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
40 16-bit platforms such as WIN16
43 *) Start cleaning up OPENSSL_NO_xxx #define's
44 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
45 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
46 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
47 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
48 Remove MS_STATIC; it's a relic from platforms <32 bits.
51 *) Experimental support for a new, fast, unbiased prime candidate generator,
52 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
53 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
55 *) New output format NSS in the sess_id command line tool. This allows
56 exporting the session id and the master key in NSS keylog format.
57 [Martin Kaiser <martin@kaiser.cx>]
59 *) Harmonize version and its documentation. -f flag is used to display
61 [mancha <mancha1@zoho.com>]
63 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
65 [mancha <mancha1@zoho.com>]
67 *) Fix some double frees. These are not thought to be exploitable.
68 [mancha <mancha1@zoho.com>]
70 *) A missing bounds check in the handling of the TLS heartbeat extension
71 can be used to reveal up to 64k of memory to a connected client or
74 Thanks for Neel Mehta of Google Security for discovering this bug and to
75 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
76 preparing the fix (CVE-2014-0160)
77 [Adam Langley, Bodo Moeller]
79 *) Fix for the attack described in the paper "Recovering OpenSSL
80 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
81 by Yuval Yarom and Naomi Benger. Details can be obtained from:
82 http://eprint.iacr.org/2014/140
84 Thanks to Yuval Yarom and Naomi Benger for discovering this
85 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
86 [Yuval Yarom and Naomi Benger]
88 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
89 this fixes a limitation in previous versions of OpenSSL.
92 *) Experimental encrypt-then-mac support.
94 Experimental support for encrypt then mac from
95 draft-gutmann-tls-encrypt-then-mac-02.txt
97 To enable it set the appropriate extension number (0x42 for the test
98 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
100 For non-compliant peers (i.e. just about everything) this should have no
103 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
107 *) Add EVP support for key wrapping algorithms, to avoid problems with
108 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
109 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
110 algorithms and include tests cases.
113 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
117 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
118 MGF1 digest and OAEP label.
121 *) Make openssl verify return errors.
122 [Chris Palmer <palmer@google.com> and Ben Laurie]
124 *) New function ASN1_TIME_diff to calculate the difference between two
125 ASN1_TIME structures or one structure and the current time.
128 *) Update fips_test_suite to support multiple command line options. New
129 test to induce all self test errors in sequence and check expected
133 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
134 sign or verify all in one operation.
137 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
138 test programs and fips_test_suite. Includes functionality to parse
139 the minimal script output of fipsalgest.pl directly.
142 *) Add authorisation parameter to FIPS_module_mode_set().
145 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
148 *) Use separate DRBG fields for internal and external flags. New function
149 FIPS_drbg_health_check() to perform on demand health checking. Add
150 generation tests to fips_test_suite with reduced health check interval to
151 demonstrate periodic health checking. Add "nodh" option to
152 fips_test_suite to skip very slow DH test.
155 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
159 *) More extensive health check for DRBG checking many more failure modes.
160 New function FIPS_selftest_drbg_all() to handle every possible DRBG
161 combination: call this in fips_test_suite.
164 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
165 and POST to handle Dual EC cases.
168 *) Add support for canonical generation of DSA parameter 'g'. See
171 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
172 POST to handle HMAC cases.
175 *) Add functions FIPS_module_version() and FIPS_module_version_text()
176 to return numerical and string versions of the FIPS module number.
179 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
180 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
181 outside the validated module in the FIPS capable OpenSSL.
184 *) Minor change to DRBG entropy callback semantics. In some cases
185 there is no multiple of the block length between min_len and
186 max_len. Allow the callback to return more than max_len bytes
187 of entropy but discard any extra: it is the callback's responsibility
188 to ensure that the extra data discarded does not impact the
189 requested amount of entropy.
192 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
193 information in FIPS186-3, SP800-57 and SP800-131A.
196 *) CCM support via EVP. Interface is very similar to GCM case except we
197 must supply all data in one chunk (i.e. no update, final) and the
198 message length must be supplied if AAD is used. Add algorithm test
202 *) Initial version of POST overhaul. Add POST callback to allow the status
203 of POST to be monitored and/or failures induced. Modify fips_test_suite
204 to use callback. Always run all selftests even if one fails.
207 *) XTS support including algorithm test driver in the fips_gcmtest program.
208 Note: this does increase the maximum key length from 32 to 64 bytes but
209 there should be no binary compatibility issues as existing applications
210 will never use XTS mode.
213 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
214 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
215 performs algorithm blocking for unapproved PRNG types. Also do not
216 set PRNG type in FIPS_mode_set(): leave this to the application.
217 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
218 the standard OpenSSL PRNG: set additional data to a date time vector.
221 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
222 This shouldn't present any incompatibility problems because applications
223 shouldn't be using these directly and any that are will need to rethink
224 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
227 *) Extensive self tests and health checking required by SP800-90 DRBG.
228 Remove strength parameter from FIPS_drbg_instantiate and always
229 instantiate at maximum supported strength.
232 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
235 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
238 *) New function DH_compute_key_padded() to compute a DH key and pad with
239 leading zeroes if needed: this complies with SP800-56A et al.
242 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
243 anything, incomplete, subject to change and largely untested at present.
246 *) Modify fipscanisteronly build option to only build the necessary object
247 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
250 *) Add experimental option FIPSSYMS to give all symbols in
251 fipscanister.o and FIPS or fips prefix. This will avoid
252 conflicts with future versions of OpenSSL. Add perl script
253 util/fipsas.pl to preprocess assembly language source files
254 and rename any affected symbols.
257 *) Add selftest checks and algorithm block of non-fips algorithms in
258 FIPS mode. Remove DES2 from selftests.
261 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
262 return internal method without any ENGINE dependencies. Add new
263 tiny fips sign and verify functions.
266 *) New build option no-ec2m to disable characteristic 2 code.
269 *) New build option "fipscanisteronly". This only builds fipscanister.o
270 and (currently) associated fips utilities. Uses the file Makefile.fips
271 instead of Makefile.org as the prototype.
274 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
275 Update fips_gcmtest to use IV generator.
278 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
279 setting output buffer to NULL. The *Final function must be
280 called although it will not retrieve any additional data. The tag
281 can be set or retrieved with a ctrl. The IV length is by default 12
282 bytes (96 bits) but can be set to an alternative value. If the IV
283 length exceeds the maximum IV length (currently 16 bytes) it cannot be
287 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
288 underlying do_cipher function handles all cipher semantics itself
289 including padding and finalisation. This is useful if (for example)
290 an ENGINE cipher handles block padding itself. The behaviour of
291 do_cipher is subtly changed if this flag is set: the return value
292 is the number of characters written to the output buffer (zero is
293 no longer an error code) or a negative error code. Also if the
294 input buffer is NULL and length 0 finalisation should be performed.
297 *) If a candidate issuer certificate is already part of the constructed
298 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
301 *) Improve forward-security support: add functions
303 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
304 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
306 for use by SSL/TLS servers; the callback function will be called whenever a
307 new session is created, and gets to decide whether the session may be
308 cached to make it resumable (return 0) or not (return 1). (As by the
309 SSL/TLS protocol specifications, the session_id sent by the server will be
310 empty to indicate that the session is not resumable; also, the server will
311 not generate RFC 4507 (RFC 5077) session tickets.)
313 A simple reasonable callback implementation is to return is_forward_secure.
314 This parameter will be set to 1 or 0 depending on the ciphersuite selected
315 by the SSL/TLS server library, indicating whether it can provide forward
317 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
319 *) New -verify_name option in command line utilities to set verification
323 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
324 Add CMAC pkey methods.
327 *) Experimental renegotiation in s_server -www mode. If the client
328 browses /reneg connection is renegotiated. If /renegcert it is
329 renegotiated requesting a certificate.
332 *) Add an "external" session cache for debugging purposes to s_server. This
333 should help trace issues which normally are only apparent in deployed
334 multi-process servers.
337 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
338 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
339 BIO_set_cipher() and some obscure PEM functions were changed so they
340 can now return an error. The RAND changes required a change to the
341 RAND_METHOD structure.
344 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
345 a gcc attribute to warn if the result of a function is ignored. This
346 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
347 whose return value is often ignored.
350 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
352 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
353 ARMv5 through ARMv8, as opposite to "locking" it to single one.
354 So far those who have to target multiple plaforms would compromise
355 and argue that binary targeting say ARMv5 would still execute on
356 ARMv8. "Universal" build resolves this compromise by providing
357 near-optimal performance even on newer platforms.
360 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
361 (other platforms pending).
362 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
364 *) Add support for the SignedCertificateTimestampList certificate and
365 OCSP response extensions from RFC6962.
368 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
369 for corner cases. (Certain input points at infinity could lead to
370 bogus results, with non-infinity inputs mapped to infinity too.)
373 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
374 This covers AES, SHA256/512 and GHASH. "Initial" means that most
375 common cases are optimized and there still is room for further
376 improvements. Vector Permutation AES for Altivec is also added.
379 *) Add support for little-endian ppc64 Linux target.
380 [Marcelo Cerri (IBM)]
382 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
383 SHA1, SHA256 and GHASH. "Initial" means that most common cases
384 are optimized and there still is room for further improvements.
385 Both 32- and 64-bit modes are supported.
386 [Andy Polyakov, Ard Biesheuvel (Linaro)]
388 *) Improved ARMv7 NEON support.
391 *) Support for SPARC Architecture 2011 crypto extensions, first
392 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
393 SHA256/512, MD5, GHASH and modular exponentiation.
394 [Andy Polyakov, David Miller]
396 *) Accelerated modular exponentiation for Intel processors, a.k.a.
398 [Shay Gueron & Vlad Krasnov (Intel Corp)]
400 *) Support for new and upcoming Intel processors, including AVX2,
401 BMI and SHA ISA extensions. This includes additional "stitched"
402 implementations, AESNI-SHA256 and GCM, and multi-buffer support
405 This work was sponsored by Intel Corp.
408 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
409 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
410 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
413 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
414 this fixes a limiation in previous versions of OpenSSL.
417 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
418 MGF1 digest and OAEP label.
421 *) Add EVP support for key wrapping algorithms, to avoid problems with
422 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
423 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
424 algorithms and include tests cases.
427 *) Add functions to allocate and set the fields of an ECDSA_METHOD
429 [Douglas E. Engert, Steve Henson]
431 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
432 difference in days and seconds between two tm or ASN1_TIME structures.
435 *) Add -rev test option to s_server to just reverse order of characters
436 received by client and send back to server. Also prints an abbreviated
437 summary of the connection parameters.
440 *) New option -brief for s_client and s_server to print out a brief summary
441 of connection parameters.
444 *) Add callbacks for arbitrary TLS extensions.
445 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
447 *) New option -crl_download in several openssl utilities to download CRLs
448 from CRLDP extension in certificates.
451 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
454 *) New function X509_CRL_diff to generate a delta CRL from the difference
455 of two full CRLs. Add support to "crl" utility.
458 *) New functions to set lookup_crls function and to retrieve
459 X509_STORE from X509_STORE_CTX.
462 *) Print out deprecated issuer and subject unique ID fields in
466 *) Extend OCSP I/O functions so they can be used for simple general purpose
467 HTTP as well as OCSP. New wrapper function which can be used to download
468 CRLs using the OCSP API.
471 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
474 *) SSL_CONF* functions. These provide a common framework for application
475 configuration using configuration files or command lines.
478 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
479 message callback and prints the results. Needs compile time option
480 "enable-ssl-trace". New options to s_client and s_server to enable
484 *) New ctrl and macro to retrieve supported points extensions.
485 Print out extension in s_server and s_client.
488 *) New functions to retrieve certificate signature and signature
492 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
496 *) New Suite B modes for TLS code. These use and enforce the requirements
497 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
498 only use Suite B curves. The Suite B modes can be set by using the
499 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
502 *) New chain verification flags for Suite B levels of security. Check
503 algorithms are acceptable when flags are set in X509_verify_cert.
506 *) Make tls1_check_chain return a set of flags indicating checks passed
507 by a certificate chain. Add additional tests to handle client
508 certificates: checks for matching certificate type and issuer name
512 *) If an attempt is made to use a signature algorithm not in the peer
513 preference list abort the handshake. If client has no suitable
514 signature algorithms in response to a certificate request do not
518 *) If server EC tmp key is not in client preference list abort handshake.
521 *) Add support for certificate stores in CERT structure. This makes it
522 possible to have different stores per SSL structure or one store in
523 the parent SSL_CTX. Include distint stores for certificate chain
524 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
525 to build and store a certificate chain in CERT structure: returing
526 an error if the chain cannot be built: this will allow applications
527 to test if a chain is correctly configured.
529 Note: if the CERT based stores are not set then the parent SSL_CTX
530 store is used to retain compatibility with existing behaviour.
534 *) New function ssl_set_client_disabled to set a ciphersuite disabled
535 mask based on the current session, check mask when sending client
536 hello and checking the requested ciphersuite.
539 *) New ctrls to retrieve and set certificate types in a certificate
540 request message. Print out received values in s_client. If certificate
541 types is not set with custom values set sensible values based on
542 supported signature algorithms.
545 *) Support for distinct client and server supported signature algorithms.
548 *) Add certificate callback. If set this is called whenever a certificate
549 is required by client or server. An application can decide which
550 certificate chain to present based on arbitrary criteria: for example
551 supported signature algorithms. Add very simple example to s_server.
552 This fixes many of the problems and restrictions of the existing client
553 certificate callback: for example you can now clear an existing
554 certificate and specify the whole chain.
557 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
558 the certificate can be used for (if anything). Set valid_flags field
559 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
560 to have similar checks in it.
562 Add new "cert_flags" field to CERT structure and include a "strict mode".
563 This enforces some TLS certificate requirements (such as only permitting
564 certificate signature algorithms contained in the supported algorithms
565 extension) which some implementations ignore: this option should be used
566 with caution as it could cause interoperability issues.
569 *) Update and tidy signature algorithm extension processing. Work out
570 shared signature algorithms based on preferences and peer algorithms
571 and print them out in s_client and s_server. Abort handshake if no
572 shared signature algorithms.
575 *) Add new functions to allow customised supported signature algorithms
576 for SSL and SSL_CTX structures. Add options to s_client and s_server
580 *) New function SSL_certs_clear() to delete all references to certificates
581 from an SSL structure. Before this once a certificate had been added
582 it couldn't be removed.
585 *) Integrate hostname, email address and IP address checking with certificate
586 verification. New verify options supporting checking in opensl utility.
589 *) Fixes and wildcard matching support to hostname and email checking
590 functions. Add manual page.
591 [Florian Weimer (Red Hat Product Security Team)]
593 *) New functions to check a hostname email or IP address against a
594 certificate. Add options x509 utility to print results of checks against
598 *) Fix OCSP checking.
599 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
601 *) Initial experimental support for explicitly trusted non-root CAs.
602 OpenSSL still tries to build a complete chain to a root but if an
603 intermediate CA has a trust setting included that is used. The first
604 setting is used: whether to trust (e.g., -addtrust option to the x509
608 *) Add -trusted_first option which attempts to find certificates in the
609 trusted store even if an untrusted chain is also supplied.
612 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
613 platform support for Linux and Android.
616 *) Support for linux-x32, ILP32 environment in x86_64 framework.
619 *) Experimental multi-implementation support for FIPS capable OpenSSL.
620 When in FIPS mode the approved implementations are used as normal,
621 when not in FIPS mode the internal unapproved versions are used instead.
622 This means that the FIPS capable OpenSSL isn't forced to use the
623 (often lower performance) FIPS implementations outside FIPS mode.
626 *) Transparently support X9.42 DH parameters when calling
627 PEM_read_bio_DHparameters. This means existing applications can handle
628 the new parameter format automatically.
631 *) Initial experimental support for X9.42 DH parameter format: mainly
632 to support use of 'q' parameter for RFC5114 parameters.
635 *) Add DH parameters from RFC5114 including test data to dhtest.
638 *) Support for automatic EC temporary key parameter selection. If enabled
639 the most preferred EC parameters are automatically used instead of
640 hardcoded fixed parameters. Now a server just has to call:
641 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
642 support ECDH and use the most appropriate parameters.
645 *) Enhance and tidy EC curve and point format TLS extension code. Use
646 static structures instead of allocation if default values are used.
647 New ctrls to set curves we wish to support and to retrieve shared curves.
648 Print out shared curves in s_server. New options to s_server and s_client
649 to set list of supported curves.
652 *) New ctrls to retrieve supported signature algorithms and
653 supported curve values as an array of NIDs. Extend openssl utility
654 to print out received values.
657 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
658 between NIDs and the more common NIST names such as "P-256". Enhance
659 ecparam utility and ECC method to recognise the NIST names for curves.
662 *) Enhance SSL/TLS certificate chain handling to support different
663 chains for each certificate instead of one chain in the parent SSL_CTX.
666 *) Support for fixed DH ciphersuite client authentication: where both
667 server and client use DH certificates with common parameters.
670 *) Support for fixed DH ciphersuites: those requiring DH server
674 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
676 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
677 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
678 X509_CINF_get_signature were reverted post internal team review.
680 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
682 *) Abort handshake if server key exchange message is omitted for ephemeral
685 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
686 reporting this issue.
690 *) Remove non-export ephemeral RSA code on client and server. This code
691 violated the TLS standard by allowing the use of temporary RSA keys in
692 non-export ciphersuites and could be used by a server to effectively
693 downgrade the RSA key length used to a value smaller than the server
694 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
695 INRIA or reporting this issue.
699 *) Ensure that the session ID context of an SSL is updated when its
700 SSL_CTX is updated via SSL_set_SSL_CTX.
702 The session ID context is typically set from the parent SSL_CTX,
703 and can vary with the CTX.
706 *) Fix various certificate fingerprint issues.
708 By using non-DER or invalid encodings outside the signed portion of a
709 certificate the fingerprint can be changed without breaking the signature.
710 Although no details of the signed portion of the certificate can be changed
711 this can cause problems with some applications: e.g. those using the
712 certificate fingerprint for blacklists.
714 1. Reject signatures with non zero unused bits.
716 If the BIT STRING containing the signature has non zero unused bits reject
717 the signature. All current signature algorithms require zero unused bits.
719 2. Check certificate algorithm consistency.
721 Check the AlgorithmIdentifier inside TBS matches the one in the
722 certificate signature. NB: this will result in signature failure
723 errors for some broken certificates.
725 Thanks to Konrad Kraszewski from Google for reporting this issue.
727 3. Check DSA/ECDSA signatures use DER.
729 Reencode DSA/ECDSA signatures and compare with the original received
730 signature. Return an error if there is a mismatch.
732 This will reject various cases including garbage after signature
733 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
734 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
735 (negative or with leading zeroes).
737 Further analysis was conducted and fixes were developed by Stephen Henson
738 of the OpenSSL core team.
743 *) Do not resume sessions on the server if the negotiated protocol
744 version does not match the session's version. Resuming with a different
745 version, while not strictly forbidden by the RFC, is of questionable
746 sanity and breaks all known clients.
747 [David Benjamin, Emilia Käsper]
749 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
750 early CCS messages during renegotiation. (Note that because
751 renegotiation is encrypted, this early CCS was not exploitable.)
754 *) Tighten client-side session ticket handling during renegotiation:
755 ensure that the client only accepts a session ticket if the server sends
756 the extension anew in the ServerHello. Previously, a TLS client would
757 reuse the old extension state and thus accept a session ticket if one was
758 announced in the initial ServerHello.
760 Similarly, ensure that the client requires a session ticket if one
761 was advertised in the ServerHello. Previously, a TLS client would
762 ignore a missing NewSessionTicket message.
765 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
769 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
770 sends a carefully crafted handshake message, to cause OpenSSL to fail
771 to free up to 64k of memory causing a memory leak. This could be
772 exploited in a Denial Of Service attack. This issue affects OpenSSL
773 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
774 whether SRTP is used or configured. Implementations of OpenSSL that
775 have been compiled with OPENSSL_NO_SRTP defined are not affected.
777 The fix was developed by the OpenSSL team.
781 *) Session Ticket Memory Leak.
783 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
784 integrity of that ticket is first verified. In the event of a session
785 ticket integrity check failing, OpenSSL will fail to free memory
786 causing a memory leak. By sending a large number of invalid session
787 tickets an attacker could exploit this issue in a Denial Of Service
792 *) Build option no-ssl3 is incomplete.
794 When OpenSSL is configured with "no-ssl3" as a build option, servers
795 could accept and complete a SSL 3.0 handshake, and clients could be
796 configured to send them.
798 [Akamai and the OpenSSL team]
800 *) Add support for TLS_FALLBACK_SCSV.
801 Client applications doing fallback retries should call
802 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
804 [Adam Langley, Bodo Moeller]
806 *) Add additional DigestInfo checks.
808 Reencode DigestInto in DER and check against the original when
809 verifying RSA signature: this will reject any improperly encoded
810 DigestInfo structures.
812 Note: this is a precautionary measure and no attacks are currently known.
816 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
818 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
819 SRP code can be overrun an internal buffer. Add sanity check that
820 g, A, B < N to SRP code.
822 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
823 Group for discovering this issue.
827 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
828 TLS 1.0 instead of higher protocol versions when the ClientHello message
829 is badly fragmented. This allows a man-in-the-middle attacker to force a
830 downgrade to TLS 1.0 even if both the server and the client support a
831 higher protocol version, by modifying the client's TLS records.
833 Thanks to David Benjamin and Adam Langley (Google) for discovering and
834 researching this issue.
838 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
839 to a denial of service attack. A malicious server can crash the client
840 with a null pointer dereference (read) by specifying an anonymous (EC)DH
841 ciphersuite and sending carefully crafted handshake messages.
843 Thanks to Felix Gröbert (Google) for discovering and researching this
848 *) By sending carefully crafted DTLS packets an attacker could cause openssl
849 to leak memory. This can be exploited through a Denial of Service attack.
850 Thanks to Adam Langley for discovering and researching this issue.
854 *) An attacker can force openssl to consume large amounts of memory whilst
855 processing DTLS handshake messages. This can be exploited through a
856 Denial of Service attack.
857 Thanks to Adam Langley for discovering and researching this issue.
861 *) An attacker can force an error condition which causes openssl to crash
862 whilst processing DTLS packets due to memory being freed twice. This
863 can be exploited through a Denial of Service attack.
864 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
869 *) If a multithreaded client connects to a malicious server using a resumed
870 session and the server sends an ec point format extension it could write
871 up to 255 bytes to freed memory.
873 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
878 *) A malicious server can crash an OpenSSL client with a null pointer
879 dereference (read) by specifying an SRP ciphersuite even though it was not
880 properly negotiated with the client. This can be exploited through a
881 Denial of Service attack.
883 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
884 discovering and researching this issue.
888 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
889 X509_name_oneline, X509_name_print_ex et al. to leak some information
890 from the stack. Applications may be affected if they echo pretty printing
891 output to the attacker.
893 Thanks to Ivan Fratric (Google) for discovering this issue.
895 [Emilia Käsper, and Steve Henson]
897 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
898 for corner cases. (Certain input points at infinity could lead to
899 bogus results, with non-infinity inputs mapped to infinity too.)
902 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
904 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
905 handshake can force the use of weak keying material in OpenSSL
906 SSL/TLS clients and servers.
908 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
909 researching this issue. (CVE-2014-0224)
910 [KIKUCHI Masashi, Steve Henson]
912 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
913 OpenSSL DTLS client the code can be made to recurse eventually crashing
916 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
918 [Imre Rad, Steve Henson]
920 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
921 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
922 client or server. This is potentially exploitable to run arbitrary
923 code on a vulnerable client or server.
925 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
926 [Jüri Aedla, Steve Henson]
928 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
929 are subject to a denial of service attack.
931 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
932 this issue. (CVE-2014-3470)
933 [Felix Gröbert, Ivan Fratric, Steve Henson]
935 *) Harmonize version and its documentation. -f flag is used to display
937 [mancha <mancha1@zoho.com>]
939 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
941 [mancha <mancha1@zoho.com>]
943 *) Fix some double frees. These are not thought to be exploitable.
944 [mancha <mancha1@zoho.com>]
946 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
948 *) A missing bounds check in the handling of the TLS heartbeat extension
949 can be used to reveal up to 64k of memory to a connected client or
952 Thanks for Neel Mehta of Google Security for discovering this bug and to
953 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
954 preparing the fix (CVE-2014-0160)
955 [Adam Langley, Bodo Moeller]
957 *) Fix for the attack described in the paper "Recovering OpenSSL
958 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
959 by Yuval Yarom and Naomi Benger. Details can be obtained from:
960 http://eprint.iacr.org/2014/140
962 Thanks to Yuval Yarom and Naomi Benger for discovering this
963 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
964 [Yuval Yarom and Naomi Benger]
966 *) TLS pad extension: draft-agl-tls-padding-03
968 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
969 TLS client Hello record length value would otherwise be > 255 and
970 less that 512 pad with a dummy extension containing zeroes so it
971 is at least 512 bytes long.
973 [Adam Langley, Steve Henson]
975 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
977 *) Fix for TLS record tampering bug. A carefully crafted invalid
978 handshake could crash OpenSSL with a NULL pointer exception.
979 Thanks to Anton Johansson for reporting this issues.
982 *) Keep original DTLS digest and encryption contexts in retransmission
983 structures so we can use the previous session parameters if they need
984 to be resent. (CVE-2013-6450)
987 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
988 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
989 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
990 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
991 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
992 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
993 [Rob Stradling, Adam Langley]
995 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
997 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
998 supporting platforms or when small records were transferred.
999 [Andy Polyakov, Steve Henson]
1001 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1003 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1005 This addresses the flaw in CBC record processing discovered by
1006 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1007 at: http://www.isg.rhul.ac.uk/tls/
1009 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1010 Security Group at Royal Holloway, University of London
1011 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1012 Emilia Käsper for the initial patch.
1014 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1016 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1017 ciphersuites which can be exploited in a denial of service attack.
1018 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1019 and detecting this bug and to Wolfgang Ettlinger
1020 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1024 *) Return an error when checking OCSP signatures when key is NULL.
1025 This fixes a DoS attack. (CVE-2013-0166)
1028 *) Make openssl verify return errors.
1029 [Chris Palmer <palmer@google.com> and Ben Laurie]
1031 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1032 the right response is stapled. Also change SSL_get_certificate()
1033 so it returns the certificate actually sent.
1034 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1035 [Rob Stradling <rob.stradling@comodo.com>]
1037 *) Fix possible deadlock when decoding public keys.
1040 *) Don't use TLS 1.0 record version number in initial client hello
1044 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1046 *) Sanity check record length before skipping explicit IV in TLS
1047 1.2, 1.1 and DTLS to fix DoS attack.
1049 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1050 fuzzing as a service testing platform.
1054 *) Initialise tkeylen properly when encrypting CMS messages.
1055 Thanks to Solar Designer of Openwall for reporting this issue.
1058 *) In FIPS mode don't try to use composite ciphers as they are not
1062 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1064 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1065 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1066 mean any application compiled against OpenSSL 1.0.0 headers setting
1067 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1068 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1069 0x10000000L Any application which was previously compiled against
1070 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1071 will need to be recompiled as a result. Letting be results in
1072 inability to disable specifically TLS 1.1 and in client context,
1073 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1076 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1077 disable just protocol X, but all protocols above X *if* there are
1078 protocols *below* X still enabled. In more practical terms it means
1079 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1080 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1081 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1085 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1087 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1088 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1089 in CRYPTO_realloc_clean.
1091 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1092 issue and to Adam Langley <agl@chromium.org> for fixing it.
1094 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1096 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1099 *) Workarounds for some broken servers that "hang" if a client hello
1100 record length exceeds 255 bytes.
1102 1. Do not use record version number > TLS 1.0 in initial client
1103 hello: some (but not all) hanging servers will now work.
1104 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1105 the number of ciphers sent in the client hello. This should be
1106 set to an even number, such as 50, for example by passing:
1107 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1108 Most broken servers should now work.
1109 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1110 TLS 1.2 client support entirely.
1113 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1116 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1118 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1119 STRING form instead of a DigestInfo.
1122 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1123 and the RSA_sign/RSA_verify functions. This was made more apparent when
1124 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1125 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1126 the correct format in RSA_verify so both forms transparently work.
1129 *) Some servers which support TLS 1.0 can choke if we initially indicate
1130 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1131 encrypted premaster secret. As a workaround use the maximum permitted
1132 client version in client hello, this should keep such servers happy
1133 and still work with previous versions of OpenSSL.
1136 *) Add support for TLS/DTLS heartbeats.
1137 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1139 *) Add support for SCTP.
1140 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1142 *) Improved PRNG seeding for VOS.
1143 [Paul Green <Paul.Green@stratus.com>]
1145 *) Extensive assembler packs updates, most notably:
1147 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1148 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1149 - x86_64: bit-sliced AES implementation;
1150 - ARM: NEON support, contemporary platforms optimizations;
1151 - s390x: z196 support;
1152 - *: GHASH and GF(2^m) multiplication implementations;
1156 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1157 (removal of unnecessary code)
1158 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1160 *) Add TLS key material exporter from RFC 5705.
1163 *) Add DTLS-SRTP negotiation from RFC 5764.
1166 *) Add Next Protocol Negotiation,
1167 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1168 disabled with a no-npn flag to config or Configure. Code donated
1170 [Adam Langley <agl@google.com> and Ben Laurie]
1172 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1173 NIST-P256, NIST-P521, with constant-time single point multiplication on
1174 typical inputs. Compiler support for the nonstandard type __uint128_t is
1175 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1176 Code made available under Apache License version 2.0.
1178 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1179 line to include this in your build of OpenSSL, and run "make depend" (or
1180 "make update"). This enables the following EC_METHODs:
1182 EC_GFp_nistp224_method()
1183 EC_GFp_nistp256_method()
1184 EC_GFp_nistp521_method()
1186 EC_GROUP_new_by_curve_name() will automatically use these (while
1187 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1189 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1191 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1192 all platforms. Move ssize_t definition from e_os.h to the public
1193 header file e_os2.h as it now appears in public header file cms.h
1196 *) New -sigopt option to the ca, req and x509 utilities. Additional
1197 signature parameters can be passed using this option and in
1201 *) Add RSA PSS signing function. This will generate and set the
1202 appropriate AlgorithmIdentifiers for PSS based on those in the
1203 corresponding EVP_MD_CTX structure. No application support yet.
1206 *) Support for companion algorithm specific ASN1 signing routines.
1207 New function ASN1_item_sign_ctx() signs a pre-initialised
1208 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1209 the appropriate parameters.
1212 *) Add new algorithm specific ASN1 verification initialisation function
1213 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1214 handling will be the same no matter what EVP_PKEY_METHOD is used.
1215 Add a PSS handler to support verification of PSS signatures: checked
1216 against a number of sample certificates.
1219 *) Add signature printing for PSS. Add PSS OIDs.
1220 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1222 *) Add algorithm specific signature printing. An individual ASN1 method
1223 can now print out signatures instead of the standard hex dump.
1225 More complex signatures (e.g. PSS) can print out more meaningful
1226 information. Include DSA version that prints out the signature
1230 *) Password based recipient info support for CMS library: implementing
1234 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1235 neatly separates the code into cipher and PBE sections and is required
1236 for some algorithms that split PBES2 into separate pieces (such as
1237 password based CMS).
1240 *) Session-handling fixes:
1241 - Fix handling of connections that are resuming with a session ID,
1242 but also support Session Tickets.
1243 - Fix a bug that suppressed issuing of a new ticket if the client
1244 presented a ticket with an expired session.
1245 - Try to set the ticket lifetime hint to something reasonable.
1246 - Make tickets shorter by excluding irrelevant information.
1247 - On the client side, don't ignore renewed tickets.
1248 [Adam Langley, Bodo Moeller (Google)]
1250 *) Fix PSK session representation.
1253 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1255 This work was sponsored by Intel.
1258 *) Add GCM support to TLS library. Some custom code is needed to split
1259 the IV between the fixed (from PRF) and explicit (from TLS record)
1260 portions. This adds all GCM ciphersuites supported by RFC5288 and
1261 RFC5289. Generalise some AES* cipherstrings to include GCM and
1262 add a special AESGCM string for GCM only.
1265 *) Expand range of ctrls for AES GCM. Permit setting invocation
1266 field on decrypt and retrieval of invocation field only on encrypt.
1269 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1270 As required by RFC5289 these ciphersuites cannot be used if for
1271 versions of TLS earlier than 1.2.
1274 *) For FIPS capable OpenSSL interpret a NULL default public key method
1275 as unset and return the appropriate default but do *not* set the default.
1276 This means we can return the appropriate method in applications that
1277 switch between FIPS and non-FIPS modes.
1280 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1281 ENGINE is used then we cannot handle that in the FIPS module so we
1282 keep original code iff non-FIPS operations are allowed.
1285 *) Add -attime option to openssl utilities.
1286 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1288 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1291 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1292 FIPS EC methods unconditionally for now.
1295 *) New build option no-ec2m to disable characteristic 2 code.
1298 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1299 all cases can be covered as some introduce binary incompatibilities.
1302 *) Redirect RSA operations to FIPS module including keygen,
1303 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1306 *) Add similar low level API blocking to ciphers.
1309 *) Low level digest APIs are not approved in FIPS mode: any attempt
1310 to use these will cause a fatal error. Applications that *really* want
1311 to use them can use the private_* version instead.
1314 *) Redirect cipher operations to FIPS module for FIPS builds.
1317 *) Redirect digest operations to FIPS module for FIPS builds.
1320 *) Update build system to add "fips" flag which will link in fipscanister.o
1321 for static and shared library builds embedding a signature if needed.
1324 *) Output TLS supported curves in preference order instead of numerical
1325 order. This is currently hardcoded for the highest order curves first.
1326 This should be configurable so applications can judge speed vs strength.
1329 *) Add TLS v1.2 server support for client authentication.
1332 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1336 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1337 FIPS modules versions.
1340 *) Add TLS v1.2 client side support for client authentication. Keep cache
1341 of handshake records longer as we don't know the hash algorithm to use
1342 until after the certificate request message is received.
1345 *) Initial TLS v1.2 client support. Add a default signature algorithms
1346 extension including all the algorithms we support. Parse new signature
1347 format in client key exchange. Relax some ECC signing restrictions for
1348 TLS v1.2 as indicated in RFC5246.
1351 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1352 to new signature format when needed using client digest preference.
1353 All server ciphersuites should now work correctly in TLS v1.2. No client
1354 support yet and no support for client certificates.
1357 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1358 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1359 ciphersuites. At present only RSA key exchange ciphersuites work with
1360 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1361 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1362 and version checking.
1365 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1366 with this defined it will not be affected by any changes to ssl internal
1367 structures. Add several utility functions to allow openssl application
1368 to work with OPENSSL_NO_SSL_INTERN defined.
1372 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1374 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1377 *) Permit abbreviated handshakes when renegotiating using the function
1378 SSL_renegotiate_abbreviated().
1379 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1381 *) Add call to ENGINE_register_all_complete() to
1382 ENGINE_load_builtin_engines(), so some implementations get used
1383 automatically instead of needing explicit application support.
1386 *) Add support for TLS key exporter as described in RFC5705.
1387 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1389 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1390 a few changes are required:
1392 Add SSL_OP_NO_TLSv1_1 flag.
1393 Add TLSv1_1 methods.
1394 Update version checking logic to handle version 1.1.
1395 Add explicit IV handling (ported from DTLS code).
1396 Add command line options to s_client/s_server.
1399 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1401 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1403 This addresses the flaw in CBC record processing discovered by
1404 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1405 at: http://www.isg.rhul.ac.uk/tls/
1407 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1408 Security Group at Royal Holloway, University of London
1409 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1410 Emilia Käsper for the initial patch.
1412 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1414 *) Return an error when checking OCSP signatures when key is NULL.
1415 This fixes a DoS attack. (CVE-2013-0166)
1418 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1419 the right response is stapled. Also change SSL_get_certificate()
1420 so it returns the certificate actually sent.
1421 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1422 (This is a backport)
1423 [Rob Stradling <rob.stradling@comodo.com>]
1425 *) Fix possible deadlock when decoding public keys.
1428 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1430 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1433 *) Sanity check record length before skipping explicit IV in DTLS
1436 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1437 fuzzing as a service testing platform.
1441 *) Initialise tkeylen properly when encrypting CMS messages.
1442 Thanks to Solar Designer of Openwall for reporting this issue.
1445 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1447 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1448 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1449 in CRYPTO_realloc_clean.
1451 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1452 issue and to Adam Langley <agl@chromium.org> for fixing it.
1454 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1456 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1458 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1459 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1460 content decryption and always return the same error. Note: this attack
1461 needs on average 2^20 messages so it only affects automated senders. The
1462 old behaviour can be reenabled in the CMS code by setting the
1463 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1464 an MMA defence is not necessary.
1465 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1466 this issue. (CVE-2012-0884)
1469 *) Fix CVE-2011-4619: make sure we really are receiving a
1470 client hello before rejecting multiple SGC restarts. Thanks to
1471 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1474 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1476 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1477 Thanks to Antonio Martin, Enterprise Secure Access Research and
1478 Development, Cisco Systems, Inc. for discovering this bug and
1479 preparing a fix. (CVE-2012-0050)
1482 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1484 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1485 of the Vaudenay padding oracle attack on CBC mode encryption
1486 which enables an efficient plaintext recovery attack against
1487 the OpenSSL implementation of DTLS. Their attack exploits timing
1488 differences arising during decryption processing. A research
1489 paper describing this attack can be found at:
1490 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1491 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1492 Security Group at Royal Holloway, University of London
1493 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1494 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1495 for preparing the fix. (CVE-2011-4108)
1496 [Robin Seggelmann, Michael Tuexen]
1498 *) Clear bytes used for block padding of SSL 3.0 records.
1500 [Adam Langley (Google)]
1502 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1503 Kadianakis <desnacked@gmail.com> for discovering this issue and
1504 Adam Langley for preparing the fix. (CVE-2011-4619)
1505 [Adam Langley (Google)]
1507 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1508 [Andrey Kulikov <amdeich@gmail.com>]
1510 *) Prevent malformed RFC3779 data triggering an assertion failure.
1511 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1512 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1513 [Rob Austein <sra@hactrn.net>]
1515 *) Improved PRNG seeding for VOS.
1516 [Paul Green <Paul.Green@stratus.com>]
1518 *) Fix ssl_ciph.c set-up race.
1519 [Adam Langley (Google)]
1521 *) Fix spurious failures in ecdsatest.c.
1522 [Emilia Käsper (Google)]
1524 *) Fix the BIO_f_buffer() implementation (which was mixing different
1525 interpretations of the '..._len' fields).
