5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Added support for TLS extended master secret from
8 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
9 initial patch which was a great help during development.
12 *) All libssl internal structures have been removed from the public header
13 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
14 now redundant). Users should not attempt to access internal structures
15 directly. Instead they should use the provided API functions.
18 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
19 Access to deprecated functions can be re-enabled by running config with
20 "enable-deprecated". In addition applications wishing to use deprecated
21 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
22 will, by default, disable some transitive includes that previously existed
23 in the header files (e.g. ec.h will no longer, by default, include bn.h)
26 *) Added support for OCB mode. OpenSSL has been granted a patent license
27 compatible with the OpenSSL license for use of OCB. Details are available
28 at https://www.openssl.org/docs/misc/OCB-patent-grant-OpenSSL.pdf. Support
29 for OCB can be removed by calling config with no-ocb.
32 *) SSLv2 support has been removed. It still supports receiving a SSLv2
33 compatible client hello.
36 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
37 done while fixing the error code for the key-too-small case.
38 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
40 *) Removed old DES API.
43 *) Remove various unsupported platforms:
49 Sinix/ReliantUNIX RM400
54 16-bit platforms such as WIN16
57 *) Clean up OPENSSL_NO_xxx #define's
58 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
59 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
60 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
61 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
62 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
63 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
64 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
65 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
66 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
67 Remove MS_STATIC; it's a relic from platforms <32 bits.
70 *) Cleaned up dead code
71 Remove all but one '#ifdef undef' which is to be looked at.
74 *) Experimental support for a new, fast, unbiased prime candidate generator,
75 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
76 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
78 *) New output format NSS in the sess_id command line tool. This allows
79 exporting the session id and the master key in NSS keylog format.
80 [Martin Kaiser <martin@kaiser.cx>]
82 *) Harmonize version and its documentation. -f flag is used to display
84 [mancha <mancha1@zoho.com>]
86 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
88 [mancha <mancha1@zoho.com>]
90 *) Fix some double frees. These are not thought to be exploitable.
91 [mancha <mancha1@zoho.com>]
93 *) A missing bounds check in the handling of the TLS heartbeat extension
94 can be used to reveal up to 64k of memory to a connected client or
97 Thanks for Neel Mehta of Google Security for discovering this bug and to
98 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
99 preparing the fix (CVE-2014-0160)
100 [Adam Langley, Bodo Moeller]
102 *) Fix for the attack described in the paper "Recovering OpenSSL
103 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
104 by Yuval Yarom and Naomi Benger. Details can be obtained from:
105 http://eprint.iacr.org/2014/140
107 Thanks to Yuval Yarom and Naomi Benger for discovering this
108 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
109 [Yuval Yarom and Naomi Benger]
111 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
112 this fixes a limitation in previous versions of OpenSSL.
115 *) Experimental encrypt-then-mac support.
117 Experimental support for encrypt then mac from
118 draft-gutmann-tls-encrypt-then-mac-02.txt
120 To enable it set the appropriate extension number (0x42 for the test
121 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
123 For non-compliant peers (i.e. just about everything) this should have no
126 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
130 *) Add EVP support for key wrapping algorithms, to avoid problems with
131 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
132 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
133 algorithms and include tests cases.
136 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
140 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
141 MGF1 digest and OAEP label.
144 *) Make openssl verify return errors.
145 [Chris Palmer <palmer@google.com> and Ben Laurie]
147 *) New function ASN1_TIME_diff to calculate the difference between two
148 ASN1_TIME structures or one structure and the current time.
151 *) Update fips_test_suite to support multiple command line options. New
152 test to induce all self test errors in sequence and check expected
156 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
157 sign or verify all in one operation.
160 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
161 test programs and fips_test_suite. Includes functionality to parse
162 the minimal script output of fipsalgest.pl directly.
165 *) Add authorisation parameter to FIPS_module_mode_set().
168 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
171 *) Use separate DRBG fields for internal and external flags. New function
172 FIPS_drbg_health_check() to perform on demand health checking. Add
173 generation tests to fips_test_suite with reduced health check interval to
174 demonstrate periodic health checking. Add "nodh" option to
175 fips_test_suite to skip very slow DH test.
178 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
182 *) More extensive health check for DRBG checking many more failure modes.
183 New function FIPS_selftest_drbg_all() to handle every possible DRBG
184 combination: call this in fips_test_suite.
187 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
188 and POST to handle Dual EC cases.
191 *) Add support for canonical generation of DSA parameter 'g'. See
194 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
195 POST to handle HMAC cases.
198 *) Add functions FIPS_module_version() and FIPS_module_version_text()
199 to return numerical and string versions of the FIPS module number.
202 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
203 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
204 outside the validated module in the FIPS capable OpenSSL.
