_server
5666 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5667 are set then the CRL is looked up in the X509_STORE structure and
5668 its validity and signature checked, then if the certificate is found
5669 in the CRL the verify fails with a revoked error.
5671 Various new CRL related callbacks added to X509_STORE_CTX structure.
5673 Command line options added to 'verify' application to support this.
5675 This needs some additional work, such as being able to handle multiple
5676 CRLs with different times, extension based lookup (rather than just
5677 by subject name) and ultimately more complete V2 CRL extension
5681 *) Add a general user interface API (crypto/ui/). This is designed
5682 to replace things like des_read_password and friends (backward
5683 compatibility functions using this new API are provided).
5684 The purpose is to remove prompting functions from the DES code
5685 section as well as provide for prompting through dialog boxes in
5686 a window system and the like.
5689 *) Add "ex_data" support to ENGINE so implementations can add state at a
5690 per-structure level rather than having to store it globally.
5693 *) Make it possible for ENGINE structures to be copied when retrieved by
5694 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5695 This causes the "original" ENGINE structure to act like a template,
5696 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5697 operational state can be localised to each ENGINE structure, despite the
5698 fact they all share the same "methods". New ENGINE structures returned in
5699 this case have no functional references and the return value is the single
5700 structural reference. This matches the single structural reference returned
5701 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5705 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5706 needs to match any other type at all we need to manually clear the
5710 *) Changes to the "openssl engine" utility to include;
5711 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5712 about an ENGINE's available control commands.
5713 - executing control commands from command line arguments using the
5714 '-pre' and '-post' switches. '-post' is only used if '-t' is
5715 specified and the ENGINE is successfully initialised. The syntax for
5716 the individual commands are colon-separated, for example;
5717 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5720 *) New dynamic control command support for ENGINEs. ENGINEs can now
5721 declare their own commands (numbers), names (strings), descriptions,
5722 and input types for run-time discovery by calling applications. A
5723 subset of these commands are implicitly classed as "executable"
5724 depending on their input type, and only these can be invoked through
5725 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5726 can be based on user input, config files, etc). The distinction is
5727 that "executable" commands cannot return anything other than a boolean
5728 result and can only support numeric or string input, whereas some
5729 discoverable commands may only be for direct use through
5730 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5731 pointers, or other custom uses. The "executable" commands are to
5732 support parameterisations of ENGINE behaviour that can be
5733 unambiguously defined by ENGINEs and used consistently across any
5734 OpenSSL-based application. Commands have been added to all the
5735 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5736 control over shared-library paths without source code alterations.
5739 *) Changed all ENGINE implementations to dynamically allocate their
5740 ENGINEs rather than declaring them statically. Apart from this being
5741 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5742 this also allows the implementations to compile without using the
5743 internal engine_int.h header.
5746 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5747 'const' value. Any code that should be able to modify a RAND_METHOD
5748 should already have non-const pointers to it (ie. they should only
5749 modify their own ones).
5752 *) Made a variety of little tweaks to the ENGINE code.
5753 - "atalla" and "ubsec" string definitions were moved from header files
5754 to C code. "nuron" string definitions were placed in variables
5755 rather than hard-coded - allowing parameterisation of these values
5756 later on via ctrl() commands.
5757 - Removed unused "#if 0"'d code.
5758 - Fixed engine list iteration code so it uses ENGINE_free() to release
5759 structural references.
5760 - Constified the RAND_METHOD element of ENGINE structures.
5761 - Constified various get/set functions as appropriate and added
5762 missing functions (including a catch-all ENGINE_cpy that duplicates
5763 all ENGINE values onto a new ENGINE except reference counts/state).
5764 - Removed NULL parameter checks in get/set functions. Setting a method
5765 or function to NULL is a way of cancelling out a previously set
5766 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5767 and doesn't justify the extra error symbols and code.
5768 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5769 flags from engine_int.h to engine.h.
5770 - Changed prototypes for ENGINE handler functions (init(), finish(),
5771 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5774 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5775 to the algorithm using long division. The binary algorithm can be
5776 used only if the modulus is odd. On 32-bit systems, it is faster
5777 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5778 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5779 up to 450 bits. In 64-bit environments, the binary algorithm
5780 appears to be advantageous for much longer moduli; here we use it
5781 for moduli up to 2048 bits.
5784 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5785 could not support the combine flag in choice fields.
5788 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5789 extensions from a certificate request to the certificate.
5792 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5793 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5794 file: this allows the display of the certificate about to be
5795 signed to be customised, to allow certain fields to be included
5796 or excluded and extension details. The old system didn't display
5797 multicharacter strings properly, omitted fields not in the policy
5798 and couldn't display additional details such as extensions.
5801 *) Function EC_POINTs_mul for multiple scalar multiplication
5802 of an arbitrary number of elliptic curve points
5803 \sum scalars[i]*points[i],
5804 optionally including the generator defined for the EC_GROUP:
5805 scalar*generator + \sum scalars[i]*points[i].
5807 EC_POINT_mul is a simple wrapper function for the typical case
5808 that the point list has just one item (besides the optional
5812 *) First EC_METHODs for curves over GF(p):
5814 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5815 operations and provides various method functions that can also
5816 operate with faster implementations of modular arithmetic.
5818 EC_GFp_mont_method() reuses most functions that are part of
5819 EC_GFp_simple_method, but uses Montgomery arithmetic.
5821 [Bodo Moeller; point addition and point doubling
5822 implementation directly derived from source code provided by
5823 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5825 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5826 crypto/ec/ec_lib.c):
5828 Curves are EC_GROUP objects (with an optional group generator)
5829 based on EC_METHODs that are built into the library.
5831 Points are EC_POINT objects based on EC_GROUP objects.
5833 Most of the framework would be able to handle curves over arbitrary
5834 finite fields, but as there are no obvious types for fields other
5835 than GF(p), some functions are limited to that for now.
5838 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5839 that the file contains a complete HTTP response.
5842 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5843 change the def and num file printf format specifier from "%-40sXXX"
5844 to "%-39s XXX". The latter will always guarantee a space after the
5845 field while the former will cause them to run together if the field
5846 is 40 of more characters long.
5849 *) Constify the cipher and digest 'method' functions and structures
5850 and modify related functions to take constant EVP_MD and EVP_CIPHER
5854 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5855 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5858 *) Modify EVP_Digest*() routines so they now return values. Although the
5859 internal software routines can never fail additional hardware versions
5863 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5865 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5866 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5869 ERR_R_NESTED_ASN1_ERROR
5871 ERR_R_MISSING_ASN1_EOS
5872 were 4 .. 9, conflicting with
5873 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5875 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5876 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5878 Add new error code 'ERR_R_INTERNAL_ERROR'.
5881 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5885 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5886 sets the subject name for a new request or supersedes the
5887 subject name in a given request. Formats that can be parsed are
5888 'CN=Some Name, OU=myOU, C=IT'
5890 'CN=Some Name/OU=myOU/C=IT'.
5892 Add options '-batch' and '-verbose' to 'openssl req'.
5893 [Massimiliano Pala <madwolf@hackmasters.net>]
5895 *) Introduce the possibility to access global variables through
5896 functions on platform were that's the best way to handle exporting
5897 global variables in shared libraries. To enable this functionality,
5898 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5899 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5900 is normally done by Configure or something similar).
5902 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5903 in the source file (foo.c) like this:
5905 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5906 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5908 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5909 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5911 OPENSSL_DECLARE_GLOBAL(int,foo);
5912 #define foo OPENSSL_GLOBAL_REF(foo)
5913 OPENSSL_DECLARE_GLOBAL(double,bar);
5914 #define bar OPENSSL_GLOBAL_REF(bar)
5916 The #defines are very important, and therefore so is including the
5917 header file everywhere where the defined globals are used.
5919 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5920 of ASN.1 items, but that structure is a bit different.
5922 The largest change is in util/mkdef.pl which has been enhanced with
5923 better and easier to understand logic to choose which symbols should
5924 go into the Windows .def files as well as a number of fixes and code
5925 cleanup (among others, algorithm keywords are now sorted
5926 lexicographically to avoid constant rewrites).
5929 *) In BN_div() keep a copy of the sign of 'num' before writing the
5930 result to 'rm' because if rm==num the value will be overwritten
5931 and produce the wrong result if 'num' is negative: this caused
5932 problems with BN_mod() and BN_nnmod().
5935 *) Function OCSP_request_verify(). This checks the signature on an
5936 OCSP request and verifies the signer certificate. The signer
5937 certificate is just checked for a generic purpose and OCSP request
5941 *) Add OCSP_check_validity() function to check the validity of OCSP
5942 responses. OCSP responses are prepared in real time and may only
5943 be a few seconds old. Simply checking that the current time lies
5944 between thisUpdate and nextUpdate max reject otherwise valid responses
5945 caused by either OCSP responder or client clock inaccuracy. Instead
5946 we allow thisUpdate and nextUpdate to fall within a certain period of
5947 the current time. The age of the response can also optionally be
5948 checked. Two new options -validity_period and -status_age added to
5952 *) If signature or public key algorithm is unrecognized print out its
5953 OID rather that just UNKNOWN.
5956 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5957 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5958 ID to be generated from the issuer certificate alone which can then be
5959 passed to OCSP_id_issuer_cmp().
5962 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5963 ASN1 modules to export functions returning ASN1_ITEM pointers
5964 instead of the ASN1_ITEM structures themselves. This adds several
5965 new macros which allow the underlying ASN1 function/structure to
5966 be accessed transparently. As a result code should not use ASN1_ITEM
5967 references directly (such as &X509_it) but instead use the relevant
5968 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5969 use of the new ASN1 code on platforms where exporting structures
5970 is problematical (for example in shared libraries) but exporting
5971 functions returning pointers to structures is not.
5974 *) Add support for overriding the generation of SSL/TLS session IDs.
5975 These callbacks can be registered either in an SSL_CTX or per SSL.
5976 The purpose of this is to allow applications to control, if they wish,
5977 the arbitrary values chosen for use as session IDs, particularly as it
5978 can be useful for session caching in multiple-server environments. A
5979 command-line switch for testing this (and any client code that wishes
5980 to use such a feature) has been added to "s_server".
5981 [Geoff Thorpe, Lutz Jaenicke]
5983 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5984 of the form '#if defined(...) || defined(...) || ...' and
5985 '#if !defined(...) && !defined(...) && ...'. This also avoids
5986 the growing number of special cases it was previously handling.
5989 *) Make all configuration macros available for application by making
5990 sure they are available in opensslconf.h, by giving them names starting
5991 with "OPENSSL_" to avoid conflicts with other packages and by making
5992 sure e_os2.h will cover all platform-specific cases together with
5994 Additionally, it is now possible to define configuration/platform-
5995 specific names (called "system identities"). In the C code, these
5996 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5997 macro with the name beginning with "OPENSSL_SYS_", which is determined
5998 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6002 *) New option -set_serial to 'req' and 'x509' this allows the serial
6003 number to use to be specified on the command line. Previously self
6004 signed certificates were hard coded with serial number 0 and the
6005 CA options of 'x509' had to use a serial number in a file which was
6009 *) New options to 'ca' utility to support V2 CRL entry extensions.
6010 Currently CRL reason, invalidity date and hold instruction are
6011 supported. Add new CRL extensions to V3 code and some new objects.
6014 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6015 disable standard block padding (aka PKCS#5 padding) in the EVP
6016 API, which was previously mandatory. This means that the data is
6017 not padded in any way and so the total length much be a multiple
6018 of the block size, otherwise an error occurs.
6021 *) Initial (incomplete) OCSP SSL support.
6024 *) New function OCSP_parse_url(). This splits up a URL into its host,
6025 port and path components: primarily to parse OCSP URLs. New -url
6026 option to ocsp utility.
6029 *) New nonce behavior. The return value of OCSP_check_nonce() now
6030 reflects the various checks performed. Applications can decide
6031 whether to tolerate certain situations such as an absent nonce
6032 in a response when one was present in a request: the ocsp application
6033 just prints out a warning. New function OCSP_add1_basic_nonce()
6034 this is to allow responders to include a nonce in a response even if
6035 the request is nonce-less.
6038 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6039 skipped when using openssl x509 multiple times on a single input file,
6040 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6043 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6044 set string type: to handle setting ASN1_TIME structures. Fix ca
6045 utility to correctly initialize revocation date of CRLs.
6048 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6049 the clients preferred ciphersuites and rather use its own preferences.
6050 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6051 Internet Explorer by ensuring unchanged hash method during stepup.
6052 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6055 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6056 to aes and add a new 'exist' option to print out symbols that don't
6060 *) Additional options to ocsp utility to allow flags to be set and
6061 additional certificates supplied.
6064 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6065 OCSP client a number of certificate to only verify the response
6069 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6070 handle the new API. Currently only ECB, CBC modes supported. Add new
6073 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6074 Encryption Standard (AES) Ciphersuites for Transport Layer
6075 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6076 not enabled by default and were not part of the "ALL" ciphersuite
6077 alias because they were not yet official; they could be
6078 explicitly requested by specifying the "AESdraft" ciphersuite
6079 group alias. In the final release of OpenSSL 0.9.7, the group
6080 alias is called "AES" and is part of "ALL".)
6081 [Ben Laurie, Steve Henson, Bodo Moeller]
6083 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6084 request to response.
6087 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6088 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6089 extract information from a certificate request. OCSP_response_create()
6090 creates a response and optionally adds a basic response structure.
6091 OCSP_basic_add1_status() adds a complete single response to a basic
6092 response and returns the OCSP_SINGLERESP structure just added (to allow
6093 extensions to be included for example). OCSP_basic_add1_cert() adds a
6094 certificate to a basic response and OCSP_basic_sign() signs a basic
6095 response with various flags. New helper functions ASN1_TIME_check()
6096 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6097 (converts ASN1_TIME to GeneralizedTime).
6100 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6101 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6102 structure from a certificate. X509_pubkey_digest() digests the public_key
6103 contents: this is used in various key identifiers.
6106 *) Make sk_sort() tolerate a NULL argument.
6107 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6109 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6110 passed by the function are trusted implicitly. If any of them signed the
6111 response then it is assumed to be valid and is not verified.
6114 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6115 to data. This was previously part of the PKCS7 ASN1 code. This
6116 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6117 [Steve Henson, reported by Kenneth R. Robinette
6118 <support@securenetterm.com>]
6120 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6121 routines: without these tracing memory leaks is very painful.
6122 Fix leaks in PKCS12 and PKCS7 routines.
6125 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6126 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6127 effectively meant GeneralizedTime would never be used. Now it
6128 is initialised to -1 but X509_time_adj() now has to check the value
6129 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6130 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6131 [Steve Henson, reported by Kenneth R. Robinette
6132 <support@securenetterm.com>]
6134 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6135 result in a zero length in the ASN1_INTEGER structure which was
6136 not consistent with the structure when d2i_ASN1_INTEGER() was used
6137 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6138 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6139 where it did not print out a minus for negative ASN1_INTEGER.
6142 *) Add summary printout to ocsp utility. The various functions which
6143 convert status values to strings have been renamed to:
6144 OCSP_response_status_str(), OCSP_cert_status_str() and
6145 OCSP_crl_reason_str() and are no longer static. New options
6146 to verify nonce values and to disable verification. OCSP response
6147 printout format cleaned up.
6150 *) Add additional OCSP certificate checks. These are those specified
6151 in RFC2560. This consists of two separate checks: the CA of the
6152 certificate being checked must either be the OCSP signer certificate
6153 or the issuer of the OCSP signer certificate. In the latter case the
6154 OCSP signer certificate must contain the OCSP signing extended key
6155 usage. This check is performed by attempting to match the OCSP
6156 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6157 in the OCSP_CERTID structures of the response.
6160 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6161 and related routines. This uses the standard OpenSSL certificate
6162 verify routines to perform initial checks (just CA validity) and
6163 to obtain the certificate chain. Then additional checks will be
6164 performed on the chain. Currently the root CA is checked to see
6165 if it is explicitly trusted for OCSP signing. This is used to set
6166 a root CA as a global signing root: that is any certificate that
6167 chains to that CA is an acceptable OCSP signing certificate.
6170 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6171 extensions from a separate configuration file.
6172 As when reading extensions from the main configuration file,
6173 the '-extensions ...' option may be used for specifying the
6175 [Massimiliano Pala <madwolf@comune.modena.it>]
6177 *) New OCSP utility. Allows OCSP requests to be generated or
6178 read. The request can be sent to a responder and the output
6179 parsed, outputed or printed in text form. Not complete yet:
6180 still needs to check the OCSP response validity.
6183 *) New subcommands for 'openssl ca':
6184 'openssl ca -status <serial>' prints the status of the cert with
6185 the given serial number (according to the index file).
6186 'openssl ca -updatedb' updates the expiry status of certificates
6188 [Massimiliano Pala <madwolf@comune.modena.it>]
6190 *) New '-newreq-nodes' command option to CA.pl. This is like
6191 '-newreq', but calls 'openssl req' with the '-nodes' option
6192 so that the resulting key is not encrypted.
6193 [Damien Miller <djm@mindrot.org>]
6195 *) New configuration for the GNU Hurd.
6196 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6198 *) Initial code to implement OCSP basic response verify. This
6199 is currently incomplete. Currently just finds the signer's
6200 certificate and verifies the signature on the response.
6203 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6204 value of OPENSSLDIR. This is available via the new '-d' option
6205 to 'openssl version', and is also included in 'openssl version -a'.
6208 *) Allowing defining memory allocation callbacks that will be given
6209 file name and line number information in additional arguments
6210 (a const char* and an int). The basic functionality remains, as
6211 well as the original possibility to just replace malloc(),
6212 realloc() and free() by functions that do not know about these
6213 additional arguments. To register and find out the current
6214 settings for extended allocation functions, the following
6215 functions are provided:
6217 CRYPTO_set_mem_ex_functions
6218 CRYPTO_set_locked_mem_ex_functions
6219 CRYPTO_get_mem_ex_functions
6220 CRYPTO_get_locked_mem_ex_functions
6222 These work the same way as CRYPTO_set_mem_functions and friends.
6223 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6224 extended allocation function is enabled.
6225 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6226 a conventional allocation function is enabled.
6227 [Richard Levitte, Bodo Moeller]
6229 *) Finish off removing the remaining LHASH function pointer casts.
6230 There should no longer be any prototype-casting required when using
6231 the LHASH abstraction, and any casts that remain are "bugs". See
6232 the callback types and macros at the head of lhash.h for details
6233 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6236 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6237 If /dev/[u]random devices are not available or do not return enough
6238 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6240 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6241 /etc/entropy will be queried once each in this sequence, quering stops
6242 when enough entropy was collected without querying more sockets.
6245 *) Change the Unix RAND_poll() variant to be able to poll several
6246 random devices, as specified by DEVRANDOM, until a sufficient amount
6247 of data has been collected. We spend at most 10 ms on each file
6248 (select timeout) and read in non-blocking mode. DEVRANDOM now
6249 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6250 (previously it was just the string "/dev/urandom"), so on typical
6251 platforms the 10 ms delay will never occur.
