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
39 *) Experimental support for a new, fast, unbiased prime candidate generator,
40 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
41 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
43 *) New output format NSS in the sess_id command line tool. This allows
44 exporting the session id and the master key in NSS keylog format.
45 [Martin Kaiser <martin@kaiser.cx>]
47 *) Harmonize version and its documentation. -f flag is used to display
49 [mancha <mancha1@zoho.com>]
51 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
53 [mancha <mancha1@zoho.com>]
55 *) Fix some double frees. These are not thought to be exploitable.
56 [mancha <mancha1@zoho.com>]
58 *) A missing bounds check in the handling of the TLS heartbeat extension
59 can be used to reveal up to 64k of memory to a connected client or
62 Thanks for Neel Mehta of Google Security for discovering this bug and to
63 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
64 preparing the fix (CVE-2014-0160)
65 [Adam Langley, Bodo Moeller]
67 *) Fix for the attack described in the paper "Recovering OpenSSL
68 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
69 by Yuval Yarom and Naomi Benger. Details can be obtained from:
70 http://eprint.iacr.org/2014/140
72 Thanks to Yuval Yarom and Naomi Benger for discovering this
73 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
74 [Yuval Yarom and Naomi Benger]
76 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
77 this fixes a limitation in previous versions of OpenSSL.
80 *) Experimental encrypt-then-mac support.
82 Experimental support for encrypt then mac from
83 draft-gutmann-tls-encrypt-then-mac-02.txt
85 To enable it set the appropriate extension number (0x42 for the test
86 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
88 For non-compliant peers (i.e. just about everything) this should have no
91 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
95 *) Add EVP support for key wrapping algorithms, to avoid problems with
96 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
97 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
98 algorithms and include tests cases.
101 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
105 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
106 MGF1 digest and OAEP label.
109 *) Make openssl verify return errors.
110 [Chris Palmer <palmer@google.com> and Ben Laurie]
112 *) New function ASN1_TIME_diff to calculate the difference between two
113 ASN1_TIME structures or one structure and the current time.
116 *) Update fips_test_suite to support multiple command line options. New
117 test to induce all self test errors in sequence and check expected
121 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
122 sign or verify all in one operation.
125 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
126 test programs and fips_test_suite. Includes functionality to parse
127 the minimal script output of fipsalgest.pl directly.
130 *) Add authorisation parameter to FIPS_module_mode_set().
133 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
136 *) Use separate DRBG fields for internal and external flags. New function
137 FIPS_drbg_health_check() to perform on demand health checking. Add
138 generation tests to fips_test_suite with reduced health check interval to
139 demonstrate periodic health checking. Add "nodh" option to
140 fips_test_suite to skip very slow DH test.
143 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
147 *) More extensive health check for DRBG checking many more failure modes.
148 New function FIPS_selftest_drbg_all() to handle every possible DRBG
149 combination: call this in fips_test_suite.
152 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
153 and POST to handle Dual EC cases.
156 *) Add support for canonical generation of DSA parameter 'g'. See
159 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
160 POST to handle HMAC cases.
163 *) Add functions FIPS_module_version() and FIPS_module_version_text()
164 to return numerical and string versions of the FIPS module number.
167 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
168 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
169 outside the validated module in the FIPS capable OpenSSL.
172 *) Minor change to DRBG entropy callback semantics. In some cases
173 there is no multiple of the block length between min_len and
174 max_len. Allow the callback to return more than max_len bytes
175 of entropy but discard any extra: it is the callback's responsibility
176 to ensure that the extra data discarded does not impact the
177 requested amount of entropy.
180 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
181 information in FIPS186-3, SP800-57 and SP800-131A.
184 *) CCM support via EVP. Interface is very similar to GCM case except we
185 must supply all data in one chunk (i.e. no update, final) and the
186 message length must be supplied if AAD is used. Add algorithm test
190 *) Initial version of POST overhaul. Add POST callback to allow the status
191 of POST to be monitored and/or failures induced. Modify fips_test_suite
192 to use callback. Always run all selftests even if one fails.
195 *) XTS support including algorithm test driver in the fips_gcmtest program.
196 Note: this does increase the maximum key length from 32 to 64 bytes but
197 there should be no binary compatibility issues as existing applications
198 will never use XTS mode.
201 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
202 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
203 performs algorithm blocking for unapproved PRNG types. Also do not
204 set PRNG type in FIPS_mode_set(): leave this to the application.
205 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
206 the standard OpenSSL PRNG: set additional data to a date time vector.
209 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
210 This shouldn't present any incompatibility problems because applications
211 shouldn't be using these directly and any that are will need to rethink
212 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
215 *) Extensive self tests and health checking required by SP800-90 DRBG.
216 Remove strength parameter from FIPS_drbg_instantiate and always
217 instantiate at maximum supported strength.
220 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
223 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
226 *) New function DH_compute_key_padded() to compute a DH key and pad with
227 leading zeroes if needed: this complies with SP800-56A et al.
230 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
231 anything, incomplete, subject to change and largely untested at present.
234 *) Modify fipscanisteronly build option to only build the necessary object
235 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
238 *) Add experimental option FIPSSYMS to give all symbols in
239 fipscanister.o and FIPS or fips prefix. This will avoid
240 conflicts with future versions of OpenSSL. Add perl script
241 util/fipsas.pl to preprocess assembly language source files
242 and rename any affected symbols.
245 *) Add selftest checks and algorithm block of non-fips algorithms in
246 FIPS mode. Remove DES2 from selftests.
249 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
250 return internal method without any ENGINE dependencies. Add new
251 tiny fips sign and verify functions.
254 *) New build option no-ec2m to disable characteristic 2 code.
257 *) New build option "fipscanisteronly". This only builds fipscanister.o
258 and (currently) associated fips utilities. Uses the file Makefile.fips
259 instead of Makefile.org as the prototype.
262 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
263 Update fips_gcmtest to use IV generator.
266 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
267 setting output buffer to NULL. The *Final function must be
268 called although it will not retrieve any additional data. The tag
269 can be set or retrieved with a ctrl. The IV length is by default 12
270 bytes (96 bits) but can be set to an alternative value. If the IV
271 length exceeds the maximum IV length (currently 16 bytes) it cannot be
275 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
276 underlying do_cipher function handles all cipher semantics itself
277 including padding and finalisation. This is useful if (for example)
278 an ENGINE cipher handles block padding itself. The behaviour of
279 do_cipher is subtly changed if this flag is set: the return value
280 is the number of characters written to the output buffer (zero is
281 no longer an error code) or a negative error code. Also if the
282 input buffer is NULL and length 0 finalisation should be performed.
285 *) If a candidate issuer certificate is already part of the constructed
286 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
289 *) Improve forward-security support: add functions
291 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
292 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
294 for use by SSL/TLS servers; the callback function will be called whenever a
295 new session is created, and gets to decide whether the session may be
296 cached to make it resumable (return 0) or not (return 1). (As by the
297 SSL/TLS protocol specifications, the session_id sent by the server will be
298 empty to indicate that the session is not resumable; also, the server will
299 not generate RFC 4507 (RFC 5077) session tickets.)
301 A simple reasonable callback implementation is to return is_forward_secure.
302 This parameter will be set to 1 or 0 depending on the ciphersuite selected
303 by the SSL/TLS server library, indicating whether it can provide forward
305 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
307 *) New -verify_name option in command line utilities to set verification
311 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
312 Add CMAC pkey methods.
315 *) Experimental renegotiation in s_server -www mode. If the client
316 browses /reneg connection is renegotiated. If /renegcert it is
317 renegotiated requesting a certificate.
320 *) Add an "external" session cache for debugging purposes to s_server. This
321 should help trace issues which normally are only apparent in deployed
322 multi-process servers.
325 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
326 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
327 BIO_set_cipher() and some obscure PEM functions were changed so they
328 can now return an error. The RAND changes required a change to the
329 RAND_METHOD structure.
332 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
333 a gcc attribute to warn if the result of a function is ignored. This
334 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
335 whose return value is often ignored.
338 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
340 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
341 (other platforms pending).
342 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
344 *) Add support for the SignedCertificateTimestampList certificate and
345 OCSP response extensions from RFC6962.
348 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
349 for corner cases. (Certain input points at infinity could lead to
350 bogus results, with non-infinity inputs mapped to infinity too.)
353 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
354 This covers AES, SHA256/512 and GHASH. "Initial" means that most
355 common cases are optimized and there still is room for further
356 improvements. Vector Permutation AES for Altivec is also added.
359 *) Add support for little-endian ppc64 Linux target.
360 [Marcelo Cerri (IBM)]
362 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
363 SHA1, SHA256 and GHASH. "Initial" means that most common cases
364 are optimized and there still is room for further improvements.
365 Both 32- and 64-bit modes are supported.
366 [Andy Polyakov, Ard Biesheuvel (Linaro)]
368 *) Improved ARMv7 NEON support.
371 *) Support for SPARC Architecture 2011 crypto extensions, first
372 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
373 SHA256/512, MD5, GHASH and modular exponentiation.
374 [Andy Polyakov, David Miller]
376 *) Accelerated modular exponentiation for Intel processors, a.k.a.
378 [Shay Gueron & Vlad Krasnov (Intel Corp)]
380 *) Support for new and upcoming Intel processors, including AVX2,
381 BMI and SHA ISA extensions. This includes additional "stitched"
382 implementations, AESNI-SHA256 and GCM, and multi-buffer support
385 This work was sponsored by Intel Corp.
388 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
389 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
390 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
393 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
394 this fixes a limiation in previous versions of OpenSSL.
397 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
398 MGF1 digest and OAEP label.
401 *) Add EVP support for key wrapping algorithms, to avoid problems with
402 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
403 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
404 algorithms and include tests cases.
407 *) Add functions to allocate and set the fields of an ECDSA_METHOD
409 [Douglas E. Engert, Steve Henson]
411 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
412 difference in days and seconds between two tm or ASN1_TIME structures.
415 *) Add -rev test option to s_server to just reverse order of characters
416 received by client and send back to server. Also prints an abbreviated
417 summary of the connection parameters.
420 *) New option -brief for s_client and s_server to print out a brief summary
421 of connection parameters.
424 *) Add callbacks for arbitrary TLS extensions.
425 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
427 *) New option -crl_download in several openssl utilities to download CRLs
428 from CRLDP extension in certificates.
431 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
434 *) New function X509_CRL_diff to generate a delta CRL from the difference
435 of two full CRLs. Add support to "crl" utility.
438 *) New functions to set lookup_crls function and to retrieve
439 X509_STORE from X509_STORE_CTX.
442 *) Print out deprecated issuer and subject unique ID fields in
446 *) Extend OCSP I/O functions so they can be used for simple general purpose
447 HTTP as well as OCSP. New wrapper function which can be used to download
448 CRLs using the OCSP API.
451 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
454 *) SSL_CONF* functions. These provide a common framework for application
455 configuration using configuration files or command lines.
458 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
459 message callback and prints the results. Needs compile time option
460 "enable-ssl-trace". New options to s_client and s_server to enable
464 *) New ctrl and macro to retrieve supported points extensions.
465 Print out extension in s_server and s_client.
468 *) New functions to retrieve certificate signature and signature
472 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
476 *) New Suite B modes for TLS code. These use and enforce the requirements
477 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
478 only use Suite B curves. The Suite B modes can be set by using the
479 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
482 *) New chain verification flags for Suite B levels of security. Check
483 algorithms are acceptable when flags are set in X509_verify_cert.
486 *) Make tls1_check_chain return a set of flags indicating checks passed
487 by a certificate chain. Add additional tests to handle client
488 certificates: checks for matching certificate type and issuer name
492 *) If an attempt is made to use a signature algorithm not in the peer
493 preference list abort the handshake. If client has no suitable
494 signature algorithms in response to a certificate request do not
498 *) If server EC tmp key is not in client preference list abort handshake.
501 *) Add support for certificate stores in CERT structure. This makes it
502 possible to have different stores per SSL structure or one store in
503 the parent SSL_CTX. Include distint stores for certificate chain
504 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
505 to build and store a certificate chain in CERT structure: returing
506 an error if the chain cannot be built: this will allow applications
507 to test if a chain is correctly configured.
509 Note: if the CERT based stores are not set then the parent SSL_CTX
510 store is used to retain compatibility with existing behaviour.
514 *) New function ssl_set_client_disabled to set a ciphersuite disabled
515 mask based on the current session, check mask when sending client
516 hello and checking the requested ciphersuite.
519 *) New ctrls to retrieve and set certificate types in a certificate
520 request message. Print out received values in s_client. If certificate
521 types is not set with custom values set sensible values based on
522 supported signature algorithms.
525 *) Support for distinct client and server supported signature algorithms.
528 *) Add certificate callback. If set this is called whenever a certificate
529 is required by client or server. An application can decide which
530 certificate chain to present based on arbitrary criteria: for example
531 supported signature algorithms. Add very simple example to s_server.
532 This fixes many of the problems and restrictions of the existing client
533 certificate callback: for example you can now clear an existing
534 certificate and specify the whole chain.
537 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
538 the certificate can be used for (if anything). Set valid_flags field
539 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
540 to have similar checks in it.
542 Add new "cert_flags" field to CERT structure and include a "strict mode".
543 This enforces some TLS certificate requirements (such as only permitting
544 certificate signature algorithms contained in the supported algorithms
545 extension) which some implementations ignore: this option should be used
546 with caution as it could cause interoperability issues.
549 *) Update and tidy signature algorithm extension processing. Work out
550 shared signature algorithms based on preferences and peer algorithms
551 and print them out in s_client and s_server. Abort handshake if no
552 shared signature algorithms.
555 *) Add new functions to allow customised supported signature algorithms
556 for SSL and SSL_CTX structures. Add options to s_client and s_server
560 *) New function SSL_certs_clear() to delete all references to certificates
561 from an SSL structure. Before this once a certificate had been added
562 it couldn't be removed.
565 *) Integrate hostname, email address and IP address checking with certificate
566 verification. New verify options supporting checking in opensl utility.
569 *) Fixes and wildcard matching support to hostname and email checking
570 functions. Add manual page.
571 [Florian Weimer (Red Hat Product Security Team)]
573 *) New functions to check a hostname email or IP address against a
574 certificate. Add options x509 utility to print results of checks against
578 *) Fix OCSP checking.
579 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
581 *) Initial experimental support for explicitly trusted non-root CAs.
582 OpenSSL still tries to build a complete chain to a root but if an
583 intermediate CA has a trust setting included that is used. The first
584 setting is used: whether to trust (e.g., -addtrust option to the x509
588 *) Add -trusted_first option which attempts to find certificates in the
589 trusted store even if an untrusted chain is also supplied.
592 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
593 platform support for Linux and Android.
596 *) Support for linux-x32, ILP32 environment in x86_64 framework.
599 *) Experimental multi-implementation support for FIPS capable OpenSSL.
600 When in FIPS mode the approved implementations are used as normal,
601 when not in FIPS mode the internal unapproved versions are used instead.
602 This means that the FIPS capable OpenSSL isn't forced to use the
603 (often lower performance) FIPS implementations outside FIPS mode.
606 *) Transparently support X9.42 DH parameters when calling
607 PEM_read_bio_DHparameters. This means existing applications can handle
608 the new parameter format automatically.
611 *) Initial experimental support for X9.42 DH parameter format: mainly
612 to support use of 'q' parameter for RFC5114 parameters.
615 *) Add DH parameters from RFC5114 including test data to dhtest.
618 *) Support for automatic EC temporary key parameter selection. If enabled
619 the most preferred EC parameters are automatically used instead of
620 hardcoded fixed parameters. Now a server just has to call:
621 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
622 support ECDH and use the most appropriate parameters.
625 *) Enhance and tidy EC curve and point format TLS extension code. Use
626 static structures instead of allocation if default values are used.
627 New ctrls to set curves we wish to support and to retrieve shared curves.
628 Print out shared curves in s_server. New options to s_server and s_client
629 to set list of supported curves.
632 *) New ctrls to retrieve supported signature algorithms and
633 supported curve values as an array of NIDs. Extend openssl utility
634 to print out received values.
637 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
638 between NIDs and the more common NIST names such as "P-256". Enhance
639 ecparam utility and ECC method to recognise the NIST names for curves.
642 *) Enhance SSL/TLS certificate chain handling to support different
643 chains for each certificate instead of one chain in the parent SSL_CTX.
646 *) Support for fixed DH ciphersuite client authentication: where both
647 server and client use DH certificates with common parameters.
650 *) Support for fixed DH ciphersuites: those requiring DH server
654 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
656 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
657 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
658 X509_CINF_get_signature were reverted post internal team review.
660 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
662 *) Abort handshake if server key exchange message is omitted for ephemeral
665 Thanks to Karthikeyan Bhargavan for reporting this issue.
669 *) Remove non-export ephemeral RSA code on client and server. This code
670 violated the TLS standard by allowing the use of temporary RSA keys in
671 non-export ciphersuites and could be used by a server to effectively
672 downgrade the RSA key length used to a value smaller than the server
673 certificate. Thanks for Karthikeyan Bhargavan for reporting this issue.
677 *) Ensure that the session ID context of an SSL is updated when its
678 SSL_CTX is updated via SSL_set_SSL_CTX.
680 The session ID context is typically set from the parent SSL_CTX,
681 and can vary with the CTX.
684 *) Fix various certificate fingerprint issues.
686 By using non-DER or invalid encodings outside the signed portion of a
687 certificate the fingerprint can be changed without breaking the signature.
688 Although no details of the signed portion of the certificate can be changed
689 this can cause problems with some applications: e.g. those using the
690 certificate fingerprint for blacklists.
692 1. Reject signatures with non zero unused bits.
694 If the BIT STRING containing the signature has non zero unused bits reject
695 the signature. All current signature algorithms require zero unused bits.
697 2. Check certificate algorithm consistency.
699 Check the AlgorithmIdentifier inside TBS matches the one in the
700 certificate signature. NB: this will result in signature failure
701 errors for some broken certificates.
703 Thanks to Konrad Kraszewski from Google for reporting this issue.
705 3. Check DSA/ECDSA signatures use DER.
707 Reencode DSA/ECDSA signatures and compare with the original received
708 signature. Return an error if there is a mismatch.
710 This will reject various cases including garbage after signature
711 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
712 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
713 (negative or with leading zeroes).
715 Further analysis was conducted and fixes were developed by Stephen Henson
716 of the OpenSSL core team.
721 *) Do not resume sessions on the server if the negotiated protocol
722 version does not match the session's version. Resuming with a different
723 version, while not strictly forbidden by the RFC, is of questionable
724 sanity and breaks all known clients.
725 [David Benjamin, Emilia Käsper]
727 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
728 early CCS messages during renegotiation. (Note that because
729 renegotiation is encrypted, this early CCS was not exploitable.)
732 *) Tighten client-side session ticket handling during renegotiation:
733 ensure that the client only accepts a session ticket if the server sends
734 the extension anew in the ServerHello. Previously, a TLS client would
735 reuse the old extension state and thus accept a session ticket if one was
736 announced in the initial ServerHello.
738 Similarly, ensure that the client requires a session ticket if one
739 was advertised in the ServerHello. Previously, a TLS client would
740 ignore a missing NewSessionTicket message.
743 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
747 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
748 sends a carefully crafted handshake message, to cause OpenSSL to fail
749 to free up to 64k of memory causing a memory leak. This could be
750 exploited in a Denial Of Service attack. This issue affects OpenSSL
751 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
752 whether SRTP is used or configured. Implementations of OpenSSL that
753 have been compiled with OPENSSL_NO_SRTP defined are not affected.
755 The fix was developed by the OpenSSL team.
759 *) Session Ticket Memory Leak.
761 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
762 integrity of that ticket is first verified. In the event of a session
763 ticket integrity check failing, OpenSSL will fail to free memory
764 causing a memory leak. By sending a large number of invalid session
765 tickets an attacker could exploit this issue in a Denial Of Service
770 *) Build option no-ssl3 is incomplete.
772 When OpenSSL is configured with "no-ssl3" as a build option, servers
773 could accept and complete a SSL 3.0 handshake, and clients could be
774 configured to send them.
776 [Akamai and the OpenSSL team]
778 *) Add support for TLS_FALLBACK_SCSV.
779 Client applications doing fallback retries should call
780 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
782 [Adam Langley, Bodo Moeller]
784 *) Add additional DigestInfo checks.
786 Reencode DigestInto in DER and check against the original when
787 verifying RSA signature: this will reject any improperly encoded
788 DigestInfo structures.
790 Note: this is a precautionary measure and no attacks are currently known.
794 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
796 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
797 SRP code can be overrun an internal buffer. Add sanity check that
798 g, A, B < N to SRP code.
800 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
801 Group for discovering this issue.
805 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
806 TLS 1.0 instead of higher protocol versions when the ClientHello message
807 is badly fragmented. This allows a man-in-the-middle attacker to force a
808 downgrade to TLS 1.0 even if both the server and the client support a
809 higher protocol version, by modifying the client's TLS records.
811 Thanks to David Benjamin and Adam Langley (Google) for discovering and
812 researching this issue.
816 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
817 to a denial of service attack. A malicious server can crash the client
818 with a null pointer dereference (read) by specifying an anonymous (EC)DH
819 ciphersuite and sending carefully crafted handshake messages.
821 Thanks to Felix Gröbert (Google) for discovering and researching this
826 *) By sending carefully crafted DTLS packets an attacker could cause openssl
827 to leak memory. This can be exploited through a Denial of Service attack.
828 Thanks to Adam Langley for discovering and researching this issue.
832 *) An attacker can force openssl to consume large amounts of memory whilst
833 processing DTLS handshake messages. This can be exploited through a
834 Denial of Service attack.
835 Thanks to Adam Langley for discovering and researching this issue.
839 *) An attacker can force an error condition which causes openssl to crash
840 whilst processing DTLS packets due to memory being freed twice. This
841 can be exploited through a Denial of Service attack.
842 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
847 *) If a multithreaded client connects to a malicious server using a resumed
848 session and the server sends an ec point format extension it could write
849 up to 255 bytes to freed memory.
851 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
856 *) A malicious server can crash an OpenSSL client with a null pointer
857 dereference (read) by specifying an SRP ciphersuite even though it was not
858 properly negotiated with the client. This can be exploited through a
859 Denial of Service attack.
861 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
862 discovering and researching this issue.
866 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
867 X509_name_oneline, X509_name_print_ex et al. to leak some information
868 from the stack. Applications may be affected if they echo pretty printing
869 output to the attacker.
871 Thanks to Ivan Fratric (Google) for discovering this issue.
873 [Emilia Käsper, and Steve Henson]
875 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
876 for corner cases. (Certain input points at infinity could lead to
877 bogus results, with non-infinity inputs mapped to infinity too.)
880 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
882 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
883 handshake can force the use of weak keying material in OpenSSL
884 SSL/TLS clients and servers.
886 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
887 researching this issue. (CVE-2014-0224)
888 [KIKUCHI Masashi, Steve Henson]
890 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
891 OpenSSL DTLS client the code can be made to recurse eventually crashing
894 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
896 [Imre Rad, Steve Henson]
898 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
899 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
900 client or server. This is potentially exploitable to run arbitrary
901 code on a vulnerable client or server.
903 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
904 [Jüri Aedla, Steve Henson]
906 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
907 are subject to a denial of service attack.
909 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
910 this issue. (CVE-2014-3470)
911 [Felix Gröbert, Ivan Fratric, Steve Henson]
913 *) Harmonize version and its documentation. -f flag is used to display
915 [mancha <mancha1@zoho.com>]
917 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
919 [mancha <mancha1@zoho.com>]
921 *) Fix some double frees. These are not thought to be exploitable.
922 [mancha <mancha1@zoho.com>]
924 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
926 *) A missing bounds check in the handling of the TLS heartbeat extension
927 can be used to reveal up to 64k of memory to a connected client or
930 Thanks for Neel Mehta of Google Security for discovering this bug and to
931 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
932 preparing the fix (CVE-2014-0160)
933 [Adam Langley, Bodo Moeller]
935 *) Fix for the attack described in the paper "Recovering OpenSSL
936 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
937 by Yuval Yarom and Naomi Benger. Details can be obtained from:
938 http://eprint.iacr.org/2014/140
940 Thanks to Yuval Yarom and Naomi Benger for discovering this
941 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
942 [Yuval Yarom and Naomi Benger]
944 *) TLS pad extension: draft-agl-tls-padding-03
946 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
947 TLS client Hello record length value would otherwise be > 255 and
948 less that 512 pad with a dummy extension containing zeroes so it
949 is at least 512 bytes long.
951 [Adam Langley, Steve Henson]
953 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
955 *) Fix for TLS record tampering bug. A carefully crafted invalid
956 handshake could crash OpenSSL with a NULL pointer exception.
957 Thanks to Anton Johansson for reporting this issues.
960 *) Keep original DTLS digest and encryption contexts in retransmission
961 structures so we can use the previous session parameters if they need
962 to be resent. (CVE-2013-6450)
965 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
966 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
967 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
968 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
969 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
970 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
971 [Rob Stradling, Adam Langley]
973 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
975 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
976 supporting platforms or when small records were transferred.
977 [Andy Polyakov, Steve Henson]
979 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
981 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
983 This addresses the flaw in CBC record processing discovered by
984 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
985 at: http://www.isg.rhul.ac.uk/tls/
987 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
988 Security Group at Royal Holloway, University of London
989 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
990 Emilia Käsper for the initial patch.
992 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
994 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
995 ciphersuites which can be exploited in a denial of service attack.
996 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
997 and detecting this bug and to Wolfgang Ettlinger
998 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1002 *) Return an error when checking OCSP signatures when key is NULL.
1003 This fixes a DoS attack. (CVE-2013-0166)
1006 *) Make openssl verify return errors.
1007 [Chris Palmer <palmer@google.com> and Ben Laurie]
1009 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1010 the right response is stapled. Also change SSL_get_certificate()
1011 so it returns the certificate actually sent.
1012 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1013 [Rob Stradling <rob.stradling@comodo.com>]
1015 *) Fix possible deadlock when decoding public keys.
1018 *) Don't use TLS 1.0 record version number in initial client hello
1022 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1024 *) Sanity check record length before skipping explicit IV in TLS
1025 1.2, 1.1 and DTLS to fix DoS attack.
1027 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1028 fuzzing as a service testing platform.
1032 *) Initialise tkeylen properly when encrypting CMS messages.
1033 Thanks to Solar Designer of Openwall for reporting this issue.
1036 *) In FIPS mode don't try to use composite ciphers as they are not
1040 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1042 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1043 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1044 mean any application compiled against OpenSSL 1.0.0 headers setting
1045 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1046 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1047 0x10000000L Any application which was previously compiled against
1048 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1049 will need to be recompiled as a result. Letting be results in
1050 inability to disable specifically TLS 1.1 and in client context,
1051 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1054 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1055 disable just protocol X, but all protocols above X *if* there are
1056 protocols *below* X still enabled. In more practical terms it means
1057 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1058 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1059 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1063 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1065 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1066 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1067 in CRYPTO_realloc_clean.
1069 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1070 issue and to Adam Langley <agl@chromium.org> for fixing it.
1072 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1074 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1077 *) Workarounds for some broken servers that "hang" if a client hello
1078 record length exceeds 255 bytes.
1080 1. Do not use record version number > TLS 1.0 in initial client
1081 hello: some (but not all) hanging servers will now work.
1082 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1083 the number of ciphers sent in the client hello. This should be
1084 set to an even number, such as 50, for example by passing:
1085 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1086 Most broken servers should now work.
1087 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1088 TLS 1.2 client support entirely.
1091 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1094 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1096 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1097 STRING form instead of a DigestInfo.
1100 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1101 and the RSA_sign/RSA_verify functions. This was made more apparent when
1102 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1103 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1104 the correct format in RSA_verify so both forms transparently work.
1107 *) Some servers which support TLS 1.0 can choke if we initially indicate
1108 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1109 encrypted premaster secret. As a workaround use the maximum permitted
1110 client version in client hello, this should keep such servers happy
1111 and still work with previous versions of OpenSSL.
1114 *) Add support for TLS/DTLS heartbeats.
1115 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1117 *) Add support for SCTP.
1118 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1120 *) Improved PRNG seeding for VOS.
1121 [Paul Green <Paul.Green@stratus.com>]
1123 *) Extensive assembler packs updates, most notably:
1125 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1126 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1127 - x86_64: bit-sliced AES implementation;
1128 - ARM: NEON support, contemporary platforms optimizations;
1129 - s390x: z196 support;
1130 - *: GHASH and GF(2^m) multiplication implementations;
1134 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1135 (removal of unnecessary code)
1136 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1138 *) Add TLS key material exporter from RFC 5705.
1141 *) Add DTLS-SRTP negotiation from RFC 5764.
1144 *) Add Next Protocol Negotiation,
1145 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1146 disabled with a no-npn flag to config or Configure. Code donated
1148 [Adam Langley <agl@google.com> and Ben Laurie]
1150 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1151 NIST-P256, NIST-P521, with constant-time single point multiplication on
1152 typical inputs. Compiler support for the nonstandard type __uint128_t is
1153 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1154 Code made available under Apache License version 2.0.
1156 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1157 line to include this in your build of OpenSSL, and run "make depend" (or
1158 "make update"). This enables the following EC_METHODs:
1160 EC_GFp_nistp224_method()
1161 EC_GFp_nistp256_method()
1162 EC_GFp_nistp521_method()
1164 EC_GROUP_new_by_curve_name() will automatically use these (while
1165 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1167 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1169 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1170 all platforms. Move ssize_t definition from e_os.h to the public
1171 header file e_os2.h as it now appears in public header file cms.h
1174 *) New -sigopt option to the ca, req and x509 utilities. Additional
1175 signature parameters can be passed using this option and in
1179 *) Add RSA PSS signing function. This will generate and set the
1180 appropriate AlgorithmIdentifiers for PSS based on those in the
1181 corresponding EVP_MD_CTX structure. No application support yet.
1184 *) Support for companion algorithm specific ASN1 signing routines.
1185 New function ASN1_item_sign_ctx() signs a pre-initialised
1186 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1187 the appropriate parameters.
1190 *) Add new algorithm specific ASN1 verification initialisation function
1191 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1192 handling will be the same no matter what EVP_PKEY_METHOD is used.
1193 Add a PSS handler to support verification of PSS signatures: checked
1194 against a number of sample certificates.
1197 *) Add signature printing for PSS. Add PSS OIDs.
1198 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1200 *) Add algorithm specific signature printing. An individual ASN1 method
1201 can now print out signatures instead of the standard hex dump.
1203 More complex signatures (e.g. PSS) can print out more meaningful
1204 information. Include DSA version that prints out the signature
1208 *) Password based recipient info support for CMS library: implementing
1212 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1213 neatly separates the code into cipher and PBE sections and is required
1214 for some algorithms that split PBES2 into separate pieces (such as
1215 password based CMS).
1218 *) Session-handling fixes:
1219 - Fix handling of connections that are resuming with a session ID,
1220 but also support Session Tickets.
1221 - Fix a bug that suppressed issuing of a new ticket if the client
1222 presented a ticket with an expired session.
1223 - Try to set the ticket lifetime hint to something reasonable.
1224 - Make tickets shorter by excluding irrelevant information.
1225 - On the client side, don't ignore renewed tickets.
1226 [Adam Langley, Bodo Moeller (Google)]
1228 *) Fix PSK session representation.
1231 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1233 This work was sponsored by Intel.
1236 *) Add GCM support to TLS library. Some custom code is needed to split
1237 the IV between the fixed (from PRF) and explicit (from TLS record)
1238 portions. This adds all GCM ciphersuites supported by RFC5288 and
1239 RFC5289. Generalise some AES* cipherstrings to include GCM and
1240 add a special AESGCM string for GCM only.
1243 *) Expand range of ctrls for AES GCM. Permit setting invocation
1244 field on decrypt and retrieval of invocation field only on encrypt.
1247 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1248 As required by RFC5289 these ciphersuites cannot be used if for
1249 versions of TLS earlier than 1.2.
1252 *) For FIPS capable OpenSSL interpret a NULL default public key method
1253 as unset and return the appropriate default but do *not* set the default.
1254 This means we can return the appropriate method in applications that
1255 switch between FIPS and non-FIPS modes.
1258 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1259 ENGINE is used then we cannot handle that in the FIPS module so we
1260 keep original code iff non-FIPS operations are allowed.
1263 *) Add -attime option to openssl utilities.
1264 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1266 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1269 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1270 FIPS EC methods unconditionally for now.
1273 *) New build option no-ec2m to disable characteristic 2 code.
1276 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1277 all cases can be covered as some introduce binary incompatibilities.
1280 *) Redirect RSA operations to FIPS module including keygen,
1281 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1284 *) Add similar low level API blocking to ciphers.
1287 *) Low level digest APIs are not approved in FIPS mode: any attempt
1288 to use these will cause a fatal error. Applications that *really* want
1289 to use them can use the private_* version instead.
1292 *) Redirect cipher operations to FIPS module for FIPS builds.
1295 *) Redirect digest operations to FIPS module for FIPS builds.
1298 *) Update build system to add "fips" flag which will link in fipscanister.o
1299 for static and shared library builds embedding a signature if needed.
1302 *) Output TLS supported curves in preference order instead of numerical
1303 order. This is currently hardcoded for the highest order curves first.
1304 This should be configurable so applications can judge speed vs strength.
1307 *) Add TLS v1.2 server support for client authentication.
1310 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1314 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1315 FIPS modules versions.
1318 *) Add TLS v1.2 client side support for client authentication. Keep cache
1319 of handshake records longer as we don't know the hash algorithm to use
1320 until after the certificate request message is received.
1323 *) Initial TLS v1.2 client support. Add a default signature algorithms
1324 extension including all the algorithms we support. Parse new signature
1325 format in client key exchange. Relax some ECC signing restrictions for
1326 TLS v1.2 as indicated in RFC5246.
1329 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1330 to new signature format when needed using client digest preference.
1331 All server ciphersuites should now work correctly in TLS v1.2. No client
1332 support yet and no support for client certificates.
1335 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1336 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1337 ciphersuites. At present only RSA key exchange ciphersuites work with
1338 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1339 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1340 and version checking.
1343 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1344 with this defined it will not be affected by any changes to ssl internal
1345 structures. Add several utility functions to allow openssl application
1346 to work with OPENSSL_NO_SSL_INTERN defined.
1350 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1352 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1355 *) Permit abbreviated handshakes when renegotiating using the function
1356 SSL_renegotiate_abbreviated().
1357 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1359 *) Add call to ENGINE_register_all_complete() to
1360 ENGINE_load_builtin_engines(), so some implementations get used
1361 automatically instead of needing explicit application support.
1364 *) Add support for TLS key exporter as described in RFC5705.
1365 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1367 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1368 a few changes are required:
1370 Add SSL_OP_NO_TLSv1_1 flag.
1371 Add TLSv1_1 methods.
1372 Update version checking logic to handle version 1.1.
1373 Add explicit IV handling (ported from DTLS code).
1374 Add command line options to s_client/s_server.
1377 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1379 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1381 This addresses the flaw in CBC record processing discovered by
1382 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1383 at: http://www.isg.rhul.ac.uk/tls/
1385 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1386 Security Group at Royal Holloway, University of London
1387 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1388 Emilia Käsper for the initial patch.
1390 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1392 *) Return an error when checking OCSP signatures when key is NULL.
1393 This fixes a DoS attack. (CVE-2013-0166)
1396 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1397 the right response is stapled. Also change SSL_get_certificate()
1398 so it returns the certificate actually sent.
1399 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1400 (This is a backport)
1401 [Rob Stradling <rob.stradling@comodo.com>]
1403 *) Fix possible deadlock when decoding public keys.
1406 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1408 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1411 *) Sanity check record length before skipping explicit IV in DTLS
1414 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1415 fuzzing as a service testing platform.
1419 *) Initialise tkeylen properly when encrypting CMS messages.
1420 Thanks to Solar Designer of Openwall for reporting this issue.
1423 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1425 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1426 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1427 in CRYPTO_realloc_clean.
1429 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1430 issue and to Adam Langley <agl@chromium.org> for fixing it.
1432 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1434 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1436 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1437 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1438 content decryption and always return the same error. Note: this attack
1439 needs on average 2^20 messages so it only affects automated senders. The
1440 old behaviour can be reenabled in the CMS code by setting the
1441 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1442 an MMA defence is not necessary.
1443 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1444 this issue. (CVE-2012-0884)
1447 *) Fix CVE-2011-4619: make sure we really are receiving a
1448 client hello before rejecting multiple SGC restarts. Thanks to
1449 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1452 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1454 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1455 Thanks to Antonio Martin, Enterprise Secure Access Research and
1456 Development, Cisco Systems, Inc. for discovering this bug and
1457 preparing a fix. (CVE-2012-0050)
1460 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1462 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1463 of the Vaudenay padding oracle attack on CBC mode encryption
1464 which enables an efficient plaintext recovery attack against
1465 the OpenSSL implementation of DTLS. Their attack exploits timing
1466 differences arising during decryption processing. A research
1467 paper describing this attack can be found at:
1468 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1469 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1470 Security Group at Royal Holloway, University of London
1471 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1472 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1473 for preparing the fix. (CVE-2011-4108)
1474 [Robin Seggelmann, Michael Tuexen]
1476 *) Clear bytes used for block padding of SSL 3.0 records.
1478 [Adam Langley (Google)]
1480 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1481 Kadianakis <desnacked@gmail.com> for discovering this issue and
1482 Adam Langley for preparing the fix. (CVE-2011-4619)
1483 [Adam Langley (Google)]
1485 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1486 [Andrey Kulikov <amdeich@gmail.com>]
1488 *) Prevent malformed RFC3779 data triggering an assertion failure.
1489 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1490 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1491 [Rob Austein <sra@hactrn.net>]
1493 *) Improved PRNG seeding for VOS.
1494 [Paul Green <Paul.Green@stratus.com>]
1496 *) Fix ssl_ciph.c set-up race.
1497 [Adam Langley (Google)]
1499 *) Fix spurious failures in ecdsatest.c.
1500 [Emilia Käsper (Google)]
1502 *) Fix the BIO_f_buffer() implementation (which was mixing different
1503 interpretations of the '..._len' fields).
1504 [Adam Langley (Google)]
1506 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1507 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1508 threads won't reuse the same blinding coefficients.
1510 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1511 lock to call BN_BLINDING_invert_ex, and avoids one use of
1512 BN_BLINDING_update for each BN_BLINDING structure (previously,
1513 the last update always remained unused).