207 *) Minor change to DRBG entropy callback semantics. In some cases
208 there is no multiple of the block length between min_len and
209 max_len. Allow the callback to return more than max_len bytes
210 of entropy but discard any extra: it is the callback's responsibility
211 to ensure that the extra data discarded does not impact the
212 requested amount of entropy.
215 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
216 information in FIPS186-3, SP800-57 and SP800-131A.
219 *) CCM support via EVP. Interface is very similar to GCM case except we
220 must supply all data in one chunk (i.e. no update, final) and the
221 message length must be supplied if AAD is used. Add algorithm test
225 *) Initial version of POST overhaul. Add POST callback to allow the status
226 of POST to be monitored and/or failures induced. Modify fips_test_suite
227 to use callback. Always run all selftests even if one fails.
230 *) XTS support including algorithm test driver in the fips_gcmtest program.
231 Note: this does increase the maximum key length from 32 to 64 bytes but
232 there should be no binary compatibility issues as existing applications
233 will never use XTS mode.
236 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
237 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
238 performs algorithm blocking for unapproved PRNG types. Also do not
239 set PRNG type in FIPS_mode_set(): leave this to the application.
240 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
241 the standard OpenSSL PRNG: set additional data to a date time vector.
244 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
245 This shouldn't present any incompatibility problems because applications
246 shouldn't be using these directly and any that are will need to rethink
247 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
250 *) Extensive self tests and health checking required by SP800-90 DRBG.
251 Remove strength parameter from FIPS_drbg_instantiate and always
252 instantiate at maximum supported strength.
255 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
258 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
261 *) New function DH_compute_key_padded() to compute a DH key and pad with
262 leading zeroes if needed: this complies with SP800-56A et al.
265 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
266 anything, incomplete, subject to change and largely untested at present.
269 *) Modify fipscanisteronly build option to only build the necessary object
270 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
273 *) Add experimental option FIPSSYMS to give all symbols in
274 fipscanister.o and FIPS or fips prefix. This will avoid
275 conflicts with future versions of OpenSSL. Add perl script
276 util/fipsas.pl to preprocess assembly language source files
277 and rename any affected symbols.
280 *) Add selftest checks and algorithm block of non-fips algorithms in
281 FIPS mode. Remove DES2 from selftests.
284 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
285 return internal method without any ENGINE dependencies. Add new
286 tiny fips sign and verify functions.
289 *) New build option no-ec2m to disable characteristic 2 code.
292 *) New build option "fipscanisteronly". This only builds fipscanister.o
293 and (currently) associated fips utilities. Uses the file Makefile.fips
294 instead of Makefile.org as the prototype.
297 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
298 Update fips_gcmtest to use IV generator.
301 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
302 setting output buffer to NULL. The *Final function must be
303 called although it will not retrieve any additional data. The tag
304 can be set or retrieved with a ctrl. The IV length is by default 12
305 bytes (96 bits) but can be set to an alternative value. If the IV
306 length exceeds the maximum IV length (currently 16 bytes) it cannot be
310 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
311 underlying do_cipher function handles all cipher semantics itself
312 including padding and finalisation. This is useful if (for example)
313 an ENGINE cipher handles block padding itself. The behaviour of
314 do_cipher is subtly changed if this flag is set: the return value
315 is the number of characters written to the output buffer (zero is
316 no longer an error code) or a negative error code. Also if the
317 input buffer is NULL and length 0 finalisation should be performed.
320 *) If a candidate issuer certificate is already part of the constructed
321 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
324 *) Improve forward-security support: add functions
326 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
327 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
329 for use by SSL/TLS servers; the callback function will be called whenever a
330 new session is created, and gets to decide whether the session may be
331 cached to make it resumable (return 0) or not (return 1). (As by the
332 SSL/TLS protocol specifications, the session_id sent by the server will be
333 empty to indicate that the session is not resumable; also, the server will
334 not generate RFC 4507 (RFC 5077) session tickets.)
336 A simple reasonable callback implementation is to return is_forward_secure.
337 This parameter will be set to 1 or 0 depending on the ciphersuite selected
338 by the SSL/TLS server library, indicating whether it can provide forward
340 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
342 *) New -verify_name option in command line utilities to set verification
346 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
347 Add CMAC pkey methods.
350 *) Experimental renegotiation in s_server -www mode. If the client
351 browses /reneg connection is renegotiated. If /renegcert it is
352 renegotiated requesting a certificate.
355 *) Add an "external" session cache for debugging purposes to s_server. This
356 should help trace issues which normally are only apparent in deployed
357 multi-process servers.
360 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
361 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
362 BIO_set_cipher() and some obscure PEM functions were changed so they
363 can now return an error. The RAND changes required a change to the
364 RAND_METHOD structure.
367 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
368 a gcc attribute to warn if the result of a function is ignored. This
369 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
370 whose return value is often ignored.