6252 Also separate out the Unix variant to its own file, rand_unix.c.
6253 For VMS, there's a currently-empty rand_vms.c.
6256 *) Move OCSP client related routines to ocsp_cl.c. These
6257 provide utility functions which an application needing
6258 to issue a request to an OCSP responder and analyse the
6259 response will typically need: as opposed to those which an
6260 OCSP responder itself would need which will be added later.
6262 OCSP_request_sign() signs an OCSP request with an API similar
6263 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6264 response. OCSP_response_get1_basic() extracts basic response
6265 from response. OCSP_resp_find_status(): finds and extracts status
6266 information from an OCSP_CERTID structure (which will be created
6267 when the request structure is built). These are built from lower
6268 level functions which work on OCSP_SINGLERESP structures but
6269 wont normally be used unless the application wishes to examine
6270 extensions in the OCSP response for example.
6272 Replace nonce routines with a pair of functions.
6273 OCSP_request_add1_nonce() adds a nonce value and optionally
6274 generates a random value. OCSP_check_nonce() checks the
6275 validity of the nonce in an OCSP response.
6278 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6279 This doesn't copy the supplied OCSP_CERTID and avoids the
6280 need to free up the newly created id. Change return type
6281 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6282 This can then be used to add extensions to the request.
6283 Deleted OCSP_request_new(), since most of its functionality
6284 is now in OCSP_REQUEST_new() (and the case insensitive name
6285 clash) apart from the ability to set the request name which
6286 will be added elsewhere.
6289 *) Update OCSP API. Remove obsolete extensions argument from
6290 various functions. Extensions are now handled using the new
6291 OCSP extension code. New simple OCSP HTTP function which
6292 can be used to send requests and parse the response.
6295 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6296 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6297 uses the special reorder version of SET OF to sort the attributes
6298 and reorder them to match the encoded order. This resolves a long
6299 standing problem: a verify on a PKCS7 structure just after signing
6300 it used to fail because the attribute order did not match the
6301 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6302 it uses the received order. This is necessary to tolerate some broken
6303 software that does not order SET OF. This is handled by encoding
6304 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6305 to produce the required SET OF.
6308 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6309 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6310 files to get correct declarations of the ASN.1 item variables.
6313 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6314 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6315 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6316 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6317 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6318 ASN1_ITEM and no wrapper functions.
6321 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6322 replace the old function pointer based I/O routines. Change most of
6323 the *_d2i_bio() and *_d2i_fp() functions to use these.
6326 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6327 lines, recognice more "algorithms" that can be deselected, and make
6328 it complain about algorithm deselection that isn't recognised.
6331 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6332 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6333 to use new functions. Add NO_ASN1_OLD which can be set to remove
6334 some old style ASN1 functions: this can be used to determine if old
6335 code will still work when these eventually go away.
6338 *) New extension functions for OCSP structures, these follow the
6339 same conventions as certificates and CRLs.
6342 *) New function X509V3_add1_i2d(). This automatically encodes and
6343 adds an extension. Its behaviour can be customised with various
6344 flags to append, replace or delete. Various wrappers added for
6345 certifcates and CRLs.
6348 *) Fix to avoid calling the underlying ASN1 print routine when
6349 an extension cannot be parsed. Correct a typo in the
6350 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6353 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6354 entries for variables.
6357 *) Add functionality to apps/openssl.c for detecting locking
6358 problems: As the program is single-threaded, all we have
6359 to do is register a locking callback using an array for
6360 storing which locks are currently held by the program.
6363 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6364 SSL_get_ex_data_X509_STORE_idx(), which is used in
6365 ssl_verify_cert_chain() and thus can be called at any time
6366 during TLS/SSL handshakes so that thread-safety is essential.
6367 Unfortunately, the ex_data design is not at all suited
6368 for multi-threaded use, so it probably should be abolished.
6371 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6372 [Broadcom, tweaked and integrated by Geoff Thorpe]
6374 *) Move common extension printing code to new function
6375 X509V3_print_extensions(). Reorganise OCSP print routines and
6376 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6379 *) New function X509_signature_print() to remove duplication in some
6383 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6384 set (this was treated exactly the same as SET OF previously). This
6385 is used to reorder the STACK representing the structure to match the
6386 encoding. This will be used to get round a problem where a PKCS7
6387 structure which was signed could not be verified because the STACK
6388 order did not reflect the encoded order.
6391 *) Reimplement the OCSP ASN1 module using the new code.
6394 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6395 for its ASN1 operations. The old style function pointers still exist
6396 for now but they will eventually go away.
6399 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6400 completely replaces the old ASN1 functionality with a table driven
6401 encoder and decoder which interprets an ASN1_ITEM structure describing
6402 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6403 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6404 has also been converted to the new form.
6407 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6408 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6409 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6410 for negative moduli.
6413 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6414 of not touching the result's sign bit.
6417 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6421 *) Changed the LHASH code to use prototypes for callbacks, and created
6422 macros to declare and implement thin (optionally static) functions
6423 that provide type-safety and avoid function pointer casting for the
6424 type-specific callbacks.
6427 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6429 [Veers Staats <staatsvr@asc.hpc.mil>,
6430 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6432 *) Reformat the FAQ so the different questions and answers can be divided
6433 in sections depending on the subject.
6436 *) Have the zlib compression code load ZLIB.DLL dynamically under
6440 *) New function BN_mod_sqrt for computing square roots modulo a prime
6441 (using the probabilistic Tonelli-Shanks algorithm unless
6442 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6443 be handled deterministically).
6444 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6446 *) Make BN_mod_inverse faster by explicitly handling small quotients
6447 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6448 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6451 *) New function BN_kronecker.
6454 *) Fix BN_gcd so that it works on negative inputs; the result is
6455 positive unless both parameters are zero.
6456 Previously something reasonably close to an infinite loop was
6457 possible because numbers could be growing instead of shrinking
6458 in the implementation of Euclid's algorithm.
6461 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6462 sign of the number in question.
6464 Fix BN_is_word(a,w) to work correctly for w == 0.
6466 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6467 because its test if the absolute value of 'a' equals 'w'.
6468 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6469 it exists mostly for use in the implementations of BN_is_zero(),
6470 BN_is_one(), and BN_is_word().
6473 *) New function BN_swap.
6476 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6477 the exponentiation functions are more likely to produce reasonable
6478 results on negative inputs.
6481 *) Change BN_mod_mul so that the result is always non-negative.
6482 Previously, it could be negative if one of the factors was negative;
6483 I don't think anyone really wanted that behaviour.
6486 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6487 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6488 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6489 and add new functions:
6498 BN_mod_lshift1_quick
6502 These functions always generate non-negative results.
6504 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6505 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6507 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6508 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6509 be reduced modulo m.
6510 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6513 The following entry accidentally appeared in the CHANGES file
6514 distributed with OpenSSL 0.9.7. The modifications described in
6515 it do *not* apply to OpenSSL 0.9.7.
6517 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6518 was actually never needed) and in BN_mul(). The removal in BN_mul()
6519 required a small change in bn_mul_part_recursive() and the addition
6520 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6521 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6522 bn_sub_words() and bn_add_words() except they take arrays with
6527 *) In 'openssl passwd', verify passwords read from the terminal
6528 unless the '-salt' option is used (which usually means that
6529 verification would just waste user's time since the resulting
6530 hash is going to be compared with some given password hash)
6531 or the new '-noverify' option is used.
6533 This is an incompatible change, but it does not affect
6534 non-interactive use of 'openssl passwd' (passwords on the command
6535 line, '-stdin' option, '-in ...' option) and thus should not
6539 *) Remove all references to RSAref, since there's no more need for it.
6542 *) Make DSO load along a path given through an environment variable
6543 (SHLIB_PATH) with shl_load().
6546 *) Constify the ENGINE code as a result of BIGNUM constification.
6547 Also constify the RSA code and most things related to it. In a
6548 few places, most notable in the depth of the ASN.1 code, ugly
6549 casts back to non-const were required (to be solved at a later
6553 *) Make it so the openssl application has all engines loaded by default.
6556 *) Constify the BIGNUM routines a little more.
6559 *) Add the following functions:
6561 ENGINE_load_cswift()
6563 ENGINE_load_atalla()
6565 ENGINE_load_builtin_engines()
6567 That way, an application can itself choose if external engines that
6568 are built-in in OpenSSL shall ever be used or not. The benefit is
6569 that applications won't have to be linked with libdl or other dso
6570 libraries unless it's really needed.
6572 Changed 'openssl engine' to load all engines on demand.
6573 Changed the engine header files to avoid the duplication of some
6574 declarations (they differed!).
6577 *) 'openssl engine' can now list capabilities.
6580 *) Better error reporting in 'openssl engine'.
6583 *) Never call load_dh_param(NULL) in s_server.
6586 *) Add engine application. It can currently list engines by name and
6587 identity, and test if they are actually available.
6590 *) Improve RPM specification file by forcing symbolic linking and making
6591 sure the installed documentation is also owned by root.root.
6592 [Damien Miller <djm@mindrot.org>]
6594 *) Give the OpenSSL applications more possibilities to make use of
6595 keys (public as well as private) handled by engines.
6598 *) Add OCSP code that comes from CertCo.
6601 *) Add VMS support for the Rijndael code.
6604 *) Added untested support for Nuron crypto accelerator.
6607 *) Add support for external cryptographic devices. This code was
6608 previously distributed separately as the "engine" branch.
6609 [Geoff Thorpe, Richard Levitte]
6611 *) Rework the filename-translation in the DSO code. It is now possible to
6612 have far greater control over how a "name" is turned into a filename
6613 depending on the operating environment and any oddities about the
6614 different shared library filenames on each system.
6617 *) Support threads on FreeBSD-elf in Configure.
6620 *) Fix for SHA1 assembly problem with MASM: it produces
6621 warnings about corrupt line number information when assembling
6622 with debugging information. This is caused by the overlapping
6624 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6627 NCONF_get_number() has no error checking at all. As a replacement,
6628 NCONF_get_number_e() is defined (_e for "error checking") and is
6629 promoted strongly. The old NCONF_get_number is kept around for
6630 binary backward compatibility.
6631 Make it possible for methods to load from something other than a BIO,
6632 by providing a function pointer that is given a name instead of a BIO.
6633 For example, this could be used to load configuration data from an
6637 *) Fix for non blocking accept BIOs. Added new I/O special reason
6638 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6639 with non blocking I/O was not possible because no retry code was
6640 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6644 *) Added the beginnings of Rijndael support.
6647 *) Fix for bug in DirectoryString mask setting. Add support for
6648 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6649 to allow certificate printing to more controllable, additional
6650 'certopt' option to 'x509' to allow new printing options to be
6654 *) Clean old EAY MD5 hack from e_os.h.
6657 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6659 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6660 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6661 [Joe Orton, Steve Henson]
6663 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6665 *) Fix additional bug revealed by the NISCC test suite:
6667 Stop bug triggering large recursion when presented with
6668 certain ASN.1 tags (CVE-2003-0851)
6671 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6673 *) Fix various bugs revealed by running the NISCC test suite:
6675 Stop out of bounds reads in the ASN1 code when presented with
6676 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6678 If verify callback ignores invalid public key errors don't try to check
6679 certificate signature with the NULL public key.
6683 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6684 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6688 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6689 extra data after the compression methods not only for TLS 1.0
6690 but also for SSL 3.0 (as required by the specification).
6691 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6693 *) Change X509_certificate_type() to mark the key as exported/exportable
6694 when it's 512 *bits* long, not 512 bytes.
6697 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6699 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6700 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6701 a protocol version number mismatch like a decryption error
6702 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6705 *) Turn on RSA blinding by default in the default implementation
6706 to avoid a timing attack. Applications that don't want it can call
6707 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6708 They would be ill-advised to do so in most cases.
6709 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6711 *) Change RSA blinding code so that it works when the PRNG is not
6712 seeded (in this case, the secret RSA exponent is abused as
6713 an unpredictable seed -- if it is not unpredictable, there
6714 is no point in blinding anyway). Make RSA blinding thread-safe
6715 by remembering the creator's thread ID in rsa->blinding and
6716 having all other threads use local one-time blinding factors
6717 (this requires more computation than sharing rsa->blinding, but
6718 avoids excessive locking; and if an RSA object is not shared
6719 between threads, blinding will still be very fast).
6722 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6724 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6725 via timing by performing a MAC computation even if incorrrect
6726 block cipher padding has been found. This is a countermeasure
6727 against active attacks where the attacker has to distinguish
6728 between bad padding and a MAC verification error. (CVE-2003-0078)
6730 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6731 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6732 Martin Vuagnoux (EPFL, Ilion)]
6734 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6736 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6737 memory from it's contents. This is done with a counter that will
6738 place alternating values in each byte. This can be used to solve
6739 two issues: 1) the removal of calls to memset() by highly optimizing
6740 compilers, and 2) cleansing with other values than 0, since those can
6741 be read through on certain media, for example a swap space on disk.
6744 *) Bugfix: client side session caching did not work with external caching,
6745 because the session->cipher setting was not restored when reloading
6746 from the external cache. This problem was masked, when
6747 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6748 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6751 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6752 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6753 [Zeev Lieber <zeev-l@yahoo.com>]
6755 *) Undo an undocumented change introduced in 0.9.6e which caused
6756 repeated calls to OpenSSL_add_all_ciphers() and
6757 OpenSSL_add_all_digests() to be ignored, even after calling
6761 *) Change the default configuration reader to deal with last line not
6762 being properly terminated.
6765 *) Change X509_NAME_cmp() so it applies the special rules on handling
6766 DN values that are of type PrintableString, as well as RDNs of type
6767 emailAddress where the value has the type ia5String.
6768 [stefank@valicert.com via Richard Levitte]
6770 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6771 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6772 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6773 the bitwise-OR of the two for use by the majority of applications
6774 wanting this behaviour, and update the docs. The documented
6775 behaviour and actual behaviour were inconsistent and had been
6776 changing anyway, so this is more a bug-fix than a behavioural
6778 [Geoff Thorpe, diagnosed by Nadav Har'El]
6780 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6781 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6784 *) Fix initialization code race conditions in
6785 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6786 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6787 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6788 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6789 ssl2_get_cipher_by_char(),
6790 ssl3_get_cipher_by_char().
6791 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6793 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6794 the cached sessions are flushed, as the remove_cb() might use ex_data
6795 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6796 (see [openssl.org #212]).
6797 [Geoff Thorpe, Lutz Jaenicke]
6799 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6800 length, instead of the encoding length to d2i_ASN1_OBJECT.
6803 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6805 *) [In 0.9.6g-engine release:]
6806 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6807 [Lynn Gazis <lgazis@rainbow.com>]
6809 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6811 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6812 and get fix the header length calculation.
6813 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6814 Alon Kantor <alonk@checkpoint.com> (and others),
6817 *) Use proper error handling instead of 'assertions' in buffer
6818 overflow checks added in 0.9.6e. This prevents DoS (the
6819 assertions could call abort()).
6820 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6822 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6824 *) Add various sanity checks to asn1_get_length() to reject
6825 the ASN1 length bytes if they exceed sizeof(long), will appear
6826 negative or the content length exceeds the length of the
6828 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6830 *) Fix cipher selection routines: ciphers without encryption had no flags
6831 for the cipher strength set and where therefore not handled correctly
6832 by the selection routines (PR #130).
6835 *) Fix EVP_dsa_sha macro.
6839 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6840 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6841 that was added in OpenSSL 0.9.6d.
6843 As the countermeasure turned out to be incompatible with some
6844 broken SSL implementations, the new option is part of SSL_OP_ALL.
6845 SSL_OP_ALL is usually employed when compatibility with weird SSL
6846 implementations is desired (e.g. '-bugs' option to 's_client' and
6847 's_server'), so the new option is automatically set in many
6851 *) Changes in security patch:
6853 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6854 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6855 Air Force Materiel Command, USAF, under agreement number
6858 *) Add various sanity checks to asn1_get_length() to reject
6859 the ASN1 length bytes if they exceed sizeof(long), will appear
6860 negative or the content length exceeds the length of the
6861 supplied buffer. (CVE-2002-0659)
6862 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6864 *) Assertions for various potential buffer overflows, not known to
6866 [Ben Laurie (CHATS)]
6868 *) Various temporary buffers to hold ASCII versions of integers were
6869 too small for 64 bit platforms. (CVE-2002-0655)
6870 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6872 *) Remote buffer overflow in SSL3 protocol - an attacker could
6873 supply an oversized session ID to a client. (CVE-2002-0656)
6874 [Ben Laurie (CHATS)]
6876 *) Remote buffer overflow in SSL2 protocol - an attacker could
6877 supply an oversized client master key. (CVE-2002-0656)
6878 [Ben Laurie (CHATS)]
6880 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6882 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6883 encoded as NULL) with id-dsa-with-sha1.
6884 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6886 *) Check various X509_...() return values in apps/req.c.
6887 [Nils Larsch <nla@trustcenter.de>]
6889 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6890 an end-of-file condition would erronously be flagged, when the CRLF
6891 was just at the end of a processed block. The bug was discovered when
6892 processing data through a buffering memory BIO handing the data to a
6893 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6894 <ptsekov@syntrex.com> and Nedelcho Stanev.
6897 *) Implement a countermeasure against a vulnerability recently found
6898 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6899 before application data chunks to avoid the use of known IVs
6900 with data potentially chosen by the attacker.
6903 *) Fix length checks in ssl3_get_client_hello().
6906 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6907 to prevent ssl3_read_internal() from incorrectly assuming that
6908 ssl3_read_bytes() found application data while handshake
6909 processing was enabled when in fact s->s3->in_read_app_data was
6910 merely automatically cleared during the initial handshake.
6911 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6913 *) Fix object definitions for Private and Enterprise: they were not
6914 recognized in their shortname (=lowercase) representation. Extend
6915 obj_dat.pl to issue an error when using undefined keywords instead
6916 of silently ignoring the problem (Svenning Sorensen
6917 <sss@sss.dnsalias.net>).
6920 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6921 generators, i.e. generators other than 2 and 5. (Previously, the
6922 code did not properly initialise the 'add' and 'rem' values to
6923 BN_generate_prime().)
6925 In the new general case, we do not insist that 'generator' is
6926 actually a primitive root: This requirement is rather pointless;
6927 a generator of the order-q subgroup is just as good, if not
6931 *) Map new X509 verification errors to alerts. Discovered and submitted by
6932 Tom Wu <tom@arcot.com>.
6935 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6936 returning non-zero before the data has been completely received
6937 when using non-blocking I/O.
6938 [Bodo Moeller; problem pointed out by John Hughes]
6940 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6941 [Ben Laurie, Lutz Jaenicke]
6943 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6944 Yoram Zahavi <YoramZ@gilian.com>).
6947 *) Add information about CygWin 1.3 and on, and preserve proper
6948 configuration for the versions before that.
6949 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6951 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6952 check whether we deal with a copy of a session and do not delete from
6953 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6954 <izhar@checkpoint.com>.
6957 *) Do not store session data into the internal session cache, if it
6958 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6959 flag is set). Proposed by Aslam <aslam@funk.com>.
6962 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6966 *) [In 0.9.6d-engine release:]
6967 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6968 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6970 *) Add the configuration target linux-s390x.
6971 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6973 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6974 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6975 variable as an indication that a ClientHello message has been
6976 received. As the flag value will be lost between multiple
6977 invocations of ssl3_accept when using non-blocking I/O, the
6978 function may not be aware that a handshake has actually taken
6979 place, thus preventing a new session from being added to the
6982 To avoid this problem, we now set s->new_session to 2 instead of
6983 using a local variable.
6984 [Lutz Jaenicke, Bodo Moeller]
6986 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6987 if the SSL_R_LENGTH_MISMATCH error is detected.