1526 [Adam Langley (Google)]
1528 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1529 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1530 threads won't reuse the same blinding coefficients.
1532 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1533 lock to call BN_BLINDING_invert_ex, and avoids one use of
1534 BN_BLINDING_update for each BN_BLINDING structure (previously,
1535 the last update always remained unused).
1536 [Emilia Käsper (Google)]
1538 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1539 [Bob Buckholz (Google)]
1541 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1543 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1544 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1545 [Kaspar Brand <ossl@velox.ch>]
1547 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1548 for multi-threaded use of ECDH. (CVE-2011-3210)
1549 [Adam Langley (Google)]
1551 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1554 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1555 signature public key algorithm by using OID xref utilities instead.
1556 Before this you could only use some ECC ciphersuites with SHA1 only.
1559 *) Add protection against ECDSA timing attacks as mentioned in the paper
1560 by Billy Bob Brumley and Nicola Tuveri, see:
1562 http://eprint.iacr.org/2011/232.pdf
1564 [Billy Bob Brumley and Nicola Tuveri]
1566 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1568 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1569 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1571 *) Fix bug in string printing code: if *any* escaping is enabled we must
1572 escape the escape character (backslash) or the resulting string is
1576 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1578 *) Disable code workaround for ancient and obsolete Netscape browsers
1579 and servers: an attacker can use it in a ciphersuite downgrade attack.
1580 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1583 *) Fixed J-PAKE implementation error, originally discovered by
1584 Sebastien Martini, further info and confirmation from Stefan
1585 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1588 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1590 *) Fix extension code to avoid race conditions which can result in a buffer
1591 overrun vulnerability: resumed sessions must not be modified as they can
1592 be shared by multiple threads. CVE-2010-3864
1595 *) Fix WIN32 build system to correctly link an ENGINE directory into
1599 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1601 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1603 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1605 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1607 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1608 context. The operation can be customised via the ctrl mechanism in
1609 case ENGINEs want to include additional functionality.
1612 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1615 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1616 output hashes compatible with older versions of OpenSSL.
1617 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1619 *) Fix compression algorithm handling: if resuming a session use the
1620 compression algorithm of the resumed session instead of determining
1621 it from client hello again. Don't allow server to change algorithm.
1624 *) Add load_crls() function to apps tidying load_certs() too. Add option
1625 to verify utility to allow additional CRLs to be included.
1628 *) Update OCSP request code to permit adding custom headers to the request:
1629 some responders need this.
1632 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1634 [Julia Lawall <julia@diku.dk>]
1636 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1637 needlessly dereferenced structures, used obsolete functions and
1638 didn't handle all updated verify codes correctly.
1641 *) Disable MD2 in the default configuration.
1644 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1645 indicate the initial BIO being pushed or popped. This makes it possible
1646 to determine whether the BIO is the one explicitly called or as a result
1647 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1648 it handles reference counts correctly and doesn't zero out the I/O bio
1649 when it is not being explicitly popped. WARNING: applications which
1650 included workarounds for the old buggy behaviour will need to be modified
1651 or they could free up already freed BIOs.
1654 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1655 renaming to all platforms (within the 0.9.8 branch, this was
1656 done conditionally on Netware platforms to avoid a name clash).
1657 [Guenter <lists@gknw.net>]
1659 *) Add ECDHE and PSK support to DTLS.
1660 [Michael Tuexen <tuexen@fh-muenster.de>]
1662 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1666 *) Add "missing" function EVP_MD_flags() (without this the only way to
1667 retrieve a digest flags is by accessing the structure directly. Update
1668 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1669 or cipher is registered as in the "from" argument. Print out all
1670 registered digests in the dgst usage message instead of manually
1671 attempting to work them out.
1674 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1675 this allows the use of compression and extensions. Change default cipher
1676 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1677 by default unless an application cipher string requests it.
1680 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1681 key ids to find matching certificates and keys but some PKCS#12 files
1682 don't follow the (somewhat unwritten) rules and this strategy fails.
1683 Now just gather all certificates together and the first private key
1684 then look for the first certificate that matches the key.
1687 *) Support use of registered digest and cipher names for dgst and cipher
1688 commands instead of having to add each one as a special case. So now
1695 openssl dgst -sha256 foo
1697 and this works for ENGINE based algorithms too.
1701 *) Update Gost ENGINE to support parameter files.
1702 [Victor B. Wagner <vitus@cryptocom.ru>]
1704 *) Support GeneralizedTime in ca utility.
1705 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1707 *) Enhance the hash format used for certificate directory links. The new
1708 form uses the canonical encoding (meaning equivalent names will work
1709 even if they aren't identical) and uses SHA1 instead of MD5. This form
1710 is incompatible with the older format and as a result c_rehash should
1711 be used to rebuild symbolic links.
1714 *) Make PKCS#8 the default write format for private keys, replacing the
1715 traditional format. This form is standardised, more secure and doesn't
1716 include an implicit MD5 dependency.
1719 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1720 committed to OpenSSL should pass this lot as a minimum.
1723 *) Add session ticket override functionality for use by EAP-FAST.
1724 [Jouni Malinen <j@w1.fi>]
1726 *) Modify HMAC functions to return a value. Since these can be implemented
1727 in an ENGINE errors can occur.
1730 *) Type-checked OBJ_bsearch_ex.
1733 *) Type-checked OBJ_bsearch. Also some constification necessitated
1734 by type-checking. Still to come: TXT_DB, bsearch(?),
1735 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1739 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1740 seconds to a tm structure directly, instead of going through OS
1741 specific date routines. This avoids any issues with OS routines such
1742 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1743 and X509_time_adj_ex() to cover the extended range. The existing
1744 X509_time_adj() is still usable and will no longer have any date issues.
1747 *) Delta CRL support. New use deltas option which will attempt to locate
1748 and search any appropriate delta CRLs available.
1750 This work was sponsored by Google.
1753 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1754 code and add additional score elements. Validate alternate CRL paths
1755 as part of the CRL checking and indicate a new error "CRL path validation
1756 error" in this case. Applications wanting additional details can use
1757 the verify callback and check the new "parent" field. If this is not
1758 NULL CRL path validation is taking place. Existing applications wont
1759 see this because it requires extended CRL support which is off by
1762 This work was sponsored by Google.
1765 *) Support for freshest CRL extension.
1767 This work was sponsored by Google.
1770 *) Initial indirect CRL support. Currently only supported in the CRLs
1771 passed directly and not via lookup. Process certificate issuer
1772 CRL entry extension and lookup CRL entries by bother issuer name
1773 and serial number. Check and process CRL issuer entry in IDP extension.
1775 This work was sponsored by Google.
1778 *) Add support for distinct certificate and CRL paths. The CRL issuer
1779 certificate is validated separately in this case. Only enabled if
1780 an extended CRL support flag is set: this flag will enable additional
1781 CRL functionality in future.
1783 This work was sponsored by Google.
1786 *) Add support for policy mappings extension.
1788 This work was sponsored by Google.
1791 *) Fixes to pathlength constraint, self issued certificate handling,
1792 policy processing to align with RFC3280 and PKITS tests.
1794 This work was sponsored by Google.
1797 *) Support for name constraints certificate extension. DN, email, DNS
1798 and URI types are currently supported.
1800 This work was sponsored by Google.
1803 *) To cater for systems that provide a pointer-based thread ID rather
1804 than numeric, deprecate the current numeric thread ID mechanism and
1805 replace it with a structure and associated callback type. This
1806 mechanism allows a numeric "hash" to be extracted from a thread ID in
1807 either case, and on platforms where pointers are larger than 'long',
1808 mixing is done to help ensure the numeric 'hash' is usable even if it
1809 can't be guaranteed unique. The default mechanism is to use "&errno"
1810 as a pointer-based thread ID to distinguish between threads.
1812 Applications that want to provide their own thread IDs should now use
1813 CRYPTO_THREADID_set_callback() to register a callback that will call
1814 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1816 Note that ERR_remove_state() is now deprecated, because it is tied
1817 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1818 to free the current thread's error state should be replaced by
1819 ERR_remove_thread_state(NULL).
1821 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1822 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1823 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1824 application was previously providing a numeric thread callback that
1825 was inappropriate for distinguishing threads, then uniqueness might
1826 have been obtained with &errno that happened immediately in the
1827 intermediate development versions of OpenSSL; this is no longer the
1828 case, the numeric thread callback will now override the automatic use
1830 [Geoff Thorpe, with help from Bodo Moeller]
1832 *) Initial support for different CRL issuing certificates. This covers a
1833 simple case where the self issued certificates in the chain exist and
1834 the real CRL issuer is higher in the existing chain.
1836 This work was sponsored by Google.
1839 *) Removed effectively defunct crypto/store from the build.
1842 *) Revamp of STACK to provide stronger type-checking. Still to come:
1843 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1844 ASN1_STRING, CONF_VALUE.
1847 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1848 RAM on SSL connections. This option can save about 34k per idle SSL.
1851 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1852 STACK, TXT_DB, bsearch, qsort.
1855 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1856 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1857 support for data, signedData, compressedData, digestedData and
1858 encryptedData, envelopedData types included. Scripts to check against
1859 RFC4134 examples draft and interop and consistency checks of many
1860 content types and variants.
1863 *) Add options to enc utility to support use of zlib compression BIO.
1866 *) Extend mk1mf to support importing of options and assembly language
1867 files from Configure script, currently only included in VC-WIN32.
1868 The assembly language rules can now optionally generate the source
1869 files from the associated perl scripts.
1872 *) Implement remaining functionality needed to support GOST ciphersuites.
1873 Interop testing has been performed using CryptoPro implementations.
1874 [Victor B. Wagner <vitus@cryptocom.ru>]
1876 *) s390x assembler pack.
1879 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1883 *) Implement Opaque PRF Input TLS extension as specified in
1884 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1885 official specification yet and no extension type assignment by
1886 IANA exists, this extension (for now) will have to be explicitly
1887 enabled when building OpenSSL by providing the extension number
1888 to use. For example, specify an option
1890 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1892 to the "config" or "Configure" script to enable the extension,
1893 assuming extension number 0x9527 (which is a completely arbitrary
1894 and unofficial assignment based on the MD5 hash of the Internet
1895 Draft). Note that by doing so, you potentially lose
1896 interoperability with other TLS implementations since these might
1897 be using the same extension number for other purposes.
1899 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1900 opaque PRF input value to use in the handshake. This will create
1901 an interal copy of the length-'len' string at 'src', and will
1902 return non-zero for success.
1904 To get more control and flexibility, provide a callback function
1907 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1908 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1912 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1915 Callback function 'cb' will be called in handshakes, and is
1916 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1917 Argument 'arg' is for application purposes (the value as given to
1918 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1919 be provided to the callback function). The callback function
1920 has to return non-zero to report success: usually 1 to use opaque
1921 PRF input just if possible, or 2 to enforce use of the opaque PRF
1922 input. In the latter case, the library will abort the handshake
1923 if opaque PRF input is not successfully negotiated.
1925 Arguments 'peerinput' and 'len' given to the callback function
1926 will always be NULL and 0 in the case of a client. A server will
1927 see the client's opaque PRF input through these variables if
1928 available (NULL and 0 otherwise). Note that if the server
1929 provides an opaque PRF input, the length must be the same as the
1930 length of the client's opaque PRF input.
1932 Note that the callback function will only be called when creating
1933 a new session (session resumption can resume whatever was
1934 previously negotiated), and will not be called in SSL 2.0
1935 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1936 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1937 for applications that need to enforce opaque PRF input.
1941 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1944 [Victor B. Wagner <vitus@cryptocom.ru>]
1946 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1947 RFC4507bis. The encrypted ticket format is an encrypted encoded
1948 SSL_SESSION structure, that way new session features are automatically
1951 If a client application caches session in an SSL_SESSION structure
1952 support is transparent because tickets are now stored in the encoded
1955 The SSL_CTX structure automatically generates keys for ticket
1956 protection in servers so again support should be possible
1957 with no application modification.
1959 If a client or server wishes to disable RFC4507 support then the option
1960 SSL_OP_NO_TICKET can be set.
1962 Add a TLS extension debugging callback to allow the contents of any client
1963 or server extensions to be examined.
1965 This work was sponsored by Google.
1968 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1969 OpenSSL should now compile cleanly on gcc 4.2
1970 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1972 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1973 support including streaming MAC support: this is required for GOST
1974 ciphersuite support.
1975 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1977 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1978 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1979 to output in BER and PEM format.
1982 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1983 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1984 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1985 ENGINE support for HMAC keys which are unextractable. New -mac and
1986 -macopt options to dgst utility.
1989 *) New option -sigopt to dgst utility. Update dgst to use
1990 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1991 alternative signing paramaters such as X9.31 or PSS in the dgst
1995 *) Change ssl_cipher_apply_rule(), the internal function that does
1996 the work each time a ciphersuite string requests enabling
1997 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1998 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1999 the order of disabled ciphersuites such that those ciphersuites
2000 that most recently went from enabled to disabled not only stay
2001 in order with respect to each other, but also have higher priority
2002 than other disabled ciphersuites the next time ciphersuites are
2005 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2006 the same ciphersuites as with "HIGH" alone, but in a specific
2007 order where the PSK ciphersuites come first (since they are the
2008 most recently disabled ciphersuites when "HIGH" is parsed).
2010 Also, change ssl_create_cipher_list() (using this new
2011 funcionality) such that between otherwise identical
2012 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2016 *) Change ssl_create_cipher_list() so that it automatically
2017 arranges the ciphersuites in reasonable order before starting
2018 to process the rule string. Thus, the definition for "DEFAULT"
2019 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2020 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2021 This makes it much easier to arrive at a reasonable default order
2022 in applications for which anonymous ciphers are OK (meaning
2023 that you can't actually use DEFAULT).
2024 [Bodo Moeller; suggested by Victor Duchovni]
2026 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2027 processing) into multiple integers instead of setting
2028 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2029 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2030 (These masks as well as the individual bit definitions are hidden
2031 away into the non-exported interface ssl/ssl_locl.h, so this
2032 change to the definition of the SSL_CIPHER structure shouldn't
2033 affect applications.) This give us more bits for each of these
2034 categories, so there is no longer a need to coagulate AES128 and
2035 AES256 into a single algorithm bit, and to coagulate Camellia128
2036 and Camellia256 into a single algorithm bit, which has led to all
2039 Thus, among other things, the kludge introduced in 0.9.7m and
2040 0.9.8e for masking out AES256 independently of AES128 or masking
2041 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2043 With the change, we also introduce new ciphersuite aliases that
2044 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2048 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2049 Use the leftmost N bytes of the signature input if the input is
2050 larger than the prime q (with N being the size in bytes of q).
2053 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2054 it yet and it is largely untested.
2057 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2060 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2061 some compilers (gcc 4.2 and later) reject their use. Safestack is
2062 reimplemented. Update ASN1 to avoid use of legacy functions.
2065 *) Win32/64 targets are linked with Winsock2.
2068 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2069 to external functions. This can be used to increase CRL handling
2070 efficiency especially when CRLs are very large by (for example) storing
2071 the CRL revoked certificates in a database.
2074 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2075 new CRLs added to a directory can be used. New command line option
2076 -verify_return_error to s_client and s_server. This causes real errors
2077 to be returned by the verify callback instead of carrying on no matter
2078 what. This reflects the way a "real world" verify callback would behave.
2081 *) GOST engine, supporting several GOST algorithms and public key formats.
2082 Kindly donated by Cryptocom.
2085 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2086 partitioned by DP are handled but no indirect CRL or reason partitioning
2087 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2088 selected via a scoring technique which handles IDP and AKID in CRLs.
2091 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2092 will ultimately be used for all verify operations: this will remove the
2093 X509_STORE dependency on certificate verification and allow alternative
2094 lookup methods. X509_STORE based implementations of these two callbacks.
2097 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2098 Modify get_crl() to find a valid (unexpired) CRL if possible.
2101 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2102 this would be called X509_CRL_cmp() but that name is already used by
2103 a function that just compares CRL issuer names. Cache several CRL
2104 extensions in X509_CRL structure and cache CRLDP in X509.
2107 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2108 this maps equivalent X509_NAME structures into a consistent structure.
2109 Name comparison can then be performed rapidly using memcmp().
2112 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2116 *) Allow digests to supply their own micalg string for S/MIME type using
2117 the ctrl EVP_MD_CTRL_MICALG.
2120 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2121 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2122 ctrl. It can then customise the structure before and/or after signing
2126 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2127 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2128 to free up any added signature OIDs.
2131 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2132 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2133 digest and cipher tables. New options added to openssl utility:
2134 list-message-digest-algorithms and list-cipher-algorithms.
2137 *) Change the array representation of binary polynomials: the list
2138 of degrees of non-zero coefficients is now terminated with -1.
2139 Previously it was terminated with 0, which was also part of the
2140 value; thus, the array representation was not applicable to
2141 polynomials where t^0 has coefficient zero. This change makes
2142 the array representation useful in a more general context.
2145 *) Various modifications and fixes to SSL/TLS cipher string
2146 handling. For ECC, the code now distinguishes between fixed ECDH
2147 with RSA certificates on the one hand and with ECDSA certificates
2148 on the other hand, since these are separate ciphersuites. The
2149 unused code for Fortezza ciphersuites has been removed.
2151 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2152 (not "ECDHE"). For consistency with the code for DH
2153 certificates, use of ECDH certificates is now considered ECDH
2154 authentication, not RSA or ECDSA authentication (the latter is
2155 merely the CA's signing algorithm and not actively used in the
2158 The temporary ciphersuite alias "ECCdraft" is no longer
2159 available, and ECC ciphersuites are no longer excluded from "ALL"
2160 and "DEFAULT". The following aliases now exist for RFC 4492
2161 ciphersuites, most of these by analogy with the DH case:
2163 kECDHr - ECDH cert, signed with RSA
2164 kECDHe - ECDH cert, signed with ECDSA
2165 kECDH - ECDH cert (signed with either RSA or ECDSA)
2166 kEECDH - ephemeral ECDH
2167 ECDH - ECDH cert or ephemeral ECDH
2173 AECDH - anonymous ECDH
2174 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2178 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2179 Use correct micalg parameters depending on digest(s) in signed message.
2182 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2183 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2186 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2187 an engine to register a method. Add ENGINE lookups for methods and
2188 functional reference processing.
2191 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2192 EVP_{Sign,Verify}* which allow an application to customise the signature
2196 *) New -resign option to smime utility. This adds one or more signers
2197 to an existing PKCS#7 signedData structure. Also -md option to use an
2198 alternative message digest algorithm for signing.
2201 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2202 create PKCS7 structures containing multiple signers. Update smime
2203 application to support multiple signers.
2206 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2210 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2211 Reorganize PBE internals to lookup from a static table using NIDs,
2212 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2213 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2214 PRF which will be automatically used with PBES2.
2217 *) Replace the algorithm specific calls to generate keys in "req" with the
2221 *) Update PKCS#7 enveloped data routines to use new API. This is now
2222 supported by any public key method supporting the encrypt operation. A
2223 ctrl is added to allow the public key algorithm to examine or modify
2224 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2228 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2229 a default digest type to use. In most cases this will be SHA1 but some
2230 algorithms (such as GOST) need to specify an alternative digest. The
2231 return value indicates how strong the preference is 1 means optional and
2232 2 is mandatory (that is it is the only supported type). Modify
2233 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2234 use the default md. Update openssl utilities to use the default digest
2235 type for signing if it is not explicitly indicated.
2238 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2239 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2240 signing method from the key type. This effectively removes the link
2241 between digests and public key types.
2244 *) Add an OID cross reference table and utility functions. Its purpose is to
2245 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2246 rsaEncryption. This will allow some of the algorithm specific hackery
2247 needed to use the correct OID to be removed.
2250 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2251 structures for PKCS7_sign(). They are now set up by the relevant public
2255 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2258 *) Add support for key derivation (agreement) in the API, DH method and
2262 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2263 public and private key formats. As a side effect these add additional
2264 command line functionality not previously available: DSA signatures can be
2265 generated and verified using pkeyutl and DH key support and generation in
2270 [Oliver Tappe <zooey@hirschkaefer.de>]
2272 *) New make target "install_html_docs" installs HTML renditions of the
2274 [Oliver Tappe <zooey@hirschkaefer.de>]
2276 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2277 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2278 support key and parameter generation and add initial key generation
2279 functionality for RSA.
2282 *) Add functions for main EVP_PKEY_method operations. The undocumented
2283 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2284 EVP_PKEY_{encrypt,decrypt}_old.
2287 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2288 key API, doesn't do much yet.
2291 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2292 public key algorithms. New option to openssl utility:
2293 "list-public-key-algorithms" to print out info.
2296 *) Implement the Supported Elliptic Curves Extension for
2297 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2300 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2301 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2304 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2305 utilities such as rsa, dsa, dsaparam etc except they process any key
2309 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2310 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2311 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2315 *) Initial support for pluggable public key ASN1.
2316 De-spaghettify the public key ASN1 handling. Move public and private
2317 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2318 algorithm specific handling to a single module within the relevant
2319 algorithm directory. Add functions to allow (near) opaque processing
2320 of public and private key structures.
2323 *) Implement the Supported Point Formats Extension for
2324 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2327 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2328 for the psk identity [hint] and the psk callback functions to the
2329 SSL_SESSION, SSL and SSL_CTX structure.
2332 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2336 SSL_CTX_use_psk_identity_hint
2337 SSL_get_psk_identity_hint
2338 SSL_get_psk_identity
2339 SSL_use_psk_identity_hint
2341 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2343 *) Add RFC 3161 compliant time stamp request creation, response generation
2344 and response verification functionality.
2345 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2347 *) Add initial support for TLS extensions, specifically for the server_name
2348 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2349 have new members for a host name. The SSL data structure has an
2350 additional member SSL_CTX *initial_ctx so that new sessions can be
2351 stored in that context to allow for session resumption, even after the
2352 SSL has been switched to a new SSL_CTX in reaction to a client's
2353 server_name extension.
2355 New functions (subject to change):
2357 SSL_get_servername()
2358 SSL_get_servername_type()
2361 New CTRL codes and macros (subject to change):
2363 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2364 - SSL_CTX_set_tlsext_servername_callback()
2365 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2366 - SSL_CTX_set_tlsext_servername_arg()
2367 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2369 openssl s_client has a new '-servername ...' option.
2371 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2372 '-key2 ...', '-servername_fatal' (subject to change). This allows
2373 testing the HostName extension for a specific single host name ('-cert'
2374 and '-key' remain fallbacks for handshakes without HostName
2375 negotiation). If the unrecognized_name alert has to be sent, this by
2376 default is a warning; it becomes fatal with the '-servername_fatal'
2379 [Peter Sylvester, Remy Allais, Christophe Renou]
2381 *) Whirlpool hash implementation is added.
2384 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2385 bn(64,32). Because of instruction set limitations it doesn't have
2386 any negative impact on performance. This was done mostly in order
2387 to make it possible to share assembler modules, such as bn_mul_mont
2388 implementations, between 32- and 64-bit builds without hassle.
2391 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2392 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2396 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2397 dedicated Montgomery multiplication procedure, is introduced.
2398 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2399 "64-bit" performance on certain 32-bit targets.
2402 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2403 in SSL structures. New SSL ctrl to set maximum send fragment size.
2404 Save memory by seeting the I/O buffer sizes dynamically instead of
2405 using the maximum available value.
2408 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2409 in addition to the text details.
2412 *) Very, very preliminary EXPERIMENTAL support for printing of general
2413 ASN1 structures. This currently produces rather ugly output and doesn't
2414 handle several customised structures at all.
2417 *) Integrated support for PVK file format and some related formats such
2418 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2419 these in the 'rsa' and 'dsa' utilities.
2422 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2425 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2426 place for the (very old) "NETSCAPE" format certificates which are now
2427 handled using new ASN1 code equivalents.
2430 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2431 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2432 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2435 *) Modify CRL distribution points extension code to print out previously
2436 unsupported fields. Enhance extension setting code to allow setting of
2440 *) Add print and set support for Issuing Distribution Point CRL extension.
2443 *) Change 'Configure' script to enable Camellia by default.
2446 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2448 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2450 This addresses the flaw in CBC record processing discovered by
2451 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2452 at: http://www.isg.rhul.ac.uk/tls/
2454 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2455 Security Group at Royal Holloway, University of London
2456 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2457 Emilia Käsper for the initial patch.
2459 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2461 *) Return an error when checking OCSP signatures when key is NULL.
2462 This fixes a DoS attack. (CVE-2013-0166)
2465 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2466 the right response is stapled. Also change SSL_get_certificate()
2467 so it returns the certificate actually sent.
2468 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2469 (This is a backport)
2470 [Rob Stradling <rob.stradling@comodo.com>]
2472 *) Fix possible deadlock when decoding public keys.
2475 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2477 *) Sanity check record length before skipping explicit IV in DTLS
2480 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2481 fuzzing as a service testing platform.
2485 *) Initialise tkeylen properly when encrypting CMS messages.
2486 Thanks to Solar Designer of Openwall for reporting this issue.
2489 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2491 *) The fix for CVE-2012-2110 did not take into account that the
2492 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2493 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2494 rejecting negative len parameter. (CVE-2012-2131)
2495 [Tomas Hoger <thoger@redhat.com>]
2497 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2499 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2500 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2501 in CRYPTO_realloc_clean.
2503 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2504 issue and to Adam Langley <agl@chromium.org> for fixing it.
2506 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2508 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2510 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2511 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2512 content decryption and always return the same error. Note: this attack
2513 needs on average 2^20 messages so it only affects automated senders. The
2514 old behaviour can be reenabled in the CMS code by setting the
2515 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2516 an MMA defence is not necessary.
2517 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2518 this issue. (CVE-2012-0884)
2521 *) Fix CVE-2011-4619: make sure we really are receiving a
2522 client hello before rejecting multiple SGC restarts. Thanks to
2523 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2526 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2528 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2529 Thanks to Antonio Martin, Enterprise Secure Access Research and
2530 Development, Cisco Systems, Inc. for discovering this bug and
2531 preparing a fix. (CVE-2012-0050)
2534 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2536 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2537 of the Vaudenay padding oracle attack on CBC mode encryption
2538 which enables an efficient plaintext recovery attack against
2539 the OpenSSL implementation of DTLS. Their attack exploits timing
2540 differences arising during decryption processing. A research
2541 paper describing this attack can be found at:
2542 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2543 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2544 Security Group at Royal Holloway, University of London
2545 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2546 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2547 for preparing the fix. (CVE-2011-4108)
2548 [Robin Seggelmann, Michael Tuexen]
2550 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2551 [Ben Laurie, Kasper <ekasper@google.com>]
2553 *) Clear bytes used for block padding of SSL 3.0 records.
2555 [Adam Langley (Google)]
2557 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2558 Kadianakis <desnacked@gmail.com> for discovering this issue and
2559 Adam Langley for preparing the fix. (CVE-2011-4619)
2560 [Adam Langley (Google)]
2562 *) Prevent malformed RFC3779 data triggering an assertion failure.
2563 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2564 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2565 [Rob Austein <sra@hactrn.net>]
2567 *) Fix ssl_ciph.c set-up race.
2568 [Adam Langley (Google)]
2570 *) Fix spurious failures in ecdsatest.c.
2571 [Emilia Käsper (Google)]
2573 *) Fix the BIO_f_buffer() implementation (which was mixing different
2574 interpretations of the '..._len' fields).
2575 [Adam Langley (Google)]
2577 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2578 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2579 threads won't reuse the same blinding coefficients.
2581 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2582 lock to call BN_BLINDING_invert_ex, and avoids one use of
2583 BN_BLINDING_update for each BN_BLINDING structure (previously,
2584 the last update always remained unused).
2585 [Emilia Käsper (Google)]
2587 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2588 for multi-threaded use of ECDH.
2589 [Adam Langley (Google)]
2591 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2594 *) Add protection against ECDSA timing attacks as mentioned in the paper
2595 by Billy Bob Brumley and Nicola Tuveri, see:
2597 http://eprint.iacr.org/2011/232.pdf
2599 [Billy Bob Brumley and Nicola Tuveri]
2601 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2603 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2604 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2606 *) Fix bug in string printing code: if *any* escaping is enabled we must
2607 escape the escape character (backslash) or the resulting string is
2611 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2613 *) Disable code workaround for ancient and obsolete Netscape browsers
2614 and servers: an attacker can use it in a ciphersuite downgrade attack.
2615 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2618 *) Fixed J-PAKE implementation error, originally discovered by
2619 Sebastien Martini, further info and confirmation from Stefan
2620 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2623 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2625 *) Fix extension code to avoid race conditions which can result in a buffer
2626 overrun vulnerability: resumed sessions must not be modified as they can
2627 be shared by multiple threads. CVE-2010-3864
2630 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2633 *) Don't reencode certificate when calculating signature: cache and use
2634 the original encoding instead. This makes signature verification of
2635 some broken encodings work correctly.
2638 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2639 is also one of the inputs.
2640 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2642 *) Don't repeatedly append PBE algorithms to table if they already exist.
2643 Sort table on each new add. This effectively makes the table read only
2644 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2648 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2650 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2653 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2654 access or freeing data twice (CVE-2010-0742)
2655 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2657 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2658 common in certificates and some applications which only call
2659 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2663 Reduce copying into .apps and .test in makevms.com
2664 Don't try to use blank CA certificate in CA.com
2665 Allow use of C files from original directories in maketests.com
2666 [Steven M. Schweda" <sms@antinode.info>]
2668 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2670 *) When rejecting SSL/TLS records due to an incorrect version number, never
2671 update s->server with a new major version number. As of
2672 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2673 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2674 the previous behavior could result in a read attempt at NULL when
2675 receiving specific incorrect SSL/TLS records once record payload
2676 protection is active. (CVE-2010-0740)
2677 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2679 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2680 could be crashed if the relevant tables were not present (e.g. chrooted).
2681 [Tomas Hoger <thoger@redhat.com>]
2683 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2685 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2686 [Martin Olsson, Neel Mehta]
2688 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2689 accommodate for stack sorting, always a write lock!).
2692 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2693 excessive delays in the RAND_poll(): over a minute. As a workaround
2694 include a time check in the inner Heap32Next loop too.
2697 *) The code that handled flushing of data in SSL/TLS originally used the
2698 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2699 the problem outlined in PR#1949. The fix suggested there however can
2700 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2701 of Apache). So instead simplify the code to flush unconditionally.
2702 This should be fine since flushing with no data to flush is a no op.
2705 *) Handle TLS versions 2.0 and later properly and correctly use the
2706 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2707 off ancient servers have a habit of sticking around for a while...
2710 *) Modify compression code so it frees up structures without using the
2711 ex_data callbacks. This works around a problem where some applications
2712 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2713 restarting) then use compression (e.g. SSL with compression) later.
2714 This results in significant per-connection memory leaks and
2715 has caused some security issues including CVE-2008-1678 and
2719 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2720 change when encrypting or decrypting.
2723 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2724 connect and renegotiate with servers which do not support RI.
2725 Until RI is more widely deployed this option is enabled by default.
2728 *) Add "missing" ssl ctrls to clear options and mode.
2731 *) If client attempts to renegotiate and doesn't support RI respond with
2732 a no_renegotiation alert as required by RFC5746. Some renegotiating
2733 TLS clients will continue a connection gracefully when they receive
2734 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2735 waiting for a server hello which it will never receive. Now we treat a
2736 received no_renegotiation alert as a fatal error. This is because
2737 applications requesting a renegotiation might well expect it to succeed
2738 and would have no code in place to handle the server denying it so the
2739 only safe thing to do is to terminate the connection.
2742 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2743 peer supports secure renegotiation and 0 otherwise. Print out peer
2744 renegotiation support in s_client/s_server.
2747 *) Replace the highly broken and deprecated SPKAC certification method with
2748 the updated NID creation version. This should correctly handle UTF8.
2751 *) Implement RFC5746. Re-enable renegotiation but require the extension
2752 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2753 turns out to be a bad idea. It has been replaced by
2754 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2755 SSL_CTX_set_options(). This is really not recommended unless you
2756 know what you are doing.
2757 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2759 *) Fixes to stateless session resumption handling. Use initial_ctx when
2760 issuing and attempting to decrypt tickets in case it has changed during
2761 servername handling. Use a non-zero length session ID when attempting
2762 stateless session resumption: this makes it possible to determine if
2763 a resumption has occurred immediately after receiving server hello
2764 (several places in OpenSSL subtly assume this) instead of later in
2768 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2769 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2770 fixes for a few places where the return code is not checked
2772 [Julia Lawall <julia@diku.dk>]
2774 *) Add --strict-warnings option to Configure script to include devteam
2775 warnings in other configurations.
2778 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2779 makes it possible to install openssl libraries in locations which
2780 have names other than "lib", for example "/usr/lib64" which some
2782 [Steve Henson, based on patch from Jeremy Utley]
2784 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2785 X690 8.9.12 and can produce some misleading textual output of OIDs.
2786 [Steve Henson, reported by Dan Kaminsky]
2788 *) Delete MD2 from algorithm tables. This follows the recommendation in
2789 several standards that it is not used in new applications due to
2790 several cryptographic weaknesses. For binary compatibility reasons
2791 the MD2 API is still compiled in by default.
2794 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2798 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2799 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2801 [Guenter <lists@gknw.net>]
2803 *) Fix the server certificate chain building code to use X509_verify_cert(),
2804 it used to have an ad-hoc builder which was unable to cope with anything
2805 other than a simple chain.
2806 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2808 *) Don't check self signed certificate signatures in X509_verify_cert()
2809 by default (a flag can override this): it just wastes time without
2810 adding any security. As a useful side effect self signed root CAs
2811 with non-FIPS digests are now usable in FIPS mode.
2814 *) In dtls1_process_out_of_seq_message() the check if the current message
2815 is already buffered was missing. For every new message was memory
2816 allocated, allowing an attacker to perform an denial of service attack
2817 with sending out of seq handshake messages until there is no memory
2818 left. Additionally every future messege was buffered, even if the
2819 sequence number made no sense and would be part of another handshake.
2820 So only messages with sequence numbers less than 10 in advance will be
2821 buffered. (CVE-2009-1378)
2822 [Robin Seggelmann, discovered by Daniel Mentz]
2824 *) Records are buffered if they arrive with a future epoch to be
2825 processed after finishing the corresponding handshake. There is
2826 currently no limitation to this buffer allowing an attacker to perform
2827 a DOS attack with sending records with future epochs until there is no
2828 memory left. This patch adds the pqueue_size() function to determine
2829 the size of a buffer and limits the record buffer to 100 entries.
2831 [Robin Seggelmann, discovered by Daniel Mentz]
2833 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2834 parent structure is freed. (CVE-2009-1379)
2837 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2838 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2841 [Ilya O. <vrghost@gmail.com>]
2843 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2845 *) Disable renegotiation completely - this fixes a severe security
2846 problem (CVE-2009-3555) at the cost of breaking all
2847 renegotiation. Renegotiation can be re-enabled by setting
2848 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2849 run-time. This is really not recommended unless you know what
2853 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2855 *) Don't set val to NULL when freeing up structures, it is freed up by
2856 underlying code. If sizeof(void *) > sizeof(long) this can result in
2857 zeroing past the valid field. (CVE-2009-0789)
2858 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2860 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2861 checked correctly. This would allow some invalid signed attributes to
2862 appear to verify correctly. (CVE-2009-0591)
2863 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2865 *) Reject UniversalString and BMPString types with invalid lengths. This
2866 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2867 a legal length. (CVE-2009-0590)
2870 *) Set S/MIME signing as the default purpose rather than setting it
2871 unconditionally. This allows applications to override it at the store
2875 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2876 to handle some structures.
2879 *) Improve efficiency of mem_gets: don't search whole buffer each time
2881 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2883 *) New -hex option for openssl rand.
2886 *) Print out UTF8String and NumericString when parsing ASN1.
2889 *) Support NumericString type for name components.
2892 *) Allow CC in the environment to override the automatically chosen
2893 compiler. Note that nothing is done to ensure flags work with the
2897 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2899 *) Properly check EVP_VerifyFinal() and similar return values
2901 [Ben Laurie, Bodo Moeller, Google Security Team]
2903 *) Enable TLS extensions by default.
2906 *) Allow the CHIL engine to be loaded, whether the application is
2907 multithreaded or not. (This does not release the developer from the
2908 obligation to set up the dynamic locking callbacks.)
2909 [Sander Temme <sander@temme.net>]
2911 *) Use correct exit code if there is an error in dgst command.
2912 [Steve Henson; problem pointed out by Roland Dirlewanger]
2914 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2915 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2918 *) Add experimental JPAKE support, including demo authentication in
2919 s_client and s_server.
2922 *) Set the comparison function in v3_addr_canonize().
2923 [Rob Austein <sra@hactrn.net>]
2925 *) Add support for XMPP STARTTLS in s_client.
2926 [Philip Paeps <philip@freebsd.org>]
2928 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2929 to ensure that even with this option, only ciphersuites in the
2930 server's preference list will be accepted. (Note that the option
2931 applies only when resuming a session, so the earlier behavior was
2932 just about the algorithm choice for symmetric cryptography.)
2935 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2937 *) Fix NULL pointer dereference if a DTLS server received
2938 ChangeCipherSpec as first record (CVE-2009-1386).
2941 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2942 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2945 *) The fix in 0.9.8c that supposedly got rid of unsafe
2946 double-checked locking was incomplete for RSA blinding,
2947 addressing just one layer of what turns out to have been
2948 doubly unsafe triple-checked locking.
2950 So now fix this for real by retiring the MONT_HELPER macro
2951 in crypto/rsa/rsa_eay.c.
2953 [Bodo Moeller; problem pointed out by Marius Schilder]
2955 *) Various precautionary measures:
2957 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2959 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2960 (NB: This would require knowledge of the secret session ticket key
2961 to exploit, in which case you'd be SOL either way.)
2963 - Change bn_nist.c so that it will properly handle input BIGNUMs
2964 outside the expected range.
2966 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2969 [Neel Mehta, Bodo Moeller]
2971 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2972 the load fails. Useful for distros.
2973 [Ben Laurie and the FreeBSD team]
2975 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2978 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2981 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2983 This work was sponsored by Logica.
2986 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2987 keystores. Support for SSL/TLS client authentication too.
2988 Not compiled unless enable-capieng specified to Configure.
2990 This work was sponsored by Logica.
2993 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2994 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2995 attribute creation routines such as certifcate requests and PKCS#12
2999 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3001 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3002 handshake which could lead to a cilent crash as found using the
3003 Codenomicon TLS test suite (CVE-2008-1672)
3004 [Steve Henson, Mark Cox]
3006 *) Fix double free in TLS server name extensions which could lead to
3007 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3010 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3012 Clear the error queue to ensure that error entries left from
3013 older function calls do not interfere with the correct operation.