373 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
375 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
376 ARMv5 through ARMv8, as opposite to "locking" it to single one.
377 So far those who have to target multiple plaforms would compromise
378 and argue that binary targeting say ARMv5 would still execute on
379 ARMv8. "Universal" build resolves this compromise by providing
380 near-optimal performance even on newer platforms.
383 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
384 (other platforms pending).
385 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
387 *) Add support for the SignedCertificateTimestampList certificate and
388 OCSP response extensions from RFC6962.
391 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
392 for corner cases. (Certain input points at infinity could lead to
393 bogus results, with non-infinity inputs mapped to infinity too.)
396 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
397 This covers AES, SHA256/512 and GHASH. "Initial" means that most
398 common cases are optimized and there still is room for further
399 improvements. Vector Permutation AES for Altivec is also added.
402 *) Add support for little-endian ppc64 Linux target.
403 [Marcelo Cerri (IBM)]
405 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
406 SHA1, SHA256 and GHASH. "Initial" means that most common cases
407 are optimized and there still is room for further improvements.
408 Both 32- and 64-bit modes are supported.
409 [Andy Polyakov, Ard Biesheuvel (Linaro)]
411 *) Improved ARMv7 NEON support.
414 *) Support for SPARC Architecture 2011 crypto extensions, first
415 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
416 SHA256/512, MD5, GHASH and modular exponentiation.
417 [Andy Polyakov, David Miller]
419 *) Accelerated modular exponentiation for Intel processors, a.k.a.
421 [Shay Gueron & Vlad Krasnov (Intel Corp)]
423 *) Support for new and upcoming Intel processors, including AVX2,
424 BMI and SHA ISA extensions. This includes additional "stitched"
425 implementations, AESNI-SHA256 and GCM, and multi-buffer support
428 This work was sponsored by Intel Corp.
431 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
432 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
433 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
436 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
437 this fixes a limiation in previous versions of OpenSSL.
440 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
441 MGF1 digest and OAEP label.
444 *) Add EVP support for key wrapping algorithms, to avoid problems with
445 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
446 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
447 algorithms and include tests cases.
450 *) Add functions to allocate and set the fields of an ECDSA_METHOD
452 [Douglas E. Engert, Steve Henson]
454 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
455 difference in days and seconds between two tm or ASN1_TIME structures.
458 *) Add -rev test option to s_server to just reverse order of characters
459 received by client and send back to server. Also prints an abbreviated
460 summary of the connection parameters.
463 *) New option -brief for s_client and s_server to print out a brief summary
464 of connection parameters.
467 *) Add callbacks for arbitrary TLS extensions.
468 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
470 *) New option -crl_download in several openssl utilities to download CRLs
471 from CRLDP extension in certificates.
474 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
477 *) New function X509_CRL_diff to generate a delta CRL from the difference
478 of two full CRLs. Add support to "crl" utility.
481 *) New functions to set lookup_crls function and to retrieve
482 X509_STORE from X509_STORE_CTX.
485 *) Print out deprecated issuer and subject unique ID fields in
489 *) Extend OCSP I/O functions so they can be used for simple general purpose
490 HTTP as well as OCSP. New wrapper function which can be used to download
491 CRLs using the OCSP API.
494 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
497 *) SSL_CONF* functions. These provide a common framework for application
498 configuration using configuration files or command lines.
501 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
502 message callback and prints the results. Needs compile time option
503 "enable-ssl-trace". New options to s_client and s_server to enable
507 *) New ctrl and macro to retrieve supported points extensions.
508 Print out extension in s_server and s_client.
511 *) New functions to retrieve certificate signature and signature
515 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
519 *) New Suite B modes for TLS code. These use and enforce the requirements
520 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
521 only use Suite B curves. The Suite B modes can be set by using the
522 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
525 *) New chain verification flags for Suite B levels of security. Check
526 algorithms are acceptable when flags are set in X509_verify_cert.
529 *) Make tls1_check_chain return a set of flags indicating checks passed
530 by a certificate chain. Add additional tests to handle client
531 certificates: checks for matching certificate type and issuer name
535 *) If an attempt is made to use a signature algorithm not in the peer
536 preference list abort the handshake. If client has no suitable
537 signature algorithms in response to a certificate request do not
541 *) If server EC tmp key is not in client preference list abort handshake.
544 *) Add support for certificate stores in CERT structure. This makes it
545 possible to have different stores per SSL structure or one store in
546 the parent SSL_CTX. Include distint stores for certificate chain
547 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
548 to build and store a certificate chain in CERT structure: returing
549 an error if the chain cannot be built: this will allow applications
550 to test if a chain is correctly configured.
552 Note: if the CERT based stores are not set then the parent SSL_CTX
553 store is used to retain compatibility with existing behaviour.