1514 [Emilia Käsper (Google)]
1516 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1517 [Bob Buckholz (Google)]
1519 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1521 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1522 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1523 [Kaspar Brand <ossl@velox.ch>]
1525 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1526 for multi-threaded use of ECDH. (CVE-2011-3210)
1527 [Adam Langley (Google)]
1529 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1532 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1533 signature public key algorithm by using OID xref utilities instead.
1534 Before this you could only use some ECC ciphersuites with SHA1 only.
1537 *) Add protection against ECDSA timing attacks as mentioned in the paper
1538 by Billy Bob Brumley and Nicola Tuveri, see:
1540 http://eprint.iacr.org/2011/232.pdf
1542 [Billy Bob Brumley and Nicola Tuveri]
1544 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1546 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1547 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1549 *) Fix bug in string printing code: if *any* escaping is enabled we must
1550 escape the escape character (backslash) or the resulting string is
1554 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1556 *) Disable code workaround for ancient and obsolete Netscape browsers
1557 and servers: an attacker can use it in a ciphersuite downgrade attack.
1558 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1561 *) Fixed J-PAKE implementation error, originally discovered by
1562 Sebastien Martini, further info and confirmation from Stefan
1563 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1566 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1568 *) Fix extension code to avoid race conditions which can result in a buffer
1569 overrun vulnerability: resumed sessions must not be modified as they can
1570 be shared by multiple threads. CVE-2010-3864
1573 *) Fix WIN32 build system to correctly link an ENGINE directory into
1577 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1579 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1581 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1583 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1585 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1586 context. The operation can be customised via the ctrl mechanism in
1587 case ENGINEs want to include additional functionality.
1590 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1593 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1594 output hashes compatible with older versions of OpenSSL.
1595 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1597 *) Fix compression algorithm handling: if resuming a session use the
1598 compression algorithm of the resumed session instead of determining
1599 it from client hello again. Don't allow server to change algorithm.
1602 *) Add load_crls() function to apps tidying load_certs() too. Add option
1603 to verify utility to allow additional CRLs to be included.
1606 *) Update OCSP request code to permit adding custom headers to the request:
1607 some responders need this.
1610 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1612 [Julia Lawall <julia@diku.dk>]
1614 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1615 needlessly dereferenced structures, used obsolete functions and
1616 didn't handle all updated verify codes correctly.
1619 *) Disable MD2 in the default configuration.
1622 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1623 indicate the initial BIO being pushed or popped. This makes it possible
1624 to determine whether the BIO is the one explicitly called or as a result
1625 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1626 it handles reference counts correctly and doesn't zero out the I/O bio
1627 when it is not being explicitly popped. WARNING: applications which
1628 included workarounds for the old buggy behaviour will need to be modified
1629 or they could free up already freed BIOs.
1632 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1633 renaming to all platforms (within the 0.9.8 branch, this was
1634 done conditionally on Netware platforms to avoid a name clash).
1635 [Guenter <lists@gknw.net>]
1637 *) Add ECDHE and PSK support to DTLS.
1638 [Michael Tuexen <tuexen@fh-muenster.de>]
1640 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1644 *) Add "missing" function EVP_MD_flags() (without this the only way to
1645 retrieve a digest flags is by accessing the structure directly. Update
1646 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1647 or cipher is registered as in the "from" argument. Print out all
1648 registered digests in the dgst usage message instead of manually
1649 attempting to work them out.
1652 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1653 this allows the use of compression and extensions. Change default cipher
1654 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1655 by default unless an application cipher string requests it.
1658 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1659 key ids to find matching certificates and keys but some PKCS#12 files
1660 don't follow the (somewhat unwritten) rules and this strategy fails.
1661 Now just gather all certificates together and the first private key
1662 then look for the first certificate that matches the key.
1665 *) Support use of registered digest and cipher names for dgst and cipher
1666 commands instead of having to add each one as a special case. So now
1673 openssl dgst -sha256 foo
1675 and this works for ENGINE based algorithms too.
1679 *) Update Gost ENGINE to support parameter files.
1680 [Victor B. Wagner <vitus@cryptocom.ru>]
1682 *) Support GeneralizedTime in ca utility.
1683 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1685 *) Enhance the hash format used for certificate directory links. The new
1686 form uses the canonical encoding (meaning equivalent names will work
1687 even if they aren't identical) and uses SHA1 instead of MD5. This form
1688 is incompatible with the older format and as a result c_rehash should
1689 be used to rebuild symbolic links.
1692 *) Make PKCS#8 the default write format for private keys, replacing the
1693 traditional format. This form is standardised, more secure and doesn't
1694 include an implicit MD5 dependency.
1697 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1698 committed to OpenSSL should pass this lot as a minimum.
1701 *) Add session ticket override functionality for use by EAP-FAST.
1702 [Jouni Malinen <j@w1.fi>]
1704 *) Modify HMAC functions to return a value. Since these can be implemented
1705 in an ENGINE errors can occur.
1708 *) Type-checked OBJ_bsearch_ex.
1711 *) Type-checked OBJ_bsearch. Also some constification necessitated
1712 by type-checking. Still to come: TXT_DB, bsearch(?),
1713 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1717 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1718 seconds to a tm structure directly, instead of going through OS
1719 specific date routines. This avoids any issues with OS routines such
1720 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1721 and X509_time_adj_ex() to cover the extended range. The existing
1722 X509_time_adj() is still usable and will no longer have any date issues.
1725 *) Delta CRL support. New use deltas option which will attempt to locate
1726 and search any appropriate delta CRLs available.
1728 This work was sponsored by Google.
1731 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1732 code and add additional score elements. Validate alternate CRL paths
1733 as part of the CRL checking and indicate a new error "CRL path validation
1734 error" in this case. Applications wanting additional details can use
1735 the verify callback and check the new "parent" field. If this is not
1736 NULL CRL path validation is taking place. Existing applications wont
1737 see this because it requires extended CRL support which is off by
1740 This work was sponsored by Google.
1743 *) Support for freshest CRL extension.
1745 This work was sponsored by Google.
1748 *) Initial indirect CRL support. Currently only supported in the CRLs
1749 passed directly and not via lookup. Process certificate issuer
1750 CRL entry extension and lookup CRL entries by bother issuer name
1751 and serial number. Check and process CRL issuer entry in IDP extension.
1753 This work was sponsored by Google.
1756 *) Add support for distinct certificate and CRL paths. The CRL issuer
1757 certificate is validated separately in this case. Only enabled if
1758 an extended CRL support flag is set: this flag will enable additional
1759 CRL functionality in future.
1761 This work was sponsored by Google.
1764 *) Add support for policy mappings extension.
1766 This work was sponsored by Google.
1769 *) Fixes to pathlength constraint, self issued certificate handling,
1770 policy processing to align with RFC3280 and PKITS tests.
1772 This work was sponsored by Google.
1775 *) Support for name constraints certificate extension. DN, email, DNS
1776 and URI types are currently supported.
1778 This work was sponsored by Google.
1781 *) To cater for systems that provide a pointer-based thread ID rather
1782 than numeric, deprecate the current numeric thread ID mechanism and
1783 replace it with a structure and associated callback type. This
1784 mechanism allows a numeric "hash" to be extracted from a thread ID in
1785 either case, and on platforms where pointers are larger than 'long',
1786 mixing is done to help ensure the numeric 'hash' is usable even if it
1787 can't be guaranteed unique. The default mechanism is to use "&errno"
1788 as a pointer-based thread ID to distinguish between threads.
1790 Applications that want to provide their own thread IDs should now use
1791 CRYPTO_THREADID_set_callback() to register a callback that will call
1792 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1794 Note that ERR_remove_state() is now deprecated, because it is tied
1795 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1796 to free the current thread's error state should be replaced by
1797 ERR_remove_thread_state(NULL).
1799 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1800 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1801 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1802 application was previously providing a numeric thread callback that
1803 was inappropriate for distinguishing threads, then uniqueness might
1804 have been obtained with &errno that happened immediately in the
1805 intermediate development versions of OpenSSL; this is no longer the
1806 case, the numeric thread callback will now override the automatic use
1808 [Geoff Thorpe, with help from Bodo Moeller]
1810 *) Initial support for different CRL issuing certificates. This covers a
1811 simple case where the self issued certificates in the chain exist and
1812 the real CRL issuer is higher in the existing chain.
1814 This work was sponsored by Google.
1817 *) Removed effectively defunct crypto/store from the build.
1820 *) Revamp of STACK to provide stronger type-checking. Still to come:
1821 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1822 ASN1_STRING, CONF_VALUE.
1825 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1826 RAM on SSL connections. This option can save about 34k per idle SSL.
1829 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1830 STACK, TXT_DB, bsearch, qsort.
1833 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1834 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1835 support for data, signedData, compressedData, digestedData and
1836 encryptedData, envelopedData types included. Scripts to check against
1837 RFC4134 examples draft and interop and consistency checks of many
1838 content types and variants.
1841 *) Add options to enc utility to support use of zlib compression BIO.
1844 *) Extend mk1mf to support importing of options and assembly language
1845 files from Configure script, currently only included in VC-WIN32.
1846 The assembly language rules can now optionally generate the source
1847 files from the associated perl scripts.
1850 *) Implement remaining functionality needed to support GOST ciphersuites.
1851 Interop testing has been performed using CryptoPro implementations.
1852 [Victor B. Wagner <vitus@cryptocom.ru>]
1854 *) s390x assembler pack.
1857 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1861 *) Implement Opaque PRF Input TLS extension as specified in
1862 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1863 official specification yet and no extension type assignment by
1864 IANA exists, this extension (for now) will have to be explicitly
1865 enabled when building OpenSSL by providing the extension number
1866 to use. For example, specify an option
1868 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1870 to the "config" or "Configure" script to enable the extension,
1871 assuming extension number 0x9527 (which is a completely arbitrary
1872 and unofficial assignment based on the MD5 hash of the Internet
1873 Draft). Note that by doing so, you potentially lose
1874 interoperability with other TLS implementations since these might
1875 be using the same extension number for other purposes.
1877 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1878 opaque PRF input value to use in the handshake. This will create
1879 an interal copy of the length-'len' string at 'src', and will
1880 return non-zero for success.
1882 To get more control and flexibility, provide a callback function
1885 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1886 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1890 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1893 Callback function 'cb' will be called in handshakes, and is
1894 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1895 Argument 'arg' is for application purposes (the value as given to
1896 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1897 be provided to the callback function). The callback function
1898 has to return non-zero to report success: usually 1 to use opaque
1899 PRF input just if possible, or 2 to enforce use of the opaque PRF
1900 input. In the latter case, the library will abort the handshake
1901 if opaque PRF input is not successfully negotiated.
1903 Arguments 'peerinput' and 'len' given to the callback function
1904 will always be NULL and 0 in the case of a client. A server will
1905 see the client's opaque PRF input through these variables if
1906 available (NULL and 0 otherwise). Note that if the server
1907 provides an opaque PRF input, the length must be the same as the
1908 length of the client's opaque PRF input.
1910 Note that the callback function will only be called when creating
1911 a new session (session resumption can resume whatever was
1912 previously negotiated), and will not be called in SSL 2.0
1913 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1914 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1915 for applications that need to enforce opaque PRF input.
1919 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1922 [Victor B. Wagner <vitus@cryptocom.ru>]
1924 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1925 RFC4507bis. The encrypted ticket format is an encrypted encoded
1926 SSL_SESSION structure, that way new session features are automatically
1929 If a client application caches session in an SSL_SESSION structure
1930 support is transparent because tickets are now stored in the encoded
1933 The SSL_CTX structure automatically generates keys for ticket
1934 protection in servers so again support should be possible
1935 with no application modification.
1937 If a client or server wishes to disable RFC4507 support then the option
1938 SSL_OP_NO_TICKET can be set.
1940 Add a TLS extension debugging callback to allow the contents of any client
1941 or server extensions to be examined.
1943 This work was sponsored by Google.
1946 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1947 OpenSSL should now compile cleanly on gcc 4.2
1948 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1950 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1951 support including streaming MAC support: this is required for GOST
1952 ciphersuite support.
1953 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1955 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1956 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1957 to output in BER and PEM format.
1960 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1961 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1962 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1963 ENGINE support for HMAC keys which are unextractable. New -mac and
1964 -macopt options to dgst utility.
1967 *) New option -sigopt to dgst utility. Update dgst to use
1968 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1969 alternative signing paramaters such as X9.31 or PSS in the dgst
1973 *) Change ssl_cipher_apply_rule(), the internal function that does
1974 the work each time a ciphersuite string requests enabling
1975 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1976 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1977 the order of disabled ciphersuites such that those ciphersuites
1978 that most recently went from enabled to disabled not only stay
1979 in order with respect to each other, but also have higher priority
1980 than other disabled ciphersuites the next time ciphersuites are
1983 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1984 the same ciphersuites as with "HIGH" alone, but in a specific
1985 order where the PSK ciphersuites come first (since they are the
1986 most recently disabled ciphersuites when "HIGH" is parsed).
1988 Also, change ssl_create_cipher_list() (using this new
1989 funcionality) such that between otherwise identical
1990 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1994 *) Change ssl_create_cipher_list() so that it automatically
1995 arranges the ciphersuites in reasonable order before starting
1996 to process the rule string. Thus, the definition for "DEFAULT"
1997 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1998 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1999 This makes it much easier to arrive at a reasonable default order
2000 in applications for which anonymous ciphers are OK (meaning
2001 that you can't actually use DEFAULT).
2002 [Bodo Moeller; suggested by Victor Duchovni]
2004 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2005 processing) into multiple integers instead of setting
2006 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2007 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2008 (These masks as well as the individual bit definitions are hidden
2009 away into the non-exported interface ssl/ssl_locl.h, so this
2010 change to the definition of the SSL_CIPHER structure shouldn't
2011 affect applications.) This give us more bits for each of these
2012 categories, so there is no longer a need to coagulate AES128 and
2013 AES256 into a single algorithm bit, and to coagulate Camellia128
2014 and Camellia256 into a single algorithm bit, which has led to all
2017 Thus, among other things, the kludge introduced in 0.9.7m and
2018 0.9.8e for masking out AES256 independently of AES128 or masking
2019 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2021 With the change, we also introduce new ciphersuite aliases that
2022 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2026 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2027 Use the leftmost N bytes of the signature input if the input is
2028 larger than the prime q (with N being the size in bytes of q).
2031 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2032 it yet and it is largely untested.
2035 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2038 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2039 some compilers (gcc 4.2 and later) reject their use. Safestack is
2040 reimplemented. Update ASN1 to avoid use of legacy functions.
2043 *) Win32/64 targets are linked with Winsock2.
2046 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2047 to external functions. This can be used to increase CRL handling
2048 efficiency especially when CRLs are very large by (for example) storing
2049 the CRL revoked certificates in a database.
2052 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2053 new CRLs added to a directory can be used. New command line option
2054 -verify_return_error to s_client and s_server. This causes real errors
2055 to be returned by the verify callback instead of carrying on no matter
2056 what. This reflects the way a "real world" verify callback would behave.
2059 *) GOST engine, supporting several GOST algorithms and public key formats.
2060 Kindly donated by Cryptocom.
2063 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2064 partitioned by DP are handled but no indirect CRL or reason partitioning
2065 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2066 selected via a scoring technique which handles IDP and AKID in CRLs.
2069 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2070 will ultimately be used for all verify operations: this will remove the
2071 X509_STORE dependency on certificate verification and allow alternative
2072 lookup methods. X509_STORE based implementations of these two callbacks.
2075 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2076 Modify get_crl() to find a valid (unexpired) CRL if possible.
2079 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2080 this would be called X509_CRL_cmp() but that name is already used by
2081 a function that just compares CRL issuer names. Cache several CRL
2082 extensions in X509_CRL structure and cache CRLDP in X509.
2085 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2086 this maps equivalent X509_NAME structures into a consistent structure.
2087 Name comparison can then be performed rapidly using memcmp().
2090 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2094 *) Allow digests to supply their own micalg string for S/MIME type using
2095 the ctrl EVP_MD_CTRL_MICALG.
2098 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2099 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2100 ctrl. It can then customise the structure before and/or after signing
2104 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2105 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2106 to free up any added signature OIDs.
2109 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2110 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2111 digest and cipher tables. New options added to openssl utility:
2112 list-message-digest-algorithms and list-cipher-algorithms.
2115 *) Change the array representation of binary polynomials: the list
2116 of degrees of non-zero coefficients is now terminated with -1.
2117 Previously it was terminated with 0, which was also part of the
2118 value; thus, the array representation was not applicable to
2119 polynomials where t^0 has coefficient zero. This change makes
2120 the array representation useful in a more general context.
2123 *) Various modifications and fixes to SSL/TLS cipher string
2124 handling. For ECC, the code now distinguishes between fixed ECDH
2125 with RSA certificates on the one hand and with ECDSA certificates
2126 on the other hand, since these are separate ciphersuites. The
2127 unused code for Fortezza ciphersuites has been removed.
2129 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2130 (not "ECDHE"). For consistency with the code for DH
2131 certificates, use of ECDH certificates is now considered ECDH
2132 authentication, not RSA or ECDSA authentication (the latter is
2133 merely the CA's signing algorithm and not actively used in the
2136 The temporary ciphersuite alias "ECCdraft" is no longer
2137 available, and ECC ciphersuites are no longer excluded from "ALL"
2138 and "DEFAULT". The following aliases now exist for RFC 4492
2139 ciphersuites, most of these by analogy with the DH case:
2141 kECDHr - ECDH cert, signed with RSA
2142 kECDHe - ECDH cert, signed with ECDSA
2143 kECDH - ECDH cert (signed with either RSA or ECDSA)
2144 kEECDH - ephemeral ECDH
2145 ECDH - ECDH cert or ephemeral ECDH
2151 AECDH - anonymous ECDH
2152 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2156 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2157 Use correct micalg parameters depending on digest(s) in signed message.
2160 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2161 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2164 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2165 an engine to register a method. Add ENGINE lookups for methods and
2166 functional reference processing.
2169 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2170 EVP_{Sign,Verify}* which allow an application to customise the signature
2174 *) New -resign option to smime utility. This adds one or more signers
2175 to an existing PKCS#7 signedData structure. Also -md option to use an
2176 alternative message digest algorithm for signing.
2179 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2180 create PKCS7 structures containing multiple signers. Update smime
2181 application to support multiple signers.
2184 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2188 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2189 Reorganize PBE internals to lookup from a static table using NIDs,
2190 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2191 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2192 PRF which will be automatically used with PBES2.
2195 *) Replace the algorithm specific calls to generate keys in "req" with the
2199 *) Update PKCS#7 enveloped data routines to use new API. This is now
2200 supported by any public key method supporting the encrypt operation. A
2201 ctrl is added to allow the public key algorithm to examine or modify
2202 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2206 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2207 a default digest type to use. In most cases this will be SHA1 but some
2208 algorithms (such as GOST) need to specify an alternative digest. The
2209 return value indicates how strong the preference is 1 means optional and
2210 2 is mandatory (that is it is the only supported type). Modify
2211 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2212 use the default md. Update openssl utilities to use the default digest
2213 type for signing if it is not explicitly indicated.
2216 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2217 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2218 signing method from the key type. This effectively removes the link
2219 between digests and public key types.
2222 *) Add an OID cross reference table and utility functions. Its purpose is to
2223 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2224 rsaEncryption. This will allow some of the algorithm specific hackery
2225 needed to use the correct OID to be removed.
2228 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2229 structures for PKCS7_sign(). They are now set up by the relevant public
2233 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2236 *) Add support for key derivation (agreement) in the API, DH method and
2240 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2241 public and private key formats. As a side effect these add additional
2242 command line functionality not previously available: DSA signatures can be
2243 generated and verified using pkeyutl and DH key support and generation in
2248 [Oliver Tappe <zooey@hirschkaefer.de>]
2250 *) New make target "install_html_docs" installs HTML renditions of the
2252 [Oliver Tappe <zooey@hirschkaefer.de>]
2254 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2255 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2256 support key and parameter generation and add initial key generation
2257 functionality for RSA.
2260 *) Add functions for main EVP_PKEY_method operations. The undocumented
2261 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2262 EVP_PKEY_{encrypt,decrypt}_old.
2265 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2266 key API, doesn't do much yet.
2269 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2270 public key algorithms. New option to openssl utility:
2271 "list-public-key-algorithms" to print out info.
2274 *) Implement the Supported Elliptic Curves Extension for
2275 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2278 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2279 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2282 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2283 utilities such as rsa, dsa, dsaparam etc except they process any key
2287 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2288 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2289 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2293 *) Initial support for pluggable public key ASN1.
2294 De-spaghettify the public key ASN1 handling. Move public and private
2295 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2296 algorithm specific handling to a single module within the relevant
2297 algorithm directory. Add functions to allow (near) opaque processing
2298 of public and private key structures.
2301 *) Implement the Supported Point Formats Extension for
2302 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2305 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2306 for the psk identity [hint] and the psk callback functions to the
2307 SSL_SESSION, SSL and SSL_CTX structure.
2310 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2314 SSL_CTX_use_psk_identity_hint
2315 SSL_get_psk_identity_hint
2316 SSL_get_psk_identity
2317 SSL_use_psk_identity_hint
2319 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2321 *) Add RFC 3161 compliant time stamp request creation, response generation
2322 and response verification functionality.
2323 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2325 *) Add initial support for TLS extensions, specifically for the server_name
2326 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2327 have new members for a host name. The SSL data structure has an
2328 additional member SSL_CTX *initial_ctx so that new sessions can be
2329 stored in that context to allow for session resumption, even after the
2330 SSL has been switched to a new SSL_CTX in reaction to a client's
2331 server_name extension.
2333 New functions (subject to change):
2335 SSL_get_servername()
2336 SSL_get_servername_type()
2339 New CTRL codes and macros (subject to change):
2341 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2342 - SSL_CTX_set_tlsext_servername_callback()
2343 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2344 - SSL_CTX_set_tlsext_servername_arg()
2345 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2347 openssl s_client has a new '-servername ...' option.
2349 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2350 '-key2 ...', '-servername_fatal' (subject to change). This allows
2351 testing the HostName extension for a specific single host name ('-cert'
2352 and '-key' remain fallbacks for handshakes without HostName
2353 negotiation). If the unrecognized_name alert has to be sent, this by
2354 default is a warning; it becomes fatal with the '-servername_fatal'
2357 [Peter Sylvester, Remy Allais, Christophe Renou]
2359 *) Whirlpool hash implementation is added.
2362 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2363 bn(64,32). Because of instruction set limitations it doesn't have
2364 any negative impact on performance. This was done mostly in order
2365 to make it possible to share assembler modules, such as bn_mul_mont
2366 implementations, between 32- and 64-bit builds without hassle.
2369 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2370 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2374 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2375 dedicated Montgomery multiplication procedure, is introduced.
2376 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2377 "64-bit" performance on certain 32-bit targets.
2380 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2381 in SSL structures. New SSL ctrl to set maximum send fragment size.
2382 Save memory by seeting the I/O buffer sizes dynamically instead of
2383 using the maximum available value.
2386 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2387 in addition to the text details.
2390 *) Very, very preliminary EXPERIMENTAL support for printing of general
2391 ASN1 structures. This currently produces rather ugly output and doesn't
2392 handle several customised structures at all.
2395 *) Integrated support for PVK file format and some related formats such
2396 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2397 these in the 'rsa' and 'dsa' utilities.
2400 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2403 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2404 place for the (very old) "NETSCAPE" format certificates which are now
2405 handled using new ASN1 code equivalents.
2408 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2409 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2410 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2413 *) Modify CRL distribution points extension code to print out previously
2414 unsupported fields. Enhance extension setting code to allow setting of
2418 *) Add print and set support for Issuing Distribution Point CRL extension.
2421 *) Change 'Configure' script to enable Camellia by default.
2424 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2426 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2428 This addresses the flaw in CBC record processing discovered by
2429 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2430 at: http://www.isg.rhul.ac.uk/tls/
2432 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2433 Security Group at Royal Holloway, University of London
2434 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2435 Emilia Käsper for the initial patch.
2437 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2439 *) Return an error when checking OCSP signatures when key is NULL.
2440 This fixes a DoS attack. (CVE-2013-0166)
2443 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2444 the right response is stapled. Also change SSL_get_certificate()
2445 so it returns the certificate actually sent.
2446 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2447 (This is a backport)
2448 [Rob Stradling <rob.stradling@comodo.com>]
2450 *) Fix possible deadlock when decoding public keys.
2453 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2455 *) Sanity check record length before skipping explicit IV in DTLS
2458 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2459 fuzzing as a service testing platform.
2463 *) Initialise tkeylen properly when encrypting CMS messages.
2464 Thanks to Solar Designer of Openwall for reporting this issue.
2467 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2469 *) The fix for CVE-2012-2110 did not take into account that the
2470 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2471 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2472 rejecting negative len parameter. (CVE-2012-2131)
2473 [Tomas Hoger <thoger@redhat.com>]
2475 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2477 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2478 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2479 in CRYPTO_realloc_clean.
2481 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2482 issue and to Adam Langley <agl@chromium.org> for fixing it.
2484 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2486 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2488 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2489 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2490 content decryption and always return the same error. Note: this attack
2491 needs on average 2^20 messages so it only affects automated senders. The
2492 old behaviour can be reenabled in the CMS code by setting the
2493 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2494 an MMA defence is not necessary.
2495 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2496 this issue. (CVE-2012-0884)
2499 *) Fix CVE-2011-4619: make sure we really are receiving a
2500 client hello before rejecting multiple SGC restarts. Thanks to
2501 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2504 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2506 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2507 Thanks to Antonio Martin, Enterprise Secure Access Research and
2508 Development, Cisco Systems, Inc. for discovering this bug and
2509 preparing a fix. (CVE-2012-0050)
2512 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2514 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2515 of the Vaudenay padding oracle attack on CBC mode encryption
2516 which enables an efficient plaintext recovery attack against
2517 the OpenSSL implementation of DTLS. Their attack exploits timing
2518 differences arising during decryption processing. A research
2519 paper describing this attack can be found at:
2520 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2521 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2522 Security Group at Royal Holloway, University of London
2523 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2524 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2525 for preparing the fix. (CVE-2011-4108)
2526 [Robin Seggelmann, Michael Tuexen]
2528 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2529 [Ben Laurie, Kasper <ekasper@google.com>]
2531 *) Clear bytes used for block padding of SSL 3.0 records.
2533 [Adam Langley (Google)]
2535 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2536 Kadianakis <desnacked@gmail.com> for discovering this issue and
2537 Adam Langley for preparing the fix. (CVE-2011-4619)
2538 [Adam Langley (Google)]
2540 *) Prevent malformed RFC3779 data triggering an assertion failure.
2541 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2542 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2543 [Rob Austein <sra@hactrn.net>]
2545 *) Fix ssl_ciph.c set-up race.
2546 [Adam Langley (Google)]
2548 *) Fix spurious failures in ecdsatest.c.
2549 [Emilia Käsper (Google)]
2551 *) Fix the BIO_f_buffer() implementation (which was mixing different
2552 interpretations of the '..._len' fields).
2553 [Adam Langley (Google)]
2555 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2556 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2557 threads won't reuse the same blinding coefficients.
2559 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2560 lock to call BN_BLINDING_invert_ex, and avoids one use of
2561 BN_BLINDING_update for each BN_BLINDING structure (previously,
2562 the last update always remained unused).
2563 [Emilia Käsper (Google)]
2565 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2566 for multi-threaded use of ECDH.
2567 [Adam Langley (Google)]
2569 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2572 *) Add protection against ECDSA timing attacks as mentioned in the paper
2573 by Billy Bob Brumley and Nicola Tuveri, see:
2575 http://eprint.iacr.org/2011/232.pdf
2577 [Billy Bob Brumley and Nicola Tuveri]
2579 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2581 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2582 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2584 *) Fix bug in string printing code: if *any* escaping is enabled we must
2585 escape the escape character (backslash) or the resulting string is
2589 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2591 *) Disable code workaround for ancient and obsolete Netscape browsers
2592 and servers: an attacker can use it in a ciphersuite downgrade attack.
2593 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2596 *) Fixed J-PAKE implementation error, originally discovered by
2597 Sebastien Martini, further info and confirmation from Stefan
2598 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2601 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2603 *) Fix extension code to avoid race conditions which can result in a buffer
2604 overrun vulnerability: resumed sessions must not be modified as they can
2605 be shared by multiple threads. CVE-2010-3864
2608 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2611 *) Don't reencode certificate when calculating signature: cache and use
2612 the original encoding instead. This makes signature verification of
2613 some broken encodings work correctly.
2616 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2617 is also one of the inputs.
2618 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2620 *) Don't repeatedly append PBE algorithms to table if they already exist.
2621 Sort table on each new add. This effectively makes the table read only
2622 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2626 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2628 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2631 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2632 access or freeing data twice (CVE-2010-0742)
2633 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2635 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2636 common in certificates and some applications which only call
2637 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2641 Reduce copying into .apps and .test in makevms.com
2642 Don't try to use blank CA certificate in CA.com
2643 Allow use of C files from original directories in maketests.com
2644 [Steven M. Schweda" <sms@antinode.info>]
2646 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2648 *) When rejecting SSL/TLS records due to an incorrect version number, never
2649 update s->server with a new major version number. As of
2650 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2651 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2652 the previous behavior could result in a read attempt at NULL when
2653 receiving specific incorrect SSL/TLS records once record payload
2654 protection is active. (CVE-2010-0740)
2655 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2657 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2658 could be crashed if the relevant tables were not present (e.g. chrooted).
2659 [Tomas Hoger <thoger@redhat.com>]
2661 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2663 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2664 [Martin Olsson, Neel Mehta]
2666 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2667 accommodate for stack sorting, always a write lock!).
2670 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2671 excessive delays in the RAND_poll(): over a minute. As a workaround
2672 include a time check in the inner Heap32Next loop too.
2675 *) The code that handled flushing of data in SSL/TLS originally used the
2676 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2677 the problem outlined in PR#1949. The fix suggested there however can
2678 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2679 of Apache). So instead simplify the code to flush unconditionally.
2680 This should be fine since flushing with no data to flush is a no op.
2683 *) Handle TLS versions 2.0 and later properly and correctly use the
2684 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2685 off ancient servers have a habit of sticking around for a while...
2688 *) Modify compression code so it frees up structures without using the
2689 ex_data callbacks. This works around a problem where some applications
2690 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2691 restarting) then use compression (e.g. SSL with compression) later.
2692 This results in significant per-connection memory leaks and
2693 has caused some security issues including CVE-2008-1678 and
2697 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2698 change when encrypting or decrypting.
2701 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2702 connect and renegotiate with servers which do not support RI.
2703 Until RI is more widely deployed this option is enabled by default.
2706 *) Add "missing" ssl ctrls to clear options and mode.
2709 *) If client attempts to renegotiate and doesn't support RI respond with
2710 a no_renegotiation alert as required by RFC5746. Some renegotiating
2711 TLS clients will continue a connection gracefully when they receive
2712 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2713 waiting for a server hello which it will never receive. Now we treat a
2714 received no_renegotiation alert as a fatal error. This is because
2715 applications requesting a renegotiation might well expect it to succeed
2716 and would have no code in place to handle the server denying it so the
2717 only safe thing to do is to terminate the connection.
2720 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2721 peer supports secure renegotiation and 0 otherwise. Print out peer
2722 renegotiation support in s_client/s_server.
2725 *) Replace the highly broken and deprecated SPKAC certification method with
2726 the updated NID creation version. This should correctly handle UTF8.
2729 *) Implement RFC5746. Re-enable renegotiation but require the extension
2730 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2731 turns out to be a bad idea. It has been replaced by
2732 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2733 SSL_CTX_set_options(). This is really not recommended unless you
2734 know what you are doing.
2735 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2737 *) Fixes to stateless session resumption handling. Use initial_ctx when
2738 issuing and attempting to decrypt tickets in case it has changed during
2739 servername handling. Use a non-zero length session ID when attempting
2740 stateless session resumption: this makes it possible to determine if
2741 a resumption has occurred immediately after receiving server hello
2742 (several places in OpenSSL subtly assume this) instead of later in
2746 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2747 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2748 fixes for a few places where the return code is not checked
2750 [Julia Lawall <julia@diku.dk>]
2752 *) Add --strict-warnings option to Configure script to include devteam
2753 warnings in other configurations.
2756 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2757 makes it possible to install openssl libraries in locations which
2758 have names other than "lib", for example "/usr/lib64" which some
2760 [Steve Henson, based on patch from Jeremy Utley]
2762 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2763 X690 8.9.12 and can produce some misleading textual output of OIDs.
2764 [Steve Henson, reported by Dan Kaminsky]
2766 *) Delete MD2 from algorithm tables. This follows the recommendation in
2767 several standards that it is not used in new applications due to
2768 several cryptographic weaknesses. For binary compatibility reasons
2769 the MD2 API is still compiled in by default.
2772 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2776 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2777 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2779 [Guenter <lists@gknw.net>]
2781 *) Fix the server certificate chain building code to use X509_verify_cert(),
2782 it used to have an ad-hoc builder which was unable to cope with anything
2783 other than a simple chain.
2784 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2786 *) Don't check self signed certificate signatures in X509_verify_cert()
2787 by default (a flag can override this): it just wastes time without
2788 adding any security. As a useful side effect self signed root CAs
2789 with non-FIPS digests are now usable in FIPS mode.
2792 *) In dtls1_process_out_of_seq_message() the check if the current message
2793 is already buffered was missing. For every new message was memory
2794 allocated, allowing an attacker to perform an denial of service attack
2795 with sending out of seq handshake messages until there is no memory
2796 left. Additionally every future messege was buffered, even if the
2797 sequence number made no sense and would be part of another handshake.
2798 So only messages with sequence numbers less than 10 in advance will be
2799 buffered. (CVE-2009-1378)
2800 [Robin Seggelmann, discovered by Daniel Mentz]
2802 *) Records are buffered if they arrive with a future epoch to be
2803 processed after finishing the corresponding handshake. There is
2804 currently no limitation to this buffer allowing an attacker to perform
2805 a DOS attack with sending records with future epochs until there is no
2806 memory left. This patch adds the pqueue_size() function to determine
2807 the size of a buffer and limits the record buffer to 100 entries.
2809 [Robin Seggelmann, discovered by Daniel Mentz]
2811 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2812 parent structure is freed. (CVE-2009-1379)
2815 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2816 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2819 [Ilya O. <vrghost@gmail.com>]
2821 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2823 *) Disable renegotiation completely - this fixes a severe security
2824 problem (CVE-2009-3555) at the cost of breaking all
2825 renegotiation. Renegotiation can be re-enabled by setting
2826 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2827 run-time. This is really not recommended unless you know what
2831 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2833 *) Don't set val to NULL when freeing up structures, it is freed up by
2834 underlying code. If sizeof(void *) > sizeof(long) this can result in
2835 zeroing past the valid field. (CVE-2009-0789)
2836 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2838 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2839 checked correctly. This would allow some invalid signed attributes to
2840 appear to verify correctly. (CVE-2009-0591)
2841 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2843 *) Reject UniversalString and BMPString types with invalid lengths. This
2844 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2845 a legal length. (CVE-2009-0590)
2848 *) Set S/MIME signing as the default purpose rather than setting it
2849 unconditionally. This allows applications to override it at the store
2853 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2854 to handle some structures.
2857 *) Improve efficiency of mem_gets: don't search whole buffer each time
2859 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2861 *) New -hex option for openssl rand.
2864 *) Print out UTF8String and NumericString when parsing ASN1.
2867 *) Support NumericString type for name components.
2870 *) Allow CC in the environment to override the automatically chosen
2871 compiler. Note that nothing is done to ensure flags work with the
2875 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2877 *) Properly check EVP_VerifyFinal() and similar return values
2879 [Ben Laurie, Bodo Moeller, Google Security Team]
2881 *) Enable TLS extensions by default.
2884 *) Allow the CHIL engine to be loaded, whether the application is
2885 multithreaded or not. (This does not release the developer from the
2886 obligation to set up the dynamic locking callbacks.)
2887 [Sander Temme <sander@temme.net>]
2889 *) Use correct exit code if there is an error in dgst command.
2890 [Steve Henson; problem pointed out by Roland Dirlewanger]
2892 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2893 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2896 *) Add experimental JPAKE support, including demo authentication in
2897 s_client and s_server.
2900 *) Set the comparison function in v3_addr_canonize().
2901 [Rob Austein <sra@hactrn.net>]
2903 *) Add support for XMPP STARTTLS in s_client.
2904 [Philip Paeps <philip@freebsd.org>]
2906 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2907 to ensure that even with this option, only ciphersuites in the
2908 server's preference list will be accepted. (Note that the option
2909 applies only when resuming a session, so the earlier behavior was
2910 just about the algorithm choice for symmetric cryptography.)
2913 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2915 *) Fix NULL pointer dereference if a DTLS server received
2916 ChangeCipherSpec as first record (CVE-2009-1386).
2919 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2920 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2923 *) The fix in 0.9.8c that supposedly got rid of unsafe
2924 double-checked locking was incomplete for RSA blinding,
2925 addressing just one layer of what turns out to have been
2926 doubly unsafe triple-checked locking.
2928 So now fix this for real by retiring the MONT_HELPER macro
2929 in crypto/rsa/rsa_eay.c.
2931 [Bodo Moeller; problem pointed out by Marius Schilder]
2933 *) Various precautionary measures:
2935 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2937 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2938 (NB: This would require knowledge of the secret session ticket key
2939 to exploit, in which case you'd be SOL either way.)
2941 - Change bn_nist.c so that it will properly handle input BIGNUMs
2942 outside the expected range.
2944 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2947 [Neel Mehta, Bodo Moeller]
2949 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2950 the load fails. Useful for distros.
2951 [Ben Laurie and the FreeBSD team]
2953 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2956 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2959 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2961 This work was sponsored by Logica.
2964 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2965 keystores. Support for SSL/TLS client authentication too.
2966 Not compiled unless enable-capieng specified to Configure.
2968 This work was sponsored by Logica.
2971 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2972 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2973 attribute creation routines such as certifcate requests and PKCS#12
2977 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2979 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2980 handshake which could lead to a cilent crash as found using the
2981 Codenomicon TLS test suite (CVE-2008-1672)
2982 [Steve Henson, Mark Cox]
2984 *) Fix double free in TLS server name extensions which could lead to
2985 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2988 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2990 Clear the error queue to ensure that error entries left from
2991 older function calls do not interfere with the correct operation.
2992 [Lutz Jaenicke, Erik de Castro Lopo]
2994 *) Remove root CA certificates of commercial CAs:
2996 The OpenSSL project does not recommend any specific CA and does not
2997 have any policy with respect to including or excluding any CA.
2998 Therefore it does not make any sense to ship an arbitrary selection
2999 of root CA certificates with the OpenSSL software.