6988 [Geoff Thorpe, Bodo Moeller]
6990 *) New 'shared_ldflag' column in Configure platform table.
6993 *) Fix EVP_CIPHER_mode macro.
6994 ["Dan S. Camper" <dan@bti.net>]
6996 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6997 type, we must throw them away by setting rr->length to 0.
6998 [D P Chang <dpc@qualys.com>]
7000 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7002 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7003 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7004 worked incorrectly for those cases where range = 10..._2 and
7005 3*range is two bits longer than range.)
7008 *) Only add signing time to PKCS7 structures if it is not already
7012 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7013 OBJ_ld_ce should be OBJ_id_ce.
7014 Also some ip-pda OIDs in crypto/objects/objects.txt were
7015 incorrect (cf. RFC 3039).
7016 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7018 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7019 returns early because it has nothing to do.
7020 [Andy Schneider <andy.schneider@bjss.co.uk>]
7022 *) [In 0.9.6c-engine release:]
7023 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7024 [Andy Schneider <andy.schneider@bjss.co.uk>]
7026 *) [In 0.9.6c-engine release:]
7027 Add support for Cryptographic Appliance's keyserver technology.
7028 (Use engine 'keyclient')
7029 [Cryptographic Appliances and Geoff Thorpe]
7031 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7032 is called via tools/c89.sh because arguments have to be
7033 rearranged (all '-L' options must appear before the first object
7035 [Richard Shapiro <rshapiro@abinitio.com>]
7037 *) [In 0.9.6c-engine release:]
7038 Add support for Broadcom crypto accelerator cards, backported
7040 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7042 *) [In 0.9.6c-engine release:]
7043 Add support for SureWare crypto accelerator cards from
7044 Baltimore Technologies. (Use engine 'sureware')
7045 [Baltimore Technologies and Mark Cox]
7047 *) [In 0.9.6c-engine release:]
7048 Add support for crypto accelerator cards from Accelerated
7049 Encryption Processing, www.aep.ie. (Use engine 'aep')
7050 [AEP Inc. and Mark Cox]
7052 *) Add a configuration entry for gcc on UnixWare.
7053 [Gary Benson <gbenson@redhat.com>]
7055 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7056 messages are stored in a single piece (fixed-length part and
7057 variable-length part combined) and fix various bugs found on the way.
7060 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7061 instead. BIO_gethostbyname() does not know what timeouts are
7062 appropriate, so entries would stay in cache even when they have
7064 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7066 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7067 faced with a pathologically small ClientHello fragment that does
7068 not contain client_version: Instead of aborting with an error,
7069 simply choose the highest available protocol version (i.e.,
7070 TLS 1.0 unless it is disabled). In practice, ClientHello
7071 messages are never sent like this, but this change gives us
7072 strictly correct behaviour at least for TLS.
7075 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7076 never resets s->method to s->ctx->method when called from within
7077 one of the SSL handshake functions.
7078 [Bodo Moeller; problem pointed out by Niko Baric]
7080 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7081 (sent using the client's version number) if client_version is
7082 smaller than the protocol version in use. Also change
7083 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7084 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7085 the client will at least see that alert.
7088 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7092 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7093 client receives HelloRequest while in a handshake.
7094 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7096 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7097 should end in 'break', not 'goto end' which circumvents various
7098 cleanups done in state SSL_ST_OK. But session related stuff
7099 must be disabled for SSL_ST_OK in the case that we just sent a
7102 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7103 before just sending a HelloRequest.
7104 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7106 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7107 reveal whether illegal block cipher padding was found or a MAC
7108 verification error occured. (Neither SSLerr() codes nor alerts
7109 are directly visible to potential attackers, but the information
7110 may leak via logfiles.)
7112 Similar changes are not required for the SSL 2.0 implementation
7113 because the number of padding bytes is sent in clear for SSL 2.0,
7114 and the extra bytes are just ignored. However ssl/s2_pkt.c
7115 failed to verify that the purported number of padding bytes is in
7119 *) Add OpenUNIX-8 support including shared libraries
7120 (Boyd Lynn Gerber <gerberb@zenez.com>).
7123 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7124 'wristwatch attack' using huge encoding parameters (cf.
7125 James H. Manger's CRYPTO 2001 paper). Note that the
7126 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7127 encoding parameters and hence was not vulnerable.
7130 *) BN_sqr() bug fix.
7131 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7133 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7134 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7135 followed by modular reduction.
7136 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7138 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7139 equivalent based on BN_pseudo_rand() instead of BN_rand().
7142 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7143 This function was broken, as the check for a new client hello message
7144 to handle SGC did not allow these large messages.
7145 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7148 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7151 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7152 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7155 *) Rework the configuration and shared library support for Tru64 Unix.
7156 The configuration part makes use of modern compiler features and
7157 still retains old compiler behavior for those that run older versions
7158 of the OS. The shared library support part includes a variant that
7159 uses the RPATH feature, and is available through the special
7160 configuration target "alpha-cc-rpath", which will never be selected
7162 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7164 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7165 with the same message size as in ssl3_get_certificate_request().
7166 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7167 messages might inadvertently be reject as too long.
7168 [Petr Lampa <lampa@fee.vutbr.cz>]
7170 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7173 *) Modified SSL library such that the verify_callback that has been set
7174 specificly for an SSL object with SSL_set_verify() is actually being
7175 used. Before the change, a verify_callback set with this function was
7176 ignored and the verify_callback() set in the SSL_CTX at the time of
7177 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7178 to allow the necessary settings.
7181 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7182 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7183 done automatically (in contradiction to the requirements of the C
7184 standard). This made problems when used from OpenSSH.
7187 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7188 dh->length and always used
7190 BN_rand_range(priv_key, dh->p).
7192 BN_rand_range() is not necessary for Diffie-Hellman, and this
7193 specific range makes Diffie-Hellman unnecessarily inefficient if
7194 dh->length (recommended exponent length) is much smaller than the
7195 length of dh->p. We could use BN_rand_range() if the order of
7196 the subgroup was stored in the DH structure, but we only have
7201 BN_rand(priv_key, l, ...)
7203 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7209 RSA_eay_public_encrypt
7210 RSA_eay_private_decrypt
7211 RSA_eay_private_encrypt (signing)
7212 RSA_eay_public_decrypt (signature verification)
7214 (default implementations for RSA_public_encrypt,
7215 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7216 always reject numbers >= n.
7219 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7220 to synchronize access to 'locking_thread'. This is necessary on
7221 systems where access to 'locking_thread' (an 'unsigned long'
7222 variable) is not atomic.
7225 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7226 *before* setting the 'crypto_lock_rand' flag. The previous code had
7227 a race condition if 0 is a valid thread ID.
7228 [Travis Vitek <vitek@roguewave.com>]
7230 *) Add support for shared libraries under Irix.
7231 [Albert Chin-A-Young <china@thewrittenword.com>]
7233 *) Add configuration option to build on Linux on both big-endian and
7235 [Ralf Baechle <ralf@uni-koblenz.de>]
7237 *) Add the possibility to create shared libraries on HP-UX.
7240 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7242 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7243 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7244 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7245 PRNG state recovery was possible based on the output of
7246 one PRNG request appropriately sized to gain knowledge on
7247 'md' followed by enough consecutive 1-byte PRNG requests
7248 to traverse all of 'state'.
7250 1. When updating 'md_local' (the current thread's copy of 'md')
7251 during PRNG output generation, hash all of the previous
7252 'md_local' value, not just the half used for PRNG output.
7254 2. Make the number of bytes from 'state' included into the hash
7255 independent from the number of PRNG bytes requested.
7257 The first measure alone would be sufficient to avoid
7258 Markku-Juhani's attack. (Actually it had never occurred
7259 to me that the half of 'md_local' used for chaining was the
7260 half from which PRNG output bytes were taken -- I had always
7261 assumed that the secret half would be used.) The second
7262 measure makes sure that additional data from 'state' is never
7263 mixed into 'md_local' in small portions; this heuristically
7264 further strengthens the PRNG.
7267 *) Fix crypto/bn/asm/mips3.s.
7270 *) When only the key is given to "enc", the IV is undefined. Print out
7271 an error message in this case.
7274 *) Handle special case when X509_NAME is empty in X509 printing routines.
7277 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7278 positive and less than q.
7281 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7282 used: it isn't thread safe and the add_lock_callback should handle
7284 [Paul Rose <Paul.Rose@bridge.com>]
7286 *) Verify that incoming data obeys the block size in
7287 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7291 [Ulf Möller, Bodo Möller]
7293 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7294 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7295 when fixing the server behaviour for backwards-compatible 'client
7296 hello' messages. (Note that the attack is impractical against
7297 SSL 3.0 and TLS 1.0 anyway because length and version checking
7298 means that the probability of guessing a valid ciphertext is
7299 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7302 Before 0.9.5, the countermeasure (hide the error by generating a
7303 random 'decryption result') did not work properly because
7304 ERR_clear_error() was missing, meaning that SSL_get_error() would
7305 detect the supposedly ignored error.
7307 Both problems are now fixed.
7310 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7311 (previously it was 1024).
7314 *) Fix for compatibility mode trust settings: ignore trust settings
7315 unless some valid trust or reject settings are present.
7318 *) Fix for blowfish EVP: its a variable length cipher.
7321 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7322 parameters in DSA public key structures and return an error in the
7323 DSA routines if parameters are absent.
7326 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7327 in the current directory if neither $RANDFILE nor $HOME was set.
7328 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7329 caused some confusion to Windows users who haven't defined $HOME.
7330 Thus RAND_file_name() is changed again: e_os.h can define a
7331 DEFAULT_HOME, which will be used if $HOME is not set.
7332 For Windows, we use "C:"; on other platforms, we still require
7333 environment variables.
7335 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7336 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7337 having multiple threads call RAND_poll() concurrently.
7340 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7341 combination of a flag and a thread ID variable.
7342 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7343 flag), *other* threads can enter ssleay_add_bytes without obeying
7344 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7345 that they do not hold after the first thread unsets add_do_not_lock).
7348 *) Change bctest again: '-x' expressions are not available in all
7352 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7354 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7355 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7357 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7358 the default extension for executables, if any. Also, make the perl
7359 scripts that use symlink() to test if it really exists and use "cp"
7360 if it doesn't. All this made OpenSSL compilable and installable in
7364 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7365 If SEQUENCE is length is indefinite just set c->slen to the total
7366 amount of data available.
7367 [Steve Henson, reported by shige@FreeBSD.org]
7368 [This change does not apply to 0.9.7.]
7370 *) Change bctest to avoid here-documents inside command substitution
7371 (workaround for FreeBSD /bin/sh bug).
7372 For compatibility with Ultrix, avoid shell functions (introduced
7373 in the bctest version that searches along $PATH).
7376 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7377 with des_encrypt() defined on some operating systems, like Solaris
7381 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7382 On the Importance of Eliminating Errors in Cryptographic
7383 Computations, J. Cryptology 14 (2001) 2, 101-119,
7384 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7387 *) MIPS assembler BIGNUM division bug fix.
7390 *) Disabled incorrect Alpha assembler code.
7393 *) Fix PKCS#7 decode routines so they correctly update the length
7394 after reading an EOC for the EXPLICIT tag.
7396 [This change does not apply to 0.9.7.]
7398 *) Fix bug in PKCS#12 key generation routines. This was triggered
7399 if a 3DES key was generated with a 0 initial byte. Include
7400 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7401 (but broken) behaviour.
7404 *) Enhance bctest to search for a working bc along $PATH and print
7406 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7408 *) Fix memory leaks in err.c: free err_data string if necessary;
7409 don't write to the wrong index in ERR_set_error_data.
7412 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7416 *) Replace rdtsc with _emit statements for VC++ version 5.
7417 [Jeremy Cooper <jeremy@baymoo.org>]
7419 *) Make it possible to reuse SSLv2 sessions.
7422 *) In copy_email() check for >= 0 as a return value for
7423 X509_NAME_get_index_by_NID() since 0 is a valid index.
7424 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7426 *) Avoid coredump with unsupported or invalid public keys by checking if
7427 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7428 PKCS7_verify() fails with non detached data.
7431 *) Don't use getenv in library functions when run as setuid/setgid.
7432 New function OPENSSL_issetugid().
7435 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7436 due to incorrect handling of multi-threading:
7438 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7440 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7442 3. Count how many times MemCheck_off() has been called so that
7443 nested use can be treated correctly. This also avoids
7444 inband-signalling in the previous code (which relied on the
7445 assumption that thread ID 0 is impossible).
7448 *) Add "-rand" option also to s_client and s_server.
7451 *) Fix CPU detection on Irix 6.x.
7452 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7453 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7455 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7458 [This change does not apply to 0.9.7.]
7460 *) Use the cached encoding of an X509_NAME structure rather than
7461 copying it. This is apparently the reason for the libsafe "errors"
7462 but the code is actually correct.
7465 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7466 Bleichenbacher's DSA attack.
7467 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7468 to be set and top=0 forces the highest bit to be set; top=-1 is new
7469 and leaves the highest bit random.
7470 [Ulf Moeller, Bodo Moeller]
7472 *) In the NCONF_...-based implementations for CONF_... queries
7473 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7474 a temporary CONF structure with the data component set to NULL
7475 (which gives segmentation faults in lh_retrieve).
7476 Instead, use NULL for the CONF pointer in CONF_get_string and
7477 CONF_get_number (which may use environment variables) and directly
7478 return NULL from CONF_get_section.
7481 *) Fix potential buffer overrun for EBCDIC.
7484 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7485 keyUsage if basicConstraints absent for a CA.
7488 *) Make SMIME_write_PKCS7() write mail header values with a format that
7489 is more generally accepted (no spaces before the semicolon), since
7490 some programs can't parse those values properly otherwise. Also make
7491 sure BIO's that break lines after each write do not create invalid
7495 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7496 macros previously used would not encode an empty SEQUENCE OF
7497 and break the signature.
7499 [This change does not apply to 0.9.7.]
7501 *) Zero the premaster secret after deriving the master secret in
7505 *) Add some EVP_add_digest_alias registrations (as found in
7506 OpenSSL_add_all_digests()) to SSL_library_init()
7507 aka OpenSSL_add_ssl_algorithms(). This provides improved
7508 compatibility with peers using X.509 certificates
7509 with unconventional AlgorithmIdentifier OIDs.
7512 *) Fix for Irix with NO_ASM.
7513 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7515 *) ./config script fixes.
7516 [Ulf Moeller, Richard Levitte]
7518 *) Fix 'openssl passwd -1'.
7521 *) Change PKCS12_key_gen_asc() so it can cope with non null
7522 terminated strings whose length is passed in the passlen
7523 parameter, for example from PEM callbacks. This was done
7524 by adding an extra length parameter to asc2uni().
7525 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7527 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7528 call failed, free the DSA structure.
7531 *) Fix to uni2asc() to cope with zero length Unicode strings.
7532 These are present in some PKCS#12 files.
7535 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7536 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7537 when writing a 32767 byte record.
7538 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7540 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7541 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7543 (RSA objects have a reference count access to which is protected
7544 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7545 so they are meant to be shared between threads.)
7546 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7547 "Reddie, Steven" <Steven.Reddie@ca.com>]
7549 *) Fix a deadlock in CRYPTO_mem_leaks().
7552 *) Use better test patterns in bntest.
7555 *) rand_win.c fix for Borland C.
7558 *) BN_rshift bugfix for n == 0.
7561 *) Add a 'bctest' script that checks for some known 'bc' bugs
7562 so that 'make test' does not abort just because 'bc' is broken.
7565 *) Store verify_result within SSL_SESSION also for client side to
7566 avoid potential security hole. (Re-used sessions on the client side
7567 always resulted in verify_result==X509_V_OK, not using the original
7568 result of the server certificate verification.)
7571 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7572 SSL3_RT_APPLICATION_DATA, return 0.
7573 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7577 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7578 releases, have been re-implemented by renaming the previous
7579 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7580 and ssl3_read_internal, respectively, and adding 'peek' parameters
7581 to them. The new ssl[23]_{read,peek} functions are calls to
7582 ssl[23]_read_internal with the 'peek' flag set appropriately.
7583 A 'peek' parameter has also been added to ssl3_read_bytes, which
7584 does the actual work for ssl3_read_internal.
7587 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7588 the method-specific "init()" handler. Also clean up ex_data after
7589 calling the method-specific "finish()" handler. Previously, this was
7590 happening the other way round.
7593 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7594 The previous value, 12, was not always sufficient for BN_mod_exp().
7597 *) Make sure that shared libraries get the internal name engine with
7598 the full version number and not just 0. This should mark the
7599 shared libraries as not backward compatible. Of course, this should
7600 be changed again when we can guarantee backward binary compatibility.
7603 *) Fix typo in get_cert_by_subject() in by_dir.c
7604 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7606 *) Rework the system to generate shared libraries:
7608 - Make note of the expected extension for the shared libraries and
7609 if there is a need for symbolic links from for example libcrypto.so.0
7610 to libcrypto.so.0.9.7. There is extended info in Configure for
7613 - Make as few rebuilds of the shared libraries as possible.
7615 - Still avoid linking the OpenSSL programs with the shared libraries.
7617 - When installing, install the shared libraries separately from the
7621 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7623 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7624 and not in SSL_clear because the latter is also used by the
7625 accept/connect functions; previously, the settings made by
7626 SSL_set_read_ahead would be lost during the handshake.
7627 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7629 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7630 Previously, it would create entries for disableed algorithms no
7634 *) Added several new manual pages for SSL_* function.
7637 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7639 *) In ssl23_get_client_hello, generate an error message when faced
7640 with an initial SSL 3.0/TLS record that is too small to contain the
7641 first two bytes of the ClientHello message, i.e. client_version.
7642 (Note that this is a pathologic case that probably has never happened
7643 in real life.) The previous approach was to use the version number
7644 from the record header as a substitute; but our protocol choice
7645 should not depend on that one because it is not authenticated
7646 by the Finished messages.
7649 *) More robust randomness gathering functions for Windows.
7650 [Jeffrey Altman <jaltman@columbia.edu>]
7652 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7653 not set then we don't setup the error code for issuer check errors
7654 to avoid possibly overwriting other errors which the callback does
7655 handle. If an application does set the flag then we assume it knows
7656 what it is doing and can handle the new informational codes
7660 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7661 a general "ANY" type, as such it should be able to decode anything
7662 including tagged types. However it didn't check the class so it would
7663 wrongly interpret tagged types in the same way as their universal
7664 counterpart and unknown types were just rejected. Changed so that the
7665 tagged and unknown types are handled in the same way as a SEQUENCE:
7666 that is the encoding is stored intact. There is also a new type
7667 "V_ASN1_OTHER" which is used when the class is not universal, in this
7668 case we have no idea what the actual type is so we just lump them all
7672 *) On VMS, stdout may very well lead to a file that is written to
7673 in a record-oriented fashion. That means that every write() will
7674 write a separate record, which will be read separately by the
7675 programs trying to read from it. This can be very confusing.
7677 The solution is to put a BIO filter in the way that will buffer
7678 text until a linefeed is reached, and then write everything a
7679 line at a time, so every record written will be an actual line,
7680 not chunks of lines and not (usually doesn't happen, but I've
7681 seen it once) several lines in one record. BIO_f_linebuffer() is
7684 Currently, it's a VMS-only method, because that's where it has
7685 been tested well enough.
7688 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7689 it can return incorrect results.