557 *) New function ssl_set_client_disabled to set a ciphersuite disabled
558 mask based on the current session, check mask when sending client
559 hello and checking the requested ciphersuite.
562 *) New ctrls to retrieve and set certificate types in a certificate
563 request message. Print out received values in s_client. If certificate
564 types is not set with custom values set sensible values based on
565 supported signature algorithms.
568 *) Support for distinct client and server supported signature algorithms.
571 *) Add certificate callback. If set this is called whenever a certificate
572 is required by client or server. An application can decide which
573 certificate chain to present based on arbitrary criteria: for example
574 supported signature algorithms. Add very simple example to s_server.
575 This fixes many of the problems and restrictions of the existing client
576 certificate callback: for example you can now clear an existing
577 certificate and specify the whole chain.
580 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
581 the certificate can be used for (if anything). Set valid_flags field
582 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
583 to have similar checks in it.
585 Add new "cert_flags" field to CERT structure and include a "strict mode".
586 This enforces some TLS certificate requirements (such as only permitting
587 certificate signature algorithms contained in the supported algorithms
588 extension) which some implementations ignore: this option should be used
589 with caution as it could cause interoperability issues.
592 *) Update and tidy signature algorithm extension processing. Work out
593 shared signature algorithms based on preferences and peer algorithms
594 and print them out in s_client and s_server. Abort handshake if no
595 shared signature algorithms.
598 *) Add new functions to allow customised supported signature algorithms
599 for SSL and SSL_CTX structures. Add options to s_client and s_server
603 *) New function SSL_certs_clear() to delete all references to certificates
604 from an SSL structure. Before this once a certificate had been added
605 it couldn't be removed.
608 *) Integrate hostname, email address and IP address checking with certificate
609 verification. New verify options supporting checking in opensl utility.
612 *) Fixes and wildcard matching support to hostname and email checking
613 functions. Add manual page.
614 [Florian Weimer (Red Hat Product Security Team)]
616 *) New functions to check a hostname email or IP address against a
617 certificate. Add options x509 utility to print results of checks against
621 *) Fix OCSP checking.
622 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
624 *) Initial experimental support for explicitly trusted non-root CAs.
625 OpenSSL still tries to build a complete chain to a root but if an
626 intermediate CA has a trust setting included that is used. The first
627 setting is used: whether to trust (e.g., -addtrust option to the x509
631 *) Add -trusted_first option which attempts to find certificates in the
632 trusted store even if an untrusted chain is also supplied.
635 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
636 platform support for Linux and Android.
639 *) Support for linux-x32, ILP32 environment in x86_64 framework.
642 *) Experimental multi-implementation support for FIPS capable OpenSSL.
643 When in FIPS mode the approved implementations are used as normal,
644 when not in FIPS mode the internal unapproved versions are used instead.
645 This means that the FIPS capable OpenSSL isn't forced to use the
646 (often lower performance) FIPS implementations outside FIPS mode.
649 *) Transparently support X9.42 DH parameters when calling
650 PEM_read_bio_DHparameters. This means existing applications can handle
651 the new parameter format automatically.
654 *) Initial experimental support for X9.42 DH parameter format: mainly
655 to support use of 'q' parameter for RFC5114 parameters.
658 *) Add DH parameters from RFC5114 including test data to dhtest.
661 *) Support for automatic EC temporary key parameter selection. If enabled
662 the most preferred EC parameters are automatically used instead of
663 hardcoded fixed parameters. Now a server just has to call:
664 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
665 support ECDH and use the most appropriate parameters.
668 *) Enhance and tidy EC curve and point format TLS extension code. Use
669 static structures instead of allocation if default values are used.
670 New ctrls to set curves we wish to support and to retrieve shared curves.
671 Print out shared curves in s_server. New options to s_server and s_client
672 to set list of supported curves.
675 *) New ctrls to retrieve supported signature algorithms and
676 supported curve values as an array of NIDs. Extend openssl utility
677 to print out received values.
680 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
681 between NIDs and the more common NIST names such as "P-256". Enhance
682 ecparam utility and ECC method to recognise the NIST names for curves.
685 *) Enhance SSL/TLS certificate chain handling to support different
686 chains for each certificate instead of one chain in the parent SSL_CTX.
689 *) Support for fixed DH ciphersuite client authentication: where both
690 server and client use DH certificates with common parameters.
693 *) Support for fixed DH ciphersuites: those requiring DH server
697 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
699 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
700 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
701 X509_CINF_get_signature were reverted post internal team review.
703 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
705 *) Abort handshake if server key exchange message is omitted for ephemeral
708 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
709 reporting this issue.