3002 *) RSA OAEP patches to fix two separate invalid memory reads.
3003 The first one involves inputs when 'lzero' is greater than
3004 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3005 before the beginning of from). The second one involves inputs where
3006 the 'db' section contains nothing but zeroes (there is a one-byte
3007 invalid read after the end of 'db').
3008 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3010 *) Partial backport from 0.9.9-dev:
3012 Introduce bn_mul_mont (dedicated Montgomery multiplication
3013 procedure) as a candidate for BIGNUM assembler implementation.
3014 While 0.9.9-dev uses assembler for various architectures, only
3015 x86_64 is available by default here in the 0.9.8 branch, and
3016 32-bit x86 is available through a compile-time setting.
3018 To try the 32-bit x86 assembler implementation, use Configure
3019 option "enable-montasm" (which exists only for this backport).
3021 As "enable-montasm" for 32-bit x86 disclaims code stability
3022 anyway, in this constellation we activate additional code
3023 backported from 0.9.9-dev for further performance improvements,
3024 namely BN_from_montgomery_word. (To enable this otherwise,
3025 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3027 [Andy Polyakov (backport partially by Bodo Moeller)]
3029 *) Add TLS session ticket callback. This allows an application to set
3030 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3031 values. This is useful for key rollover for example where several key
3032 sets may exist with different names.
3035 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3036 This was broken until now in 0.9.8 releases, such that the only way
3037 a registered ENGINE could be used (assuming it initialises
3038 successfully on the host) was to explicitly set it as the default
3039 for the relevant algorithms. This is in contradiction with 0.9.7
3040 behaviour and the documentation. With this fix, when an ENGINE is
3041 registered into a given algorithm's table of implementations, the
3042 'uptodate' flag is reset so that auto-discovery will be used next
3043 time a new context for that algorithm attempts to select an
3045 [Ian Lister (tweaked by Geoff Thorpe)]
3047 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3048 implemention in the following ways:
3050 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3053 Lack of BER streaming support means one pass streaming processing is
3054 only supported if data is detached: setting the streaming flag is
3055 ignored for embedded content.
3057 CMS support is disabled by default and must be explicitly enabled
3058 with the enable-cms configuration option.
3061 *) Update the GMP engine glue to do direct copies between BIGNUM and
3062 mpz_t when openssl and GMP use the same limb size. Otherwise the
3063 existing "conversion via a text string export" trick is still used.
3064 [Paul Sheer <paulsheer@gmail.com>]
3066 *) Zlib compression BIO. This is a filter BIO which compressed and
3067 uncompresses any data passed through it.
3070 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3071 RFC3394 compatible AES key wrapping.
3074 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3075 sets string data without copying. X509_ALGOR_set0() and
3076 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3077 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3078 from an X509_ATTRIBUTE structure optionally checking it occurs only
3079 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3083 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3084 to get the expected BN_FLG_CONSTTIME behavior.
3085 [Bodo Moeller (Google)]
3089 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3090 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3091 - added some more tests to do_tests.pl
3092 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3093 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3094 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3095 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3096 - various changes to netware.pl to enable gcc-cross builds on Win32
3098 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3099 - various changes to fix missing prototype warnings
3100 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3101 - added AES, WHIRLPOOL and CPUID assembler code to build files
3102 - added missing AES assembler make rules to mk1mf.pl
3103 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3104 [Guenter Knauf <eflash@gmx.net>]
3106 *) Implement certificate status request TLS extension defined in RFC3546.
3107 A client can set the appropriate parameters and receive the encoded
3108 OCSP response via a callback. A server can query the supplied parameters
3109 and set the encoded OCSP response in the callback. Add simplified examples
3110 to s_client and s_server.
3113 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3115 *) Fix various bugs:
3116 + Binary incompatibility of ssl_ctx_st structure
3117 + DTLS interoperation with non-compliant servers
3118 + Don't call get_session_cb() without proposed session
3119 + Fix ia64 assembler code
3120 [Andy Polyakov, Steve Henson]
3122 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3124 *) DTLS Handshake overhaul. There were longstanding issues with
3125 OpenSSL DTLS implementation, which were making it impossible for
3126 RFC 4347 compliant client to communicate with OpenSSL server.
3127 Unfortunately just fixing these incompatibilities would "cut off"
3128 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3129 server keeps tolerating non RFC compliant syntax. The opposite is
3130 not true, 0.9.8f client can not communicate with earlier server.
3131 This update even addresses CVE-2007-4995.
3134 *) Changes to avoid need for function casts in OpenSSL: some compilers
3135 (gcc 4.2 and later) reject their use.
3136 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3139 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3140 RFC4507bis. The encrypted ticket format is an encrypted encoded
3141 SSL_SESSION structure, that way new session features are automatically
3144 If a client application caches session in an SSL_SESSION structure
3145 support is transparent because tickets are now stored in the encoded
3148 The SSL_CTX structure automatically generates keys for ticket
3149 protection in servers so again support should be possible
3150 with no application modification.
3152 If a client or server wishes to disable RFC4507 support then the option
3153 SSL_OP_NO_TICKET can be set.
3155 Add a TLS extension debugging callback to allow the contents of any client
3156 or server extensions to be examined.
3158 This work was sponsored by Google.
3161 *) Add initial support for TLS extensions, specifically for the server_name
3162 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3163 have new members for a host name. The SSL data structure has an
3164 additional member SSL_CTX *initial_ctx so that new sessions can be
3165 stored in that context to allow for session resumption, even after the
3166 SSL has been switched to a new SSL_CTX in reaction to a client's
3167 server_name extension.
3169 New functions (subject to change):
3171 SSL_get_servername()
3172 SSL_get_servername_type()
3175 New CTRL codes and macros (subject to change):
3177 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3178 - SSL_CTX_set_tlsext_servername_callback()
3179 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3180 - SSL_CTX_set_tlsext_servername_arg()
3181 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3183 openssl s_client has a new '-servername ...' option.
3185 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3186 '-key2 ...', '-servername_fatal' (subject to change). This allows
3187 testing the HostName extension for a specific single host name ('-cert'
3188 and '-key' remain fallbacks for handshakes without HostName
3189 negotiation). If the unrecognized_name alert has to be sent, this by
3190 default is a warning; it becomes fatal with the '-servername_fatal'
3193 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3195 *) Add AES and SSE2 assembly language support to VC++ build.
3198 *) Mitigate attack on final subtraction in Montgomery reduction.
3201 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3202 (which previously caused an internal error).
3205 *) Squeeze another 10% out of IGE mode when in != out.
3208 *) AES IGE mode speedup.
3209 [Dean Gaudet (Google)]
3211 *) Add the Korean symmetric 128-bit cipher SEED (see
3212 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3213 add SEED ciphersuites from RFC 4162:
3215 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3216 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3217 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3218 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3220 To minimize changes between patchlevels in the OpenSSL 0.9.8
3221 series, SEED remains excluded from compilation unless OpenSSL
3222 is configured with 'enable-seed'.
3223 [KISA, Bodo Moeller]
3225 *) Mitigate branch prediction attacks, which can be practical if a
3226 single processor is shared, allowing a spy process to extract
3227 information. For detailed background information, see
3228 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3229 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3230 and Necessary Software Countermeasures"). The core of the change
3231 are new versions BN_div_no_branch() and
3232 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3233 respectively, which are slower, but avoid the security-relevant
3234 conditional branches. These are automatically called by BN_div()
3235 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3236 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3237 remove a conditional branch.
3239 BN_FLG_CONSTTIME is the new name for the previous
3240 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3241 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3242 in the exponent causes BN_mod_exp_mont() to use the alternative
3243 implementation in BN_mod_exp_mont_consttime().) The old name
3244 remains as a deprecated alias.
3246 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3247 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3248 constant-time implementations for more than just exponentiation.
3249 Here too the old name is kept as a deprecated alias.
3251 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3252 the BN_BLINDING structure gets an independent copy of the
3253 modulus. This means that the previous "BIGNUM *m" argument to
3254 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3255 essentially becomes "const BIGNUM *m", although we can't actually
3256 change this in the header file before 0.9.9. It allows
3257 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3258 enable BN_FLG_CONSTTIME.
3260 [Matthew D Wood (Intel Corp)]
3262 *) In the SSL/TLS server implementation, be strict about session ID
3263 context matching (which matters if an application uses a single
3264 external cache for different purposes). Previously,
3265 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3266 set. This did ensure strict client verification, but meant that,
3267 with applications using a single external cache for quite
3268 different requirements, clients could circumvent ciphersuite
3269 restrictions for a given session ID context by starting a session
3270 in a different context.
3273 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3274 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3275 authentication-only ciphersuites.
3278 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3279 not complete and could lead to a possible single byte overflow
3280 (CVE-2007-5135) [Ben Laurie]
3282 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3284 *) Since AES128 and AES256 (and similarly Camellia128 and
3285 Camellia256) share a single mask bit in the logic of
3286 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3287 kludge to work properly if AES128 is available and AES256 isn't
3288 (or if Camellia128 is available and Camellia256 isn't).
3291 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3292 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3293 When a point or a seed is encoded in a BIT STRING, we need to
3294 prevent the removal of trailing zero bits to get the proper DER
3295 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3296 of a NamedBitList, for which trailing 0 bits need to be removed.)
3299 *) Have SSL/TLS server implementation tolerate "mismatched" record
3300 protocol version while receiving ClientHello even if the
3301 ClientHello is fragmented. (The server can't insist on the
3302 particular protocol version it has chosen before the ServerHello
3303 message has informed the client about his choice.)
3306 *) Add RFC 3779 support.
3307 [Rob Austein for ARIN, Ben Laurie]
3309 *) Load error codes if they are not already present instead of using a
3310 static variable. This allows them to be cleanly unloaded and reloaded.
3311 Improve header file function name parsing.
3314 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3315 or CAPABILITY handshake as required by RFCs.
3318 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3320 *) Introduce limits to prevent malicious keys being able to
3321 cause a denial of service. (CVE-2006-2940)
3322 [Steve Henson, Bodo Moeller]
3324 *) Fix ASN.1 parsing of certain invalid structures that can result
3325 in a denial of service. (CVE-2006-2937) [Steve Henson]
3327 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3328 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3330 *) Fix SSL client code which could crash if connecting to a
3331 malicious SSLv2 server. (CVE-2006-4343)
3332 [Tavis Ormandy and Will Drewry, Google Security Team]
3334 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3335 match only those. Before that, "AES256-SHA" would be interpreted
3336 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3337 the same strength classification in 0.9.7h) as we currently only
3338 have a single AES bit in the ciphersuite description bitmap.
3339 That change, however, also applied to ciphersuite strings such as
3340 "RC4-MD5" that intentionally matched multiple ciphersuites --
3341 namely, SSL 2.0 ciphersuites in addition to the more common ones
3342 from SSL 3.0/TLS 1.0.
3344 So we change the selection algorithm again: Naming an explicit
3345 ciphersuite selects this one ciphersuite, and any other similar
3346 ciphersuite (same bitmap) from *other* protocol versions.
3347 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3348 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3350 Since SSL 2.0 does not have any ciphersuites for which the
3351 128/256 bit distinction would be relevant, this works for now.
3352 The proper fix will be to use different bits for AES128 and
3353 AES256, which would have avoided the problems from the beginning;
3354 however, bits are scarce, so we can only do this in a new release
3355 (not just a patchlevel) when we can change the SSL_CIPHER
3356 definition to split the single 'unsigned long mask' bitmap into
3357 multiple values to extend the available space.
3361 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3363 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3364 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3366 *) Add AES IGE and biIGE modes.
3369 *) Change the Unix randomness entropy gathering to use poll() when
3370 possible instead of select(), since the latter has some
3371 undesirable limitations.
3372 [Darryl Miles via Richard Levitte and Bodo Moeller]
3374 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3375 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3376 cannot be implicitly activated as part of, e.g., the "AES" alias.
3377 However, please upgrade to OpenSSL 0.9.9[-dev] for
3378 non-experimental use of the ECC ciphersuites to get TLS extension
3379 support, which is required for curve and point format negotiation
3380 to avoid potential handshake problems.
3383 *) Disable rogue ciphersuites:
3385 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3386 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3387 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3389 The latter two were purportedly from
3390 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3393 Also deactivate the remaining ciphersuites from
3394 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3395 unofficial, and the ID has long expired.
3398 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3399 dual-core machines) and other potential thread-safety issues.
3402 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3403 versions), which is now available for royalty-free use
3404 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3405 Also, add Camellia TLS ciphersuites from RFC 4132.
3407 To minimize changes between patchlevels in the OpenSSL 0.9.8
3408 series, Camellia remains excluded from compilation unless OpenSSL
3409 is configured with 'enable-camellia'.
3412 *) Disable the padding bug check when compression is in use. The padding
3413 bug check assumes the first packet is of even length, this is not
3414 necessarily true if compresssion is enabled and can result in false
3415 positives causing handshake failure. The actual bug test is ancient
3416 code so it is hoped that implementations will either have fixed it by
3417 now or any which still have the bug do not support compression.
3420 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3422 *) When applying a cipher rule check to see if string match is an explicit
3423 cipher suite and only match that one cipher suite if it is.
3426 *) Link in manifests for VC++ if needed.
3427 [Austin Ziegler <halostatue@gmail.com>]
3429 *) Update support for ECC-based TLS ciphersuites according to
3430 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3431 TLS extensions, which are supported starting with the 0.9.9
3432 branch, not in the OpenSSL 0.9.8 branch).
3435 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3436 opaque EVP_CIPHER_CTX handling.
3439 *) Fixes and enhancements to zlib compression code. We now only use
3440 "zlib1.dll" and use the default __cdecl calling convention on Win32
3441 to conform with the standards mentioned here:
3442 http://www.zlib.net/DLL_FAQ.txt
3443 Static zlib linking now works on Windows and the new --with-zlib-include
3444 --with-zlib-lib options to Configure can be used to supply the location
3445 of the headers and library. Gracefully handle case where zlib library
3449 *) Several fixes and enhancements to the OID generation code. The old code
3450 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3451 handle numbers larger than ULONG_MAX, truncated printing and had a
3452 non standard OBJ_obj2txt() behaviour.
3455 *) Add support for building of engines under engine/ as shared libraries
3456 under VC++ build system.
3459 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3460 Hopefully, we will not see any false combination of paths any more.
3463 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3465 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3466 (part of SSL_OP_ALL). This option used to disable the
3467 countermeasure against man-in-the-middle protocol-version
3468 rollback in the SSL 2.0 server implementation, which is a bad
3469 idea. (CVE-2005-2969)
3471 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3472 for Information Security, National Institute of Advanced Industrial
3473 Science and Technology [AIST], Japan)]
3475 *) Add two function to clear and return the verify parameter flags.
3478 *) Keep cipherlists sorted in the source instead of sorting them at
3479 runtime, thus removing the need for a lock.
3482 *) Avoid some small subgroup attacks in Diffie-Hellman.
3483 [Nick Mathewson and Ben Laurie]
3485 *) Add functions for well-known primes.
3488 *) Extended Windows CE support.
3489 [Satoshi Nakamura and Andy Polyakov]
3491 *) Initialize SSL_METHOD structures at compile time instead of during
3492 runtime, thus removing the need for a lock.
3495 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3496 attempting to decrypt each encrypted key in turn. Add support to
3500 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3502 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3505 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3508 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3509 key into the same file any more.
3512 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3515 *) Add -utf8 command line and config file option to 'ca'.
3516 [Stefan <stf@udoma.org]
3518 *) Removed the macro des_crypt(), as it seems to conflict with some
3519 libraries. Use DES_crypt().
3522 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3523 involves renaming the source and generated shared-libs for
3524 both. The engines will accept the corrected or legacy ids
3525 ('ncipher' and '4758_cca' respectively) when binding. NB,
3526 this only applies when building 'shared'.
3527 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3529 *) Add attribute functions to EVP_PKEY structure. Modify
3530 PKCS12_create() to recognize a CSP name attribute and
3531 use it. Make -CSP option work again in pkcs12 utility.
3534 *) Add new functionality to the bn blinding code:
3535 - automatic re-creation of the BN_BLINDING parameters after
3536 a fixed number of uses (currently 32)
3537 - add new function for parameter creation
3538 - introduce flags to control the update behaviour of the
3539 BN_BLINDING parameters
3540 - hide BN_BLINDING structure
3541 Add a second BN_BLINDING slot to the RSA structure to improve
3542 performance when a single RSA object is shared among several
3546 *) Add support for DTLS.
3547 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3549 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3550 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3553 *) Remove buggy and incomplete DH cert support from
3554 ssl/ssl_rsa.c and ssl/s3_both.c
3557 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3558 the apps/openssl applications.
3561 *) Compile clean with "-Wall -Wmissing-prototypes
3562 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3563 DEBUG_SAFESTACK must also be set.
3566 *) Change ./Configure so that certain algorithms can be disabled by default.
3567 The new counterpiece to "no-xxx" is "enable-xxx".
3569 The patented RC5 and MDC2 algorithms will now be disabled unless
3570 "enable-rc5" and "enable-mdc2", respectively, are specified.
3572 (IDEA remains enabled despite being patented. This is because IDEA
3573 is frequently required for interoperability, and there is no license
3574 fee for non-commercial use. As before, "no-idea" can be used to
3575 avoid this algorithm.)
3579 *) Add processing of proxy certificates (see RFC 3820). This work was
3580 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3581 EGEE (Enabling Grids for E-science in Europe).
3584 *) RC4 performance overhaul on modern architectures/implementations, such
3585 as Intel P4, IA-64 and AMD64.
3588 *) New utility extract-section.pl. This can be used specify an alternative
3589 section number in a pod file instead of having to treat each file as
3590 a separate case in Makefile. This can be done by adding two lines to the
3593 =for comment openssl_section:XXX
3595 The blank line is mandatory.
3599 *) New arguments -certform, -keyform and -pass for s_client and s_server
3600 to allow alternative format key and certificate files and passphrase
3604 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3605 update associated structures and add various utility functions.
3607 Add new policy related verify parameters, include policy checking in
3608 standard verify code. Enhance 'smime' application with extra parameters
3609 to support policy checking and print out.
3612 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3613 Nehemiah processors. These extensions support AES encryption in hardware
3614 as well as RNG (though RNG support is currently disabled).
3615 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3617 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3620 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3621 [Andy Polyakov and a number of other people]
3623 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3624 implementation contributed by IBM.
3625 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3627 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3628 exponent rather than 'unsigned long'. There is a corresponding change to
3629 the new 'rsa_keygen' element of the RSA_METHOD structure.
3630 [Jelte Jansen, Geoff Thorpe]
3632 *) Functionality for creating the initial serial number file is now
3633 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3635 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3636 number file to 1, which is bound to cause problems. To avoid
3637 the problems while respecting compatibility between different 0.9.7
3638 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3639 CA.pl for serial number initialization. With the new release 0.9.8,
3640 we can fix the problem directly in the 'ca' utility.)
3643 *) Reduced header interdepencies by declaring more opaque objects in
3644 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3645 give fewer recursive includes, which could break lazy source code - so
3646 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3647 developers should define this symbol when building and using openssl to
3648 ensure they track the recommended behaviour, interfaces, [etc], but
3649 backwards-compatible behaviour prevails when this isn't defined.
3652 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3655 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3656 This will generate a random key of the appropriate length based on the
3657 cipher context. The EVP_CIPHER can provide its own random key generation
3658 routine to support keys of a specific form. This is used in the des and
3659 3des routines to generate a key of the correct parity. Update S/MIME
3660 code to use new functions and hence generate correct parity DES keys.
3661 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3662 valid (weak or incorrect parity).
3665 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3666 as looking them up. This is useful when the verified structure may contain
3667 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3668 present unless the new PKCS7_NO_CRL flag is asserted.
3671 *) Extend ASN1 oid configuration module. It now additionally accepts the
3674 shortName = some long name, 1.2.3.4
3677 *) Reimplemented the BN_CTX implementation. There is now no more static
3678 limitation on the number of variables it can handle nor the depth of the
3679 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3680 information can now expand as required, and rather than having a single
3681 static array of bignums, BN_CTX now uses a linked-list of such arrays
3682 allowing it to expand on demand whilst maintaining the usefulness of
3683 BN_CTX's "bundling".
3686 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3687 to allow all RSA operations to function using a single BN_CTX.
3690 *) Preliminary support for certificate policy evaluation and checking. This
3691 is initially intended to pass the tests outlined in "Conformance Testing
3692 of Relying Party Client Certificate Path Processing Logic" v1.07.
3695 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3696 remained unused and not that useful. A variety of other little bignum
3697 tweaks and fixes have also been made continuing on from the audit (see
3701 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3702 associated ASN1, EVP and SSL functions and old ASN1 macros.
3705 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3706 and this should never fail. So the return value from the use of
3707 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3708 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3711 *) BN_CTX_get() should return zero-valued bignums, providing the same
3712 initialised value as BN_new().
3713 [Geoff Thorpe, suggested by Ulf Möller]
3715 *) Support for inhibitAnyPolicy certificate extension.
3718 *) An audit of the BIGNUM code is underway, for which debugging code is
3719 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3720 is considered valid when processing BIGNUMs, and causes execution to
3721 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3722 further steps are taken to deliberately pollute unused data in BIGNUM
3723 structures to try and expose faulty code further on. For now, openssl will
3724 (in its default mode of operation) continue to tolerate the inconsistent
3725 forms that it has tolerated in the past, but authors and packagers should
3726 consider trying openssl and their own applications when compiled with
3727 these debugging symbols defined. It will help highlight potential bugs in
3728 their own code, and will improve the test coverage for OpenSSL itself. At
3729 some point, these tighter rules will become openssl's default to improve
3730 maintainability, though the assert()s and other overheads will remain only
3731 in debugging configurations. See bn.h for more details.
3732 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3734 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3735 that can only be obtained through BN_CTX_new() (which implicitly
3736 initialises it). The presence of this function only made it possible
3737 to overwrite an existing structure (and cause memory leaks).
3740 *) Because of the callback-based approach for implementing LHASH as a
3741 template type, lh_insert() adds opaque objects to hash-tables and
3742 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3743 to clean up those corresponding objects before destroying the hash table
3744 (and losing the object pointers). So some over-zealous constifications in
3745 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3746 objects as "const" and the lh_doall[_arg] callback wrappers are not
3747 prototyped to have "const" restrictions on the object pointers they are
3748 given (and so aren't required to cast them away any more).
3751 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3752 (speed) prefers to use its own implementation. The two implementations
3753 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3754 its object type properly exposed (MS_TM) instead of casting to/from "char
3755 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3756 aren't necessarily the greatest nomenclatures - but this is what was used
3757 internally to the implementation so I've used that for now.
3760 *) Ensure that deprecated functions do not get compiled when
3761 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3762 the self-tests were still using deprecated key-generation functions so
3763 these have been updated also.
3766 *) Reorganise PKCS#7 code to separate the digest location functionality
3767 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3768 New function PKCS7_set_digest() to set the digest type for PKCS#7
3769 digestedData type. Add additional code to correctly generate the
3770 digestedData type and add support for this type in PKCS7 initialization
3774 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3775 structure of type "other".
3778 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3779 sure the loop does correctly stop and breaking ("division by zero")
3780 modulus operations are not performed. The (pre-generated) prime
3781 table crypto/bn/bn_prime.h was already correct, but it could not be
3782 re-generated on some platforms because of the "division by zero"
3783 situation in the script.
3784 [Ralf S. Engelschall]
3786 *) Update support for ECC-based TLS ciphersuites according to
3787 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3788 SHA-1 now is only used for "small" curves (where the
3789 representation of a field element takes up to 24 bytes); for
3790 larger curves, the field element resulting from ECDH is directly
3791 used as premaster secret.
3792 [Douglas Stebila (Sun Microsystems Laboratories)]
3794 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3795 curve secp160r1 to the tests.
3796 [Douglas Stebila (Sun Microsystems Laboratories)]
3798 *) Add the possibility to load symbols globally with DSO.
3799 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3801 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3802 control of the error stack.
3805 *) Add support for STORE in ENGINE.
3808 *) Add the STORE type. The intention is to provide a common interface
3809 to certificate and key stores, be they simple file-based stores, or
3810 HSM-type store, or LDAP stores, or...
3811 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3814 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3815 pass a list of arguments to any function as well as provide a way
3816 for a function to pass data back to the caller.
3819 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3820 works like BUF_strdup() but can be used to duplicate a portion of
3821 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3825 *) Add the function sk_find_ex() which works like sk_find(), but will
3826 return an index to an element even if an exact match couldn't be
3827 found. The index is guaranteed to point at the element where the
3828 searched-for key would be inserted to preserve sorting order.
3831 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3832 takes an extra flags argument for optional functionality. Currently,
3833 the following flags are defined:
3835 OBJ_BSEARCH_VALUE_ON_NOMATCH
3836 This one gets OBJ_bsearch_ex() to return a pointer to the first
3837 element where the comparing function returns a negative or zero
3840 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3841 This one gets OBJ_bsearch_ex() to return a pointer to the first
3842 element where the comparing function returns zero. This is useful
3843 if there are more than one element where the comparing function
3847 *) Make it possible to create self-signed certificates with 'openssl ca'
3848 in such a way that the self-signed certificate becomes part of the
3849 CA database and uses the same mechanisms for serial number generation
3850 as all other certificate signing. The new flag '-selfsign' enables
3851 this functionality. Adapt CA.sh and CA.pl.in.
3854 *) Add functionality to check the public key of a certificate request
3855 against a given private. This is useful to check that a certificate
3856 request can be signed by that key (self-signing).
3859 *) Make it possible to have multiple active certificates with the same
3860 subject in the CA index file. This is done only if the keyword
3861 'unique_subject' is set to 'no' in the main CA section (default
3862 if 'CA_default') of the configuration file. The value is saved
3863 with the database itself in a separate index attribute file,
3864 named like the index file with '.attr' appended to the name.
3867 *) Generate muti valued AVAs using '+' notation in config files for
3871 *) Support for nameConstraints certificate extension.
3874 *) Support for policyConstraints certificate extension.
3877 *) Support for policyMappings certificate extension.
3880 *) Make sure the default DSA_METHOD implementation only uses its
3881 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3882 and change its own handlers to be NULL so as to remove unnecessary
3883 indirection. This lets alternative implementations fallback to the
3884 default implementation more easily.
3887 *) Support for directoryName in GeneralName related extensions
3891 *) Make it possible to link applications using Makefile.shared.
3892 Make that possible even when linking against static libraries!
3895 *) Support for single pass processing for S/MIME signing. This now
3896 means that S/MIME signing can be done from a pipe, in addition
3897 cleartext signing (multipart/signed type) is effectively streaming
3898 and the signed data does not need to be all held in memory.
3900 This is done with a new flag PKCS7_STREAM. When this flag is set
3901 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3902 is done after the data is output (and digests calculated) in
3903 SMIME_write_PKCS7().
3906 *) Add full support for -rpath/-R, both in shared libraries and
3907 applications, at least on the platforms where it's known how
3911 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3912 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3913 will now compute a table of multiples of the generator that
3914 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3915 faster (notably in the case of a single point multiplication,
3916 scalar * generator).
3917 [Nils Larsch, Bodo Moeller]
3919 *) IPv6 support for certificate extensions. The various extensions
3920 which use the IP:a.b.c.d can now take IPv6 addresses using the
3921 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3925 *) Added an ENGINE that implements RSA by performing private key
3926 exponentiations with the GMP library. The conversions to and from
3927 GMP's mpz_t format aren't optimised nor are any montgomery forms
3928 cached, and on x86 it appears OpenSSL's own performance has caught up.
3929 However there are likely to be other architectures where GMP could
3930 provide a boost. This ENGINE is not built in by default, but it can be
3931 specified at Configure time and should be accompanied by the necessary
3932 linker additions, eg;
3933 ./config -DOPENSSL_USE_GMP -lgmp
3936 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3937 testing availability of engines with "-t" - the old behaviour is
3938 produced by increasing the feature's verbosity with "-tt".
3941 *) ECDSA routines: under certain error conditions uninitialized BN objects
3942 could be freed. Solution: make sure initialization is performed early
3943 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3947 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3948 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3949 software implementations. For DSA and DH, parameter generation can
3950 also be overriden by providing the appropriate method callbacks.
3953 *) Change the "progress" mechanism used in key-generation and
3954 primality testing to functions that take a new BN_GENCB pointer in
3955 place of callback/argument pairs. The new API functions have "_ex"
3956 postfixes and the older functions are reimplemented as wrappers for
3957 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3958 declarations of the old functions to help (graceful) attempts to
3959 migrate to the new functions. Also, the new key-generation API
3960 functions operate on a caller-supplied key-structure and return
3961 success/failure rather than returning a key or NULL - this is to
3962 help make "keygen" another member function of RSA_METHOD etc.
3964 Example for using the new callback interface:
3966 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3970 BN_GENCB_set(&my_cb, my_callback, my_arg);
3972 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3973 /* For the meaning of a, b in calls to my_callback(), see the
3974 * documentation of the function that calls the callback.
3975 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3976 * my_callback should return 1 if it wants BN_is_prime_ex()
3977 * to continue, or 0 to stop.
3982 *) Change the ZLIB compression method to be stateful, and make it
3983 available to TLS with the number defined in
3984 draft-ietf-tls-compression-04.txt.
3987 *) Add the ASN.1 structures and functions for CertificatePair, which
3988 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3990 CertificatePair ::= SEQUENCE {
3991 forward [0] Certificate OPTIONAL,
3992 reverse [1] Certificate OPTIONAL,
3993 -- at least one of the pair shall be present -- }
3995 Also implement the PEM functions to read and write certificate
3996 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3998 This needed to be defined, mostly for the sake of the LDAP
3999 attribute crossCertificatePair, but may prove useful elsewhere as
4003 *) Make it possible to inhibit symlinking of shared libraries in
4004 Makefile.shared, for Cygwin's sake.
4007 *) Extend the BIGNUM API by creating a function
4008 void BN_set_negative(BIGNUM *a, int neg);
4009 and a macro that behave like
4010 int BN_is_negative(const BIGNUM *a);
4012 to avoid the need to access 'a->neg' directly in applications.
4015 *) Implement fast modular reduction for pseudo-Mersenne primes
4016 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4017 EC_GROUP_new_curve_GFp() will now automatically use this
4019 [Nils Larsch <nla@trustcenter.de>]
4021 *) Add new lock type (CRYPTO_LOCK_BN).
4024 *) Change the ENGINE framework to automatically load engines
4025 dynamically from specific directories unless they could be
4026 found to already be built in or loaded. Move all the
4027 current engines except for the cryptodev one to a new
4029 The engines in engines/ are built as shared libraries if
4030 the "shared" options was given to ./Configure or ./config.
4031 Otherwise, they are inserted in libcrypto.a.
4032 /usr/local/ssl/engines is the default directory for dynamic
4033 engines, but that can be overriden at configure time through
4034 the usual use of --prefix and/or --openssldir, and at run
4035 time with the environment variable OPENSSL_ENGINES.
4036 [Geoff Thorpe and Richard Levitte]
4038 *) Add Makefile.shared, a helper makefile to build shared
4039 libraries. Adapt Makefile.org.
4042 *) Add version info to Win32 DLLs.
4043 [Peter 'Luna' Runestig" <peter@runestig.com>]
4045 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4046 can be added using this API to created arbitrary PKCS#12
4047 files while avoiding the low level API.
4049 New options to PKCS12_create(), key or cert can be NULL and
4050 will then be omitted from the output file. The encryption
4051 algorithm NIDs can be set to -1 for no encryption, the mac
4052 iteration count can be set to 0 to omit the mac.
4054 Enhance pkcs12 utility by making the -nokeys and -nocerts
4055 options work when creating a PKCS#12 file. New option -nomac
4056 to omit the mac, NONE can be set for an encryption algorithm.
4057 New code is modified to use the enhanced PKCS12_create()
4058 instead of the low level API.
4061 *) Extend ASN1 encoder to support indefinite length constructed
4062 encoding. This can output sequences tags and octet strings in
4063 this form. Modify pk7_asn1.c to support indefinite length
4064 encoding. This is experimental and needs additional code to
4065 be useful, such as an ASN1 bio and some enhanced streaming
4068 Extend template encode functionality so that tagging is passed
4069 down to the template encoder.
4072 *) Let 'openssl req' fail if an argument to '-newkey' is not
4073 recognized instead of using RSA as a default.
4076 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4077 As these are not official, they are not included in "ALL";
4078 the "ECCdraft" ciphersuite group alias can be used to select them.
4079 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4081 *) Add ECDH engine support.
4082 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4084 *) Add ECDH in new directory crypto/ecdh/.
4085 [Douglas Stebila (Sun Microsystems Laboratories)]
4087 *) Let BN_rand_range() abort with an error after 100 iterations
4088 without success (which indicates a broken PRNG).
4091 *) Change BN_mod_sqrt() so that it verifies that the input value
4092 is really the square of the return value. (Previously,
4093 BN_mod_sqrt would show GIGO behaviour.)
4096 *) Add named elliptic curves over binary fields from X9.62, SECG,
4097 and WAP/WTLS; add OIDs that were still missing.
4099 [Sheueling Chang Shantz and Douglas Stebila
4100 (Sun Microsystems Laboratories)]
4102 *) Extend the EC library for elliptic curves over binary fields
4103 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4106 EC_GF2m_simple_method
4110 EC_GROUP_new_curve_GF2m
4111 EC_GROUP_set_curve_GF2m
4112 EC_GROUP_get_curve_GF2m
4113 EC_POINT_set_affine_coordinates_GF2m
4114 EC_POINT_get_affine_coordinates_GF2m
4115 EC_POINT_set_compressed_coordinates_GF2m
4117 Point compression for binary fields is disabled by default for
4118 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4121 As binary polynomials are represented as BIGNUMs, various members
4122 of the EC_GROUP and EC_POINT data structures can be shared
4123 between the implementations for prime fields and binary fields;
4124 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4125 are essentially identical to their ..._GFp counterparts.
4126 (For simplicity, the '..._GFp' prefix has been dropped from
4127 various internal method names.)
4129 An internal 'field_div' method (similar to 'field_mul' and
4130 'field_sqr') has been added; this is used only for binary fields.
4132 [Sheueling Chang Shantz and Douglas Stebila
4133 (Sun Microsystems Laboratories)]
4135 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4136 through methods ('mul', 'precompute_mult').
4138 The generic implementations (now internally called 'ec_wNAF_mul'
4139 and 'ec_wNAF_precomputed_mult') remain the default if these
4140 methods are undefined.
4142 [Sheueling Chang Shantz and Douglas Stebila
4143 (Sun Microsystems Laboratories)]
4145 *) New function EC_GROUP_get_degree, which is defined through
4146 EC_METHOD. For curves over prime fields, this returns the bit
4147 length of the modulus.
4149 [Sheueling Chang Shantz and Douglas Stebila
4150 (Sun Microsystems Laboratories)]
4152 *) New functions EC_GROUP_dup, EC_POINT_dup.
4153 (These simply call ..._new and ..._copy).
4155 [Sheueling Chang Shantz and Douglas Stebila
4156 (Sun Microsystems Laboratories)]
4158 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4159 Polynomials are represented as BIGNUMs (where the sign bit is not
4160 used) in the following functions [macros]:
4163 BN_GF2m_sub [= BN_GF2m_add]
4164 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4165 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4166 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4168 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4169 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4170 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4171 BN_GF2m_cmp [= BN_ucmp]
4173 (Note that only the 'mod' functions are actually for fields GF(2^m).
4174 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4176 For some functions, an the irreducible polynomial defining a
4177 field can be given as an 'unsigned int[]' with strictly
4178 decreasing elements giving the indices of those bits that are set;
4179 i.e., p[] represents the polynomial
4180 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4182 p[0] > p[1] > ... > p[k] = 0.
4183 This applies to the following functions:
4188 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4189 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4191 BN_GF2m_mod_sqrt_arr
4192 BN_GF2m_mod_solve_quad_arr
4196 Conversion can be performed by the following functions:
4201 bntest.c has additional tests for binary polynomial arithmetic.
4203 Two implementations for BN_GF2m_mod_div() are available.
4204 The default algorithm simply uses BN_GF2m_mod_inv() and
4205 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4206 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4207 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4209 [Sheueling Chang Shantz and Douglas Stebila
4210 (Sun Microsystems Laboratories)]
4212 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4213 functionality is disabled at compile-time.
4214 [Douglas Stebila <douglas.stebila@sun.com>]
4216 *) Change default behaviour of 'openssl asn1parse' so that more
4217 information is visible when viewing, e.g., a certificate:
4219 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4220 mode the content of non-printable OCTET STRINGs is output in a
4221 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4222 avoid the appearance of a printable string.
4223 [Nils Larsch <nla@trustcenter.de>]
4225 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4227 EC_GROUP_set_asn1_flag()
4228 EC_GROUP_get_asn1_flag()
4229 EC_GROUP_set_point_conversion_form()
4230 EC_GROUP_get_point_conversion_form()
4231 These control ASN1 encoding details:
4232 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4233 has been set to OPENSSL_EC_NAMED_CURVE.
4234 - Points are encoded in uncompressed form by default; options for
4235 asn1_for are as for point2oct, namely
4236 POINT_CONVERSION_COMPRESSED
4237 POINT_CONVERSION_UNCOMPRESSED
4238 POINT_CONVERSION_HYBRID
4240 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4243 EC_GROUP_get0_seed()
4244 EC_GROUP_get_seed_len()
4245 This is used only for ASN1 purposes (so far).
4246 [Nils Larsch <nla@trustcenter.de>]
4248 *) Add 'field_type' member to EC_METHOD, which holds the NID
4249 of the appropriate field type OID. The new function
4250 EC_METHOD_get_field_type() returns this value.
4251 [Nils Larsch <nla@trustcenter.de>]
4256 EC_POINT_point2hex()
4257 EC_POINT_hex2point()
4258 providing useful interfaces to EC_POINT_point2oct() and
4259 EC_POINT_oct2point().
4260 [Nils Larsch <nla@trustcenter.de>]
4262 *) Change internals of the EC library so that the functions
4263 EC_GROUP_set_generator()
4264 EC_GROUP_get_generator()
4265 EC_GROUP_get_order()
4266 EC_GROUP_get_cofactor()
4267 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4268 to methods, which would lead to unnecessary code duplication when
4269 adding different types of curves.
4270 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4272 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4273 arithmetic, and such that modified wNAFs are generated
4274 (which avoid length expansion in many cases).