7690 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7691 but it was in 0.9.6-beta[12].)
7694 *) Disable the check for content being present when verifying detached
7695 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7696 include zero length content when signing messages.
7699 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7700 BIO_ctrl (for BIO pairs).
7703 *) Add DSO method for VMS.
7706 *) Bug fix: Montgomery multiplication could produce results with the
7710 *) Add RPM specification openssl.spec and modify it to build three
7711 packages. The default package contains applications, application
7712 documentation and run-time libraries. The devel package contains
7713 include files, static libraries and function documentation. The
7714 doc package contains the contents of the doc directory. The original
7715 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7718 *) Add a large number of documentation files for many SSL routines.
7719 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7721 *) Add a configuration entry for Sony News 4.
7722 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7724 *) Don't set the two most significant bits to one when generating a
7725 random number < q in the DSA library.
7728 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7729 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7730 the underlying transport is blocking) if a handshake took place.
7731 (The default behaviour is needed by applications such as s_client
7732 and s_server that use select() to determine when to use SSL_read;
7733 but for applications that know in advance when to expect data, it
7734 just makes things more complicated.)
7737 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7741 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7742 work better on such systems.
7743 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7745 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7746 Update PKCS12_parse() so it copies the friendlyName and the
7747 keyid to the certificates aux info.
7750 *) Fix bug in PKCS7_verify() which caused an infinite loop
7751 if there was more than one signature.
7752 [Sven Uszpelkat <su@celocom.de>]
7754 *) Major change in util/mkdef.pl to include extra information
7755 about each symbol, as well as presenting variables as well
7756 as functions. This change means that there's n more need
7757 to rebuild the .num files when some algorithms are excluded.
7760 *) Allow the verify time to be set by an application,
7761 rather than always using the current time.
7764 *) Phase 2 verify code reorganisation. The certificate
7765 verify code now looks up an issuer certificate by a
7766 number of criteria: subject name, authority key id
7767 and key usage. It also verifies self signed certificates
7768 by the same criteria. The main comparison function is
7769 X509_check_issued() which performs these checks.
7771 Lot of changes were necessary in order to support this
7772 without completely rewriting the lookup code.
7774 Authority and subject key identifier are now cached.
7776 The LHASH 'certs' is X509_STORE has now been replaced
7777 by a STACK_OF(X509_OBJECT). This is mainly because an
7778 LHASH can't store or retrieve multiple objects with
7779 the same hash value.
7781 As a result various functions (which were all internal
7782 use only) have changed to handle the new X509_STORE
7783 structure. This will break anything that messed round
7784 with X509_STORE internally.
7786 The functions X509_STORE_add_cert() now checks for an
7787 exact match, rather than just subject name.
7789 The X509_STORE API doesn't directly support the retrieval
7790 of multiple certificates matching a given criteria, however
7791 this can be worked round by performing a lookup first
7792 (which will fill the cache with candidate certificates)
7793 and then examining the cache for matches. This is probably
7794 the best we can do without throwing out X509_LOOKUP
7795 entirely (maybe later...).
7797 The X509_VERIFY_CTX structure has been enhanced considerably.
7799 All certificate lookup operations now go via a get_issuer()
7800 callback. Although this currently uses an X509_STORE it
7801 can be replaced by custom lookups. This is a simple way
7802 to bypass the X509_STORE hackery necessary to make this
7803 work and makes it possible to use more efficient techniques
7804 in future. A very simple version which uses a simple
7805 STACK for its trusted certificate store is also provided
7806 using X509_STORE_CTX_trusted_stack().
7808 The verify_cb() and verify() callbacks now have equivalents
7809 in the X509_STORE_CTX structure.
7811 X509_STORE_CTX also has a 'flags' field which can be used
7812 to customise the verify behaviour.
7815 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7816 excludes S/MIME capabilities.
7819 *) When a certificate request is read in keep a copy of the
7820 original encoding of the signed data and use it when outputing
7821 again. Signatures then use the original encoding rather than
7822 a decoded, encoded version which may cause problems if the
7823 request is improperly encoded.
7826 *) For consistency with other BIO_puts implementations, call
7827 buffer_write(b, ...) directly in buffer_puts instead of calling
7830 In BIO_puts, increment b->num_write as in BIO_write.
7831 [Peter.Sylvester@EdelWeb.fr]
7833 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7834 BN_zero, we may not return a BIGNUM with an array consisting of
7838 *) Avoid calling abort() from within the library when problems are
7839 detected, except if preprocessor symbols have been defined
7840 (such as REF_CHECK, BN_DEBUG etc.).
7843 *) New openssl application 'rsautl'. This utility can be
7844 used for low level RSA operations. DER public key
7845 BIO/fp routines also added.
7848 *) New Configure entry and patches for compiling on QNX 4.
7849 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7851 *) A demo state-machine implementation was sponsored by
7852 Nuron (http://www.nuron.com/) and is now available in
7853 demos/state_machine.
7856 *) New options added to the 'dgst' utility for signature
7857 generation and verification.
7860 *) Unrecognized PKCS#7 content types are now handled via a
7861 catch all ASN1_TYPE structure. This allows unsupported
7862 types to be stored as a "blob" and an application can
7863 encode and decode it manually.
7866 *) Fix various signed/unsigned issues to make a_strex.c
7868 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7870 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7871 length if passed a buffer. ASN1_INTEGER_to_BN failed
7872 if passed a NULL BN and its argument was negative.
7873 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7875 *) Modification to PKCS#7 encoding routines to output definite
7876 length encoding. Since currently the whole structures are in
7877 memory there's not real point in using indefinite length
7878 constructed encoding. However if OpenSSL is compiled with
7879 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7882 *) Added BIO_vprintf() and BIO_vsnprintf().
7885 *) Added more prefixes to parse for in the the strings written
7886 through a logging bio, to cover all the levels that are available
7887 through syslog. The prefixes are now:
7889 PANIC, EMERG, EMR => LOG_EMERG
7890 ALERT, ALR => LOG_ALERT
7891 CRIT, CRI => LOG_CRIT
7892 ERROR, ERR => LOG_ERR
7893 WARNING, WARN, WAR => LOG_WARNING
7894 NOTICE, NOTE, NOT => LOG_NOTICE
7895 INFO, INF => LOG_INFO
7896 DEBUG, DBG => LOG_DEBUG
7898 and as before, if none of those prefixes are present at the
7899 beginning of the string, LOG_ERR is chosen.
7901 On Win32, the LOG_* levels are mapped according to this:
7903 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7904 LOG_WARNING => EVENTLOG_WARNING_TYPE
7905 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7909 *) Made it possible to reconfigure with just the configuration
7910 argument "reconf" or "reconfigure". The command line arguments
7911 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7912 and are retrieved from there when reconfiguring.
7916 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7918 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7921 *) The obj_dat.pl script was messing up the sorting of object
7922 names. The reason was that it compared the quoted version
7923 of strings as a result "OCSP" > "OCSP Signing" because
7924 " > SPACE. Changed script to store unquoted versions of
7925 names and add quotes on output. It was also omitting some
7926 names from the lookup table if they were given a default
7927 value (that is if SN is missing it is given the same
7928 value as LN and vice versa), these are now added on the
7929 grounds that if an object has a name we should be able to
7930 look it up. Finally added warning output when duplicate
7931 short or long names are found.
7934 *) Changes needed for Tandem NSK.
7935 [Scott Uroff <scott@xypro.com>]
7937 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7938 RSA_padding_check_SSLv23(), special padding was never detected
7939 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7940 version rollback attacks was not effective.
7942 In s23_clnt.c, don't use special rollback-attack detection padding
7943 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7944 client; similarly, in s23_srvr.c, don't do the rollback check if
7945 SSL 2.0 is the only protocol enabled in the server.
7948 *) Make it possible to get hexdumps of unprintable data with 'openssl
7949 asn1parse'. By implication, the functions ASN1_parse_dump() and
7950 BIO_dump_indent() are added.
7953 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7954 these print out strings and name structures based on various
7955 flags including RFC2253 support and proper handling of
7956 multibyte characters. Added options to the 'x509' utility
7957 to allow the various flags to be set.
7960 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7961 Also change the functions X509_cmp_current_time() and
7962 X509_gmtime_adj() work with an ASN1_TIME structure,
7963 this will enable certificates using GeneralizedTime in validity
7964 dates to be checked.
7967 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7968 negative public key encodings) on by default,
7969 NO_NEG_PUBKEY_BUG can be set to disable it.
7972 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7973 content octets. An i2c_ASN1_OBJECT is unnecessary because
7974 the encoding can be trivially obtained from the structure.
7977 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7978 not read locks (CRYPTO_r_[un]lock).
7981 *) A first attempt at creating official support for shared
7982 libraries through configuration. I've kept it so the
7983 default is static libraries only, and the OpenSSL programs
7984 are always statically linked for now, but there are
7985 preparations for dynamic linking in place.
7986 This has been tested on Linux and Tru64.
7989 *) Randomness polling function for Win9x, as described in:
7990 Peter Gutmann, Software Generation of Practically Strong
7994 *) Fix so PRNG is seeded in req if using an already existing
7998 *) New options to smime application. -inform and -outform
7999 allow alternative formats for the S/MIME message including
8000 PEM and DER. The -content option allows the content to be
8001 specified separately. This should allow things like Netscape
8002 form signing output easier to verify.
8005 *) Fix the ASN1 encoding of tags using the 'long form'.
8008 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8009 STRING types. These convert content octets to and from the
8010 underlying type. The actual tag and length octets are
8011 already assumed to have been read in and checked. These
8012 are needed because all other string types have virtually
8013 identical handling apart from the tag. By having versions
8014 of the ASN1 functions that just operate on content octets
8015 IMPLICIT tagging can be handled properly. It also allows
8016 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8017 and ASN1_INTEGER are identical apart from the tag.
8020 *) Change the handling of OID objects as follows:
8022 - New object identifiers are inserted in objects.txt, following
8023 the syntax given in objects.README.
8024 - objects.pl is used to process obj_mac.num and create a new
8026 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8029 This is currently kind of a hack, and the perl code in objects.pl
8030 isn't very elegant, but it works as I intended. The simplest way
8031 to check that it worked correctly is to look in obj_dat.h and
8032 check the array nid_objs and make sure the objects haven't moved
8033 around (this is important!). Additions are OK, as well as
8034 consistent name changes.
8037 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8040 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8041 The given file adds to whatever has already been seeded into the
8042 random pool through the RANDFILE configuration file option or
8043 environment variable, or the default random state file.
8046 *) mkstack.pl now sorts each macro group into lexical order.
8047 Previously the output order depended on the order the files
8048 appeared in the directory, resulting in needless rewriting
8052 *) Patches to make OpenSSL compile under Win32 again. Mostly
8053 work arounds for the VC++ problem that it treats func() as
8054 func(void). Also stripped out the parts of mkdef.pl that
8055 added extra typesafe functions: these no longer exist.
8058 *) Reorganisation of the stack code. The macros are now all
8059 collected in safestack.h . Each macro is defined in terms of
8060 a "stack macro" of the form SKM_<name>(type, a, b). The
8061 DEBUG_SAFESTACK is now handled in terms of function casts,
8062 this has the advantage of retaining type safety without the
8063 use of additional functions. If DEBUG_SAFESTACK is not defined
8064 then the non typesafe macros are used instead. Also modified the
8065 mkstack.pl script to handle the new form. Needs testing to see
8066 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8067 the default if no major problems. Similar behaviour for ASN1_SET_OF
8068 and PKCS12_STACK_OF.
8071 *) When some versions of IIS use the 'NET' form of private key the
8072 key derivation algorithm is different. Normally MD5(password) is
8073 used as a 128 bit RC4 key. In the modified case
8074 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8075 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8076 as the old Netscape_RSA functions except they have an additional
8077 'sgckey' parameter which uses the modified algorithm. Also added
8078 an -sgckey command line option to the rsa utility. Thanks to
8079 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8080 algorithm to openssl-dev.
8083 *) The evp_local.h macros were using 'c.##kname' which resulted in
8084 invalid expansion on some systems (SCO 5.0.5 for example).
8085 Corrected to 'c.kname'.
8086 [Phillip Porch <root@theporch.com>]
8088 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8089 a STACK of email addresses from a certificate or request, these look
8090 in the subject name and the subject alternative name extensions and
8091 omit any duplicate addresses.
8094 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8095 This makes DSA verification about 2 % faster.
8098 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8099 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8100 plus overhead for 1024 bit moduli).
8101 This makes exponentiations about 0.5 % faster for 1024 bit
8102 exponents (as measured by "openssl speed rsa2048").
8105 *) Rename memory handling macros to avoid conflicts with other
8107 Malloc => OPENSSL_malloc
8108 Malloc_locked => OPENSSL_malloc_locked
8109 Realloc => OPENSSL_realloc
8110 Free => OPENSSL_free
8113 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8114 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8117 *) CygWin32 support.
8118 [John Jarvie <jjarvie@newsguy.com>]
8120 *) The type-safe stack code has been rejigged. It is now only compiled
8121 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8122 by default all type-specific stack functions are "#define"d back to
8123 standard stack functions. This results in more streamlined output
8124 but retains the type-safety checking possibilities of the original
8128 *) The STACK code has been cleaned up, and certain type declarations
8129 that didn't make a lot of sense have been brought in line. This has
8130 also involved a cleanup of sorts in safestack.h to more correctly
8131 map type-safe stack functions onto their plain stack counterparts.
8132 This work has also resulted in a variety of "const"ifications of
8133 lots of the code, especially "_cmp" operations which should normally
8134 be prototyped with "const" parameters anyway.
8137 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8138 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8139 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8140 where all of 'md' is used each time the PRNG is used, but 'state'
8141 is used only indexed by a cyclic counter. As entropy may not be
8142 well distributed from the beginning, 'md' is important as a
8143 chaining variable. However, the output function chains only half
8144 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8145 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8146 in all of 'state' being rewritten, with the new values depending
8147 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8150 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8151 the handshake is continued after ssl_verify_cert_chain();
8152 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8153 can lead to 'unexplainable' connection aborts later.
8154 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8156 *) Major EVP API cipher revision.
8157 Add hooks for extra EVP features. This allows various cipher
8158 parameters to be set in the EVP interface. Support added for variable
8159 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8160 setting of RC2 and RC5 parameters.
8162 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8165 Remove lots of duplicated code from the EVP library. For example *every*
8166 cipher init() function handles the 'iv' in the same way according to the
8167 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8168 for CFB and OFB modes they zero ctx->num.
8170 New functionality allows removal of S/MIME code RC2 hack.
8172 Most of the routines have the same form and so can be declared in terms
8175 By shifting this to the top level EVP_CipherInit() it can be removed from
8176 all individual ciphers. If the cipher wants to handle IVs or keys
8177 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8180 Change lots of functions like EVP_EncryptUpdate() to now return a
8181 value: although software versions of the algorithms cannot fail
8182 any installed hardware versions can.
8185 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8186 this option is set, tolerate broken clients that send the negotiated
8187 protocol version number instead of the requested protocol version
8191 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8192 i.e. non-zero for export ciphersuites, zero otherwise.
8193 Previous versions had this flag inverted, inconsistent with
8194 rsa_tmp_cb (..._TMP_RSA_CB).
8195 [Bodo Moeller; problem reported by Amit Chopra]
8197 *) Add missing DSA library text string. Work around for some IIS
8198 key files with invalid SEQUENCE encoding.
8201 *) Add a document (doc/standards.txt) that list all kinds of standards
8202 and so on that are implemented in OpenSSL.
8205 *) Enhance c_rehash script. Old version would mishandle certificates
8206 with the same subject name hash and wouldn't handle CRLs at all.
8207 Added -fingerprint option to crl utility, to support new c_rehash
8211 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8214 *) Fix for SSL server purpose checking. Server checking was
8215 rejecting certificates which had extended key usage present
8216 but no ssl client purpose.
8217 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8219 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8220 is a little unclear about how a blank password is handled.
8221 Since the password in encoded as a BMPString with terminating
8222 double NULL a zero length password would end up as just the
8223 double NULL. However no password at all is different and is
8224 handled differently in the PKCS#12 key generation code. NS
8225 treats a blank password as zero length. MSIE treats it as no
8226 password on export: but it will try both on import. We now do
8227 the same: PKCS12_parse() tries zero length and no password if
8228 the password is set to "" or NULL (NULL is now a valid password:
8229 it wasn't before) as does the pkcs12 application.
8232 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8233 perror when PEM_read_bio_X509_REQ fails, the error message must
8234 be obtained from the error queue.
8237 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8238 it in ERR_remove_state if appropriate, and change ERR_get_state
8239 accordingly to avoid race conditions (this is necessary because
8240 thread_hash is no longer constant once set).
8243 *) Bugfix for linux-elf makefile.one.
8246 *) RSA_get_default_method() will now cause a default
8247 RSA_METHOD to be chosen if one doesn't exist already.
8248 Previously this was only set during a call to RSA_new()
8249 or RSA_new_method(NULL) meaning it was possible for
8250 RSA_get_default_method() to return NULL.
8253 *) Added native name translation to the existing DSO code
8254 that will convert (if the flag to do so is set) filenames
8255 that are sufficiently small and have no path information
8256 into a canonical native form. Eg. "blah" converted to
8257 "libblah.so" or "blah.dll" etc.
8260 *) New function ERR_error_string_n(e, buf, len) which is like
8261 ERR_error_string(e, buf), but writes at most 'len' bytes
8262 including the 0 terminator. For ERR_error_string_n, 'buf'
8264 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8266 *) CONF library reworked to become more general. A new CONF
8267 configuration file reader "class" is implemented as well as a
8268 new functions (NCONF_*, for "New CONF") to handle it. The now
8269 old CONF_* functions are still there, but are reimplemented to
8270 work in terms of the new functions. Also, a set of functions
8271 to handle the internal storage of the configuration data is
8272 provided to make it easier to write new configuration file
8273 reader "classes" (I can definitely see something reading a
8274 configuration file in XML format, for example), called _CONF_*,
8275 or "the configuration storage API"...
8277 The new configuration file reading functions are:
8279 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8280 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8282 NCONF_default, NCONF_WIN32
8284 NCONF_dump_fp, NCONF_dump_bio
8286 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8287 NCONF_new creates a new CONF object. This works in the same way
8288 as other interfaces in OpenSSL, like the BIO interface.
8289 NCONF_dump_* dump the internal storage of the configuration file,
8290 which is useful for debugging. All other functions take the same
8291 arguments as the old CONF_* functions wth the exception of the
8292 first that must be a `CONF *' instead of a `LHASH *'.
8294 To make it easer to use the new classes with the old CONF_* functions,
8295 the function CONF_set_default_method is provided.
8298 *) Add '-tls1' option to 'openssl ciphers', which was already
8299 mentioned in the documentation but had not been implemented.
8300 (This option is not yet really useful because even the additional
8301 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8304 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8305 OpenSSL-based applications) load shared libraries and bind to
8306 them in a portable way.
8307 [Geoff Thorpe, with contributions from Richard Levitte]
8309 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8311 *) Make sure _lrotl and _lrotr are only used with MSVC.
8313 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8314 (the default implementation of RAND_status).
8316 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8317 to '-clrext' (= clear extensions), as intended and documented.
8318 [Bodo Moeller; inconsistency pointed out by Michael Attili
8321 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8322 was larger than the MD block size.