713 *) Remove non-export ephemeral RSA code on client and server. This code
714 violated the TLS standard by allowing the use of temporary RSA keys in
715 non-export ciphersuites and could be used by a server to effectively
716 downgrade the RSA key length used to a value smaller than the server
717 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
718 INRIA or reporting this issue.
722 *) Ensure that the session ID context of an SSL is updated when its
723 SSL_CTX is updated via SSL_set_SSL_CTX.
725 The session ID context is typically set from the parent SSL_CTX,
726 and can vary with the CTX.
729 *) Fix various certificate fingerprint issues.
731 By using non-DER or invalid encodings outside the signed portion of a
732 certificate the fingerprint can be changed without breaking the signature.
733 Although no details of the signed portion of the certificate can be changed
734 this can cause problems with some applications: e.g. those using the
735 certificate fingerprint for blacklists.
737 1. Reject signatures with non zero unused bits.
739 If the BIT STRING containing the signature has non zero unused bits reject
740 the signature. All current signature algorithms require zero unused bits.
742 2. Check certificate algorithm consistency.
744 Check the AlgorithmIdentifier inside TBS matches the one in the
745 certificate signature. NB: this will result in signature failure
746 errors for some broken certificates.
748 Thanks to Konrad Kraszewski from Google for reporting this issue.
750 3. Check DSA/ECDSA signatures use DER.
752 Reencode DSA/ECDSA signatures and compare with the original received
753 signature. Return an error if there is a mismatch.
755 This will reject various cases including garbage after signature
756 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
757 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
758 (negative or with leading zeroes).
760 Further analysis was conducted and fixes were developed by Stephen Henson
761 of the OpenSSL core team.
766 *) Do not resume sessions on the server if the negotiated protocol
767 version does not match the session's version. Resuming with a different
768 version, while not strictly forbidden by the RFC, is of questionable
769 sanity and breaks all known clients.
770 [David Benjamin, Emilia Käsper]
772 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
773 early CCS messages during renegotiation. (Note that because
774 renegotiation is encrypted, this early CCS was not exploitable.)
777 *) Tighten client-side session ticket handling during renegotiation:
778 ensure that the client only accepts a session ticket if the server sends
779 the extension anew in the ServerHello. Previously, a TLS client would
780 reuse the old extension state and thus accept a session ticket if one was
781 announced in the initial ServerHello.
783 Similarly, ensure that the client requires a session ticket if one
784 was advertised in the ServerHello. Previously, a TLS client would
785 ignore a missing NewSessionTicket message.
788 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
792 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
793 sends a carefully crafted handshake message, to cause OpenSSL to fail
794 to free up to 64k of memory causing a memory leak. This could be
795 exploited in a Denial Of Service attack. This issue affects OpenSSL
796 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
797 whether SRTP is used or configured. Implementations of OpenSSL that
798 have been compiled with OPENSSL_NO_SRTP defined are not affected.
800 The fix was developed by the OpenSSL team.
804 *) Session Ticket Memory Leak.
806 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
807 integrity of that ticket is first verified. In the event of a session
808 ticket integrity check failing, OpenSSL will fail to free memory
809 causing a memory leak. By sending a large number of invalid session
810 tickets an attacker could exploit this issue in a Denial Of Service
815 *) Build option no-ssl3 is incomplete.
817 When OpenSSL is configured with "no-ssl3" as a build option, servers
818 could accept and complete a SSL 3.0 handshake, and clients could be
819 configured to send them.
821 [Akamai and the OpenSSL team]
823 *) Add support for TLS_FALLBACK_SCSV.
824 Client applications doing fallback retries should call
825 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
827 [Adam Langley, Bodo Moeller]
829 *) Add additional DigestInfo checks.
831 Reencode DigestInto in DER and check against the original when
832 verifying RSA signature: this will reject any improperly encoded
833 DigestInfo structures.
835 Note: this is a precautionary measure and no attacks are currently known.
839 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
841 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
842 SRP code can be overrun an internal buffer. Add sanity check that
843 g, A, B < N to SRP code.
845 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
846 Group for discovering this issue.
850 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
851 TLS 1.0 instead of higher protocol versions when the ClientHello message
852 is badly fragmented. This allows a man-in-the-middle attacker to force a
853 downgrade to TLS 1.0 even if both the server and the client support a
854 higher protocol version, by modifying the client's TLS records.
856 Thanks to David Benjamin and Adam Langley (Google) for discovering and
857 researching this issue.
861 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
862 to a denial of service attack. A malicious server can crash the client
863 with a null pointer dereference (read) by specifying an anonymous (EC)DH
864 ciphersuite and sending carefully crafted handshake messages.
866 Thanks to Felix Gröbert (Google) for discovering and researching this
871 *) By sending carefully crafted DTLS packets an attacker could cause openssl
872 to leak memory. This can be exploited through a Denial of Service attack.