4277 *) Add a function EC_GROUP_check_discriminant() (defined via
4278 EC_METHOD) that verifies that the curve discriminant is non-zero.
4280 Add a function EC_GROUP_check() that makes some sanity tests
4281 on a EC_GROUP, its generator and order. This includes
4282 EC_GROUP_check_discriminant().
4283 [Nils Larsch <nla@trustcenter.de>]
4285 *) Add ECDSA in new directory crypto/ecdsa/.
4287 Add applications 'openssl ecparam' and 'openssl ecdsa'
4288 (these are based on 'openssl dsaparam' and 'openssl dsa').
4290 ECDSA support is also included in various other files across the
4291 library. Most notably,
4292 - 'openssl req' now has a '-newkey ecdsa:file' option;
4293 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4294 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4295 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4296 them suitable for ECDSA where domain parameters must be
4297 extracted before the specific public key;
4298 - ECDSA engine support has been added.
4299 [Nils Larsch <nla@trustcenter.de>]
4301 *) Include some named elliptic curves, and add OIDs from X9.62,
4302 SECG, and WAP/WTLS. Each curve can be obtained from the new
4304 EC_GROUP_new_by_curve_name(),
4305 and the list of available named curves can be obtained with
4306 EC_get_builtin_curves().
4307 Also add a 'curve_name' member to EC_GROUP objects, which can be
4309 EC_GROUP_set_curve_name()
4310 EC_GROUP_get_curve_name()
4311 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4313 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4314 was actually never needed) and in BN_mul(). The removal in BN_mul()
4315 required a small change in bn_mul_part_recursive() and the addition
4316 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4317 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4318 bn_sub_words() and bn_add_words() except they take arrays with
4322 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4324 *) Cleanse PEM buffers before freeing them since they may contain
4326 [Benjamin Bennett <ben@psc.edu>]
4328 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4329 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4330 authentication-only ciphersuites.
4333 *) Since AES128 and AES256 share a single mask bit in the logic of
4334 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4335 kludge to work properly if AES128 is available and AES256 isn't.
4338 *) Expand security boundary to match 1.1.1 module.
4341 *) Remove redundant features: hash file source, editing of test vectors
4342 modify fipsld to use external fips_premain.c signature.
4345 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4346 run algorithm test programs.
4349 *) Make algorithm test programs more tolerant of whitespace.
4352 *) Have SSL/TLS server implementation tolerate "mismatched" record
4353 protocol version while receiving ClientHello even if the
4354 ClientHello is fragmented. (The server can't insist on the
4355 particular protocol version it has chosen before the ServerHello
4356 message has informed the client about his choice.)
4359 *) Load error codes if they are not already present instead of using a
4360 static variable. This allows them to be cleanly unloaded and reloaded.
4363 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4365 *) Introduce limits to prevent malicious keys being able to
4366 cause a denial of service. (CVE-2006-2940)
4367 [Steve Henson, Bodo Moeller]
4369 *) Fix ASN.1 parsing of certain invalid structures that can result
4370 in a denial of service. (CVE-2006-2937) [Steve Henson]
4372 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4373 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4375 *) Fix SSL client code which could crash if connecting to a
4376 malicious SSLv2 server. (CVE-2006-4343)
4377 [Tavis Ormandy and Will Drewry, Google Security Team]
4379 *) Change ciphersuite string processing so that an explicit
4380 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4381 will no longer include "AES128-SHA"), and any other similar
4382 ciphersuite (same bitmap) from *other* protocol versions (so that
4383 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4384 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4385 changes from 0.9.8b and 0.9.8d.
4388 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4390 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4391 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4393 *) Change the Unix randomness entropy gathering to use poll() when
4394 possible instead of select(), since the latter has some
4395 undesirable limitations.
4396 [Darryl Miles via Richard Levitte and Bodo Moeller]
4398 *) Disable rogue ciphersuites:
4400 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4401 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4402 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4404 The latter two were purportedly from
4405 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4408 Also deactive the remaining ciphersuites from
4409 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4410 unofficial, and the ID has long expired.
4413 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4414 dual-core machines) and other potential thread-safety issues.
4417 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4419 *) Adapt fipsld and the build system to link against the validated FIPS
4420 module in FIPS mode.
4423 *) Fixes for VC++ 2005 build under Windows.
4426 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4427 from a Windows bash shell such as MSYS. It is autodetected from the
4428 "config" script when run from a VC++ environment. Modify standard VC++
4429 build to use fipscanister.o from the GNU make build.
4432 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4434 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4435 The value now differs depending on if you build for FIPS or not.
4436 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4437 safely run with a non-FIPSed libcrypto, as it may crash because of
4438 the difference induced by this change.
4441 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4443 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4444 (part of SSL_OP_ALL). This option used to disable the
4445 countermeasure against man-in-the-middle protocol-version
4446 rollback in the SSL 2.0 server implementation, which is a bad
4447 idea. (CVE-2005-2969)
4449 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4450 for Information Security, National Institute of Advanced Industrial
4451 Science and Technology [AIST], Japan)]
4453 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4454 mainly for FIPS compliance and not fully integrated at this stage.
4457 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4458 the exponentiation using a fixed-length exponent. (Otherwise,
4459 the information leaked through timing could expose the secret key
4460 after many signatures; cf. Bleichenbacher's attack on DSA with
4464 *) Make a new fixed-window mod_exp implementation the default for
4465 RSA, DSA, and DH private-key operations so that the sequence of
4466 squares and multiplies and the memory access pattern are
4467 independent of the particular secret key. This will mitigate
4468 cache-timing and potential related attacks.
4470 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4471 and this is automatically used by BN_mod_exp_mont() if the new flag
4472 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4473 will use this BN flag for private exponents unless the flag
4474 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4475 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4477 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4479 *) Change the client implementation for SSLv23_method() and
4480 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4481 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4482 (Previously, the SSL 2.0 backwards compatible Client Hello
4483 message format would be used even with SSL_OP_NO_SSLv2.)
4486 *) Add support for smime-type MIME parameter in S/MIME messages which some
4490 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4491 a threadsafe manner. Modify rsa code to use new function and add calls
4492 to dsa and dh code (which had race conditions before).
4495 *) Include the fixed error library code in the C error file definitions
4496 instead of fixing them up at runtime. This keeps the error code
4497 structures constant.
4500 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4502 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4505 *) Fixes for newer kerberos headers. NB: the casts are needed because
4506 the 'length' field is signed on one version and unsigned on another
4507 with no (?) obvious way to tell the difference, without these VC++
4508 complains. Also the "definition" of FAR (blank) is no longer included
4509 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4510 some needed definitions.
4513 *) Undo Cygwin change.
4516 *) Added support for proxy certificates according to RFC 3820.
4517 Because they may be a security thread to unaware applications,
4518 they must be explicitely allowed in run-time. See
4519 docs/HOWTO/proxy_certificates.txt for further information.
4522 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4524 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4525 server and client random values. Previously
4526 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4527 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4529 This change has negligible security impact because:
4531 1. Server and client random values still have 24 bytes of pseudo random
4534 2. Server and client random values are sent in the clear in the initial
4537 3. The master secret is derived using the premaster secret (48 bytes in
4538 size for static RSA ciphersuites) as well as client server and random
4541 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4544 [Stephen Henson, reported by UK NISCC]
4546 *) Use Windows randomness collection on Cygwin.
4549 *) Fix hang in EGD/PRNGD query when communication socket is closed
4550 prematurely by EGD/PRNGD.
4551 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4553 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4556 *) Back-port of selected performance improvements from development
4557 branch, as well as improved support for PowerPC platforms.
4560 *) Add lots of checks for memory allocation failure, error codes to indicate
4561 failure and freeing up memory if a failure occurs.
4562 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4564 *) Add new -passin argument to dgst.
4567 *) Perform some character comparisons of different types in X509_NAME_cmp:
4568 this is needed for some certificates that reencode DNs into UTF8Strings
4569 (in violation of RFC3280) and can't or wont issue name rollover
4573 *) Make an explicit check during certificate validation to see that
4574 the CA setting in each certificate on the chain is correct. As a
4575 side effect always do the following basic checks on extensions,
4576 not just when there's an associated purpose to the check:
4578 - if there is an unhandled critical extension (unless the user
4579 has chosen to ignore this fault)
4580 - if the path length has been exceeded (if one is set at all)
4581 - that certain extensions fit the associated purpose (if one has
4585 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4587 *) Avoid a race condition when CRLs are checked in a multi threaded
4588 environment. This would happen due to the reordering of the revoked
4589 entries during signature checking and serial number lookup. Now the
4590 encoding is cached and the serial number sort performed under a lock.
4591 Add new STACK function sk_is_sorted().
4594 *) Add Delta CRL to the extension code.
4597 *) Various fixes to s3_pkt.c so alerts are sent properly.
4598 [David Holmes <d.holmes@f5.com>]
4600 *) Reduce the chances of duplicate issuer name and serial numbers (in
4601 violation of RFC3280) using the OpenSSL certificate creation utilities.
4602 This is done by creating a random 64 bit value for the initial serial
4603 number when a serial number file is created or when a self signed
4604 certificate is created using 'openssl req -x509'. The initial serial
4605 number file is created using 'openssl x509 -next_serial' in CA.pl
4606 rather than being initialized to 1.
4609 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4611 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4612 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4613 [Joe Orton, Steve Henson]
4615 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4617 [Joe Orton, Steve Henson]
4619 *) Make it possible to have multiple active certificates with the same
4620 subject in the CA index file. This is done only if the keyword
4621 'unique_subject' is set to 'no' in the main CA section (default
4622 if 'CA_default') of the configuration file. The value is saved
4623 with the database itself in a separate index attribute file,
4624 named like the index file with '.attr' appended to the name.
4627 *) X509 verify fixes. Disable broken certificate workarounds when
4628 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4629 keyUsage extension present. Don't accept CRLs with unhandled critical
4630 extensions: since verify currently doesn't process CRL extensions this
4631 rejects a CRL with *any* critical extensions. Add new verify error codes
4635 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4636 A clarification of RFC2560 will require the use of OCTET STRINGs and
4637 some implementations cannot handle the current raw format. Since OpenSSL
4638 copies and compares OCSP nonces as opaque blobs without any attempt at
4639 parsing them this should not create any compatibility issues.
4642 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4643 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4644 this HMAC (and other) operations are several times slower than OpenSSL
4648 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4649 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4651 *) Use the correct content when signing type "other".
4654 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4656 *) Fix various bugs revealed by running the NISCC test suite:
4658 Stop out of bounds reads in the ASN1 code when presented with
4659 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4661 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4663 If verify callback ignores invalid public key errors don't try to check
4664 certificate signature with the NULL public key.
4668 *) New -ignore_err option in ocsp application to stop the server
4669 exiting on the first error in a request.
4672 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4673 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4677 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4678 extra data after the compression methods not only for TLS 1.0
4679 but also for SSL 3.0 (as required by the specification).
4680 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4682 *) Change X509_certificate_type() to mark the key as exported/exportable
4683 when it's 512 *bits* long, not 512 bytes.
4686 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4687 blocks during encryption.
4690 *) Various fixes to base64 BIO and non blocking I/O. On write
4691 flushes were not handled properly if the BIO retried. On read
4692 data was not being buffered properly and had various logic bugs.
4693 This also affects blocking I/O when the data being decoded is a
4697 *) Various S/MIME bugfixes and compatibility changes:
4698 output correct application/pkcs7 MIME type if
4699 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4700 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4701 of files as .eml work). Correctly handle very long lines in MIME
4705 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4707 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4708 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4709 a protocol version number mismatch like a decryption error
4710 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4713 *) Turn on RSA blinding by default in the default implementation
4714 to avoid a timing attack. Applications that don't want it can call
4715 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4716 They would be ill-advised to do so in most cases.
4717 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4719 *) Change RSA blinding code so that it works when the PRNG is not
4720 seeded (in this case, the secret RSA exponent is abused as
4721 an unpredictable seed -- if it is not unpredictable, there
4722 is no point in blinding anyway). Make RSA blinding thread-safe
4723 by remembering the creator's thread ID in rsa->blinding and
4724 having all other threads use local one-time blinding factors
4725 (this requires more computation than sharing rsa->blinding, but
4726 avoids excessive locking; and if an RSA object is not shared
4727 between threads, blinding will still be very fast).
4730 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4731 ENGINE as defaults for all supported algorithms irrespective of
4732 the 'flags' parameter. 'flags' is now honoured, so applications
4733 should make sure they are passing it correctly.
4736 *) Target "mingw" now allows native Windows code to be generated in
4737 the Cygwin environment as well as with the MinGW compiler.
4740 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4742 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4743 via timing by performing a MAC computation even if incorrect
4744 block cipher padding has been found. This is a countermeasure
4745 against active attacks where the attacker has to distinguish
4746 between bad padding and a MAC verification error. (CVE-2003-0078)
4748 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4749 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4750 Martin Vuagnoux (EPFL, Ilion)]
4752 *) Make the no-err option work as intended. The intention with no-err
4753 is not to have the whole error stack handling routines removed from
4754 libcrypto, it's only intended to remove all the function name and
4755 reason texts, thereby removing some of the footprint that may not
4756 be interesting if those errors aren't displayed anyway.
4758 NOTE: it's still possible for any application or module to have it's
4759 own set of error texts inserted. The routines are there, just not
4760 used by default when no-err is given.
4763 *) Add support for FreeBSD on IA64.
4764 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4766 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4767 Kerberos function mit_des_cbc_cksum(). Before this change,
4768 the value returned by DES_cbc_cksum() was like the one from
4769 mit_des_cbc_cksum(), except the bytes were swapped.
4770 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4772 *) Allow an application to disable the automatic SSL chain building.
4773 Before this a rather primitive chain build was always performed in
4774 ssl3_output_cert_chain(): an application had no way to send the
4775 correct chain if the automatic operation produced an incorrect result.
4777 Now the chain builder is disabled if either:
4779 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4781 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4783 The reasoning behind this is that an application would not want the
4784 auto chain building to take place if extra chain certificates are
4785 present and it might also want a means of sending no additional
4786 certificates (for example the chain has two certificates and the
4790 *) Add the possibility to build without the ENGINE framework.
4791 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4793 *) Under Win32 gmtime() can return NULL: check return value in
4794 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4797 *) DSA routines: under certain error conditions uninitialized BN objects
4798 could be freed. Solution: make sure initialization is performed early
4799 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4800 Nils Larsch <nla@trustcenter.de> via PR#459)
4803 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4804 checked on reconnect on the client side, therefore session resumption
4805 could still fail with a "ssl session id is different" error. This
4806 behaviour is masked when SSL_OP_ALL is used due to
4807 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4808 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4809 followup to PR #377.
4812 *) IA-32 assembler support enhancements: unified ELF targets, support
4813 for SCO/Caldera platforms, fix for Cygwin shared build.
4816 *) Add support for FreeBSD on sparc64. As a consequence, support for
4817 FreeBSD on non-x86 processors is separate from x86 processors on
4818 the config script, much like the NetBSD support.
4819 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4821 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4823 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4826 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4827 code (06) was taken as the first octet of the session ID and the last
4828 octet was ignored consequently. As a result SSLv2 client side session
4829 caching could not have worked due to the session ID mismatch between
4831 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4835 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4836 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4840 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4841 seems that in spite of existing for more than a year, many application
4842 author have done nothing to provide the necessary callbacks, which
4843 means that this particular engine will not work properly anywhere.
4844 This is a very unfortunate situation which forces us, in the name
4845 of usability, to give the hw_ncipher.c a static lock, which is part
4847 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4848 appear in 0.9.8 or later. We EXPECT application authors to have
4849 dealt properly with this when 0.9.8 is released (unless we actually
4850 make such changes in the libcrypto locking code that changes will
4851 have to be made anyway).
4854 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4855 octets have been read, EOF or an error occurs. Without this change
4856 some truncated ASN1 structures will not produce an error.
4859 *) Disable Heimdal support, since it hasn't been fully implemented.
4860 Still give the possibility to force the use of Heimdal, but with
4861 warnings and a request that patches get sent to openssl-dev.
4864 *) Add the VC-CE target, introduce the WINCE sysname, and add
4865 INSTALL.WCE and appropriate conditionals to make it build.
4866 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4868 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4869 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4870 edit numbers of the version.
4871 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4873 *) Introduce safe string copy and catenation functions
4874 (BUF_strlcpy() and BUF_strlcat()).
4875 [Ben Laurie (CHATS) and Richard Levitte]
4877 *) Avoid using fixed-size buffers for one-line DNs.
4878 [Ben Laurie (CHATS)]
4880 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4881 resizing buffers containing secrets, and use where appropriate.
4882 [Ben Laurie (CHATS)]
4884 *) Avoid using fixed size buffers for configuration file location.
4885 [Ben Laurie (CHATS)]
4887 *) Avoid filename truncation for various CA files.
4888 [Ben Laurie (CHATS)]
4890 *) Use sizeof in preference to magic numbers.
4891 [Ben Laurie (CHATS)]
4893 *) Avoid filename truncation in cert requests.
4894 [Ben Laurie (CHATS)]
4896 *) Add assertions to check for (supposedly impossible) buffer
4898 [Ben Laurie (CHATS)]
4900 *) Don't cache truncated DNS entries in the local cache (this could
4901 potentially lead to a spoofing attack).
4902 [Ben Laurie (CHATS)]
4904 *) Fix various buffers to be large enough for hex/decimal
4905 representations in a platform independent manner.
4906 [Ben Laurie (CHATS)]
4908 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4909 resizing buffers containing secrets, and use where appropriate.
4910 [Ben Laurie (CHATS)]
4912 *) Add BIO_indent() to avoid much slightly worrying code to do
4914 [Ben Laurie (CHATS)]
4916 *) Convert sprintf()/BIO_puts() to BIO_printf().
4917 [Ben Laurie (CHATS)]
4919 *) buffer_gets() could terminate with the buffer only half
4921 [Ben Laurie (CHATS)]
4923 *) Add assertions to prevent user-supplied crypto functions from
4924 overflowing internal buffers by having large block sizes, etc.
4925 [Ben Laurie (CHATS)]
4927 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4929 [Ben Laurie (CHATS)]
4931 *) Eliminate unused copy of key in RC4.
4932 [Ben Laurie (CHATS)]
4934 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4935 [Ben Laurie (CHATS)]
4937 *) Fix off-by-one error in EGD path.
4938 [Ben Laurie (CHATS)]
4940 *) If RANDFILE path is too long, ignore instead of truncating.
4941 [Ben Laurie (CHATS)]
4943 *) Eliminate unused and incorrectly sized X.509 structure
4945 [Ben Laurie (CHATS)]
4947 *) Eliminate unused and dangerous function knumber().
4948 [Ben Laurie (CHATS)]
4950 *) Eliminate unused and dangerous structure, KSSL_ERR.
4951 [Ben Laurie (CHATS)]
4953 *) Protect against overlong session ID context length in an encoded
4954 session object. Since these are local, this does not appear to be
4956 [Ben Laurie (CHATS)]
4958 *) Change from security patch (see 0.9.6e below) that did not affect
4959 the 0.9.6 release series:
4961 Remote buffer overflow in SSL3 protocol - an attacker could
4962 supply an oversized master key in Kerberos-enabled versions.
4964 [Ben Laurie (CHATS)]
4966 *) Change the SSL kerb5 codes to match RFC 2712.
4969 *) Make -nameopt work fully for req and add -reqopt switch.
4970 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4972 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4973 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4975 *) Make sure tests can be performed even if the corresponding algorithms
4976 have been removed entirely. This was also the last step to make
4977 OpenSSL compilable with DJGPP under all reasonable conditions.
4978 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4980 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4981 to allow version independent disabling of normally unselected ciphers,
4982 which may be activated as a side-effect of selecting a single cipher.
4984 (E.g., cipher list string "RSA" enables ciphersuites that are left
4985 out of "ALL" because they do not provide symmetric encryption.
4986 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4987 [Lutz Jaenicke, Bodo Moeller]
4989 *) Add appropriate support for separate platform-dependent build
4990 directories. The recommended way to make a platform-dependent
4991 build directory is the following (tested on Linux), maybe with
4994 # Place yourself outside of the OpenSSL source tree. In
4995 # this example, the environment variable OPENSSL_SOURCE
4996 # is assumed to contain the absolute OpenSSL source directory.
4997 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4998 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4999 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5000 mkdir -p `dirname $F`
5001 ln -s $OPENSSL_SOURCE/$F $F
5004 To be absolutely sure not to disturb the source tree, a "make clean"
5005 is a good thing. If it isn't successfull, don't worry about it,
5006 it probably means the source directory is very clean.
5009 *) Make sure any ENGINE control commands make local copies of string
5010 pointers passed to them whenever necessary. Otherwise it is possible
5011 the caller may have overwritten (or deallocated) the original string
5012 data when a later ENGINE operation tries to use the stored values.
5013 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5015 *) Improve diagnostics in file reading and command-line digests.
5016 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5018 *) Add AES modes CFB and OFB to the object database. Correct an
5019 error in AES-CFB decryption.
5022 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5023 allows existing EVP_CIPHER_CTX structures to be reused after
5024 calling EVP_*Final(). This behaviour is used by encryption
5025 BIOs and some applications. This has the side effect that
5026 applications must explicitly clean up cipher contexts with
5027 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5030 *) Check the values of dna and dnb in bn_mul_recursive before calling
5031 bn_mul_comba (a non zero value means the a or b arrays do not contain
5032 n2 elements) and fallback to bn_mul_normal if either is not zero.
5035 *) Fix escaping of non-ASCII characters when using the -subj option
5036 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5039 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5040 form for "surname", serialNumber has no short form.
5041 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5042 therefore remove "mail" short name for "internet 7".
5043 The OID for unique identifiers in X509 certificates is
5044 x500UniqueIdentifier, not uniqueIdentifier.
5045 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5048 *) Add an "init" command to the ENGINE config module and auto initialize
5049 ENGINEs. Without any "init" command the ENGINE will be initialized
5050 after all ctrl commands have been executed on it. If init=1 the
5051 ENGINE is initailized at that point (ctrls before that point are run
5052 on the uninitialized ENGINE and after on the initialized one). If
5053 init=0 then the ENGINE will not be iniatialized at all.
5056 *) Fix the 'app_verify_callback' interface so that the user-defined
5057 argument is actually passed to the callback: In the
5058 SSL_CTX_set_cert_verify_callback() prototype, the callback
5059 declaration has been changed from
5062 int (*cb)(X509_STORE_CTX *,void *);
5063 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5064 i=s->ctx->app_verify_callback(&ctx)
5065 has been changed into
5066 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5068 To update applications using SSL_CTX_set_cert_verify_callback(),
5069 a dummy argument can be added to their callback functions.
5070 [D. K. Smetters <smetters@parc.xerox.com>]
5072 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5073 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5075 *) Add and OPENSSL_LOAD_CONF define which will cause
5076 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5077 This allows older applications to transparently support certain
5078 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5079 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5080 load the config file and OPENSSL_add_all_algorithms_conf() which will
5081 always load it have also been added.
5084 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5085 Adjust NIDs and EVP layer.
5086 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5088 *) Config modules support in openssl utility.
5090 Most commands now load modules from the config file,
5091 though in a few (such as version) this isn't done
5092 because it couldn't be used for anything.
5094 In the case of ca and req the config file used is
5095 the same as the utility itself: that is the -config
5096 command line option can be used to specify an
5100 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5101 use "openssl_conf" if filename is NULL use default openssl config file.
5104 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5105 config section name. Add a new flag to tolerate a missing config file
5106 and move code to CONF_modules_load_file().
5109 *) Support for crypto accelerator cards from Accelerated Encryption
5110 Processing, www.aep.ie. (Use engine 'aep')
5111 The support was copied from 0.9.6c [engine] and adapted/corrected
5112 to work with the new engine framework.
5113 [AEP Inc. and Richard Levitte]
5115 *) Support for SureWare crypto accelerator cards from Baltimore
5116 Technologies. (Use engine 'sureware')
5117 The support was copied from 0.9.6c [engine] and adapted
5118 to work with the new engine framework.
5121 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5122 make the newer ENGINE framework commands for the CHIL engine work.
5123 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5125 *) Make it possible to produce shared libraries on ReliantUNIX.
5126 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5128 *) Add the configuration target debug-linux-ppro.
5129 Make 'openssl rsa' use the general key loading routines
5130 implemented in apps.c, and make those routines able to
5131 handle the key format FORMAT_NETSCAPE and the variant
5133 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5135 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5136 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5138 *) Add -keyform to rsautl, and document -engine.
5139 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5141 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5142 BIO_R_NO_SUCH_FILE error code rather than the generic
5143 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5146 *) Add new functions
5148 ERR_peek_last_error_line
5149 ERR_peek_last_error_line_data.
5150 These are similar to
5153 ERR_peek_error_line_data,
5154 but report on the latest error recorded rather than the first one
5155 still in the error queue.
5156 [Ben Laurie, Bodo Moeller]
5158 *) default_algorithms option in ENGINE config module. This allows things
5160 default_algorithms = ALL
5161 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5164 *) Preliminary ENGINE config module.
5167 *) New experimental application configuration code.
5170 *) Change the AES code to follow the same name structure as all other
5171 symmetric ciphers, and behave the same way. Move everything to
5172 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5173 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5175 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5176 [Ben Laurie and Theo de Raadt]
5178 *) Add option to output public keys in req command.
5179 [Massimiliano Pala madwolf@openca.org]
5181 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5182 (up to about 10% better than before for P-192 and P-224).
5185 *) New functions/macros
5187 SSL_CTX_set_msg_callback(ctx, cb)
5188 SSL_CTX_set_msg_callback_arg(ctx, arg)
5189 SSL_set_msg_callback(ssl, cb)
5190 SSL_set_msg_callback_arg(ssl, arg)
5192 to request calling a callback function
5194 void cb(int write_p, int version, int content_type,
5195 const void *buf, size_t len, SSL *ssl, void *arg)
5197 whenever a protocol message has been completely received
5198 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5199 protocol version according to which the SSL library interprets
5200 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5201 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5202 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5203 specification (change_cipher_spec(20), alert(21), handshake(22)).
5204 'buf' and 'len' point to the actual message, 'ssl' to the
5205 SSL object, and 'arg' is the application-defined value set by
5206 SSL[_CTX]_set_msg_callback_arg().
5208 'openssl s_client' and 'openssl s_server' have new '-msg' options
5209 to enable a callback that displays all protocol messages.
5212 *) Change the shared library support so shared libraries are built as
5213 soon as the corresponding static library is finished, and thereby get
5214 openssl and the test programs linked against the shared library.
5215 This still only happens when the keyword "shard" has been given to
5216 the configuration scripts.
5218 NOTE: shared library support is still an experimental thing, and
5219 backward binary compatibility is still not guaranteed.
5220 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5222 *) Add support for Subject Information Access extension.
5223 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5225 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5226 additional bytes when new memory had to be allocated, not just
5227 when reusing an existing buffer.
5230 *) New command line and configuration option 'utf8' for the req command.
5231 This allows field values to be specified as UTF8 strings.
5234 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5235 runs for the former and machine-readable output for the latter.
5238 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5239 of the e-mail address in the DN (i.e., it will go into a certificate
5240 extension only). The new configuration file option 'email_in_dn = no'
5241 has the same effect.
5242 [Massimiliano Pala madwolf@openca.org]
5244 *) Change all functions with names starting with des_ to be starting
5245 with DES_ instead. Add wrappers that are compatible with libdes,
5246 but are named _ossl_old_des_*. Finally, add macros that map the
5247 des_* symbols to the corresponding _ossl_old_des_* if libdes
5248 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5249 desired, the des_* symbols will be mapped to DES_*, with one
5252 Since we provide two compatibility mappings, the user needs to
5253 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5254 compatibility is desired. The default (i.e., when that macro
5255 isn't defined) is OpenSSL 0.9.6c compatibility.
5257 There are also macros that enable and disable the support of old
5258 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5259 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5260 are defined, the default will apply: to support the old des routines.
5262 In either case, one must include openssl/des.h to get the correct
5263 definitions. Do not try to just include openssl/des_old.h, that
5266 NOTE: This is a major break of an old API into a new one. Software
5267 authors are encouraged to switch to the DES_ style functions. Some
5268 time in the future, des_old.h and the libdes compatibility functions
5269 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5270 default), and then completely removed.
5273 *) Test for certificates which contain unsupported critical extensions.
5274 If such a certificate is found during a verify operation it is
5275 rejected by default: this behaviour can be overridden by either
5276 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5277 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5278 X509_supported_extension() has also been added which returns 1 if a
5279 particular extension is supported.
5282 *) Modify the behaviour of EVP cipher functions in similar way to digests
5283 to retain compatibility with existing code.
5286 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5287 compatibility with existing code. In particular the 'ctx' parameter does
5288 not have to be to be initialized before the call to EVP_DigestInit() and
5289 it is tidied up after a call to EVP_DigestFinal(). New function
5290 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5291 EVP_MD_CTX_copy() changed to not require the destination to be
5292 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5293 requires the destination to be valid.
5295 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5296 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5299 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5300 so that complete 'Handshake' protocol structures are kept in memory
5301 instead of overwriting 'msg_type' and 'length' with 'body' data.
5304 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5305 [Massimo Santin via Richard Levitte]
5307 *) Major restructuring to the underlying ENGINE code. This includes
5308 reduction of linker bloat, separation of pure "ENGINE" manipulation
5309 (initialisation, etc) from functionality dealing with implementations
5310 of specific crypto iterfaces. This change also introduces integrated
5311 support for symmetric ciphers and digest implementations - so ENGINEs
5312 can now accelerate these by providing EVP_CIPHER and EVP_MD
5313 implementations of their own. This is detailed in crypto/engine/README
5314 as it couldn't be adequately described here. However, there are a few
5315 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5316 were changed in the original introduction of ENGINE code have now
5317 reverted back - the hooking from this code to ENGINE is now a good
5318 deal more passive and at run-time, operations deal directly with
5319 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5320 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5321 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5322 they were not being used by the framework as there is no concept of a
5323 BIGNUM_METHOD and they could not be generalised to the new
5324 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5325 ENGINE_cpy() has been removed as it cannot be consistently defined in
5329 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5332 *) Change mkdef.pl to sort symbols that get the same entry number,
5333 and make sure the automatically generated functions ERR_load_*
5334 become part of libeay.num as well.
5337 *) New function SSL_renegotiate_pending(). This returns true once
5338 renegotiation has been requested (either SSL_renegotiate() call
5339 or HelloRequest/ClientHello received from the peer) and becomes
5340 false once a handshake has been completed.
5341 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5342 sends a HelloRequest, but does not ensure that a handshake takes
5343 place. SSL_renegotiate_pending() is useful for checking if the
5344 client has followed the request.)
5347 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5348 By default, clients may request session resumption even during
5349 renegotiation (if session ID contexts permit); with this option,
5350 session resumption is possible only in the first handshake.
5352 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5353 more bits available for options that should not be part of
5354 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5357 *) Add some demos for certificate and certificate request creation.
5360 *) Make maximum certificate chain size accepted from the peer application
5361 settable (SSL*_get/set_max_cert_list()), as proposed by
5362 "Douglas E. Engert" <deengert@anl.gov>.
5365 *) Add support for shared libraries for Unixware-7
5366 (Boyd Lynn Gerber <gerberb@zenez.com>).
5369 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5370 be done prior to destruction. Use this to unload error strings from
5371 ENGINEs that load their own error strings. NB: This adds two new API
5372 functions to "get" and "set" this destroy handler in an ENGINE.
5375 *) Alter all existing ENGINE implementations (except "openssl" and
5376 "openbsd") to dynamically instantiate their own error strings. This
5377 makes them more flexible to be built both as statically-linked ENGINEs
5378 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5379 Also, add stub code to each that makes building them as self-contained
5380 shared-libraries easier (see README.ENGINE).
5383 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5384 implementations into applications that are completely implemented in
5385 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5386 commands that can be used to configure what shared-library to load and
5387 to control aspects of the way it is handled. Also, made an update to
5388 the README.ENGINE file that brings its information up-to-date and
5389 provides some information and instructions on the "dynamic" ENGINE
5390 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5393 *) Make it possible to unload ranges of ERR strings with a new
5394 "ERR_unload_strings" function.
5397 *) Add a copy() function to EVP_MD.
5400 *) Make EVP_MD routines take a context pointer instead of just the
5401 md_data void pointer.
5404 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5405 that the digest can only process a single chunk of data
5406 (typically because it is provided by a piece of
5407 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5408 is only going to provide a single chunk of data, and hence the
5409 framework needn't accumulate the data for oneshot drivers.
5412 *) As with "ERR", make it possible to replace the underlying "ex_data"
5413 functions. This change also alters the storage and management of global
5414 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5415 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5416 index counters. The API functions that use this state have been changed
5417 to take a "class_index" rather than pointers to the class's local STACK
5418 and counter, and there is now an API function to dynamically create new
5419 classes. This centralisation allows us to (a) plug a lot of the
5420 thread-safety problems that existed, and (b) makes it possible to clean
5421 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5422 such data would previously have always leaked in application code and
5423 workarounds were in place to make the memory debugging turn a blind eye
5424 to it. Application code that doesn't use this new function will still
5425 leak as before, but their memory debugging output will announce it now
5426 rather than letting it slide.
5428 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5429 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5430 has a return value to indicate success or failure.
5433 *) Make it possible to replace the underlying "ERR" functions such that the
5434 global state (2 LHASH tables and 2 locks) is only used by the "default"
5435 implementation. This change also adds two functions to "get" and "set"
5436 the implementation prior to it being automatically set the first time
5437 any other ERR function takes place. Ie. an application can call "get",
5438 pass the return value to a module it has just loaded, and that module
5439 can call its own "set" function using that value. This means the
5440 module's "ERR" operations will use (and modify) the error state in the
5441 application and not in its own statically linked copy of OpenSSL code.
5444 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5445 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5446 the operation, and provides a more encapsulated way for external code
5447 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5448 to use these functions rather than manually incrementing the counts.
5450 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5453 *) Add EVP test program.
5456 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5459 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5460 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5461 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5462 These allow a CRL to be built without having to access X509_CRL fields
5463 directly. Modify 'ca' application to use new functions.
5466 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5467 bug workarounds. Rollback attack detection is a security feature.
5468 The problem will only arise on OpenSSL servers when TLSv1 is not
5469 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5470 Software authors not wanting to support TLSv1 will have special reasons
5471 for their choice and can explicitly enable this option.
5472 [Bodo Moeller, Lutz Jaenicke]
5474 *) Rationalise EVP so it can be extended: don't include a union of
5475 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5476 (similar to those existing for EVP_CIPHER_CTX).
5481 EVP_MD_CTX_init(&md); /* new function call */
5482 EVP_DigestInit(&md, EVP_sha1());
5483 EVP_DigestUpdate(&md, in, len);
5484 EVP_DigestFinal(&md, out, NULL);
5485 EVP_MD_CTX_cleanup(&md); /* new function call */
5489 *) Make DES key schedule conform to the usual scheme, as well as
5490 correcting its structure. This means that calls to DES functions
5491 now have to pass a pointer to a des_key_schedule instead of a
5492 plain des_key_schedule (which was actually always a pointer
5495 des_key_schedule ks;
5497 des_set_key_checked(..., &ks);
5498 des_ncbc_encrypt(..., &ks, ...);
5500 (Note that a later change renames 'des_...' into 'DES_...'.)
5503 *) Initial reduction of linker bloat: the use of some functions, such as
5504 PEM causes large amounts of unused functions to be linked in due to
5505 poor organisation. For example pem_all.c contains every PEM function
5506 which has a knock on effect of linking in large amounts of (unused)
5507 ASN1 code. Grouping together similar functions and splitting unrelated
5508 functions prevents this.
5511 *) Cleanup of EVP macros.
5514 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5515 correct _ecb suffix.
5518 *) Add initial OCSP responder support to ocsp application. The
5519 revocation information is handled using the text based index
5520 use by the ca application. The responder can either handle
5521 requests generated internally, supplied in files (for example
5522 via a CGI script) or using an internal minimal server.
5525 *) Add configuration choices to get zlib compression for TLS.
5528 *) Changes to Kerberos SSL for RFC 2712 compliance:
5529 1. Implemented real KerberosWrapper, instead of just using
5530 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5531 2. Implemented optional authenticator field of KerberosWrapper.
5533 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5534 and authenticator structs; see crypto/krb5/.
5536 Generalized Kerberos calls to support multiple Kerberos libraries.
5537 [Vern Staats <staatsvr@asc.hpc.mil>,
5538 Jeffrey Altman <jaltman@columbia.edu>
5539 via Richard Levitte]
5541 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5542 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5543 values for each of the key sizes rather than having just
5544 parameters (and 'speed' generating keys each time).
5547 *) Speed up EVP routines.
5550 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5551 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5552 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5553 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5555 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5556 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5557 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5560 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5562 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5565 *) Added the OS2-EMX target.
5566 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5568 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5569 to support NCONF routines in extension code. New function CONF_set_nconf()
5570 to allow functions which take an NCONF to also handle the old LHASH
5571 structure: this means that the old CONF compatible routines can be
5572 retained (in particular wrt extensions) without having to duplicate the
5573 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5576 *) Enhance the general user interface with mechanisms for inner control
5577 and with possibilities to have yes/no kind of prompts.
5580 *) Change all calls to low level digest routines in the library and
5581 applications to use EVP. Add missing calls to HMAC_cleanup() and
5582 don't assume HMAC_CTX can be copied using memcpy().
5583 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5585 *) Add the possibility to control engines through control names but with
5586 arbitrary arguments instead of just a string.
5587 Change the key loaders to take a UI_METHOD instead of a callback
5588 function pointer. NOTE: this breaks binary compatibility with earlier
5589 versions of OpenSSL [engine].
5590 Adapt the nCipher code for these new conditions and add a card insertion
5594 *) Enhance the general user interface with mechanisms to better support
5595 dialog box interfaces, application-defined prompts, the possibility
5596 to use defaults (for example default passwords from somewhere else)
5597 and interrupts/cancellations.
5600 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5601 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5604 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5605 tidy up some unnecessarily weird code in 'sk_new()').
5606 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5608 *) Change the key loading routines for ENGINEs to use the same kind
5609 callback (pem_password_cb) as all other routines that need this
5613 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5614 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5615 than this minimum value is recommended.
5618 *) New random seeder for OpenVMS, using the system process statistics
5619 that are easily reachable.
5622 *) Windows apparently can't transparently handle global
5623 variables defined in DLLs. Initialisations such as:
5625 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5627 wont compile. This is used by the any applications that need to
5628 declare their own ASN1 modules. This was fixed by adding the option
5629 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5630 needed for static libraries under Win32.
5633 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5634 setting of purpose and trust fields. New X509_STORE trust and
5635 purpose functions and tidy up setting in other SSL functions.