8323 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8325 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8326 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8327 using the passed key: if the passed key was a private key the result
8328 of X509_print(), for example, would be to print out all the private key
8332 *) des_quad_cksum() byte order bug fix.
8333 [Ulf Möller, using the problem description in krb4-0.9.7, where
8334 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8336 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8338 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8340 *) For easily testing in shell scripts whether some command
8341 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8342 returns with exit code 0 iff no command of the given name is available.
8343 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8344 the output goes to stdout and nothing is printed to stderr.
8345 Additional arguments are always ignored.
8347 Since for each cipher there is a command of the same name,
8348 the 'no-cipher' compilation switches can be tested this way.
8350 ('openssl no-XXX' is not able to detect pseudo-commands such
8351 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8354 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8357 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8358 is set; it will be thrown away anyway because each handshake creates
8360 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8361 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8362 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8363 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8366 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8367 'Q' and 'R' lose their special meanings (quit/renegotiate).
8368 This is part of what -quiet does; unlike -quiet, -ign_eof
8369 does not suppress any output.
8372 *) Add compatibility options to the purpose and trust code. The
8373 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8374 accepts a certificate or CA, this was the previous behaviour,
8375 with all the associated security issues.
8377 X509_TRUST_COMPAT is the old trust behaviour: only and
8378 automatically trust self signed roots in certificate store. A
8379 new trust setting X509_TRUST_DEFAULT is used to specify that
8380 a purpose has no associated trust setting and it should instead
8381 use the value in the default purpose.
8384 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8385 and fix a memory leak.
8388 *) In util/mkerr.pl (which implements 'make errors'), preserve
8389 reason strings from the previous version of the .c file, as
8390 the default to have only downcase letters (and digits) in
8391 automatically generated reasons codes is not always appropriate.
8394 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8395 using strerror. Previously, ERR_reason_error_string() returned
8396 library names as reason strings for SYSerr; but SYSerr is a special
8397 case where small numbers are errno values, not library numbers.
8400 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8401 converts DSA parameters into DH parameters. (When creating parameters,
8402 DSA_generate_parameters is used.)
8405 *) Include 'length' (recommended exponent length) in C code generated
8406 by 'openssl dhparam -C'.
8409 *) The second argument to set_label in perlasm was already being used
8410 so couldn't be used as a "file scope" flag. Moved to third argument
8414 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8415 instead of RAND_bytes for encryption IVs and salts.
8418 *) Include RAND_status() into RAND_METHOD instead of implementing
8419 it only for md_rand.c Otherwise replacing the PRNG by calling
8420 RAND_set_rand_method would be impossible.
8423 *) Don't let DSA_generate_key() enter an infinite loop if the random
8424 number generation fails.
8427 *) New 'rand' application for creating pseudo-random output.
8430 *) Added configuration support for Linux/IA64
8431 [Rolf Haberrecker <rolf@suse.de>]
8433 *) Assembler module support for Mingw32.
8436 *) Shared library support for HPUX (in shlib/).
8437 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8439 *) Shared library support for Solaris gcc.
8440 [Lutz Behnke <behnke@trustcenter.de>]
8442 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8444 *) PKCS7_encrypt() was adding text MIME headers twice because they
8445 were added manually and by SMIME_crlf_copy().
8448 *) In bntest.c don't call BN_rand with zero bits argument.
8449 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8451 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8452 case was implemented. This caused BN_div_recp() to fail occasionally.
8455 *) Add an optional second argument to the set_label() in the perl
8456 assembly language builder. If this argument exists and is set
8457 to 1 it signals that the assembler should use a symbol whose
8458 scope is the entire file, not just the current function. This
8459 is needed with MASM which uses the format label:: for this scope.
8460 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8462 *) Change the ASN1 types so they are typedefs by default. Before
8463 almost all types were #define'd to ASN1_STRING which was causing
8464 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8468 *) Change names of new functions to the new get1/get0 naming
8469 convention: After 'get1', the caller owns a reference count
8470 and has to call ..._free; 'get0' returns a pointer to some
8471 data structure without incrementing reference counters.
8472 (Some of the existing 'get' functions increment a reference
8473 counter, some don't.)
8474 Similarly, 'set1' and 'add1' functions increase reference
8475 counters or duplicate objects.
8478 *) Allow for the possibility of temp RSA key generation failure:
8479 the code used to assume it always worked and crashed on failure.
8482 *) Fix potential buffer overrun problem in BIO_printf().
8483 [Ulf Möller, using public domain code by Patrick Powell; problem
8484 pointed out by David Sacerdote <das33@cornell.edu>]
8486 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8487 RAND_egd() and RAND_status(). In the command line application,
8488 the EGD socket can be specified like a seed file using RANDFILE
8492 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8493 Some CAs (e.g. Verisign) distribute certificates in this form.
8496 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8497 list to exclude them. This means that no special compilation option
8498 is needed to use anonymous DH: it just needs to be included in the
8502 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8503 EVP_MD_type. The old functionality is available in a new macro called
8504 EVP_MD_md(). Change code that uses it and update docs.
8507 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8508 where the 'void *' argument is replaced by a function pointer argument.
8509 Previously 'void *' was abused to point to functions, which works on
8510 many platforms, but is not correct. As these functions are usually
8511 called by macros defined in OpenSSL header files, most source code
8512 should work without changes.
8515 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8516 sections with information on -D... compiler switches used for
8517 compiling the library so that applications can see them. To enable
8518 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8519 must be defined. E.g.,
8520 #define OPENSSL_ALGORITHM_DEFINES
8521 #include <openssl/opensslconf.h>
8522 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8523 [Richard Levitte, Ulf and Bodo Möller]
8525 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8529 *) Change the 'other' type in certificate aux info to a STACK_OF
8530 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8531 the required ASN1 format: arbitrary types determined by an OID.
8534 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8535 argument to 'req'. This is not because the function is newer or
8536 better than others it just uses the work 'NEW' in the certificate
8537 request header lines. Some software needs this.
8540 *) Reorganise password command line arguments: now passwords can be
8541 obtained from various sources. Delete the PEM_cb function and make
8542 it the default behaviour: i.e. if the callback is NULL and the
8543 usrdata argument is not NULL interpret it as a null terminated pass
8544 phrase. If usrdata and the callback are NULL then the pass phrase
8545 is prompted for as usual.
8548 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8549 the support is automatically enabled. The resulting binaries will
8550 autodetect the card and use it if present.
8551 [Ben Laurie and Compaq Inc.]
8553 *) Work around for Netscape hang bug. This sends certificate request
8554 and server done in one record. Since this is perfectly legal in the
8555 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8556 the bugs/SSLv3 entry for more info.
8559 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8562 *) Add -rand argument to smime and pkcs12 applications and read/write
8566 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8569 *) Add command line password options to the remaining applications.
8572 *) Bug fix for BN_div_recp() for numerators with an even number of
8576 *) More tests in bntest.c, and changed test_bn output.
8579 *) ./config recognizes MacOS X now.
8582 *) Bug fix for BN_div() when the first words of num and divsor are
8583 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8586 *) Add support for various broken PKCS#8 formats, and command line
8587 options to produce them.
8590 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8591 get temporary BIGNUMs from a BN_CTX.
8594 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8598 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8599 include a #define from the old name to the new. The original intent
8600 was that statically linked binaries could for example just call
8601 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8602 link with digests. This never worked because SSLeay_add_all_digests()
8603 and SSLeay_add_all_ciphers() were in the same source file so calling
8604 one would link with the other. They are now in separate source files.
8607 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8610 *) Use a less unusual form of the Miller-Rabin primality test (it used
8611 a binary algorithm for exponentiation integrated into the Miller-Rabin
8612 loop, our standard modexp algorithms are faster).
8615 *) Support for the EBCDIC character set completed.
8616 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8618 *) Source code cleanups: use const where appropriate, eliminate casts,
8619 use void * instead of char * in lhash.
8622 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8623 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8624 this the server could overwrite ephemeral keys that the client
8628 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8629 using 50 iterations of the Rabin-Miller test.
8631 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8632 iterations of the Rabin-Miller test as required by the appendix
8633 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8634 As BN_is_prime_fasttest includes trial division, DSA parameter
8635 generation becomes much faster.
8637 This implies a change for the callback functions in DSA_is_prime
8638 and DSA_generate_parameters: The callback function is called once
8639 for each positive witness in the Rabin-Miller test, not just
8640 occasionally in the inner loop; and the parameters to the
8641 callback function now provide an iteration count for the outer
8642 loop rather than for the current invocation of the inner loop.
8643 DSA_generate_parameters additionally can call the callback
8644 function with an 'iteration count' of -1, meaning that a
8645 candidate has passed the trial division test (when q is generated
8646 from an application-provided seed, trial division is skipped).
8649 *) New function BN_is_prime_fasttest that optionally does trial
8650 division before starting the Rabin-Miller test and has
8651 an additional BN_CTX * argument (whereas BN_is_prime always
8652 has to allocate at least one BN_CTX).
8653 'callback(1, -1, cb_arg)' is called when a number has passed the
8654 trial division stage.
8657 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8661 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8664 *) New function BN_pseudo_rand().
8667 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8668 bignum version of BN_from_montgomery() with the working code from
8669 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8673 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8674 made it impossible to use the same SSL_SESSION data structure in
8675 SSL2 clients in multiple threads.
8678 *) The return value of RAND_load_file() no longer counts bytes obtained
8679 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8680 to seed the PRNG (previously an explicit byte count was required).
8681 [Ulf Möller, Bodo Möller]
8683 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8684 used (char *) instead of (void *) and had casts all over the place.
8687 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8690 *) Retain source code compatibility for BN_prime_checks macro:
8691 BN_is_prime(..., BN_prime_checks, ...) now uses
8692 BN_prime_checks_for_size to determine the appropriate number of
8693 Rabin-Miller iterations.
8696 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8697 DH_CHECK_P_NOT_SAFE_PRIME.
8698 (Check if this is true? OpenPGP calls them "strong".)
8701 *) Merge the functionality of "dh" and "gendh" programs into a new program
8702 "dhparam". The old programs are retained for now but will handle DH keys
8703 (instead of parameters) in future.
8706 *) Make the ciphers, s_server and s_client programs check the return values
8707 when a new cipher list is set.
8710 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8711 ciphers. Before when the 56bit ciphers were enabled the sorting was
8714 The syntax for the cipher sorting has been extended to support sorting by
8715 cipher-strength (using the strength_bits hard coded in the tables).
8716 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8718 Fix a bug in the cipher-command parser: when supplying a cipher command
8719 string with an "undefined" symbol (neither command nor alphanumeric
8720 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8721 an error is flagged.
8723 Due to the strength-sorting extension, the code of the
8724 ssl_create_cipher_list() function was completely rearranged. I hope that
8725 the readability was also increased :-)
8726 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8728 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8729 for the first serial number and places 2 in the serial number file. This
8730 avoids problems when the root CA is created with serial number zero and
8731 the first user certificate has the same issuer name and serial number
8735 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8736 the new code. Add documentation for this stuff.
8739 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8740 X509_*() to X509at_*() on the grounds that they don't handle X509
8741 structures and behave in an analogous way to the X509v3 functions:
8742 they shouldn't be called directly but wrapper functions should be used
8745 So we also now have some wrapper functions that call the X509at functions
8746 when passed certificate requests. (TO DO: similar things can be done with
8747 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8748 things. Some of these need some d2i or i2d and print functionality
8749 because they handle more complex structures.)
8752 *) Add missing #ifndefs that caused missing symbols when building libssl
8753 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8754 NO_RSA in ssl/s2*.c.
8755 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8757 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8758 has a return value which indicates the quality of the random data
8759 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8760 error queue. New function RAND_pseudo_bytes() generates output that is
8761 guaranteed to be unique but not unpredictable. RAND_add is like
8762 RAND_seed, but takes an extra argument for an entropy estimate
8763 (RAND_seed always assumes full entropy).
8766 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8767 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8768 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8769 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8770 false-positive rate of at most 2^-80 for random input.
8773 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8776 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8777 in the 0.9.5 release), this returns the chain
8778 from an X509_CTX structure with a dup of the stack and all
8779 the X509 reference counts upped: so the stack will exist
8780 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8783 Also make SSL_SESSION_print() print out the verify return
8787 *) Add manpage for the pkcs12 command. Also change the default
8788 behaviour so MAC iteration counts are used unless the new
8789 -nomaciter option is used. This improves file security and
8790 only older versions of MSIE (4.0 for example) need it.
8793 *) Honor the no-xxx Configure options when creating .DEF files.
8796 *) Add PKCS#10 attributes to field table: challengePassword,
8797 unstructuredName and unstructuredAddress. These are taken from
8798 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8799 international characters are used.
8801 More changes to X509_ATTRIBUTE code: allow the setting of types
8802 based on strings. Remove the 'loc' parameter when adding
8803 attributes because these will be a SET OF encoding which is sorted
8807 *) Initial changes to the 'req' utility to allow request generation
8808 automation. This will allow an application to just generate a template
8809 file containing all the field values and have req construct the
8812 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8813 used all over the place including certificate requests and PKCS#7
8814 structures. They are currently handled manually where necessary with
8815 some primitive wrappers for PKCS#7. The new functions behave in a
8816 manner analogous to the X509 extension functions: they allow
8817 attributes to be looked up by NID and added.
8819 Later something similar to the X509V3 code would be desirable to
8820 automatically handle the encoding, decoding and printing of the
8821 more complex types. The string types like challengePassword can
8822 be handled by the string table functions.
8824 Also modified the multi byte string table handling. Now there is
8825 a 'global mask' which masks out certain types. The table itself
8826 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8827 is useful when for example there is only one permissible type
8828 (as in countryName) and using the mask might result in no valid
8832 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8833 SSL_get_peer_finished to allow applications to obtain the latest
8834 Finished messages sent to the peer or expected from the peer,
8835 respectively. (SSL_get_peer_finished is usually the Finished message
8836 actually received from the peer, otherwise the protocol will be aborted.)
8838 As the Finished message are message digests of the complete handshake
8839 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8840 be used for external authentication procedures when the authentication
8841 provided by SSL/TLS is not desired or is not enough.
8844 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8845 the host supports BWX extension and if Compaq C is present on the
8846 $PATH. Just exploiting of the BWX extension results in 20-30%
8847 performance kick for some algorithms, e.g. DES and RC4 to mention
8848 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8852 *) Add support for MS "fast SGC". This is arguably a violation of the
8853 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8854 weak crypto and after checking the certificate is SGC a second one
8855 with strong crypto. MS SGC stops the first handshake after receiving
8856 the server certificate message and sends a second client hello. Since
8857 a server will typically do all the time consuming operations before
8858 expecting any further messages from the client (server key exchange
8859 is the most expensive) there is little difference between the two.
8861 To get OpenSSL to support MS SGC we have to permit a second client
8862 hello message after we have sent server done. In addition we have to
8863 reset the MAC if we do get this second client hello.
8866 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8867 if a DER encoded private key is RSA or DSA traditional format. Changed
8868 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8869 format DER encoded private key. Newer code should use PKCS#8 format which
8870 has the key type encoded in the ASN1 structure. Added DER private key
8871 support to pkcs8 application.
8874 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8875 ciphersuites has been selected (as required by the SSL 3/TLS 1
8876 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8877 is set, we interpret this as a request to violate the specification
8878 (the worst that can happen is a handshake failure, and 'correct'
8879 behaviour would result in a handshake failure anyway).
8882 *) In SSL_CTX_add_session, take into account that there might be multiple
8883 SSL_SESSION structures with the same session ID (e.g. when two threads
8884 concurrently obtain them from an external cache).
8885 The internal cache can handle only one SSL_SESSION with a given ID,
8886 so if there's a conflict, we now throw out the old one to achieve
8890 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8891 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8892 some routines that use cipher OIDs: some ciphers do not have OIDs
8893 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8897 *) Simplify the trust setting structure and code. Now we just have
8898 two sequences of OIDs for trusted and rejected settings. These will
8899 typically have values the same as the extended key usage extension
8900 and any application specific purposes.
8902 The trust checking code now has a default behaviour: it will just
8903 check for an object with the same NID as the passed id. Functions can
8904 be provided to override either the default behaviour or the behaviour
8905 for a given id. SSL client, server and email already have functions
8906 in place for compatibility: they check the NID and also return "trusted"
8907 if the certificate is self signed.
8910 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8911 traditional format into an EVP_PKEY structure.
8914 *) Add a password callback function PEM_cb() which either prompts for
8915 a password if usr_data is NULL or otherwise assumes it is a null
8916 terminated password. Allow passwords to be passed on command line
8917 environment or config files in a few more utilities.
8920 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8921 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8922 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8923 Update documentation.
8926 *) Support for ASN1 "NULL" type. This could be handled before by using
8927 ASN1_TYPE but there wasn't any function that would try to read a NULL
8928 and produce an error if it couldn't. For compatibility we also have
8929 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8930 don't allocate anything because they don't need to.
8933 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8935 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8937 *) Rebuild of the memory allocation routines used by OpenSSL code and
8938 possibly others as well. The purpose is to make an interface that
8939 provide hooks so anyone can build a separate set of allocation and
8940 deallocation routines to be used by OpenSSL, for example memory
8941 pool implementations, or something else, which was previously hard
8942 since Malloc(), Realloc() and Free() were defined as macros having
8943 the values malloc, realloc and free, respectively (except for Win32
8944 compilations). The same is provided for memory debugging code.
8945 OpenSSL already comes with functionality to find memory leaks, but
8946 this gives people a chance to debug other memory problems.
8948 With these changes, a new set of functions and macros have appeared:
8950 CRYPTO_set_mem_debug_functions() [F]
8951 CRYPTO_get_mem_debug_functions() [F]
8952 CRYPTO_dbg_set_options() [F]
8953 CRYPTO_dbg_get_options() [F]
8954 CRYPTO_malloc_debug_init() [M]
8956 The memory debug functions are NULL by default, unless the library
8957 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8958 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8959 gives the standard debugging functions that come with OpenSSL) or
8960 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8961 provided by the library user) must be used. When the standard
8962 debugging functions are used, CRYPTO_dbg_set_options can be used to
8963 request additional information:
8964 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8965 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8967 Also, things like CRYPTO_set_mem_functions will always give the
8968 expected result (the new set of functions is used for allocation
8969 and deallocation) at all times, regardless of platform and compiler
8972 To finish it up, some functions that were never use in any other
8973 way than through macros have a new API and new semantic:
8976 CRYPTO_dbg_realloc()
8979 All macros of value have retained their old syntax.
8980 [Richard Levitte and Bodo Moeller]
8982 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8983 ordering of SMIMECapabilities wasn't in "strength order" and there
8984 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8988 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8989 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8990 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8992 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8993 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8994 functionality to handle multipart/signed properly) and a utility
8995 called 'smime' to call all this stuff. This is based on code I
8996 originally wrote for Celo who have kindly allowed it to be
8997 included in OpenSSL.
9000 *) Add variants des_set_key_checked and des_set_key_unchecked of
9001 des_set_key (aka des_key_sched). Global variable des_check_key
9002 decides which of these is called by des_set_key; this way
9003 des_check_key behaves as it always did, but applications and
9004 the library itself, which was buggy for des_check_key == 1,
9005 have a cleaner way to pick the version they need.