873 Thanks to Adam Langley for discovering and researching this issue.
877 *) An attacker can force openssl to consume large amounts of memory whilst
878 processing DTLS handshake messages. This can be exploited through a
879 Denial of Service attack.
880 Thanks to Adam Langley for discovering and researching this issue.
884 *) An attacker can force an error condition which causes openssl to crash
885 whilst processing DTLS packets due to memory being freed twice. This
886 can be exploited through a Denial of Service attack.
887 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
892 *) If a multithreaded client connects to a malicious server using a resumed
893 session and the server sends an ec point format extension it could write
894 up to 255 bytes to freed memory.
896 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
901 *) A malicious server can crash an OpenSSL client with a null pointer
902 dereference (read) by specifying an SRP ciphersuite even though it was not
903 properly negotiated with the client. This can be exploited through a
904 Denial of Service attack.
906 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
907 discovering and researching this issue.
911 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
912 X509_name_oneline, X509_name_print_ex et al. to leak some information
913 from the stack. Applications may be affected if they echo pretty printing
914 output to the attacker.
916 Thanks to Ivan Fratric (Google) for discovering this issue.
918 [Emilia Käsper, and Steve Henson]
920 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
921 for corner cases. (Certain input points at infinity could lead to
922 bogus results, with non-infinity inputs mapped to infinity too.)
925 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
927 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
928 handshake can force the use of weak keying material in OpenSSL
929 SSL/TLS clients and servers.
931 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
932 researching this issue. (CVE-2014-0224)
933 [KIKUCHI Masashi, Steve Henson]
935 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
936 OpenSSL DTLS client the code can be made to recurse eventually crashing
939 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
941 [Imre Rad, Steve Henson]
943 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
944 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
945 client or server. This is potentially exploitable to run arbitrary
946 code on a vulnerable client or server.
948 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
949 [Jüri Aedla, Steve Henson]
951 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
952 are subject to a denial of service attack.
954 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
955 this issue. (CVE-2014-3470)
956 [Felix Gröbert, Ivan Fratric, Steve Henson]
958 *) Harmonize version and its documentation. -f flag is used to display
960 [mancha <mancha1@zoho.com>]
962 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
964 [mancha <mancha1@zoho.com>]
966 *) Fix some double frees. These are not thought to be exploitable.
967 [mancha <mancha1@zoho.com>]
969 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
971 *) A missing bounds check in the handling of the TLS heartbeat extension
972 can be used to reveal up to 64k of memory to a connected client or
975 Thanks for Neel Mehta of Google Security for discovering this bug and to
976 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
977 preparing the fix (CVE-2014-0160)
978 [Adam Langley, Bodo Moeller]
980 *) Fix for the attack described in the paper "Recovering OpenSSL
981 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
982 by Yuval Yarom and Naomi Benger. Details can be obtained from:
983 http://eprint.iacr.org/2014/140
985 Thanks to Yuval Yarom and Naomi Benger for discovering this
986 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
987 [Yuval Yarom and Naomi Benger]
989 *) TLS pad extension: draft-agl-tls-padding-03
991 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
992 TLS client Hello record length value would otherwise be > 255 and
993 less that 512 pad with a dummy extension containing zeroes so it
994 is at least 512 bytes long.
996 [Adam Langley, Steve Henson]
998 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
1000 *) Fix for TLS record tampering bug. A carefully crafted invalid
1001 handshake could crash OpenSSL with a NULL pointer exception.
1002 Thanks to Anton Johansson for reporting this issues.
1005 *) Keep original DTLS digest and encryption contexts in retransmission
1006 structures so we can use the previous session parameters if they need
1007 to be resent. (CVE-2013-6450)
1010 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
1011 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
1012 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
1013 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
1014 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
1015 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
1016 [Rob Stradling, Adam Langley]
1018 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
1020 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
1021 supporting platforms or when small records were transferred.
1022 [Andy Polyakov, Steve Henson]
1024 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
1026 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1028 This addresses the flaw in CBC record processing discovered by
1029 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1030 at: http://www.isg.rhul.ac.uk/tls/
1032 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1033 Security Group at Royal Holloway, University of London
1034 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1035 Emilia Käsper for the initial patch.
1037 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1039 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
1040 ciphersuites which can be exploited in a denial of service attack.
1041 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
1042 and detecting this bug and to Wolfgang Ettlinger
1043 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
1047 *) Return an error when checking OCSP signatures when key is NULL.
1048 This fixes a DoS attack. (CVE-2013-0166)
1051 *) Make openssl verify return errors.
1052 [Chris Palmer <palmer@google.com> and Ben Laurie]
1054 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1055 the right response is stapled. Also change SSL_get_certificate()
1056 so it returns the certificate actually sent.
1057 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1058 [Rob Stradling <rob.stradling@comodo.com>]
1060 *) Fix possible deadlock when decoding public keys.