5638 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5639 structure. These are inherited by X509_STORE_CTX when it is
5640 initialised. This allows various defaults to be set in the
5641 X509_STORE structure (such as flags for CRL checking and custom
5642 purpose or trust settings) for functions which only use X509_STORE_CTX
5643 internally such as S/MIME.
5645 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5646 trust settings if they are not set in X509_STORE. This allows X509_STORE
5647 purposes and trust (in S/MIME for example) to override any set by default.
5649 Add command line options for CRL checking to smime, s_client and s_server
5653 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5654 are set then the CRL is looked up in the X509_STORE structure and
5655 its validity and signature checked, then if the certificate is found
5656 in the CRL the verify fails with a revoked error.
5658 Various new CRL related callbacks added to X509_STORE_CTX structure.
5660 Command line options added to 'verify' application to support this.
5662 This needs some additional work, such as being able to handle multiple
5663 CRLs with different times, extension based lookup (rather than just
5664 by subject name) and ultimately more complete V2 CRL extension
5668 *) Add a general user interface API (crypto/ui/). This is designed
5669 to replace things like des_read_password and friends (backward
5670 compatibility functions using this new API are provided).
5671 The purpose is to remove prompting functions from the DES code
5672 section as well as provide for prompting through dialog boxes in
5673 a window system and the like.
5676 *) Add "ex_data" support to ENGINE so implementations can add state at a
5677 per-structure level rather than having to store it globally.
5680 *) Make it possible for ENGINE structures to be copied when retrieved by
5681 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5682 This causes the "original" ENGINE structure to act like a template,
5683 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5684 operational state can be localised to each ENGINE structure, despite the
5685 fact they all share the same "methods". New ENGINE structures returned in
5686 this case have no functional references and the return value is the single
5687 structural reference. This matches the single structural reference returned
5688 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5692 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5693 needs to match any other type at all we need to manually clear the
5697 *) Changes to the "openssl engine" utility to include;
5698 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5699 about an ENGINE's available control commands.
5700 - executing control commands from command line arguments using the
5701 '-pre' and '-post' switches. '-post' is only used if '-t' is
5702 specified and the ENGINE is successfully initialised. The syntax for
5703 the individual commands are colon-separated, for example;
5704 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5707 *) New dynamic control command support for ENGINEs. ENGINEs can now
5708 declare their own commands (numbers), names (strings), descriptions,
5709 and input types for run-time discovery by calling applications. A
5710 subset of these commands are implicitly classed as "executable"
5711 depending on their input type, and only these can be invoked through
5712 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5713 can be based on user input, config files, etc). The distinction is
5714 that "executable" commands cannot return anything other than a boolean
5715 result and can only support numeric or string input, whereas some
5716 discoverable commands may only be for direct use through
5717 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5718 pointers, or other custom uses. The "executable" commands are to
5719 support parameterisations of ENGINE behaviour that can be
5720 unambiguously defined by ENGINEs and used consistently across any
5721 OpenSSL-based application. Commands have been added to all the
5722 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5723 control over shared-library paths without source code alterations.
5726 *) Changed all ENGINE implementations to dynamically allocate their
5727 ENGINEs rather than declaring them statically. Apart from this being
5728 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5729 this also allows the implementations to compile without using the
5730 internal engine_int.h header.
5733 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5734 'const' value. Any code that should be able to modify a RAND_METHOD
5735 should already have non-const pointers to it (ie. they should only
5736 modify their own ones).
5739 *) Made a variety of little tweaks to the ENGINE code.
5740 - "atalla" and "ubsec" string definitions were moved from header files
5741 to C code. "nuron" string definitions were placed in variables
5742 rather than hard-coded - allowing parameterisation of these values
5743 later on via ctrl() commands.
5744 - Removed unused "#if 0"'d code.
5745 - Fixed engine list iteration code so it uses ENGINE_free() to release
5746 structural references.
5747 - Constified the RAND_METHOD element of ENGINE structures.
5748 - Constified various get/set functions as appropriate and added
5749 missing functions (including a catch-all ENGINE_cpy that duplicates
5750 all ENGINE values onto a new ENGINE except reference counts/state).
5751 - Removed NULL parameter checks in get/set functions. Setting a method
5752 or function to NULL is a way of cancelling out a previously set
5753 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5754 and doesn't justify the extra error symbols and code.
5755 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5756 flags from engine_int.h to engine.h.
5757 - Changed prototypes for ENGINE handler functions (init(), finish(),
5758 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5761 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5762 to the algorithm using long division. The binary algorithm can be
5763 used only if the modulus is odd. On 32-bit systems, it is faster
5764 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5765 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5766 up to 450 bits. In 64-bit environments, the binary algorithm
5767 appears to be advantageous for much longer moduli; here we use it
5768 for moduli up to 2048 bits.
5771 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5772 could not support the combine flag in choice fields.
5775 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5776 extensions from a certificate request to the certificate.
5779 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5780 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5781 file: this allows the display of the certificate about to be
5782 signed to be customised, to allow certain fields to be included
5783 or excluded and extension details. The old system didn't display
5784 multicharacter strings properly, omitted fields not in the policy
5785 and couldn't display additional details such as extensions.
5788 *) Function EC_POINTs_mul for multiple scalar multiplication
5789 of an arbitrary number of elliptic curve points
5790 \sum scalars[i]*points[i],
5791 optionally including the generator defined for the EC_GROUP:
5792 scalar*generator + \sum scalars[i]*points[i].
5794 EC_POINT_mul is a simple wrapper function for the typical case
5795 that the point list has just one item (besides the optional
5799 *) First EC_METHODs for curves over GF(p):
5801 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5802 operations and provides various method functions that can also
5803 operate with faster implementations of modular arithmetic.
5805 EC_GFp_mont_method() reuses most functions that are part of
5806 EC_GFp_simple_method, but uses Montgomery arithmetic.
5808 [Bodo Moeller; point addition and point doubling
5809 implementation directly derived from source code provided by
5810 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5812 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5813 crypto/ec/ec_lib.c):
5815 Curves are EC_GROUP objects (with an optional group generator)
5816 based on EC_METHODs that are built into the library.
5818 Points are EC_POINT objects based on EC_GROUP objects.
5820 Most of the framework would be able to handle curves over arbitrary
5821 finite fields, but as there are no obvious types for fields other
5822 than GF(p), some functions are limited to that for now.
5825 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5826 that the file contains a complete HTTP response.
5829 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5830 change the def and num file printf format specifier from "%-40sXXX"
5831 to "%-39s XXX". The latter will always guarantee a space after the
5832 field while the former will cause them to run together if the field
5833 is 40 of more characters long.
5836 *) Constify the cipher and digest 'method' functions and structures
5837 and modify related functions to take constant EVP_MD and EVP_CIPHER
5841 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5842 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5845 *) Modify EVP_Digest*() routines so they now return values. Although the
5846 internal software routines can never fail additional hardware versions
5850 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5852 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5853 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5856 ERR_R_NESTED_ASN1_ERROR
5858 ERR_R_MISSING_ASN1_EOS
5859 were 4 .. 9, conflicting with
5860 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5862 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5863 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5865 Add new error code 'ERR_R_INTERNAL_ERROR'.
5868 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5872 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5873 sets the subject name for a new request or supersedes the
5874 subject name in a given request. Formats that can be parsed are
5875 'CN=Some Name, OU=myOU, C=IT'
5877 'CN=Some Name/OU=myOU/C=IT'.
5879 Add options '-batch' and '-verbose' to 'openssl req'.
5880 [Massimiliano Pala <madwolf@hackmasters.net>]
5882 *) Introduce the possibility to access global variables through
5883 functions on platform were that's the best way to handle exporting
5884 global variables in shared libraries. To enable this functionality,
5885 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5886 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5887 is normally done by Configure or something similar).
5889 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5890 in the source file (foo.c) like this:
5892 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5893 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5895 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5896 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5898 OPENSSL_DECLARE_GLOBAL(int,foo);
5899 #define foo OPENSSL_GLOBAL_REF(foo)
5900 OPENSSL_DECLARE_GLOBAL(double,bar);
5901 #define bar OPENSSL_GLOBAL_REF(bar)
5903 The #defines are very important, and therefore so is including the
5904 header file everywhere where the defined globals are used.
5906 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5907 of ASN.1 items, but that structure is a bit different.
5909 The largest change is in util/mkdef.pl which has been enhanced with
5910 better and easier to understand logic to choose which symbols should
5911 go into the Windows .def files as well as a number of fixes and code
5912 cleanup (among others, algorithm keywords are now sorted
5913 lexicographically to avoid constant rewrites).
5916 *) In BN_div() keep a copy of the sign of 'num' before writing the
5917 result to 'rm' because if rm==num the value will be overwritten
5918 and produce the wrong result if 'num' is negative: this caused
5919 problems with BN_mod() and BN_nnmod().
5922 *) Function OCSP_request_verify(). This checks the signature on an
5923 OCSP request and verifies the signer certificate. The signer
5924 certificate is just checked for a generic purpose and OCSP request
5928 *) Add OCSP_check_validity() function to check the validity of OCSP
5929 responses. OCSP responses are prepared in real time and may only
5930 be a few seconds old. Simply checking that the current time lies
5931 between thisUpdate and nextUpdate max reject otherwise valid responses
5932 caused by either OCSP responder or client clock inaccuracy. Instead
5933 we allow thisUpdate and nextUpdate to fall within a certain period of
5934 the current time. The age of the response can also optionally be
5935 checked. Two new options -validity_period and -status_age added to
5939 *) If signature or public key algorithm is unrecognized print out its
5940 OID rather that just UNKNOWN.
5943 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5944 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5945 ID to be generated from the issuer certificate alone which can then be
5946 passed to OCSP_id_issuer_cmp().
5949 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5950 ASN1 modules to export functions returning ASN1_ITEM pointers
5951 instead of the ASN1_ITEM structures themselves. This adds several
5952 new macros which allow the underlying ASN1 function/structure to
5953 be accessed transparently. As a result code should not use ASN1_ITEM
5954 references directly (such as &X509_it) but instead use the relevant
5955 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5956 use of the new ASN1 code on platforms where exporting structures
5957 is problematical (for example in shared libraries) but exporting
5958 functions returning pointers to structures is not.
5961 *) Add support for overriding the generation of SSL/TLS session IDs.
5962 These callbacks can be registered either in an SSL_CTX or per SSL.
5963 The purpose of this is to allow applications to control, if they wish,
5964 the arbitrary values chosen for use as session IDs, particularly as it
5965 can be useful for session caching in multiple-server environments. A
5966 command-line switch for testing this (and any client code that wishes
5967 to use such a feature) has been added to "s_server".
5968 [Geoff Thorpe, Lutz Jaenicke]
5970 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5971 of the form '#if defined(...) || defined(...) || ...' and
5972 '#if !defined(...) && !defined(...) && ...'. This also avoids
5973 the growing number of special cases it was previously handling.
5976 *) Make all configuration macros available for application by making
5977 sure they are available in opensslconf.h, by giving them names starting
5978 with "OPENSSL_" to avoid conflicts with other packages and by making
5979 sure e_os2.h will cover all platform-specific cases together with
5981 Additionally, it is now possible to define configuration/platform-
5982 specific names (called "system identities"). In the C code, these
5983 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5984 macro with the name beginning with "OPENSSL_SYS_", which is determined
5985 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5989 *) New option -set_serial to 'req' and 'x509' this allows the serial
5990 number to use to be specified on the command line. Previously self
5991 signed certificates were hard coded with serial number 0 and the
5992 CA options of 'x509' had to use a serial number in a file which was
5996 *) New options to 'ca' utility to support V2 CRL entry extensions.
5997 Currently CRL reason, invalidity date and hold instruction are
5998 supported. Add new CRL extensions to V3 code and some new objects.
6001 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6002 disable standard block padding (aka PKCS#5 padding) in the EVP
6003 API, which was previously mandatory. This means that the data is
6004 not padded in any way and so the total length much be a multiple
6005 of the block size, otherwise an error occurs.
6008 *) Initial (incomplete) OCSP SSL support.
6011 *) New function OCSP_parse_url(). This splits up a URL into its host,
6012 port and path components: primarily to parse OCSP URLs. New -url
6013 option to ocsp utility.
6016 *) New nonce behavior. The return value of OCSP_check_nonce() now
6017 reflects the various checks performed. Applications can decide
6018 whether to tolerate certain situations such as an absent nonce
6019 in a response when one was present in a request: the ocsp application
6020 just prints out a warning. New function OCSP_add1_basic_nonce()
6021 this is to allow responders to include a nonce in a response even if
6022 the request is nonce-less.
6025 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6026 skipped when using openssl x509 multiple times on a single input file,
6027 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6030 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6031 set string type: to handle setting ASN1_TIME structures. Fix ca
6032 utility to correctly initialize revocation date of CRLs.
6035 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6036 the clients preferred ciphersuites and rather use its own preferences.
6037 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6038 Internet Explorer by ensuring unchanged hash method during stepup.
6039 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6042 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6043 to aes and add a new 'exist' option to print out symbols that don't
6047 *) Additional options to ocsp utility to allow flags to be set and
6048 additional certificates supplied.
6051 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6052 OCSP client a number of certificate to only verify the response
6056 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6057 handle the new API. Currently only ECB, CBC modes supported. Add new
6060 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6061 Encryption Standard (AES) Ciphersuites for Transport Layer
6062 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6063 not enabled by default and were not part of the "ALL" ciphersuite
6064 alias because they were not yet official; they could be
6065 explicitly requested by specifying the "AESdraft" ciphersuite
6066 group alias. In the final release of OpenSSL 0.9.7, the group
6067 alias is called "AES" and is part of "ALL".)
6068 [Ben Laurie, Steve Henson, Bodo Moeller]
6070 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6071 request to response.
6074 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6075 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6076 extract information from a certificate request. OCSP_response_create()
6077 creates a response and optionally adds a basic response structure.
6078 OCSP_basic_add1_status() adds a complete single response to a basic
6079 response and returns the OCSP_SINGLERESP structure just added (to allow
6080 extensions to be included for example). OCSP_basic_add1_cert() adds a
6081 certificate to a basic response and OCSP_basic_sign() signs a basic
6082 response with various flags. New helper functions ASN1_TIME_check()
6083 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6084 (converts ASN1_TIME to GeneralizedTime).
6087 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6088 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6089 structure from a certificate. X509_pubkey_digest() digests the public_key
6090 contents: this is used in various key identifiers.
6093 *) Make sk_sort() tolerate a NULL argument.
6094 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6096 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6097 passed by the function are trusted implicitly. If any of them signed the
6098 response then it is assumed to be valid and is not verified.
6101 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6102 to data. This was previously part of the PKCS7 ASN1 code. This
6103 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6104 [Steve Henson, reported by Kenneth R. Robinette
6105 <support@securenetterm.com>]
6107 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6108 routines: without these tracing memory leaks is very painful.
6109 Fix leaks in PKCS12 and PKCS7 routines.
6112 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6113 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6114 effectively meant GeneralizedTime would never be used. Now it
6115 is initialised to -1 but X509_time_adj() now has to check the value
6116 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6117 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6118 [Steve Henson, reported by Kenneth R. Robinette
6119 <support@securenetterm.com>]
6121 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6122 result in a zero length in the ASN1_INTEGER structure which was
6123 not consistent with the structure when d2i_ASN1_INTEGER() was used
6124 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6125 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6126 where it did not print out a minus for negative ASN1_INTEGER.
6129 *) Add summary printout to ocsp utility. The various functions which
6130 convert status values to strings have been renamed to:
6131 OCSP_response_status_str(), OCSP_cert_status_str() and
6132 OCSP_crl_reason_str() and are no longer static. New options
6133 to verify nonce values and to disable verification. OCSP response
6134 printout format cleaned up.
6137 *) Add additional OCSP certificate checks. These are those specified
6138 in RFC2560. This consists of two separate checks: the CA of the
6139 certificate being checked must either be the OCSP signer certificate
6140 or the issuer of the OCSP signer certificate. In the latter case the
6141 OCSP signer certificate must contain the OCSP signing extended key
6142 usage. This check is performed by attempting to match the OCSP
6143 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6144 in the OCSP_CERTID structures of the response.
6147 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6148 and related routines. This uses the standard OpenSSL certificate
6149 verify routines to perform initial checks (just CA validity) and
6150 to obtain the certificate chain. Then additional checks will be
6151 performed on the chain. Currently the root CA is checked to see
6152 if it is explicitly trusted for OCSP signing. This is used to set
6153 a root CA as a global signing root: that is any certificate that
6154 chains to that CA is an acceptable OCSP signing certificate.
6157 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6158 extensions from a separate configuration file.
6159 As when reading extensions from the main configuration file,
6160 the '-extensions ...' option may be used for specifying the
6162 [Massimiliano Pala <madwolf@comune.modena.it>]
6164 *) New OCSP utility. Allows OCSP requests to be generated or
6165 read. The request can be sent to a responder and the output
6166 parsed, outputed or printed in text form. Not complete yet:
6167 still needs to check the OCSP response validity.
6170 *) New subcommands for 'openssl ca':
6171 'openssl ca -status <serial>' prints the status of the cert with
6172 the given serial number (according to the index file).
6173 'openssl ca -updatedb' updates the expiry status of certificates
6175 [Massimiliano Pala <madwolf@comune.modena.it>]
6177 *) New '-newreq-nodes' command option to CA.pl. This is like
6178 '-newreq', but calls 'openssl req' with the '-nodes' option
6179 so that the resulting key is not encrypted.
6180 [Damien Miller <djm@mindrot.org>]
6182 *) New configuration for the GNU Hurd.
6183 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6185 *) Initial code to implement OCSP basic response verify. This
6186 is currently incomplete. Currently just finds the signer's
6187 certificate and verifies the signature on the response.
6190 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6191 value of OPENSSLDIR. This is available via the new '-d' option
6192 to 'openssl version', and is also included in 'openssl version -a'.
6195 *) Allowing defining memory allocation callbacks that will be given
6196 file name and line number information in additional arguments
6197 (a const char* and an int). The basic functionality remains, as
6198 well as the original possibility to just replace malloc(),
6199 realloc() and free() by functions that do not know about these
6200 additional arguments. To register and find out the current
6201 settings for extended allocation functions, the following
6202 functions are provided:
6204 CRYPTO_set_mem_ex_functions
6205 CRYPTO_set_locked_mem_ex_functions
6206 CRYPTO_get_mem_ex_functions
6207 CRYPTO_get_locked_mem_ex_functions
6209 These work the same way as CRYPTO_set_mem_functions and friends.
6210 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6211 extended allocation function is enabled.
6212 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6213 a conventional allocation function is enabled.
6214 [Richard Levitte, Bodo Moeller]
6216 *) Finish off removing the remaining LHASH function pointer casts.
6217 There should no longer be any prototype-casting required when using
6218 the LHASH abstraction, and any casts that remain are "bugs". See
6219 the callback types and macros at the head of lhash.h for details
6220 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6223 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6224 If /dev/[u]random devices are not available or do not return enough
6225 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6227 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6228 /etc/entropy will be queried once each in this sequence, quering stops
6229 when enough entropy was collected without querying more sockets.
6232 *) Change the Unix RAND_poll() variant to be able to poll several
6233 random devices, as specified by DEVRANDOM, until a sufficient amount
6234 of data has been collected. We spend at most 10 ms on each file
6235 (select timeout) and read in non-blocking mode. DEVRANDOM now
6236 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6237 (previously it was just the string "/dev/urandom"), so on typical
6238 platforms the 10 ms delay will never occur.
6239 Also separate out the Unix variant to its own file, rand_unix.c.
6240 For VMS, there's a currently-empty rand_vms.c.
6243 *) Move OCSP client related routines to ocsp_cl.c. These
6244 provide utility functions which an application needing
6245 to issue a request to an OCSP responder and analyse the
6246 response will typically need: as opposed to those which an
6247 OCSP responder itself would need which will be added later.
6249 OCSP_request_sign() signs an OCSP request with an API similar
6250 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6251 response. OCSP_response_get1_basic() extracts basic response
6252 from response. OCSP_resp_find_status(): finds and extracts status
6253 information from an OCSP_CERTID structure (which will be created
6254 when the request structure is built). These are built from lower
6255 level functions which work on OCSP_SINGLERESP structures but
6256 wont normally be used unless the application wishes to examine
6257 extensions in the OCSP response for example.
6259 Replace nonce routines with a pair of functions.
6260 OCSP_request_add1_nonce() adds a nonce value and optionally
6261 generates a random value. OCSP_check_nonce() checks the
6262 validity of the nonce in an OCSP response.
6265 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6266 This doesn't copy the supplied OCSP_CERTID and avoids the
6267 need to free up the newly created id. Change return type
6268 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6269 This can then be used to add extensions to the request.
6270 Deleted OCSP_request_new(), since most of its functionality
6271 is now in OCSP_REQUEST_new() (and the case insensitive name
6272 clash) apart from the ability to set the request name which
6273 will be added elsewhere.
6276 *) Update OCSP API. Remove obsolete extensions argument from
6277 various functions. Extensions are now handled using the new
6278 OCSP extension code. New simple OCSP HTTP function which
6279 can be used to send requests and parse the response.
6282 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6283 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6284 uses the special reorder version of SET OF to sort the attributes
6285 and reorder them to match the encoded order. This resolves a long
6286 standing problem: a verify on a PKCS7 structure just after signing
6287 it used to fail because the attribute order did not match the
6288 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6289 it uses the received order. This is necessary to tolerate some broken
6290 software that does not order SET OF. This is handled by encoding
6291 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6292 to produce the required SET OF.
6295 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6296 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6297 files to get correct declarations of the ASN.1 item variables.
6300 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6301 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6302 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6303 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6304 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6305 ASN1_ITEM and no wrapper functions.
6308 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6309 replace the old function pointer based I/O routines. Change most of
6310 the *_d2i_bio() and *_d2i_fp() functions to use these.
6313 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6314 lines, recognice more "algorithms" that can be deselected, and make
6315 it complain about algorithm deselection that isn't recognised.
6318 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6319 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6320 to use new functions. Add NO_ASN1_OLD which can be set to remove
6321 some old style ASN1 functions: this can be used to determine if old
6322 code will still work when these eventually go away.
6325 *) New extension functions for OCSP structures, these follow the
6326 same conventions as certificates and CRLs.
6329 *) New function X509V3_add1_i2d(). This automatically encodes and
6330 adds an extension. Its behaviour can be customised with various
6331 flags to append, replace or delete. Various wrappers added for
6332 certifcates and CRLs.
6335 *) Fix to avoid calling the underlying ASN1 print routine when
6336 an extension cannot be parsed. Correct a typo in the
6337 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6340 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6341 entries for variables.
6344 *) Add functionality to apps/openssl.c for detecting locking
6345 problems: As the program is single-threaded, all we have
6346 to do is register a locking callback using an array for
6347 storing which locks are currently held by the program.
6350 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6351 SSL_get_ex_data_X509_STORE_idx(), which is used in
6352 ssl_verify_cert_chain() and thus can be called at any time
6353 during TLS/SSL handshakes so that thread-safety is essential.
6354 Unfortunately, the ex_data design is not at all suited
6355 for multi-threaded use, so it probably should be abolished.
6358 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6359 [Broadcom, tweaked and integrated by Geoff Thorpe]
6361 *) Move common extension printing code to new function
6362 X509V3_print_extensions(). Reorganise OCSP print routines and
6363 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6366 *) New function X509_signature_print() to remove duplication in some
6370 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6371 set (this was treated exactly the same as SET OF previously). This
6372 is used to reorder the STACK representing the structure to match the
6373 encoding. This will be used to get round a problem where a PKCS7
6374 structure which was signed could not be verified because the STACK
6375 order did not reflect the encoded order.
6378 *) Reimplement the OCSP ASN1 module using the new code.
6381 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6382 for its ASN1 operations. The old style function pointers still exist
6383 for now but they will eventually go away.
6386 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6387 completely replaces the old ASN1 functionality with a table driven
6388 encoder and decoder which interprets an ASN1_ITEM structure describing
6389 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6390 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6391 has also been converted to the new form.
6394 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6395 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6396 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6397 for negative moduli.
6400 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6401 of not touching the result's sign bit.
6404 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6408 *) Changed the LHASH code to use prototypes for callbacks, and created
6409 macros to declare and implement thin (optionally static) functions
6410 that provide type-safety and avoid function pointer casting for the
6411 type-specific callbacks.
6414 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6416 [Veers Staats <staatsvr@asc.hpc.mil>,
6417 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6419 *) Reformat the FAQ so the different questions and answers can be divided
6420 in sections depending on the subject.
6423 *) Have the zlib compression code load ZLIB.DLL dynamically under
6427 *) New function BN_mod_sqrt for computing square roots modulo a prime
6428 (using the probabilistic Tonelli-Shanks algorithm unless
6429 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6430 be handled deterministically).
6431 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6433 *) Make BN_mod_inverse faster by explicitly handling small quotients
6434 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6435 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6438 *) New function BN_kronecker.
6441 *) Fix BN_gcd so that it works on negative inputs; the result is
6442 positive unless both parameters are zero.
6443 Previously something reasonably close to an infinite loop was
6444 possible because numbers could be growing instead of shrinking
6445 in the implementation of Euclid's algorithm.
6448 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6449 sign of the number in question.
6451 Fix BN_is_word(a,w) to work correctly for w == 0.
6453 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6454 because its test if the absolute value of 'a' equals 'w'.
6455 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6456 it exists mostly for use in the implementations of BN_is_zero(),
6457 BN_is_one(), and BN_is_word().
6460 *) New function BN_swap.
6463 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6464 the exponentiation functions are more likely to produce reasonable
6465 results on negative inputs.
6468 *) Change BN_mod_mul so that the result is always non-negative.
6469 Previously, it could be negative if one of the factors was negative;
6470 I don't think anyone really wanted that behaviour.
6473 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6474 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6475 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6476 and add new functions:
6485 BN_mod_lshift1_quick
6489 These functions always generate non-negative results.
6491 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6492 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6494 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6495 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6496 be reduced modulo m.
6497 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6500 The following entry accidentally appeared in the CHANGES file
6501 distributed with OpenSSL 0.9.7. The modifications described in
6502 it do *not* apply to OpenSSL 0.9.7.
6504 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6505 was actually never needed) and in BN_mul(). The removal in BN_mul()
6506 required a small change in bn_mul_part_recursive() and the addition
6507 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6508 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6509 bn_sub_words() and bn_add_words() except they take arrays with
6514 *) In 'openssl passwd', verify passwords read from the terminal
6515 unless the '-salt' option is used (which usually means that
6516 verification would just waste user's time since the resulting
6517 hash is going to be compared with some given password hash)
6518 or the new '-noverify' option is used.
6520 This is an incompatible change, but it does not affect
6521 non-interactive use of 'openssl passwd' (passwords on the command
6522 line, '-stdin' option, '-in ...' option) and thus should not
6526 *) Remove all references to RSAref, since there's no more need for it.
6529 *) Make DSO load along a path given through an environment variable
6530 (SHLIB_PATH) with shl_load().
6533 *) Constify the ENGINE code as a result of BIGNUM constification.
6534 Also constify the RSA code and most things related to it. In a
6535 few places, most notable in the depth of the ASN.1 code, ugly
6536 casts back to non-const were required (to be solved at a later
6540 *) Make it so the openssl application has all engines loaded by default.
6543 *) Constify the BIGNUM routines a little more.
6546 *) Add the following functions:
6548 ENGINE_load_cswift()
6550 ENGINE_load_atalla()
6552 ENGINE_load_builtin_engines()
6554 That way, an application can itself choose if external engines that
6555 are built-in in OpenSSL shall ever be used or not. The benefit is
6556 that applications won't have to be linked with libdl or other dso
6557 libraries unless it's really needed.
6559 Changed 'openssl engine' to load all engines on demand.
6560 Changed the engine header files to avoid the duplication of some
6561 declarations (they differed!).
6564 *) 'openssl engine' can now list capabilities.
6567 *) Better error reporting in 'openssl engine'.
6570 *) Never call load_dh_param(NULL) in s_server.
6573 *) Add engine application. It can currently list engines by name and
6574 identity, and test if they are actually available.
6577 *) Improve RPM specification file by forcing symbolic linking and making
6578 sure the installed documentation is also owned by root.root.
6579 [Damien Miller <djm@mindrot.org>]
6581 *) Give the OpenSSL applications more possibilities to make use of
6582 keys (public as well as private) handled by engines.
6585 *) Add OCSP code that comes from CertCo.
6588 *) Add VMS support for the Rijndael code.
6591 *) Added untested support for Nuron crypto accelerator.
6594 *) Add support for external cryptographic devices. This code was
6595 previously distributed separately as the "engine" branch.
6596 [Geoff Thorpe, Richard Levitte]
6598 *) Rework the filename-translation in the DSO code. It is now possible to
6599 have far greater control over how a "name" is turned into a filename
6600 depending on the operating environment and any oddities about the
6601 different shared library filenames on each system.
6604 *) Support threads on FreeBSD-elf in Configure.
6607 *) Fix for SHA1 assembly problem with MASM: it produces
6608 warnings about corrupt line number information when assembling
6609 with debugging information. This is caused by the overlapping
6611 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6614 NCONF_get_number() has no error checking at all. As a replacement,
6615 NCONF_get_number_e() is defined (_e for "error checking") and is
6616 promoted strongly. The old NCONF_get_number is kept around for
6617 binary backward compatibility.
6618 Make it possible for methods to load from something other than a BIO,
6619 by providing a function pointer that is given a name instead of a BIO.
6620 For example, this could be used to load configuration data from an
6624 *) Fix for non blocking accept BIOs. Added new I/O special reason
6625 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6626 with non blocking I/O was not possible because no retry code was
6627 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6631 *) Added the beginnings of Rijndael support.
6634 *) Fix for bug in DirectoryString mask setting. Add support for
6635 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6636 to allow certificate printing to more controllable, additional
6637 'certopt' option to 'x509' to allow new printing options to be
6641 *) Clean old EAY MD5 hack from e_os.h.
6644 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6646 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6647 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6648 [Joe Orton, Steve Henson]
6650 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6652 *) Fix additional bug revealed by the NISCC test suite:
6654 Stop bug triggering large recursion when presented with
6655 certain ASN.1 tags (CVE-2003-0851)
6658 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6660 *) Fix various bugs revealed by running the NISCC test suite:
6662 Stop out of bounds reads in the ASN1 code when presented with
6663 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6665 If verify callback ignores invalid public key errors don't try to check
6666 certificate signature with the NULL public key.
6670 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6671 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6675 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6676 extra data after the compression methods not only for TLS 1.0
6677 but also for SSL 3.0 (as required by the specification).
6678 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6680 *) Change X509_certificate_type() to mark the key as exported/exportable
6681 when it's 512 *bits* long, not 512 bytes.
6684 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6686 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6687 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6688 a protocol version number mismatch like a decryption error
6689 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6692 *) Turn on RSA blinding by default in the default implementation
6693 to avoid a timing attack. Applications that don't want it can call
6694 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6695 They would be ill-advised to do so in most cases.
6696 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6698 *) Change RSA blinding code so that it works when the PRNG is not
6699 seeded (in this case, the secret RSA exponent is abused as
6700 an unpredictable seed -- if it is not unpredictable, there
6701 is no point in blinding anyway). Make RSA blinding thread-safe
6702 by remembering the creator's thread ID in rsa->blinding and
6703 having all other threads use local one-time blinding factors
6704 (this requires more computation than sharing rsa->blinding, but
6705 avoids excessive locking; and if an RSA object is not shared
6706 between threads, blinding will still be very fast).
6709 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6711 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6712 via timing by performing a MAC computation even if incorrrect
6713 block cipher padding has been found. This is a countermeasure
6714 against active attacks where the attacker has to distinguish
6715 between bad padding and a MAC verification error. (CVE-2003-0078)
6717 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6718 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6719 Martin Vuagnoux (EPFL, Ilion)]
6721 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6723 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6724 memory from it's contents. This is done with a counter that will
6725 place alternating values in each byte. This can be used to solve
6726 two issues: 1) the removal of calls to memset() by highly optimizing
6727 compilers, and 2) cleansing with other values than 0, since those can
6728 be read through on certain media, for example a swap space on disk.
6731 *) Bugfix: client side session caching did not work with external caching,
6732 because the session->cipher setting was not restored when reloading
6733 from the external cache. This problem was masked, when
6734 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6735 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6738 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6739 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6740 [Zeev Lieber <zeev-l@yahoo.com>]
6742 *) Undo an undocumented change introduced in 0.9.6e which caused
6743 repeated calls to OpenSSL_add_all_ciphers() and
6744 OpenSSL_add_all_digests() to be ignored, even after calling
6748 *) Change the default configuration reader to deal with last line not
6749 being properly terminated.
6752 *) Change X509_NAME_cmp() so it applies the special rules on handling
6753 DN values that are of type PrintableString, as well as RDNs of type
6754 emailAddress where the value has the type ia5String.
6755 [stefank@valicert.com via Richard Levitte]
6757 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6758 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6759 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6760 the bitwise-OR of the two for use by the majority of applications
6761 wanting this behaviour, and update the docs. The documented
6762 behaviour and actual behaviour were inconsistent and had been
6763 changing anyway, so this is more a bug-fix than a behavioural
6765 [Geoff Thorpe, diagnosed by Nadav Har'El]
6767 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6768 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6771 *) Fix initialization code race conditions in
6772 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6773 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6774 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6775 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6776 ssl2_get_cipher_by_char(),
6777 ssl3_get_cipher_by_char().
6778 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6780 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6781 the cached sessions are flushed, as the remove_cb() might use ex_data
6782 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6783 (see [openssl.org #212]).
6784 [Geoff Thorpe, Lutz Jaenicke]
6786 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6787 length, instead of the encoding length to d2i_ASN1_OBJECT.
6790 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6792 *) [In 0.9.6g-engine release:]
6793 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6794 [Lynn Gazis <lgazis@rainbow.com>]
6796 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6798 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6799 and get fix the header length calculation.
6800 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6801 Alon Kantor <alonk@checkpoint.com> (and others),
6804 *) Use proper error handling instead of 'assertions' in buffer
6805 overflow checks added in 0.9.6e. This prevents DoS (the
6806 assertions could call abort()).
6807 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6809 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6811 *) Add various sanity checks to asn1_get_length() to reject
6812 the ASN1 length bytes if they exceed sizeof(long), will appear
6813 negative or the content length exceeds the length of the
6815 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6817 *) Fix cipher selection routines: ciphers without encryption had no flags
6818 for the cipher strength set and where therefore not handled correctly
6819 by the selection routines (PR #130).
6822 *) Fix EVP_dsa_sha macro.
6826 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6827 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6828 that was added in OpenSSL 0.9.6d.
6830 As the countermeasure turned out to be incompatible with some
6831 broken SSL implementations, the new option is part of SSL_OP_ALL.
6832 SSL_OP_ALL is usually employed when compatibility with weird SSL
6833 implementations is desired (e.g. '-bugs' option to 's_client' and
6834 's_server'), so the new option is automatically set in many
6838 *) Changes in security patch:
6840 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6841 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6842 Air Force Materiel Command, USAF, under agreement number
6845 *) Add various sanity checks to asn1_get_length() to reject
6846 the ASN1 length bytes if they exceed sizeof(long), will appear
6847 negative or the content length exceeds the length of the
6848 supplied buffer. (CVE-2002-0659)
6849 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6851 *) Assertions for various potential buffer overflows, not known to
6853 [Ben Laurie (CHATS)]
6855 *) Various temporary buffers to hold ASCII versions of integers were
6856 too small for 64 bit platforms. (CVE-2002-0655)
6857 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6859 *) Remote buffer overflow in SSL3 protocol - an attacker could
6860 supply an oversized session ID to a client. (CVE-2002-0656)
6861 [Ben Laurie (CHATS)]
6863 *) Remote buffer overflow in SSL2 protocol - an attacker could
6864 supply an oversized client master key. (CVE-2002-0656)
6865 [Ben Laurie (CHATS)]
6867 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6869 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6870 encoded as NULL) with id-dsa-with-sha1.
6871 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6873 *) Check various X509_...() return values in apps/req.c.
6874 [Nils Larsch <nla@trustcenter.de>]
6876 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6877 an end-of-file condition would erronously be flagged, when the CRLF
6878 was just at the end of a processed block. The bug was discovered when
6879 processing data through a buffering memory BIO handing the data to a
6880 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6881 <ptsekov@syntrex.com> and Nedelcho Stanev.
6884 *) Implement a countermeasure against a vulnerability recently found
6885 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6886 before application data chunks to avoid the use of known IVs
6887 with data potentially chosen by the attacker.
6890 *) Fix length checks in ssl3_get_client_hello().
6893 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6894 to prevent ssl3_read_internal() from incorrectly assuming that
6895 ssl3_read_bytes() found application data while handshake
6896 processing was enabled when in fact s->s3->in_read_app_data was
6897 merely automatically cleared during the initial handshake.
6898 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6900 *) Fix object definitions for Private and Enterprise: they were not
6901 recognized in their shortname (=lowercase) representation. Extend
6902 obj_dat.pl to issue an error when using undefined keywords instead
6903 of silently ignoring the problem (Svenning Sorensen
6904 <sss@sss.dnsalias.net>).
6907 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6908 generators, i.e. generators other than 2 and 5. (Previously, the
6909 code did not properly initialise the 'add' and 'rem' values to
6910 BN_generate_prime().)
6912 In the new general case, we do not insist that 'generator' is
6913 actually a primitive root: This requirement is rather pointless;
6914 a generator of the order-q subgroup is just as good, if not
6918 *) Map new X509 verification errors to alerts. Discovered and submitted by
6919 Tom Wu <tom@arcot.com>.
6922 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6923 returning non-zero before the data has been completely received
6924 when using non-blocking I/O.
6925 [Bodo Moeller; problem pointed out by John Hughes]
6927 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6928 [Ben Laurie, Lutz Jaenicke]
6930 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6931 Yoram Zahavi <YoramZ@gilian.com>).
6934 *) Add information about CygWin 1.3 and on, and preserve proper
6935 configuration for the versions before that.
6936 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6938 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6939 check whether we deal with a copy of a session and do not delete from
6940 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6941 <izhar@checkpoint.com>.
6944 *) Do not store session data into the internal session cache, if it
6945 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6946 flag is set). Proposed by Aslam <aslam@funk.com>.
6949 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6953 *) [In 0.9.6d-engine release:]
6954 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6955 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6957 *) Add the configuration target linux-s390x.