9008 *) New function PKCS12_newpass() which changes the password of a
9012 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9013 dynamic mix. In both cases the ids can be used as an index into the
9014 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9015 functions so they accept a list of the field values and the
9016 application doesn't need to directly manipulate the X509_TRUST
9020 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9024 *) Modify the way the V3 extension code looks up extensions. This now
9025 works in a similar way to the object code: we have some "standard"
9026 extensions in a static table which is searched with OBJ_bsearch()
9027 and the application can add dynamic ones if needed. The file
9028 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9029 updated whenever a new extension is added to the core code and kept
9030 in ext_nid order. There is a simple program 'tabtest.c' which checks
9031 this. New extensions are not added too often so this file can readily
9032 be maintained manually.
9034 There are two big advantages in doing things this way. The extensions
9035 can be looked up immediately and no longer need to be "added" using
9036 X509V3_add_standard_extensions(): this function now does nothing.
9037 [Side note: I get *lots* of email saying the extension code doesn't
9038 work because people forget to call this function]
9039 Also no dynamic allocation is done unless new extensions are added:
9040 so if we don't add custom extensions there is no need to call
9041 X509V3_EXT_cleanup().
9044 *) Modify enc utility's salting as follows: make salting the default. Add a
9045 magic header, so unsalted files fail gracefully instead of just decrypting
9046 to garbage. This is because not salting is a big security hole, so people
9047 should be discouraged from doing it.
9050 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9051 digest to be passed on the command line but it only used this
9052 parameter when signing a certificate. Modified so all relevant
9053 operations are affected by the digest parameter including the
9054 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9055 DSA key was used because it didn't fix the digest.
9058 *) Initial certificate chain verify code. Currently tests the untrusted
9059 certificates for consistency with the verify purpose (which is set
9060 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9062 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9063 this is because it will reject chains with invalid extensions whereas
9064 every previous version of OpenSSL and SSLeay made no checks at all.
9066 Trust code: checks the root CA for the relevant trust settings. Trust
9067 settings have an initial value consistent with the verify purpose: e.g.
9068 if the verify purpose is for SSL client use it expects the CA to be
9069 trusted for SSL client use. However the default value can be changed to
9070 permit custom trust settings: one example of this would be to only trust
9071 certificates from a specific "secure" set of CAs.
9073 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9074 which should be used for version portability: especially since the
9075 verify structure is likely to change more often now.
9077 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9078 to set them. If not set then assume SSL clients will verify SSL servers
9081 Two new options to the verify program: -untrusted allows a set of
9082 untrusted certificates to be passed in and -purpose which sets the
9083 intended purpose of the certificate. If a purpose is set then the
9084 new chain verify code is used to check extension consistency.
9087 *) Support for the authority information access extension.
9090 *) Modify RSA and DSA PEM read routines to transparently handle
9091 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9092 public keys in a format compatible with certificate
9093 SubjectPublicKeyInfo structures. Unfortunately there were already
9094 functions called *_PublicKey_* which used various odd formats so
9095 these are retained for compatibility: however the DSA variants were
9096 never in a public release so they have been deleted. Changed dsa/rsa
9097 utilities to handle the new format: note no releases ever handled public
9098 keys so we should be OK.
9100 The primary motivation for this change is to avoid the same fiasco
9101 that dogs private keys: there are several incompatible private key
9102 formats some of which are standard and some OpenSSL specific and
9103 require various evil hacks to allow partial transparent handling and
9104 even then it doesn't work with DER formats. Given the option anything
9105 other than PKCS#8 should be dumped: but the other formats have to
9106 stay in the name of compatibility.
9108 With public keys and the benefit of hindsight one standard format
9109 is used which works with EVP_PKEY, RSA or DSA structures: though
9110 it clearly returns an error if you try to read the wrong kind of key.
9112 Added a -pubkey option to the 'x509' utility to output the public key.
9113 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9114 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9115 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9116 that do the same as the EVP_PKEY_assign_*() except they up the
9117 reference count of the added key (they don't "swallow" the
9121 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9122 CRLs would fail if the file contained no certificates or no CRLs:
9123 added a new function to read in both types and return the number
9124 read: this means that if none are read it will be an error. The
9125 DER versions of the certificate and CRL reader would always fail
9126 because it isn't possible to mix certificates and CRLs in DER format
9127 without choking one or the other routine. Changed this to just read
9128 a certificate: this is the best we can do. Also modified the code
9129 in apps/verify.c to take notice of return codes: it was previously
9130 attempting to read in certificates from NULL pointers and ignoring
9131 any errors: this is one reason why the cert and CRL reader seemed
9132 to work. It doesn't check return codes from the default certificate
9133 routines: these may well fail if the certificates aren't installed.
9136 *) Code to support otherName option in GeneralName.
9139 *) First update to verify code. Change the verify utility
9140 so it warns if it is passed a self signed certificate:
9141 for consistency with the normal behaviour. X509_verify
9142 has been modified to it will now verify a self signed
9143 certificate if *exactly* the same certificate appears
9144 in the store: it was previously impossible to trust a
9145 single self signed certificate. This means that:
9146 openssl verify ss.pem
9147 now gives a warning about a self signed certificate but
9148 openssl verify -CAfile ss.pem ss.pem
9152 *) For servers, store verify_result in SSL_SESSION data structure
9153 (and add it to external session representation).
9154 This is needed when client certificate verifications fails,
9155 but an application-provided verification callback (set by
9156 SSL_CTX_set_cert_verify_callback) allows accepting the session
9157 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9158 but returns 1): When the session is reused, we have to set
9159 ssl->verify_result to the appropriate error code to avoid
9161 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9163 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9164 case in PKCS7_dataInit() where the signed PKCS7 structure
9165 didn't contain any existing data because it was being created.
9166 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9168 *) Add a salt to the key derivation routines in enc.c. This
9169 forms the first 8 bytes of the encrypted file. Also add a
9170 -S option to allow a salt to be input on the command line.
9173 *) New function X509_cmp(). Oddly enough there wasn't a function
9174 to compare two certificates. We do this by working out the SHA1
9175 hash and comparing that. X509_cmp() will be needed by the trust
9179 *) SSL_get1_session() is like SSL_get_session(), but increments
9180 the reference count in the SSL_SESSION returned.
9181 [Geoff Thorpe <geoff@eu.c2.net>]
9183 *) Fix for 'req': it was adding a null to request attributes.
9184 Also change the X509_LOOKUP and X509_INFO code to handle
9185 certificate auxiliary information.
9188 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9192 *) Add the possibility to add extra information to the memory leak
9193 detecting output, to form tracebacks, showing from where each
9194 allocation was originated: CRYPTO_push_info("constant string") adds
9195 the string plus current file name and line number to a per-thread
9196 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9197 is like calling CYRPTO_pop_info() until the stack is empty.
9198 Also updated memory leak detection code to be multi-thread-safe.
9201 *) Add options -text and -noout to pkcs7 utility and delete the
9202 encryption options which never did anything. Update docs.
9205 *) Add options to some of the utilities to allow the pass phrase
9206 to be included on either the command line (not recommended on
9207 OSes like Unix) or read from the environment. Update the
9208 manpages and fix a few bugs.
9211 *) Add a few manpages for some of the openssl commands.
9214 *) Fix the -revoke option in ca. It was freeing up memory twice,
9215 leaking and not finding already revoked certificates.
9218 *) Extensive changes to support certificate auxiliary information.
9219 This involves the use of X509_CERT_AUX structure and X509_AUX
9220 functions. An X509_AUX function such as PEM_read_X509_AUX()
9221 can still read in a certificate file in the usual way but it
9222 will also read in any additional "auxiliary information". By
9223 doing things this way a fair degree of compatibility can be
9224 retained: existing certificates can have this information added
9225 using the new 'x509' options.
9227 Current auxiliary information includes an "alias" and some trust
9228 settings. The trust settings will ultimately be used in enhanced
9229 certificate chain verification routines: currently a certificate
9230 can only be trusted if it is self signed and then it is trusted
9234 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9235 The problem was that one of the replacement routines had not been working
9236 since SSLeay releases. For now the offending routine has been replaced
9237 with non-optimised assembler. Even so, this now gives around 95%
9238 performance improvement for 1024 bit RSA signs.
9241 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9242 handling. Most clients have the effective key size in bits equal to
9243 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9244 A few however don't do this and instead use the size of the decrypted key
9245 to determine the RC2 key length and the AlgorithmIdentifier to determine
9246 the effective key length. In this case the effective key length can still
9247 be 40 bits but the key length can be 168 bits for example. This is fixed
9248 by manually forcing an RC2 key into the EVP_PKEY structure because the
9249 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9250 the key length and effective key length are equal.
9253 *) Add a bunch of functions that should simplify the creation of
9254 X509_NAME structures. Now you should be able to do:
9255 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9256 and have it automatically work out the correct field type and fill in
9257 the structures. The more adventurous can try:
9258 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9259 and it will (hopefully) work out the correct multibyte encoding.
9262 *) Change the 'req' utility to use the new field handling and multibyte
9263 copy routines. Before the DN field creation was handled in an ad hoc
9264 way in req, ca, and x509 which was rather broken and didn't support
9265 BMPStrings or UTF8Strings. Since some software doesn't implement
9266 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9267 using the dirstring_type option. See the new comment in the default
9268 openssl.cnf for more info.
9271 *) Make crypto/rand/md_rand.c more robust:
9272 - Assure unique random numbers after fork().
9273 - Make sure that concurrent threads access the global counter and
9274 md serializably so that we never lose entropy in them
9275 or use exactly the same state in multiple threads.
9276 Access to the large state is not always serializable because
9277 the additional locking could be a performance killer, and
9278 md should be large enough anyway.
9281 *) New file apps/app_rand.c with commonly needed functionality
9282 for handling the random seed file.
9284 Use the random seed file in some applications that previously did not:
9286 dsaparam -genkey (which also ignored its '-rand' option),
9289 x509 (when signing).
9290 Except on systems with /dev/urandom, it is crucial to have a random
9291 seed file at least for key creation, DSA signing, and for DH exchanges;
9292 for RSA signatures we could do without one.
9294 gendh and gendsa (unlike genrsa) used to read only the first byte
9295 of each file listed in the '-rand' option. The function as previously
9296 found in genrsa is now in app_rand.c and is used by all programs
9297 that support '-rand'.
9300 *) In RAND_write_file, use mode 0600 for creating files;
9301 don't just chmod when it may be too late.
9304 *) Report an error from X509_STORE_load_locations
9305 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9308 *) New function ASN1_mbstring_copy() this copies a string in either
9309 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9310 into an ASN1_STRING type. A mask of permissible types is passed
9311 and it chooses the "minimal" type to use or an error if not type
9315 *) Add function equivalents to the various macros in asn1.h. The old
9316 macros are retained with an M_ prefix. Code inside the library can
9317 use the M_ macros. External code (including the openssl utility)
9318 should *NOT* in order to be "shared library friendly".
9321 *) Add various functions that can check a certificate's extensions
9322 to see if it usable for various purposes such as SSL client,
9323 server or S/MIME and CAs of these types. This is currently
9324 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9325 verification. Also added a -purpose flag to x509 utility to
9326 print out all the purposes.
9329 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9333 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9334 for, obtain and decode and extension and obtain its critical flag.
9335 This allows all the necessary extension code to be handled in a
9336 single function call.
9339 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9340 platforms. See crypto/rc4/rc4_enc.c for further details.
9343 *) New -noout option to asn1parse. This causes no output to be produced
9344 its main use is when combined with -strparse and -out to extract data
9345 from a file (which may not be in ASN.1 format).
9348 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9349 when producing the local key id.
9350 [Richard Levitte <levitte@stacken.kth.se>]
9352 *) New option -dhparam in s_server. This allows a DH parameter file to be
9353 stated explicitly. If it is not stated then it tries the first server
9354 certificate file. The previous behaviour hard coded the filename
9358 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9359 a public key to be input or output. For example:
9360 openssl rsa -in key.pem -pubout -out pubkey.pem
9361 Also added necessary DSA public key functions to handle this.
9364 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9365 in the message. This was handled by allowing
9366 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9367 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9369 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9370 to the end of the strings whereas this didn't. This would cause problems
9371 if strings read with d2i_ASN1_bytes() were later modified.
9372 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9374 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9375 data and it contains EOF it will end up returning an error. This is
9376 caused by input 46 bytes long. The cause is due to the way base64
9377 BIOs find the start of base64 encoded data. They do this by trying a
9378 trial decode on each line until they find one that works. When they
9379 do a flag is set and it starts again knowing it can pass all the
9380 data directly through the decoder. Unfortunately it doesn't reset
9381 the context it uses. This means that if EOF is reached an attempt
9382 is made to pass two EOFs through the context and this causes the
9383 resulting error. This can also cause other problems as well. As is
9384 usual with these problems it takes *ages* to find and the fix is
9385 trivial: move one line.
9386 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9388 *) Ugly workaround to get s_client and s_server working under Windows. The
9389 old code wouldn't work because it needed to select() on sockets and the
9390 tty (for keypresses and to see if data could be written). Win32 only
9391 supports select() on sockets so we select() with a 1s timeout on the
9392 sockets and then see if any characters are waiting to be read, if none
9393 are present then we retry, we also assume we can always write data to
9394 the tty. This isn't nice because the code then blocks until we've
9395 received a complete line of data and it is effectively polling the
9396 keyboard at 1s intervals: however it's quite a bit better than not
9397 working at all :-) A dedicated Windows application might handle this
9398 with an event loop for example.
9401 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9402 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9403 will be called when RSA_sign() and RSA_verify() are used. This is useful
9404 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9405 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9406 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9407 This necessitated the support of an extra signature type NID_md5_sha1
9408 for SSL signatures and modifications to the SSL library to use it instead
9409 of calling RSA_public_decrypt() and RSA_private_encrypt().
9412 *) Add new -verify -CAfile and -CApath options to the crl program, these
9413 will lookup a CRL issuers certificate and verify the signature in a
9414 similar way to the verify program. Tidy up the crl program so it
9415 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9416 less strict. It will now permit CRL extensions even if it is not
9417 a V2 CRL: this will allow it to tolerate some broken CRLs.
9420 *) Initialize all non-automatic variables each time one of the openssl
9421 sub-programs is started (this is necessary as they may be started
9422 multiple times from the "OpenSSL>" prompt).
9423 [Lennart Bang, Bodo Moeller]
9425 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9426 removing all other RSA functionality (this is what NO_RSA does). This
9427 is so (for example) those in the US can disable those operations covered
9428 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9432 *) Non-copying interface to BIO pairs.
9433 (still largely untested)
9436 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9437 ASCII string. This was handled independently in various places before.
9440 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9441 UTF8 strings a character at a time.
9444 *) Use client_version from client hello to select the protocol
9445 (s23_srvr.c) and for RSA client key exchange verification
9446 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9449 *) Add various utility functions to handle SPKACs, these were previously
9450 handled by poking round in the structure internals. Added new function
9451 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9452 print, verify and generate SPKACs. Based on an original idea from
9453 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9456 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9459 *) Allow the config file extension section to be overwritten on the
9460 command line. Based on an original idea from Massimiliano Pala
9461 <madwolf@comune.modena.it>. The new option is called -extensions
9462 and can be applied to ca, req and x509. Also -reqexts to override
9463 the request extensions in req and -crlexts to override the crl extensions
9467 *) Add new feature to the SPKAC handling in ca. Now you can include
9468 the same field multiple times by preceding it by "XXXX." for example:
9471 this is the same syntax as used in the req config file.
9474 *) Allow certificate extensions to be added to certificate requests. These
9475 are specified in a 'req_extensions' option of the req section of the
9476 config file. They can be printed out with the -text option to req but
9477 are otherwise ignored at present.
9480 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9481 data read consists of only the final block it would not decrypted because
9482 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9483 A misplaced 'break' also meant the decrypted final block might not be
9484 copied until the next read.
9487 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9488 a few extra parameters to the DH structure: these will be useful if
9489 for example we want the value of 'q' or implement X9.42 DH.
9492 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9493 provides hooks that allow the default DSA functions or functions on a
9494 "per key" basis to be replaced. This allows hardware acceleration and
9495 hardware key storage to be handled without major modification to the
9496 library. Also added low level modexp hooks and CRYPTO_EX structure and
9497 associated functions.
9500 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9501 as "read only": it can't be written to and the buffer it points to will
9502 not be freed. Reading from a read only BIO is much more efficient than
9503 a normal memory BIO. This was added because there are several times when
9504 an area of memory needs to be read from a BIO. The previous method was
9505 to create a memory BIO and write the data to it, this results in two
9506 copies of the data and an O(n^2) reading algorithm. There is a new
9507 function BIO_new_mem_buf() which creates a read only memory BIO from
9508 an area of memory. Also modified the PKCS#7 routines to use read only
9512 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9513 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9514 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9515 but a retry condition occured while trying to read the rest.
9518 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9519 NID_pkcs7_encrypted by default: this was wrong since this should almost
9520 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9521 the encrypted data type: this is a more sensible place to put it and it
9522 allows the PKCS#12 code to be tidied up that duplicated this
9526 *) Changed obj_dat.pl script so it takes its input and output files on
9527 the command line. This should avoid shell escape redirection problems
9531 *) Initial support for certificate extension requests, these are included
9532 in things like Xenroll certificate requests. Included functions to allow
9533 extensions to be obtained and added.
9536 *) -crlf option to s_client and s_server for sending newlines as
9537 CRLF (as required by many protocols).
9540 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9542 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9543 [Ralf S. Engelschall]
9545 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9546 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9548 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9552 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9553 DH parameters/keys (q is lost during that conversion, but the resulting
9554 DH parameters contain its length).
9556 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9557 much faster than DH_generate_parameters (which creates parameters
9558 where p = 2*q + 1), and also the smaller q makes DH computations
9559 much more efficient (160-bit exponentiation instead of 1024-bit
9560 exponentiation); so this provides a convenient way to support DHE
9561 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9562 utter importance to use
9563 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9565 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9566 when such DH parameters are used, because otherwise small subgroup
9567 attacks may become possible!
9570 *) Avoid memory leak in i2d_DHparams.
9573 *) Allow the -k option to be used more than once in the enc program:
9574 this allows the same encrypted message to be read by multiple recipients.
9577 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9578 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9579 it will always use the numerical form of the OID, even if it has a short
9583 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9584 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9585 otherwise bn_mod_exp was called. In the case of hardware keys for example
9586 no private key components need be present and it might store extra data
9587 in the RSA structure, which cannot be accessed from bn_mod_exp.
9588 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9589 private key operations.
9592 *) Added support for SPARC Linux.
9595 *) pem_password_cb function type incompatibly changed from
9596 typedef int pem_password_cb(char *buf, int size, int rwflag);
9598 ....(char *buf, int size, int rwflag, void *userdata);
9599 so that applications can pass data to their callbacks:
9600 The PEM[_ASN1]_{read,write}... functions and macros now take an
9601 additional void * argument, which is just handed through whenever
9602 the password callback is called.
9603 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9605 New function SSL_CTX_set_default_passwd_cb_userdata.
9607 Compatibility note: As many C implementations push function arguments
9608 onto the stack in reverse order, the new library version is likely to
9609 interoperate with programs that have been compiled with the old
9610 pem_password_cb definition (PEM_whatever takes some data that
9611 happens to be on the stack as its last argument, and the callback
9612 just ignores this garbage); but there is no guarantee whatsoever that
9615 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9616 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9617 problems not only on Windows, but also on some Unix platforms.