3014 [Lutz Jaenicke, Erik de Castro Lopo]
3016 *) Remove root CA certificates of commercial CAs:
3018 The OpenSSL project does not recommend any specific CA and does not
3019 have any policy with respect to including or excluding any CA.
3020 Therefore it does not make any sense to ship an arbitrary selection
3021 of root CA certificates with the OpenSSL software.
3024 *) RSA OAEP patches to fix two separate invalid memory reads.
3025 The first one involves inputs when 'lzero' is greater than
3026 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3027 before the beginning of from). The second one involves inputs where
3028 the 'db' section contains nothing but zeroes (there is a one-byte
3029 invalid read after the end of 'db').
3030 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3032 *) Partial backport from 0.9.9-dev:
3034 Introduce bn_mul_mont (dedicated Montgomery multiplication
3035 procedure) as a candidate for BIGNUM assembler implementation.
3036 While 0.9.9-dev uses assembler for various architectures, only
3037 x86_64 is available by default here in the 0.9.8 branch, and
3038 32-bit x86 is available through a compile-time setting.
3040 To try the 32-bit x86 assembler implementation, use Configure
3041 option "enable-montasm" (which exists only for this backport).
3043 As "enable-montasm" for 32-bit x86 disclaims code stability
3044 anyway, in this constellation we activate additional code
3045 backported from 0.9.9-dev for further performance improvements,
3046 namely BN_from_montgomery_word. (To enable this otherwise,
3047 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3049 [Andy Polyakov (backport partially by Bodo Moeller)]
3051 *) Add TLS session ticket callback. This allows an application to set
3052 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3053 values. This is useful for key rollover for example where several key
3054 sets may exist with different names.
3057 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3058 This was broken until now in 0.9.8 releases, such that the only way
3059 a registered ENGINE could be used (assuming it initialises
3060 successfully on the host) was to explicitly set it as the default
3061 for the relevant algorithms. This is in contradiction with 0.9.7
3062 behaviour and the documentation. With this fix, when an ENGINE is
3063 registered into a given algorithm's table of implementations, the
3064 'uptodate' flag is reset so that auto-discovery will be used next
3065 time a new context for that algorithm attempts to select an
3067 [Ian Lister (tweaked by Geoff Thorpe)]
3069 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3070 implemention in the following ways:
3072 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3075 Lack of BER streaming support means one pass streaming processing is
3076 only supported if data is detached: setting the streaming flag is
3077 ignored for embedded content.
3079 CMS support is disabled by default and must be explicitly enabled
3080 with the enable-cms configuration option.
3083 *) Update the GMP engine glue to do direct copies between BIGNUM and
3084 mpz_t when openssl and GMP use the same limb size. Otherwise the
3085 existing "conversion via a text string export" trick is still used.
3086 [Paul Sheer <paulsheer@gmail.com>]
3088 *) Zlib compression BIO. This is a filter BIO which compressed and
3089 uncompresses any data passed through it.
3092 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3093 RFC3394 compatible AES key wrapping.
3096 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3097 sets string data without copying. X509_ALGOR_set0() and
3098 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3099 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3100 from an X509_ATTRIBUTE structure optionally checking it occurs only
3101 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3105 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3106 to get the expected BN_FLG_CONSTTIME behavior.
3107 [Bodo Moeller (Google)]
3111 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3112 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3113 - added some more tests to do_tests.pl
3114 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3115 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3116 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3117 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3118 - various changes to netware.pl to enable gcc-cross builds on Win32
3120 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3121 - various changes to fix missing prototype warnings
3122 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3123 - added AES, WHIRLPOOL and CPUID assembler code to build files
3124 - added missing AES assembler make rules to mk1mf.pl
3125 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3126 [Guenter Knauf <eflash@gmx.net>]
3128 *) Implement certificate status request TLS extension defined in RFC3546.
3129 A client can set the appropriate parameters and receive the encoded
3130 OCSP response via a callback. A server can query the supplied parameters
3131 and set the encoded OCSP response in the callback. Add simplified examples
3132 to s_client and s_server.
3135 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3137 *) Fix various bugs:
3138 + Binary incompatibility of ssl_ctx_st structure
3139 + DTLS interoperation with non-compliant servers
3140 + Don't call get_session_cb() without proposed session
3141 + Fix ia64 assembler code
3142 [Andy Polyakov, Steve Henson]
3144 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3146 *) DTLS Handshake overhaul. There were longstanding issues with
3147 OpenSSL DTLS implementation, which were making it impossible for
3148 RFC 4347 compliant client to communicate with OpenSSL server.
3149 Unfortunately just fixing these incompatibilities would "cut off"
3150 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3151 server keeps tolerating non RFC compliant syntax. The opposite is
3152 not true, 0.9.8f client can not communicate with earlier server.
3153 This update even addresses CVE-2007-4995.
3156 *) Changes to avoid need for function casts in OpenSSL: some compilers
3157 (gcc 4.2 and later) reject their use.
3158 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3161 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3162 RFC4507bis. The encrypted ticket format is an encrypted encoded
3163 SSL_SESSION structure, that way new session features are automatically
3166 If a client application caches session in an SSL_SESSION structure
3167 support is transparent because tickets are now stored in the encoded
3170 The SSL_CTX structure automatically generates keys for ticket
3171 protection in servers so again support should be possible
3172 with no application modification.
3174 If a client or server wishes to disable RFC4507 support then the option
3175 SSL_OP_NO_TICKET can be set.
3177 Add a TLS extension debugging callback to allow the contents of any client
3178 or server extensions to be examined.
3180 This work was sponsored by Google.
3183 *) Add initial support for TLS extensions, specifically for the server_name
3184 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3185 have new members for a host name. The SSL data structure has an
3186 additional member SSL_CTX *initial_ctx so that new sessions can be
3187 stored in that context to allow for session resumption, even after the
3188 SSL has been switched to a new SSL_CTX in reaction to a client's
3189 server_name extension.
3191 New functions (subject to change):
3193 SSL_get_servername()
3194 SSL_get_servername_type()
3197 New CTRL codes and macros (subject to change):
3199 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3200 - SSL_CTX_set_tlsext_servername_callback()
3201 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3202 - SSL_CTX_set_tlsext_servername_arg()
3203 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3205 openssl s_client has a new '-servername ...' option.
3207 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3208 '-key2 ...', '-servername_fatal' (subject to change). This allows
3209 testing the HostName extension for a specific single host name ('-cert'
3210 and '-key' remain fallbacks for handshakes without HostName
3211 negotiation). If the unrecognized_name alert has to be sent, this by
3212 default is a warning; it becomes fatal with the '-servername_fatal'
3215 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3217 *) Add AES and SSE2 assembly language support to VC++ build.
3220 *) Mitigate attack on final subtraction in Montgomery reduction.
3223 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3224 (which previously caused an internal error).
3227 *) Squeeze another 10% out of IGE mode when in != out.
3230 *) AES IGE mode speedup.
3231 [Dean Gaudet (Google)]
3233 *) Add the Korean symmetric 128-bit cipher SEED (see
3234 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3235 add SEED ciphersuites from RFC 4162:
3237 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3238 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3239 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3240 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3242 To minimize changes between patchlevels in the OpenSSL 0.9.8
3243 series, SEED remains excluded from compilation unless OpenSSL
3244 is configured with 'enable-seed'.
3245 [KISA, Bodo Moeller]
3247 *) Mitigate branch prediction attacks, which can be practical if a
3248 single processor is shared, allowing a spy process to extract
3249 information. For detailed background information, see
3250 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3251 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3252 and Necessary Software Countermeasures"). The core of the change
3253 are new versions BN_div_no_branch() and
3254 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3255 respectively, which are slower, but avoid the security-relevant
3256 conditional branches. These are automatically called by BN_div()
3257 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3258 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3259 remove a conditional branch.
3261 BN_FLG_CONSTTIME is the new name for the previous
3262 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3263 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3264 in the exponent causes BN_mod_exp_mont() to use the alternative
3265 implementation in BN_mod_exp_mont_consttime().) The old name
3266 remains as a deprecated alias.
3268 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3269 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3270 constant-time implementations for more than just exponentiation.
3271 Here too the old name is kept as a deprecated alias.
3273 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3274 the BN_BLINDING structure gets an independent copy of the
3275 modulus. This means that the previous "BIGNUM *m" argument to
3276 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3277 essentially becomes "const BIGNUM *m", although we can't actually
3278 change this in the header file before 0.9.9. It allows
3279 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3280 enable BN_FLG_CONSTTIME.
3282 [Matthew D Wood (Intel Corp)]
3284 *) In the SSL/TLS server implementation, be strict about session ID
3285 context matching (which matters if an application uses a single
3286 external cache for different purposes). Previously,
3287 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3288 set. This did ensure strict client verification, but meant that,
3289 with applications using a single external cache for quite
3290 different requirements, clients could circumvent ciphersuite
3291 restrictions for a given session ID context by starting a session
3292 in a different context.
3295 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3296 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3297 authentication-only ciphersuites.
3300 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3301 not complete and could lead to a possible single byte overflow
3302 (CVE-2007-5135) [Ben Laurie]
3304 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3306 *) Since AES128 and AES256 (and similarly Camellia128 and
3307 Camellia256) share a single mask bit in the logic of
3308 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3309 kludge to work properly if AES128 is available and AES256 isn't
3310 (or if Camellia128 is available and Camellia256 isn't).
3313 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3314 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3315 When a point or a seed is encoded in a BIT STRING, we need to
3316 prevent the removal of trailing zero bits to get the proper DER
3317 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3318 of a NamedBitList, for which trailing 0 bits need to be removed.)
3321 *) Have SSL/TLS server implementation tolerate "mismatched" record
3322 protocol version while receiving ClientHello even if the
3323 ClientHello is fragmented. (The server can't insist on the
3324 particular protocol version it has chosen before the ServerHello
3325 message has informed the client about his choice.)
3328 *) Add RFC 3779 support.
3329 [Rob Austein for ARIN, Ben Laurie]
3331 *) Load error codes if they are not already present instead of using a
3332 static variable. This allows them to be cleanly unloaded and reloaded.
3333 Improve header file function name parsing.
3336 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3337 or CAPABILITY handshake as required by RFCs.
3340 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3342 *) Introduce limits to prevent malicious keys being able to
3343 cause a denial of service. (CVE-2006-2940)
3344 [Steve Henson, Bodo Moeller]
3346 *) Fix ASN.1 parsing of certain invalid structures that can result
3347 in a denial of service. (CVE-2006-2937) [Steve Henson]
3349 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3350 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3352 *) Fix SSL client code which could crash if connecting to a
3353 malicious SSLv2 server. (CVE-2006-4343)
3354 [Tavis Ormandy and Will Drewry, Google Security Team]
3356 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3357 match only those. Before that, "AES256-SHA" would be interpreted
3358 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3359 the same strength classification in 0.9.7h) as we currently only
3360 have a single AES bit in the ciphersuite description bitmap.
3361 That change, however, also applied to ciphersuite strings such as
3362 "RC4-MD5" that intentionally matched multiple ciphersuites --
3363 namely, SSL 2.0 ciphersuites in addition to the more common ones
3364 from SSL 3.0/TLS 1.0.
3366 So we change the selection algorithm again: Naming an explicit
3367 ciphersuite selects this one ciphersuite, and any other similar
3368 ciphersuite (same bitmap) from *other* protocol versions.
3369 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3370 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3372 Since SSL 2.0 does not have any ciphersuites for which the
3373 128/256 bit distinction would be relevant, this works for now.
3374 The proper fix will be to use different bits for AES128 and
3375 AES256, which would have avoided the problems from the beginning;
3376 however, bits are scarce, so we can only do this in a new release
3377 (not just a patchlevel) when we can change the SSL_CIPHER
3378 definition to split the single 'unsigned long mask' bitmap into
3379 multiple values to extend the available space.
3383 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3385 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3386 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3388 *) Add AES IGE and biIGE modes.
3391 *) Change the Unix randomness entropy gathering to use poll() when
3392 possible instead of select(), since the latter has some
3393 undesirable limitations.
3394 [Darryl Miles via Richard Levitte and Bodo Moeller]
3396 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3397 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3398 cannot be implicitly activated as part of, e.g., the "AES" alias.
3399 However, please upgrade to OpenSSL 0.9.9[-dev] for
3400 non-experimental use of the ECC ciphersuites to get TLS extension
3401 support, which is required for curve and point format negotiation
3402 to avoid potential handshake problems.
3405 *) Disable rogue ciphersuites:
3407 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3408 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3409 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3411 The latter two were purportedly from
3412 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3415 Also deactivate the remaining ciphersuites from
3416 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3417 unofficial, and the ID has long expired.
3420 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3421 dual-core machines) and other potential thread-safety issues.
3424 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3425 versions), which is now available for royalty-free use
3426 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3427 Also, add Camellia TLS ciphersuites from RFC 4132.
3429 To minimize changes between patchlevels in the OpenSSL 0.9.8
3430 series, Camellia remains excluded from compilation unless OpenSSL
3431 is configured with 'enable-camellia'.
3434 *) Disable the padding bug check when compression is in use. The padding
3435 bug check assumes the first packet is of even length, this is not
3436 necessarily true if compresssion is enabled and can result in false
3437 positives causing handshake failure. The actual bug test is ancient
3438 code so it is hoped that implementations will either have fixed it by
3439 now or any which still have the bug do not support compression.
3442 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3444 *) When applying a cipher rule check to see if string match is an explicit
3445 cipher suite and only match that one cipher suite if it is.
3448 *) Link in manifests for VC++ if needed.
3449 [Austin Ziegler <halostatue@gmail.com>]
3451 *) Update support for ECC-based TLS ciphersuites according to
3452 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3453 TLS extensions, which are supported starting with the 0.9.9
3454 branch, not in the OpenSSL 0.9.8 branch).
3457 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3458 opaque EVP_CIPHER_CTX handling.
3461 *) Fixes and enhancements to zlib compression code. We now only use
3462 "zlib1.dll" and use the default __cdecl calling convention on Win32
3463 to conform with the standards mentioned here:
3464 http://www.zlib.net/DLL_FAQ.txt
3465 Static zlib linking now works on Windows and the new --with-zlib-include
3466 --with-zlib-lib options to Configure can be used to supply the location
3467 of the headers and library. Gracefully handle case where zlib library
3471 *) Several fixes and enhancements to the OID generation code. The old code
3472 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3473 handle numbers larger than ULONG_MAX, truncated printing and had a
3474 non standard OBJ_obj2txt() behaviour.
3477 *) Add support for building of engines under engine/ as shared libraries
3478 under VC++ build system.
3481 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3482 Hopefully, we will not see any false combination of paths any more.
3485 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3487 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3488 (part of SSL_OP_ALL). This option used to disable the
3489 countermeasure against man-in-the-middle protocol-version
3490 rollback in the SSL 2.0 server implementation, which is a bad
3491 idea. (CVE-2005-2969)
3493 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3494 for Information Security, National Institute of Advanced Industrial
3495 Science and Technology [AIST], Japan)]
3497 *) Add two function to clear and return the verify parameter flags.
3500 *) Keep cipherlists sorted in the source instead of sorting them at
3501 runtime, thus removing the need for a lock.
3504 *) Avoid some small subgroup attacks in Diffie-Hellman.
3505 [Nick Mathewson and Ben Laurie]
3507 *) Add functions for well-known primes.
3510 *) Extended Windows CE support.
3511 [Satoshi Nakamura and Andy Polyakov]
3513 *) Initialize SSL_METHOD structures at compile time instead of during
3514 runtime, thus removing the need for a lock.
3517 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3518 attempting to decrypt each encrypted key in turn. Add support to
3522 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3524 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3527 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3530 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3531 key into the same file any more.
3534 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3537 *) Add -utf8 command line and config file option to 'ca'.
3538 [Stefan <stf@udoma.org]
3540 *) Removed the macro des_crypt(), as it seems to conflict with some
3541 libraries. Use DES_crypt().
3544 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3545 involves renaming the source and generated shared-libs for
3546 both. The engines will accept the corrected or legacy ids
3547 ('ncipher' and '4758_cca' respectively) when binding. NB,
3548 this only applies when building 'shared'.
3549 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3551 *) Add attribute functions to EVP_PKEY structure. Modify
3552 PKCS12_create() to recognize a CSP name attribute and
3553 use it. Make -CSP option work again in pkcs12 utility.
3556 *) Add new functionality to the bn blinding code:
3557 - automatic re-creation of the BN_BLINDING parameters after
3558 a fixed number of uses (currently 32)
3559 - add new function for parameter creation
3560 - introduce flags to control the update behaviour of the
3561 BN_BLINDING parameters
3562 - hide BN_BLINDING structure
3563 Add a second BN_BLINDING slot to the RSA structure to improve
3564 performance when a single RSA object is shared among several
3568 *) Add support for DTLS.
3569 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3571 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3572 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3575 *) Remove buggy and incomplete DH cert support from
3576 ssl/ssl_rsa.c and ssl/s3_both.c
3579 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3580 the apps/openssl applications.
3583 *) Compile clean with "-Wall -Wmissing-prototypes
3584 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3585 DEBUG_SAFESTACK must also be set.
3588 *) Change ./Configure so that certain algorithms can be disabled by default.
3589 The new counterpiece to "no-xxx" is "enable-xxx".
3591 The patented RC5 and MDC2 algorithms will now be disabled unless
3592 "enable-rc5" and "enable-mdc2", respectively, are specified.
3594 (IDEA remains enabled despite being patented. This is because IDEA
3595 is frequently required for interoperability, and there is no license
3596 fee for non-commercial use. As before, "no-idea" can be used to
3597 avoid this algorithm.)
3601 *) Add processing of proxy certificates (see RFC 3820). This work was
3602 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3603 EGEE (Enabling Grids for E-science in Europe).
3606 *) RC4 performance overhaul on modern architectures/implementations, such
3607 as Intel P4, IA-64 and AMD64.
3610 *) New utility extract-section.pl. This can be used specify an alternative
3611 section number in a pod file instead of having to treat each file as
3612 a separate case in Makefile. This can be done by adding two lines to the
3615 =for comment openssl_section:XXX
3617 The blank line is mandatory.
3621 *) New arguments -certform, -keyform and -pass for s_client and s_server
3622 to allow alternative format key and certificate files and passphrase
3626 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3627 update associated structures and add various utility functions.
3629 Add new policy related verify parameters, include policy checking in
3630 standard verify code. Enhance 'smime' application with extra parameters
3631 to support policy checking and print out.
3634 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3635 Nehemiah processors. These extensions support AES encryption in hardware
3636 as well as RNG (though RNG support is currently disabled).
3637 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3639 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3642 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3643 [Andy Polyakov and a number of other people]
3645 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3646 implementation contributed by IBM.
3647 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3649 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3650 exponent rather than 'unsigned long'. There is a corresponding change to
3651 the new 'rsa_keygen' element of the RSA_METHOD structure.
3652 [Jelte Jansen, Geoff Thorpe]
3654 *) Functionality for creating the initial serial number file is now
3655 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3657 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3658 number file to 1, which is bound to cause problems. To avoid
3659 the problems while respecting compatibility between different 0.9.7
3660 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3661 CA.pl for serial number initialization. With the new release 0.9.8,
3662 we can fix the problem directly in the 'ca' utility.)
3665 *) Reduced header interdepencies by declaring more opaque objects in
3666 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3667 give fewer recursive includes, which could break lazy source code - so
3668 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3669 developers should define this symbol when building and using openssl to
3670 ensure they track the recommended behaviour, interfaces, [etc], but
3671 backwards-compatible behaviour prevails when this isn't defined.
3674 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3677 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3678 This will generate a random key of the appropriate length based on the
3679 cipher context. The EVP_CIPHER can provide its own random key generation
3680 routine to support keys of a specific form. This is used in the des and
3681 3des routines to generate a key of the correct parity. Update S/MIME
3682 code to use new functions and hence generate correct parity DES keys.
3683 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3684 valid (weak or incorrect parity).
3687 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3688 as looking them up. This is useful when the verified structure may contain
3689 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3690 present unless the new PKCS7_NO_CRL flag is asserted.
3693 *) Extend ASN1 oid configuration module. It now additionally accepts the
3696 shortName = some long name, 1.2.3.4
3699 *) Reimplemented the BN_CTX implementation. There is now no more static
3700 limitation on the number of variables it can handle nor the depth of the
3701 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3702 information can now expand as required, and rather than having a single
3703 static array of bignums, BN_CTX now uses a linked-list of such arrays
3704 allowing it to expand on demand whilst maintaining the usefulness of
3705 BN_CTX's "bundling".
3708 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3709 to allow all RSA operations to function using a single BN_CTX.
3712 *) Preliminary support for certificate policy evaluation and checking. This
3713 is initially intended to pass the tests outlined in "Conformance Testing
3714 of Relying Party Client Certificate Path Processing Logic" v1.07.
3717 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3718 remained unused and not that useful. A variety of other little bignum
3719 tweaks and fixes have also been made continuing on from the audit (see
3723 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3724 associated ASN1, EVP and SSL functions and old ASN1 macros.
3727 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3728 and this should never fail. So the return value from the use of
3729 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3730 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3733 *) BN_CTX_get() should return zero-valued bignums, providing the same
3734 initialised value as BN_new().
3735 [Geoff Thorpe, suggested by Ulf Möller]
3737 *) Support for inhibitAnyPolicy certificate extension.
3740 *) An audit of the BIGNUM code is underway, for which debugging code is
3741 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3742 is considered valid when processing BIGNUMs, and causes execution to
3743 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3744 further steps are taken to deliberately pollute unused data in BIGNUM
3745 structures to try and expose faulty code further on. For now, openssl will
3746 (in its default mode of operation) continue to tolerate the inconsistent
3747 forms that it has tolerated in the past, but authors and packagers should
3748 consider trying openssl and their own applications when compiled with
3749 these debugging symbols defined. It will help highlight potential bugs in
3750 their own code, and will improve the test coverage for OpenSSL itself. At
3751 some point, these tighter rules will become openssl's default to improve
3752 maintainability, though the assert()s and other overheads will remain only
3753 in debugging configurations. See bn.h for more details.
3754 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3756 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3757 that can only be obtained through BN_CTX_new() (which implicitly
3758 initialises it). The presence of this function only made it possible
3759 to overwrite an existing structure (and cause memory leaks).
3762 *) Because of the callback-based approach for implementing LHASH as a
3763 template type, lh_insert() adds opaque objects to hash-tables and
3764 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3765 to clean up those corresponding objects before destroying the hash table
3766 (and losing the object pointers). So some over-zealous constifications in
3767 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3768 objects as "const" and the lh_doall[_arg] callback wrappers are not
3769 prototyped to have "const" restrictions on the object pointers they are
3770 given (and so aren't required to cast them away any more).
3773 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3774 (speed) prefers to use its own implementation. The two implementations
3775 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3776 its object type properly exposed (MS_TM) instead of casting to/from "char
3777 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3778 aren't necessarily the greatest nomenclatures - but this is what was used
3779 internally to the implementation so I've used that for now.
3782 *) Ensure that deprecated functions do not get compiled when
3783 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3784 the self-tests were still using deprecated key-generation functions so
3785 these have been updated also.
3788 *) Reorganise PKCS#7 code to separate the digest location functionality
3789 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3790 New function PKCS7_set_digest() to set the digest type for PKCS#7
3791 digestedData type. Add additional code to correctly generate the
3792 digestedData type and add support for this type in PKCS7 initialization
3796 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3797 structure of type "other".
3800 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3801 sure the loop does correctly stop and breaking ("division by zero")
3802 modulus operations are not performed. The (pre-generated) prime
3803 table crypto/bn/bn_prime.h was already correct, but it could not be
3804 re-generated on some platforms because of the "division by zero"
3805 situation in the script.
3806 [Ralf S. Engelschall]
3808 *) Update support for ECC-based TLS ciphersuites according to
3809 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3810 SHA-1 now is only used for "small" curves (where the
3811 representation of a field element takes up to 24 bytes); for
3812 larger curves, the field element resulting from ECDH is directly
3813 used as premaster secret.
3814 [Douglas Stebila (Sun Microsystems Laboratories)]
3816 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3817 curve secp160r1 to the tests.
3818 [Douglas Stebila (Sun Microsystems Laboratories)]
3820 *) Add the possibility to load symbols globally with DSO.
3821 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3823 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3824 control of the error stack.
3827 *) Add support for STORE in ENGINE.
3830 *) Add the STORE type. The intention is to provide a common interface
3831 to certificate and key stores, be they simple file-based stores, or
3832 HSM-type store, or LDAP stores, or...
3833 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3836 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3837 pass a list of arguments to any function as well as provide a way
3838 for a function to pass data back to the caller.
3841 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3842 works like BUF_strdup() but can be used to duplicate a portion of
3843 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3847 *) Add the function sk_find_ex() which works like sk_find(), but will
3848 return an index to an element even if an exact match couldn't be
3849 found. The index is guaranteed to point at the element where the
3850 searched-for key would be inserted to preserve sorting order.
3853 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3854 takes an extra flags argument for optional functionality. Currently,
3855 the following flags are defined:
3857 OBJ_BSEARCH_VALUE_ON_NOMATCH
3858 This one gets OBJ_bsearch_ex() to return a pointer to the first
3859 element where the comparing function returns a negative or zero
3862 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3863 This one gets OBJ_bsearch_ex() to return a pointer to the first
3864 element where the comparing function returns zero. This is useful
3865 if there are more than one element where the comparing function
3869 *) Make it possible to create self-signed certificates with 'openssl ca'
3870 in such a way that the self-signed certificate becomes part of the
3871 CA database and uses the same mechanisms for serial number generation
3872 as all other certificate signing. The new flag '-selfsign' enables
3873 this functionality. Adapt CA.sh and CA.pl.in.
3876 *) Add functionality to check the public key of a certificate request
3877 against a given private. This is useful to check that a certificate
3878 request can be signed by that key (self-signing).
3881 *) Make it possible to have multiple active certificates with the same
3882 subject in the CA index file. This is done only if the keyword
3883 'unique_subject' is set to 'no' in the main CA section (default
3884 if 'CA_default') of the configuration file. The value is saved
3885 with the database itself in a separate index attribute file,
3886 named like the index file with '.attr' appended to the name.
3889 *) Generate muti valued AVAs using '+' notation in config files for
3893 *) Support for nameConstraints certificate extension.
3896 *) Support for policyConstraints certificate extension.
3899 *) Support for policyMappings certificate extension.
3902 *) Make sure the default DSA_METHOD implementation only uses its
3903 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3904 and change its own handlers to be NULL so as to remove unnecessary
3905 indirection. This lets alternative implementations fallback to the
3906 default implementation more easily.
3909 *) Support for directoryName in GeneralName related extensions
3913 *) Make it possible to link applications using Makefile.shared.
3914 Make that possible even when linking against static libraries!
3917 *) Support for single pass processing for S/MIME signing. This now
3918 means that S/MIME signing can be done from a pipe, in addition
3919 cleartext signing (multipart/signed type) is effectively streaming
3920 and the signed data does not need to be all held in memory.
3922 This is done with a new flag PKCS7_STREAM. When this flag is set
3923 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3924 is done after the data is output (and digests calculated) in
3925 SMIME_write_PKCS7().
3928 *) Add full support for -rpath/-R, both in shared libraries and
3929 applications, at least on the platforms where it's known how
3933 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3934 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3935 will now compute a table of multiples of the generator that
3936 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3937 faster (notably in the case of a single point multiplication,
3938 scalar * generator).
3939 [Nils Larsch, Bodo Moeller]
3941 *) IPv6 support for certificate extensions. The various extensions
3942 which use the IP:a.b.c.d can now take IPv6 addresses using the
3943 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3947 *) Added an ENGINE that implements RSA by performing private key
3948 exponentiations with the GMP library. The conversions to and from
3949 GMP's mpz_t format aren't optimised nor are any montgomery forms
3950 cached, and on x86 it appears OpenSSL's own performance has caught up.
3951 However there are likely to be other architectures where GMP could
3952 provide a boost. This ENGINE is not built in by default, but it can be
3953 specified at Configure time and should be accompanied by the necessary
3954 linker additions, eg;
3955 ./config -DOPENSSL_USE_GMP -lgmp
3958 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3959 testing availability of engines with "-t" - the old behaviour is
3960 produced by increasing the feature's verbosity with "-tt".
3963 *) ECDSA routines: under certain error conditions uninitialized BN objects
3964 could be freed. Solution: make sure initialization is performed early
3965 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3969 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3970 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3971 software implementations. For DSA and DH, parameter generation can
3972 also be overriden by providing the appropriate method callbacks.
3975 *) Change the "progress" mechanism used in key-generation and
3976 primality testing to functions that take a new BN_GENCB pointer in
3977 place of callback/argument pairs. The new API functions have "_ex"
3978 postfixes and the older functions are reimplemented as wrappers for
3979 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3980 declarations of the old functions to help (graceful) attempts to
3981 migrate to the new functions. Also, the new key-generation API
3982 functions operate on a caller-supplied key-structure and return
3983 success/failure rather than returning a key or NULL - this is to
3984 help make "keygen" another member function of RSA_METHOD etc.
3986 Example for using the new callback interface:
3988 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3992 BN_GENCB_set(&my_cb, my_callback, my_arg);
3994 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3995 /* For the meaning of a, b in calls to my_callback(), see the
3996 * documentation of the function that calls the callback.
3997 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3998 * my_callback should return 1 if it wants BN_is_prime_ex()
3999 * to continue, or 0 to stop.
4004 *) Change the ZLIB compression method to be stateful, and make it
4005 available to TLS with the number defined in
4006 draft-ietf-tls-compression-04.txt.
4009 *) Add the ASN.1 structures and functions for CertificatePair, which
4010 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4012 CertificatePair ::= SEQUENCE {
4013 forward [0] Certificate OPTIONAL,
4014 reverse [1] Certificate OPTIONAL,
4015 -- at least one of the pair shall be present -- }
4017 Also implement the PEM functions to read and write certificate
4018 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4020 This needed to be defined, mostly for the sake of the LDAP
4021 attribute crossCertificatePair, but may prove useful elsewhere as
4025 *) Make it possible to inhibit symlinking of shared libraries in
4026 Makefile.shared, for Cygwin's sake.
4029 *) Extend the BIGNUM API by creating a function
4030 void BN_set_negative(BIGNUM *a, int neg);
4031 and a macro that behave like
4032 int BN_is_negative(const BIGNUM *a);
4034 to avoid the need to access 'a->neg' directly in applications.
4037 *) Implement fast modular reduction for pseudo-Mersenne primes
4038 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4039 EC_GROUP_new_curve_GFp() will now automatically use this
4041 [Nils Larsch <nla@trustcenter.de>]
4043 *) Add new lock type (CRYPTO_LOCK_BN).
4046 *) Change the ENGINE framework to automatically load engines
4047 dynamically from specific directories unless they could be
4048 found to already be built in or loaded. Move all the
4049 current engines except for the cryptodev one to a new
4051 The engines in engines/ are built as shared libraries if
4052 the "shared" options was given to ./Configure or ./config.
4053 Otherwise, they are inserted in libcrypto.a.
4054 /usr/local/ssl/engines is the default directory for dynamic
4055 engines, but that can be overriden at configure time through
4056 the usual use of --prefix and/or --openssldir, and at run
4057 time with the environment variable OPENSSL_ENGINES.
4058 [Geoff Thorpe and Richard Levitte]
4060 *) Add Makefile.shared, a helper makefile to build shared
4061 libraries. Adapt Makefile.org.
4064 *) Add version info to Win32 DLLs.
4065 [Peter 'Luna' Runestig" <peter@runestig.com>]
4067 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4068 can be added using this API to created arbitrary PKCS#12
4069 files while avoiding the low level API.
4071 New options to PKCS12_create(), key or cert can be NULL and
4072 will then be omitted from the output file. The encryption
4073 algorithm NIDs can be set to -1 for no encryption, the mac
4074 iteration count can be set to 0 to omit the mac.
4076 Enhance pkcs12 utility by making the -nokeys and -nocerts
4077 options work when creating a PKCS#12 file. New option -nomac
4078 to omit the mac, NONE can be set for an encryption algorithm.
4079 New code is modified to use the enhanced PKCS12_create()
4080 instead of the low level API.
4083 *) Extend ASN1 encoder to support indefinite length constructed
4084 encoding. This can output sequences tags and octet strings in
4085 this form. Modify pk7_asn1.c to support indefinite length
4086 encoding. This is experimental and needs additional code to
4087 be useful, such as an ASN1 bio and some enhanced streaming
4090 Extend template encode functionality so that tagging is passed
4091 down to the template encoder.
4094 *) Let 'openssl req' fail if an argument to '-newkey' is not
4095 recognized instead of using RSA as a default.
4098 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4099 As these are not official, they are not included in "ALL";
4100 the "ECCdraft" ciphersuite group alias can be used to select them.
4101 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4103 *) Add ECDH engine support.
4104 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4106 *) Add ECDH in new directory crypto/ecdh/.
4107 [Douglas Stebila (Sun Microsystems Laboratories)]
4109 *) Let BN_rand_range() abort with an error after 100 iterations
4110 without success (which indicates a broken PRNG).
4113 *) Change BN_mod_sqrt() so that it verifies that the input value
4114 is really the square of the return value. (Previously,
4115 BN_mod_sqrt would show GIGO behaviour.)
4118 *) Add named elliptic curves over binary fields from X9.62, SECG,
4119 and WAP/WTLS; add OIDs that were still missing.
4121 [Sheueling Chang Shantz and Douglas Stebila
4122 (Sun Microsystems Laboratories)]
4124 *) Extend the EC library for elliptic curves over binary fields
4125 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4128 EC_GF2m_simple_method
4132 EC_GROUP_new_curve_GF2m
4133 EC_GROUP_set_curve_GF2m
4134 EC_GROUP_get_curve_GF2m
4135 EC_POINT_set_affine_coordinates_GF2m
4136 EC_POINT_get_affine_coordinates_GF2m
4137 EC_POINT_set_compressed_coordinates_GF2m
4139 Point compression for binary fields is disabled by default for
4140 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4143 As binary polynomials are represented as BIGNUMs, various members
4144 of the EC_GROUP and EC_POINT data structures can be shared
4145 between the implementations for prime fields and binary fields;
4146 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4147 are essentially identical to their ..._GFp counterparts.
4148 (For simplicity, the '..._GFp' prefix has been dropped from
4149 various internal method names.)
4151 An internal 'field_div' method (similar to 'field_mul' and
4152 'field_sqr') has been added; this is used only for binary fields.
4154 [Sheueling Chang Shantz and Douglas Stebila
4155 (Sun Microsystems Laboratories)]
4157 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4158 through methods ('mul', 'precompute_mult').
4160 The generic implementations (now internally called 'ec_wNAF_mul'
4161 and 'ec_wNAF_precomputed_mult') remain the default if these
4162 methods are undefined.
4164 [Sheueling Chang Shantz and Douglas Stebila
4165 (Sun Microsystems Laboratories)]
4167 *) New function EC_GROUP_get_degree, which is defined through
4168 EC_METHOD. For curves over prime fields, this returns the bit
4169 length of the modulus.
4171 [Sheueling Chang Shantz and Douglas Stebila
4172 (Sun Microsystems Laboratories)]
4174 *) New functions EC_GROUP_dup, EC_POINT_dup.
4175 (These simply call ..._new and ..._copy).
4177 [Sheueling Chang Shantz and Douglas Stebila
4178 (Sun Microsystems Laboratories)]
4180 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4181 Polynomials are represented as BIGNUMs (where the sign bit is not
4182 used) in the following functions [macros]:
4185 BN_GF2m_sub [= BN_GF2m_add]
4186 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4187 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4188 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4190 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4191 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4192 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4193 BN_GF2m_cmp [= BN_ucmp]
4195 (Note that only the 'mod' functions are actually for fields GF(2^m).
4196 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4198 For some functions, an the irreducible polynomial defining a
4199 field can be given as an 'unsigned int[]' with strictly
4200 decreasing elements giving the indices of those bits that are set;
4201 i.e., p[] represents the polynomial
4202 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4204 p[0] > p[1] > ... > p[k] = 0.
4205 This applies to the following functions:
4210 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4211 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4213 BN_GF2m_mod_sqrt_arr
4214 BN_GF2m_mod_solve_quad_arr
4218 Conversion can be performed by the following functions:
4223 bntest.c has additional tests for binary polynomial arithmetic.
4225 Two implementations for BN_GF2m_mod_div() are available.
4226 The default algorithm simply uses BN_GF2m_mod_inv() and
4227 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4228 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4229 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4231 [Sheueling Chang Shantz and Douglas Stebila
4232 (Sun Microsystems Laboratories)]
4234 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4235 functionality is disabled at compile-time.
4236 [Douglas Stebila <douglas.stebila@sun.com>]
4238 *) Change default behaviour of 'openssl asn1parse' so that more
4239 information is visible when viewing, e.g., a certificate:
4241 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4242 mode the content of non-printable OCTET STRINGs is output in a
4243 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4244 avoid the appearance of a printable string.
4245 [Nils Larsch <nla@trustcenter.de>]
4247 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4249 EC_GROUP_set_asn1_flag()
4250 EC_GROUP_get_asn1_flag()
4251 EC_GROUP_set_point_conversion_form()
4252 EC_GROUP_get_point_conversion_form()
4253 These control ASN1 encoding details:
4254 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4255 has been set to OPENSSL_EC_NAMED_CURVE.
4256 - Points are encoded in uncompressed form by default; options for
4257 asn1_for are as for point2oct, namely
4258 POINT_CONVERSION_COMPRESSED
4259 POINT_CONVERSION_UNCOMPRESSED
4260 POINT_CONVERSION_HYBRID
4262 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4265 EC_GROUP_get0_seed()
4266 EC_GROUP_get_seed_len()
4267 This is used only for ASN1 purposes (so far).
4268 [Nils Larsch <nla@trustcenter.de>]
4270 *) Add 'field_type' member to EC_METHOD, which holds the NID
4271 of the appropriate field type OID. The new function
4272 EC_METHOD_get_field_type() returns this value.
4273 [Nils Larsch <nla@trustcenter.de>]
4278 EC_POINT_point2hex()
4279 EC_POINT_hex2point()
4280 providing useful interfaces to EC_POINT_point2oct() and
4281 EC_POINT_oct2point().
4282 [Nils Larsch <nla@trustcenter.de>]
4284 *) Change internals of the EC library so that the functions
4285 EC_GROUP_set_generator()
4286 EC_GROUP_get_generator()
4287 EC_GROUP_get_order()
4288 EC_GROUP_get_cofactor()
4289 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4290 to methods, which would lead to unnecessary code duplication when
4291 adding different types of curves.