6958 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6960 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6961 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6962 variable as an indication that a ClientHello message has been
6963 received. As the flag value will be lost between multiple
6964 invocations of ssl3_accept when using non-blocking I/O, the
6965 function may not be aware that a handshake has actually taken
6966 place, thus preventing a new session from being added to the
6969 To avoid this problem, we now set s->new_session to 2 instead of
6970 using a local variable.
6971 [Lutz Jaenicke, Bodo Moeller]
6973 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6974 if the SSL_R_LENGTH_MISMATCH error is detected.
6975 [Geoff Thorpe, Bodo Moeller]
6977 *) New 'shared_ldflag' column in Configure platform table.
6980 *) Fix EVP_CIPHER_mode macro.
6981 ["Dan S. Camper" <dan@bti.net>]
6983 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6984 type, we must throw them away by setting rr->length to 0.
6985 [D P Chang <dpc@qualys.com>]
6987 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6989 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6990 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6991 worked incorrectly for those cases where range = 10..._2 and
6992 3*range is two bits longer than range.)
6995 *) Only add signing time to PKCS7 structures if it is not already
6999 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7000 OBJ_ld_ce should be OBJ_id_ce.
7001 Also some ip-pda OIDs in crypto/objects/objects.txt were
7002 incorrect (cf. RFC 3039).
7003 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7005 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7006 returns early because it has nothing to do.
7007 [Andy Schneider <andy.schneider@bjss.co.uk>]
7009 *) [In 0.9.6c-engine release:]
7010 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7011 [Andy Schneider <andy.schneider@bjss.co.uk>]
7013 *) [In 0.9.6c-engine release:]
7014 Add support for Cryptographic Appliance's keyserver technology.
7015 (Use engine 'keyclient')
7016 [Cryptographic Appliances and Geoff Thorpe]
7018 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7019 is called via tools/c89.sh because arguments have to be
7020 rearranged (all '-L' options must appear before the first object
7022 [Richard Shapiro <rshapiro@abinitio.com>]
7024 *) [In 0.9.6c-engine release:]
7025 Add support for Broadcom crypto accelerator cards, backported
7027 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7029 *) [In 0.9.6c-engine release:]
7030 Add support for SureWare crypto accelerator cards from
7031 Baltimore Technologies. (Use engine 'sureware')
7032 [Baltimore Technologies and Mark Cox]
7034 *) [In 0.9.6c-engine release:]
7035 Add support for crypto accelerator cards from Accelerated
7036 Encryption Processing, www.aep.ie. (Use engine 'aep')
7037 [AEP Inc. and Mark Cox]
7039 *) Add a configuration entry for gcc on UnixWare.
7040 [Gary Benson <gbenson@redhat.com>]
7042 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7043 messages are stored in a single piece (fixed-length part and
7044 variable-length part combined) and fix various bugs found on the way.
7047 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7048 instead. BIO_gethostbyname() does not know what timeouts are
7049 appropriate, so entries would stay in cache even when they have
7051 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7053 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7054 faced with a pathologically small ClientHello fragment that does
7055 not contain client_version: Instead of aborting with an error,
7056 simply choose the highest available protocol version (i.e.,
7057 TLS 1.0 unless it is disabled). In practice, ClientHello
7058 messages are never sent like this, but this change gives us
7059 strictly correct behaviour at least for TLS.
7062 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7063 never resets s->method to s->ctx->method when called from within
7064 one of the SSL handshake functions.
7065 [Bodo Moeller; problem pointed out by Niko Baric]
7067 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7068 (sent using the client's version number) if client_version is
7069 smaller than the protocol version in use. Also change
7070 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7071 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7072 the client will at least see that alert.
7075 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7079 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7080 client receives HelloRequest while in a handshake.
7081 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7083 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7084 should end in 'break', not 'goto end' which circumvents various
7085 cleanups done in state SSL_ST_OK. But session related stuff
7086 must be disabled for SSL_ST_OK in the case that we just sent a
7089 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7090 before just sending a HelloRequest.
7091 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7093 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7094 reveal whether illegal block cipher padding was found or a MAC
7095 verification error occured. (Neither SSLerr() codes nor alerts
7096 are directly visible to potential attackers, but the information
7097 may leak via logfiles.)
7099 Similar changes are not required for the SSL 2.0 implementation
7100 because the number of padding bytes is sent in clear for SSL 2.0,
7101 and the extra bytes are just ignored. However ssl/s2_pkt.c
7102 failed to verify that the purported number of padding bytes is in
7106 *) Add OpenUNIX-8 support including shared libraries
7107 (Boyd Lynn Gerber <gerberb@zenez.com>).
7110 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7111 'wristwatch attack' using huge encoding parameters (cf.
7112 James H. Manger's CRYPTO 2001 paper). Note that the
7113 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7114 encoding parameters and hence was not vulnerable.
7117 *) BN_sqr() bug fix.
7118 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7120 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7121 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7122 followed by modular reduction.
7123 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7125 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7126 equivalent based on BN_pseudo_rand() instead of BN_rand().
7129 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7130 This function was broken, as the check for a new client hello message
7131 to handle SGC did not allow these large messages.
7132 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7135 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7138 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7139 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7142 *) Rework the configuration and shared library support for Tru64 Unix.
7143 The configuration part makes use of modern compiler features and
7144 still retains old compiler behavior for those that run older versions
7145 of the OS. The shared library support part includes a variant that
7146 uses the RPATH feature, and is available through the special
7147 configuration target "alpha-cc-rpath", which will never be selected
7149 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7151 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7152 with the same message size as in ssl3_get_certificate_request().
7153 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7154 messages might inadvertently be reject as too long.
7155 [Petr Lampa <lampa@fee.vutbr.cz>]
7157 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7160 *) Modified SSL library such that the verify_callback that has been set
7161 specificly for an SSL object with SSL_set_verify() is actually being
7162 used. Before the change, a verify_callback set with this function was
7163 ignored and the verify_callback() set in the SSL_CTX at the time of
7164 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7165 to allow the necessary settings.
7168 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7169 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7170 done automatically (in contradiction to the requirements of the C
7171 standard). This made problems when used from OpenSSH.
7174 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7175 dh->length and always used
7177 BN_rand_range(priv_key, dh->p).
7179 BN_rand_range() is not necessary for Diffie-Hellman, and this
7180 specific range makes Diffie-Hellman unnecessarily inefficient if
7181 dh->length (recommended exponent length) is much smaller than the
7182 length of dh->p. We could use BN_rand_range() if the order of
7183 the subgroup was stored in the DH structure, but we only have
7188 BN_rand(priv_key, l, ...)
7190 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7196 RSA_eay_public_encrypt
7197 RSA_eay_private_decrypt
7198 RSA_eay_private_encrypt (signing)
7199 RSA_eay_public_decrypt (signature verification)
7201 (default implementations for RSA_public_encrypt,
7202 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7203 always reject numbers >= n.
7206 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7207 to synchronize access to 'locking_thread'. This is necessary on
7208 systems where access to 'locking_thread' (an 'unsigned long'
7209 variable) is not atomic.
7212 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7213 *before* setting the 'crypto_lock_rand' flag. The previous code had
7214 a race condition if 0 is a valid thread ID.
7215 [Travis Vitek <vitek@roguewave.com>]
7217 *) Add support for shared libraries under Irix.
7218 [Albert Chin-A-Young <china@thewrittenword.com>]
7220 *) Add configuration option to build on Linux on both big-endian and
7222 [Ralf Baechle <ralf@uni-koblenz.de>]
7224 *) Add the possibility to create shared libraries on HP-UX.
7227 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7229 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7230 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7231 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7232 PRNG state recovery was possible based on the output of
7233 one PRNG request appropriately sized to gain knowledge on
7234 'md' followed by enough consecutive 1-byte PRNG requests
7235 to traverse all of 'state'.
7237 1. When updating 'md_local' (the current thread's copy of 'md')
7238 during PRNG output generation, hash all of the previous
7239 'md_local' value, not just the half used for PRNG output.
7241 2. Make the number of bytes from 'state' included into the hash
7242 independent from the number of PRNG bytes requested.
7244 The first measure alone would be sufficient to avoid
7245 Markku-Juhani's attack. (Actually it had never occurred
7246 to me that the half of 'md_local' used for chaining was the
7247 half from which PRNG output bytes were taken -- I had always
7248 assumed that the secret half would be used.) The second
7249 measure makes sure that additional data from 'state' is never
7250 mixed into 'md_local' in small portions; this heuristically
7251 further strengthens the PRNG.
7254 *) Fix crypto/bn/asm/mips3.s.
7257 *) When only the key is given to "enc", the IV is undefined. Print out
7258 an error message in this case.
7261 *) Handle special case when X509_NAME is empty in X509 printing routines.
7264 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7265 positive and less than q.
7268 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7269 used: it isn't thread safe and the add_lock_callback should handle
7271 [Paul Rose <Paul.Rose@bridge.com>]
7273 *) Verify that incoming data obeys the block size in
7274 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7278 [Ulf Möller, Bodo Möller]
7280 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7281 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7282 when fixing the server behaviour for backwards-compatible 'client
7283 hello' messages. (Note that the attack is impractical against
7284 SSL 3.0 and TLS 1.0 anyway because length and version checking
7285 means that the probability of guessing a valid ciphertext is
7286 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7289 Before 0.9.5, the countermeasure (hide the error by generating a
7290 random 'decryption result') did not work properly because
7291 ERR_clear_error() was missing, meaning that SSL_get_error() would
7292 detect the supposedly ignored error.
7294 Both problems are now fixed.
7297 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7298 (previously it was 1024).
7301 *) Fix for compatibility mode trust settings: ignore trust settings
7302 unless some valid trust or reject settings are present.
7305 *) Fix for blowfish EVP: its a variable length cipher.
7308 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7309 parameters in DSA public key structures and return an error in the
7310 DSA routines if parameters are absent.
7313 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7314 in the current directory if neither $RANDFILE nor $HOME was set.
7315 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7316 caused some confusion to Windows users who haven't defined $HOME.
7317 Thus RAND_file_name() is changed again: e_os.h can define a
7318 DEFAULT_HOME, which will be used if $HOME is not set.
7319 For Windows, we use "C:"; on other platforms, we still require
7320 environment variables.
7322 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7323 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7324 having multiple threads call RAND_poll() concurrently.
7327 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7328 combination of a flag and a thread ID variable.
7329 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7330 flag), *other* threads can enter ssleay_add_bytes without obeying
7331 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7332 that they do not hold after the first thread unsets add_do_not_lock).
7335 *) Change bctest again: '-x' expressions are not available in all
7339 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7341 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7342 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7344 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7345 the default extension for executables, if any. Also, make the perl
7346 scripts that use symlink() to test if it really exists and use "cp"
7347 if it doesn't. All this made OpenSSL compilable and installable in
7351 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7352 If SEQUENCE is length is indefinite just set c->slen to the total
7353 amount of data available.
7354 [Steve Henson, reported by shige@FreeBSD.org]
7355 [This change does not apply to 0.9.7.]
7357 *) Change bctest to avoid here-documents inside command substitution
7358 (workaround for FreeBSD /bin/sh bug).
7359 For compatibility with Ultrix, avoid shell functions (introduced
7360 in the bctest version that searches along $PATH).
7363 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7364 with des_encrypt() defined on some operating systems, like Solaris
7368 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7369 On the Importance of Eliminating Errors in Cryptographic
7370 Computations, J. Cryptology 14 (2001) 2, 101-119,
7371 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7374 *) MIPS assembler BIGNUM division bug fix.
7377 *) Disabled incorrect Alpha assembler code.
7380 *) Fix PKCS#7 decode routines so they correctly update the length
7381 after reading an EOC for the EXPLICIT tag.
7383 [This change does not apply to 0.9.7.]
7385 *) Fix bug in PKCS#12 key generation routines. This was triggered
7386 if a 3DES key was generated with a 0 initial byte. Include
7387 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7388 (but broken) behaviour.
7391 *) Enhance bctest to search for a working bc along $PATH and print
7393 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7395 *) Fix memory leaks in err.c: free err_data string if necessary;
7396 don't write to the wrong index in ERR_set_error_data.
7399 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7403 *) Replace rdtsc with _emit statements for VC++ version 5.
7404 [Jeremy Cooper <jeremy@baymoo.org>]
7406 *) Make it possible to reuse SSLv2 sessions.
7409 *) In copy_email() check for >= 0 as a return value for
7410 X509_NAME_get_index_by_NID() since 0 is a valid index.
7411 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7413 *) Avoid coredump with unsupported or invalid public keys by checking if
7414 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7415 PKCS7_verify() fails with non detached data.
7418 *) Don't use getenv in library functions when run as setuid/setgid.
7419 New function OPENSSL_issetugid().
7422 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7423 due to incorrect handling of multi-threading:
7425 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7427 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7429 3. Count how many times MemCheck_off() has been called so that
7430 nested use can be treated correctly. This also avoids
7431 inband-signalling in the previous code (which relied on the
7432 assumption that thread ID 0 is impossible).
7435 *) Add "-rand" option also to s_client and s_server.
7438 *) Fix CPU detection on Irix 6.x.
7439 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7440 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7442 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7445 [This change does not apply to 0.9.7.]
7447 *) Use the cached encoding of an X509_NAME structure rather than
7448 copying it. This is apparently the reason for the libsafe "errors"
7449 but the code is actually correct.
7452 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7453 Bleichenbacher's DSA attack.
7454 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7455 to be set and top=0 forces the highest bit to be set; top=-1 is new
7456 and leaves the highest bit random.
7457 [Ulf Moeller, Bodo Moeller]
7459 *) In the NCONF_...-based implementations for CONF_... queries
7460 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7461 a temporary CONF structure with the data component set to NULL
7462 (which gives segmentation faults in lh_retrieve).
7463 Instead, use NULL for the CONF pointer in CONF_get_string and
7464 CONF_get_number (which may use environment variables) and directly
7465 return NULL from CONF_get_section.
7468 *) Fix potential buffer overrun for EBCDIC.
7471 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7472 keyUsage if basicConstraints absent for a CA.
7475 *) Make SMIME_write_PKCS7() write mail header values with a format that
7476 is more generally accepted (no spaces before the semicolon), since
7477 some programs can't parse those values properly otherwise. Also make
7478 sure BIO's that break lines after each write do not create invalid
7482 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7483 macros previously used would not encode an empty SEQUENCE OF
7484 and break the signature.
7486 [This change does not apply to 0.9.7.]
7488 *) Zero the premaster secret after deriving the master secret in
7492 *) Add some EVP_add_digest_alias registrations (as found in
7493 OpenSSL_add_all_digests()) to SSL_library_init()
7494 aka OpenSSL_add_ssl_algorithms(). This provides improved
7495 compatibility with peers using X.509 certificates
7496 with unconventional AlgorithmIdentifier OIDs.
7499 *) Fix for Irix with NO_ASM.
7500 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7502 *) ./config script fixes.
7503 [Ulf Moeller, Richard Levitte]
7505 *) Fix 'openssl passwd -1'.
7508 *) Change PKCS12_key_gen_asc() so it can cope with non null
7509 terminated strings whose length is passed in the passlen
7510 parameter, for example from PEM callbacks. This was done
7511 by adding an extra length parameter to asc2uni().
7512 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7514 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7515 call failed, free the DSA structure.
7518 *) Fix to uni2asc() to cope with zero length Unicode strings.
7519 These are present in some PKCS#12 files.
7522 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7523 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7524 when writing a 32767 byte record.
7525 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7527 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7528 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7530 (RSA objects have a reference count access to which is protected
7531 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7532 so they are meant to be shared between threads.)
7533 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7534 "Reddie, Steven" <Steven.Reddie@ca.com>]
7536 *) Fix a deadlock in CRYPTO_mem_leaks().
7539 *) Use better test patterns in bntest.
7542 *) rand_win.c fix for Borland C.
7545 *) BN_rshift bugfix for n == 0.
7548 *) Add a 'bctest' script that checks for some known 'bc' bugs
7549 so that 'make test' does not abort just because 'bc' is broken.
7552 *) Store verify_result within SSL_SESSION also for client side to
7553 avoid potential security hole. (Re-used sessions on the client side
7554 always resulted in verify_result==X509_V_OK, not using the original
7555 result of the server certificate verification.)
7558 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7559 SSL3_RT_APPLICATION_DATA, return 0.
7560 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7564 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7565 releases, have been re-implemented by renaming the previous
7566 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7567 and ssl3_read_internal, respectively, and adding 'peek' parameters
7568 to them. The new ssl[23]_{read,peek} functions are calls to
7569 ssl[23]_read_internal with the 'peek' flag set appropriately.
7570 A 'peek' parameter has also been added to ssl3_read_bytes, which
7571 does the actual work for ssl3_read_internal.
7574 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7575 the method-specific "init()" handler. Also clean up ex_data after
7576 calling the method-specific "finish()" handler. Previously, this was
7577 happening the other way round.
7580 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7581 The previous value, 12, was not always sufficient for BN_mod_exp().
7584 *) Make sure that shared libraries get the internal name engine with
7585 the full version number and not just 0. This should mark the
7586 shared libraries as not backward compatible. Of course, this should
7587 be changed again when we can guarantee backward binary compatibility.
7590 *) Fix typo in get_cert_by_subject() in by_dir.c
7591 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7593 *) Rework the system to generate shared libraries:
7595 - Make note of the expected extension for the shared libraries and
7596 if there is a need for symbolic links from for example libcrypto.so.0
7597 to libcrypto.so.0.9.7. There is extended info in Configure for
7600 - Make as few rebuilds of the shared libraries as possible.
7602 - Still avoid linking the OpenSSL programs with the shared libraries.
7604 - When installing, install the shared libraries separately from the
7608 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7610 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7611 and not in SSL_clear because the latter is also used by the
7612 accept/connect functions; previously, the settings made by
7613 SSL_set_read_ahead would be lost during the handshake.
7614 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7616 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7617 Previously, it would create entries for disableed algorithms no
7621 *) Added several new manual pages for SSL_* function.
7624 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7626 *) In ssl23_get_client_hello, generate an error message when faced
7627 with an initial SSL 3.0/TLS record that is too small to contain the
7628 first two bytes of the ClientHello message, i.e. client_version.
7629 (Note that this is a pathologic case that probably has never happened
7630 in real life.) The previous approach was to use the version number
7631 from the record header as a substitute; but our protocol choice
7632 should not depend on that one because it is not authenticated
7633 by the Finished messages.
7636 *) More robust randomness gathering functions for Windows.
7637 [Jeffrey Altman <jaltman@columbia.edu>]
7639 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7640 not set then we don't setup the error code for issuer check errors
7641 to avoid possibly overwriting other errors which the callback does
7642 handle. If an application does set the flag then we assume it knows
7643 what it is doing and can handle the new informational codes
7647 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7648 a general "ANY" type, as such it should be able to decode anything
7649 including tagged types. However it didn't check the class so it would
7650 wrongly interpret tagged types in the same way as their universal
7651 counterpart and unknown types were just rejected. Changed so that the
7652 tagged and unknown types are handled in the same way as a SEQUENCE:
7653 that is the encoding is stored intact. There is also a new type
7654 "V_ASN1_OTHER" which is used when the class is not universal, in this
7655 case we have no idea what the actual type is so we just lump them all
7659 *) On VMS, stdout may very well lead to a file that is written to
7660 in a record-oriented fashion. That means that every write() will
7661 write a separate record, which will be read separately by the
7662 programs trying to read from it. This can be very confusing.
7664 The solution is to put a BIO filter in the way that will buffer
7665 text until a linefeed is reached, and then write everything a
7666 line at a time, so every record written will be an actual line,
7667 not chunks of lines and not (usually doesn't happen, but I've
7668 seen it once) several lines in one record. BIO_f_linebuffer() is
7671 Currently, it's a VMS-only method, because that's where it has
7672 been tested well enough.
7675 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7676 it can return incorrect results.
7677 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7678 but it was in 0.9.6-beta[12].)
7681 *) Disable the check for content being present when verifying detached
7682 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7683 include zero length content when signing messages.
7686 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7687 BIO_ctrl (for BIO pairs).
7690 *) Add DSO method for VMS.
7693 *) Bug fix: Montgomery multiplication could produce results with the
7697 *) Add RPM specification openssl.spec and modify it to build three
7698 packages. The default package contains applications, application
7699 documentation and run-time libraries. The devel package contains
7700 include files, static libraries and function documentation. The
7701 doc package contains the contents of the doc directory. The original
7702 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7705 *) Add a large number of documentation files for many SSL routines.
7706 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7708 *) Add a configuration entry for Sony News 4.
7709 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7711 *) Don't set the two most significant bits to one when generating a
7712 random number < q in the DSA library.
7715 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7716 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7717 the underlying transport is blocking) if a handshake took place.
7718 (The default behaviour is needed by applications such as s_client
7719 and s_server that use select() to determine when to use SSL_read;
7720 but for applications that know in advance when to expect data, it
7721 just makes things more complicated.)
7724 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7728 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7729 work better on such systems.
7730 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7732 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7733 Update PKCS12_parse() so it copies the friendlyName and the
7734 keyid to the certificates aux info.
7737 *) Fix bug in PKCS7_verify() which caused an infinite loop
7738 if there was more than one signature.
7739 [Sven Uszpelkat <su@celocom.de>]
7741 *) Major change in util/mkdef.pl to include extra information
7742 about each symbol, as well as presenting variables as well
7743 as functions. This change means that there's n more need
7744 to rebuild the .num files when some algorithms are excluded.
7747 *) Allow the verify time to be set by an application,
7748 rather than always using the current time.
7751 *) Phase 2 verify code reorganisation. The certificate
7752 verify code now looks up an issuer certificate by a
7753 number of criteria: subject name, authority key id
7754 and key usage. It also verifies self signed certificates
7755 by the same criteria. The main comparison function is
7756 X509_check_issued() which performs these checks.
7758 Lot of changes were necessary in order to support this
7759 without completely rewriting the lookup code.
7761 Authority and subject key identifier are now cached.
7763 The LHASH 'certs' is X509_STORE has now been replaced
7764 by a STACK_OF(X509_OBJECT). This is mainly because an
7765 LHASH can't store or retrieve multiple objects with
7766 the same hash value.
7768 As a result various functions (which were all internal
7769 use only) have changed to handle the new X509_STORE
7770 structure. This will break anything that messed round
7771 with X509_STORE internally.
7773 The functions X509_STORE_add_cert() now checks for an
7774 exact match, rather than just subject name.
7776 The X509_STORE API doesn't directly support the retrieval
7777 of multiple certificates matching a given criteria, however
7778 this can be worked round by performing a lookup first
7779 (which will fill the cache with candidate certificates)
7780 and then examining the cache for matches. This is probably
7781 the best we can do without throwing out X509_LOOKUP
7782 entirely (maybe later...).
7784 The X509_VERIFY_CTX structure has been enhanced considerably.
7786 All certificate lookup operations now go via a get_issuer()
7787 callback. Although this currently uses an X509_STORE it
7788 can be replaced by custom lookups. This is a simple way
7789 to bypass the X509_STORE hackery necessary to make this
7790 work and makes it possible to use more efficient techniques
7791 in future. A very simple version which uses a simple
7792 STACK for its trusted certificate store is also provided
7793 using X509_STORE_CTX_trusted_stack().
7795 The verify_cb() and verify() callbacks now have equivalents
7796 in the X509_STORE_CTX structure.
7798 X509_STORE_CTX also has a 'flags' field which can be used
7799 to customise the verify behaviour.
7802 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7803 excludes S/MIME capabilities.
7806 *) When a certificate request is read in keep a copy of the
7807 original encoding of the signed data and use it when outputing
7808 again. Signatures then use the original encoding rather than
7809 a decoded, encoded version which may cause problems if the
7810 request is improperly encoded.
7813 *) For consistency with other BIO_puts implementations, call
7814 buffer_write(b, ...) directly in buffer_puts instead of calling
7817 In BIO_puts, increment b->num_write as in BIO_write.
7818 [Peter.Sylvester@EdelWeb.fr]
7820 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7821 BN_zero, we may not return a BIGNUM with an array consisting of
7825 *) Avoid calling abort() from within the library when problems are
7826 detected, except if preprocessor symbols have been defined
7827 (such as REF_CHECK, BN_DEBUG etc.).
7830 *) New openssl application 'rsautl'. This utility can be
7831 used for low level RSA operations. DER public key
7832 BIO/fp routines also added.
7835 *) New Configure entry and patches for compiling on QNX 4.
7836 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7838 *) A demo state-machine implementation was sponsored by
7839 Nuron (http://www.nuron.com/) and is now available in
7840 demos/state_machine.
7843 *) New options added to the 'dgst' utility for signature
7844 generation and verification.
7847 *) Unrecognized PKCS#7 content types are now handled via a
7848 catch all ASN1_TYPE structure. This allows unsupported
7849 types to be stored as a "blob" and an application can
7850 encode and decode it manually.
7853 *) Fix various signed/unsigned issues to make a_strex.c
7855 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7857 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7858 length if passed a buffer. ASN1_INTEGER_to_BN failed
7859 if passed a NULL BN and its argument was negative.
7860 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7862 *) Modification to PKCS#7 encoding routines to output definite
7863 length encoding. Since currently the whole structures are in
7864 memory there's not real point in using indefinite length
7865 constructed encoding. However if OpenSSL is compiled with
7866 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7869 *) Added BIO_vprintf() and BIO_vsnprintf().
7872 *) Added more prefixes to parse for in the the strings written
7873 through a logging bio, to cover all the levels that are available
7874 through syslog. The prefixes are now:
7876 PANIC, EMERG, EMR => LOG_EMERG
7877 ALERT, ALR => LOG_ALERT
7878 CRIT, CRI => LOG_CRIT
7879 ERROR, ERR => LOG_ERR
7880 WARNING, WARN, WAR => LOG_WARNING
7881 NOTICE, NOTE, NOT => LOG_NOTICE
7882 INFO, INF => LOG_INFO
7883 DEBUG, DBG => LOG_DEBUG
7885 and as before, if none of those prefixes are present at the
7886 beginning of the string, LOG_ERR is chosen.
7888 On Win32, the LOG_* levels are mapped according to this:
7890 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7891 LOG_WARNING => EVENTLOG_WARNING_TYPE
7892 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7896 *) Made it possible to reconfigure with just the configuration
7897 argument "reconf" or "reconfigure". The command line arguments
7898 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7899 and are retrieved from there when reconfiguring.
7903 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7905 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7908 *) The obj_dat.pl script was messing up the sorting of object
7909 names. The reason was that it compared the quoted version
7910 of strings as a result "OCSP" > "OCSP Signing" because
7911 " > SPACE. Changed script to store unquoted versions of
7912 names and add quotes on output. It was also omitting some
7913 names from the lookup table if they were given a default
7914 value (that is if SN is missing it is given the same
7915 value as LN and vice versa), these are now added on the
7916 grounds that if an object has a name we should be able to
7917 look it up. Finally added warning output when duplicate
7918 short or long names are found.
7921 *) Changes needed for Tandem NSK.
7922 [Scott Uroff <scott@xypro.com>]
7924 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7925 RSA_padding_check_SSLv23(), special padding was never detected
7926 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7927 version rollback attacks was not effective.
7929 In s23_clnt.c, don't use special rollback-attack detection padding
7930 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7931 client; similarly, in s23_srvr.c, don't do the rollback check if
7932 SSL 2.0 is the only protocol enabled in the server.
7935 *) Make it possible to get hexdumps of unprintable data with 'openssl
7936 asn1parse'. By implication, the functions ASN1_parse_dump() and
7937 BIO_dump_indent() are added.
7940 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7941 these print out strings and name structures based on various
7942 flags including RFC2253 support and proper handling of
7943 multibyte characters. Added options to the 'x509' utility
7944 to allow the various flags to be set.
7947 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7948 Also change the functions X509_cmp_current_time() and
7949 X509_gmtime_adj() work with an ASN1_TIME structure,
7950 this will enable certificates using GeneralizedTime in validity
7951 dates to be checked.
7954 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7955 negative public key encodings) on by default,
7956 NO_NEG_PUBKEY_BUG can be set to disable it.
7959 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7960 content octets. An i2c_ASN1_OBJECT is unnecessary because
7961 the encoding can be trivially obtained from the structure.
7964 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7965 not read locks (CRYPTO_r_[un]lock).
7968 *) A first attempt at creating official support for shared
7969 libraries through configuration. I've kept it so the
7970 default is static libraries only, and the OpenSSL programs
7971 are always statically linked for now, but there are
7972 preparations for dynamic linking in place.
7973 This has been tested on Linux and Tru64.
7976 *) Randomness polling function for Win9x, as described in:
7977 Peter Gutmann, Software Generation of Practically Strong
7981 *) Fix so PRNG is seeded in req if using an already existing
7985 *) New options to smime application. -inform and -outform
7986 allow alternative formats for the S/MIME message including
7987 PEM and DER. The -content option allows the content to be
7988 specified separately. This should allow things like Netscape
7989 form signing output easier to verify.
7992 *) Fix the ASN1 encoding of tags using the 'long form'.
7995 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7996 STRING types. These convert content octets to and from the
7997 underlying type. The actual tag and length octets are
7998 already assumed to have been read in and checked. These
7999 are needed because all other string types have virtually
8000 identical handling apart from the tag. By having versions
8001 of the ASN1 functions that just operate on content octets
8002 IMPLICIT tagging can be handled properly. It also allows
8003 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8004 and ASN1_INTEGER are identical apart from the tag.
8007 *) Change the handling of OID objects as follows:
8009 - New object identifiers are inserted in objects.txt, following
8010 the syntax given in objects.README.
8011 - objects.pl is used to process obj_mac.num and create a new
8013 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8016 This is currently kind of a hack, and the perl code in objects.pl
8017 isn't very elegant, but it works as I intended. The simplest way
8018 to check that it worked correctly is to look in obj_dat.h and
8019 check the array nid_objs and make sure the objects haven't moved
8020 around (this is important!). Additions are OK, as well as
8021 consistent name changes.
8024 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8027 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8028 The given file adds to whatever has already been seeded into the
8029 random pool through the RANDFILE configuration file option or
8030 environment variable, or the default random state file.
8033 *) mkstack.pl now sorts each macro group into lexical order.
8034 Previously the output order depended on the order the files
8035 appeared in the directory, resulting in needless rewriting
8039 *) Patches to make OpenSSL compile under Win32 again. Mostly
8040 work arounds for the VC++ problem that it treats func() as
8041 func(void). Also stripped out the parts of mkdef.pl that
8042 added extra typesafe functions: these no longer exist.
8045 *) Reorganisation of the stack code. The macros are now all
8046 collected in safestack.h . Each macro is defined in terms of
8047 a "stack macro" of the form SKM_<name>(type, a, b). The
8048 DEBUG_SAFESTACK is now handled in terms of function casts,
8049 this has the advantage of retaining type safety without the
8050 use of additional functions. If DEBUG_SAFESTACK is not defined
8051 then the non typesafe macros are used instead. Also modified the
8052 mkstack.pl script to handle the new form. Needs testing to see
8053 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8054 the default if no major problems. Similar behaviour for ASN1_SET_OF
8055 and PKCS12_STACK_OF.
8058 *) When some versions of IIS use the 'NET' form of private key the
8059 key derivation algorithm is different. Normally MD5(password) is
8060 used as a 128 bit RC4 key. In the modified case
8061 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8062 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8063 as the old Netscape_RSA functions except they have an additional
8064 'sgckey' parameter which uses the modified algorithm. Also added
8065 an -sgckey command line option to the rsa utility. Thanks to
8066 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8067 algorithm to openssl-dev.
8070 *) The evp_local.h macros were using 'c.##kname' which resulted in
8071 invalid expansion on some systems (SCO 5.0.5 for example).
8072 Corrected to 'c.kname'.
8073 [Phillip Porch <root@theporch.com>]
8075 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8076 a STACK of email addresses from a certificate or request, these look
8077 in the subject name and the subject alternative name extensions and
8078 omit any duplicate addresses.
8081 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8082 This makes DSA verification about 2 % faster.
8085 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8086 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8087 plus overhead for 1024 bit moduli).
8088 This makes exponentiations about 0.5 % faster for 1024 bit
8089 exponents (as measured by "openssl speed rsa2048").
8092 *) Rename memory handling macros to avoid conflicts with other
8094 Malloc => OPENSSL_malloc
8095 Malloc_locked => OPENSSL_malloc_locked
8096 Realloc => OPENSSL_realloc
8097 Free => OPENSSL_free
8100 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8101 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8104 *) CygWin32 support.
8105 [John Jarvie <jjarvie@newsguy.com>]
8107 *) The type-safe stack code has been rejigged. It is now only compiled
8108 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8109 by default all type-specific stack functions are "#define"d back to
8110 standard stack functions. This results in more streamlined output
8111 but retains the type-safety checking possibilities of the original
8115 *) The STACK code has been cleaned up, and certain type declarations
8116 that didn't make a lot of sense have been brought in line. This has
8117 also involved a cleanup of sorts in safestack.h to more correctly
8118 map type-safe stack functions onto their plain stack counterparts.
8119 This work has also resulted in a variety of "const"ifications of
8120 lots of the code, especially "_cmp" operations which should normally
8121 be prototyped with "const" parameters anyway.
8124 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8125 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8126 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8127 where all of 'md' is used each time the PRNG is used, but 'state'
8128 is used only indexed by a cyclic counter. As entropy may not be
8129 well distributed from the beginning, 'md' is important as a
8130 chaining variable. However, the output function chains only half
8131 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8132 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8133 in all of 'state' being rewritten, with the new values depending
8134 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8137 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8138 the handshake is continued after ssl_verify_cert_chain();
8139 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8140 can lead to 'unexplainable' connection aborts later.
8141 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8143 *) Major EVP API cipher revision.
8144 Add hooks for extra EVP features. This allows various cipher
8145 parameters to be set in the EVP interface. Support added for variable
8146 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8147 setting of RC2 and RC5 parameters.
8149 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8152 Remove lots of duplicated code from the EVP library. For example *every*
8153 cipher init() function handles the 'iv' in the same way according to the
8154 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8155 for CFB and OFB modes they zero ctx->num.
8157 New functionality allows removal of S/MIME code RC2 hack.
8159 Most of the routines have the same form and so can be declared in terms
8162 By shifting this to the top level EVP_CipherInit() it can be removed from
8163 all individual ciphers. If the cipher wants to handle IVs or keys
8164 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8167 Change lots of functions like EVP_EncryptUpdate() to now return a
8168 value: although software versions of the algorithms cannot fail
8169 any installed hardware versions can.
8172 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8173 this option is set, tolerate broken clients that send the negotiated
8174 protocol version number instead of the requested protocol version
8178 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8179 i.e. non-zero for export ciphersuites, zero otherwise.
8180 Previous versions had this flag inverted, inconsistent with
8181 rsa_tmp_cb (..._TMP_RSA_CB).
8182 [Bodo Moeller; problem reported by Amit Chopra]
8184 *) Add missing DSA library text string. Work around for some IIS
8185 key files with invalid SEQUENCE encoding.
8188 *) Add a document (doc/standards.txt) that list all kinds of standards
8189 and so on that are implemented in OpenSSL.
8192 *) Enhance c_rehash script. Old version would mishandle certificates
8193 with the same subject name hash and wouldn't handle CRLs at all.
8194 Added -fingerprint option to crl utility, to support new c_rehash
8198 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8201 *) Fix for SSL server purpose checking. Server checking was
8202 rejecting certificates which had extended key usage present
8203 but no ssl client purpose.
8204 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8206 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8207 is a little unclear about how a blank password is handled.
8208 Since the password in encoded as a BMPString with terminating
8209 double NULL a zero length password would end up as just the
8210 double NULL. However no password at all is different and is
8211 handled differently in the PKCS#12 key generation code. NS
8212 treats a blank password as zero length. MSIE treats it as no
8213 password on export: but it will try both on import. We now do
8214 the same: PKCS12_parse() tries zero length and no password if
8215 the password is set to "" or NULL (NULL is now a valid password:
8216 it wasn't before) as does the pkcs12 application.
8219 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8220 perror when PEM_read_bio_X509_REQ fails, the error message must
8221 be obtained from the error queue.
8224 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8225 it in ERR_remove_state if appropriate, and change ERR_get_state
8226 accordingly to avoid race conditions (this is necessary because
8227 thread_hash is no longer constant once set).
8230 *) Bugfix for linux-elf makefile.one.
8233 *) RSA_get_default_method() will now cause a default
8234 RSA_METHOD to be chosen if one doesn't exist already.
8235 Previously this was only set during a call to RSA_new()
8236 or RSA_new_method(NULL) meaning it was possible for
8237 RSA_get_default_method() to return NULL.
8240 *) Added native name translation to the existing DSO code
8241 that will convert (if the flag to do so is set) filenames
8242 that are sufficiently small and have no path information
8243 into a canonical native form. Eg. "blah" converted to
8244 "libblah.so" or "blah.dll" etc.
8247 *) New function ERR_error_string_n(e, buf, len) which is like
8248 ERR_error_string(e, buf), but writes at most 'len' bytes
8249 including the 0 terminator. For ERR_error_string_n, 'buf'
8251 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8253 *) CONF library reworked to become more general. A new CONF
8254 configuration file reader "class" is implemented as well as a
8255 new functions (NCONF_*, for "New CONF") to handle it. The now
8256 old CONF_* functions are still there, but are reimplemented to
8257 work in terms of the new functions. Also, a set of functions
8258 to handle the internal storage of the configuration data is
8259 provided to make it easier to write new configuration file
8260 reader "classes" (I can definitely see something reading a
8261 configuration file in XML format, for example), called _CONF_*,
8262 or "the configuration storage API"...
8264 The new configuration file reading functions are:
8266 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8267 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8269 NCONF_default, NCONF_WIN32
8271 NCONF_dump_fp, NCONF_dump_bio
8273 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8274 NCONF_new creates a new CONF object. This works in the same way
8275 as other interfaces in OpenSSL, like the BIO interface.
8276 NCONF_dump_* dump the internal storage of the configuration file,
8277 which is useful for debugging. All other functions take the same
8278 arguments as the old CONF_* functions wth the exception of the
8279 first that must be a `CONF *' instead of a `LHASH *'.
8281 To make it easer to use the new classes with the old CONF_* functions,
8282 the function CONF_set_default_method is provided.
8285 *) Add '-tls1' option to 'openssl ciphers', which was already
8286 mentioned in the documentation but had not been implemented.
8287 (This option is not yet really useful because even the additional
8288 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8291 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8292 OpenSSL-based applications) load shared libraries and bind to
8293 them in a portable way.
8294 [Geoff Thorpe, with contributions from Richard Levitte]
8296 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8298 *) Make sure _lrotl and _lrotr are only used with MSVC.