1063 *) Don't use TLS 1.0 record version number in initial client hello
1067 Changes between 1.0.1b and 1.0.1c [10 May 2012]
1069 *) Sanity check record length before skipping explicit IV in TLS
1070 1.2, 1.1 and DTLS to fix DoS attack.
1072 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1073 fuzzing as a service testing platform.
1077 *) Initialise tkeylen properly when encrypting CMS messages.
1078 Thanks to Solar Designer of Openwall for reporting this issue.
1081 *) In FIPS mode don't try to use composite ciphers as they are not
1085 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
1087 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
1088 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
1089 mean any application compiled against OpenSSL 1.0.0 headers setting
1090 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
1091 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
1092 0x10000000L Any application which was previously compiled against
1093 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
1094 will need to be recompiled as a result. Letting be results in
1095 inability to disable specifically TLS 1.1 and in client context,
1096 in unlike event, limit maximum offered version to TLS 1.0 [see below].
1099 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
1100 disable just protocol X, but all protocols above X *if* there are
1101 protocols *below* X still enabled. In more practical terms it means
1102 that if application wants to disable TLS1.0 in favor of TLS1.1 and
1103 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
1104 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
1108 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
1110 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1111 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1112 in CRYPTO_realloc_clean.
1114 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1115 issue and to Adam Langley <agl@chromium.org> for fixing it.
1117 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1119 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
1122 *) Workarounds for some broken servers that "hang" if a client hello
1123 record length exceeds 255 bytes.
1125 1. Do not use record version number > TLS 1.0 in initial client
1126 hello: some (but not all) hanging servers will now work.
1127 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
1128 the number of ciphers sent in the client hello. This should be
1129 set to an even number, such as 50, for example by passing:
1130 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
1131 Most broken servers should now work.
1132 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1133 TLS 1.2 client support entirely.
1136 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1139 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1141 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1142 STRING form instead of a DigestInfo.
1145 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1146 and the RSA_sign/RSA_verify functions. This was made more apparent when
1147 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1148 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1149 the correct format in RSA_verify so both forms transparently work.
1152 *) Some servers which support TLS 1.0 can choke if we initially indicate
1153 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1154 encrypted premaster secret. As a workaround use the maximum permitted
1155 client version in client hello, this should keep such servers happy
1156 and still work with previous versions of OpenSSL.
1159 *) Add support for TLS/DTLS heartbeats.
1160 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1162 *) Add support for SCTP.
1163 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1165 *) Improved PRNG seeding for VOS.
1166 [Paul Green <Paul.Green@stratus.com>]
1168 *) Extensive assembler packs updates, most notably:
1170 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1171 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1172 - x86_64: bit-sliced AES implementation;
1173 - ARM: NEON support, contemporary platforms optimizations;
1174 - s390x: z196 support;
1175 - *: GHASH and GF(2^m) multiplication implementations;
1179 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1180 (removal of unnecessary code)
1181 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1183 *) Add TLS key material exporter from RFC 5705.
1186 *) Add DTLS-SRTP negotiation from RFC 5764.
1189 *) Add Next Protocol Negotiation,
1190 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1191 disabled with a no-npn flag to config or Configure. Code donated
1193 [Adam Langley <agl@google.com> and Ben Laurie]
1195 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1196 NIST-P256, NIST-P521, with constant-time single point multiplication on
1197 typical inputs. Compiler support for the nonstandard type __uint128_t is
1198 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1199 Code made available under Apache License version 2.0.
1201 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1202 line to include this in your build of OpenSSL, and run "make depend" (or
1203 "make update"). This enables the following EC_METHODs:
1205 EC_GFp_nistp224_method()
1206 EC_GFp_nistp256_method()
1207 EC_GFp_nistp521_method()
1209 EC_GROUP_new_by_curve_name() will automatically use these (while
1210 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1212 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1214 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1215 all platforms. Move ssize_t definition from e_os.h to the public
1216 header file e_os2.h as it now appears in public header file cms.h
1219 *) New -sigopt option to the ca, req and x509 utilities. Additional
1220 signature parameters can be passed using this option and in
1224 *) Add RSA PSS signing function. This will generate and set the
1225 appropriate AlgorithmIdentifiers for PSS based on those in the
1226 corresponding EVP_MD_CTX structure. No application support yet.
1229 *) Support for companion algorithm specific ASN1 signing routines.
1230 New function ASN1_item_sign_ctx() signs a pre-initialised
1231 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1232 the appropriate parameters.
1235 *) Add new algorithm specific ASN1 verification initialisation function
1236 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1237 handling will be the same no matter what EVP_PKEY_METHOD is used.
1238 Add a PSS handler to support verification of PSS signatures: checked
1239 against a number of sample certificates.
1242 *) Add signature printing for PSS. Add PSS OIDs.