4292 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4294 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4295 arithmetic, and such that modified wNAFs are generated
4296 (which avoid length expansion in many cases).
4299 *) Add a function EC_GROUP_check_discriminant() (defined via
4300 EC_METHOD) that verifies that the curve discriminant is non-zero.
4302 Add a function EC_GROUP_check() that makes some sanity tests
4303 on a EC_GROUP, its generator and order. This includes
4304 EC_GROUP_check_discriminant().
4305 [Nils Larsch <nla@trustcenter.de>]
4307 *) Add ECDSA in new directory crypto/ecdsa/.
4309 Add applications 'openssl ecparam' and 'openssl ecdsa'
4310 (these are based on 'openssl dsaparam' and 'openssl dsa').
4312 ECDSA support is also included in various other files across the
4313 library. Most notably,
4314 - 'openssl req' now has a '-newkey ecdsa:file' option;
4315 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4316 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4317 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4318 them suitable for ECDSA where domain parameters must be
4319 extracted before the specific public key;
4320 - ECDSA engine support has been added.
4321 [Nils Larsch <nla@trustcenter.de>]
4323 *) Include some named elliptic curves, and add OIDs from X9.62,
4324 SECG, and WAP/WTLS. Each curve can be obtained from the new
4326 EC_GROUP_new_by_curve_name(),
4327 and the list of available named curves can be obtained with
4328 EC_get_builtin_curves().
4329 Also add a 'curve_name' member to EC_GROUP objects, which can be
4331 EC_GROUP_set_curve_name()
4332 EC_GROUP_get_curve_name()
4333 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4335 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4336 was actually never needed) and in BN_mul(). The removal in BN_mul()
4337 required a small change in bn_mul_part_recursive() and the addition
4338 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4339 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4340 bn_sub_words() and bn_add_words() except they take arrays with
4344 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4346 *) Cleanse PEM buffers before freeing them since they may contain
4348 [Benjamin Bennett <ben@psc.edu>]
4350 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4351 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4352 authentication-only ciphersuites.
4355 *) Since AES128 and AES256 share a single mask bit in the logic of
4356 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4357 kludge to work properly if AES128 is available and AES256 isn't.
4360 *) Expand security boundary to match 1.1.1 module.
4363 *) Remove redundant features: hash file source, editing of test vectors
4364 modify fipsld to use external fips_premain.c signature.
4367 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4368 run algorithm test programs.
4371 *) Make algorithm test programs more tolerant of whitespace.
4374 *) Have SSL/TLS server implementation tolerate "mismatched" record
4375 protocol version while receiving ClientHello even if the
4376 ClientHello is fragmented. (The server can't insist on the
4377 particular protocol version it has chosen before the ServerHello
4378 message has informed the client about his choice.)
4381 *) Load error codes if they are not already present instead of using a
4382 static variable. This allows them to be cleanly unloaded and reloaded.
4385 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4387 *) Introduce limits to prevent malicious keys being able to
4388 cause a denial of service. (CVE-2006-2940)
4389 [Steve Henson, Bodo Moeller]
4391 *) Fix ASN.1 parsing of certain invalid structures that can result
4392 in a denial of service. (CVE-2006-2937) [Steve Henson]
4394 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4395 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4397 *) Fix SSL client code which could crash if connecting to a
4398 malicious SSLv2 server. (CVE-2006-4343)
4399 [Tavis Ormandy and Will Drewry, Google Security Team]
4401 *) Change ciphersuite string processing so that an explicit
4402 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4403 will no longer include "AES128-SHA"), and any other similar
4404 ciphersuite (same bitmap) from *other* protocol versions (so that
4405 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4406 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4407 changes from 0.9.8b and 0.9.8d.
4410 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4412 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4413 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4415 *) Change the Unix randomness entropy gathering to use poll() when
4416 possible instead of select(), since the latter has some
4417 undesirable limitations.
4418 [Darryl Miles via Richard Levitte and Bodo Moeller]
4420 *) Disable rogue ciphersuites:
4422 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4423 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4424 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4426 The latter two were purportedly from
4427 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4430 Also deactive the remaining ciphersuites from
4431 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4432 unofficial, and the ID has long expired.
4435 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4436 dual-core machines) and other potential thread-safety issues.
4439 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4441 *) Adapt fipsld and the build system to link against the validated FIPS
4442 module in FIPS mode.
4445 *) Fixes for VC++ 2005 build under Windows.
4448 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4449 from a Windows bash shell such as MSYS. It is autodetected from the
4450 "config" script when run from a VC++ environment. Modify standard VC++
4451 build to use fipscanister.o from the GNU make build.
4454 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4456 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4457 The value now differs depending on if you build for FIPS or not.
4458 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4459 safely run with a non-FIPSed libcrypto, as it may crash because of
4460 the difference induced by this change.
4463 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4465 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4466 (part of SSL_OP_ALL). This option used to disable the
4467 countermeasure against man-in-the-middle protocol-version
4468 rollback in the SSL 2.0 server implementation, which is a bad
4469 idea. (CVE-2005-2969)
4471 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4472 for Information Security, National Institute of Advanced Industrial
4473 Science and Technology [AIST], Japan)]
4475 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4476 mainly for FIPS compliance and not fully integrated at this stage.
4479 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4480 the exponentiation using a fixed-length exponent. (Otherwise,
4481 the information leaked through timing could expose the secret key
4482 after many signatures; cf. Bleichenbacher's attack on DSA with
4486 *) Make a new fixed-window mod_exp implementation the default for
4487 RSA, DSA, and DH private-key operations so that the sequence of
4488 squares and multiplies and the memory access pattern are
4489 independent of the particular secret key. This will mitigate
4490 cache-timing and potential related attacks.
4492 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4493 and this is automatically used by BN_mod_exp_mont() if the new flag
4494 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4495 will use this BN flag for private exponents unless the flag
4496 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4497 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4499 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4501 *) Change the client implementation for SSLv23_method() and
4502 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4503 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4504 (Previously, the SSL 2.0 backwards compatible Client Hello
4505 message format would be used even with SSL_OP_NO_SSLv2.)
4508 *) Add support for smime-type MIME parameter in S/MIME messages which some
4512 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4513 a threadsafe manner. Modify rsa code to use new function and add calls
4514 to dsa and dh code (which had race conditions before).
4517 *) Include the fixed error library code in the C error file definitions
4518 instead of fixing them up at runtime. This keeps the error code
4519 structures constant.
4522 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4524 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4527 *) Fixes for newer kerberos headers. NB: the casts are needed because
4528 the 'length' field is signed on one version and unsigned on another
4529 with no (?) obvious way to tell the difference, without these VC++
4530 complains. Also the "definition" of FAR (blank) is no longer included
4531 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4532 some needed definitions.
4535 *) Undo Cygwin change.
4538 *) Added support for proxy certificates according to RFC 3820.
4539 Because they may be a security thread to unaware applications,
4540 they must be explicitely allowed in run-time. See
4541 docs/HOWTO/proxy_certificates.txt for further information.
4544 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4546 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4547 server and client random values. Previously
4548 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4549 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4551 This change has negligible security impact because:
4553 1. Server and client random values still have 24 bytes of pseudo random
4556 2. Server and client random values are sent in the clear in the initial
4559 3. The master secret is derived using the premaster secret (48 bytes in
4560 size for static RSA ciphersuites) as well as client server and random
4563 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4566 [Stephen Henson, reported by UK NISCC]
4568 *) Use Windows randomness collection on Cygwin.
4571 *) Fix hang in EGD/PRNGD query when communication socket is closed
4572 prematurely by EGD/PRNGD.
4573 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4575 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4578 *) Back-port of selected performance improvements from development
4579 branch, as well as improved support for PowerPC platforms.
4582 *) Add lots of checks for memory allocation failure, error codes to indicate
4583 failure and freeing up memory if a failure occurs.
4584 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4586 *) Add new -passin argument to dgst.
4589 *) Perform some character comparisons of different types in X509_NAME_cmp:
4590 this is needed for some certificates that reencode DNs into UTF8Strings
4591 (in violation of RFC3280) and can't or wont issue name rollover
4595 *) Make an explicit check during certificate validation to see that
4596 the CA setting in each certificate on the chain is correct. As a
4597 side effect always do the following basic checks on extensions,
4598 not just when there's an associated purpose to the check:
4600 - if there is an unhandled critical extension (unless the user
4601 has chosen to ignore this fault)
4602 - if the path length has been exceeded (if one is set at all)
4603 - that certain extensions fit the associated purpose (if one has
4607 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4609 *) Avoid a race condition when CRLs are checked in a multi threaded
4610 environment. This would happen due to the reordering of the revoked
4611 entries during signature checking and serial number lookup. Now the
4612 encoding is cached and the serial number sort performed under a lock.
4613 Add new STACK function sk_is_sorted().
4616 *) Add Delta CRL to the extension code.
4619 *) Various fixes to s3_pkt.c so alerts are sent properly.
4620 [David Holmes <d.holmes@f5.com>]
4622 *) Reduce the chances of duplicate issuer name and serial numbers (in
4623 violation of RFC3280) using the OpenSSL certificate creation utilities.
4624 This is done by creating a random 64 bit value for the initial serial
4625 number when a serial number file is created or when a self signed
4626 certificate is created using 'openssl req -x509'. The initial serial
4627 number file is created using 'openssl x509 -next_serial' in CA.pl
4628 rather than being initialized to 1.
4631 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4633 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4634 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4635 [Joe Orton, Steve Henson]
4637 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4639 [Joe Orton, Steve Henson]
4641 *) Make it possible to have multiple active certificates with the same
4642 subject in the CA index file. This is done only if the keyword
4643 'unique_subject' is set to 'no' in the main CA section (default
4644 if 'CA_default') of the configuration file. The value is saved
4645 with the database itself in a separate index attribute file,
4646 named like the index file with '.attr' appended to the name.
4649 *) X509 verify fixes. Disable broken certificate workarounds when
4650 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4651 keyUsage extension present. Don't accept CRLs with unhandled critical
4652 extensions: since verify currently doesn't process CRL extensions this
4653 rejects a CRL with *any* critical extensions. Add new verify error codes
4657 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4658 A clarification of RFC2560 will require the use of OCTET STRINGs and
4659 some implementations cannot handle the current raw format. Since OpenSSL
4660 copies and compares OCSP nonces as opaque blobs without any attempt at
4661 parsing them this should not create any compatibility issues.
4664 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4665 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4666 this HMAC (and other) operations are several times slower than OpenSSL
4670 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4671 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4673 *) Use the correct content when signing type "other".
4676 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4678 *) Fix various bugs revealed by running the NISCC test suite:
4680 Stop out of bounds reads in the ASN1 code when presented with
4681 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4683 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4685 If verify callback ignores invalid public key errors don't try to check
4686 certificate signature with the NULL public key.
4690 *) New -ignore_err option in ocsp application to stop the server
4691 exiting on the first error in a request.
4694 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4695 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4699 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4700 extra data after the compression methods not only for TLS 1.0
4701 but also for SSL 3.0 (as required by the specification).
4702 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4704 *) Change X509_certificate_type() to mark the key as exported/exportable
4705 when it's 512 *bits* long, not 512 bytes.
4708 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4709 blocks during encryption.
4712 *) Various fixes to base64 BIO and non blocking I/O. On write
4713 flushes were not handled properly if the BIO retried. On read
4714 data was not being buffered properly and had various logic bugs.
4715 This also affects blocking I/O when the data being decoded is a
4719 *) Various S/MIME bugfixes and compatibility changes:
4720 output correct application/pkcs7 MIME type if
4721 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4722 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4723 of files as .eml work). Correctly handle very long lines in MIME
4727 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4729 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4730 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4731 a protocol version number mismatch like a decryption error
4732 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4735 *) Turn on RSA blinding by default in the default implementation
4736 to avoid a timing attack. Applications that don't want it can call
4737 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4738 They would be ill-advised to do so in most cases.
4739 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4741 *) Change RSA blinding code so that it works when the PRNG is not
4742 seeded (in this case, the secret RSA exponent is abused as
4743 an unpredictable seed -- if it is not unpredictable, there
4744 is no point in blinding anyway). Make RSA blinding thread-safe
4745 by remembering the creator's thread ID in rsa->blinding and
4746 having all other threads use local one-time blinding factors
4747 (this requires more computation than sharing rsa->blinding, but
4748 avoids excessive locking; and if an RSA object is not shared
4749 between threads, blinding will still be very fast).
4752 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4753 ENGINE as defaults for all supported algorithms irrespective of
4754 the 'flags' parameter. 'flags' is now honoured, so applications
4755 should make sure they are passing it correctly.
4758 *) Target "mingw" now allows native Windows code to be generated in
4759 the Cygwin environment as well as with the MinGW compiler.
4762 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4764 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4765 via timing by performing a MAC computation even if incorrect
4766 block cipher padding has been found. This is a countermeasure
4767 against active attacks where the attacker has to distinguish
4768 between bad padding and a MAC verification error. (CVE-2003-0078)
4770 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4771 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4772 Martin Vuagnoux (EPFL, Ilion)]
4774 *) Make the no-err option work as intended. The intention with no-err
4775 is not to have the whole error stack handling routines removed from
4776 libcrypto, it's only intended to remove all the function name and
4777 reason texts, thereby removing some of the footprint that may not
4778 be interesting if those errors aren't displayed anyway.
4780 NOTE: it's still possible for any application or module to have it's
4781 own set of error texts inserted. The routines are there, just not
4782 used by default when no-err is given.
4785 *) Add support for FreeBSD on IA64.
4786 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4788 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4789 Kerberos function mit_des_cbc_cksum(). Before this change,
4790 the value returned by DES_cbc_cksum() was like the one from
4791 mit_des_cbc_cksum(), except the bytes were swapped.
4792 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4794 *) Allow an application to disable the automatic SSL chain building.
4795 Before this a rather primitive chain build was always performed in
4796 ssl3_output_cert_chain(): an application had no way to send the
4797 correct chain if the automatic operation produced an incorrect result.
4799 Now the chain builder is disabled if either:
4801 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4803 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4805 The reasoning behind this is that an application would not want the
4806 auto chain building to take place if extra chain certificates are
4807 present and it might also want a means of sending no additional
4808 certificates (for example the chain has two certificates and the
4812 *) Add the possibility to build without the ENGINE framework.
4813 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4815 *) Under Win32 gmtime() can return NULL: check return value in
4816 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4819 *) DSA routines: under certain error conditions uninitialized BN objects
4820 could be freed. Solution: make sure initialization is performed early
4821 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4822 Nils Larsch <nla@trustcenter.de> via PR#459)
4825 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4826 checked on reconnect on the client side, therefore session resumption
4827 could still fail with a "ssl session id is different" error. This
4828 behaviour is masked when SSL_OP_ALL is used due to
4829 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4830 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4831 followup to PR #377.
4834 *) IA-32 assembler support enhancements: unified ELF targets, support
4835 for SCO/Caldera platforms, fix for Cygwin shared build.
4838 *) Add support for FreeBSD on sparc64. As a consequence, support for
4839 FreeBSD on non-x86 processors is separate from x86 processors on
4840 the config script, much like the NetBSD support.
4841 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4843 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4845 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4848 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4849 code (06) was taken as the first octet of the session ID and the last
4850 octet was ignored consequently. As a result SSLv2 client side session
4851 caching could not have worked due to the session ID mismatch between
4853 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4857 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4858 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4862 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4863 seems that in spite of existing for more than a year, many application
4864 author have done nothing to provide the necessary callbacks, which
4865 means that this particular engine will not work properly anywhere.
4866 This is a very unfortunate situation which forces us, in the name
4867 of usability, to give the hw_ncipher.c a static lock, which is part
4869 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4870 appear in 0.9.8 or later. We EXPECT application authors to have
4871 dealt properly with this when 0.9.8 is released (unless we actually
4872 make such changes in the libcrypto locking code that changes will
4873 have to be made anyway).
4876 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4877 octets have been read, EOF or an error occurs. Without this change
4878 some truncated ASN1 structures will not produce an error.
4881 *) Disable Heimdal support, since it hasn't been fully implemented.
4882 Still give the possibility to force the use of Heimdal, but with
4883 warnings and a request that patches get sent to openssl-dev.
4886 *) Add the VC-CE target, introduce the WINCE sysname, and add
4887 INSTALL.WCE and appropriate conditionals to make it build.
4888 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4890 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4891 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4892 edit numbers of the version.
4893 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4895 *) Introduce safe string copy and catenation functions
4896 (BUF_strlcpy() and BUF_strlcat()).
4897 [Ben Laurie (CHATS) and Richard Levitte]
4899 *) Avoid using fixed-size buffers for one-line DNs.
4900 [Ben Laurie (CHATS)]
4902 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4903 resizing buffers containing secrets, and use where appropriate.
4904 [Ben Laurie (CHATS)]
4906 *) Avoid using fixed size buffers for configuration file location.
4907 [Ben Laurie (CHATS)]
4909 *) Avoid filename truncation for various CA files.
4910 [Ben Laurie (CHATS)]
4912 *) Use sizeof in preference to magic numbers.
4913 [Ben Laurie (CHATS)]
4915 *) Avoid filename truncation in cert requests.
4916 [Ben Laurie (CHATS)]
4918 *) Add assertions to check for (supposedly impossible) buffer
4920 [Ben Laurie (CHATS)]
4922 *) Don't cache truncated DNS entries in the local cache (this could
4923 potentially lead to a spoofing attack).
4924 [Ben Laurie (CHATS)]
4926 *) Fix various buffers to be large enough for hex/decimal
4927 representations in a platform independent manner.
4928 [Ben Laurie (CHATS)]
4930 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4931 resizing buffers containing secrets, and use where appropriate.
4932 [Ben Laurie (CHATS)]
4934 *) Add BIO_indent() to avoid much slightly worrying code to do
4936 [Ben Laurie (CHATS)]
4938 *) Convert sprintf()/BIO_puts() to BIO_printf().
4939 [Ben Laurie (CHATS)]
4941 *) buffer_gets() could terminate with the buffer only half
4943 [Ben Laurie (CHATS)]
4945 *) Add assertions to prevent user-supplied crypto functions from
4946 overflowing internal buffers by having large block sizes, etc.
4947 [Ben Laurie (CHATS)]
4949 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4951 [Ben Laurie (CHATS)]
4953 *) Eliminate unused copy of key in RC4.
4954 [Ben Laurie (CHATS)]
4956 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4957 [Ben Laurie (CHATS)]
4959 *) Fix off-by-one error in EGD path.
4960 [Ben Laurie (CHATS)]
4962 *) If RANDFILE path is too long, ignore instead of truncating.
4963 [Ben Laurie (CHATS)]
4965 *) Eliminate unused and incorrectly sized X.509 structure
4967 [Ben Laurie (CHATS)]
4969 *) Eliminate unused and dangerous function knumber().
4970 [Ben Laurie (CHATS)]
4972 *) Eliminate unused and dangerous structure, KSSL_ERR.
4973 [Ben Laurie (CHATS)]
4975 *) Protect against overlong session ID context length in an encoded
4976 session object. Since these are local, this does not appear to be
4978 [Ben Laurie (CHATS)]
4980 *) Change from security patch (see 0.9.6e below) that did not affect
4981 the 0.9.6 release series:
4983 Remote buffer overflow in SSL3 protocol - an attacker could
4984 supply an oversized master key in Kerberos-enabled versions.
4986 [Ben Laurie (CHATS)]
4988 *) Change the SSL kerb5 codes to match RFC 2712.
4991 *) Make -nameopt work fully for req and add -reqopt switch.
4992 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4994 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4995 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4997 *) Make sure tests can be performed even if the corresponding algorithms
4998 have been removed entirely. This was also the last step to make
4999 OpenSSL compilable with DJGPP under all reasonable conditions.
5000 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5002 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5003 to allow version independent disabling of normally unselected ciphers,
5004 which may be activated as a side-effect of selecting a single cipher.
5006 (E.g., cipher list string "RSA" enables ciphersuites that are left
5007 out of "ALL" because they do not provide symmetric encryption.
5008 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5009 [Lutz Jaenicke, Bodo Moeller]
5011 *) Add appropriate support for separate platform-dependent build
5012 directories. The recommended way to make a platform-dependent
5013 build directory is the following (tested on Linux), maybe with
5016 # Place yourself outside of the OpenSSL source tree. In
5017 # this example, the environment variable OPENSSL_SOURCE
5018 # is assumed to contain the absolute OpenSSL source directory.
5019 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5020 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5021 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5022 mkdir -p `dirname $F`
5023 ln -s $OPENSSL_SOURCE/$F $F
5026 To be absolutely sure not to disturb the source tree, a "make clean"
5027 is a good thing. If it isn't successfull, don't worry about it,
5028 it probably means the source directory is very clean.
5031 *) Make sure any ENGINE control commands make local copies of string
5032 pointers passed to them whenever necessary. Otherwise it is possible
5033 the caller may have overwritten (or deallocated) the original string
5034 data when a later ENGINE operation tries to use the stored values.
5035 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5037 *) Improve diagnostics in file reading and command-line digests.
5038 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5040 *) Add AES modes CFB and OFB to the object database. Correct an
5041 error in AES-CFB decryption.
5044 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5045 allows existing EVP_CIPHER_CTX structures to be reused after
5046 calling EVP_*Final(). This behaviour is used by encryption
5047 BIOs and some applications. This has the side effect that
5048 applications must explicitly clean up cipher contexts with
5049 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5052 *) Check the values of dna and dnb in bn_mul_recursive before calling
5053 bn_mul_comba (a non zero value means the a or b arrays do not contain
5054 n2 elements) and fallback to bn_mul_normal if either is not zero.
5057 *) Fix escaping of non-ASCII characters when using the -subj option
5058 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5061 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5062 form for "surname", serialNumber has no short form.
5063 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5064 therefore remove "mail" short name for "internet 7".
5065 The OID for unique identifiers in X509 certificates is
5066 x500UniqueIdentifier, not uniqueIdentifier.
5067 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5070 *) Add an "init" command to the ENGINE config module and auto initialize
5071 ENGINEs. Without any "init" command the ENGINE will be initialized
5072 after all ctrl commands have been executed on it. If init=1 the
5073 ENGINE is initailized at that point (ctrls before that point are run
5074 on the uninitialized ENGINE and after on the initialized one). If
5075 init=0 then the ENGINE will not be iniatialized at all.
5078 *) Fix the 'app_verify_callback' interface so that the user-defined
5079 argument is actually passed to the callback: In the
5080 SSL_CTX_set_cert_verify_callback() prototype, the callback
5081 declaration has been changed from
5084 int (*cb)(X509_STORE_CTX *,void *);
5085 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5086 i=s->ctx->app_verify_callback(&ctx)
5087 has been changed into
5088 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5090 To update applications using SSL_CTX_set_cert_verify_callback(),
5091 a dummy argument can be added to their callback functions.
5092 [D. K. Smetters <smetters@parc.xerox.com>]
5094 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5095 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5097 *) Add and OPENSSL_LOAD_CONF define which will cause
5098 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5099 This allows older applications to transparently support certain
5100 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5101 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5102 load the config file and OPENSSL_add_all_algorithms_conf() which will
5103 always load it have also been added.
5106 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5107 Adjust NIDs and EVP layer.
5108 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5110 *) Config modules support in openssl utility.
5112 Most commands now load modules from the config file,
5113 though in a few (such as version) this isn't done
5114 because it couldn't be used for anything.
5116 In the case of ca and req the config file used is
5117 the same as the utility itself: that is the -config
5118 command line option can be used to specify an
5122 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5123 use "openssl_conf" if filename is NULL use default openssl config file.
5126 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5127 config section name. Add a new flag to tolerate a missing config file
5128 and move code to CONF_modules_load_file().
5131 *) Support for crypto accelerator cards from Accelerated Encryption
5132 Processing, www.aep.ie. (Use engine 'aep')
5133 The support was copied from 0.9.6c [engine] and adapted/corrected
5134 to work with the new engine framework.
5135 [AEP Inc. and Richard Levitte]
5137 *) Support for SureWare crypto accelerator cards from Baltimore
5138 Technologies. (Use engine 'sureware')
5139 The support was copied from 0.9.6c [engine] and adapted
5140 to work with the new engine framework.
5143 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5144 make the newer ENGINE framework commands for the CHIL engine work.
5145 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5147 *) Make it possible to produce shared libraries on ReliantUNIX.
5148 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5150 *) Add the configuration target debug-linux-ppro.
5151 Make 'openssl rsa' use the general key loading routines
5152 implemented in apps.c, and make those routines able to
5153 handle the key format FORMAT_NETSCAPE and the variant
5155 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5157 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5158 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5160 *) Add -keyform to rsautl, and document -engine.
5161 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5163 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5164 BIO_R_NO_SUCH_FILE error code rather than the generic
5165 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5168 *) Add new functions
5170 ERR_peek_last_error_line
5171 ERR_peek_last_error_line_data.
5172 These are similar to
5175 ERR_peek_error_line_data,
5176 but report on the latest error recorded rather than the first one
5177 still in the error queue.
5178 [Ben Laurie, Bodo Moeller]
5180 *) default_algorithms option in ENGINE config module. This allows things
5182 default_algorithms = ALL
5183 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5186 *) Preliminary ENGINE config module.
5189 *) New experimental application configuration code.
5192 *) Change the AES code to follow the same name structure as all other
5193 symmetric ciphers, and behave the same way. Move everything to
5194 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5195 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5197 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5198 [Ben Laurie and Theo de Raadt]
5200 *) Add option to output public keys in req command.
5201 [Massimiliano Pala madwolf@openca.org]
5203 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5204 (up to about 10% better than before for P-192 and P-224).
5207 *) New functions/macros
5209 SSL_CTX_set_msg_callback(ctx, cb)
5210 SSL_CTX_set_msg_callback_arg(ctx, arg)
5211 SSL_set_msg_callback(ssl, cb)
5212 SSL_set_msg_callback_arg(ssl, arg)
5214 to request calling a callback function
5216 void cb(int write_p, int version, int content_type,
5217 const void *buf, size_t len, SSL *ssl, void *arg)
5219 whenever a protocol message has been completely received
5220 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5221 protocol version according to which the SSL library interprets
5222 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5223 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5224 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5225 specification (change_cipher_spec(20), alert(21), handshake(22)).
5226 'buf' and 'len' point to the actual message, 'ssl' to the
5227 SSL object, and 'arg' is the application-defined value set by
5228 SSL[_CTX]_set_msg_callback_arg().
5230 'openssl s_client' and 'openssl s_server' have new '-msg' options
5231 to enable a callback that displays all protocol messages.
5234 *) Change the shared library support so shared libraries are built as
5235 soon as the corresponding static library is finished, and thereby get
5236 openssl and the test programs linked against the shared library.
5237 This still only happens when the keyword "shard" has been given to
5238 the configuration scripts.
5240 NOTE: shared library support is still an experimental thing, and
5241 backward binary compatibility is still not guaranteed.
5242 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5244 *) Add support for Subject Information Access extension.
5245 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5247 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5248 additional bytes when new memory had to be allocated, not just
5249 when reusing an existing buffer.
5252 *) New command line and configuration option 'utf8' for the req command.
5253 This allows field values to be specified as UTF8 strings.
5256 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5257 runs for the former and machine-readable output for the latter.
5260 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5261 of the e-mail address in the DN (i.e., it will go into a certificate
5262 extension only). The new configuration file option 'email_in_dn = no'
5263 has the same effect.
5264 [Massimiliano Pala madwolf@openca.org]
5266 *) Change all functions with names starting with des_ to be starting
5267 with DES_ instead. Add wrappers that are compatible with libdes,
5268 but are named _ossl_old_des_*. Finally, add macros that map the
5269 des_* symbols to the corresponding _ossl_old_des_* if libdes
5270 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5271 desired, the des_* symbols will be mapped to DES_*, with one
5274 Since we provide two compatibility mappings, the user needs to
5275 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5276 compatibility is desired. The default (i.e., when that macro
5277 isn't defined) is OpenSSL 0.9.6c compatibility.
5279 There are also macros that enable and disable the support of old
5280 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5281 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5282 are defined, the default will apply: to support the old des routines.
5284 In either case, one must include openssl/des.h to get the correct
5285 definitions. Do not try to just include openssl/des_old.h, that
5288 NOTE: This is a major break of an old API into a new one. Software
5289 authors are encouraged to switch to the DES_ style functions. Some
5290 time in the future, des_old.h and the libdes compatibility functions
5291 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5292 default), and then completely removed.
5295 *) Test for certificates which contain unsupported critical extensions.
5296 If such a certificate is found during a verify operation it is
5297 rejected by default: this behaviour can be overridden by either
5298 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5299 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5300 X509_supported_extension() has also been added which returns 1 if a
5301 particular extension is supported.
5304 *) Modify the behaviour of EVP cipher functions in similar way to digests
5305 to retain compatibility with existing code.
5308 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5309 compatibility with existing code. In particular the 'ctx' parameter does
5310 not have to be to be initialized before the call to EVP_DigestInit() and
5311 it is tidied up after a call to EVP_DigestFinal(). New function
5312 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5313 EVP_MD_CTX_copy() changed to not require the destination to be
5314 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5315 requires the destination to be valid.
5317 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5318 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5321 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5322 so that complete 'Handshake' protocol structures are kept in memory
5323 instead of overwriting 'msg_type' and 'length' with 'body' data.
5326 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5327 [Massimo Santin via Richard Levitte]
5329 *) Major restructuring to the underlying ENGINE code. This includes
5330 reduction of linker bloat, separation of pure "ENGINE" manipulation
5331 (initialisation, etc) from functionality dealing with implementations
5332 of specific crypto iterfaces. This change also introduces integrated
5333 support for symmetric ciphers and digest implementations - so ENGINEs
5334 can now accelerate these by providing EVP_CIPHER and EVP_MD
5335 implementations of their own. This is detailed in crypto/engine/README
5336 as it couldn't be adequately described here. However, there are a few
5337 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5338 were changed in the original introduction of ENGINE code have now
5339 reverted back - the hooking from this code to ENGINE is now a good
5340 deal more passive and at run-time, operations deal directly with
5341 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5342 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5343 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5344 they were not being used by the framework as there is no concept of a
5345 BIGNUM_METHOD and they could not be generalised to the new
5346 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5347 ENGINE_cpy() has been removed as it cannot be consistently defined in
5351 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5354 *) Change mkdef.pl to sort symbols that get the same entry number,
5355 and make sure the automatically generated functions ERR_load_*
5356 become part of libeay.num as well.
5359 *) New function SSL_renegotiate_pending(). This returns true once
5360 renegotiation has been requested (either SSL_renegotiate() call
5361 or HelloRequest/ClientHello received from the peer) and becomes
5362 false once a handshake has been completed.
5363 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5364 sends a HelloRequest, but does not ensure that a handshake takes
5365 place. SSL_renegotiate_pending() is useful for checking if the
5366 client has followed the request.)
5369 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5370 By default, clients may request session resumption even during
5371 renegotiation (if session ID contexts permit); with this option,
5372 session resumption is possible only in the first handshake.
5374 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5375 more bits available for options that should not be part of
5376 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5379 *) Add some demos for certificate and certificate request creation.
5382 *) Make maximum certificate chain size accepted from the peer application
5383 settable (SSL*_get/set_max_cert_list()), as proposed by
5384 "Douglas E. Engert" <deengert@anl.gov>.
5387 *) Add support for shared libraries for Unixware-7
5388 (Boyd Lynn Gerber <gerberb@zenez.com>).
5391 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5392 be done prior to destruction. Use this to unload error strings from
5393 ENGINEs that load their own error strings. NB: This adds two new API
5394 functions to "get" and "set" this destroy handler in an ENGINE.
5397 *) Alter all existing ENGINE implementations (except "openssl" and
5398 "openbsd") to dynamically instantiate their own error strings. This
5399 makes them more flexible to be built both as statically-linked ENGINEs
5400 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5401 Also, add stub code to each that makes building them as self-contained
5402 shared-libraries easier (see README.ENGINE).
5405 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5406 implementations into applications that are completely implemented in
5407 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5408 commands that can be used to configure what shared-library to load and
5409 to control aspects of the way it is handled. Also, made an update to
5410 the README.ENGINE file that brings its information up-to-date and
5411 provides some information and instructions on the "dynamic" ENGINE
5412 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5415 *) Make it possible to unload ranges of ERR strings with a new
5416 "ERR_unload_strings" function.
5419 *) Add a copy() function to EVP_MD.
5422 *) Make EVP_MD routines take a context pointer instead of just the
5423 md_data void pointer.
5426 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5427 that the digest can only process a single chunk of data
5428 (typically because it is provided by a piece of
5429 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5430 is only going to provide a single chunk of data, and hence the
5431 framework needn't accumulate the data for oneshot drivers.
5434 *) As with "ERR", make it possible to replace the underlying "ex_data"
5435 functions. This change also alters the storage and management of global
5436 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5437 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5438 index counters. The API functions that use this state have been changed
5439 to take a "class_index" rather than pointers to the class's local STACK
5440 and counter, and there is now an API function to dynamically create new
5441 classes. This centralisation allows us to (a) plug a lot of the
5442 thread-safety problems that existed, and (b) makes it possible to clean
5443 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5444 such data would previously have always leaked in application code and
5445 workarounds were in place to make the memory debugging turn a blind eye
5446 to it. Application code that doesn't use this new function will still
5447 leak as before, but their memory debugging output will announce it now
5448 rather than letting it slide.
5450 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5451 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5452 has a return value to indicate success or failure.
5455 *) Make it possible to replace the underlying "ERR" functions such that the
5456 global state (2 LHASH tables and 2 locks) is only used by the "default"
5457 implementation. This change also adds two functions to "get" and "set"
5458 the implementation prior to it being automatically set the first time
5459 any other ERR function takes place. Ie. an application can call "get",
5460 pass the return value to a module it has just loaded, and that module
5461 can call its own "set" function using that value. This means the
5462 module's "ERR" operations will use (and modify) the error state in the
5463 application and not in its own statically linked copy of OpenSSL code.
5466 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5467 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5468 the operation, and provides a more encapsulated way for external code
5469 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5470 to use these functions rather than manually incrementing the counts.
5472 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5475 *) Add EVP test program.
5478 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5481 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5482 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5483 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5484 These allow a CRL to be built without having to access X509_CRL fields
5485 directly. Modify 'ca' application to use new functions.
5488 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5489 bug workarounds. Rollback attack detection is a security feature.
5490 The problem will only arise on OpenSSL servers when TLSv1 is not
5491 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5492 Software authors not wanting to support TLSv1 will have special reasons
5493 for their choice and can explicitly enable this option.
5494 [Bodo Moeller, Lutz Jaenicke]
5496 *) Rationalise EVP so it can be extended: don't include a union of
5497 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5498 (similar to those existing for EVP_CIPHER_CTX).
5503 EVP_MD_CTX_init(&md); /* new function call */
5504 EVP_DigestInit(&md, EVP_sha1());
5505 EVP_DigestUpdate(&md, in, len);
5506 EVP_DigestFinal(&md, out, NULL);
5507 EVP_MD_CTX_cleanup(&md); /* new function call */
5511 *) Make DES key schedule conform to the usual scheme, as well as
5512 correcting its structure. This means that calls to DES functions
5513 now have to pass a pointer to a des_key_schedule instead of a
5514 plain des_key_schedule (which was actually always a pointer
5517 des_key_schedule ks;
5519 des_set_key_checked(..., &ks);
5520 des_ncbc_encrypt(..., &ks, ...);
5522 (Note that a later change renames 'des_...' into 'DES_...'.)
5525 *) Initial reduction of linker bloat: the use of some functions, such as
5526 PEM causes large amounts of unused functions to be linked in due to
5527 poor organisation. For example pem_all.c contains every PEM function
5528 which has a knock on effect of linking in large amounts of (unused)
5529 ASN1 code. Grouping together similar functions and splitting unrelated
5530 functions prevents this.
5533 *) Cleanup of EVP macros.
5536 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5537 correct _ecb suffix.
5540 *) Add initial OCSP responder support to ocsp application. The
5541 revocation information is handled using the text based index
5542 use by the ca application. The responder can either handle
5543 requests generated internally, supplied in files (for example
5544 via a CGI script) or using an internal minimal server.
5547 *) Add configuration choices to get zlib compression for TLS.
5550 *) Changes to Kerberos SSL for RFC 2712 compliance:
5551 1. Implemented real KerberosWrapper, instead of just using
5552 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5553 2. Implemented optional authenticator field of KerberosWrapper.
5555 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5556 and authenticator structs; see crypto/krb5/.
5558 Generalized Kerberos calls to support multiple Kerberos libraries.
5559 [Vern Staats <staatsvr@asc.hpc.mil>,
5560 Jeffrey Altman <jaltman@columbia.edu>
5561 via Richard Levitte]
5563 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5564 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5565 values for each of the key sizes rather than having just
5566 parameters (and 'speed' generating keys each time).
5569 *) Speed up EVP routines.
5572 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5573 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5574 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5575 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5577 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5578 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5579 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5582 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5584 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5587 *) Added the OS2-EMX target.
5588 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5590 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5591 to support NCONF routines in extension code. New function CONF_set_nconf()
5592 to allow functions which take an NCONF to also handle the old LHASH
5593 structure: this means that the old CONF compatible routines can be
5594 retained (in particular wrt extensions) without having to duplicate the
5595 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5598 *) Enhance the general user interface with mechanisms for inner control
5599 and with possibilities to have yes/no kind of prompts.
5602 *) Change all calls to low level digest routines in the library and
5603 applications to use EVP. Add missing calls to HMAC_cleanup() and
5604 don't assume HMAC_CTX can be copied using memcpy().
5605 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5607 *) Add the possibility to control engines through control names but with
5608 arbitrary arguments instead of just a string.
5609 Change the key loaders to take a UI_METHOD instead of a callback
5610 function pointer. NOTE: this breaks binary compatibility with earlier
5611 versions of OpenSSL [engine].
5612 Adapt the nCipher code for these new conditions and add a card insertion
5616 *) Enhance the general user interface with mechanisms to better support
5617 dialog box interfaces, application-defined prompts, the possibility
5618 to use defaults (for example default passwords from somewhere else)
5619 and interrupts/cancellations.
5622 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5623 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5626 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5627 tidy up some unnecessarily weird code in 'sk_new()').
5628 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5630 *) Change the key loading routines for ENGINEs to use the same kind
5631 callback (pem_password_cb) as all other routines that need this
5635 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5636 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5637 than this minimum value is recommended.
5640 *) New random seeder for OpenVMS, using the system process statistics
5641 that are easily reachable.