9618 To avoid problematic command lines, these definitions are now in an
9619 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9620 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9623 *) MIPS III/IV assembler module is reimplemented.
9626 *) More DES library cleanups: remove references to srand/rand and
9627 delete an unused file.
9630 *) Add support for the the free Netwide assembler (NASM) under Win32,
9631 since not many people have MASM (ml) and it can be hard to obtain.
9632 This is currently experimental but it seems to work OK and pass all
9633 the tests. Check out INSTALL.W32 for info.
9636 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9637 without temporary keys kept an extra copy of the server key,
9638 and connections with temporary keys did not free everything in case
9642 *) New function RSA_check_key and new openssl rsa option -check
9643 for verifying the consistency of RSA keys.
9644 [Ulf Moeller, Bodo Moeller]
9646 *) Various changes to make Win32 compile work:
9647 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9648 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9649 comparison" warnings.
9650 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9653 *) Add a debugging option to PKCS#5 v2 key generation function: when
9654 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9655 derived keys are printed to stderr.
9658 *) Copy the flags in ASN1_STRING_dup().
9659 [Roman E. Pavlov <pre@mo.msk.ru>]
9661 *) The x509 application mishandled signing requests containing DSA
9662 keys when the signing key was also DSA and the parameters didn't match.
9664 It was supposed to omit the parameters when they matched the signing key:
9665 the verifying software was then supposed to automatically use the CA's
9666 parameters if they were absent from the end user certificate.
9668 Omitting parameters is no longer recommended. The test was also
9669 the wrong way round! This was probably due to unusual behaviour in
9670 EVP_cmp_parameters() which returns 1 if the parameters match.
9671 This meant that parameters were omitted when they *didn't* match and
9672 the certificate was useless. Certificates signed with 'ca' didn't have
9674 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9676 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9677 The interface is as follows:
9678 Applications can use
9679 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9680 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9681 "off" is now the default.
9682 The library internally uses
9683 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9684 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9685 to disable memory-checking temporarily.
9687 Some inconsistent states that previously were possible (and were
9688 even the default) are now avoided.
9690 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9691 with each memory chunk allocated; this is occasionally more helpful
9692 than just having a counter.
9694 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9696 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9700 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9701 which largely parallels "options", but is for changing API behaviour,
9702 whereas "options" are about protocol behaviour.
9703 Initial "mode" flags are:
9705 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9706 a single record has been written.
9707 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9708 retries use the same buffer location.
9709 (But all of the contents must be
9713 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9716 *) Fix problems with no-hmac etc.
9717 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9719 *) New functions RSA_get_default_method(), RSA_set_method() and
9720 RSA_get_method(). These allows replacement of RSA_METHODs without having
9721 to mess around with the internals of an RSA structure.
9724 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9725 Also really enable memory leak checks in openssl.c and in some
9727 [Chad C. Mulligan, Bodo Moeller]
9729 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9730 up the length of negative integers. This has now been simplified to just
9731 store the length when it is first determined and use it later, rather
9732 than trying to keep track of where data is copied and updating it to
9734 [Steve Henson, reported by Brien Wheeler
9735 <bwheeler@authentica-security.com>]
9737 *) Add a new function PKCS7_signatureVerify. This allows the verification
9738 of a PKCS#7 signature but with the signing certificate passed to the
9739 function itself. This contrasts with PKCS7_dataVerify which assumes the
9740 certificate is present in the PKCS#7 structure. This isn't always the
9741 case: certificates can be omitted from a PKCS#7 structure and be
9742 distributed by "out of band" means (such as a certificate database).
9745 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9746 function prototypes in pem.h, also change util/mkdef.pl to add the
9747 necessary function names.
9750 *) mk1mf.pl (used by Windows builds) did not properly read the
9751 options set by Configure in the top level Makefile, and Configure
9752 was not even able to write more than one option correctly.
9753 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9756 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9757 file to be loaded from a BIO or FILE pointer. The BIO version will
9758 for example allow memory BIOs to contain config info.
9761 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9762 Whoever hopes to achieve shared-library compatibility across versions
9763 must use this, not the compile-time macro.
9764 (Exercise 0.9.4: Which is the minimum library version required by
9766 Note: All this applies only to multi-threaded programs, others don't
9770 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9771 through a BIO pair triggered the default case, i.e.
9772 SSLerr(...,SSL_R_UNKNOWN_STATE).
9775 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9776 can use the SSL library even if none of the specific BIOs is
9780 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9781 for the encoded length.
9782 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9784 *) Add initial documentation of the X509V3 functions.
9787 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9788 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9789 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9790 secure PKCS#8 private key format with a high iteration count.
9793 *) Fix determination of Perl interpreter: A perl or perl5
9794 _directory_ in $PATH was also accepted as the interpreter.
9795 [Ralf S. Engelschall]
9797 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9798 wrong with it but it was very old and did things like calling
9799 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9803 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9804 to use the new extension code.
9807 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9808 with macros. This should make it easier to change their form, add extra
9809 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9813 *) Add to configuration table a new entry that can specify an alternative
9814 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9815 according to Mark Crispin <MRC@Panda.COM>.
9819 *) DES CBC did not update the IV. Weird.
9822 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9823 Changing the behaviour of the former might break existing programs --
9824 where IV updating is needed, des_ncbc_encrypt can be used.
9827 *) When bntest is run from "make test" it drives bc to check its
9828 calculations, as well as internally checking them. If an internal check
9829 fails, it needs to cause bc to give a non-zero result or make test carries
9830 on without noticing the failure. Fixed.
9833 *) DES library cleanups.
9836 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9837 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9838 ciphers. NOTE: although the key derivation function has been verified
9839 against some published test vectors it has not been extensively tested
9840 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9844 *) Instead of "mkdir -p", which is not fully portable, use new
9845 Perl script "util/mkdir-p.pl".
9848 *) Rewrite the way password based encryption (PBE) is handled. It used to
9849 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9850 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9851 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9852 the 'parameter' field of the AlgorithmIdentifier is passed to the
9853 underlying key generation function so it must do its own ASN1 parsing.
9854 This has also changed the EVP_PBE_CipherInit() function which now has a
9855 'parameter' argument instead of literal salt and iteration count values
9856 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9859 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9860 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9861 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9862 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9863 value was just used as a "magic string" and not used directly its
9864 value doesn't matter.
9867 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9871 *) "linux-sparc64" configuration (ultrapenguin).
9872 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9873 "linux-sparc" configuration.
9874 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9876 *) config now generates no-xxx options for missing ciphers.
9879 *) Support the EBCDIC character set (work in progress).
9880 File ebcdic.c not yet included because it has a different license.
9881 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9883 *) Support BS2000/OSD-POSIX.
9884 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9886 *) Make callbacks for key generation use void * instead of char *.
9889 *) Make S/MIME samples compile (not yet tested).
9892 *) Additional typesafe stacks.
9895 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9899 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9901 *) New configuration variant "sco5-gcc".
9903 *) Updated some demos.
9904 [Sean O Riordain, Wade Scholine]
9906 *) Add missing BIO_free at exit of pkcs12 application.
9909 *) Fix memory leak in conf.c.
9912 *) Updates for Win32 to assembler version of MD5.
9915 *) Set #! path to perl in apps/der_chop to where we found it
9916 instead of using a fixed path.
9919 *) SHA library changes for irix64-mips4-cc.
9922 *) Improvements for VMS support.
9926 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9928 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9929 This also avoids the problems with SC4.2 and unpatched SC5.
9930 [Andy Polyakov <appro@fy.chalmers.se>]
9932 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9933 These are required because of the typesafe stack would otherwise break
9934 existing code. If old code used a structure member which used to be STACK
9935 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9936 sk_num or sk_value it would produce an error because the num, data members
9937 are not present in STACK_OF. Now it just produces a warning. sk_set
9938 replaces the old method of assigning a value to sk_value
9939 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9940 that does this will no longer work (and should use sk_set instead) but
9941 this could be regarded as a "questionable" behaviour anyway.
9944 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9945 correctly handle encrypted S/MIME data.
9948 *) Change type of various DES function arguments from des_cblock
9949 (which means, in function argument declarations, pointer to char)
9950 to des_cblock * (meaning pointer to array with 8 char elements),
9951 which allows the compiler to do more typechecking; it was like
9952 that back in SSLeay, but with lots of ugly casts.
9954 Introduce new type const_des_cblock.
9957 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9958 problems: find RecipientInfo structure that matches recipient certificate
9959 and initialise the ASN1 structures properly based on passed cipher.
9962 *) Belatedly make the BN tests actually check the results.
9965 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9966 to and from BNs: it was completely broken. New compilation option
9967 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9968 key elements as negative integers.
9971 *) Reorganize and speed up MD5.
9972 [Andy Polyakov <appro@fy.chalmers.se>]
9975 [Richard Levitte <richard@levitte.org>]
9977 *) New option -out to asn1parse to allow the parsed structure to be
9978 output to a file. This is most useful when combined with the -strparse
9979 option to examine the output of things like OCTET STRINGS.
9982 *) Make SSL library a little more fool-proof by not requiring any longer
9983 that SSL_set_{accept,connect}_state be called before
9984 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9985 in many applications because usually everything *appeared* to work as
9986 intended anyway -- now it really works as intended).
9989 *) Move openssl.cnf out of lib/.
9992 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9993 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9994 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9995 [Ralf S. Engelschall]
9997 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9998 handle PKCS#7 enveloped data properly.
9999 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10001 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10002 copying pointers. The cert_st handling is changed by this in
10003 various ways (and thus what used to be known as ctx->default_cert
10004 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10005 any longer when s->cert does not give us what we need).
10006 ssl_cert_instantiate becomes obsolete by this change.
10007 As soon as we've got the new code right (possibly it already is?),
10008 we have solved a couple of bugs of the earlier code where s->cert
10009 was used as if it could not have been shared with other SSL structures.
10011 Note that using the SSL API in certain dirty ways now will result
10012 in different behaviour than observed with earlier library versions:
10013 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10014 does not influence s as it used to.
10016 In order to clean up things more thoroughly, inside SSL_SESSION
10017 we don't use CERT any longer, but a new structure SESS_CERT
10018 that holds per-session data (if available); currently, this is
10019 the peer's certificate chain and, for clients, the server's certificate
10020 and temporary key. CERT holds only those values that can have
10021 meaningful defaults in an SSL_CTX.
10024 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10025 from the internal representation. Various PKCS#7 fixes: remove some
10026 evil casts and set the enc_dig_alg field properly based on the signing
10030 *) Allow PKCS#12 password to be set from the command line or the
10031 environment. Let 'ca' get its config file name from the environment
10032 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10036 *) Allow certificate policies extension to use an IA5STRING for the
10037 organization field. This is contrary to the PKIX definition but
10038 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10042 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10043 without disallowing inline assembler and the like for non-pedantic builds.
10046 *) Support Borland C++ builder.
10047 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10049 *) Support Mingw32.
10052 *) SHA-1 cleanups and performance enhancements.
10053 [Andy Polyakov <appro@fy.chalmers.se>]
10055 *) Sparc v8plus assembler for the bignum library.
10056 [Andy Polyakov <appro@fy.chalmers.se>]
10058 *) Accept any -xxx and +xxx compiler options in Configure.
10061 *) Update HPUX configuration.
10064 *) Add missing sk_<type>_unshift() function to safestack.h
10065 [Ralf S. Engelschall]
10067 *) New function SSL_CTX_use_certificate_chain_file that sets the
10068 "extra_cert"s in addition to the certificate. (This makes sense
10069 only for "PEM" format files, as chains as a whole are not
10073 *) Support verify_depth from the SSL API.
10074 x509_vfy.c had what can be considered an off-by-one-error:
10075 Its depth (which was not part of the external interface)
10076 was actually counting the number of certificates in a chain;
10077 now it really counts the depth.
10080 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10081 instead of X509err, which often resulted in confusing error
10082 messages since the error codes are not globally unique
10083 (e.g. an alleged error in ssl3_accept when a certificate
10084 didn't match the private key).
10086 *) New function SSL_CTX_set_session_id_context that allows to set a default
10087 value (so that you don't need SSL_set_session_id_context for each
10088 connection using the SSL_CTX).
10091 *) OAEP decoding bug fix.
10094 *) Support INSTALL_PREFIX for package builders, as proposed by
10098 *) New Configure options "threads" and "no-threads". For systems
10099 where the proper compiler options are known (currently Solaris
10100 and Linux), "threads" is the default.
10103 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10106 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10107 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10108 such as /usr/local/bin.
10111 *) "make linux-shared" to build shared libraries.
10112 [Niels Poppe <niels@netbox.org>]
10114 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10117 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10118 extension adding in x509 utility.
10121 *) Remove NOPROTO sections and error code comments.
10124 *) Partial rewrite of the DEF file generator to now parse the ANSI
10128 *) New Configure options --prefix=DIR and --openssldir=DIR.
10131 *) Complete rewrite of the error code script(s). It is all now handled
10132 by one script at the top level which handles error code gathering,
10133 header rewriting and C source file generation. It should be much better
10134 than the old method: it now uses a modified version of Ulf's parser to
10135 read the ANSI prototypes in all header files (thus the old K&R definitions
10136 aren't needed for error creation any more) and do a better job of
10137 translating function codes into names. The old 'ASN1 error code imbedded
10138 in a comment' is no longer necessary and it doesn't use .err files which
10139 have now been deleted. Also the error code call doesn't have to appear all
10140 on one line (which resulted in some large lines...).
10143 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10146 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10147 0 (which usually indicates a closed connection), but continue reading.
10150 *) Fix some race conditions.
10153 *) Add support for CRL distribution points extension. Add Certificate
10154 Policies and CRL distribution points documentation.
10157 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10160 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10161 8 of keying material. Merlin has also confirmed interop with this fix
10162 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10163 [Merlin Hughes <merlin@baltimore.ie>]
10165 *) Fix lots of warnings.
10166 [Richard Levitte <levitte@stacken.kth.se>]
10168 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10169 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10170 [Richard Levitte <levitte@stacken.kth.se>]
10172 *) Fix problems with sizeof(long) == 8.
10173 [Andy Polyakov <appro@fy.chalmers.se>]
10175 *) Change functions to ANSI C.
10178 *) Fix typos in error codes.
10179 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10181 *) Remove defunct assembler files from Configure.
10184 *) SPARC v8 assembler BIGNUM implementation.
10185 [Andy Polyakov <appro@fy.chalmers.se>]
10187 *) Support for Certificate Policies extension: both print and set.
10188 Various additions to support the r2i method this uses.
10191 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10192 return a const string when you are expecting an allocated buffer.
10195 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10196 types DirectoryString and DisplayText.
10199 *) Add code to allow r2i extensions to access the configuration database,
10200 add an LHASH database driver and add several ctx helper functions.
10203 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10204 fail when they extended the size of a BIGNUM.
10207 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10208 support typesafe stack.
10211 *) Fix typo in SSL_[gs]et_options().
10212 [Nils Frostberg <nils@medcom.se>]
10214 *) Delete various functions and files that belonged to the (now obsolete)
10215 old X509V3 handling code.
10218 *) New Configure option "rsaref".
10221 *) Don't auto-generate pem.h.
10224 *) Introduce type-safe ASN.1 SETs.
10227 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10228 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10230 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10231 that links with OpenSSL (well at least cause lots of warnings), but fear
10232 not: the conversion is trivial, and it eliminates loads of evil casts. A
10233 few STACKed things have been converted already. Feel free to convert more.
10234 In the fullness of time, I'll do away with the STACK type altogether.
10237 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10238 specified in <certfile> by updating the entry in the index.txt file.
10239 This way one no longer has to edit the index.txt file manually for
10240 revoking a certificate. The -revoke option does the gory details now.
10241 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10243 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10244 `-text' option at all and this way the `-noout -text' combination was
10245 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10246 [Ralf S. Engelschall]
10248 *) Make sure a corresponding plain text error message exists for the
10249 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10250 verify callback function determined that a certificate was revoked.
10251 [Ralf S. Engelschall]
10253 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10254 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10255 all available cipers including rc5, which was forgotten until now.
10256 In order to let the testing shell script know which algorithms
10257 are available, a new (up to now undocumented) command
10258 "openssl list-cipher-commands" is used.
10261 *) Bugfix: s_client occasionally would sleep in select() when
10262 it should have checked SSL_pending() first.
10265 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10266 the raw DSA values prior to ASN.1 encoding.
10269 *) Tweaks to Configure
10270 [Niels Poppe <niels@netbox.org>]
10272 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10276 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10279 *) New config option to avoid instructions that are illegal on the 80386.
10280 The default code is faster, but requires at least a 486.
10283 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10284 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10285 same as SSL2_VERSION anyway.
10288 *) New "-showcerts" option for s_client.
10291 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10292 application. Various cleanups and fixes.
10295 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10296 modify error routines to work internally. Add error codes and PBE init
10297 to library startup routines.
10300 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10301 packing functions to asn1 and evp. Changed function names and error
10302 codes along the way.
10305 *) PKCS12 integration: and so it begins... First of several patches to
10306 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10307 objects to objects.h
10310 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10311 and display support for Thawte strong extranet extension.
10314 *) Add LinuxPPC support.
10315 [Jeff Dubrule <igor@pobox.org>]
10317 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10318 bn_div_words in alpha.s.
10319 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10321 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10322 OAEP isn't supported when OpenSSL is built with RSAref.
10323 [Ulf Moeller <ulf@fitug.de>]
10325 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10326 so they no longer are missing under -DNOPROTO.
10327 [Soren S. Jorvang <soren@t.dk>]
10330 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10332 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10333 doesn't work when the session is reused. Coming soon!
10336 *) Fix a security hole, that allows sessions to be reused in the wrong
10337 context thus bypassing client cert protection! All software that uses
10338 client certs and session caches in multiple contexts NEEDS PATCHING to
10339 allow session reuse! A fuller solution is in the works.
10340 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10342 *) Some more source tree cleanups (removed obsolete files
10343 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10344 permission on "config" script to be executable) and a fix for the INSTALL
10346 [Ulf Moeller <ulf@fitug.de>]
10348 *) Remove some legacy and erroneous uses of malloc, free instead of
10350 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10352 *) Make rsa_oaep_test return non-zero on error.
10353 [Ulf Moeller <ulf@fitug.de>]
10355 *) Add support for native Solaris shared libraries. Configure
10356 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10357 if someone would make that last step automatic.
10358 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10360 *) ctx_size was not built with the right compiler during "make links". Fixed.
10363 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10364 except NULL ciphers". This means the default cipher list will no longer
10365 enable NULL ciphers. They need to be specifically enabled e.g. with
10366 the string "DEFAULT:eNULL".
10369 *) Fix to RSA private encryption routines: if p < q then it would
10370 occasionally produce an invalid result. This will only happen with
10371 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10374 *) Be less restrictive and allow also `perl util/perlpath.pl
10375 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10376 because this way one can also use an interpreter named `perl5' (which is
10377 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10378 installed as `perl').
10379 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10381 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10382 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10384 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10385 advapi32.lib to Win32 build and change the pem test comparison
10386 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10387 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10388 and crypto/des/ede_cbcm_enc.c.
10391 *) DES quad checksum was broken on big-endian architectures. Fixed.
10394 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10395 Win32 test batch file so it (might) work again. The Win32 test batch file
10396 is horrible: I feel ill....
10399 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10400 in e_os.h. Audit of header files to check ANSI and non ANSI
10401 sections: 10 functions were absent from non ANSI section and not exported
10402 from Windows DLLs. Fixed up libeay.num for new functions.