8300 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8301 (the default implementation of RAND_status).
8303 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8304 to '-clrext' (= clear extensions), as intended and documented.
8305 [Bodo Moeller; inconsistency pointed out by Michael Attili
8308 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8309 was larger than the MD block size.
8310 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8312 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8313 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8314 using the passed key: if the passed key was a private key the result
8315 of X509_print(), for example, would be to print out all the private key
8319 *) des_quad_cksum() byte order bug fix.
8320 [Ulf Möller, using the problem description in krb4-0.9.7, where
8321 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8323 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8325 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8327 *) For easily testing in shell scripts whether some command
8328 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8329 returns with exit code 0 iff no command of the given name is available.
8330 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8331 the output goes to stdout and nothing is printed to stderr.
8332 Additional arguments are always ignored.
8334 Since for each cipher there is a command of the same name,
8335 the 'no-cipher' compilation switches can be tested this way.
8337 ('openssl no-XXX' is not able to detect pseudo-commands such
8338 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8341 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8344 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8345 is set; it will be thrown away anyway because each handshake creates
8347 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8348 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8349 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8350 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8353 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8354 'Q' and 'R' lose their special meanings (quit/renegotiate).
8355 This is part of what -quiet does; unlike -quiet, -ign_eof
8356 does not suppress any output.
8359 *) Add compatibility options to the purpose and trust code. The
8360 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8361 accepts a certificate or CA, this was the previous behaviour,
8362 with all the associated security issues.
8364 X509_TRUST_COMPAT is the old trust behaviour: only and
8365 automatically trust self signed roots in certificate store. A
8366 new trust setting X509_TRUST_DEFAULT is used to specify that
8367 a purpose has no associated trust setting and it should instead
8368 use the value in the default purpose.
8371 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8372 and fix a memory leak.
8375 *) In util/mkerr.pl (which implements 'make errors'), preserve
8376 reason strings from the previous version of the .c file, as
8377 the default to have only downcase letters (and digits) in
8378 automatically generated reasons codes is not always appropriate.
8381 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8382 using strerror. Previously, ERR_reason_error_string() returned
8383 library names as reason strings for SYSerr; but SYSerr is a special
8384 case where small numbers are errno values, not library numbers.
8387 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8388 converts DSA parameters into DH parameters. (When creating parameters,
8389 DSA_generate_parameters is used.)
8392 *) Include 'length' (recommended exponent length) in C code generated
8393 by 'openssl dhparam -C'.
8396 *) The second argument to set_label in perlasm was already being used
8397 so couldn't be used as a "file scope" flag. Moved to third argument
8401 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8402 instead of RAND_bytes for encryption IVs and salts.
8405 *) Include RAND_status() into RAND_METHOD instead of implementing
8406 it only for md_rand.c Otherwise replacing the PRNG by calling
8407 RAND_set_rand_method would be impossible.
8410 *) Don't let DSA_generate_key() enter an infinite loop if the random
8411 number generation fails.
8414 *) New 'rand' application for creating pseudo-random output.
8417 *) Added configuration support for Linux/IA64
8418 [Rolf Haberrecker <rolf@suse.de>]
8420 *) Assembler module support for Mingw32.
8423 *) Shared library support for HPUX (in shlib/).
8424 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8426 *) Shared library support for Solaris gcc.
8427 [Lutz Behnke <behnke@trustcenter.de>]
8429 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8431 *) PKCS7_encrypt() was adding text MIME headers twice because they
8432 were added manually and by SMIME_crlf_copy().
8435 *) In bntest.c don't call BN_rand with zero bits argument.
8436 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8438 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8439 case was implemented. This caused BN_div_recp() to fail occasionally.
8442 *) Add an optional second argument to the set_label() in the perl
8443 assembly language builder. If this argument exists and is set
8444 to 1 it signals that the assembler should use a symbol whose
8445 scope is the entire file, not just the current function. This
8446 is needed with MASM which uses the format label:: for this scope.
8447 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8449 *) Change the ASN1 types so they are typedefs by default. Before
8450 almost all types were #define'd to ASN1_STRING which was causing
8451 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8455 *) Change names of new functions to the new get1/get0 naming
8456 convention: After 'get1', the caller owns a reference count
8457 and has to call ..._free; 'get0' returns a pointer to some
8458 data structure without incrementing reference counters.
8459 (Some of the existing 'get' functions increment a reference
8460 counter, some don't.)
8461 Similarly, 'set1' and 'add1' functions increase reference
8462 counters or duplicate objects.
8465 *) Allow for the possibility of temp RSA key generation failure:
8466 the code used to assume it always worked and crashed on failure.
8469 *) Fix potential buffer overrun problem in BIO_printf().
8470 [Ulf Möller, using public domain code by Patrick Powell; problem
8471 pointed out by David Sacerdote <das33@cornell.edu>]
8473 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8474 RAND_egd() and RAND_status(). In the command line application,
8475 the EGD socket can be specified like a seed file using RANDFILE
8479 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8480 Some CAs (e.g. Verisign) distribute certificates in this form.
8483 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8484 list to exclude them. This means that no special compilation option
8485 is needed to use anonymous DH: it just needs to be included in the
8489 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8490 EVP_MD_type. The old functionality is available in a new macro called
8491 EVP_MD_md(). Change code that uses it and update docs.
8494 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8495 where the 'void *' argument is replaced by a function pointer argument.
8496 Previously 'void *' was abused to point to functions, which works on
8497 many platforms, but is not correct. As these functions are usually
8498 called by macros defined in OpenSSL header files, most source code
8499 should work without changes.
8502 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8503 sections with information on -D... compiler switches used for
8504 compiling the library so that applications can see them. To enable
8505 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8506 must be defined. E.g.,
8507 #define OPENSSL_ALGORITHM_DEFINES
8508 #include <openssl/opensslconf.h>
8509 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8510 [Richard Levitte, Ulf and Bodo Möller]
8512 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8516 *) Change the 'other' type in certificate aux info to a STACK_OF
8517 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8518 the required ASN1 format: arbitrary types determined by an OID.
8521 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8522 argument to 'req'. This is not because the function is newer or
8523 better than others it just uses the work 'NEW' in the certificate
8524 request header lines. Some software needs this.
8527 *) Reorganise password command line arguments: now passwords can be
8528 obtained from various sources. Delete the PEM_cb function and make
8529 it the default behaviour: i.e. if the callback is NULL and the
8530 usrdata argument is not NULL interpret it as a null terminated pass
8531 phrase. If usrdata and the callback are NULL then the pass phrase
8532 is prompted for as usual.
8535 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8536 the support is automatically enabled. The resulting binaries will
8537 autodetect the card and use it if present.
8538 [Ben Laurie and Compaq Inc.]
8540 *) Work around for Netscape hang bug. This sends certificate request
8541 and server done in one record. Since this is perfectly legal in the
8542 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8543 the bugs/SSLv3 entry for more info.
8546 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8549 *) Add -rand argument to smime and pkcs12 applications and read/write
8553 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8556 *) Add command line password options to the remaining applications.
8559 *) Bug fix for BN_div_recp() for numerators with an even number of
8563 *) More tests in bntest.c, and changed test_bn output.
8566 *) ./config recognizes MacOS X now.
8569 *) Bug fix for BN_div() when the first words of num and divsor are
8570 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8573 *) Add support for various broken PKCS#8 formats, and command line
8574 options to produce them.
8577 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8578 get temporary BIGNUMs from a BN_CTX.
8581 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8585 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8586 include a #define from the old name to the new. The original intent
8587 was that statically linked binaries could for example just call
8588 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8589 link with digests. This never worked because SSLeay_add_all_digests()
8590 and SSLeay_add_all_ciphers() were in the same source file so calling
8591 one would link with the other. They are now in separate source files.
8594 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8597 *) Use a less unusual form of the Miller-Rabin primality test (it used
8598 a binary algorithm for exponentiation integrated into the Miller-Rabin
8599 loop, our standard modexp algorithms are faster).
8602 *) Support for the EBCDIC character set completed.
8603 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8605 *) Source code cleanups: use const where appropriate, eliminate casts,
8606 use void * instead of char * in lhash.
8609 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8610 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8611 this the server could overwrite ephemeral keys that the client
8615 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8616 using 50 iterations of the Rabin-Miller test.
8618 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8619 iterations of the Rabin-Miller test as required by the appendix
8620 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8621 As BN_is_prime_fasttest includes trial division, DSA parameter
8622 generation becomes much faster.
8624 This implies a change for the callback functions in DSA_is_prime
8625 and DSA_generate_parameters: The callback function is called once
8626 for each positive witness in the Rabin-Miller test, not just
8627 occasionally in the inner loop; and the parameters to the
8628 callback function now provide an iteration count for the outer
8629 loop rather than for the current invocation of the inner loop.
8630 DSA_generate_parameters additionally can call the callback
8631 function with an 'iteration count' of -1, meaning that a
8632 candidate has passed the trial division test (when q is generated
8633 from an application-provided seed, trial division is skipped).
8636 *) New function BN_is_prime_fasttest that optionally does trial
8637 division before starting the Rabin-Miller test and has
8638 an additional BN_CTX * argument (whereas BN_is_prime always
8639 has to allocate at least one BN_CTX).
8640 'callback(1, -1, cb_arg)' is called when a number has passed the
8641 trial division stage.
8644 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8648 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8651 *) New function BN_pseudo_rand().
8654 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8655 bignum version of BN_from_montgomery() with the working code from
8656 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8660 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8661 made it impossible to use the same SSL_SESSION data structure in
8662 SSL2 clients in multiple threads.
8665 *) The return value of RAND_load_file() no longer counts bytes obtained
8666 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8667 to seed the PRNG (previously an explicit byte count was required).
8668 [Ulf Möller, Bodo Möller]
8670 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8671 used (char *) instead of (void *) and had casts all over the place.
8674 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8677 *) Retain source code compatibility for BN_prime_checks macro:
8678 BN_is_prime(..., BN_prime_checks, ...) now uses
8679 BN_prime_checks_for_size to determine the appropriate number of
8680 Rabin-Miller iterations.
8683 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8684 DH_CHECK_P_NOT_SAFE_PRIME.
8685 (Check if this is true? OpenPGP calls them "strong".)
8688 *) Merge the functionality of "dh" and "gendh" programs into a new program
8689 "dhparam". The old programs are retained for now but will handle DH keys
8690 (instead of parameters) in future.
8693 *) Make the ciphers, s_server and s_client programs check the return values
8694 when a new cipher list is set.
8697 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8698 ciphers. Before when the 56bit ciphers were enabled the sorting was
8701 The syntax for the cipher sorting has been extended to support sorting by
8702 cipher-strength (using the strength_bits hard coded in the tables).
8703 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8705 Fix a bug in the cipher-command parser: when supplying a cipher command
8706 string with an "undefined" symbol (neither command nor alphanumeric
8707 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8708 an error is flagged.
8710 Due to the strength-sorting extension, the code of the
8711 ssl_create_cipher_list() function was completely rearranged. I hope that
8712 the readability was also increased :-)
8713 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8715 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8716 for the first serial number and places 2 in the serial number file. This
8717 avoids problems when the root CA is created with serial number zero and
8718 the first user certificate has the same issuer name and serial number
8722 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8723 the new code. Add documentation for this stuff.
8726 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8727 X509_*() to X509at_*() on the grounds that they don't handle X509
8728 structures and behave in an analogous way to the X509v3 functions:
8729 they shouldn't be called directly but wrapper functions should be used
8732 So we also now have some wrapper functions that call the X509at functions
8733 when passed certificate requests. (TO DO: similar things can be done with
8734 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8735 things. Some of these need some d2i or i2d and print functionality
8736 because they handle more complex structures.)
8739 *) Add missing #ifndefs that caused missing symbols when building libssl
8740 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8741 NO_RSA in ssl/s2*.c.
8742 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8744 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8745 has a return value which indicates the quality of the random data
8746 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8747 error queue. New function RAND_pseudo_bytes() generates output that is
8748 guaranteed to be unique but not unpredictable. RAND_add is like
8749 RAND_seed, but takes an extra argument for an entropy estimate
8750 (RAND_seed always assumes full entropy).
8753 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8754 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8755 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8756 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8757 false-positive rate of at most 2^-80 for random input.
8760 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8763 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8764 in the 0.9.5 release), this returns the chain
8765 from an X509_CTX structure with a dup of the stack and all
8766 the X509 reference counts upped: so the stack will exist
8767 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8770 Also make SSL_SESSION_print() print out the verify return
8774 *) Add manpage for the pkcs12 command. Also change the default
8775 behaviour so MAC iteration counts are used unless the new
8776 -nomaciter option is used. This improves file security and
8777 only older versions of MSIE (4.0 for example) need it.
8780 *) Honor the no-xxx Configure options when creating .DEF files.
8783 *) Add PKCS#10 attributes to field table: challengePassword,
8784 unstructuredName and unstructuredAddress. These are taken from
8785 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8786 international characters are used.
8788 More changes to X509_ATTRIBUTE code: allow the setting of types
8789 based on strings. Remove the 'loc' parameter when adding
8790 attributes because these will be a SET OF encoding which is sorted
8794 *) Initial changes to the 'req' utility to allow request generation
8795 automation. This will allow an application to just generate a template
8796 file containing all the field values and have req construct the
8799 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8800 used all over the place including certificate requests and PKCS#7
8801 structures. They are currently handled manually where necessary with
8802 some primitive wrappers for PKCS#7. The new functions behave in a
8803 manner analogous to the X509 extension functions: they allow
8804 attributes to be looked up by NID and added.
8806 Later something similar to the X509V3 code would be desirable to
8807 automatically handle the encoding, decoding and printing of the
8808 more complex types. The string types like challengePassword can
8809 be handled by the string table functions.
8811 Also modified the multi byte string table handling. Now there is
8812 a 'global mask' which masks out certain types. The table itself
8813 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8814 is useful when for example there is only one permissible type
8815 (as in countryName) and using the mask might result in no valid
8819 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8820 SSL_get_peer_finished to allow applications to obtain the latest
8821 Finished messages sent to the peer or expected from the peer,
8822 respectively. (SSL_get_peer_finished is usually the Finished message
8823 actually received from the peer, otherwise the protocol will be aborted.)
8825 As the Finished message are message digests of the complete handshake
8826 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8827 be used for external authentication procedures when the authentication
8828 provided by SSL/TLS is not desired or is not enough.
8831 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8832 the host supports BWX extension and if Compaq C is present on the
8833 $PATH. Just exploiting of the BWX extension results in 20-30%
8834 performance kick for some algorithms, e.g. DES and RC4 to mention
8835 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8839 *) Add support for MS "fast SGC". This is arguably a violation of the
8840 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8841 weak crypto and after checking the certificate is SGC a second one
8842 with strong crypto. MS SGC stops the first handshake after receiving
8843 the server certificate message and sends a second client hello. Since
8844 a server will typically do all the time consuming operations before
8845 expecting any further messages from the client (server key exchange
8846 is the most expensive) there is little difference between the two.
8848 To get OpenSSL to support MS SGC we have to permit a second client
8849 hello message after we have sent server done. In addition we have to
8850 reset the MAC if we do get this second client hello.
8853 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8854 if a DER encoded private key is RSA or DSA traditional format. Changed
8855 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8856 format DER encoded private key. Newer code should use PKCS#8 format which
8857 has the key type encoded in the ASN1 structure. Added DER private key
8858 support to pkcs8 application.
8861 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8862 ciphersuites has been selected (as required by the SSL 3/TLS 1
8863 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8864 is set, we interpret this as a request to violate the specification
8865 (the worst that can happen is a handshake failure, and 'correct'
8866 behaviour would result in a handshake failure anyway).
8869 *) In SSL_CTX_add_session, take into account that there might be multiple
8870 SSL_SESSION structures with the same session ID (e.g. when two threads
8871 concurrently obtain them from an external cache).
8872 The internal cache can handle only one SSL_SESSION with a given ID,
8873 so if there's a conflict, we now throw out the old one to achieve
8877 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8878 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8879 some routines that use cipher OIDs: some ciphers do not have OIDs
8880 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8884 *) Simplify the trust setting structure and code. Now we just have
8885 two sequences of OIDs for trusted and rejected settings. These will
8886 typically have values the same as the extended key usage extension
8887 and any application specific purposes.
8889 The trust checking code now has a default behaviour: it will just
8890 check for an object with the same NID as the passed id. Functions can
8891 be provided to override either the default behaviour or the behaviour
8892 for a given id. SSL client, server and email already have functions
8893 in place for compatibility: they check the NID and also return "trusted"
8894 if the certificate is self signed.
8897 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8898 traditional format into an EVP_PKEY structure.
8901 *) Add a password callback function PEM_cb() which either prompts for
8902 a password if usr_data is NULL or otherwise assumes it is a null
8903 terminated password. Allow passwords to be passed on command line
8904 environment or config files in a few more utilities.
8907 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8908 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8909 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8910 Update documentation.
8913 *) Support for ASN1 "NULL" type. This could be handled before by using
8914 ASN1_TYPE but there wasn't any function that would try to read a NULL
8915 and produce an error if it couldn't. For compatibility we also have
8916 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8917 don't allocate anything because they don't need to.
8920 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8922 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8924 *) Rebuild of the memory allocation routines used by OpenSSL code and
8925 possibly others as well. The purpose is to make an interface that
8926 provide hooks so anyone can build a separate set of allocation and
8927 deallocation routines to be used by OpenSSL, for example memory
8928 pool implementations, or something else, which was previously hard
8929 since Malloc(), Realloc() and Free() were defined as macros having
8930 the values malloc, realloc and free, respectively (except for Win32
8931 compilations). The same is provided for memory debugging code.
8932 OpenSSL already comes with functionality to find memory leaks, but
8933 this gives people a chance to debug other memory problems.
8935 With these changes, a new set of functions and macros have appeared:
8937 CRYPTO_set_mem_debug_functions() [F]
8938 CRYPTO_get_mem_debug_functions() [F]
8939 CRYPTO_dbg_set_options() [F]
8940 CRYPTO_dbg_get_options() [F]
8941 CRYPTO_malloc_debug_init() [M]
8943 The memory debug functions are NULL by default, unless the library
8944 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8945 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8946 gives the standard debugging functions that come with OpenSSL) or
8947 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8948 provided by the library user) must be used. When the standard
8949 debugging functions are used, CRYPTO_dbg_set_options can be used to
8950 request additional information:
8951 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8952 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8954 Also, things like CRYPTO_set_mem_functions will always give the
8955 expected result (the new set of functions is used for allocation
8956 and deallocation) at all times, regardless of platform and compiler
8959 To finish it up, some functions that were never use in any other
8960 way than through macros have a new API and new semantic:
8963 CRYPTO_dbg_realloc()
8966 All macros of value have retained their old syntax.
8967 [Richard Levitte and Bodo Moeller]
8969 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8970 ordering of SMIMECapabilities wasn't in "strength order" and there
8971 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8975 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8976 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8977 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8979 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8980 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8981 functionality to handle multipart/signed properly) and a utility
8982 called 'smime' to call all this stuff. This is based on code I
8983 originally wrote for Celo who have kindly allowed it to be
8984 included in OpenSSL.
8987 *) Add variants des_set_key_checked and des_set_key_unchecked of
8988 des_set_key (aka des_key_sched). Global variable des_check_key
8989 decides which of these is called by des_set_key; this way
8990 des_check_key behaves as it always did, but applications and
8991 the library itself, which was buggy for des_check_key == 1,
8992 have a cleaner way to pick the version they need.
8995 *) New function PKCS12_newpass() which changes the password of a
8999 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9000 dynamic mix. In both cases the ids can be used as an index into the
9001 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9002 functions so they accept a list of the field values and the
9003 application doesn't need to directly manipulate the X509_TRUST
9007 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9011 *) Modify the way the V3 extension code looks up extensions. This now
9012 works in a similar way to the object code: we have some "standard"
9013 extensions in a static table which is searched with OBJ_bsearch()
9014 and the application can add dynamic ones if needed. The file
9015 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9016 updated whenever a new extension is added to the core code and kept
9017 in ext_nid order. There is a simple program 'tabtest.c' which checks
9018 this. New extensions are not added too often so this file can readily
9019 be maintained manually.
9021 There are two big advantages in doing things this way. The extensions
9022 can be looked up immediately and no longer need to be "added" using
9023 X509V3_add_standard_extensions(): this function now does nothing.
9024 [Side note: I get *lots* of email saying the extension code doesn't
9025 work because people forget to call this function]
9026 Also no dynamic allocation is done unless new extensions are added:
9027 so if we don't add custom extensions there is no need to call
9028 X509V3_EXT_cleanup().
9031 *) Modify enc utility's salting as follows: make salting the default. Add a
9032 magic header, so unsalted files fail gracefully instead of just decrypting
9033 to garbage. This is because not salting is a big security hole, so people
9034 should be discouraged from doing it.
9037 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9038 digest to be passed on the command line but it only used this
9039 parameter when signing a certificate. Modified so all relevant
9040 operations are affected by the digest parameter including the
9041 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9042 DSA key was used because it didn't fix the digest.
9045 *) Initial certificate chain verify code. Currently tests the untrusted
9046 certificates for consistency with the verify purpose (which is set
9047 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9049 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9050 this is because it will reject chains with invalid extensions whereas
9051 every previous version of OpenSSL and SSLeay made no checks at all.
9053 Trust code: checks the root CA for the relevant trust settings. Trust
9054 settings have an initial value consistent with the verify purpose: e.g.
9055 if the verify purpose is for SSL client use it expects the CA to be
9056 trusted for SSL client use. However the default value can be changed to
9057 permit custom trust settings: one example of this would be to only trust
9058 certificates from a specific "secure" set of CAs.
9060 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9061 which should be used for version portability: especially since the
9062 verify structure is likely to change more often now.
9064 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9065 to set them. If not set then assume SSL clients will verify SSL servers
9068 Two new options to the verify program: -untrusted allows a set of
9069 untrusted certificates to be passed in and -purpose which sets the
9070 intended purpose of the certificate. If a purpose is set then the
9071 new chain verify code is used to check extension consistency.
9074 *) Support for the authority information access extension.
9077 *) Modify RSA and DSA PEM read routines to transparently handle
9078 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9079 public keys in a format compatible with certificate
9080 SubjectPublicKeyInfo structures. Unfortunately there were already
9081 functions called *_PublicKey_* which used various odd formats so
9082 these are retained for compatibility: however the DSA variants were
9083 never in a public release so they have been deleted. Changed dsa/rsa
9084 utilities to handle the new format: note no releases ever handled public
9085 keys so we should be OK.
9087 The primary motivation for this change is to avoid the same fiasco
9088 that dogs private keys: there are several incompatible private key
9089 formats some of which are standard and some OpenSSL specific and
9090 require various evil hacks to allow partial transparent handling and
9091 even then it doesn't work with DER formats. Given the option anything
9092 other than PKCS#8 should be dumped: but the other formats have to
9093 stay in the name of compatibility.
9095 With public keys and the benefit of hindsight one standard format
9096 is used which works with EVP_PKEY, RSA or DSA structures: though
9097 it clearly returns an error if you try to read the wrong kind of key.
9099 Added a -pubkey option to the 'x509' utility to output the public key.
9100 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9101 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9102 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9103 that do the same as the EVP_PKEY_assign_*() except they up the
9104 reference count of the added key (they don't "swallow" the
9108 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9109 CRLs would fail if the file contained no certificates or no CRLs:
9110 added a new function to read in both types and return the number
9111 read: this means that if none are read it will be an error. The
9112 DER versions of the certificate and CRL reader would always fail
9113 because it isn't possible to mix certificates and CRLs in DER format
9114 without choking one or the other routine. Changed this to just read
9115 a certificate: this is the best we can do. Also modified the code
9116 in apps/verify.c to take notice of return codes: it was previously
9117 attempting to read in certificates from NULL pointers and ignoring
9118 any errors: this is one reason why the cert and CRL reader seemed
9119 to work. It doesn't check return codes from the default certificate
9120 routines: these may well fail if the certificates aren't installed.
9123 *) Code to support otherName option in GeneralName.
9126 *) First update to verify code. Change the verify utility
9127 so it warns if it is passed a self signed certificate:
9128 for consistency with the normal behaviour. X509_verify
9129 has been modified to it will now verify a self signed
9130 certificate if *exactly* the same certificate appears
9131 in the store: it was previously impossible to trust a
9132 single self signed certificate. This means that:
9133 openssl verify ss.pem
9134 now gives a warning about a self signed certificate but
9135 openssl verify -CAfile ss.pem ss.pem
9139 *) For servers, store verify_result in SSL_SESSION data structure
9140 (and add it to external session representation).
9141 This is needed when client certificate verifications fails,
9142 but an application-provided verification callback (set by
9143 SSL_CTX_set_cert_verify_callback) allows accepting the session
9144 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9145 but returns 1): When the session is reused, we have to set
9146 ssl->verify_result to the appropriate error code to avoid
9148 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9150 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9151 case in PKCS7_dataInit() where the signed PKCS7 structure
9152 didn't contain any existing data because it was being created.
9153 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9155 *) Add a salt to the key derivation routines in enc.c. This
9156 forms the first 8 bytes of the encrypted file. Also add a
9157 -S option to allow a salt to be input on the command line.
9160 *) New function X509_cmp(). Oddly enough there wasn't a function
9161 to compare two certificates. We do this by working out the SHA1
9162 hash and comparing that. X509_cmp() will be needed by the trust
9166 *) SSL_get1_session() is like SSL_get_session(), but increments
9167 the reference count in the SSL_SESSION returned.
9168 [Geoff Thorpe <geoff@eu.c2.net>]
9170 *) Fix for 'req': it was adding a null to request attributes.
9171 Also change the X509_LOOKUP and X509_INFO code to handle
9172 certificate auxiliary information.
9175 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9179 *) Add the possibility to add extra information to the memory leak
9180 detecting output, to form tracebacks, showing from where each
9181 allocation was originated: CRYPTO_push_info("constant string") adds
9182 the string plus current file name and line number to a per-thread
9183 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9184 is like calling CYRPTO_pop_info() until the stack is empty.
9185 Also updated memory leak detection code to be multi-thread-safe.
9188 *) Add options -text and -noout to pkcs7 utility and delete the
9189 encryption options which never did anything. Update docs.
9192 *) Add options to some of the utilities to allow the pass phrase
9193 to be included on either the command line (not recommended on
9194 OSes like Unix) or read from the environment. Update the
9195 manpages and fix a few bugs.
9198 *) Add a few manpages for some of the openssl commands.
9201 *) Fix the -revoke option in ca. It was freeing up memory twice,
9202 leaking and not finding already revoked certificates.
9205 *) Extensive changes to support certificate auxiliary information.
9206 This involves the use of X509_CERT_AUX structure and X509_AUX
9207 functions. An X509_AUX function such as PEM_read_X509_AUX()
9208 can still read in a certificate file in the usual way but it
9209 will also read in any additional "auxiliary information". By
9210 doing things this way a fair degree of compatibility can be
9211 retained: existing certificates can have this information added
9212 using the new 'x509' options.
9214 Current auxiliary information includes an "alias" and some trust
9215 settings. The trust settings will ultimately be used in enhanced
9216 certificate chain verification routines: currently a certificate
9217 can only be trusted if it is self signed and then it is trusted
9221 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9222 The problem was that one of the replacement routines had not been working
9223 since SSLeay releases. For now the offending routine has been replaced
9224 with non-optimised assembler. Even so, this now gives around 95%
9225 performance improvement for 1024 bit RSA signs.
9228 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9229 handling. Most clients have the effective key size in bits equal to
9230 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9231 A few however don't do this and instead use the size of the decrypted key
9232 to determine the RC2 key length and the AlgorithmIdentifier to determine
9233 the effective key length. In this case the effective key length can still
9234 be 40 bits but the key length can be 168 bits for example. This is fixed
9235 by manually forcing an RC2 key into the EVP_PKEY structure because the
9236 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9237 the key length and effective key length are equal.
9240 *) Add a bunch of functions that should simplify the creation of
9241 X509_NAME structures. Now you should be able to do:
9242 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9243 and have it automatically work out the correct field type and fill in
9244 the structures. The more adventurous can try:
9245 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9246 and it will (hopefully) work out the correct multibyte encoding.
9249 *) Change the 'req' utility to use the new field handling and multibyte
9250 copy routines. Before the DN field creation was handled in an ad hoc
9251 way in req, ca, and x509 which was rather broken and didn't support
9252 BMPStrings or UTF8Strings. Since some software doesn't implement
9253 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9254 using the dirstring_type option. See the new comment in the default
9255 openssl.cnf for more info.
9258 *) Make crypto/rand/md_rand.c more robust:
9259 - Assure unique random numbers after fork().
9260 - Make sure that concurrent threads access the global counter and
9261 md serializably so that we never lose entropy in them
9262 or use exactly the same state in multiple threads.
9263 Access to the large state is not always serializable because
9264 the additional locking could be a performance killer, and
9265 md should be large enough anyway.
9268 *) New file apps/app_rand.c with commonly needed functionality
9269 for handling the random seed file.
9271 Use the random seed file in some applications that previously did not:
9273 dsaparam -genkey (which also ignored its '-rand' option),
9276 x509 (when signing).
9277 Except on systems with /dev/urandom, it is crucial to have a random
9278 seed file at least for key creation, DSA signing, and for DH exchanges;
9279 for RSA signatures we could do without one.
9281 gendh and gendsa (unlike genrsa) used to read only the first byte
9282 of each file listed in the '-rand' option. The function as previously
9283 found in genrsa is now in app_rand.c and is used by all programs
9284 that support '-rand'.
9287 *) In RAND_write_file, use mode 0600 for creating files;
9288 don't just chmod when it may be too late.
9291 *) Report an error from X509_STORE_load_locations
9292 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9295 *) New function ASN1_mbstring_copy() this copies a string in either
9296 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9297 into an ASN1_STRING type. A mask of permissible types is passed
9298 and it chooses the "minimal" type to use or an error if not type
9302 *) Add function equivalents to the various macros in asn1.h. The old
9303 macros are retained with an M_ prefix. Code inside the library can
9304 use the M_ macros. External code (including the openssl utility)
9305 should *NOT* in order to be "shared library friendly".
9308 *) Add various functions that can check a certificate's extensions
9309 to see if it usable for various purposes such as SSL client,
9310 server or S/MIME and CAs of these types. This is currently
9311 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9312 verification. Also added a -purpose flag to x509 utility to
9313 print out all the purposes.
9316 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9320 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9321 for, obtain and decode and extension and obtain its critical flag.
9322 This allows all the necessary extension code to be handled in a
9323 single function call.
9326 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9327 platforms. See crypto/rc4/rc4_enc.c for further details.
9330 *) New -noout option to asn1parse. This causes no output to be produced
9331 its main use is when combined with -strparse and -out to extract data
9332 from a file (which may not be in ASN.1 format).
9335 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9336 when producing the local key id.
9337 [Richard Levitte <levitte@stacken.kth.se>]
9339 *) New option -dhparam in s_server. This allows a DH parameter file to be
9340 stated explicitly. If it is not stated then it tries the first server
9341 certificate file. The previous behaviour hard coded the filename
9345 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9346 a public key to be input or output. For example:
9347 openssl rsa -in key.pem -pubout -out pubkey.pem
9348 Also added necessary DSA public key functions to handle this.
9351 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9352 in the message. This was handled by allowing
9353 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9354 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9356 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9357 to the end of the strings whereas this didn't. This would cause problems
9358 if strings read with d2i_ASN1_bytes() were later modified.
9359 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9361 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9362 data and it contains EOF it will end up returning an error. This is
9363 caused by input 46 bytes long. The cause is due to the way base64
9364 BIOs find the start of base64 encoded data. They do this by trying a
9365 trial decode on each line until they find one that works. When they
9366 do a flag is set and it starts again knowing it can pass all the
9367 data directly through the decoder. Unfortunately it doesn't reset
9368 the context it uses. This means that if EOF is reached an attempt
9369 is made to pass two EOFs through the context and this causes the
9370 resulting error. This can also cause other problems as well. As is
9371 usual with these problems it takes *ages* to find and the fix is
9372 trivial: move one line.
9373 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9375 *) Ugly workaround to get s_client and s_server working under Windows. The
9376 old code wouldn't work because it needed to select() on sockets and the
9377 tty (for keypresses and to see if data could be written). Win32 only
9378 supports select() on sockets so we select() with a 1s timeout on the
9379 sockets and then see if any characters are waiting to be read, if none
9380 are present then we retry, we also assume we can always write data to
9381 the tty. This isn't nice because the code then blocks until we've
9382 received a complete line of data and it is effectively polling the
9383 keyboard at 1s intervals: however it's quite a bit better than not
9384 working at all :-) A dedicated Windows application might handle this
9385 with an event loop for example.
9388 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9389 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9390 will be called when RSA_sign() and RSA_verify() are used. This is useful
9391 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9392 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9393 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9394 This necessitated the support of an extra signature type NID_md5_sha1
9395 for SSL signatures and modifications to the SSL library to use it instead
9396 of calling RSA_public_decrypt() and RSA_private_encrypt().
9399 *) Add new -verify -CAfile and -CApath options to the crl program, these
9400 will lookup a CRL issuers certificate and verify the signature in a
9401 similar way to the verify program. Tidy up the crl program so it
9402 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9403 less strict. It will now permit CRL extensions even if it is not
9404 a V2 CRL: this will allow it to tolerate some broken CRLs.
9407 *) Initialize all non-automatic variables each time one of the openssl
9408 sub-programs is started (this is necessary as they may be started
9409 multiple times from the "OpenSSL>" prompt).
9410 [Lennart Bang, Bodo Moeller]
9412 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9413 removing all other RSA functionality (this is what NO_RSA does). This
9414 is so (for example) those in the US can disable those operations covered
9415 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9419 *) Non-copying interface to BIO pairs.
9420 (still largely untested)
9423 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9424 ASCII string. This was handled independently in various places before.
9427 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9428 UTF8 strings a character at a time.
9431 *) Use client_version from client hello to select the protocol
9432 (s23_srvr.c) and for RSA client key exchange verification
9433 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9436 *) Add various utility functions to handle SPKACs, these were previously
9437 handled by poking round in the structure internals. Added new function
9438 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9439 print, verify and generate SPKACs. Based on an original idea from
9440 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9443 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9446 *) Allow the config file extension section to be overwritten on the
9447 command line. Based on an original idea from Massimiliano Pala
9448 <madwolf@comune.modena.it>. The new option is called -extensions
9449 and can be applied to ca, req and x509. Also -reqexts to override
9450 the request extensions in req and -crlexts to override the crl extensions
9454 *) Add new feature to the SPKAC handling in ca. Now you can include
9455 the same field multiple times by preceding it by "XXXX." for example:
9458 this is the same syntax as used in the req config file.
9461 *) Allow certificate extensions to be added to certificate requests. These
9462 are specified in a 'req_extensions' option of the req section of the
9463 config file. They can be printed out with the -text option to req but
9464 are otherwise ignored at present.
9467 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9468 data read consists of only the final block it would not decrypted because
9469 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9470 A misplaced 'break' also meant the decrypted final block might not be
9471 copied until the next read.
9474 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9475 a few extra parameters to the DH structure: these will be useful if
9476 for example we want the value of 'q' or implement X9.42 DH.
9479 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9480 provides hooks that allow the default DSA functions or functions on a
9481 "per key" basis to be replaced. This allows hardware acceleration and
9482 hardware key storage to be handled without major modification to the
9483 library. Also added low level modexp hooks and CRYPTO_EX structure and
9484 associated functions.
9487 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9488 as "read only": it can't be written to and the buffer it points to will
9489 not be freed. Reading from a read only BIO is much more efficient than
9490 a normal memory BIO. This was added because there are several times when
9491 an area of memory needs to be read from a BIO. The previous method was
9492 to create a memory BIO and write the data to it, this results in two
9493 copies of the data and an O(n^2) reading algorithm. There is a new
9494 function BIO_new_mem_buf() which creates a read only memory BIO from
9495 an area of memory. Also modified the PKCS#7 routines to use read only
9499 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9500 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9501 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9502 but a retry condition occured while trying to read the rest.
9505 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9506 NID_pkcs7_encrypted by default: this was wrong since this should almost
9507 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9508 the encrypted data type: this is a more sensible place to put it and it
9509 allows the PKCS#12 code to be tidied up that duplicated this
9513 *) Changed obj_dat.pl script so it takes its input and output files on
9514 the command line. This should avoid shell escape redirection problems
9518 *) Initial support for certificate extension requests, these are included
9519 in things like Xenroll certificate requests. Included functions to allow
9520 extensions to be obtained and added.
9523 *) -crlf option to s_client and s_server for sending newlines as
9524 CRLF (as required by many protocols).
9527 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9529 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9530 [Ralf S. Engelschall]
9532 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9533 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9535 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9539 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9540 DH parameters/keys (q is lost during that conversion, but the resulting
9541 DH parameters contain its length).
9543 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9544 much faster than DH_generate_parameters (which creates parameters
9545 where p = 2*q + 1), and also the smaller q makes DH computations
9546 much more efficient (160-bit exponentiation instead of 1024-bit
9547 exponentiation); so this provides a convenient way to support DHE
9548 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9549 utter importance to use
9550 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9552 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9553 when such DH parameters are used, because otherwise small subgroup
9554 attacks may become possible!
9557 *) Avoid memory leak in i2d_DHparams.
9560 *) Allow the -k option to be used more than once in the enc program:
9561 this allows the same encrypted message to be read by multiple recipients.
9564 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9565 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9566 it will always use the numerical form of the OID, even if it has a short
9570 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9571 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9572 otherwise bn_mod_exp was called. In the case of hardware keys for example
9573 no private key components need be present and it might store extra data
9574 in the RSA structure, which cannot be accessed from bn_mod_exp.
9575 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9576 private key operations.
9579 *) Added support for SPARC Linux.
9582 *) pem_password_cb function type incompatibly changed from
9583 typedef int pem_password_cb(char *buf, int size, int rwflag);
9585 ....(char *buf, int size, int rwflag, void *userdata);
9586 so that applications can pass data to their callbacks:
9587 The PEM[_ASN1]_{read,write}... functions and macros now take an
9588 additional void * argument, which is just handed through whenever
9589 the password callback is called.
9590 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9592 New function SSL_CTX_set_default_passwd_cb_userdata.
9594 Compatibility note: As many C implementations push function arguments
9595 onto the stack in reverse order, the new library version is likely to
9596 interoperate with programs that have been compiled with the old
9597 pem_password_cb definition (PEM_whatever takes some data that
9598 happens to be on the stack as its last argument, and the callback
9599 just ignores this garbage); but there is no guarantee whatsoever that
9602 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9603 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9604 problems not only on Windows, but also on some Unix platforms.