5644 *) Windows apparently can't transparently handle global
5645 variables defined in DLLs. Initialisations such as:
5647 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5649 wont compile. This is used by the any applications that need to
5650 declare their own ASN1 modules. This was fixed by adding the option
5651 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5652 needed for static libraries under Win32.
5655 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5656 setting of purpose and trust fields. New X509_STORE trust and
5657 purpose functions and tidy up setting in other SSL functions.
5660 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5661 structure. These are inherited by X509_STORE_CTX when it is
5662 initialised. This allows various defaults to be set in the
5663 X509_STORE structure (such as flags for CRL checking and custom
5664 purpose or trust settings) for functions which only use X509_STORE_CTX
5665 internally such as S/MIME.
5667 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5668 trust settings if they are not set in X509_STORE. This allows X509_STORE
5669 purposes and trust (in S/MIME for example) to override any set by default.
5671 Add command line options for CRL checking to smime, s_client and s_server
5675 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5676 are set then the CRL is looked up in the X509_STORE structure and
5677 its validity and signature checked, then if the certificate is found
5678 in the CRL the verify fails with a revoked error.
5680 Various new CRL related callbacks added to X509_STORE_CTX structure.
5682 Command line options added to 'verify' application to support this.
5684 This needs some additional work, such as being able to handle multiple
5685 CRLs with different times, extension based lookup (rather than just
5686 by subject name) and ultimately more complete V2 CRL extension
5690 *) Add a general user interface API (crypto/ui/). This is designed
5691 to replace things like des_read_password and friends (backward
5692 compatibility functions using this new API are provided).
5693 The purpose is to remove prompting functions from the DES code
5694 section as well as provide for prompting through dialog boxes in
5695 a window system and the like.
5698 *) Add "ex_data" support to ENGINE so implementations can add state at a
5699 per-structure level rather than having to store it globally.
5702 *) Make it possible for ENGINE structures to be copied when retrieved by
5703 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5704 This causes the "original" ENGINE structure to act like a template,
5705 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5706 operational state can be localised to each ENGINE structure, despite the
5707 fact they all share the same "methods". New ENGINE structures returned in
5708 this case have no functional references and the return value is the single
5709 structural reference. This matches the single structural reference returned
5710 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5714 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5715 needs to match any other type at all we need to manually clear the
5719 *) Changes to the "openssl engine" utility to include;
5720 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5721 about an ENGINE's available control commands.
5722 - executing control commands from command line arguments using the
5723 '-pre' and '-post' switches. '-post' is only used if '-t' is
5724 specified and the ENGINE is successfully initialised. The syntax for
5725 the individual commands are colon-separated, for example;
5726 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5729 *) New dynamic control command support for ENGINEs. ENGINEs can now
5730 declare their own commands (numbers), names (strings), descriptions,
5731 and input types for run-time discovery by calling applications. A
5732 subset of these commands are implicitly classed as "executable"
5733 depending on their input type, and only these can be invoked through
5734 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5735 can be based on user input, config files, etc). The distinction is
5736 that "executable" commands cannot return anything other than a boolean
5737 result and can only support numeric or string input, whereas some
5738 discoverable commands may only be for direct use through
5739 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5740 pointers, or other custom uses. The "executable" commands are to
5741 support parameterisations of ENGINE behaviour that can be
5742 unambiguously defined by ENGINEs and used consistently across any
5743 OpenSSL-based application. Commands have been added to all the
5744 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5745 control over shared-library paths without source code alterations.
5748 *) Changed all ENGINE implementations to dynamically allocate their
5749 ENGINEs rather than declaring them statically. Apart from this being
5750 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5751 this also allows the implementations to compile without using the
5752 internal engine_int.h header.
5755 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5756 'const' value. Any code that should be able to modify a RAND_METHOD
5757 should already have non-const pointers to it (ie. they should only
5758 modify their own ones).
5761 *) Made a variety of little tweaks to the ENGINE code.
5762 - "atalla" and "ubsec" string definitions were moved from header files
5763 to C code. "nuron" string definitions were placed in variables
5764 rather than hard-coded - allowing parameterisation of these values
5765 later on via ctrl() commands.
5766 - Removed unused "#if 0"'d code.
5767 - Fixed engine list iteration code so it uses ENGINE_free() to release
5768 structural references.
5769 - Constified the RAND_METHOD element of ENGINE structures.
5770 - Constified various get/set functions as appropriate and added
5771 missing functions (including a catch-all ENGINE_cpy that duplicates
5772 all ENGINE values onto a new ENGINE except reference counts/state).
5773 - Removed NULL parameter checks in get/set functions. Setting a method
5774 or function to NULL is a way of cancelling out a previously set
5775 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5776 and doesn't justify the extra error symbols and code.
5777 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5778 flags from engine_int.h to engine.h.
5779 - Changed prototypes for ENGINE handler functions (init(), finish(),
5780 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5783 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5784 to the algorithm using long division. The binary algorithm can be
5785 used only if the modulus is odd. On 32-bit systems, it is faster
5786 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5787 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5788 up to 450 bits. In 64-bit environments, the binary algorithm
5789 appears to be advantageous for much longer moduli; here we use it
5790 for moduli up to 2048 bits.
5793 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5794 could not support the combine flag in choice fields.
5797 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5798 extensions from a certificate request to the certificate.
5801 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5802 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5803 file: this allows the display of the certificate about to be
5804 signed to be customised, to allow certain fields to be included
5805 or excluded and extension details. The old system didn't display
5806 multicharacter strings properly, omitted fields not in the policy
5807 and couldn't display additional details such as extensions.
5810 *) Function EC_POINTs_mul for multiple scalar multiplication
5811 of an arbitrary number of elliptic curve points
5812 \sum scalars[i]*points[i],
5813 optionally including the generator defined for the EC_GROUP:
5814 scalar*generator + \sum scalars[i]*points[i].
5816 EC_POINT_mul is a simple wrapper function for the typical case
5817 that the point list has just one item (besides the optional
5821 *) First EC_METHODs for curves over GF(p):
5823 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5824 operations and provides various method functions that can also
5825 operate with faster implementations of modular arithmetic.
5827 EC_GFp_mont_method() reuses most functions that are part of
5828 EC_GFp_simple_method, but uses Montgomery arithmetic.
5830 [Bodo Moeller; point addition and point doubling
5831 implementation directly derived from source code provided by
5832 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5834 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5835 crypto/ec/ec_lib.c):
5837 Curves are EC_GROUP objects (with an optional group generator)
5838 based on EC_METHODs that are built into the library.
5840 Points are EC_POINT objects based on EC_GROUP objects.
5842 Most of the framework would be able to handle curves over arbitrary
5843 finite fields, but as there are no obvious types for fields other
5844 than GF(p), some functions are limited to that for now.
5847 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5848 that the file contains a complete HTTP response.
5851 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5852 change the def and num file printf format specifier from "%-40sXXX"
5853 to "%-39s XXX". The latter will always guarantee a space after the
5854 field while the former will cause them to run together if the field
5855 is 40 of more characters long.
5858 *) Constify the cipher and digest 'method' functions and structures
5859 and modify related functions to take constant EVP_MD and EVP_CIPHER
5863 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5864 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5867 *) Modify EVP_Digest*() routines so they now return values. Although the
5868 internal software routines can never fail additional hardware versions
5872 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5874 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5875 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5878 ERR_R_NESTED_ASN1_ERROR
5880 ERR_R_MISSING_ASN1_EOS
5881 were 4 .. 9, conflicting with
5882 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5884 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5885 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5887 Add new error code 'ERR_R_INTERNAL_ERROR'.
5890 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5894 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5895 sets the subject name for a new request or supersedes the
5896 subject name in a given request. Formats that can be parsed are
5897 'CN=Some Name, OU=myOU, C=IT'
5899 'CN=Some Name/OU=myOU/C=IT'.
5901 Add options '-batch' and '-verbose' to 'openssl req'.
5902 [Massimiliano Pala <madwolf@hackmasters.net>]
5904 *) Introduce the possibility to access global variables through
5905 functions on platform were that's the best way to handle exporting
5906 global variables in shared libraries. To enable this functionality,
5907 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5908 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5909 is normally done by Configure or something similar).
5911 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5912 in the source file (foo.c) like this:
5914 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5915 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5917 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5918 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5920 OPENSSL_DECLARE_GLOBAL(int,foo);
5921 #define foo OPENSSL_GLOBAL_REF(foo)
5922 OPENSSL_DECLARE_GLOBAL(double,bar);
5923 #define bar OPENSSL_GLOBAL_REF(bar)
5925 The #defines are very important, and therefore so is including the
5926 header file everywhere where the defined globals are used.
5928 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5929 of ASN.1 items, but that structure is a bit different.
5931 The largest change is in util/mkdef.pl which has been enhanced with
5932 better and easier to understand logic to choose which symbols should
5933 go into the Windows .def files as well as a number of fixes and code
5934 cleanup (among others, algorithm keywords are now sorted
5935 lexicographically to avoid constant rewrites).
5938 *) In BN_div() keep a copy of the sign of 'num' before writing the
5939 result to 'rm' because if rm==num the value will be overwritten
5940 and produce the wrong result if 'num' is negative: this caused
5941 problems with BN_mod() and BN_nnmod().
5944 *) Function OCSP_request_verify(). This checks the signature on an
5945 OCSP request and verifies the signer certificate. The signer
5946 certificate is just checked for a generic purpose and OCSP request
5950 *) Add OCSP_check_validity() function to check the validity of OCSP
5951 responses. OCSP responses are prepared in real time and may only
5952 be a few seconds old. Simply checking that the current time lies
5953 between thisUpdate and nextUpdate max reject otherwise valid responses
5954 caused by either OCSP responder or client clock inaccuracy. Instead
5955 we allow thisUpdate and nextUpdate to fall within a certain period of
5956 the current time. The age of the response can also optionally be
5957 checked. Two new options -validity_period and -status_age added to
5961 *) If signature or public key algorithm is unrecognized print out its
5962 OID rather that just UNKNOWN.
5965 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5966 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5967 ID to be generated from the issuer certificate alone which can then be
5968 passed to OCSP_id_issuer_cmp().
5971 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5972 ASN1 modules to export functions returning ASN1_ITEM pointers
5973 instead of the ASN1_ITEM structures themselves. This adds several
5974 new macros which allow the underlying ASN1 function/structure to
5975 be accessed transparently. As a result code should not use ASN1_ITEM
5976 references directly (such as &X509_it) but instead use the relevant
5977 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5978 use of the new ASN1 code on platforms where exporting structures
5979 is problematical (for example in shared libraries) but exporting
5980 functions returning pointers to structures is not.
5983 *) Add support for overriding the generation of SSL/TLS session IDs.
5984 These callbacks can be registered either in an SSL_CTX or per SSL.
5985 The purpose of this is to allow applications to control, if they wish,
5986 the arbitrary values chosen for use as session IDs, particularly as it
5987 can be useful for session caching in multiple-server environments. A
5988 command-line switch for testing this (and any client code that wishes
5989 to use such a feature) has been added to "s_server".
5990 [Geoff Thorpe, Lutz Jaenicke]
5992 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5993 of the form '#if defined(...) || defined(...) || ...' and
5994 '#if !defined(...) && !defined(...) && ...'. This also avoids
5995 the growing number of special cases it was previously handling.
5998 *) Make all configuration macros available for application by making
5999 sure they are available in opensslconf.h, by giving them names starting
6000 with "OPENSSL_" to avoid conflicts with other packages and by making
6001 sure e_os2.h will cover all platform-specific cases together with
6003 Additionally, it is now possible to define configuration/platform-
6004 specific names (called "system identities"). In the C code, these
6005 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6006 macro with the name beginning with "OPENSSL_SYS_", which is determined
6007 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6011 *) New option -set_serial to 'req' and 'x509' this allows the serial
6012 number to use to be specified on the command line. Previously self
6013 signed certificates were hard coded with serial number 0 and the
6014 CA options of 'x509' had to use a serial number in a file which was
6018 *) New options to 'ca' utility to support V2 CRL entry extensions.
6019 Currently CRL reason, invalidity date and hold instruction are
6020 supported. Add new CRL extensions to V3 code and some new objects.
6023 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6024 disable standard block padding (aka PKCS#5 padding) in the EVP
6025 API, which was previously mandatory. This means that the data is
6026 not padded in any way and so the total length much be a multiple
6027 of the block size, otherwise an error occurs.
6030 *) Initial (incomplete) OCSP SSL support.
6033 *) New function OCSP_parse_url(). This splits up a URL into its host,
6034 port and path components: primarily to parse OCSP URLs. New -url
6035 option to ocsp utility.
6038 *) New nonce behavior. The return value of OCSP_check_nonce() now
6039 reflects the various checks performed. Applications can decide
6040 whether to tolerate certain situations such as an absent nonce
6041 in a response when one was present in a request: the ocsp application
6042 just prints out a warning. New function OCSP_add1_basic_nonce()
6043 this is to allow responders to include a nonce in a response even if
6044 the request is nonce-less.
6047 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6048 skipped when using openssl x509 multiple times on a single input file,
6049 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6052 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6053 set string type: to handle setting ASN1_TIME structures. Fix ca
6054 utility to correctly initialize revocation date of CRLs.
6057 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6058 the clients preferred ciphersuites and rather use its own preferences.
6059 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6060 Internet Explorer by ensuring unchanged hash method during stepup.
6061 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6064 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6065 to aes and add a new 'exist' option to print out symbols that don't
6069 *) Additional options to ocsp utility to allow flags to be set and
6070 additional certificates supplied.
6073 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6074 OCSP client a number of certificate to only verify the response
6078 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6079 handle the new API. Currently only ECB, CBC modes supported. Add new
6082 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6083 Encryption Standard (AES) Ciphersuites for Transport Layer
6084 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6085 not enabled by default and were not part of the "ALL" ciphersuite
6086 alias because they were not yet official; they could be
6087 explicitly requested by specifying the "AESdraft" ciphersuite
6088 group alias. In the final release of OpenSSL 0.9.7, the group
6089 alias is called "AES" and is part of "ALL".)
6090 [Ben Laurie, Steve Henson, Bodo Moeller]
6092 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6093 request to response.
6096 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6097 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6098 extract information from a certificate request. OCSP_response_create()
6099 creates a response and optionally adds a basic response structure.
6100 OCSP_basic_add1_status() adds a complete single response to a basic
6101 response and returns the OCSP_SINGLERESP structure just added (to allow
6102 extensions to be included for example). OCSP_basic_add1_cert() adds a
6103 certificate to a basic response and OCSP_basic_sign() signs a basic
6104 response with various flags. New helper functions ASN1_TIME_check()
6105 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6106 (converts ASN1_TIME to GeneralizedTime).
6109 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6110 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6111 structure from a certificate. X509_pubkey_digest() digests the public_key
6112 contents: this is used in various key identifiers.
6115 *) Make sk_sort() tolerate a NULL argument.
6116 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6118 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6119 passed by the function are trusted implicitly. If any of them signed the
6120 response then it is assumed to be valid and is not verified.
6123 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6124 to data. This was previously part of the PKCS7 ASN1 code. This
6125 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6126 [Steve Henson, reported by Kenneth R. Robinette
6127 <support@securenetterm.com>]
6129 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6130 routines: without these tracing memory leaks is very painful.
6131 Fix leaks in PKCS12 and PKCS7 routines.
6134 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6135 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6136 effectively meant GeneralizedTime would never be used. Now it
6137 is initialised to -1 but X509_time_adj() now has to check the value
6138 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6139 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6140 [Steve Henson, reported by Kenneth R. Robinette
6141 <support@securenetterm.com>]
6143 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6144 result in a zero length in the ASN1_INTEGER structure which was
6145 not consistent with the structure when d2i_ASN1_INTEGER() was used
6146 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6147 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6148 where it did not print out a minus for negative ASN1_INTEGER.
6151 *) Add summary printout to ocsp utility. The various functions which
6152 convert status values to strings have been renamed to:
6153 OCSP_response_status_str(), OCSP_cert_status_str() and
6154 OCSP_crl_reason_str() and are no longer static. New options
6155 to verify nonce values and to disable verification. OCSP response
6156 printout format cleaned up.
6159 *) Add additional OCSP certificate checks. These are those specified
6160 in RFC2560. This consists of two separate checks: the CA of the
6161 certificate being checked must either be the OCSP signer certificate
6162 or the issuer of the OCSP signer certificate. In the latter case the
6163 OCSP signer certificate must contain the OCSP signing extended key
6164 usage. This check is performed by attempting to match the OCSP
6165 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6166 in the OCSP_CERTID structures of the response.
6169 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6170 and related routines. This uses the standard OpenSSL certificate
6171 verify routines to perform initial checks (just CA validity) and
6172 to obtain the certificate chain. Then additional checks will be
6173 performed on the chain. Currently the root CA is checked to see
6174 if it is explicitly trusted for OCSP signing. This is used to set
6175 a root CA as a global signing root: that is any certificate that
6176 chains to that CA is an acceptable OCSP signing certificate.
6179 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6180 extensions from a separate configuration file.
6181 As when reading extensions from the main configuration file,
6182 the '-extensions ...' option may be used for specifying the
6184 [Massimiliano Pala <madwolf@comune.modena.it>]
6186 *) New OCSP utility. Allows OCSP requests to be generated or
6187 read. The request can be sent to a responder and the output
6188 parsed, outputed or printed in text form. Not complete yet:
6189 still needs to check the OCSP response validity.
6192 *) New subcommands for 'openssl ca':
6193 'openssl ca -status <serial>' prints the status of the cert with
6194 the given serial number (according to the index file).
6195 'openssl ca -updatedb' updates the expiry status of certificates
6197 [Massimiliano Pala <madwolf@comune.modena.it>]
6199 *) New '-newreq-nodes' command option to CA.pl. This is like
6200 '-newreq', but calls 'openssl req' with the '-nodes' option
6201 so that the resulting key is not encrypted.
6202 [Damien Miller <djm@mindrot.org>]
6204 *) New configuration for the GNU Hurd.
6205 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6207 *) Initial code to implement OCSP basic response verify. This
6208 is currently incomplete. Currently just finds the signer's
6209 certificate and verifies the signature on the response.
6212 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6213 value of OPENSSLDIR. This is available via the new '-d' option
6214 to 'openssl version', and is also included in 'openssl version -a'.
6217 *) Allowing defining memory allocation callbacks that will be given
6218 file name and line number information in additional arguments
6219 (a const char* and an int). The basic functionality remains, as
6220 well as the original possibility to just replace malloc(),
6221 realloc() and free() by functions that do not know about these
6222 additional arguments. To register and find out the current
6223 settings for extended allocation functions, the following
6224 functions are provided:
6226 CRYPTO_set_mem_ex_functions
6227 CRYPTO_set_locked_mem_ex_functions
6228 CRYPTO_get_mem_ex_functions
6229 CRYPTO_get_locked_mem_ex_functions
6231 These work the same way as CRYPTO_set_mem_functions and friends.
6232 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6233 extended allocation function is enabled.
6234 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6235 a conventional allocation function is enabled.
6236 [Richard Levitte, Bodo Moeller]
6238 *) Finish off removing the remaining LHASH function pointer casts.
6239 There should no longer be any prototype-casting required when using
6240 the LHASH abstraction, and any casts that remain are "bugs". See
6241 the callback types and macros at the head of lhash.h for details
6242 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6245 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6246 If /dev/[u]random devices are not available or do not return enough
6247 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6249 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6250 /etc/entropy will be queried once each in this sequence, quering stops
6251 when enough entropy was collected without querying more sockets.
6254 *) Change the Unix RAND_poll() variant to be able to poll several
6255 random devices, as specified by DEVRANDOM, until a sufficient amount
6256 of data has been collected. We spend at most 10 ms on each file
6257 (select timeout) and read in non-blocking mode. DEVRANDOM now
6258 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6259 (previously it was just the string "/dev/urandom"), so on typical
6260 platforms the 10 ms delay will never occur.
6261 Also separate out the Unix variant to its own file, rand_unix.c.
6262 For VMS, there's a currently-empty rand_vms.c.
6265 *) Move OCSP client related routines to ocsp_cl.c. These
6266 provide utility functions which an application needing
6267 to issue a request to an OCSP responder and analyse the
6268 response will typically need: as opposed to those which an
6269 OCSP responder itself would need which will be added later.
6271 OCSP_request_sign() signs an OCSP request with an API similar
6272 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6273 response. OCSP_response_get1_basic() extracts basic response
6274 from response. OCSP_resp_find_status(): finds and extracts status
6275 information from an OCSP_CERTID structure (which will be created
6276 when the request structure is built). These are built from lower
6277 level functions which work on OCSP_SINGLERESP structures but
6278 wont normally be used unless the application wishes to examine
6279 extensions in the OCSP response for example.
6281 Replace nonce routines with a pair of functions.
6282 OCSP_request_add1_nonce() adds a nonce value and optionally
6283 generates a random value. OCSP_check_nonce() checks the
6284 validity of the nonce in an OCSP response.
6287 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6288 This doesn't copy the supplied OCSP_CERTID and avoids the
6289 need to free up the newly created id. Change return type
6290 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6291 This can then be used to add extensions to the request.
6292 Deleted OCSP_request_new(), since most of its functionality
6293 is now in OCSP_REQUEST_new() (and the case insensitive name
6294 clash) apart from the ability to set the request name which
6295 will be added elsewhere.
6298 *) Update OCSP API. Remove obsolete extensions argument from
6299 various functions. Extensions are now handled using the new
6300 OCSP extension code. New simple OCSP HTTP function which
6301 can be used to send requests and parse the response.
6304 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6305 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6306 uses the special reorder version of SET OF to sort the attributes
6307 and reorder them to match the encoded order. This resolves a long
6308 standing problem: a verify on a PKCS7 structure just after signing
6309 it used to fail because the attribute order did not match the
6310 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6311 it uses the received order. This is necessary to tolerate some broken
6312 software that does not order SET OF. This is handled by encoding
6313 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6314 to produce the required SET OF.
6317 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6318 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6319 files to get correct declarations of the ASN.1 item variables.
6322 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6323 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6324 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6325 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6326 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6327 ASN1_ITEM and no wrapper functions.
6330 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6331 replace the old function pointer based I/O routines. Change most of
6332 the *_d2i_bio() and *_d2i_fp() functions to use these.
6335 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6336 lines, recognice more "algorithms" that can be deselected, and make
6337 it complain about algorithm deselection that isn't recognised.
6340 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6341 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6342 to use new functions. Add NO_ASN1_OLD which can be set to remove
6343 some old style ASN1 functions: this can be used to determine if old
6344 code will still work when these eventually go away.
6347 *) New extension functions for OCSP structures, these follow the
6348 same conventions as certificates and CRLs.
6351 *) New function X509V3_add1_i2d(). This automatically encodes and
6352 adds an extension. Its behaviour can be customised with various
6353 flags to append, replace or delete. Various wrappers added for
6354 certifcates and CRLs.
6357 *) Fix to avoid calling the underlying ASN1 print routine when
6358 an extension cannot be parsed. Correct a typo in the
6359 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6362 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6363 entries for variables.
6366 *) Add functionality to apps/openssl.c for detecting locking
6367 problems: As the program is single-threaded, all we have
6368 to do is register a locking callback using an array for
6369 storing which locks are currently held by the program.
6372 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6373 SSL_get_ex_data_X509_STORE_idx(), which is used in
6374 ssl_verify_cert_chain() and thus can be called at any time
6375 during TLS/SSL handshakes so that thread-safety is essential.
6376 Unfortunately, the ex_data design is not at all suited
6377 for multi-threaded use, so it probably should be abolished.
6380 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6381 [Broadcom, tweaked and integrated by Geoff Thorpe]
6383 *) Move common extension printing code to new function
6384 X509V3_print_extensions(). Reorganise OCSP print routines and
6385 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6388 *) New function X509_signature_print() to remove duplication in some
6392 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6393 set (this was treated exactly the same as SET OF previously). This
6394 is used to reorder the STACK representing the structure to match the
6395 encoding. This will be used to get round a problem where a PKCS7
6396 structure which was signed could not be verified because the STACK
6397 order did not reflect the encoded order.
6400 *) Reimplement the OCSP ASN1 module using the new code.
6403 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6404 for its ASN1 operations. The old style function pointers still exist
6405 for now but they will eventually go away.
6408 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6409 completely replaces the old ASN1 functionality with a table driven
6410 encoder and decoder which interprets an ASN1_ITEM structure describing
6411 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6412 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6413 has also been converted to the new form.
6416 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6417 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6418 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6419 for negative moduli.
6422 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6423 of not touching the result's sign bit.
6426 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6430 *) Changed the LHASH code to use prototypes for callbacks, and created
6431 macros to declare and implement thin (optionally static) functions
6432 that provide type-safety and avoid function pointer casting for the
6433 type-specific callbacks.
6436 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6438 [Veers Staats <staatsvr@asc.hpc.mil>,
6439 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6441 *) Reformat the FAQ so the different questions and answers can be divided
6442 in sections depending on the subject.
6445 *) Have the zlib compression code load ZLIB.DLL dynamically under
6449 *) New function BN_mod_sqrt for computing square roots modulo a prime
6450 (using the probabilistic Tonelli-Shanks algorithm unless
6451 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6452 be handled deterministically).
6453 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6455 *) Make BN_mod_inverse faster by explicitly handling small quotients
6456 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6457 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6460 *) New function BN_kronecker.
6463 *) Fix BN_gcd so that it works on negative inputs; the result is
6464 positive unless both parameters are zero.
6465 Previously something reasonably close to an infinite loop was
6466 possible because numbers could be growing instead of shrinking
6467 in the implementation of Euclid's algorithm.
6470 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6471 sign of the number in question.
6473 Fix BN_is_word(a,w) to work correctly for w == 0.
6475 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6476 because its test if the absolute value of 'a' equals 'w'.
6477 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6478 it exists mostly for use in the implementations of BN_is_zero(),
6479 BN_is_one(), and BN_is_word().
6482 *) New function BN_swap.
6485 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6486 the exponentiation functions are more likely to produce reasonable
6487 results on negative inputs.
6490 *) Change BN_mod_mul so that the result is always non-negative.
6491 Previously, it could be negative if one of the factors was negative;
6492 I don't think anyone really wanted that behaviour.
6495 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6496 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6497 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6498 and add new functions:
6507 BN_mod_lshift1_quick
6511 These functions always generate non-negative results.
6513 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6514 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6516 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6517 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6518 be reduced modulo m.
6519 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6522 The following entry accidentally appeared in the CHANGES file
6523 distributed with OpenSSL 0.9.7. The modifications described in
6524 it do *not* apply to OpenSSL 0.9.7.
6526 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6527 was actually never needed) and in BN_mul(). The removal in BN_mul()
6528 required a small change in bn_mul_part_recursive() and the addition
6529 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6530 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6531 bn_sub_words() and bn_add_words() except they take arrays with
6536 *) In 'openssl passwd', verify passwords read from the terminal
6537 unless the '-salt' option is used (which usually means that
6538 verification would just waste user's time since the resulting
6539 hash is going to be compared with some given password hash)
6540 or the new '-noverify' option is used.
6542 This is an incompatible change, but it does not affect
6543 non-interactive use of 'openssl passwd' (passwords on the command
6544 line, '-stdin' option, '-in ...' option) and thus should not
6548 *) Remove all references to RSAref, since there's no more need for it.
6551 *) Make DSO load along a path given through an environment variable
6552 (SHLIB_PATH) with shl_load().
6555 *) Constify the ENGINE code as a result of BIGNUM constification.
6556 Also constify the RSA code and most things related to it. In a
6557 few places, most notable in the depth of the ASN.1 code, ugly
6558 casts back to non-const were required (to be solved at a later
6562 *) Make it so the openssl application has all engines loaded by default.
6565 *) Constify the BIGNUM routines a little more.
6568 *) Add the following functions:
6570 ENGINE_load_cswift()
6572 ENGINE_load_atalla()
6574 ENGINE_load_builtin_engines()
6576 That way, an application can itself choose if external engines that
6577 are built-in in OpenSSL shall ever be used or not. The benefit is
6578 that applications won't have to be linked with libdl or other dso
6579 libraries unless it's really needed.
6581 Changed 'openssl engine' to load all engines on demand.
6582 Changed the engine header files to avoid the duplication of some
6583 declarations (they differed!).
6586 *) 'openssl engine' can now list capabilities.
6589 *) Better error reporting in 'openssl engine'.
6592 *) Never call load_dh_param(NULL) in s_server.
6595 *) Add engine application. It can currently list engines by name and
6596 identity, and test if they are actually available.
6599 *) Improve RPM specification file by forcing symbolic linking and making
6600 sure the installed documentation is also owned by root.root.
6601 [Damien Miller <djm@mindrot.org>]
6603 *) Give the OpenSSL applications more possibilities to make use of
6604 keys (public as well as private) handled by engines.
6607 *) Add OCSP code that comes from CertCo.
6610 *) Add VMS support for the Rijndael code.
6613 *) Added untested support for Nuron crypto accelerator.
6616 *) Add support for external cryptographic devices. This code was
6617 previously distributed separately as the "engine" branch.
6618 [Geoff Thorpe, Richard Levitte]
6620 *) Rework the filename-translation in the DSO code. It is now possible to
6621 have far greater control over how a "name" is turned into a filename
6622 depending on the operating environment and any oddities about the
6623 different shared library filenames on each system.
6626 *) Support threads on FreeBSD-elf in Configure.
6629 *) Fix for SHA1 assembly problem with MASM: it produces
6630 warnings about corrupt line number information when assembling
6631 with debugging information. This is caused by the overlapping
6633 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6636 NCONF_get_number() has no error checking at all. As a replacement,
6637 NCONF_get_number_e() is defined (_e for "error checking") and is
6638 promoted strongly. The old NCONF_get_number is kept around for
6639 binary backward compatibility.
6640 Make it possible for methods to load from something other than a BIO,
6641 by providing a function pointer that is given a name instead of a BIO.
6642 For example, this could be used to load configuration data from an
6646 *) Fix for non blocking accept BIOs. Added new I/O special reason
6647 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6648 with non blocking I/O was not possible because no retry code was
6649 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6653 *) Added the beginnings of Rijndael support.
6656 *) Fix for bug in DirectoryString mask setting. Add support for
6657 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6658 to allow certificate printing to more controllable, additional
6659 'certopt' option to 'x509' to allow new printing options to be
6663 *) Clean old EAY MD5 hack from e_os.h.
6666 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6668 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6669 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6670 [Joe Orton, Steve Henson]
6672 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6674 *) Fix additional bug revealed by the NISCC test suite:
6676 Stop bug triggering large recursion when presented with
6677 certain ASN.1 tags (CVE-2003-0851)
6680 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6682 *) Fix various bugs revealed by running the NISCC test suite:
6684 Stop out of bounds reads in the ASN1 code when presented with
6685 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6687 If verify callback ignores invalid public key errors don't try to check
6688 certificate signature with the NULL public key.
6692 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6693 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6697 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6698 extra data after the compression methods not only for TLS 1.0
6699 but also for SSL 3.0 (as required by the specification).
6700 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6702 *) Change X509_certificate_type() to mark the key as exported/exportable
6703 when it's 512 *bits* long, not 512 bytes.
6706 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6708 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6709 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6710 a protocol version number mismatch like a decryption error
6711 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6714 *) Turn on RSA blinding by default in the default implementation
6715 to avoid a timing attack. Applications that don't want it can call
6716 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6717 They would be ill-advised to do so in most cases.
6718 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6720 *) Change RSA blinding code so that it works when the PRNG is not
6721 seeded (in this case, the secret RSA exponent is abused as
6722 an unpredictable seed -- if it is not unpredictable, there
6723 is no point in blinding anyway). Make RSA blinding thread-safe
6724 by remembering the creator's thread ID in rsa->blinding and
6725 having all other threads use local one-time blinding factors
6726 (this requires more computation than sharing rsa->blinding, but
6727 avoids excessive locking; and if an RSA object is not shared
6728 between threads, blinding will still be very fast).
6731 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6733 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6734 via timing by performing a MAC computation even if incorrrect
6735 block cipher padding has been found. This is a countermeasure
6736 against active attacks where the attacker has to distinguish
6737 between bad padding and a MAC verification error. (CVE-2003-0078)
6739 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6740 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6741 Martin Vuagnoux (EPFL, Ilion)]
6743 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6745 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6746 memory from it's contents. This is done with a counter that will
6747 place alternating values in each byte. This can be used to solve
6748 two issues: 1) the removal of calls to memset() by highly optimizing
6749 compilers, and 2) cleansing with other values than 0, since those can
6750 be read through on certain media, for example a swap space on disk.
6753 *) Bugfix: client side session caching did not work with external caching,
6754 because the session->cipher setting was not restored when reloading
6755 from the external cache. This problem was masked, when
6756 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6757 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6760 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6761 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6762 [Zeev Lieber <zeev-l@yahoo.com>]
6764 *) Undo an undocumented change introduced in 0.9.6e which caused
6765 repeated calls to OpenSSL_add_all_ciphers() and
6766 OpenSSL_add_all_digests() to be ignored, even after calling
6770 *) Change the default configuration reader to deal with last line not
6771 being properly terminated.
6774 *) Change X509_NAME_cmp() so it applies the special rules on handling
6775 DN values that are of type PrintableString, as well as RDNs of type
6776 emailAddress where the value has the type ia5String.
6777 [stefank@valicert.com via Richard Levitte]
6779 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6780 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6781 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6782 the bitwise-OR of the two for use by the majority of applications
6783 wanting this behaviour, and update the docs. The documented
6784 behaviour and actual behaviour were inconsistent and had been
6785 changing anyway, so this is more a bug-fix than a behavioural
6787 [Geoff Thorpe, diagnosed by Nadav Har'El]
6789 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6790 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6793 *) Fix initialization code race conditions in
6794 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6795 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6796 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6797 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6798 ssl2_get_cipher_by_char(),
6799 ssl3_get_cipher_by_char().
6800 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6802 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6803 the cached sessions are flushed, as the remove_cb() might use ex_data
6804 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6805 (see [openssl.org #212]).
6806 [Geoff Thorpe, Lutz Jaenicke]
6808 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6809 length, instead of the encoding length to d2i_ASN1_OBJECT.
6812 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6814 *) [In 0.9.6g-engine release:]
6815 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6816 [Lynn Gazis <lgazis@rainbow.com>]
6818 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6820 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6821 and get fix the header length calculation.
6822 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6823 Alon Kantor <alonk@checkpoint.com> (and others),
6826 *) Use proper error handling instead of 'assertions' in buffer
6827 overflow checks added in 0.9.6e. This prevents DoS (the
6828 assertions could call abort()).
6829 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6831 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6833 *) Add various sanity checks to asn1_get_length() to reject
6834 the ASN1 length bytes if they exceed sizeof(long), will appear
6835 negative or the content length exceeds the length of the
6837 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6839 *) Fix cipher selection routines: ciphers without encryption had no flags
6840 for the cipher strength set and where therefore not handled correctly
6841 by the selection routines (PR #130).
6844 *) Fix EVP_dsa_sha macro.
6848 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6849 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6850 that was added in OpenSSL 0.9.6d.
6852 As the countermeasure turned out to be incompatible with some
6853 broken SSL implementations, the new option is part of SSL_OP_ALL.
6854 SSL_OP_ALL is usually employed when compatibility with weird SSL
6855 implementations is desired (e.g. '-bugs' option to 's_client' and
6856 's_server'), so the new option is automatically set in many
6860 *) Changes in security patch:
6862 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6863 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6864 Air Force Materiel Command, USAF, under agreement number
6867 *) Add various sanity checks to asn1_get_length() to reject
6868 the ASN1 length bytes if they exceed sizeof(long), will appear
6869 negative or the content length exceeds the length of the
6870 supplied buffer. (CVE-2002-0659)
6871 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6873 *) Assertions for various potential buffer overflows, not known to
6875 [Ben Laurie (CHATS)]
6877 *) Various temporary buffers to hold ASCII versions of integers were
6878 too small for 64 bit platforms. (CVE-2002-0655)
6879 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6881 *) Remote buffer overflow in SSL3 protocol - an attacker could
6882 supply an oversized session ID to a client. (CVE-2002-0656)
6883 [Ben Laurie (CHATS)]
6885 *) Remote buffer overflow in SSL2 protocol - an attacker could
6886 supply an oversized client master key. (CVE-2002-0656)
6887 [Ben Laurie (CHATS)]
6889 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6891 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6892 encoded as NULL) with id-dsa-with-sha1.
6893 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6895 *) Check various X509_...() return values in apps/req.c.
6896 [Nils Larsch <nla@trustcenter.de>]
6898 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6899 an end-of-file condition would erronously be flagged, when the CRLF
6900 was just at the end of a processed block. The bug was discovered when
6901 processing data through a buffering memory BIO handing the data to a
6902 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6903 <ptsekov@syntrex.com> and Nedelcho Stanev.
6906 *) Implement a countermeasure against a vulnerability recently found
6907 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6908 before application data chunks to avoid the use of known IVs
6909 with data potentially chosen by the attacker.
6912 *) Fix length checks in ssl3_get_client_hello().
6915 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6916 to prevent ssl3_read_internal() from incorrectly assuming that
6917 ssl3_read_bytes() found application data while handshake
6918 processing was enabled when in fact s->s3->in_read_app_data was
6919 merely automatically cleared during the initial handshake.
6920 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6922 *) Fix object definitions for Private and Enterprise: they were not
6923 recognized in their shortname (=lowercase) representation. Extend
6924 obj_dat.pl to issue an error when using undefined keywords instead
6925 of silently ignoring the problem (Svenning Sorensen
6926 <sss@sss.dnsalias.net>).
6929 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6930 generators, i.e. generators other than 2 and 5. (Previously, the
6931 code did not properly initialise the 'add' and 'rem' values to
6932 BN_generate_prime().)
6934 In the new general case, we do not insist that 'generator' is
6935 actually a primitive root: This requirement is rather pointless;
6936 a generator of the order-q subgroup is just as good, if not
6940 *) Map new X509 verification errors to alerts. Discovered and submitted by
6941 Tom Wu <tom@arcot.com>.
6944 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6945 returning non-zero before the data has been completely received
6946 when using non-blocking I/O.
6947 [Bodo Moeller; problem pointed out by John Hughes]
6949 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6950 [Ben Laurie, Lutz Jaenicke]
6952 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6953 Yoram Zahavi <YoramZ@gilian.com>).
6956 *) Add information about CygWin 1.3 and on, and preserve proper
6957 configuration for the versions before that.
6958 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6960 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6961 check whether we deal with a copy of a session and do not delete from
6962 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6963 <izhar@checkpoint.com>.
6966 *) Do not store session data into the internal session cache, if it
6967 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6968 flag is set). Proposed by Aslam <aslam@funk.com>.