10405 *) Make `openssl version' output lines consistent.
10406 [Ralf S. Engelschall]
10408 *) Fix Win32 symbol export lists for BIO functions: Added
10409 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10410 to ms/libeay{16,32}.def.
10411 [Ralf S. Engelschall]
10413 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10414 fine under Unix and passes some trivial tests I've now added. But the
10415 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10416 added to make sure no one expects that this stuff really works in the
10417 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10418 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10420 [Ralf S. Engelschall]
10422 *) Fix the generation of two part addresses in perl.
10423 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10425 *) Add config entry for Linux on MIPS.
10426 [John Tobey <jtobey@channel1.com>]
10428 *) Make links whenever Configure is run, unless we are on Windoze.
10431 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10432 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10436 *) Add a useful kludge to allow package maintainers to specify compiler and
10437 other platforms details on the command line without having to patch the
10438 Configure script everytime: One now can use ``perl Configure
10439 <id>:<details>'', i.e. platform ids are allowed to have details appended
10440 to them (separated by colons). This is treated as there would be a static
10441 pre-configured entry in Configure's %table under key <id> with value
10442 <details> and ``perl Configure <id>'' is called. So, when you want to
10443 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10444 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10445 now, which overrides the FreeBSD-elf entry on-the-fly.
10446 [Ralf S. Engelschall]
10448 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10451 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10452 on the `perl Configure ...' command line. This way one can compile
10453 OpenSSL libraries with Position Independent Code (PIC) which is needed
10454 for linking it into DSOs.
10455 [Ralf S. Engelschall]
10457 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10461 *) Cleaned up the LICENSE document: The official contact for any license
10462 questions now is the OpenSSL core team under openssl-core@openssl.org.
10463 And add a paragraph about the dual-license situation to make sure people
10464 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10465 to the OpenSSL toolkit.
10466 [Ralf S. Engelschall]
10468 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10469 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10470 Additionally cleaned up the `make links' target: Remove unnecessary
10471 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10472 to speed processing and no longer clutter the display with confusing
10473 stuff. Instead only the actually done links are displayed.
10474 [Ralf S. Engelschall]
10476 *) Permit null encryption ciphersuites, used for authentication only. It used
10477 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10478 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10482 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10483 signed attributes when verifying signatures (this would break them),
10484 the detached data encoding was wrong and public keys obtained using
10485 X509_get_pubkey() weren't freed.
10488 *) Add text documentation for the BUFFER functions. Also added a work around
10489 to a Win95 console bug. This was triggered by the password read stuff: the
10490 last character typed gets carried over to the next fread(). If you were
10491 generating a new cert request using 'req' for example then the last
10492 character of the passphrase would be CR which would then enter the first
10496 *) Added the new `Includes OpenSSL Cryptography Software' button as
10497 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10498 button and can be used by applications based on OpenSSL to show the
10499 relationship to the OpenSSL project.
10500 [Ralf S. Engelschall]
10502 *) Remove confusing variables in function signatures in files
10503 ssl/ssl_lib.c and ssl/ssl.h.
10504 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10506 *) Don't install bss_file.c under PREFIX/include/
10507 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10509 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10510 functions that return function pointers and has support for NT specific
10511 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10512 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10513 unsigned to signed types: this was killing the Win32 compile.
10516 *) Add new certificate file to stack functions,
10517 SSL_add_dir_cert_subjects_to_stack() and
10518 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10519 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10520 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10521 This means that Apache-SSL and similar packages don't have to mess around
10522 to add as many CAs as they want to the preferred list.
10525 *) Experiment with doxygen documentation. Currently only partially applied to
10527 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10528 openssl.doxy as the configuration file.
10531 *) Get rid of remaining C++-style comments which strict C compilers hate.
10532 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10534 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10535 compiled in by default: it has problems with large keys.
10538 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10539 DH private keys and/or callback functions which directly correspond to
10540 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10541 is needed for applications which have to configure certificates on a
10542 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10544 For the RSA certificate situation is makes no difference, but
10545 for the DSA certificate situation this fixes the "no shared cipher"
10546 problem where the OpenSSL cipher selection procedure failed because the
10547 temporary keys were not overtaken from the context and the API provided
10548 no way to reconfigure them.
10549 The new functions now let applications reconfigure the stuff and they
10550 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10551 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10552 non-public-API function ssl_cert_instantiate() is used as a helper
10553 function and also to reduce code redundancy inside ssl_rsa.c.
10554 [Ralf S. Engelschall]
10556 *) Move s_server -dcert and -dkey options out of the undocumented feature
10557 area because they are useful for the DSA situation and should be
10558 recognized by the users.
10559 [Ralf S. Engelschall]
10561 *) Fix the cipher decision scheme for export ciphers: the export bits are
10562 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10563 SSL_EXP_MASK. So, the original variable has to be used instead of the
10564 already masked variable.
10565 [Richard Levitte <levitte@stacken.kth.se>]
10567 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10568 [Richard Levitte <levitte@stacken.kth.se>]
10570 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10571 from `int' to `unsigned int' because it's a length and initialized by
10572 EVP_DigestFinal() which expects an `unsigned int *'.
10573 [Richard Levitte <levitte@stacken.kth.se>]
10575 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10576 script. Instead use the usual Shell->Perl transition trick.
10577 [Ralf S. Engelschall]
10579 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10580 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10581 -noout -modulus' as it's already the case for `openssl rsa -noout
10582 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10583 currently the public key is printed (a decision which was already done by
10584 `openssl dsa -modulus' in the past) which serves a similar purpose.
10585 Additionally the NO_RSA no longer completely removes the whole -modulus
10586 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10588 [Ralf S. Engelschall]
10590 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10591 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10592 [Arne Ansper <arne@ats.cyber.ee>]
10594 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10595 to be added. Now both 'req' and 'ca' can use new objects defined in the
10599 *) Add cool BIO that does syslog (or event log on NT).
10600 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10602 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10603 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10604 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10605 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10608 *) Add preliminary config info for new extension code.
10611 *) Make RSA_NO_PADDING really use no padding.
10612 [Ulf Moeller <ulf@fitug.de>]
10614 *) Generate errors when private/public key check is done.
10617 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10618 for some CRL extensions and new objects added.
10621 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10622 key usage extension and fuller support for authority key id.
10625 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10626 padding method for RSA, which is recommended for new applications in PKCS
10627 #1 v2.0 (RFC 2437, October 1998).
10628 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10629 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10630 against Bleichbacher's attack on RSA.
10631 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10634 *) Updates to the new SSL compression code
10635 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10637 *) Fix so that the version number in the master secret, when passed
10638 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10639 (because the server will not accept higher), that the version number
10640 is 0x03,0x01, not 0x03,0x00
10641 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10643 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10644 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10645 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10648 *) Support for RAW extensions where an arbitrary extension can be
10649 created by including its DER encoding. See apps/openssl.cnf for
10653 *) Make sure latest Perl versions don't interpret some generated C array
10654 code as Perl array code in the crypto/err/err_genc.pl script.
10655 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10657 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10658 not many people have the assembler. Various Win32 compilation fixes and
10659 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10660 build instructions.
10663 *) Modify configure script 'Configure' to automatically create crypto/date.h
10664 file under Win32 and also build pem.h from pem.org. New script
10665 util/mkfiles.pl to create the MINFO file on environments that can't do a
10666 'make files': perl util/mkfiles.pl >MINFO should work.
10669 *) Major rework of DES function declarations, in the pursuit of correctness
10670 and purity. As a result, many evil casts evaporated, and some weirdness,
10671 too. You may find this causes warnings in your code. Zapping your evil
10672 casts will probably fix them. Mostly.
10675 *) Fix for a typo in asn1.h. Bug fix to object creation script
10676 obj_dat.pl. It considered a zero in an object definition to mean
10677 "end of object": none of the objects in objects.h have any zeros
10678 so it wasn't spotted.
10679 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10681 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10682 Masking (CBCM). In the absence of test vectors, the best I have been able
10683 to do is check that the decrypt undoes the encrypt, so far. Send me test
10684 vectors if you have them.
10687 *) Correct calculation of key length for export ciphers (too much space was
10688 allocated for null ciphers). This has not been tested!
10691 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10692 message is now correct (it understands "crypto" and "ssl" on its
10693 command line). There is also now an "update" option. This will update
10694 the util/ssleay.num and util/libeay.num files with any new functions.
10696 perl util/mkdef.pl crypto ssl update
10697 it will update them.
10700 *) Overhauled the Perl interface (perl/*):
10701 - ported BN stuff to OpenSSL's different BN library
10702 - made the perl/ source tree CVS-aware
10703 - renamed the package from SSLeay to OpenSSL (the files still contain
10704 their history because I've copied them in the repository)
10705 - removed obsolete files (the test scripts will be replaced
10706 by better Test::Harness variants in the future)
10707 [Ralf S. Engelschall]
10709 *) First cut for a very conservative source tree cleanup:
10710 1. merge various obsolete readme texts into doc/ssleay.txt
10711 where we collect the old documents and readme texts.
10712 2. remove the first part of files where I'm already sure that we no
10713 longer need them because of three reasons: either they are just temporary
10714 files which were left by Eric or they are preserved original files where
10715 I've verified that the diff is also available in the CVS via "cvs diff
10716 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10717 the crypto/md/ stuff).
10718 [Ralf S. Engelschall]
10720 *) More extension code. Incomplete support for subject and issuer alt
10721 name, issuer and authority key id. Change the i2v function parameters
10722 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10723 what that's for :-) Fix to ASN1 macro which messed up
10724 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10727 *) Preliminary support for ENUMERATED type. This is largely copied from the
10731 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10732 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10734 *) Make sure `make rehash' target really finds the `openssl' program.
10735 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10737 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10738 like to hear about it if this slows down other processors.
10741 *) Add CygWin32 platform information to Configure script.
10742 [Alan Batie <batie@aahz.jf.intel.com>]
10744 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10745 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10747 *) New program nseq to manipulate netscape certificate sequences
10750 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10754 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10755 but the BN code had some problems that would cause failures when
10756 doing certificate verification and some other functions.
10757 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10759 *) Add ASN1 and PEM code to support netscape certificate sequences.
10762 *) Add ASN1 and PEM code to support netscape certificate sequences.
10765 *) Add several PKIX and private extended key usage OIDs.
10768 *) Modify the 'ca' program to handle the new extension code. Modify
10769 openssl.cnf for new extension format, add comments.
10772 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10773 and add a sample to openssl.cnf so req -x509 now adds appropriate
10777 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10778 error code, add initial support to X509_print() and x509 application.
10781 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10782 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10783 stuff is currently isolated and isn't even compiled yet.
10786 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10787 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10788 Removed the versions check from X509 routines when loading extensions:
10789 this allows certain broken certificates that don't set the version
10790 properly to be processed.
10793 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10794 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10795 can still be regenerated with "make depend".
10798 *) Spelling mistake in C version of CAST-128.
10799 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10801 *) Changes to the error generation code. The perl script err-code.pl
10802 now reads in the old error codes and retains the old numbers, only
10803 adding new ones if necessary. It also only changes the .err files if new
10804 codes are added. The makefiles have been modified to only insert errors
10805 when needed (to avoid needlessly modifying header files). This is done
10806 by only inserting errors if the .err file is newer than the auto generated
10807 C file. To rebuild all the error codes from scratch (the old behaviour)
10808 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10809 or delete all the .err files.
10812 *) CAST-128 was incorrectly implemented for short keys. The C version has
10813 been fixed, but is untested. The assembler versions are also fixed, but
10814 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10815 to regenerate it if needed.
10816 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10817 Hagino <itojun@kame.net>]
10819 *) File was opened incorrectly in randfile.c.
10820 [Ulf Möller <ulf@fitug.de>]
10822 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10823 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10824 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10825 al: it's just almost always a UTCTime. Note this patch adds new error
10826 codes so do a "make errors" if there are problems.
10829 *) Correct Linux 1 recognition in config.
10830 [Ulf Möller <ulf@fitug.de>]
10832 *) Remove pointless MD5 hash when using DSA keys in ca.
10833 [Anonymous <nobody@replay.com>]
10835 *) Generate an error if given an empty string as a cert directory. Also
10836 generate an error if handed NULL (previously returned 0 to indicate an
10837 error, but didn't set one).
10838 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10840 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10843 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10844 parameters. This was causing a warning which killed off the Win32 compile.
10847 *) Remove C++ style comments from crypto/bn/bn_local.h.
10848 [Neil Costigan <neil.costigan@celocom.com>]
10850 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10851 based on a text string, looking up short and long names and finally
10852 "dot" format. The "dot" format stuff didn't work. Added new function
10853 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10854 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10855 OID is not part of the table.
10858 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10859 X509_LOOKUP_by_alias().
10862 *) Sort openssl functions by name.
10865 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10866 encryption from sample DSA keys (in case anyone is interested the password
10870 *) Make _all_ *_free functions accept a NULL pointer.
10871 [Frans Heymans <fheymans@isaserver.be>]
10873 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10875 [Anonymous <nobody@replay.com>]
10877 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10878 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10880 *) Don't blow it for numeric -newkey arguments to apps/req.
10881 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10883 *) Temp key "for export" tests were wrong in s3_srvr.c.
10884 [Anonymous <nobody@replay.com>]
10886 *) Add prototype for temp key callback functions
10887 SSL_CTX_set_tmp_{rsa,dh}_callback().
10890 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10891 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10894 *) X509_name_add_entry() freed the wrong thing after an error.
10895 [Arne Ansper <arne@ats.cyber.ee>]
10897 *) rsa_eay.c would attempt to free a NULL context.
10898 [Arne Ansper <arne@ats.cyber.ee>]
10900 *) BIO_s_socket() had a broken should_retry() on Windoze.
10901 [Arne Ansper <arne@ats.cyber.ee>]
10903 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10904 [Arne Ansper <arne@ats.cyber.ee>]
10906 *) Make sure the already existing X509_STORE->depth variable is initialized
10907 in X509_STORE_new(), but document the fact that this variable is still
10908 unused in the certificate verification process.
10909 [Ralf S. Engelschall]
10911 *) Fix the various library and apps files to free up pkeys obtained from
10912 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10915 *) Fix reference counting in X509_PUBKEY_get(). This makes
10916 demos/maurice/example2.c work, amongst others, probably.
10917 [Steve Henson and Ben Laurie]
10919 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10920 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10921 are no longer created. This way we have a single and consistent command
10922 line interface `openssl <command>', similar to `cvs <command>'.
10923 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10925 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10926 BIT STRING wrapper always have zero unused bits.
10929 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10932 *) Make the top-level INSTALL documentation easier to understand.
10935 *) Makefiles updated to exit if an error occurs in a sub-directory
10936 make (including if user presses ^C) [Paul Sutton]
10938 *) Make Montgomery context stuff explicit in RSA data structure.
10941 *) Fix build order of pem and err to allow for generated pem.h.
10944 *) Fix renumbering bug in X509_NAME_delete_entry().
10947 *) Enhanced the err-ins.pl script so it makes the error library number
10948 global and can add a library name. This is needed for external ASN1 and
10949 other error libraries.
10952 *) Fixed sk_insert which never worked properly.
10955 *) Fix ASN1 macros so they can handle indefinite length constructed
10956 EXPLICIT tags. Some non standard certificates use these: they can now
10960 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10961 into a single doc/ssleay.txt bundle. This way the information is still
10962 preserved but no longer messes up this directory. Now it's new room for
10963 the new set of documentation files.
10964 [Ralf S. Engelschall]
10966 *) SETs were incorrectly DER encoded. This was a major pain, because they
10967 shared code with SEQUENCEs, which aren't coded the same. This means that
10968 almost everything to do with SETs or SEQUENCEs has either changed name or
10969 number of arguments.
10970 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10972 *) Fix test data to work with the above.
10975 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10976 was already fixed by Eric for 0.9.1 it seems.
10977 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10979 *) Autodetect FreeBSD3.
10982 *) Fix various bugs in Configure. This affects the following platforms:
10986 unixware-2.0-pentium
10990 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10991 before they are needed.
10994 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10998 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11000 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11001 changed SSLeay to OpenSSL in version strings.
11002 [Ralf S. Engelschall]
11004 *) Some fixups to the top-level documents.
11007 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11008 because the symlink to include/ was missing.
11009 [Ralf S. Engelschall]
11011 *) Incorporated the popular no-RSA/DSA-only patches
11012 which allow to compile a RSA-free SSLeay.
11013 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11015 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11016 when "ssleay" is still not found.
11017 [Ralf S. Engelschall]
11019 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11020 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11022 *) Updated the README file.
11023 [Ralf S. Engelschall]
11025 *) Added various .cvsignore files in the CVS repository subdirs
11026 to make a "cvs update" really silent.
11027 [Ralf S. Engelschall]
11029 *) Recompiled the error-definition header files and added
11030 missing symbols to the Win32 linker tables.
11031 [Ralf S. Engelschall]
11033 *) Cleaned up the top-level documents;
11034 o new files: CHANGES and LICENSE
11035 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11036 o merged COPYRIGHT into LICENSE
11037 o removed obsolete TODO file
11038 o renamed MICROSOFT to INSTALL.W32
11039 [Ralf S. Engelschall]
11041 *) Removed dummy files from the 0.9.1b source tree:
11042 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11043 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11044 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11045 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11046 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11047 [Ralf S. Engelschall]
11049 *) Added various platform portability fixes.
11052 *) The Genesis of the OpenSSL rpject:
11053 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11054 Young and Tim J. Hudson created while they were working for C2Net until
11056 [The OpenSSL Project]
11059 Changes between 0.9.0b and 0.9.1b [not released]
11061 *) Updated a few CA certificates under certs/
11064 *) Changed some BIGNUM api stuff.
11067 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11068 DGUX x86, Linux Alpha, etc.
11071 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11072 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11076 *) Add -strparse option to asn1pars program which parses nested
11078 [Dr Stephen Henson <shenson@bigfoot.com>]
11080 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11083 *) DSA fix for "ca" program.
11086 *) Added "-genkey" option to "dsaparam" program.
11089 *) Added RIPE MD160 (rmd160) message digest.
11092 *) Added -a (all) option to "ssleay version" command.
11095 *) Added PLATFORM define which is the id given to Configure.
11098 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11101 *) Extended the ASN.1 parser routines.
11104 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11107 *) Added a BN_CTX to the BN library.
11110 *) Fixed the weak key values in DES library
11113 *) Changed API in EVP library for cipher aliases.
11116 *) Added support for RC2/64bit cipher.
11119 *) Converted the lhash library to the crypto/mem.c functions.
11122 *) Added more recognized ASN.1 object ids.
11125 *) Added more RSA padding checks for SSL/TLS.
11128 *) Added BIO proxy/filter functionality.
11131 *) Added extra_certs to SSL_CTX which can be used
11132 send extra CA certificates to the client in the CA cert chain sending
11133 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11136 *) Now Fortezza is denied in the authentication phase because
11137 this is key exchange mechanism is not supported by SSLeay at all.
11140 *) Additional PKCS1 checks.
11143 *) Support the string "TLSv1" for all TLS v1 ciphers.
11146 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11147 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11150 *) Fixed a few memory leaks.
11153 *) Fixed various code and comment typos.
11156 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11157 bytes sent in the client random.
11158 [Edward Bishop <ebishop@spyglass.com>]