9605 To avoid problematic command lines, these definitions are now in an
9606 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9607 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9610 *) MIPS III/IV assembler module is reimplemented.
9613 *) More DES library cleanups: remove references to srand/rand and
9614 delete an unused file.
9617 *) Add support for the the free Netwide assembler (NASM) under Win32,
9618 since not many people have MASM (ml) and it can be hard to obtain.
9619 This is currently experimental but it seems to work OK and pass all
9620 the tests. Check out INSTALL.W32 for info.
9623 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9624 without temporary keys kept an extra copy of the server key,
9625 and connections with temporary keys did not free everything in case
9629 *) New function RSA_check_key and new openssl rsa option -check
9630 for verifying the consistency of RSA keys.
9631 [Ulf Moeller, Bodo Moeller]
9633 *) Various changes to make Win32 compile work:
9634 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9635 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9636 comparison" warnings.
9637 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9640 *) Add a debugging option to PKCS#5 v2 key generation function: when
9641 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9642 derived keys are printed to stderr.
9645 *) Copy the flags in ASN1_STRING_dup().
9646 [Roman E. Pavlov <pre@mo.msk.ru>]
9648 *) The x509 application mishandled signing requests containing DSA
9649 keys when the signing key was also DSA and the parameters didn't match.
9651 It was supposed to omit the parameters when they matched the signing key:
9652 the verifying software was then supposed to automatically use the CA's
9653 parameters if they were absent from the end user certificate.
9655 Omitting parameters is no longer recommended. The test was also
9656 the wrong way round! This was probably due to unusual behaviour in
9657 EVP_cmp_parameters() which returns 1 if the parameters match.
9658 This meant that parameters were omitted when they *didn't* match and
9659 the certificate was useless. Certificates signed with 'ca' didn't have
9661 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9663 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9664 The interface is as follows:
9665 Applications can use
9666 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9667 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9668 "off" is now the default.
9669 The library internally uses
9670 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9671 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9672 to disable memory-checking temporarily.
9674 Some inconsistent states that previously were possible (and were
9675 even the default) are now avoided.
9677 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9678 with each memory chunk allocated; this is occasionally more helpful
9679 than just having a counter.
9681 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9683 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9687 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9688 which largely parallels "options", but is for changing API behaviour,
9689 whereas "options" are about protocol behaviour.
9690 Initial "mode" flags are:
9692 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9693 a single record has been written.
9694 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9695 retries use the same buffer location.
9696 (But all of the contents must be
9700 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9703 *) Fix problems with no-hmac etc.
9704 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9706 *) New functions RSA_get_default_method(), RSA_set_method() and
9707 RSA_get_method(). These allows replacement of RSA_METHODs without having
9708 to mess around with the internals of an RSA structure.
9711 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9712 Also really enable memory leak checks in openssl.c and in some
9714 [Chad C. Mulligan, Bodo Moeller]
9716 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9717 up the length of negative integers. This has now been simplified to just
9718 store the length when it is first determined and use it later, rather
9719 than trying to keep track of where data is copied and updating it to
9721 [Steve Henson, reported by Brien Wheeler
9722 <bwheeler@authentica-security.com>]
9724 *) Add a new function PKCS7_signatureVerify. This allows the verification
9725 of a PKCS#7 signature but with the signing certificate passed to the
9726 function itself. This contrasts with PKCS7_dataVerify which assumes the
9727 certificate is present in the PKCS#7 structure. This isn't always the
9728 case: certificates can be omitted from a PKCS#7 structure and be
9729 distributed by "out of band" means (such as a certificate database).
9732 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9733 function prototypes in pem.h, also change util/mkdef.pl to add the
9734 necessary function names.
9737 *) mk1mf.pl (used by Windows builds) did not properly read the
9738 options set by Configure in the top level Makefile, and Configure
9739 was not even able to write more than one option correctly.
9740 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9743 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9744 file to be loaded from a BIO or FILE pointer. The BIO version will
9745 for example allow memory BIOs to contain config info.
9748 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9749 Whoever hopes to achieve shared-library compatibility across versions
9750 must use this, not the compile-time macro.
9751 (Exercise 0.9.4: Which is the minimum library version required by
9753 Note: All this applies only to multi-threaded programs, others don't
9757 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9758 through a BIO pair triggered the default case, i.e.
9759 SSLerr(...,SSL_R_UNKNOWN_STATE).
9762 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9763 can use the SSL library even if none of the specific BIOs is
9767 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9768 for the encoded length.
9769 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9771 *) Add initial documentation of the X509V3 functions.
9774 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9775 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9776 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9777 secure PKCS#8 private key format with a high iteration count.
9780 *) Fix determination of Perl interpreter: A perl or perl5
9781 _directory_ in $PATH was also accepted as the interpreter.
9782 [Ralf S. Engelschall]
9784 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9785 wrong with it but it was very old and did things like calling
9786 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9790 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9791 to use the new extension code.
9794 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9795 with macros. This should make it easier to change their form, add extra
9796 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9800 *) Add to configuration table a new entry that can specify an alternative
9801 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9802 according to Mark Crispin <MRC@Panda.COM>.
9806 *) DES CBC did not update the IV. Weird.
9809 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9810 Changing the behaviour of the former might break existing programs --
9811 where IV updating is needed, des_ncbc_encrypt can be used.
9814 *) When bntest is run from "make test" it drives bc to check its
9815 calculations, as well as internally checking them. If an internal check
9816 fails, it needs to cause bc to give a non-zero result or make test carries
9817 on without noticing the failure. Fixed.
9820 *) DES library cleanups.
9823 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9824 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9825 ciphers. NOTE: although the key derivation function has been verified
9826 against some published test vectors it has not been extensively tested
9827 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9831 *) Instead of "mkdir -p", which is not fully portable, use new
9832 Perl script "util/mkdir-p.pl".
9835 *) Rewrite the way password based encryption (PBE) is handled. It used to
9836 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9837 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9838 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9839 the 'parameter' field of the AlgorithmIdentifier is passed to the
9840 underlying key generation function so it must do its own ASN1 parsing.
9841 This has also changed the EVP_PBE_CipherInit() function which now has a
9842 'parameter' argument instead of literal salt and iteration count values
9843 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9846 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9847 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9848 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9849 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9850 value was just used as a "magic string" and not used directly its
9851 value doesn't matter.
9854 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9858 *) "linux-sparc64" configuration (ultrapenguin).
9859 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9860 "linux-sparc" configuration.
9861 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9863 *) config now generates no-xxx options for missing ciphers.
9866 *) Support the EBCDIC character set (work in progress).
9867 File ebcdic.c not yet included because it has a different license.
9868 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9870 *) Support BS2000/OSD-POSIX.
9871 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9873 *) Make callbacks for key generation use void * instead of char *.
9876 *) Make S/MIME samples compile (not yet tested).
9879 *) Additional typesafe stacks.
9882 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9886 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9888 *) New configuration variant "sco5-gcc".
9890 *) Updated some demos.
9891 [Sean O Riordain, Wade Scholine]
9893 *) Add missing BIO_free at exit of pkcs12 application.
9896 *) Fix memory leak in conf.c.
9899 *) Updates for Win32 to assembler version of MD5.
9902 *) Set #! path to perl in apps/der_chop to where we found it
9903 instead of using a fixed path.
9906 *) SHA library changes for irix64-mips4-cc.
9909 *) Improvements for VMS support.
9913 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9915 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9916 This also avoids the problems with SC4.2 and unpatched SC5.
9917 [Andy Polyakov <appro@fy.chalmers.se>]
9919 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9920 These are required because of the typesafe stack would otherwise break
9921 existing code. If old code used a structure member which used to be STACK
9922 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9923 sk_num or sk_value it would produce an error because the num, data members
9924 are not present in STACK_OF. Now it just produces a warning. sk_set
9925 replaces the old method of assigning a value to sk_value
9926 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9927 that does this will no longer work (and should use sk_set instead) but
9928 this could be regarded as a "questionable" behaviour anyway.
9931 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9932 correctly handle encrypted S/MIME data.
9935 *) Change type of various DES function arguments from des_cblock
9936 (which means, in function argument declarations, pointer to char)
9937 to des_cblock * (meaning pointer to array with 8 char elements),
9938 which allows the compiler to do more typechecking; it was like
9939 that back in SSLeay, but with lots of ugly casts.
9941 Introduce new type const_des_cblock.
9944 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9945 problems: find RecipientInfo structure that matches recipient certificate
9946 and initialise the ASN1 structures properly based on passed cipher.
9949 *) Belatedly make the BN tests actually check the results.
9952 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9953 to and from BNs: it was completely broken. New compilation option
9954 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9955 key elements as negative integers.
9958 *) Reorganize and speed up MD5.
9959 [Andy Polyakov <appro@fy.chalmers.se>]
9962 [Richard Levitte <richard@levitte.org>]
9964 *) New option -out to asn1parse to allow the parsed structure to be
9965 output to a file. This is most useful when combined with the -strparse
9966 option to examine the output of things like OCTET STRINGS.
9969 *) Make SSL library a little more fool-proof by not requiring any longer
9970 that SSL_set_{accept,connect}_state be called before
9971 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9972 in many applications because usually everything *appeared* to work as
9973 intended anyway -- now it really works as intended).
9976 *) Move openssl.cnf out of lib/.
9979 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9980 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9981 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9982 [Ralf S. Engelschall]
9984 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9985 handle PKCS#7 enveloped data properly.
9986 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9988 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9989 copying pointers. The cert_st handling is changed by this in
9990 various ways (and thus what used to be known as ctx->default_cert
9991 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9992 any longer when s->cert does not give us what we need).
9993 ssl_cert_instantiate becomes obsolete by this change.
9994 As soon as we've got the new code right (possibly it already is?),
9995 we have solved a couple of bugs of the earlier code where s->cert
9996 was used as if it could not have been shared with other SSL structures.
9998 Note that using the SSL API in certain dirty ways now will result
9999 in different behaviour than observed with earlier library versions:
10000 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10001 does not influence s as it used to.
10003 In order to clean up things more thoroughly, inside SSL_SESSION
10004 we don't use CERT any longer, but a new structure SESS_CERT
10005 that holds per-session data (if available); currently, this is
10006 the peer's certificate chain and, for clients, the server's certificate
10007 and temporary key. CERT holds only those values that can have
10008 meaningful defaults in an SSL_CTX.
10011 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10012 from the internal representation. Various PKCS#7 fixes: remove some
10013 evil casts and set the enc_dig_alg field properly based on the signing
10017 *) Allow PKCS#12 password to be set from the command line or the
10018 environment. Let 'ca' get its config file name from the environment
10019 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10023 *) Allow certificate policies extension to use an IA5STRING for the
10024 organization field. This is contrary to the PKIX definition but
10025 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10029 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10030 without disallowing inline assembler and the like for non-pedantic builds.
10033 *) Support Borland C++ builder.
10034 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10036 *) Support Mingw32.
10039 *) SHA-1 cleanups and performance enhancements.
10040 [Andy Polyakov <appro@fy.chalmers.se>]
10042 *) Sparc v8plus assembler for the bignum library.
10043 [Andy Polyakov <appro@fy.chalmers.se>]
10045 *) Accept any -xxx and +xxx compiler options in Configure.
10048 *) Update HPUX configuration.
10051 *) Add missing sk_<type>_unshift() function to safestack.h
10052 [Ralf S. Engelschall]
10054 *) New function SSL_CTX_use_certificate_chain_file that sets the
10055 "extra_cert"s in addition to the certificate. (This makes sense
10056 only for "PEM" format files, as chains as a whole are not
10060 *) Support verify_depth from the SSL API.
10061 x509_vfy.c had what can be considered an off-by-one-error:
10062 Its depth (which was not part of the external interface)
10063 was actually counting the number of certificates in a chain;
10064 now it really counts the depth.
10067 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10068 instead of X509err, which often resulted in confusing error
10069 messages since the error codes are not globally unique
10070 (e.g. an alleged error in ssl3_accept when a certificate
10071 didn't match the private key).
10073 *) New function SSL_CTX_set_session_id_context that allows to set a default
10074 value (so that you don't need SSL_set_session_id_context for each
10075 connection using the SSL_CTX).
10078 *) OAEP decoding bug fix.
10081 *) Support INSTALL_PREFIX for package builders, as proposed by
10085 *) New Configure options "threads" and "no-threads". For systems
10086 where the proper compiler options are known (currently Solaris
10087 and Linux), "threads" is the default.
10090 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10093 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10094 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10095 such as /usr/local/bin.
10098 *) "make linux-shared" to build shared libraries.
10099 [Niels Poppe <niels@netbox.org>]
10101 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10104 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10105 extension adding in x509 utility.
10108 *) Remove NOPROTO sections and error code comments.
10111 *) Partial rewrite of the DEF file generator to now parse the ANSI
10115 *) New Configure options --prefix=DIR and --openssldir=DIR.
10118 *) Complete rewrite of the error code script(s). It is all now handled
10119 by one script at the top level which handles error code gathering,
10120 header rewriting and C source file generation. It should be much better
10121 than the old method: it now uses a modified version of Ulf's parser to
10122 read the ANSI prototypes in all header files (thus the old K&R definitions
10123 aren't needed for error creation any more) and do a better job of
10124 translating function codes into names. The old 'ASN1 error code imbedded
10125 in a comment' is no longer necessary and it doesn't use .err files which
10126 have now been deleted. Also the error code call doesn't have to appear all
10127 on one line (which resulted in some large lines...).
10130 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10133 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10134 0 (which usually indicates a closed connection), but continue reading.
10137 *) Fix some race conditions.
10140 *) Add support for CRL distribution points extension. Add Certificate
10141 Policies and CRL distribution points documentation.
10144 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10147 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10148 8 of keying material. Merlin has also confirmed interop with this fix
10149 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10150 [Merlin Hughes <merlin@baltimore.ie>]
10152 *) Fix lots of warnings.
10153 [Richard Levitte <levitte@stacken.kth.se>]
10155 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10156 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10157 [Richard Levitte <levitte@stacken.kth.se>]
10159 *) Fix problems with sizeof(long) == 8.
10160 [Andy Polyakov <appro@fy.chalmers.se>]
10162 *) Change functions to ANSI C.
10165 *) Fix typos in error codes.
10166 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10168 *) Remove defunct assembler files from Configure.
10171 *) SPARC v8 assembler BIGNUM implementation.
10172 [Andy Polyakov <appro@fy.chalmers.se>]
10174 *) Support for Certificate Policies extension: both print and set.
10175 Various additions to support the r2i method this uses.
10178 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10179 return a const string when you are expecting an allocated buffer.
10182 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10183 types DirectoryString and DisplayText.
10186 *) Add code to allow r2i extensions to access the configuration database,
10187 add an LHASH database driver and add several ctx helper functions.
10190 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10191 fail when they extended the size of a BIGNUM.
10194 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10195 support typesafe stack.
10198 *) Fix typo in SSL_[gs]et_options().
10199 [Nils Frostberg <nils@medcom.se>]
10201 *) Delete various functions and files that belonged to the (now obsolete)
10202 old X509V3 handling code.
10205 *) New Configure option "rsaref".
10208 *) Don't auto-generate pem.h.
10211 *) Introduce type-safe ASN.1 SETs.
10214 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10215 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10217 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10218 that links with OpenSSL (well at least cause lots of warnings), but fear
10219 not: the conversion is trivial, and it eliminates loads of evil casts. A
10220 few STACKed things have been converted already. Feel free to convert more.
10221 In the fullness of time, I'll do away with the STACK type altogether.
10224 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10225 specified in <certfile> by updating the entry in the index.txt file.
10226 This way one no longer has to edit the index.txt file manually for
10227 revoking a certificate. The -revoke option does the gory details now.
10228 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10230 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10231 `-text' option at all and this way the `-noout -text' combination was
10232 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10233 [Ralf S. Engelschall]
10235 *) Make sure a corresponding plain text error message exists for the
10236 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10237 verify callback function determined that a certificate was revoked.
10238 [Ralf S. Engelschall]
10240 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10241 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10242 all available cipers including rc5, which was forgotten until now.
10243 In order to let the testing shell script know which algorithms
10244 are available, a new (up to now undocumented) command
10245 "openssl list-cipher-commands" is used.
10248 *) Bugfix: s_client occasionally would sleep in select() when
10249 it should have checked SSL_pending() first.
10252 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10253 the raw DSA values prior to ASN.1 encoding.
10256 *) Tweaks to Configure
10257 [Niels Poppe <niels@netbox.org>]
10259 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10263 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10266 *) New config option to avoid instructions that are illegal on the 80386.
10267 The default code is faster, but requires at least a 486.
10270 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10271 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10272 same as SSL2_VERSION anyway.
10275 *) New "-showcerts" option for s_client.
10278 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10279 application. Various cleanups and fixes.
10282 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10283 modify error routines to work internally. Add error codes and PBE init
10284 to library startup routines.
10287 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10288 packing functions to asn1 and evp. Changed function names and error
10289 codes along the way.
10292 *) PKCS12 integration: and so it begins... First of several patches to
10293 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10294 objects to objects.h
10297 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10298 and display support for Thawte strong extranet extension.
10301 *) Add LinuxPPC support.
10302 [Jeff Dubrule <igor@pobox.org>]
10304 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10305 bn_div_words in alpha.s.
10306 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10308 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10309 OAEP isn't supported when OpenSSL is built with RSAref.
10310 [Ulf Moeller <ulf@fitug.de>]
10312 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10313 so they no longer are missing under -DNOPROTO.
10314 [Soren S. Jorvang <soren@t.dk>]
10317 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10319 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10320 doesn't work when the session is reused. Coming soon!
10323 *) Fix a security hole, that allows sessions to be reused in the wrong
10324 context thus bypassing client cert protection! All software that uses
10325 client certs and session caches in multiple contexts NEEDS PATCHING to
10326 allow session reuse! A fuller solution is in the works.
10327 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10329 *) Some more source tree cleanups (removed obsolete files
10330 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10331 permission on "config" script to be executable) and a fix for the INSTALL
10333 [Ulf Moeller <ulf@fitug.de>]
10335 *) Remove some legacy and erroneous uses of malloc, free instead of
10337 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10339 *) Make rsa_oaep_test return non-zero on error.
10340 [Ulf Moeller <ulf@fitug.de>]
10342 *) Add support for native Solaris shared libraries. Configure
10343 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10344 if someone would make that last step automatic.
10345 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10347 *) ctx_size was not built with the right compiler during "make links". Fixed.
10350 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10351 except NULL ciphers". This means the default cipher list will no longer
10352 enable NULL ciphers. They need to be specifically enabled e.g. with
10353 the string "DEFAULT:eNULL".
10356 *) Fix to RSA private encryption routines: if p < q then it would
10357 occasionally produce an invalid result. This will only happen with
10358 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10361 *) Be less restrictive and allow also `perl util/perlpath.pl
10362 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10363 because this way one can also use an interpreter named `perl5' (which is
10364 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10365 installed as `perl').
10366 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10368 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10369 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10371 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10372 advapi32.lib to Win32 build and change the pem test comparison
10373 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10374 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10375 and crypto/des/ede_cbcm_enc.c.
10378 *) DES quad checksum was broken on big-endian architectures. Fixed.
10381 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10382 Win32 test batch file so it (might) work again. The Win32 test batch file
10383 is horrible: I feel ill....
10386 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10387 in e_os.h. Audit of header files to check ANSI and non ANSI
10388 sections: 10 functions were absent from non ANSI section and not exported
10389 from Windows DLLs. Fixed up libeay.num for new functions.
10392 *) Make `openssl version' output lines consistent.
10393 [Ralf S. Engelschall]
10395 *) Fix Win32 symbol export lists for BIO functions: Added
10396 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10397 to ms/libeay{16,32}.def.
10398 [Ralf S. Engelschall]
10400 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10401 fine under Unix and passes some trivial tests I've now added. But the
10402 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10403 added to make sure no one expects that this stuff really works in the
10404 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10405 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10407 [Ralf S. Engelschall]
10409 *) Fix the generation of two part addresses in perl.
10410 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10412 *) Add config entry for Linux on MIPS.
10413 [John Tobey <jtobey@channel1.com>]
10415 *) Make links whenever Configure is run, unless we are on Windoze.
10418 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10419 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10423 *) Add a useful kludge to allow package maintainers to specify compiler and
10424 other platforms details on the command line without having to patch the
10425 Configure script everytime: One now can use ``perl Configure
10426 <id>:<details>'', i.e. platform ids are allowed to have details appended
10427 to them (separated by colons). This is treated as there would be a static
10428 pre-configured entry in Configure's %table under key <id> with value
10429 <details> and ``perl Configure <id>'' is called. So, when you want to
10430 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10431 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10432 now, which overrides the FreeBSD-elf entry on-the-fly.
10433 [Ralf S. Engelschall]
10435 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10438 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10439 on the `perl Configure ...' command line. This way one can compile
10440 OpenSSL libraries with Position Independent Code (PIC) which is needed
10441 for linking it into DSOs.
10442 [Ralf S. Engelschall]
10444 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10448 *) Cleaned up the LICENSE document: The official contact for any license
10449 questions now is the OpenSSL core team under openssl-core@openssl.org.
10450 And add a paragraph about the dual-license situation to make sure people
10451 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10452 to the OpenSSL toolkit.
10453 [Ralf S. Engelschall]
10455 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10456 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10457 Additionally cleaned up the `make links' target: Remove unnecessary
10458 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10459 to speed processing and no longer clutter the display with confusing
10460 stuff. Instead only the actually done links are displayed.
10461 [Ralf S. Engelschall]
10463 *) Permit null encryption ciphersuites, used for authentication only. It used
10464 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10465 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10469 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10470 signed attributes when verifying signatures (this would break them),
10471 the detached data encoding was wrong and public keys obtained using
10472 X509_get_pubkey() weren't freed.
10475 *) Add text documentation for the BUFFER functions. Also added a work around
10476 to a Win95 console bug. This was triggered by the password read stuff: the
10477 last character typed gets carried over to the next fread(). If you were
10478 generating a new cert request using 'req' for example then the last
10479 character of the passphrase would be CR which would then enter the first
10483 *) Added the new `Includes OpenSSL Cryptography Software' button as
10484 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10485 button and can be used by applications based on OpenSSL to show the
10486 relationship to the OpenSSL project.
10487 [Ralf S. Engelschall]
10489 *) Remove confusing variables in function signatures in files
10490 ssl/ssl_lib.c and ssl/ssl.h.
10491 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10493 *) Don't install bss_file.c under PREFIX/include/
10494 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10496 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10497 functions that return function pointers and has support for NT specific
10498 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10499 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10500 unsigned to signed types: this was killing the Win32 compile.
10503 *) Add new certificate file to stack functions,
10504 SSL_add_dir_cert_subjects_to_stack() and
10505 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10506 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10507 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10508 This means that Apache-SSL and similar packages don't have to mess around
10509 to add as many CAs as they want to the preferred list.
10512 *) Experiment with doxygen documentation. Currently only partially applied to
10514 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10515 openssl.doxy as the configuration file.
10518 *) Get rid of remaining C++-style comments which strict C compilers hate.
10519 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10521 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10522 compiled in by default: it has problems with large keys.
10525 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10526 DH private keys and/or callback functions which directly correspond to
10527 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10528 is needed for applications which have to configure certificates on a
10529 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10531 For the RSA certificate situation is makes no difference, but
10532 for the DSA certificate situation this fixes the "no shared cipher"
10533 problem where the OpenSSL cipher selection procedure failed because the
10534 temporary keys were not overtaken from the context and the API provided
10535 no way to reconfigure them.
10536 The new functions now let applications reconfigure the stuff and they
10537 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10538 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10539 non-public-API function ssl_cert_instantiate() is used as a helper
10540 function and also to reduce code redundancy inside ssl_rsa.c.
10541 [Ralf S. Engelschall]
10543 *) Move s_server -dcert and -dkey options out of the undocumented feature
10544 area because they are useful for the DSA situation and should be
10545 recognized by the users.
10546 [Ralf S. Engelschall]
10548 *) Fix the cipher decision scheme for export ciphers: the export bits are
10549 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10550 SSL_EXP_MASK. So, the original variable has to be used instead of the
10551 already masked variable.
10552 [Richard Levitte <levitte@stacken.kth.se>]
10554 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10555 [Richard Levitte <levitte@stacken.kth.se>]
10557 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10558 from `int' to `unsigned int' because it's a length and initialized by
10559 EVP_DigestFinal() which expects an `unsigned int *'.
10560 [Richard Levitte <levitte@stacken.kth.se>]
10562 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10563 script. Instead use the usual Shell->Perl transition trick.
10564 [Ralf S. Engelschall]
10566 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10567 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10568 -noout -modulus' as it's already the case for `openssl rsa -noout
10569 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10570 currently the public key is printed (a decision which was already done by
10571 `openssl dsa -modulus' in the past) which serves a similar purpose.
10572 Additionally the NO_RSA no longer completely removes the whole -modulus
10573 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10575 [Ralf S. Engelschall]
10577 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10578 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10579 [Arne Ansper <arne@ats.cyber.ee>]
10581 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10582 to be added. Now both 'req' and 'ca' can use new objects defined in the
10586 *) Add cool BIO that does syslog (or event log on NT).
10587 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10589 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10590 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10591 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10592 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10595 *) Add preliminary config info for new extension code.
10598 *) Make RSA_NO_PADDING really use no padding.
10599 [Ulf Moeller <ulf@fitug.de>]
10601 *) Generate errors when private/public key check is done.
10604 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10605 for some CRL extensions and new objects added.
10608 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10609 key usage extension and fuller support for authority key id.
10612 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10613 padding method for RSA, which is recommended for new applications in PKCS
10614 #1 v2.0 (RFC 2437, October 1998).
10615 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10616 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10617 against Bleichbacher's attack on RSA.
10618 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10621 *) Updates to the new SSL compression code
10622 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10624 *) Fix so that the version number in the master secret, when passed
10625 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10626 (because the server will not accept higher), that the version number
10627 is 0x03,0x01, not 0x03,0x00
10628 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10630 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10631 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10632 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10635 *) Support for RAW extensions where an arbitrary extension can be
10636 created by including its DER encoding. See apps/openssl.cnf for
10640 *) Make sure latest Perl versions don't interpret some generated C array
10641 code as Perl array code in the crypto/err/err_genc.pl script.
10642 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10644 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10645 not many people have the assembler. Various Win32 compilation fixes and
10646 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10647 build instructions.
10650 *) Modify configure script 'Configure' to automatically create crypto/date.h
10651 file under Win32 and also build pem.h from pem.org. New script
10652 util/mkfiles.pl to create the MINFO file on environments that can't do a
10653 'make files': perl util/mkfiles.pl >MINFO should work.
10656 *) Major rework of DES function declarations, in the pursuit of correctness
10657 and purity. As a result, many evil casts evaporated, and some weirdness,
10658 too. You may find this causes warnings in your code. Zapping your evil
10659 casts will probably fix them. Mostly.
10662 *) Fix for a typo in asn1.h. Bug fix to object creation script
10663 obj_dat.pl. It considered a zero in an object definition to mean
10664 "end of object": none of the objects in objects.h have any zeros
10665 so it wasn't spotted.
10666 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10668 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10669 Masking (CBCM). In the absence of test vectors, the best I have been able
10670 to do is check that the decrypt undoes the encrypt, so far. Send me test
10671 vectors if you have them.
10674 *) Correct calculation of key length for export ciphers (too much space was
10675 allocated for null ciphers). This has not been tested!
10678 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10679 message is now correct (it understands "crypto" and "ssl" on its
10680 command line). There is also now an "update" option. This will update
10681 the util/ssleay.num and util/libeay.num files with any new functions.
10683 perl util/mkdef.pl crypto ssl update
10684 it will update them.
10687 *) Overhauled the Perl interface (perl/*):
10688 - ported BN stuff to OpenSSL's different BN library
10689 - made the perl/ source tree CVS-aware
10690 - renamed the package from SSLeay to OpenSSL (the files still contain
10691 their history because I've copied them in the repository)
10692 - removed obsolete files (the test scripts will be replaced
10693 by better Test::Harness variants in the future)
10694 [Ralf S. Engelschall]
10696 *) First cut for a very conservative source tree cleanup:
10697 1. merge various obsolete readme texts into doc/ssleay.txt
10698 where we collect the old documents and readme texts.
10699 2. remove the first part of files where I'm already sure that we no
10700 longer need them because of three reasons: either they are just temporary
10701 files which were left by Eric or they are preserved original files where
10702 I've verified that the diff is also available in the CVS via "cvs diff
10703 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10704 the crypto/md/ stuff).
10705 [Ralf S. Engelschall]
10707 *) More extension code. Incomplete support for subject and issuer alt
10708 name, issuer and authority key id. Change the i2v function parameters
10709 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10710 what that's for :-) Fix to ASN1 macro which messed up
10711 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10714 *) Preliminary support for ENUMERATED type. This is largely copied from the
10718 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10719 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10721 *) Make sure `make rehash' target really finds the `openssl' program.
10722 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10724 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10725 like to hear about it if this slows down other processors.
10728 *) Add CygWin32 platform information to Configure script.
10729 [Alan Batie <batie@aahz.jf.intel.com>]
10731 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10732 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10734 *) New program nseq to manipulate netscape certificate sequences
10737 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10741 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10742 but the BN code had some problems that would cause failures when
10743 doing certificate verification and some other functions.
10744 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10746 *) Add ASN1 and PEM code to support netscape certificate sequences.
10749 *) Add ASN1 and PEM code to support netscape certificate sequences.
10752 *) Add several PKIX and private extended key usage OIDs.
10755 *) Modify the 'ca' program to handle the new extension code. Modify
10756 openssl.cnf for new extension format, add comments.
10759 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10760 and add a sample to openssl.cnf so req -x509 now adds appropriate
10764 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10765 error code, add initial support to X509_print() and x509 application.
10768 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10769 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10770 stuff is currently isolated and isn't even compiled yet.
10773 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10774 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10775 Removed the versions check from X509 routines when loading extensions:
10776 this allows certain broken certificates that don't set the version
10777 properly to be processed.
10780 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10781 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10782 can still be regenerated with "make depend".
10785 *) Spelling mistake in C version of CAST-128.
10786 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10788 *) Changes to the error generation code. The perl script err-code.pl
10789 now reads in the old error codes and retains the old numbers, only
10790 adding new ones if necessary. It also only changes the .err files if new
10791 codes are added. The makefiles have been modified to only insert errors
10792 when needed (to avoid needlessly modifying header files). This is done
10793 by only inserting errors if the .err file is newer than the auto generated
10794 C file. To rebuild all the error codes from scratch (the old behaviour)
10795 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10796 or delete all the .err files.
10799 *) CAST-128 was incorrectly implemented for short keys. The C version has
10800 been fixed, but is untested. The assembler versions are also fixed, but
10801 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10802 to regenerate it if needed.
10803 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10804 Hagino <itojun@kame.net>]
10806 *) File was opened incorrectly in randfile.c.
10807 [Ulf Möller <ulf@fitug.de>]
10809 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10810 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10811 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10812 al: it's just almost always a UTCTime. Note this patch adds new error
10813 codes so do a "make errors" if there are problems.
10816 *) Correct Linux 1 recognition in config.
10817 [Ulf Möller <ulf@fitug.de>]
10819 *) Remove pointless MD5 hash when using DSA keys in ca.
10820 [Anonymous <nobody@replay.com>]
10822 *) Generate an error if given an empty string as a cert directory. Also
10823 generate an error if handed NULL (previously returned 0 to indicate an
10824 error, but didn't set one).
10825 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10827 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10830 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10831 parameters. This was causing a warning which killed off the Win32 compile.
10834 *) Remove C++ style comments from crypto/bn/bn_local.h.
10835 [Neil Costigan <neil.costigan@celocom.com>]
10837 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10838 based on a text string, looking up short and long names and finally
10839 "dot" format. The "dot" format stuff didn't work. Added new function
10840 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10841 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10842 OID is not part of the table.
10845 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10846 X509_LOOKUP_by_alias().
10849 *) Sort openssl functions by name.
10852 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10853 encryption from sample DSA keys (in case anyone is interested the password
10857 *) Make _all_ *_free functions accept a NULL pointer.
10858 [Frans Heymans <fheymans@isaserver.be>]
10860 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10862 [Anonymous <nobody@replay.com>]
10864 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10865 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10867 *) Don't blow it for numeric -newkey arguments to apps/req.
10868 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10870 *) Temp key "for export" tests were wrong in s3_srvr.c.
10871 [Anonymous <nobody@replay.com>]
10873 *) Add prototype for temp key callback functions
10874 SSL_CTX_set_tmp_{rsa,dh}_callback().
10877 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10878 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10881 *) X509_name_add_entry() freed the wrong thing after an error.
10882 [Arne Ansper <arne@ats.cyber.ee>]
10884 *) rsa_eay.c would attempt to free a NULL context.
10885 [Arne Ansper <arne@ats.cyber.ee>]
10887 *) BIO_s_socket() had a broken should_retry() on Windoze.
10888 [Arne Ansper <arne@ats.cyber.ee>]
10890 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10891 [Arne Ansper <arne@ats.cyber.ee>]
10893 *) Make sure the already existing X509_STORE->depth variable is initialized
10894 in X509_STORE_new(), but document the fact that this variable is still
10895 unused in the certificate verification process.
10896 [Ralf S. Engelschall]
10898 *) Fix the various library and apps files to free up pkeys obtained from
10899 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10902 *) Fix reference counting in X509_PUBKEY_get(). This makes
10903 demos/maurice/example2.c work, amongst others, probably.
10904 [Steve Henson and Ben Laurie]
10906 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10907 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10908 are no longer created. This way we have a single and consistent command
10909 line interface `openssl <command>', similar to `cvs <command>'.
10910 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10912 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10913 BIT STRING wrapper always have zero unused bits.
10916 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10919 *) Make the top-level INSTALL documentation easier to understand.
10922 *) Makefiles updated to exit if an error occurs in a sub-directory
10923 make (including if user presses ^C) [Paul Sutton]
10925 *) Make Montgomery context stuff explicit in RSA data structure.
10928 *) Fix build order of pem and err to allow for generated pem.h.
10931 *) Fix renumbering bug in X509_NAME_delete_entry().
10934 *) Enhanced the err-ins.pl script so it makes the error library number
10935 global and can add a library name. This is needed for external ASN1 and
10936 other error libraries.
10939 *) Fixed sk_insert which never worked properly.
10942 *) Fix ASN1 macros so they can handle indefinite length constructed
10943 EXPLICIT tags. Some non standard certificates use these: they can now
10947 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10948 into a single doc/ssleay.txt bundle. This way the information is still
10949 preserved but no longer messes up this directory. Now it's new room for
10950 the new set of documentation files.
10951 [Ralf S. Engelschall]
10953 *) SETs were incorrectly DER encoded. This was a major pain, because they
10954 shared code with SEQUENCEs, which aren't coded the same. This means that
10955 almost everything to do with SETs or SEQUENCEs has either changed name or
10956 number of arguments.
10957 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10959 *) Fix test data to work with the above.
10962 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10963 was already fixed by Eric for 0.9.1 it seems.
10964 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10966 *) Autodetect FreeBSD3.
10969 *) Fix various bugs in Configure. This affects the following platforms:
10973 unixware-2.0-pentium
10977 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10978 before they are needed.
10981 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10985 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10987 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10988 changed SSLeay to OpenSSL in version strings.
10989 [Ralf S. Engelschall]
10991 *) Some fixups to the top-level documents.
10994 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10995 because the symlink to include/ was missing.
10996 [Ralf S. Engelschall]
10998 *) Incorporated the popular no-RSA/DSA-only patches
10999 which allow to compile a RSA-free SSLeay.
11000 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11002 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11003 when "ssleay" is still not found.
11004 [Ralf S. Engelschall]
11006 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11007 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11009 *) Updated the README file.
11010 [Ralf S. Engelschall]
11012 *) Added various .cvsignore files in the CVS repository subdirs
11013 to make a "cvs update" really silent.
11014 [Ralf S. Engelschall]
11016 *) Recompiled the error-definition header files and added
11017 missing symbols to the Win32 linker tables.
11018 [Ralf S. Engelschall]
11020 *) Cleaned up the top-level documents;
11021 o new files: CHANGES and LICENSE
11022 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11023 o merged COPYRIGHT into LICENSE
11024 o removed obsolete TODO file
11025 o renamed MICROSOFT to INSTALL.W32
11026 [Ralf S. Engelschall]
11028 *) Removed dummy files from the 0.9.1b source tree:
11029 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11030 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11031 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11032 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11033 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11034 [Ralf S. Engelschall]
11036 *) Added various platform portability fixes.
11039 *) The Genesis of the OpenSSL rpject:
11040 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11041 Young and Tim J. Hudson created while they were working for C2Net until
11043 [The OpenSSL Project]
11046 Changes between 0.9.0b and 0.9.1b [not released]
11048 *) Updated a few CA certificates under certs/
11051 *) Changed some BIGNUM api stuff.
11054 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11055 DGUX x86, Linux Alpha, etc.
11058 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11059 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11063 *) Add -strparse option to asn1pars program which parses nested
11065 [Dr Stephen Henson <shenson@bigfoot.com>]
11067 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11070 *) DSA fix for "ca" program.
11073 *) Added "-genkey" option to "dsaparam" program.
11076 *) Added RIPE MD160 (rmd160) message digest.
11079 *) Added -a (all) option to "ssleay version" command.
11082 *) Added PLATFORM define which is the id given to Configure.
11085 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11088 *) Extended the ASN.1 parser routines.
11091 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11094 *) Added a BN_CTX to the BN library.
11097 *) Fixed the weak key values in DES library
11100 *) Changed API in EVP library for cipher aliases.
11103 *) Added support for RC2/64bit cipher.
11106 *) Converted the lhash library to the crypto/mem.c functions.
11109 *) Added more recognized ASN.1 object ids.
11112 *) Added more RSA padding checks for SSL/TLS.
11115 *) Added BIO proxy/filter functionality.
11118 *) Added extra_certs to SSL_CTX which can be used
11119 send extra CA certificates to the client in the CA cert chain sending
11120 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11123 *) Now Fortezza is denied in the authentication phase because
11124 this is key exchange mechanism is not supported by SSLeay at all.
11127 *) Additional PKCS1 checks.
11130 *) Support the string "TLSv1" for all TLS v1 ciphers.
11133 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11134 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11137 *) Fixed a few memory leaks.
11140 *) Fixed various code and comment typos.
11143 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11144 bytes sent in the client random.
11145 [Edward Bishop <ebishop@spyglass.com>]