6971 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6975 *) [In 0.9.6d-engine release:]
6976 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6977 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6979 *) Add the configuration target linux-s390x.
6980 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6982 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6983 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6984 variable as an indication that a ClientHello message has been
6985 received. As the flag value will be lost between multiple
6986 invocations of ssl3_accept when using non-blocking I/O, the
6987 function may not be aware that a handshake has actually taken
6988 place, thus preventing a new session from being added to the
6991 To avoid this problem, we now set s->new_session to 2 instead of
6992 using a local variable.
6993 [Lutz Jaenicke, Bodo Moeller]
6995 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6996 if the SSL_R_LENGTH_MISMATCH error is detected.
6997 [Geoff Thorpe, Bodo Moeller]
6999 *) New 'shared_ldflag' column in Configure platform table.
7002 *) Fix EVP_CIPHER_mode macro.
7003 ["Dan S. Camper" <dan@bti.net>]
7005 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7006 type, we must throw them away by setting rr->length to 0.
7007 [D P Chang <dpc@qualys.com>]
7009 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7011 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7012 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7013 worked incorrectly for those cases where range = 10..._2 and
7014 3*range is two bits longer than range.)
7017 *) Only add signing time to PKCS7 structures if it is not already
7021 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7022 OBJ_ld_ce should be OBJ_id_ce.
7023 Also some ip-pda OIDs in crypto/objects/objects.txt were
7024 incorrect (cf. RFC 3039).
7025 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7027 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7028 returns early because it has nothing to do.
7029 [Andy Schneider <andy.schneider@bjss.co.uk>]
7031 *) [In 0.9.6c-engine release:]
7032 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7033 [Andy Schneider <andy.schneider@bjss.co.uk>]
7035 *) [In 0.9.6c-engine release:]
7036 Add support for Cryptographic Appliance's keyserver technology.
7037 (Use engine 'keyclient')
7038 [Cryptographic Appliances and Geoff Thorpe]
7040 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7041 is called via tools/c89.sh because arguments have to be
7042 rearranged (all '-L' options must appear before the first object
7044 [Richard Shapiro <rshapiro@abinitio.com>]
7046 *) [In 0.9.6c-engine release:]
7047 Add support for Broadcom crypto accelerator cards, backported
7049 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7051 *) [In 0.9.6c-engine release:]
7052 Add support for SureWare crypto accelerator cards from
7053 Baltimore Technologies. (Use engine 'sureware')
7054 [Baltimore Technologies and Mark Cox]
7056 *) [In 0.9.6c-engine release:]
7057 Add support for crypto accelerator cards from Accelerated
7058 Encryption Processing, www.aep.ie. (Use engine 'aep')
7059 [AEP Inc. and Mark Cox]
7061 *) Add a configuration entry for gcc on UnixWare.
7062 [Gary Benson <gbenson@redhat.com>]
7064 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7065 messages are stored in a single piece (fixed-length part and
7066 variable-length part combined) and fix various bugs found on the way.
7069 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7070 instead. BIO_gethostbyname() does not know what timeouts are
7071 appropriate, so entries would stay in cache even when they have
7073 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7075 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7076 faced with a pathologically small ClientHello fragment that does
7077 not contain client_version: Instead of aborting with an error,
7078 simply choose the highest available protocol version (i.e.,
7079 TLS 1.0 unless it is disabled). In practice, ClientHello
7080 messages are never sent like this, but this change gives us
7081 strictly correct behaviour at least for TLS.
7084 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7085 never resets s->method to s->ctx->method when called from within
7086 one of the SSL handshake functions.
7087 [Bodo Moeller; problem pointed out by Niko Baric]
7089 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7090 (sent using the client's version number) if client_version is
7091 smaller than the protocol version in use. Also change
7092 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7093 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7094 the client will at least see that alert.
7097 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7101 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7102 client receives HelloRequest while in a handshake.
7103 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7105 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7106 should end in 'break', not 'goto end' which circumvents various
7107 cleanups done in state SSL_ST_OK. But session related stuff
7108 must be disabled for SSL_ST_OK in the case that we just sent a
7111 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7112 before just sending a HelloRequest.
7113 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7115 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7116 reveal whether illegal block cipher padding was found or a MAC
7117 verification error occured. (Neither SSLerr() codes nor alerts
7118 are directly visible to potential attackers, but the information
7119 may leak via logfiles.)
7121 Similar changes are not required for the SSL 2.0 implementation
7122 because the number of padding bytes is sent in clear for SSL 2.0,
7123 and the extra bytes are just ignored. However ssl/s2_pkt.c
7124 failed to verify that the purported number of padding bytes is in
7128 *) Add OpenUNIX-8 support including shared libraries
7129 (Boyd Lynn Gerber <gerberb@zenez.com>).
7132 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7133 'wristwatch attack' using huge encoding parameters (cf.
7134 James H. Manger's CRYPTO 2001 paper). Note that the
7135 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7136 encoding parameters and hence was not vulnerable.
7139 *) BN_sqr() bug fix.
7140 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7142 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7143 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7144 followed by modular reduction.
7145 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7147 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7148 equivalent based on BN_pseudo_rand() instead of BN_rand().
7151 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7152 This function was broken, as the check for a new client hello message
7153 to handle SGC did not allow these large messages.
7154 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7157 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7160 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7161 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7164 *) Rework the configuration and shared library support for Tru64 Unix.
7165 The configuration part makes use of modern compiler features and
7166 still retains old compiler behavior for those that run older versions
7167 of the OS. The shared library support part includes a variant that
7168 uses the RPATH feature, and is available through the special
7169 configuration target "alpha-cc-rpath", which will never be selected
7171 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7173 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7174 with the same message size as in ssl3_get_certificate_request().
7175 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7176 messages might inadvertently be reject as too long.
7177 [Petr Lampa <lampa@fee.vutbr.cz>]
7179 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7182 *) Modified SSL library such that the verify_callback that has been set
7183 specificly for an SSL object with SSL_set_verify() is actually being
7184 used. Before the change, a verify_callback set with this function was
7185 ignored and the verify_callback() set in the SSL_CTX at the time of
7186 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7187 to allow the necessary settings.
7190 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7191 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7192 done automatically (in contradiction to the requirements of the C
7193 standard). This made problems when used from OpenSSH.
7196 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7197 dh->length and always used
7199 BN_rand_range(priv_key, dh->p).
7201 BN_rand_range() is not necessary for Diffie-Hellman, and this
7202 specific range makes Diffie-Hellman unnecessarily inefficient if
7203 dh->length (recommended exponent length) is much smaller than the
7204 length of dh->p. We could use BN_rand_range() if the order of
7205 the subgroup was stored in the DH structure, but we only have
7210 BN_rand(priv_key, l, ...)
7212 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7218 RSA_eay_public_encrypt
7219 RSA_eay_private_decrypt
7220 RSA_eay_private_encrypt (signing)
7221 RSA_eay_public_decrypt (signature verification)
7223 (default implementations for RSA_public_encrypt,
7224 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7225 always reject numbers >= n.
7228 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7229 to synchronize access to 'locking_thread'. This is necessary on
7230 systems where access to 'locking_thread' (an 'unsigned long'
7231 variable) is not atomic.
7234 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7235 *before* setting the 'crypto_lock_rand' flag. The previous code had
7236 a race condition if 0 is a valid thread ID.
7237 [Travis Vitek <vitek@roguewave.com>]
7239 *) Add support for shared libraries under Irix.
7240 [Albert Chin-A-Young <china@thewrittenword.com>]
7242 *) Add configuration option to build on Linux on both big-endian and
7244 [Ralf Baechle <ralf@uni-koblenz.de>]
7246 *) Add the possibility to create shared libraries on HP-UX.
7249 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7251 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7252 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7253 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7254 PRNG state recovery was possible based on the output of
7255 one PRNG request appropriately sized to gain knowledge on
7256 'md' followed by enough consecutive 1-byte PRNG requests
7257 to traverse all of 'state'.
7259 1. When updating 'md_local' (the current thread's copy of 'md')
7260 during PRNG output generation, hash all of the previous
7261 'md_local' value, not just the half used for PRNG output.
7263 2. Make the number of bytes from 'state' included into the hash
7264 independent from the number of PRNG bytes requested.
7266 The first measure alone would be sufficient to avoid
7267 Markku-Juhani's attack. (Actually it had never occurred
7268 to me that the half of 'md_local' used for chaining was the
7269 half from which PRNG output bytes were taken -- I had always
7270 assumed that the secret half would be used.) The second
7271 measure makes sure that additional data from 'state' is never
7272 mixed into 'md_local' in small portions; this heuristically
7273 further strengthens the PRNG.
7276 *) Fix crypto/bn/asm/mips3.s.
7279 *) When only the key is given to "enc", the IV is undefined. Print out
7280 an error message in this case.
7283 *) Handle special case when X509_NAME is empty in X509 printing routines.
7286 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7287 positive and less than q.
7290 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7291 used: it isn't thread safe and the add_lock_callback should handle
7293 [Paul Rose <Paul.Rose@bridge.com>]
7295 *) Verify that incoming data obeys the block size in
7296 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7300 [Ulf Möller, Bodo Möller]
7302 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7303 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7304 when fixing the server behaviour for backwards-compatible 'client
7305 hello' messages. (Note that the attack is impractical against
7306 SSL 3.0 and TLS 1.0 anyway because length and version checking
7307 means that the probability of guessing a valid ciphertext is
7308 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7311 Before 0.9.5, the countermeasure (hide the error by generating a
7312 random 'decryption result') did not work properly because
7313 ERR_clear_error() was missing, meaning that SSL_get_error() would
7314 detect the supposedly ignored error.
7316 Both problems are now fixed.
7319 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7320 (previously it was 1024).
7323 *) Fix for compatibility mode trust settings: ignore trust settings
7324 unless some valid trust or reject settings are present.
7327 *) Fix for blowfish EVP: its a variable length cipher.
7330 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7331 parameters in DSA public key structures and return an error in the
7332 DSA routines if parameters are absent.
7335 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7336 in the current directory if neither $RANDFILE nor $HOME was set.
7337 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7338 caused some confusion to Windows users who haven't defined $HOME.
7339 Thus RAND_file_name() is changed again: e_os.h can define a
7340 DEFAULT_HOME, which will be used if $HOME is not set.
7341 For Windows, we use "C:"; on other platforms, we still require
7342 environment variables.
7344 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7345 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7346 having multiple threads call RAND_poll() concurrently.
7349 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7350 combination of a flag and a thread ID variable.
7351 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7352 flag), *other* threads can enter ssleay_add_bytes without obeying
7353 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7354 that they do not hold after the first thread unsets add_do_not_lock).
7357 *) Change bctest again: '-x' expressions are not available in all
7361 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7363 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7364 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7366 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7367 the default extension for executables, if any. Also, make the perl
7368 scripts that use symlink() to test if it really exists and use "cp"
7369 if it doesn't. All this made OpenSSL compilable and installable in
7373 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7374 If SEQUENCE is length is indefinite just set c->slen to the total
7375 amount of data available.
7376 [Steve Henson, reported by shige@FreeBSD.org]
7377 [This change does not apply to 0.9.7.]
7379 *) Change bctest to avoid here-documents inside command substitution
7380 (workaround for FreeBSD /bin/sh bug).
7381 For compatibility with Ultrix, avoid shell functions (introduced
7382 in the bctest version that searches along $PATH).
7385 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7386 with des_encrypt() defined on some operating systems, like Solaris
7390 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7391 On the Importance of Eliminating Errors in Cryptographic
7392 Computations, J. Cryptology 14 (2001) 2, 101-119,
7393 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7396 *) MIPS assembler BIGNUM division bug fix.
7399 *) Disabled incorrect Alpha assembler code.
7402 *) Fix PKCS#7 decode routines so they correctly update the length
7403 after reading an EOC for the EXPLICIT tag.
7405 [This change does not apply to 0.9.7.]
7407 *) Fix bug in PKCS#12 key generation routines. This was triggered
7408 if a 3DES key was generated with a 0 initial byte. Include
7409 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7410 (but broken) behaviour.
7413 *) Enhance bctest to search for a working bc along $PATH and print
7415 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7417 *) Fix memory leaks in err.c: free err_data string if necessary;
7418 don't write to the wrong index in ERR_set_error_data.
7421 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7425 *) Replace rdtsc with _emit statements for VC++ version 5.
7426 [Jeremy Cooper <jeremy@baymoo.org>]
7428 *) Make it possible to reuse SSLv2 sessions.
7431 *) In copy_email() check for >= 0 as a return value for
7432 X509_NAME_get_index_by_NID() since 0 is a valid index.
7433 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7435 *) Avoid coredump with unsupported or invalid public keys by checking if
7436 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7437 PKCS7_verify() fails with non detached data.
7440 *) Don't use getenv in library functions when run as setuid/setgid.
7441 New function OPENSSL_issetugid().
7444 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7445 due to incorrect handling of multi-threading:
7447 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7449 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7451 3. Count how many times MemCheck_off() has been called so that
7452 nested use can be treated correctly. This also avoids
7453 inband-signalling in the previous code (which relied on the
7454 assumption that thread ID 0 is impossible).
7457 *) Add "-rand" option also to s_client and s_server.
7460 *) Fix CPU detection on Irix 6.x.
7461 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7462 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7464 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7467 [This change does not apply to 0.9.7.]
7469 *) Use the cached encoding of an X509_NAME structure rather than
7470 copying it. This is apparently the reason for the libsafe "errors"
7471 but the code is actually correct.
7474 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7475 Bleichenbacher's DSA attack.
7476 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7477 to be set and top=0 forces the highest bit to be set; top=-1 is new
7478 and leaves the highest bit random.
7479 [Ulf Moeller, Bodo Moeller]
7481 *) In the NCONF_...-based implementations for CONF_... queries
7482 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7483 a temporary CONF structure with the data component set to NULL
7484 (which gives segmentation faults in lh_retrieve).
7485 Instead, use NULL for the CONF pointer in CONF_get_string and
7486 CONF_get_number (which may use environment variables) and directly
7487 return NULL from CONF_get_section.
7490 *) Fix potential buffer overrun for EBCDIC.
7493 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7494 keyUsage if basicConstraints absent for a CA.
7497 *) Make SMIME_write_PKCS7() write mail header values with a format that
7498 is more generally accepted (no spaces before the semicolon), since
7499 some programs can't parse those values properly otherwise. Also make
7500 sure BIO's that break lines after each write do not create invalid
7504 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7505 macros previously used would not encode an empty SEQUENCE OF
7506 and break the signature.
7508 [This change does not apply to 0.9.7.]
7510 *) Zero the premaster secret after deriving the master secret in
7514 *) Add some EVP_add_digest_alias registrations (as found in
7515 OpenSSL_add_all_digests()) to SSL_library_init()
7516 aka OpenSSL_add_ssl_algorithms(). This provides improved
7517 compatibility with peers using X.509 certificates
7518 with unconventional AlgorithmIdentifier OIDs.
7521 *) Fix for Irix with NO_ASM.
7522 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7524 *) ./config script fixes.
7525 [Ulf Moeller, Richard Levitte]
7527 *) Fix 'openssl passwd -1'.
7530 *) Change PKCS12_key_gen_asc() so it can cope with non null
7531 terminated strings whose length is passed in the passlen
7532 parameter, for example from PEM callbacks. This was done
7533 by adding an extra length parameter to asc2uni().
7534 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7536 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7537 call failed, free the DSA structure.
7540 *) Fix to uni2asc() to cope with zero length Unicode strings.
7541 These are present in some PKCS#12 files.
7544 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7545 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7546 when writing a 32767 byte record.
7547 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7549 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7550 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7552 (RSA objects have a reference count access to which is protected
7553 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7554 so they are meant to be shared between threads.)
7555 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7556 "Reddie, Steven" <Steven.Reddie@ca.com>]
7558 *) Fix a deadlock in CRYPTO_mem_leaks().
7561 *) Use better test patterns in bntest.
7564 *) rand_win.c fix for Borland C.
7567 *) BN_rshift bugfix for n == 0.
7570 *) Add a 'bctest' script that checks for some known 'bc' bugs
7571 so that 'make test' does not abort just because 'bc' is broken.
7574 *) Store verify_result within SSL_SESSION also for client side to
7575 avoid potential security hole. (Re-used sessions on the client side
7576 always resulted in verify_result==X509_V_OK, not using the original
7577 result of the server certificate verification.)
7580 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7581 SSL3_RT_APPLICATION_DATA, return 0.
7582 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7586 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7587 releases, have been re-implemented by renaming the previous
7588 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7589 and ssl3_read_internal, respectively, and adding 'peek' parameters
7590 to them. The new ssl[23]_{read,peek} functions are calls to
7591 ssl[23]_read_internal with the 'peek' flag set appropriately.
7592 A 'peek' parameter has also been added to ssl3_read_bytes, which
7593 does the actual work for ssl3_read_internal.
7596 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7597 the method-specific "init()" handler. Also clean up ex_data after
7598 calling the method-specific "finish()" handler. Previously, this was
7599 happening the other way round.
7602 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7603 The previous value, 12, was not always sufficient for BN_mod_exp().
7606 *) Make sure that shared libraries get the internal name engine with
7607 the full version number and not just 0. This should mark the
7608 shared libraries as not backward compatible. Of course, this should
7609 be changed again when we can guarantee backward binary compatibility.
7612 *) Fix typo in get_cert_by_subject() in by_dir.c
7613 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7615 *) Rework the system to generate shared libraries:
7617 - Make note of the expected extension for the shared libraries and
7618 if there is a need for symbolic links from for example libcrypto.so.0
7619 to libcrypto.so.0.9.7. There is extended info in Configure for
7622 - Make as few rebuilds of the shared libraries as possible.
7624 - Still avoid linking the OpenSSL programs with the shared libraries.
7626 - When installing, install the shared libraries separately from the
7630 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7632 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7633 and not in SSL_clear because the latter is also used by the
7634 accept/connect functions; previously, the settings made by
7635 SSL_set_read_ahead would be lost during the handshake.
7636 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7638 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7639 Previously, it would create entries for disableed algorithms no
7643 *) Added several new manual pages for SSL_* function.
7646 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7648 *) In ssl23_get_client_hello, generate an error message when faced
7649 with an initial SSL 3.0/TLS record that is too small to contain the
7650 first two bytes of the ClientHello message, i.e. client_version.
7651 (Note that this is a pathologic case that probably has never happened
7652 in real life.) The previous approach was to use the version number
7653 from the record header as a substitute; but our protocol choice
7654 should not depend on that one because it is not authenticated
7655 by the Finished messages.
7658 *) More robust randomness gathering functions for Windows.
7659 [Jeffrey Altman <jaltman@columbia.edu>]
7661 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7662 not set then we don't setup the error code for issuer check errors
7663 to avoid possibly overwriting other errors which the callback does
7664 handle. If an application does set the flag then we assume it knows
7665 what it is doing and can handle the new informational codes
7669 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7670 a general "ANY" type, as such it should be able to decode anything
7671 including tagged types. However it didn't check the class so it would
7672 wrongly interpret tagged types in the same way as their universal
7673 counterpart and unknown types were just rejected. Changed so that the
7674 tagged and unknown types are handled in the same way as a SEQUENCE:
7675 that is the encoding is stored intact. There is also a new type
7676 "V_ASN1_OTHER" which is used when the class is not universal, in this
7677 case we have no idea what the actual type is so we just lump them all
7681 *) On VMS, stdout may very well lead to a file that is written to
7682 in a record-oriented fashion. That means that every write() will
7683 write a separate record, which will be read separately by the
7684 programs trying to read from it. This can be very confusing.
7686 The solution is to put a BIO filter in the way that will buffer
7687 text until a linefeed is reached, and then write everything a
7688 line at a time, so every record written will be an actual line,
7689 not chunks of lines and not (usually doesn't happen, but I've
7690 seen it once) several lines in one record. BIO_f_linebuffer() is
7693 Currently, it's a VMS-only method, because that's where it has
7694 been tested well enough.
7697 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7698 it can return incorrect results.
7699 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7700 but it was in 0.9.6-beta[12].)
7703 *) Disable the check for content being present when verifying detached
7704 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7705 include zero length content when signing messages.
7708 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7709 BIO_ctrl (for BIO pairs).
7712 *) Add DSO method for VMS.
7715 *) Bug fix: Montgomery multiplication could produce results with the
7719 *) Add RPM specification openssl.spec and modify it to build three
7720 packages. The default package contains applications, application
7721 documentation and run-time libraries. The devel package contains
7722 include files, static libraries and function documentation. The
7723 doc package contains the contents of the doc directory. The original
7724 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7727 *) Add a large number of documentation files for many SSL routines.
7728 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7730 *) Add a configuration entry for Sony News 4.
7731 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7733 *) Don't set the two most significant bits to one when generating a
7734 random number < q in the DSA library.
7737 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7738 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7739 the underlying transport is blocking) if a handshake took place.
7740 (The default behaviour is needed by applications such as s_client
7741 and s_server that use select() to determine when to use SSL_read;
7742 but for applications that know in advance when to expect data, it
7743 just makes things more complicated.)
7746 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7750 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7751 work better on such systems.
7752 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7754 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7755 Update PKCS12_parse() so it copies the friendlyName and the
7756 keyid to the certificates aux info.
7759 *) Fix bug in PKCS7_verify() which caused an infinite loop
7760 if there was more than one signature.
7761 [Sven Uszpelkat <su@celocom.de>]
7763 *) Major change in util/mkdef.pl to include extra information
7764 about each symbol, as well as presenting variables as well
7765 as functions. This change means that there's n more need
7766 to rebuild the .num files when some algorithms are excluded.
7769 *) Allow the verify time to be set by an application,
7770 rather than always using the current time.
7773 *) Phase 2 verify code reorganisation. The certificate
7774 verify code now looks up an issuer certificate by a
7775 number of criteria: subject name, authority key id
7776 and key usage. It also verifies self signed certificates
7777 by the same criteria. The main comparison function is
7778 X509_check_issued() which performs these checks.
7780 Lot of changes were necessary in order to support this
7781 without completely rewriting the lookup code.
7783 Authority and subject key identifier are now cached.
7785 The LHASH 'certs' is X509_STORE has now been replaced
7786 by a STACK_OF(X509_OBJECT). This is mainly because an
7787 LHASH can't store or retrieve multiple objects with
7788 the same hash value.
7790 As a result various functions (which were all internal
7791 use only) have changed to handle the new X509_STORE
7792 structure. This will break anything that messed round
7793 with X509_STORE internally.
7795 The functions X509_STORE_add_cert() now checks for an
7796 exact match, rather than just subject name.
7798 The X509_STORE API doesn't directly support the retrieval
7799 of multiple certificates matching a given criteria, however
7800 this can be worked round by performing a lookup first
7801 (which will fill the cache with candidate certificates)
7802 and then examining the cache for matches. This is probably
7803 the best we can do without throwing out X509_LOOKUP
7804 entirely (maybe later...).
7806 The X509_VERIFY_CTX structure has been enhanced considerably.
7808 All certificate lookup operations now go via a get_issuer()
7809 callback. Although this currently uses an X509_STORE it
7810 can be replaced by custom lookups. This is a simple way
7811 to bypass the X509_STORE hackery necessary to make this
7812 work and makes it possible to use more efficient techniques
7813 in future. A very simple version which uses a simple
7814 STACK for its trusted certificate store is also provided
7815 using X509_STORE_CTX_trusted_stack().
7817 The verify_cb() and verify() callbacks now have equivalents
7818 in the X509_STORE_CTX structure.
7820 X509_STORE_CTX also has a 'flags' field which can be used
7821 to customise the verify behaviour.
7824 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7825 excludes S/MIME capabilities.
7828 *) When a certificate request is read in keep a copy of the
7829 original encoding of the signed data and use it when outputing
7830 again. Signatures then use the original encoding rather than
7831 a decoded, encoded version which may cause problems if the
7832 request is improperly encoded.
7835 *) For consistency with other BIO_puts implementations, call
7836 buffer_write(b, ...) directly in buffer_puts instead of calling
7839 In BIO_puts, increment b->num_write as in BIO_write.
7840 [Peter.Sylvester@EdelWeb.fr]
7842 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7843 BN_zero, we may not return a BIGNUM with an array consisting of
7847 *) Avoid calling abort() from within the library when problems are
7848 detected, except if preprocessor symbols have been defined
7849 (such as REF_CHECK, BN_DEBUG etc.).
7852 *) New openssl application 'rsautl'. This utility can be
7853 used for low level RSA operations. DER public key
7854 BIO/fp routines also added.
7857 *) New Configure entry and patches for compiling on QNX 4.
7858 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7860 *) A demo state-machine implementation was sponsored by
7861 Nuron (http://www.nuron.com/) and is now available in
7862 demos/state_machine.
7865 *) New options added to the 'dgst' utility for signature
7866 generation and verification.
7869 *) Unrecognized PKCS#7 content types are now handled via a
7870 catch all ASN1_TYPE structure. This allows unsupported
7871 types to be stored as a "blob" and an application can
7872 encode and decode it manually.
7875 *) Fix various signed/unsigned issues to make a_strex.c
7877 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7879 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7880 length if passed a buffer. ASN1_INTEGER_to_BN failed
7881 if passed a NULL BN and its argument was negative.
7882 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7884 *) Modification to PKCS#7 encoding routines to output definite
7885 length encoding. Since currently the whole structures are in
7886 memory there's not real point in using indefinite length
7887 constructed encoding. However if OpenSSL is compiled with
7888 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7891 *) Added BIO_vprintf() and BIO_vsnprintf().
7894 *) Added more prefixes to parse for in the the strings written
7895 through a logging bio, to cover all the levels that are available
7896 through syslog. The prefixes are now:
7898 PANIC, EMERG, EMR => LOG_EMERG
7899 ALERT, ALR => LOG_ALERT
7900 CRIT, CRI => LOG_CRIT
7901 ERROR, ERR => LOG_ERR
7902 WARNING, WARN, WAR => LOG_WARNING
7903 NOTICE, NOTE, NOT => LOG_NOTICE
7904 INFO, INF => LOG_INFO
7905 DEBUG, DBG => LOG_DEBUG
7907 and as before, if none of those prefixes are present at the
7908 beginning of the string, LOG_ERR is chosen.
7910 On Win32, the LOG_* levels are mapped according to this:
7912 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7913 LOG_WARNING => EVENTLOG_WARNING_TYPE
7914 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7918 *) Made it possible to reconfigure with just the configuration
7919 argument "reconf" or "reconfigure". The command line arguments
7920 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7921 and are retrieved from there when reconfiguring.
7925 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7927 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7930 *) The obj_dat.pl script was messing up the sorting of object
7931 names. The reason was that it compared the quoted version
7932 of strings as a result "OCSP" > "OCSP Signing" because
7933 " > SPACE. Changed script to store unquoted versions of
7934 names and add quotes on output. It was also omitting some
7935 names from the lookup table if they were given a default
7936 value (that is if SN is missing it is given the same
7937 value as LN and vice versa), these are now added on the
7938 grounds that if an object has a name we should be able to
7939 look it up. Finally added warning output when duplicate
7940 short or long names are found.
7943 *) Changes needed for Tandem NSK.
7944 [Scott Uroff <scott@xypro.com>]
7946 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7947 RSA_padding_check_SSLv23(), special padding was never detected
7948 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7949 version rollback attacks was not effective.
7951 In s23_clnt.c, don't use special rollback-attack detection padding
7952 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7953 client; similarly, in s23_srvr.c, don't do the rollback check if
7954 SSL 2.0 is the only protocol enabled in the server.
7957 *) Make it possible to get hexdumps of unprintable data with 'openssl
7958 asn1parse'. By implication, the functions ASN1_parse_dump() and
7959 BIO_dump_indent() are added.
7962 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7963 these print out strings and name structures based on various
7964 flags including RFC2253 support and proper handling of
7965 multibyte characters. Added options to the 'x509' utility
7966 to allow the various flags to be set.
7969 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7970 Also change the functions X509_cmp_current_time() and
7971 X509_gmtime_adj() work with an ASN1_TIME structure,
7972 this will enable certificates using GeneralizedTime in validity
7973 dates to be checked.
7976 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7977 negative public key encodings) on by default,
7978 NO_NEG_PUBKEY_BUG can be set to disable it.
7981 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7982 content octets. An i2c_ASN1_OBJECT is unnecessary because
7983 the encoding can be trivially obtained from the structure.
7986 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7987 not read locks (CRYPTO_r_[un]lock).
7990 *) A first attempt at creating official support for shared
7991 libraries through configuration. I've kept it so the
7992 default is static libraries only, and the OpenSSL programs
7993 are always statically linked for now, but there are
7994 preparations for dynamic linking in place.
7995 This has been tested on Linux and Tru64.
7998 *) Randomness polling function for Win9x, as described in:
7999 Peter Gutmann, Software Generation of Practically Strong
8003 *) Fix so PRNG is seeded in req if using an already existing
8007 *) New options to smime application. -inform and -outform
8008 allow alternative formats for the S/MIME message including
8009 PEM and DER. The -content option allows the content to be
8010 specified separately. This should allow things like Netscape
8011 form signing output easier to verify.
8014 *) Fix the ASN1 encoding of tags using the 'long form'.
8017 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8018 STRING types. These convert content octets to and from the
8019 underlying type. The actual tag and length octets are
8020 already assumed to have been read in and checked. These
8021 are needed because all other string types have virtually
8022 identical handling apart from the tag. By having versions
8023 of the ASN1 functions that just operate on content octets
8024 IMPLICIT tagging can be handled properly. It also allows
8025 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8026 and ASN1_INTEGER are identical apart from the tag.
8029 *) Change the handling of OID objects as follows:
8031 - New object identifiers are inserted in objects.txt, following
8032 the syntax given in objects.README.
8033 - objects.pl is used to process obj_mac.num and create a new
8035 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8038 This is currently kind of a hack, and the perl code in objects.pl
8039 isn't very elegant, but it works as I intended. The simplest way
8040 to check that it worked correctly is to look in obj_dat.h and
8041 check the array nid_objs and make sure the objects haven't moved
8042 around (this is important!). Additions are OK, as well as
8043 consistent name changes.
8046 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8049 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8050 The given file adds to whatever has already been seeded into the
8051 random pool through the RANDFILE configuration file option or
8052 environment variable, or the default random state file.
8055 *) mkstack.pl now sorts each macro group into lexical order.
8056 Previously the output order depended on the order the files
8057 appeared in the directory, resulting in needless rewriting
8061 *) Patches to make OpenSSL compile under Win32 again. Mostly
8062 work arounds for the VC++ problem that it treats func() as
8063 func(void). Also stripped out the parts of mkdef.pl that
8064 added extra typesafe functions: these no longer exist.
8067 *) Reorganisation of the stack code. The macros are now all
8068 collected in safestack.h . Each macro is defined in terms of
8069 a "stack macro" of the form SKM_<name>(type, a, b). The
8070 DEBUG_SAFESTACK is now handled in terms of function casts,
8071 this has the advantage of retaining type safety without the
8072 use of additional functions. If DEBUG_SAFESTACK is not defined
8073 then the non typesafe macros are used instead. Also modified the
8074 mkstack.pl script to handle the new form. Needs testing to see
8075 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8076 the default if no major problems. Similar behaviour for ASN1_SET_OF
8077 and PKCS12_STACK_OF.
8080 *) When some versions of IIS use the 'NET' form of private key the
8081 key derivation algorithm is different. Normally MD5(password) is
8082 used as a 128 bit RC4 key. In the modified case
8083 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8084 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8085 as the old Netscape_RSA functions except they have an additional
8086 'sgckey' parameter which uses the modified algorithm. Also added
8087 an -sgckey command line option to the rsa utility. Thanks to
8088 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8089 algorithm to openssl-dev.
8092 *) The evp_local.h macros were using 'c.##kname' which resulted in
8093 invalid expansion on some systems (SCO 5.0.5 for example).
8094 Corrected to 'c.kname'.
8095 [Phillip Porch <root@theporch.com>]
8097 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8098 a STACK of email addresses from a certificate or request, these look
8099 in the subject name and the subject alternative name extensions and
8100 omit any duplicate addresses.
8103 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8104 This makes DSA verification about 2 % faster.
8107 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8108 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8109 plus overhead for 1024 bit moduli).
8110 This makes exponentiations about 0.5 % faster for 1024 bit
8111 exponents (as measured by "openssl speed rsa2048").
8114 *) Rename memory handling macros to avoid conflicts with other
8116 Malloc => OPENSSL_malloc
8117 Malloc_locked => OPENSSL_malloc_locked
8118 Realloc => OPENSSL_realloc
8119 Free => OPENSSL_free
8122 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8123 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8126 *) CygWin32 support.
8127 [John Jarvie <jjarvie@newsguy.com>]
8129 *) The type-safe stack code has been rejigged. It is now only compiled
8130 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8131 by default all type-specific stack functions are "#define"d back to
8132 standard stack functions. This results in more streamlined output
8133 but retains the type-safety checking possibilities of the original
8137 *) The STACK code has been cleaned up, and certain type declarations
8138 that didn't make a lot of sense have been brought in line. This has
8139 also involved a cleanup of sorts in safestack.h to more correctly
8140 map type-safe stack functions onto their plain stack counterparts.
8141 This work has also resulted in a variety of "const"ifications of
8142 lots of the code, especially "_cmp" operations which should normally
8143 be prototyped with "const" parameters anyway.
8146 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8147 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8148 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8149 where all of 'md' is used each time the PRNG is used, but 'state'
8150 is used only indexed by a cyclic counter. As entropy may not be
8151 well distributed from the beginning, 'md' is important as a
8152 chaining variable. However, the output function chains only half
8153 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8154 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8155 in all of 'state' being rewritten, with the new values depending
8156 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8159 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8160 the handshake is continued after ssl_verify_cert_chain();
8161 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8162 can lead to 'unexplainable' connection aborts later.
8163 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8165 *) Major EVP API cipher revision.
8166 Add hooks for extra EVP features. This allows various cipher
8167 parameters to be set in the EVP interface. Support added for variable
8168 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8169 setting of RC2 and RC5 parameters.
8171 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8174 Remove lots of duplicated code from the EVP library. For example *every*
8175 cipher init() function handles the 'iv' in the same way according to the
8176 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8177 for CFB and OFB modes they zero ctx->num.
8179 New functionality allows removal of S/MIME code RC2 hack.
8181 Most of the routines have the same form and so can be declared in terms
8184 By shifting this to the top level EVP_CipherInit() it can be removed from
8185 all individual ciphers. If the cipher wants to handle IVs or keys
8186 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8189 Change lots of functions like EVP_EncryptUpdate() to now return a
8190 value: although software versions of the algorithms cannot fail
8191 any installed hardware versions can.
8194 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8195 this option is set, tolerate broken clients that send the negotiated
8196 protocol version number instead of the requested protocol version
8200 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8201 i.e. non-zero for export ciphersuites, zero otherwise.
8202 Previous versions had this flag inverted, inconsistent with
8203 rsa_tmp_cb (..._TMP_RSA_CB).
8204 [Bodo Moeller; problem reported by Amit Chopra]
8206 *) Add missing DSA library text string. Work around for some IIS
8207 key files with invalid SEQUENCE encoding.
8210 *) Add a document (doc/standards.txt) that list all kinds of standards
8211 and so on that are implemented in OpenSSL.
8214 *) Enhance c_rehash script. Old version would mishandle certificates
8215 with the same subject name hash and wouldn't handle CRLs at all.
8216 Added -fingerprint option to crl utility, to support new c_rehash
8220 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8223 *) Fix for SSL server purpose checking. Server checking was
8224 rejecting certificates which had extended key usage present
8225 but no ssl client purpose.
8226 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8228 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8229 is a little unclear about how a blank password is handled.
8230 Since the password in encoded as a BMPString with terminating
8231 double NULL a zero length password would end up as just the
8232 double NULL. However no password at all is different and is
8233 handled differently in the PKCS#12 key generation code. NS
8234 treats a blank password as zero length. MSIE treats it as no
8235 password on export: but it will try both on import. We now do
8236 the same: PKCS12_parse() tries zero length and no password if
8237 the password is set to "" or NULL (NULL is now a valid password:
8238 it wasn't before) as does the pkcs12 application.
8241 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8242 perror when PEM_read_bio_X509_REQ fails, the error message must
8243 be obtained from the error queue.
8246 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8247 it in ERR_remove_state if appropriate, and change ERR_get_state
8248 accordingly to avoid race conditions (this is necessary because
8249 thread_hash is no longer constant once set).
8252 *) Bugfix for linux-elf makefile.one.
8255 *) RSA_get_default_method() will now cause a default
8256 RSA_METHOD to be chosen if one doesn't exist already.
8257 Previously this was only set during a call to RSA_new()
8258 or RSA_new_method(NULL) meaning it was possible for
8259 RSA_get_default_method() to return NULL.
8262 *) Added native name translation to the existing DSO code
8263 that will convert (if the flag to do so is set) filenames
8264 that are sufficiently small and have no path information
8265 into a canonical native form. Eg. "blah" converted to
8266 "libblah.so" or "blah.dll" etc.
8269 *) New function ERR_error_string_n(e, buf, len) which is like
8270 ERR_error_string(e, buf), but writes at most 'len' bytes
8271 including the 0 terminator. For ERR_error_string_n, 'buf'
8273 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8275 *) CONF library reworked to become more general. A new CONF
8276 configuration file reader "class" is implemented as well as a
8277 new functions (NCONF_*, for "New CONF") to handle it. The now
8278 old CONF_* functions are still there, but are reimplemented to
8279 work in terms of the new functions. Also, a set of functions
8280 to handle the internal storage of the configuration data is
8281 provided to make it easier to write new configuration file
8282 reader "classes" (I can definitely see something reading a
8283 configuration file in XML format, for example), called _CONF_*,
8284 or "the configuration storage API"...
8286 The new configuration file reading functions are:
8288 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8289 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8291 NCONF_default, NCONF_WIN32
8293 NCONF_dump_fp, NCONF_dump_bio
8295 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8296 NCONF_new creates a new CONF object. This works in the same way
8297 as other interfaces in OpenSSL, like the BIO interface.
8298 NCONF_dump_* dump the internal storage of the configuration file,
8299 which is useful for debugging. All other functions take the same
8300 arguments as the old CONF_* functions wth the exception of the
8301 first that must be a `CONF *' instead of a `LHASH *'.
8303 To make it easer to use the new classes with the old CONF_* functions,
8304 the function CONF_set_default_method is provided.
8307 *) Add '-tls1' option to 'openssl ciphers', which was already
8308 mentioned in the documentation but had not been implemented.
8309 (This option is not yet really useful because even the additional
8310 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8313 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8314 OpenSSL-based applications) load shared libraries and bind to
8315 them in a portable way.
8316 [Geoff Thorpe, with contributions from Richard Levitte]
8318 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8320 *) Make sure _lrotl and _lrotr are only used with MSVC.
8322 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8323 (the default implementation of RAND_status).
8325 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8326 to '-clrext' (= clear extensions), as intended and documented.
8327 [Bodo Moeller; inconsistency pointed out by Michael Attili
8330 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8331 was larger than the MD block size.
8332 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8334 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8335 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8336 using the passed key: if the passed key was a private key the result
8337 of X509_print(), for example, would be to print out all the private key
8341 *) des_quad_cksum() byte order bug fix.
8342 [Ulf Möller, using the problem description in krb4-0.9.7, where
8343 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8345 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8347 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8349 *) For easily testing in shell scripts whether some command
8350 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8351 returns with exit code 0 iff no command of the given name is available.
8352 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8353 the output goes to stdout and nothing is printed to stderr.
8354 Additional arguments are always ignored.
8356 Since for each cipher there is a command of the same name,
8357 the 'no-cipher' compilation switches can be tested this way.
8359 ('openssl no-XXX' is not able to detect pseudo-commands such
8360 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8363 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8366 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8367 is set; it will be thrown away anyway because each handshake creates
8369 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8370 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8371 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8372 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8375 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8376 'Q' and 'R' lose their special meanings (quit/renegotiate).
8377 This is part of what -quiet does; unlike -quiet, -ign_eof
8378 does not suppress any output.
8381 *) Add compatibility options to the purpose and trust code. The
8382 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8383 accepts a certificate or CA, this was the previous behaviour,
8384 with all the associated security issues.
8386 X509_TRUST_COMPAT is the old trust behaviour: only and
8387 automatically trust self signed roots in certificate store. A
8388 new trust setting X509_TRUST_DEFAULT is used to specify that
8389 a purpose has no associated trust setting and it should instead
8390 use the value in the default purpose.
8393 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8394 and fix a memory leak.
8397 *) In util/mkerr.pl (which implements 'make errors'), preserve
8398 reason strings from the previous version of the .c file, as
8399 the default to have only downcase letters (and digits) in
8400 automatically generated reasons codes is not always appropriate.
8403 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8404 using strerror. Previously, ERR_reason_error_string() returned
8405 library names as reason strings for SYSerr; but SYSerr is a special
8406 case where small numbers are errno values, not library numbers.
8409 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8410 converts DSA parameters into DH parameters. (When creating parameters,
8411 DSA_generate_parameters is used.)
8414 *) Include 'length' (recommended exponent length) in C code generated
8415 by 'openssl dhparam -C'.
8418 *) The second argument to set_label in perlasm was already being used
8419 so couldn't be used as a "file scope" flag. Moved to third argument
8423 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8424 instead of RAND_bytes for encryption IVs and salts.
8427 *) Include RAND_status() into RAND_METHOD instead of implementing
8428 it only for md_rand.c Otherwise replacing the PRNG by calling
8429 RAND_set_rand_method would be impossible.
8432 *) Don't let DSA_generate_key() enter an infinite loop if the random
8433 number generation fails.
8436 *) New 'rand' application for creating pseudo-random output.
8439 *) Added configuration support for Linux/IA64
8440 [Rolf Haberrecker <rolf@suse.de>]
8442 *) Assembler module support for Mingw32.
8445 *) Shared library support for HPUX (in shlib/).
8446 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8448 *) Shared library support for Solaris gcc.
8449 [Lutz Behnke <behnke@trustcenter.de>]
8451 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8453 *) PKCS7_encrypt() was adding text MIME headers twice because they
8454 were added manually and by SMIME_crlf_copy().
8457 *) In bntest.c don't call BN_rand with zero bits argument.
8458 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8460 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8461 case was implemented. This caused BN_div_recp() to fail occasionally.
8464 *) Add an optional second argument to the set_label() in the perl
8465 assembly language builder. If this argument exists and is set
8466 to 1 it signals that the assembler should use a symbol whose
8467 scope is the entire file, not just the current function. This
8468 is needed with MASM which uses the format label:: for this scope.
8469 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8471 *) Change the ASN1 types so they are typedefs by default. Before
8472 almost all types were #define'd to ASN1_STRING which was causing
8473 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8477 *) Change names of new functions to the new get1/get0 naming
8478 convention: After 'get1', the caller owns a reference count
8479 and has to call ..._free; 'get0' returns a pointer to some
8480 data structure without incrementing reference counters.
8481 (Some of the existing 'get' functions increment a reference
8482 counter, some don't.)
8483 Similarly, 'set1' and 'add1' functions increase reference
8484 counters or duplicate objects.
8487 *) Allow for the possibility of temp RSA key generation failure:
8488 the code used to assume it always worked and crashed on failure.
8491 *) Fix potential buffer overrun problem in BIO_printf().
8492 [Ulf Möller, using public domain code by Patrick Powell; problem
8493 pointed out by David Sacerdote <das33@cornell.edu>]
8495 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8496 RAND_egd() and RAND_status(). In the command line application,
8497 the EGD socket can be specified like a seed file using RANDFILE
8501 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8502 Some CAs (e.g. Verisign) distribute certificates in this form.
8505 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8506 list to exclude them. This means that no special compilation option
8507 is needed to use anonymous DH: it just needs to be included in the
8511 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8512 EVP_MD_type. The old functionality is available in a new macro called
8513 EVP_MD_md(). Change code that uses it and update docs.
8516 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8517 where the 'void *' argument is replaced by a function pointer argument.
8518 Previously 'void *' was abused to point to functions, which works on
8519 many platforms, but is not correct. As these functions are usually
8520 called by macros defined in OpenSSL header files, most source code
8521 should work without changes.
8524 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8525 sections with information on -D... compiler switches used for
8526 compiling the library so that applications can see them. To enable
8527 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8528 must be defined. E.g.,
8529 #define OPENSSL_ALGORITHM_DEFINES
8530 #include <openssl/opensslconf.h>
8531 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8532 [Richard Levitte, Ulf and Bodo Möller]
8534 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8538 *) Change the 'other' type in certificate aux info to a STACK_OF
8539 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8540 the required ASN1 format: arbitrary types determined by an OID.
8543 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8544 argument to 'req'. This is not because the function is newer or
8545 better than others it just uses the work 'NEW' in the certificate
8546 request header lines. Some software needs this.
8549 *) Reorganise password command line arguments: now passwords can be
8550 obtained from various sources. Delete the PEM_cb function and make
8551 it the default behaviour: i.e. if the callback is NULL and the
8552 usrdata argument is not NULL interpret it as a null terminated pass
8553 phrase. If usrdata and the callback are NULL then the pass phrase
8554 is prompted for as usual.
8557 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8558 the support is automatically enabled. The resulting binaries will
8559 autodetect the card and use it if present.
8560 [Ben Laurie and Compaq Inc.]
8562 *) Work around for Netscape hang bug. This sends certificate request
8563 and server done in one record. Since this is perfectly legal in the
8564 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8565 the bugs/SSLv3 entry for more info.
8568 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8571 *) Add -rand argument to smime and pkcs12 applications and read/write
8575 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8578 *) Add command line password options to the remaining applications.
8581 *) Bug fix for BN_div_recp() for numerators with an even number of
8585 *) More tests in bntest.c, and changed test_bn output.
8588 *) ./config recognizes MacOS X now.
8591 *) Bug fix for BN_div() when the first words of num and divsor are
8592 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8595 *) Add support for various broken PKCS#8 formats, and command line
8596 options to produce them.
8599 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8600 get temporary BIGNUMs from a BN_CTX.
8603 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8607 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8608 include a #define from the old name to the new. The original intent
8609 was that statically linked binaries could for example just call
8610 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8611 link with digests. This never worked because SSLeay_add_all_digests()
8612 and SSLeay_add_all_ciphers() were in the same source file so calling
8613 one would link with the other. They are now in separate source files.
8616 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8619 *) Use a less unusual form of the Miller-Rabin primality test (it used
8620 a binary algorithm for exponentiation integrated into the Miller-Rabin
8621 loop, our standard modexp algorithms are faster).
8624 *) Support for the EBCDIC character set completed.
8625 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8627 *) Source code cleanups: use const where appropriate, eliminate casts,
8628 use void * instead of char * in lhash.
8631 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8632 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8633 this the server could overwrite ephemeral keys that the client
8637 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8638 using 50 iterations of the Rabin-Miller test.
8640 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8641 iterations of the Rabin-Miller test as required by the appendix
8642 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8643 As BN_is_prime_fasttest includes trial division, DSA parameter
8644 generation becomes much faster.
8646 This implies a change for the callback functions in DSA_is_prime
8647 and DSA_generate_parameters: The callback function is called once
8648 for each positive witness in the Rabin-Miller test, not just
8649 occasionally in the inner loop; and the parameters to the
8650 callback function now provide an iteration count for the outer
8651 loop rather than for the current invocation of the inner loop.
8652 DSA_generate_parameters additionally can call the callback
8653 function with an 'iteration count' of -1, meaning that a
8654 candidate has passed the trial division test (when q is generated
8655 from an application-provided seed, trial division is skipped).
8658 *) New function BN_is_prime_fasttest that optionally does trial
8659 division before starting the Rabin-Miller test and has
8660 an additional BN_CTX * argument (whereas BN_is_prime always
8661 has to allocate at least one BN_CTX).
8662 'callback(1, -1, cb_arg)' is called when a number has passed the
8663 trial division stage.
8666 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8670 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8673 *) New function BN_pseudo_rand().
8676 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8677 bignum version of BN_from_montgomery() with the working code from
8678 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8682 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8683 made it impossible to use the same SSL_SESSION data structure in
8684 SSL2 clients in multiple threads.
8687 *) The return value of RAND_load_file() no longer counts bytes obtained
8688 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8689 to seed the PRNG (previously an explicit byte count was required).
8690 [Ulf Möller, Bodo Möller]
8692 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8693 used (char *) instead of (void *) and had casts all over the place.
8696 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8699 *) Retain source code compatibility for BN_prime_checks macro:
8700 BN_is_prime(..., BN_prime_checks, ...) now uses
8701 BN_prime_checks_for_size to determine the appropriate number of
8702 Rabin-Miller iterations.
8705 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8706 DH_CHECK_P_NOT_SAFE_PRIME.
8707 (Check if this is true? OpenPGP calls them "strong".)
8710 *) Merge the functionality of "dh" and "gendh" programs into a new program
8711 "dhparam". The old programs are retained for now but will handle DH keys
8712 (instead of parameters) in future.
8715 *) Make the ciphers, s_server and s_client programs check the return values
8716 when a new cipher list is set.
8719 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8720 ciphers. Before when the 56bit ciphers were enabled the sorting was
8723 The syntax for the cipher sorting has been extended to support sorting by
8724 cipher-strength (using the strength_bits hard coded in the tables).
8725 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8727 Fix a bug in the cipher-command parser: when supplying a cipher command
8728 string with an "undefined" symbol (neither command nor alphanumeric
8729 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8730 an error is flagged.
8732 Due to the strength-sorting extension, the code of the
8733 ssl_create_cipher_list() function was completely rearranged. I hope that
8734 the readability was also increased :-)
8735 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8737 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8738 for the first serial number and places 2 in the serial number file. This
8739 avoids problems when the root CA is created with serial number zero and
8740 the first user certificate has the same issuer name and serial number
8744 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8745 the new code. Add documentation for this stuff.
8748 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8749 X509_*() to X509at_*() on the grounds that they don't handle X509
8750 structures and behave in an analogous way to the X509v3 functions:
8751 they shouldn't be called directly but wrapper functions should be used
8754 So we also now have some wrapper functions that call the X509at functions
8755 when passed certificate requests. (TO DO: similar things can be done with
8756 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8757 things. Some of these need some d2i or i2d and print functionality
8758 because they handle more complex structures.)
8761 *) Add missing #ifndefs that caused missing symbols when building libssl
8762 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8763 NO_RSA in ssl/s2*.c.
8764 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8766 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8767 has a return value which indicates the quality of the random data
8768 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8769 error queue. New function RAND_pseudo_bytes() generates output that is
8770 guaranteed to be unique but not unpredictable. RAND_add is like
8771 RAND_seed, but takes an extra argument for an entropy estimate
8772 (RAND_seed always assumes full entropy).
8775 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8776 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8777 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8778 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8779 false-positive rate of at most 2^-80 for random input.
8782 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8785 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8786 in the 0.9.5 release), this returns the chain
8787 from an X509_CTX structure with a dup of the stack and all
8788 the X509 reference counts upped: so the stack will exist
8789 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8792 Also make SSL_SESSION_print() print out the verify return
8796 *) Add manpage for the pkcs12 command. Also change the default
8797 behaviour so MAC iteration counts are used unless the new
8798 -nomaciter option is used. This improves file security and
8799 only older versions of MSIE (4.0 for example) need it.
8802 *) Honor the no-xxx Configure options when creating .DEF files.
8805 *) Add PKCS#10 attributes to field table: challengePassword,
8806 unstructuredName and unstructuredAddress. These are taken from
8807 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8808 international characters are used.
8810 More changes to X509_ATTRIBUTE code: allow the setting of types
8811 based on strings. Remove the 'loc' parameter when adding
8812 attributes because these will be a SET OF encoding which is sorted
8816 *) Initial changes to the 'req' utility to allow request generation
8817 automation. This will allow an application to just generate a template
8818 file containing all the field values and have req construct the
8821 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8822 used all over the place including certificate requests and PKCS#7
8823 structures. They are currently handled manually where necessary with
8824 some primitive wrappers for PKCS#7. The new functions behave in a
8825 manner analogous to the X509 extension functions: they allow
8826 attributes to be looked up by NID and added.
8828 Later something similar to the X509V3 code would be desirable to
8829 automatically handle the encoding, decoding and printing of the
8830 more complex types. The string types like challengePassword can
8831 be handled by the string table functions.
8833 Also modified the multi byte string table handling. Now there is
8834 a 'global mask' which masks out certain types. The table itself
8835 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8836 is useful when for example there is only one permissible type
8837 (as in countryName) and using the mask might result in no valid
8841 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8842 SSL_get_peer_finished to allow applications to obtain the latest
8843 Finished messages sent to the peer or expected from the peer,
8844 respectively. (SSL_get_peer_finished is usually the Finished message
8845 actually received from the peer, otherwise the protocol will be aborted.)
8847 As the Finished message are message digests of the complete handshake
8848 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8849 be used for external authentication procedures when the authentication
8850 provided by SSL/TLS is not desired or is not enough.
8853 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8854 the host supports BWX extension and if Compaq C is present on the
8855 $PATH. Just exploiting of the BWX extension results in 20-30%
8856 performance kick for some algorithms, e.g. DES and RC4 to mention
8857 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8861 *) Add support for MS "fast SGC". This is arguably a violation of the
8862 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8863 weak crypto and after checking the certificate is SGC a second one
8864 with strong crypto. MS SGC stops the first handshake after receiving
8865 the server certificate message and sends a second client hello. Since
8866 a server will typically do all the time consuming operations before
8867 expecting any further messages from the client (server key exchange
8868 is the most expensive) there is little difference between the two.
8870 To get OpenSSL to support MS SGC we have to permit a second client
8871 hello message after we have sent server done. In addition we have to
8872 reset the MAC if we do get this second client hello.
8875 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8876 if a DER encoded private key is RSA or DSA traditional format. Changed
8877 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8878 format DER encoded private key. Newer code should use PKCS#8 format which
8879 has the key type encoded in the ASN1 structure. Added DER private key
8880 support to pkcs8 application.
8883 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8884 ciphersuites has been selected (as required by the SSL 3/TLS 1
8885 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8886 is set, we interpret this as a request to violate the specification
8887 (the worst that can happen is a handshake failure, and 'correct'
8888 behaviour would result in a handshake failure anyway).
8891 *) In SSL_CTX_add_session, take into account that there might be multiple
8892 SSL_SESSION structures with the same session ID (e.g. when two threads
8893 concurrently obtain them from an external cache).
8894 The internal cache can handle only one SSL_SESSION with a given ID,
8895 so if there's a conflict, we now throw out the old one to achieve
8899 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8900 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8901 some routines that use cipher OIDs: some ciphers do not have OIDs
8902 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8906 *) Simplify the trust setting structure and code. Now we just have
8907 two sequences of OIDs for trusted and rejected settings. These will
8908 typically have values the same as the extended key usage extension
8909 and any application specific purposes.
8911 The trust checking code now has a default behaviour: it will just
8912 check for an object with the same NID as the passed id. Functions can
8913 be provided to override either the default behaviour or the behaviour
8914 for a given id. SSL client, server and email already have functions
8915 in place for compatibility: they check the NID and also return "trusted"
8916 if the certificate is self signed.
8919 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8920 traditional format into an EVP_PKEY structure.
8923 *) Add a password callback function PEM_cb() which either prompts for
8924 a password if usr_data is NULL or otherwise assumes it is a null
8925 terminated password. Allow passwords to be passed on command line
8926 environment or config files in a few more utilities.
8929 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8930 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8931 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8932 Update documentation.
8935 *) Support for ASN1 "NULL" type. This could be handled before by using
8936 ASN1_TYPE but there wasn't any function that would try to read a NULL
8937 and produce an error if it couldn't. For compatibility we also have
8938 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8939 don't allocate anything because they don't need to.
8942 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8944 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8946 *) Rebuild of the memory allocation routines used by OpenSSL code and
8947 possibly others as well. The purpose is to make an interface that
8948 provide hooks so anyone can build a separate set of allocation and
8949 deallocation routines to be used by OpenSSL, for example memory
8950 pool implementations, or something else, which was previously hard
8951 since Malloc(), Realloc() and Free() were defined as macros having
8952 the values malloc, realloc and free, respectively (except for Win32
8953 compilations). The same is provided for memory debugging code.
8954 OpenSSL already comes with functionality to find memory leaks, but
8955 this gives people a chance to debug other memory problems.
8957 With these changes, a new set of functions and macros have appeared:
8959 CRYPTO_set_mem_debug_functions() [F]
8960 CRYPTO_get_mem_debug_functions() [F]
8961 CRYPTO_dbg_set_options() [F]
8962 CRYPTO_dbg_get_options() [F]
8963 CRYPTO_malloc_debug_init() [M]
8965 The memory debug functions are NULL by default, unless the library
8966 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8967 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8968 gives the standard debugging functions that come with OpenSSL) or
8969 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8970 provided by the library user) must be used. When the standard
8971 debugging functions are used, CRYPTO_dbg_set_options can be used to
8972 request additional information:
8973 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8974 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8976 Also, things like CRYPTO_set_mem_functions will always give the
8977 expected result (the new set of functions is used for allocation
8978 and deallocation) at all times, regardless of platform and compiler
8981 To finish it up, some functions that were never use in any other
8982 way than through macros have a new API and new semantic:
8985 CRYPTO_dbg_realloc()
8988 All macros of value have retained their old syntax.
8989 [Richard Levitte and Bodo Moeller]
8991 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8992 ordering of SMIMECapabilities wasn't in "strength order" and there
8993 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8997 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8998 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8999 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9001 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9002 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9003 functionality to handle multipart/signed properly) and a utility
9004 called 'smime' to call all this stuff. This is based on code I
9005 originally wrote for Celo who have kindly allowed it to be
9006 included in OpenSSL.
9009 *) Add variants des_set_key_checked and des_set_key_unchecked of
9010 des_set_key (aka des_key_sched). Global variable des_check_key
9011 decides which of these is called by des_set_key; this way
9012 des_check_key behaves as it always did, but applications and
9013 the library itself, which was buggy for des_check_key == 1,
9014 have a cleaner way to pick the version they need.
9017 *) New function PKCS12_newpass() which changes the password of a
9021 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9022 dynamic mix. In both cases the ids can be used as an index into the
9023 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9024 functions so they accept a list of the field values and the
9025 application doesn't need to directly manipulate the X509_TRUST
9029 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9033 *) Modify the way the V3 extension code looks up extensions. This now
9034 works in a similar way to the object code: we have some "standard"
9035 extensions in a static table which is searched with OBJ_bsearch()
9036 and the application can add dynamic ones if needed. The file
9037 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9038 updated whenever a new extension is added to the core code and kept
9039 in ext_nid order. There is a simple program 'tabtest.c' which checks
9040 this. New extensions are not added too often so this file can readily
9041 be maintained manually.
9043 There are two big advantages in doing things this way. The extensions
9044 can be looked up immediately and no longer need to be "added" using
9045 X509V3_add_standard_extensions(): this function now does nothing.
9046 [Side note: I get *lots* of email saying the extension code doesn't
9047 work because people forget to call this function]
9048 Also no dynamic allocation is done unless new extensions are added:
9049 so if we don't add custom extensions there is no need to call
9050 X509V3_EXT_cleanup().
9053 *) Modify enc utility's salting as follows: make salting the default. Add a
9054 magic header, so unsalted files fail gracefully instead of just decrypting
9055 to garbage. This is because not salting is a big security hole, so people
9056 should be discouraged from doing it.
9059 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9060 digest to be passed on the command line but it only used this
9061 parameter when signing a certificate. Modified so all relevant
9062 operations are affected by the digest parameter including the
9063 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9064 DSA key was used because it didn't fix the digest.
9067 *) Initial certificate chain verify code. Currently tests the untrusted
9068 certificates for consistency with the verify purpose (which is set
9069 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9071 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9072 this is because it will reject chains with invalid extensions whereas
9073 every previous version of OpenSSL and SSLeay made no checks at all.
9075 Trust code: checks the root CA for the relevant trust settings. Trust
9076 settings have an initial value consistent with the verify purpose: e.g.
9077 if the verify purpose is for SSL client use it expects the CA to be
9078 trusted for SSL client use. However the default value can be changed to
9079 permit custom trust settings: one example of this would be to only trust
9080 certificates from a specific "secure" set of CAs.
9082 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9083 which should be used for version portability: especially since the
9084 verify structure is likely to change more often now.
9086 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9087 to set them. If not set then assume SSL clients will verify SSL servers
9090 Two new options to the verify program: -untrusted allows a set of
9091 untrusted certificates to be passed in and -purpose which sets the
9092 intended purpose of the certificate. If a purpose is set then the
9093 new chain verify code is used to check extension consistency.
9096 *) Support for the authority information access extension.
9099 *) Modify RSA and DSA PEM read routines to transparently handle
9100 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9101 public keys in a format compatible with certificate
9102 SubjectPublicKeyInfo structures. Unfortunately there were already
9103 functions called *_PublicKey_* which used various odd formats so
9104 these are retained for compatibility: however the DSA variants were
9105 never in a public release so they have been deleted. Changed dsa/rsa
9106 utilities to handle the new format: note no releases ever handled public
9107 keys so we should be OK.
9109 The primary motivation for this change is to avoid the same fiasco
9110 that dogs private keys: there are several incompatible private key
9111 formats some of which are standard and some OpenSSL specific and
9112 require various evil hacks to allow partial transparent handling and
9113 even then it doesn't work with DER formats. Given the option anything
9114 other than PKCS#8 should be dumped: but the other formats have to
9115 stay in the name of compatibility.
9117 With public keys and the benefit of hindsight one standard format
9118 is used which works with EVP_PKEY, RSA or DSA structures: though
9119 it clearly returns an error if you try to read the wrong kind of key.
9121 Added a -pubkey option to the 'x509' utility to output the public key.
9122 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9123 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9124 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9125 that do the same as the EVP_PKEY_assign_*() except they up the
9126 reference count of the added key (they don't "swallow" the
9130 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9131 CRLs would fail if the file contained no certificates or no CRLs:
9132 added a new function to read in both types and return the number
9133 read: this means that if none are read it will be an error. The
9134 DER versions of the certificate and CRL reader would always fail
9135 because it isn't possible to mix certificates and CRLs in DER format
9136 without choking one or the other routine. Changed this to just read
9137 a certificate: this is the best we can do. Also modified the code
9138 in apps/verify.c to take notice of return codes: it was previously
9139 attempting to read in certificates from NULL pointers and ignoring
9140 any errors: this is one reason why the cert and CRL reader seemed
9141 to work. It doesn't check return codes from the default certificate
9142 routines: these may well fail if the certificates aren't installed.
9145 *) Code to support otherName option in GeneralName.
9148 *) First update to verify code. Change the verify utility
9149 so it warns if it is passed a self signed certificate:
9150 for consistency with the normal behaviour. X509_verify
9151 has been modified to it will now verify a self signed
9152 certificate if *exactly* the same certificate appears
9153 in the store: it was previously impossible to trust a
9154 single self signed certificate. This means that:
9155 openssl verify ss.pem
9156 now gives a warning about a self signed certificate but
9157 openssl verify -CAfile ss.pem ss.pem
9161 *) For servers, store verify_result in SSL_SESSION data structure
9162 (and add it to external session representation).
9163 This is needed when client certificate verifications fails,
9164 but an application-provided verification callback (set by
9165 SSL_CTX_set_cert_verify_callback) allows accepting the session
9166 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9167 but returns 1): When the session is reused, we have to set
9168 ssl->verify_result to the appropriate error code to avoid
9170 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9172 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9173 case in PKCS7_dataInit() where the signed PKCS7 structure
9174 didn't contain any existing data because it was being created.
9175 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9177 *) Add a salt to the key derivation routines in enc.c. This
9178 forms the first 8 bytes of the encrypted file. Also add a
9179 -S option to allow a salt to be input on the command line.
9182 *) New function X509_cmp(). Oddly enough there wasn't a function
9183 to compare two certificates. We do this by working out the SHA1
9184 hash and comparing that. X509_cmp() will be needed by the trust
9188 *) SSL_get1_session() is like SSL_get_session(), but increments
9189 the reference count in the SSL_SESSION returned.
9190 [Geoff Thorpe <geoff@eu.c2.net>]
9192 *) Fix for 'req': it was adding a null to request attributes.
9193 Also change the X509_LOOKUP and X509_INFO code to handle
9194 certificate auxiliary information.
9197 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9201 *) Add the possibility to add extra information to the memory leak
9202 detecting output, to form tracebacks, showing from where each
9203 allocation was originated: CRYPTO_push_info("constant string") adds
9204 the string plus current file name and line number to a per-thread
9205 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9206 is like calling CYRPTO_pop_info() until the stack is empty.
9207 Also updated memory leak detection code to be multi-thread-safe.
9210 *) Add options -text and -noout to pkcs7 utility and delete the
9211 encryption options which never did anything. Update docs.
9214 *) Add options to some of the utilities to allow the pass phrase
9215 to be included on either the command line (not recommended on
9216 OSes like Unix) or read from the environment. Update the
9217 manpages and fix a few bugs.
9220 *) Add a few manpages for some of the openssl commands.
9223 *) Fix the -revoke option in ca. It was freeing up memory twice,
9224 leaking and not finding already revoked certificates.
9227 *) Extensive changes to support certificate auxiliary information.
9228 This involves the use of X509_CERT_AUX structure and X509_AUX
9229 functions. An X509_AUX function such as PEM_read_X509_AUX()
9230 can still read in a certificate file in the usual way but it
9231 will also read in any additional "auxiliary information". By
9232 doing things this way a fair degree of compatibility can be
9233 retained: existing certificates can have this information added
9234 using the new 'x509' options.
9236 Current auxiliary information includes an "alias" and some trust
9237 settings. The trust settings will ultimately be used in enhanced
9238 certificate chain verification routines: currently a certificate
9239 can only be trusted if it is self signed and then it is trusted
9243 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9244 The problem was that one of the replacement routines had not been working
9245 since SSLeay releases. For now the offending routine has been replaced
9246 with non-optimised assembler. Even so, this now gives around 95%
9247 performance improvement for 1024 bit RSA signs.
9250 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9251 handling. Most clients have the effective key size in bits equal to
9252 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9253 A few however don't do this and instead use the size of the decrypted key
9254 to determine the RC2 key length and the AlgorithmIdentifier to determine
9255 the effective key length. In this case the effective key length can still
9256 be 40 bits but the key length can be 168 bits for example. This is fixed
9257 by manually forcing an RC2 key into the EVP_PKEY structure because the
9258 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9259 the key length and effective key length are equal.
9262 *) Add a bunch of functions that should simplify the creation of
9263 X509_NAME structures. Now you should be able to do:
9264 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9265 and have it automatically work out the correct field type and fill in
9266 the structures. The more adventurous can try:
9267 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9268 and it will (hopefully) work out the correct multibyte encoding.
9271 *) Change the 'req' utility to use the new field handling and multibyte
9272 copy routines. Before the DN field creation was handled in an ad hoc
9273 way in req, ca, and x509 which was rather broken and didn't support
9274 BMPStrings or UTF8Strings. Since some software doesn't implement
9275 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9276 using the dirstring_type option. See the new comment in the default
9277 openssl.cnf for more info.
9280 *) Make crypto/rand/md_rand.c more robust:
9281 - Assure unique random numbers after fork().
9282 - Make sure that concurrent threads access the global counter and
9283 md serializably so that we never lose entropy in them
9284 or use exactly the same state in multiple threads.
9285 Access to the large state is not always serializable because
9286 the additional locking could be a performance killer, and
9287 md should be large enough anyway.
9290 *) New file apps/app_rand.c with commonly needed functionality
9291 for handling the random seed file.
9293 Use the random seed file in some applications that previously did not:
9295 dsaparam -genkey (which also ignored its '-rand' option),
9298 x509 (when signing).
9299 Except on systems with /dev/urandom, it is crucial to have a random
9300 seed file at least for key creation, DSA signing, and for DH exchanges;
9301 for RSA signatures we could do without one.
9303 gendh and gendsa (unlike genrsa) used to read only the first byte
9304 of each file listed in the '-rand' option. The function as previously
9305 found in genrsa is now in app_rand.c and is used by all programs
9306 that support '-rand'.
9309 *) In RAND_write_file, use mode 0600 for creating files;
9310 don't just chmod when it may be too late.
9313 *) Report an error from X509_STORE_load_locations
9314 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9317 *) New function ASN1_mbstring_copy() this copies a string in either
9318 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9319 into an ASN1_STRING type. A mask of permissible types is passed
9320 and it chooses the "minimal" type to use or an error if not type
9324 *) Add function equivalents to the various macros in asn1.h. The old
9325 macros are retained with an M_ prefix. Code inside the library can
9326 use the M_ macros. External code (including the openssl utility)
9327 should *NOT* in order to be "shared library friendly".
9330 *) Add various functions that can check a certificate's extensions
9331 to see if it usable for various purposes such as SSL client,
9332 server or S/MIME and CAs of these types. This is currently
9333 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9334 verification. Also added a -purpose flag to x509 utility to
9335 print out all the purposes.
9338 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9342 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9343 for, obtain and decode and extension and obtain its critical flag.
9344 This allows all the necessary extension code to be handled in a
9345 single function call.
9348 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9349 platforms. See crypto/rc4/rc4_enc.c for further details.
9352 *) New -noout option to asn1parse. This causes no output to be produced
9353 its main use is when combined with -strparse and -out to extract data
9354 from a file (which may not be in ASN.1 format).
9357 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9358 when producing the local key id.
9359 [Richard Levitte <levitte@stacken.kth.se>]
9361 *) New option -dhparam in s_server. This allows a DH parameter file to be
9362 stated explicitly. If it is not stated then it tries the first server
9363 certificate file. The previous behaviour hard coded the filename
9367 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9368 a public key to be input or output. For example:
9369 openssl rsa -in key.pem -pubout -out pubkey.pem
9370 Also added necessary DSA public key functions to handle this.
9373 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9374 in the message. This was handled by allowing
9375 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9376 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9378 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9379 to the end of the strings whereas this didn't. This would cause problems
9380 if strings read with d2i_ASN1_bytes() were later modified.
9381 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9383 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9384 data and it contains EOF it will end up returning an error. This is
9385 caused by input 46 bytes long. The cause is due to the way base64
9386 BIOs find the start of base64 encoded data. They do this by trying a
9387 trial decode on each line until they find one that works. When they
9388 do a flag is set and it starts again knowing it can pass all the
9389 data directly through the decoder. Unfortunately it doesn't reset
9390 the context it uses. This means that if EOF is reached an attempt
9391 is made to pass two EOFs through the context and this causes the
9392 resulting error. This can also cause other problems as well. As is
9393 usual with these problems it takes *ages* to find and the fix is
9394 trivial: move one line.
9395 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9397 *) Ugly workaround to get s_client and s_server working under Windows. The
9398 old code wouldn't work because it needed to select() on sockets and the
9399 tty (for keypresses and to see if data could be written). Win32 only
9400 supports select() on sockets so we select() with a 1s timeout on the
9401 sockets and then see if any characters are waiting to be read, if none
9402 are present then we retry, we also assume we can always write data to
9403 the tty. This isn't nice because the code then blocks until we've
9404 received a complete line of data and it is effectively polling the
9405 keyboard at 1s intervals: however it's quite a bit better than not
9406 working at all :-) A dedicated Windows application might handle this
9407 with an event loop for example.
9410 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9411 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9412 will be called when RSA_sign() and RSA_verify() are used. This is useful
9413 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9414 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9415 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9416 This necessitated the support of an extra signature type NID_md5_sha1
9417 for SSL signatures and modifications to the SSL library to use it instead
9418 of calling RSA_public_decrypt() and RSA_private_encrypt().
9421 *) Add new -verify -CAfile and -CApath options to the crl program, these
9422 will lookup a CRL issuers certificate and verify the signature in a
9423 similar way to the verify program. Tidy up the crl program so it
9424 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9425 less strict. It will now permit CRL extensions even if it is not
9426 a V2 CRL: this will allow it to tolerate some broken CRLs.
9429 *) Initialize all non-automatic variables each time one of the openssl
9430 sub-programs is started (this is necessary as they may be started
9431 multiple times from the "OpenSSL>" prompt).
9432 [Lennart Bang, Bodo Moeller]
9434 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9435 removing all other RSA functionality (this is what NO_RSA does). This
9436 is so (for example) those in the US can disable those operations covered
9437 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9441 *) Non-copying interface to BIO pairs.
9442 (still largely untested)
9445 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9446 ASCII string. This was handled independently in various places before.
9449 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9450 UTF8 strings a character at a time.
9453 *) Use client_version from client hello to select the protocol
9454 (s23_srvr.c) and for RSA client key exchange verification
9455 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9458 *) Add various utility functions to handle SPKACs, these were previously
9459 handled by poking round in the structure internals. Added new function
9460 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9461 print, verify and generate SPKACs. Based on an original idea from
9462 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9465 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9468 *) Allow the config file extension section to be overwritten on the
9469 command line. Based on an original idea from Massimiliano Pala
9470 <madwolf@comune.modena.it>. The new option is called -extensions
9471 and can be applied to ca, req and x509. Also -reqexts to override
9472 the request extensions in req and -crlexts to override the crl extensions
9476 *) Add new feature to the SPKAC handling in ca. Now you can include
9477 the same field multiple times by preceding it by "XXXX." for example:
9480 this is the same syntax as used in the req config file.
9483 *) Allow certificate extensions to be added to certificate requests. These
9484 are specified in a 'req_extensions' option of the req section of the
9485 config file. They can be printed out with the -text option to req but
9486 are otherwise ignored at present.
9489 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9490 data read consists of only the final block it would not decrypted because
9491 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9492 A misplaced 'break' also meant the decrypted final block might not be
9493 copied until the next read.
9496 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9497 a few extra parameters to the DH structure: these will be useful if
9498 for example we want the value of 'q' or implement X9.42 DH.
9501 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9502 provides hooks that allow the default DSA functions or functions on a
9503 "per key" basis to be replaced. This allows hardware acceleration and
9504 hardware key storage to be handled without major modification to the
9505 library. Also added low level modexp hooks and CRYPTO_EX structure and
9506 associated functions.
9509 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9510 as "read only": it can't be written to and the buffer it points to will
9511 not be freed. Reading from a read only BIO is much more efficient than
9512 a normal memory BIO. This was added because there are several times when
9513 an area of memory needs to be read from a BIO. The previous method was
9514 to create a memory BIO and write the data to it, this results in two
9515 copies of the data and an O(n^2) reading algorithm. There is a new
9516 function BIO_new_mem_buf() which creates a read only memory BIO from
9517 an area of memory. Also modified the PKCS#7 routines to use read only
9521 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9522 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9523 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9524 but a retry condition occured while trying to read the rest.
9527 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9528 NID_pkcs7_encrypted by default: this was wrong since this should almost
9529 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9530 the encrypted data type: this is a more sensible place to put it and it
9531 allows the PKCS#12 code to be tidied up that duplicated this
9535 *) Changed obj_dat.pl script so it takes its input and output files on
9536 the command line. This should avoid shell escape redirection problems
9540 *) Initial support for certificate extension requests, these are included
9541 in things like Xenroll certificate requests. Included functions to allow
9542 extensions to be obtained and added.
9545 *) -crlf option to s_client and s_server for sending newlines as
9546 CRLF (as required by many protocols).
9549 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9551 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9552 [Ralf S. Engelschall]
9554 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9555 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9557 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9561 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9562 DH parameters/keys (q is lost during that conversion, but the resulting
9563 DH parameters contain its length).
9565 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9566 much faster than DH_generate_parameters (which creates parameters
9567 where p = 2*q + 1), and also the smaller q makes DH computations
9568 much more efficient (160-bit exponentiation instead of 1024-bit
9569 exponentiation); so this provides a convenient way to support DHE
9570 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9571 utter importance to use
9572 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9574 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9575 when such DH parameters are used, because otherwise small subgroup
9576 attacks may become possible!
9579 *) Avoid memory leak in i2d_DHparams.
9582 *) Allow the -k option to be used more than once in the enc program:
9583 this allows the same encrypted message to be read by multiple recipients.
9586 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9587 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9588 it will always use the numerical form of the OID, even if it has a short
9592 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9593 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9594 otherwise bn_mod_exp was called. In the case of hardware keys for example
9595 no private key components need be present and it might store extra data
9596 in the RSA structure, which cannot be accessed from bn_mod_exp.
9597 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9598 private key operations.
9601 *) Added support for SPARC Linux.
9604 *) pem_password_cb function type incompatibly changed from
9605 typedef int pem_password_cb(char *buf, int size, int rwflag);
9607 ....(char *buf, int size, int rwflag, void *userdata);
9608 so that applications can pass data to their callbacks:
9609 The PEM[_ASN1]_{read,write}... functions and macros now take an
9610 additional void * argument, which is just handed through whenever
9611 the password callback is called.
9612 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9614 New function SSL_CTX_set_default_passwd_cb_userdata.
9616 Compatibility note: As many C implementations push function arguments
9617 onto the stack in reverse order, the new library version is likely to
9618 interoperate with programs that have been compiled with the old
9619 pem_password_cb definition (PEM_whatever takes some data that
9620 happens to be on the stack as its last argument, and the callback
9621 just ignores this garbage); but there is no guarantee whatsoever that
9624 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9625 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9626 problems not only on Windows, but also on some Unix platforms.
9627 To avoid problematic command lines, these definitions are now in an
9628 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9629 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9632 *) MIPS III/IV assembler module is reimplemented.
9635 *) More DES library cleanups: remove references to srand/rand and
9636 delete an unused file.
9639 *) Add support for the the free Netwide assembler (NASM) under Win32,
9640 since not many people have MASM (ml) and it can be hard to obtain.
9641 This is currently experimental but it seems to work OK and pass all
9642 the tests. Check out INSTALL.W32 for info.
9645 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9646 without temporary keys kept an extra copy of the server key,
9647 and connections with temporary keys did not free everything in case
9651 *) New function RSA_check_key and new openssl rsa option -check
9652 for verifying the consistency of RSA keys.
9653 [Ulf Moeller, Bodo Moeller]
9655 *) Various changes to make Win32 compile work:
9656 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9657 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9658 comparison" warnings.
9659 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9662 *) Add a debugging option to PKCS#5 v2 key generation function: when
9663 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9664 derived keys are printed to stderr.
9667 *) Copy the flags in ASN1_STRING_dup().
9668 [Roman E. Pavlov <pre@mo.msk.ru>]
9670 *) The x509 application mishandled signing requests containing DSA
9671 keys when the signing key was also DSA and the parameters didn't match.
9673 It was supposed to omit the parameters when they matched the signing key:
9674 the verifying software was then supposed to automatically use the CA's
9675 parameters if they were absent from the end user certificate.
9677 Omitting parameters is no longer recommended. The test was also
9678 the wrong way round! This was probably due to unusual behaviour in
9679 EVP_cmp_parameters() which returns 1 if the parameters match.
9680 This meant that parameters were omitted when they *didn't* match and
9681 the certificate was useless. Certificates signed with 'ca' didn't have
9683 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9685 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9686 The interface is as follows:
9687 Applications can use
9688 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9689 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9690 "off" is now the default.
9691 The library internally uses
9692 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9693 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9694 to disable memory-checking temporarily.
9696 Some inconsistent states that previously were possible (and were
9697 even the default) are now avoided.
9699 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9700 with each memory chunk allocated; this is occasionally more helpful
9701 than just having a counter.
9703 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9705 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9709 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9710 which largely parallels "options", but is for changing API behaviour,
9711 whereas "options" are about protocol behaviour.
9712 Initial "mode" flags are:
9714 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9715 a single record has been written.
9716 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9717 retries use the same buffer location.
9718 (But all of the contents must be
9722 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9725 *) Fix problems with no-hmac etc.
9726 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9728 *) New functions RSA_get_default_method(), RSA_set_method() and
9729 RSA_get_method(). These allows replacement of RSA_METHODs without having
9730 to mess around with the internals of an RSA structure.
9733 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9734 Also really enable memory leak checks in openssl.c and in some
9736 [Chad C. Mulligan, Bodo Moeller]
9738 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9739 up the length of negative integers. This has now been simplified to just
9740 store the length when it is first determined and use it later, rather
9741 than trying to keep track of where data is copied and updating it to
9743 [Steve Henson, reported by Brien Wheeler
9744 <bwheeler@authentica-security.com>]
9746 *) Add a new function PKCS7_signatureVerify. This allows the verification
9747 of a PKCS#7 signature but with the signing certificate passed to the
9748 function itself. This contrasts with PKCS7_dataVerify which assumes the
9749 certificate is present in the PKCS#7 structure. This isn't always the
9750 case: certificates can be omitted from a PKCS#7 structure and be
9751 distributed by "out of band" means (such as a certificate database).
9754 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9755 function prototypes in pem.h, also change util/mkdef.pl to add the
9756 necessary function names.
9759 *) mk1mf.pl (used by Windows builds) did not properly read the
9760 options set by Configure in the top level Makefile, and Configure
9761 was not even able to write more than one option correctly.
9762 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9765 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9766 file to be loaded from a BIO or FILE pointer. The BIO version will
9767 for example allow memory BIOs to contain config info.
9770 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9771 Whoever hopes to achieve shared-library compatibility across versions
9772 must use this, not the compile-time macro.
9773 (Exercise 0.9.4: Which is the minimum library version required by
9775 Note: All this applies only to multi-threaded programs, others don't
9779 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9780 through a BIO pair triggered the default case, i.e.
9781 SSLerr(...,SSL_R_UNKNOWN_STATE).
9784 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9785 can use the SSL library even if none of the specific BIOs is
9789 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9790 for the encoded length.
9791 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9793 *) Add initial documentation of the X509V3 functions.
9796 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9797 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9798 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9799 secure PKCS#8 private key format with a high iteration count.
9802 *) Fix determination of Perl interpreter: A perl or perl5
9803 _directory_ in $PATH was also accepted as the interpreter.
9804 [Ralf S. Engelschall]
9806 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9807 wrong with it but it was very old and did things like calling
9808 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9812 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9813 to use the new extension code.
9816 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9817 with macros. This should make it easier to change their form, add extra
9818 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9822 *) Add to configuration table a new entry that can specify an alternative
9823 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9824 according to Mark Crispin <MRC@Panda.COM>.
9828 *) DES CBC did not update the IV. Weird.
9831 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9832 Changing the behaviour of the former might break existing programs --
9833 where IV updating is needed, des_ncbc_encrypt can be used.
9836 *) When bntest is run from "make test" it drives bc to check its
9837 calculations, as well as internally checking them. If an internal check
9838 fails, it needs to cause bc to give a non-zero result or make test carries
9839 on without noticing the failure. Fixed.
9842 *) DES library cleanups.
9845 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9846 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9847 ciphers. NOTE: although the key derivation function has been verified
9848 against some published test vectors it has not been extensively tested
9849 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9853 *) Instead of "mkdir -p", which is not fully portable, use new
9854 Perl script "util/mkdir-p.pl".
9857 *) Rewrite the way password based encryption (PBE) is handled. It used to
9858 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9859 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9860 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9861 the 'parameter' field of the AlgorithmIdentifier is passed to the
9862 underlying key generation function so it must do its own ASN1 parsing.
9863 This has also changed the EVP_PBE_CipherInit() function which now has a
9864 'parameter' argument instead of literal salt and iteration count values
9865 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9868 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9869 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9870 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9871 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9872 value was just used as a "magic string" and not used directly its
9873 value doesn't matter.
9876 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9880 *) "linux-sparc64" configuration (ultrapenguin).
9881 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9882 "linux-sparc" configuration.
9883 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9885 *) config now generates no-xxx options for missing ciphers.
9888 *) Support the EBCDIC character set (work in progress).
9889 File ebcdic.c not yet included because it has a different license.
9890 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9892 *) Support BS2000/OSD-POSIX.
9893 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9895 *) Make callbacks for key generation use void * instead of char *.
9898 *) Make S/MIME samples compile (not yet tested).
9901 *) Additional typesafe stacks.
9904 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9908 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9910 *) New configuration variant "sco5-gcc".
9912 *) Updated some demos.
9913 [Sean O Riordain, Wade Scholine]
9915 *) Add missing BIO_free at exit of pkcs12 application.
9918 *) Fix memory leak in conf.c.
9921 *) Updates for Win32 to assembler version of MD5.
9924 *) Set #! path to perl in apps/der_chop to where we found it
9925 instead of using a fixed path.
9928 *) SHA library changes for irix64-mips4-cc.
9931 *) Improvements for VMS support.
9935 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9937 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9938 This also avoids the problems with SC4.2 and unpatched SC5.
9939 [Andy Polyakov <appro@fy.chalmers.se>]
9941 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9942 These are required because of the typesafe stack would otherwise break
9943 existing code. If old code used a structure member which used to be STACK
9944 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9945 sk_num or sk_value it would produce an error because the num, data members
9946 are not present in STACK_OF. Now it just produces a warning. sk_set
9947 replaces the old method of assigning a value to sk_value
9948 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9949 that does this will no longer work (and should use sk_set instead) but
9950 this could be regarded as a "questionable" behaviour anyway.
9953 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9954 correctly handle encrypted S/MIME data.
9957 *) Change type of various DES function arguments from des_cblock
9958 (which means, in function argument declarations, pointer to char)
9959 to des_cblock * (meaning pointer to array with 8 char elements),
9960 which allows the compiler to do more typechecking; it was like
9961 that back in SSLeay, but with lots of ugly casts.
9963 Introduce new type const_des_cblock.
9966 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9967 problems: find RecipientInfo structure that matches recipient certificate
9968 and initialise the ASN1 structures properly based on passed cipher.
9971 *) Belatedly make the BN tests actually check the results.
9974 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9975 to and from BNs: it was completely broken. New compilation option
9976 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9977 key elements as negative integers.
9980 *) Reorganize and speed up MD5.
9981 [Andy Polyakov <appro@fy.chalmers.se>]
9984 [Richard Levitte <richard@levitte.org>]
9986 *) New option -out to asn1parse to allow the parsed structure to be
9987 output to a file. This is most useful when combined with the -strparse
9988 option to examine the output of things like OCTET STRINGS.
9991 *) Make SSL library a little more fool-proof by not requiring any longer
9992 that SSL_set_{accept,connect}_state be called before
9993 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9994 in many applications because usually everything *appeared* to work as
9995 intended anyway -- now it really works as intended).
9998 *) Move openssl.cnf out of lib/.
10001 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10002 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10003 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10004 [Ralf S. Engelschall]
10006 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10007 handle PKCS#7 enveloped data properly.
10008 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10010 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10011 copying pointers. The cert_st handling is changed by this in
10012 various ways (and thus what used to be known as ctx->default_cert
10013 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10014 any longer when s->cert does not give us what we need).
10015 ssl_cert_instantiate becomes obsolete by this change.
10016 As soon as we've got the new code right (possibly it already is?),
10017 we have solved a couple of bugs of the earlier code where s->cert
10018 was used as if it could not have been shared with other SSL structures.
10020 Note that using the SSL API in certain dirty ways now will result
10021 in different behaviour than observed with earlier library versions:
10022 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10023 does not influence s as it used to.
10025 In order to clean up things more thoroughly, inside SSL_SESSION
10026 we don't use CERT any longer, but a new structure SESS_CERT
10027 that holds per-session data (if available); currently, this is
10028 the peer's certificate chain and, for clients, the server's certificate
10029 and temporary key. CERT holds only those values that can have
10030 meaningful defaults in an SSL_CTX.
10033 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10034 from the internal representation. Various PKCS#7 fixes: remove some
10035 evil casts and set the enc_dig_alg field properly based on the signing
10039 *) Allow PKCS#12 password to be set from the command line or the
10040 environment. Let 'ca' get its config file name from the environment
10041 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10045 *) Allow certificate policies extension to use an IA5STRING for the
10046 organization field. This is contrary to the PKIX definition but
10047 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10051 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10052 without disallowing inline assembler and the like for non-pedantic builds.
10055 *) Support Borland C++ builder.
10056 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10058 *) Support Mingw32.
10061 *) SHA-1 cleanups and performance enhancements.
10062 [Andy Polyakov <appro@fy.chalmers.se>]
10064 *) Sparc v8plus assembler for the bignum library.
10065 [Andy Polyakov <appro@fy.chalmers.se>]
10067 *) Accept any -xxx and +xxx compiler options in Configure.
10070 *) Update HPUX configuration.
10073 *) Add missing sk_<type>_unshift() function to safestack.h
10074 [Ralf S. Engelschall]
10076 *) New function SSL_CTX_use_certificate_chain_file that sets the
10077 "extra_cert"s in addition to the certificate. (This makes sense
10078 only for "PEM" format files, as chains as a whole are not
10082 *) Support verify_depth from the SSL API.
10083 x509_vfy.c had what can be considered an off-by-one-error:
10084 Its depth (which was not part of the external interface)
10085 was actually counting the number of certificates in a chain;
10086 now it really counts the depth.
10089 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10090 instead of X509err, which often resulted in confusing error
10091 messages since the error codes are not globally unique
10092 (e.g. an alleged error in ssl3_accept when a certificate
10093 didn't match the private key).
10095 *) New function SSL_CTX_set_session_id_context that allows to set a default
10096 value (so that you don't need SSL_set_session_id_context for each
10097 connection using the SSL_CTX).
10100 *) OAEP decoding bug fix.
10103 *) Support INSTALL_PREFIX for package builders, as proposed by
10107 *) New Configure options "threads" and "no-threads". For systems
10108 where the proper compiler options are known (currently Solaris
10109 and Linux), "threads" is the default.
10112 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10115 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10116 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10117 such as /usr/local/bin.
10120 *) "make linux-shared" to build shared libraries.
10121 [Niels Poppe <niels@netbox.org>]
10123 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10126 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10127 extension adding in x509 utility.
10130 *) Remove NOPROTO sections and error code comments.
10133 *) Partial rewrite of the DEF file generator to now parse the ANSI
10137 *) New Configure options --prefix=DIR and --openssldir=DIR.
10140 *) Complete rewrite of the error code script(s). It is all now handled
10141 by one script at the top level which handles error code gathering,
10142 header rewriting and C source file generation. It should be much better
10143 than the old method: it now uses a modified version of Ulf's parser to
10144 read the ANSI prototypes in all header files (thus the old K&R definitions
10145 aren't needed for error creation any more) and do a better job of
10146 translating function codes into names. The old 'ASN1 error code imbedded
10147 in a comment' is no longer necessary and it doesn't use .err files which
10148 have now been deleted. Also the error code call doesn't have to appear all
10149 on one line (which resulted in some large lines...).
10152 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10155 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10156 0 (which usually indicates a closed connection), but continue reading.
10159 *) Fix some race conditions.
10162 *) Add support for CRL distribution points extension. Add Certificate
10163 Policies and CRL distribution points documentation.
10166 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10169 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10170 8 of keying material. Merlin has also confirmed interop with this fix
10171 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10172 [Merlin Hughes <merlin@baltimore.ie>]
10174 *) Fix lots of warnings.
10175 [Richard Levitte <levitte@stacken.kth.se>]
10177 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10178 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10179 [Richard Levitte <levitte@stacken.kth.se>]
10181 *) Fix problems with sizeof(long) == 8.
10182 [Andy Polyakov <appro@fy.chalmers.se>]
10184 *) Change functions to ANSI C.
10187 *) Fix typos in error codes.
10188 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10190 *) Remove defunct assembler files from Configure.
10193 *) SPARC v8 assembler BIGNUM implementation.
10194 [Andy Polyakov <appro@fy.chalmers.se>]
10196 *) Support for Certificate Policies extension: both print and set.
10197 Various additions to support the r2i method this uses.
10200 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10201 return a const string when you are expecting an allocated buffer.
10204 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10205 types DirectoryString and DisplayText.
10208 *) Add code to allow r2i extensions to access the configuration database,
10209 add an LHASH database driver and add several ctx helper functions.
10212 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10213 fail when they extended the size of a BIGNUM.
10216 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10217 support typesafe stack.
10220 *) Fix typo in SSL_[gs]et_options().
10221 [Nils Frostberg <nils@medcom.se>]
10223 *) Delete various functions and files that belonged to the (now obsolete)
10224 old X509V3 handling code.
10227 *) New Configure option "rsaref".
10230 *) Don't auto-generate pem.h.
10233 *) Introduce type-safe ASN.1 SETs.
10236 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10237 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10239 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10240 that links with OpenSSL (well at least cause lots of warnings), but fear
10241 not: the conversion is trivial, and it eliminates loads of evil casts. A
10242 few STACKed things have been converted already. Feel free to convert more.
10243 In the fullness of time, I'll do away with the STACK type altogether.
10246 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10247 specified in <certfile> by updating the entry in the index.txt file.
10248 This way one no longer has to edit the index.txt file manually for
10249 revoking a certificate. The -revoke option does the gory details now.
10250 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10252 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10253 `-text' option at all and this way the `-noout -text' combination was
10254 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10255 [Ralf S. Engelschall]
10257 *) Make sure a corresponding plain text error message exists for the
10258 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10259 verify callback function determined that a certificate was revoked.
10260 [Ralf S. Engelschall]
10262 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10263 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10264 all available cipers including rc5, which was forgotten until now.
10265 In order to let the testing shell script know which algorithms
10266 are available, a new (up to now undocumented) command
10267 "openssl list-cipher-commands" is used.
10270 *) Bugfix: s_client occasionally would sleep in select() when
10271 it should have checked SSL_pending() first.
10274 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10275 the raw DSA values prior to ASN.1 encoding.
10278 *) Tweaks to Configure
10279 [Niels Poppe <niels@netbox.org>]
10281 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10285 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10288 *) New config option to avoid instructions that are illegal on the 80386.
10289 The default code is faster, but requires at least a 486.
10292 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10293 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10294 same as SSL2_VERSION anyway.
10297 *) New "-showcerts" option for s_client.
10300 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10301 application. Various cleanups and fixes.
10304 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10305 modify error routines to work internally. Add error codes and PBE init
10306 to library startup routines.
10309 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10310 packing functions to asn1 and evp. Changed function names and error
10311 codes along the way.
10314 *) PKCS12 integration: and so it begins... First of several patches to
10315 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10316 objects to objects.h
10319 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10320 and display support for Thawte strong extranet extension.
10323 *) Add LinuxPPC support.
10324 [Jeff Dubrule <igor@pobox.org>]
10326 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10327 bn_div_words in alpha.s.
10328 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10330 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10331 OAEP isn't supported when OpenSSL is built with RSAref.
10332 [Ulf Moeller <ulf@fitug.de>]
10334 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10335 so they no longer are missing under -DNOPROTO.
10336 [Soren S. Jorvang <soren@t.dk>]
10339 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10341 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10342 doesn't work when the session is reused. Coming soon!
10345 *) Fix a security hole, that allows sessions to be reused in the wrong
10346 context thus bypassing client cert protection! All software that uses
10347 client certs and session caches in multiple contexts NEEDS PATCHING to
10348 allow session reuse! A fuller solution is in the works.
10349 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10351 *) Some more source tree cleanups (removed obsolete files
10352 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10353 permission on "config" script to be executable) and a fix for the INSTALL
10355 [Ulf Moeller <ulf@fitug.de>]
10357 *) Remove some legacy and erroneous uses of malloc, free instead of
10359 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10361 *) Make rsa_oaep_test return non-zero on error.
10362 [Ulf Moeller <ulf@fitug.de>]
10364 *) Add support for native Solaris shared libraries. Configure
10365 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10366 if someone would make that last step automatic.
10367 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10369 *) ctx_size was not built with the right compiler during "make links". Fixed.
10372 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10373 except NULL ciphers". This means the default cipher list will no longer
10374 enable NULL ciphers. They need to be specifically enabled e.g. with
10375 the string "DEFAULT:eNULL".
10378 *) Fix to RSA private encryption routines: if p < q then it would
10379 occasionally produce an invalid result. This will only happen with
10380 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10383 *) Be less restrictive and allow also `perl util/perlpath.pl
10384 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10385 because this way one can also use an interpreter named `perl5' (which is
10386 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10387 installed as `perl').
10388 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10390 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10391 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10393 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10394 advapi32.lib to Win32 build and change the pem test comparison
10395 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10396 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10397 and crypto/des/ede_cbcm_enc.c.
10400 *) DES quad checksum was broken on big-endian architectures. Fixed.
10403 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10404 Win32 test batch file so it (might) work again. The Win32 test batch file
10405 is horrible: I feel ill....
10408 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10409 in e_os.h. Audit of header files to check ANSI and non ANSI
10410 sections: 10 functions were absent from non ANSI section and not exported
10411 from Windows DLLs. Fixed up libeay.num for new functions.
10414 *) Make `openssl version' output lines consistent.
10415 [Ralf S. Engelschall]
10417 *) Fix Win32 symbol export lists for BIO functions: Added
10418 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10419 to ms/libeay{16,32}.def.
10420 [Ralf S. Engelschall]
10422 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10423 fine under Unix and passes some trivial tests I've now added. But the
10424 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10425 added to make sure no one expects that this stuff really works in the
10426 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10427 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10429 [Ralf S. Engelschall]
10431 *) Fix the generation of two part addresses in perl.
10432 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10434 *) Add config entry for Linux on MIPS.
10435 [John Tobey <jtobey@channel1.com>]
10437 *) Make links whenever Configure is run, unless we are on Windoze.
10440 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10441 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10445 *) Add a useful kludge to allow package maintainers to specify compiler and
10446 other platforms details on the command line without having to patch the
10447 Configure script everytime: One now can use ``perl Configure
10448 <id>:<details>'', i.e. platform ids are allowed to have details appended
10449 to them (separated by colons). This is treated as there would be a static
10450 pre-configured entry in Configure's %table under key <id> with value
10451 <details> and ``perl Configure <id>'' is called. So, when you want to
10452 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10453 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10454 now, which overrides the FreeBSD-elf entry on-the-fly.
10455 [Ralf S. Engelschall]
10457 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10460 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10461 on the `perl Configure ...' command line. This way one can compile
10462 OpenSSL libraries with Position Independent Code (PIC) which is needed
10463 for linking it into DSOs.
10464 [Ralf S. Engelschall]
10466 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10470 *) Cleaned up the LICENSE document: The official contact for any license
10471 questions now is the OpenSSL core team under openssl-core@openssl.org.
10472 And add a paragraph about the dual-license situation to make sure people
10473 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10474 to the OpenSSL toolkit.
10475 [Ralf S. Engelschall]
10477 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10478 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10479 Additionally cleaned up the `make links' target: Remove unnecessary
10480 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10481 to speed processing and no longer clutter the display with confusing
10482 stuff. Instead only the actually done links are displayed.
10483 [Ralf S. Engelschall]
10485 *) Permit null encryption ciphersuites, used for authentication only. It used
10486 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10487 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10491 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10492 signed attributes when verifying signatures (this would break them),
10493 the detached data encoding was wrong and public keys obtained using
10494 X509_get_pubkey() weren't freed.
10497 *) Add text documentation for the BUFFER functions. Also added a work around
10498 to a Win95 console bug. This was triggered by the password read stuff: the
10499 last character typed gets carried over to the next fread(). If you were
10500 generating a new cert request using 'req' for example then the last
10501 character of the passphrase would be CR which would then enter the first
10505 *) Added the new `Includes OpenSSL Cryptography Software' button as
10506 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10507 button and can be used by applications based on OpenSSL to show the
10508 relationship to the OpenSSL project.
10509 [Ralf S. Engelschall]
10511 *) Remove confusing variables in function signatures in files
10512 ssl/ssl_lib.c and ssl/ssl.h.
10513 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10515 *) Don't install bss_file.c under PREFIX/include/
10516 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10518 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10519 functions that return function pointers and has support for NT specific
10520 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10521 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10522 unsigned to signed types: this was killing the Win32 compile.
10525 *) Add new certificate file to stack functions,
10526 SSL_add_dir_cert_subjects_to_stack() and
10527 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10528 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10529 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10530 This means that Apache-SSL and similar packages don't have to mess around
10531 to add as many CAs as they want to the preferred list.
10534 *) Experiment with doxygen documentation. Currently only partially applied to
10536 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10537 openssl.doxy as the configuration file.
10540 *) Get rid of remaining C++-style comments which strict C compilers hate.
10541 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10543 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10544 compiled in by default: it has problems with large keys.
10547 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10548 DH private keys and/or callback functions which directly correspond to
10549 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10550 is needed for applications which have to configure certificates on a
10551 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10553 For the RSA certificate situation is makes no difference, but
10554 for the DSA certificate situation this fixes the "no shared cipher"
10555 problem where the OpenSSL cipher selection procedure failed because the
10556 temporary keys were not overtaken from the context and the API provided
10557 no way to reconfigure them.
10558 The new functions now let applications reconfigure the stuff and they
10559 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10560 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10561 non-public-API function ssl_cert_instantiate() is used as a helper
10562 function and also to reduce code redundancy inside ssl_rsa.c.
10563 [Ralf S. Engelschall]
10565 *) Move s_server -dcert and -dkey options out of the undocumented feature
10566 area because they are useful for the DSA situation and should be
10567 recognized by the users.
10568 [Ralf S. Engelschall]
10570 *) Fix the cipher decision scheme for export ciphers: the export bits are
10571 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10572 SSL_EXP_MASK. So, the original variable has to be used instead of the
10573 already masked variable.
10574 [Richard Levitte <levitte@stacken.kth.se>]
10576 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10577 [Richard Levitte <levitte@stacken.kth.se>]
10579 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10580 from `int' to `unsigned int' because it's a length and initialized by
10581 EVP_DigestFinal() which expects an `unsigned int *'.
10582 [Richard Levitte <levitte@stacken.kth.se>]
10584 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10585 script. Instead use the usual Shell->Perl transition trick.
10586 [Ralf S. Engelschall]
10588 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10589 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10590 -noout -modulus' as it's already the case for `openssl rsa -noout
10591 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10592 currently the public key is printed (a decision which was already done by
10593 `openssl dsa -modulus' in the past) which serves a similar purpose.
10594 Additionally the NO_RSA no longer completely removes the whole -modulus
10595 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10597 [Ralf S. Engelschall]
10599 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10600 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10601 [Arne Ansper <arne@ats.cyber.ee>]
10603 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10604 to be added. Now both 'req' and 'ca' can use new objects defined in the
10608 *) Add cool BIO that does syslog (or event log on NT).
10609 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10611 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10612 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10613 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10614 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10617 *) Add preliminary config info for new extension code.
10620 *) Make RSA_NO_PADDING really use no padding.
10621 [Ulf Moeller <ulf@fitug.de>]
10623 *) Generate errors when private/public key check is done.
10626 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10627 for some CRL extensions and new objects added.
10630 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10631 key usage extension and fuller support for authority key id.
10634 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10635 padding method for RSA, which is recommended for new applications in PKCS
10636 #1 v2.0 (RFC 2437, October 1998).
10637 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10638 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10639 against Bleichbacher's attack on RSA.
10640 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10643 *) Updates to the new SSL compression code
10644 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10646 *) Fix so that the version number in the master secret, when passed
10647 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10648 (because the server will not accept higher), that the version number
10649 is 0x03,0x01, not 0x03,0x00
10650 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10652 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10653 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10654 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10657 *) Support for RAW extensions where an arbitrary extension can be
10658 created by including its DER encoding. See apps/openssl.cnf for
10662 *) Make sure latest Perl versions don't interpret some generated C array
10663 code as Perl array code in the crypto/err/err_genc.pl script.
10664 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10666 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10667 not many people have the assembler. Various Win32 compilation fixes and
10668 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10669 build instructions.
10672 *) Modify configure script 'Configure' to automatically create crypto/date.h
10673 file under Win32 and also build pem.h from pem.org. New script
10674 util/mkfiles.pl to create the MINFO file on environments that can't do a
10675 'make files': perl util/mkfiles.pl >MINFO should work.
10678 *) Major rework of DES function declarations, in the pursuit of correctness
10679 and purity. As a result, many evil casts evaporated, and some weirdness,
10680 too. You may find this causes warnings in your code. Zapping your evil
10681 casts will probably fix them. Mostly.
10684 *) Fix for a typo in asn1.h. Bug fix to object creation script
10685 obj_dat.pl. It considered a zero in an object definition to mean
10686 "end of object": none of the objects in objects.h have any zeros
10687 so it wasn't spotted.
10688 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10690 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10691 Masking (CBCM). In the absence of test vectors, the best I have been able
10692 to do is check that the decrypt undoes the encrypt, so far. Send me test
10693 vectors if you have them.
10696 *) Correct calculation of key length for export ciphers (too much space was
10697 allocated for null ciphers). This has not been tested!
10700 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10701 message is now correct (it understands "crypto" and "ssl" on its
10702 command line). There is also now an "update" option. This will update
10703 the util/ssleay.num and util/libeay.num files with any new functions.
10705 perl util/mkdef.pl crypto ssl update
10706 it will update them.
10709 *) Overhauled the Perl interface (perl/*):
10710 - ported BN stuff to OpenSSL's different BN library
10711 - made the perl/ source tree CVS-aware
10712 - renamed the package from SSLeay to OpenSSL (the files still contain
10713 their history because I've copied them in the repository)
10714 - removed obsolete files (the test scripts will be replaced
10715 by better Test::Harness variants in the future)
10716 [Ralf S. Engelschall]
10718 *) First cut for a very conservative source tree cleanup:
10719 1. merge various obsolete readme texts into doc/ssleay.txt
10720 where we collect the old documents and readme texts.
10721 2. remove the first part of files where I'm already sure that we no
10722 longer need them because of three reasons: either they are just temporary
10723 files which were left by Eric or they are preserved original files where
10724 I've verified that the diff is also available in the CVS via "cvs diff
10725 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10726 the crypto/md/ stuff).
10727 [Ralf S. Engelschall]
10729 *) More extension code. Incomplete support for subject and issuer alt
10730 name, issuer and authority key id. Change the i2v function parameters
10731 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10732 what that's for :-) Fix to ASN1 macro which messed up
10733 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10736 *) Preliminary support for ENUMERATED type. This is largely copied from the
10740 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10741 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10743 *) Make sure `make rehash' target really finds the `openssl' program.
10744 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10746 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10747 like to hear about it if this slows down other processors.
10750 *) Add CygWin32 platform information to Configure script.
10751 [Alan Batie <batie@aahz.jf.intel.com>]
10753 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10754 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10756 *) New program nseq to manipulate netscape certificate sequences
10759 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10763 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10764 but the BN code had some problems that would cause failures when
10765 doing certificate verification and some other functions.
10766 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10768 *) Add ASN1 and PEM code to support netscape certificate sequences.
10771 *) Add ASN1 and PEM code to support netscape certificate sequences.
10774 *) Add several PKIX and private extended key usage OIDs.
10777 *) Modify the 'ca' program to handle the new extension code. Modify
10778 openssl.cnf for new extension format, add comments.
10781 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10782 and add a sample to openssl.cnf so req -x509 now adds appropriate
10786 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10787 error code, add initial support to X509_print() and x509 application.
10790 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10791 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10792 stuff is currently isolated and isn't even compiled yet.
10795 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10796 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10797 Removed the versions check from X509 routines when loading extensions:
10798 this allows certain broken certificates that don't set the version
10799 properly to be processed.
10802 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10803 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10804 can still be regenerated with "make depend".
10807 *) Spelling mistake in C version of CAST-128.
10808 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10810 *) Changes to the error generation code. The perl script err-code.pl
10811 now reads in the old error codes and retains the old numbers, only
10812 adding new ones if necessary. It also only changes the .err files if new
10813 codes are added. The makefiles have been modified to only insert errors
10814 when needed (to avoid needlessly modifying header files). This is done
10815 by only inserting errors if the .err file is newer than the auto generated
10816 C file. To rebuild all the error codes from scratch (the old behaviour)
10817 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10818 or delete all the .err files.
10821 *) CAST-128 was incorrectly implemented for short keys. The C version has
10822 been fixed, but is untested. The assembler versions are also fixed, but
10823 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10824 to regenerate it if needed.
10825 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10826 Hagino <itojun@kame.net>]
10828 *) File was opened incorrectly in randfile.c.
10829 [Ulf Möller <ulf@fitug.de>]
10831 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10832 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10833 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10834 al: it's just almost always a UTCTime. Note this patch adds new error
10835 codes so do a "make errors" if there are problems.
10838 *) Correct Linux 1 recognition in config.
10839 [Ulf Möller <ulf@fitug.de>]
10841 *) Remove pointless MD5 hash when using DSA keys in ca.
10842 [Anonymous <nobody@replay.com>]
10844 *) Generate an error if given an empty string as a cert directory. Also
10845 generate an error if handed NULL (previously returned 0 to indicate an
10846 error, but didn't set one).
10847 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10849 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10852 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10853 parameters. This was causing a warning which killed off the Win32 compile.
10856 *) Remove C++ style comments from crypto/bn/bn_local.h.
10857 [Neil Costigan <neil.costigan@celocom.com>]
10859 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10860 based on a text string, looking up short and long names and finally
10861 "dot" format. The "dot" format stuff didn't work. Added new function
10862 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10863 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10864 OID is not part of the table.
10867 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10868 X509_LOOKUP_by_alias().
10871 *) Sort openssl functions by name.
10874 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10875 encryption from sample DSA keys (in case anyone is interested the password
10879 *) Make _all_ *_free functions accept a NULL pointer.
10880 [Frans Heymans <fheymans@isaserver.be>]
10882 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10884 [Anonymous <nobody@replay.com>]
10886 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10887 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10889 *) Don't blow it for numeric -newkey arguments to apps/req.
10890 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10892 *) Temp key "for export" tests were wrong in s3_srvr.c.
10893 [Anonymous <nobody@replay.com>]
10895 *) Add prototype for temp key callback functions
10896 SSL_CTX_set_tmp_{rsa,dh}_callback().
10899 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10900 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10903 *) X509_name_add_entry() freed the wrong thing after an error.
10904 [Arne Ansper <arne@ats.cyber.ee>]
10906 *) rsa_eay.c would attempt to free a NULL context.
10907 [Arne Ansper <arne@ats.cyber.ee>]
10909 *) BIO_s_socket() had a broken should_retry() on Windoze.
10910 [Arne Ansper <arne@ats.cyber.ee>]
10912 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10913 [Arne Ansper <arne@ats.cyber.ee>]
10915 *) Make sure the already existing X509_STORE->depth variable is initialized
10916 in X509_STORE_new(), but document the fact that this variable is still
10917 unused in the certificate verification process.
10918 [Ralf S. Engelschall]
10920 *) Fix the various library and apps files to free up pkeys obtained from
10921 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10924 *) Fix reference counting in X509_PUBKEY_get(). This makes
10925 demos/maurice/example2.c work, amongst others, probably.
10926 [Steve Henson and Ben Laurie]
10928 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10929 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10930 are no longer created. This way we have a single and consistent command
10931 line interface `openssl <command>', similar to `cvs <command>'.
10932 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10934 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10935 BIT STRING wrapper always have zero unused bits.
10938 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10941 *) Make the top-level INSTALL documentation easier to understand.
10944 *) Makefiles updated to exit if an error occurs in a sub-directory
10945 make (including if user presses ^C) [Paul Sutton]
10947 *) Make Montgomery context stuff explicit in RSA data structure.
10950 *) Fix build order of pem and err to allow for generated pem.h.
10953 *) Fix renumbering bug in X509_NAME_delete_entry().
10956 *) Enhanced the err-ins.pl script so it makes the error library number
10957 global and can add a library name. This is needed for external ASN1 and
10958 other error libraries.
10961 *) Fixed sk_insert which never worked properly.
10964 *) Fix ASN1 macros so they can handle indefinite length constructed
10965 EXPLICIT tags. Some non standard certificates use these: they can now
10969 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10970 into a single doc/ssleay.txt bundle. This way the information is still
10971 preserved but no longer messes up this directory. Now it's new room for
10972 the new set of documentation files.
10973 [Ralf S. Engelschall]
10975 *) SETs were incorrectly DER encoded. This was a major pain, because they
10976 shared code with SEQUENCEs, which aren't coded the same. This means that
10977 almost everything to do with SETs or SEQUENCEs has either changed name or
10978 number of arguments.
10979 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10981 *) Fix test data to work with the above.
10984 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10985 was already fixed by Eric for 0.9.1 it seems.
10986 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10988 *) Autodetect FreeBSD3.
10991 *) Fix various bugs in Configure. This affects the following platforms:
10995 unixware-2.0-pentium
10999 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11000 before they are needed.
11003 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11007 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11009 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11010 changed SSLeay to OpenSSL in version strings.
11011 [Ralf S. Engelschall]
11013 *) Some fixups to the top-level documents.
11016 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11017 because the symlink to include/ was missing.
11018 [Ralf S. Engelschall]
11020 *) Incorporated the popular no-RSA/DSA-only patches
11021 which allow to compile a RSA-free SSLeay.
11022 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11024 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11025 when "ssleay" is still not found.
11026 [Ralf S. Engelschall]
11028 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11029 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11031 *) Updated the README file.
11032 [Ralf S. Engelschall]
11034 *) Added various .cvsignore files in the CVS repository subdirs
11035 to make a "cvs update" really silent.
11036 [Ralf S. Engelschall]
11038 *) Recompiled the error-definition header files and added
11039 missing symbols to the Win32 linker tables.
11040 [Ralf S. Engelschall]
11042 *) Cleaned up the top-level documents;
11043 o new files: CHANGES and LICENSE
11044 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11045 o merged COPYRIGHT into LICENSE
11046 o removed obsolete TODO file
11047 o renamed MICROSOFT to INSTALL.W32
11048 [Ralf S. Engelschall]
11050 *) Removed dummy files from the 0.9.1b source tree:
11051 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11052 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11053 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11054 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11055 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11056 [Ralf S. Engelschall]
11058 *) Added various platform portability fixes.
11061 *) The Genesis of the OpenSSL rpject:
11062 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11063 Young and Tim J. Hudson created while they were working for C2Net until
11065 [The OpenSSL Project]
11068 Changes between 0.9.0b and 0.9.1b [not released]
11070 *) Updated a few CA certificates under certs/
11073 *) Changed some BIGNUM api stuff.
11076 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11077 DGUX x86, Linux Alpha, etc.
11080 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11081 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11085 *) Add -strparse option to asn1pars program which parses nested
11087 [Dr Stephen Henson <shenson@bigfoot.com>]
11089 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11092 *) DSA fix for "ca" program.
11095 *) Added "-genkey" option to "dsaparam" program.
11098 *) Added RIPE MD160 (rmd160) message digest.
11101 *) Added -a (all) option to "ssleay version" command.
11104 *) Added PLATFORM define which is the id given to Configure.
11107 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11110 *) Extended the ASN.1 parser routines.
11113 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11116 *) Added a BN_CTX to the BN library.
11119 *) Fixed the weak key values in DES library
11122 *) Changed API in EVP library for cipher aliases.
11125 *) Added support for RC2/64bit cipher.
11128 *) Converted the lhash library to the crypto/mem.c functions.
11131 *) Added more recognized ASN.1 object ids.
11134 *) Added more RSA padding checks for SSL/TLS.
11137 *) Added BIO proxy/filter functionality.
11140 *) Added extra_certs to SSL_CTX which can be used
11141 send extra CA certificates to the client in the CA cert chain sending
11142 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11145 *) Now Fortezza is denied in the authentication phase because
11146 this is key exchange mechanism is not supported by SSLeay at all.
11149 *) Additional PKCS1 checks.
11152 *) Support the string "TLSv1" for all TLS v1 ciphers.
11155 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11156 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11159 *) Fixed a few memory leaks.
11162 *) Fixed various code and comment typos.
11165 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11166 bytes sent in the client random.
11167 [Edward Bishop <ebishop@spyglass.com>]