1243 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1245 *) Add algorithm specific signature printing. An individual ASN1 method
1246 can now print out signatures instead of the standard hex dump.
1248 More complex signatures (e.g. PSS) can print out more meaningful
1249 information. Include DSA version that prints out the signature
1253 *) Password based recipient info support for CMS library: implementing
1257 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1258 neatly separates the code into cipher and PBE sections and is required
1259 for some algorithms that split PBES2 into separate pieces (such as
1260 password based CMS).
1263 *) Session-handling fixes:
1264 - Fix handling of connections that are resuming with a session ID,
1265 but also support Session Tickets.
1266 - Fix a bug that suppressed issuing of a new ticket if the client
1267 presented a ticket with an expired session.
1268 - Try to set the ticket lifetime hint to something reasonable.
1269 - Make tickets shorter by excluding irrelevant information.
1270 - On the client side, don't ignore renewed tickets.
1271 [Adam Langley, Bodo Moeller (Google)]
1273 *) Fix PSK session representation.
1276 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1278 This work was sponsored by Intel.
1281 *) Add GCM support to TLS library. Some custom code is needed to split
1282 the IV between the fixed (from PRF) and explicit (from TLS record)
1283 portions. This adds all GCM ciphersuites supported by RFC5288 and
1284 RFC5289. Generalise some AES* cipherstrings to include GCM and
1285 add a special AESGCM string for GCM only.
1288 *) Expand range of ctrls for AES GCM. Permit setting invocation
1289 field on decrypt and retrieval of invocation field only on encrypt.
1292 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1293 As required by RFC5289 these ciphersuites cannot be used if for
1294 versions of TLS earlier than 1.2.
1297 *) For FIPS capable OpenSSL interpret a NULL default public key method
1298 as unset and return the appropriate default but do *not* set the default.
1299 This means we can return the appropriate method in applications that
1300 switch between FIPS and non-FIPS modes.
1303 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1304 ENGINE is used then we cannot handle that in the FIPS module so we
1305 keep original code iff non-FIPS operations are allowed.
1308 *) Add -attime option to openssl utilities.
1309 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1311 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1314 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1315 FIPS EC methods unconditionally for now.
1318 *) New build option no-ec2m to disable characteristic 2 code.
1321 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1322 all cases can be covered as some introduce binary incompatibilities.
1325 *) Redirect RSA operations to FIPS module including keygen,
1326 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1329 *) Add similar low level API blocking to ciphers.
1332 *) Low level digest APIs are not approved in FIPS mode: any attempt
1333 to use these will cause a fatal error. Applications that *really* want
1334 to use them can use the private_* version instead.
1337 *) Redirect cipher operations to FIPS module for FIPS builds.
1340 *) Redirect digest operations to FIPS module for FIPS builds.
1343 *) Update build system to add "fips" flag which will link in fipscanister.o
1344 for static and shared library builds embedding a signature if needed.
1347 *) Output TLS supported curves in preference order instead of numerical
1348 order. This is currently hardcoded for the highest order curves first.
1349 This should be configurable so applications can judge speed vs strength.
1352 *) Add TLS v1.2 server support for client authentication.
1355 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1359 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1360 FIPS modules versions.
1363 *) Add TLS v1.2 client side support for client authentication. Keep cache
1364 of handshake records longer as we don't know the hash algorithm to use
1365 until after the certificate request message is received.
1368 *) Initial TLS v1.2 client support. Add a default signature algorithms
1369 extension including all the algorithms we support. Parse new signature
1370 format in client key exchange. Relax some ECC signing restrictions for
1371 TLS v1.2 as indicated in RFC5246.
1374 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1375 to new signature format when needed using client digest preference.
1376 All server ciphersuites should now work correctly in TLS v1.2. No client
1377 support yet and no support for client certificates.
1380 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1381 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1382 ciphersuites. At present only RSA key exchange ciphersuites work with
1383 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1384 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1385 and version checking.
1388 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1389 with this defined it will not be affected by any changes to ssl internal
1390 structures. Add several utility functions to allow openssl application
1391 to work with OPENSSL_NO_SSL_INTERN defined.
1395 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1397 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1400 *) Permit abbreviated handshakes when renegotiating using the function
1401 SSL_renegotiate_abbreviated().
1402 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1404 *) Add call to ENGINE_register_all_complete() to
1405 ENGINE_load_builtin_engines(), so some implementations get used
1406 automatically instead of needing explicit application support.
1409 *) Add support for TLS key exporter as described in RFC5705.
1410 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1412 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1413 a few changes are required:
1415 Add SSL_OP_NO_TLSv1_1 flag.
1416 Add TLSv1_1 methods.
1417 Update version checking logic to handle version 1.1.
1418 Add explicit IV handling (ported from DTLS code).
1419 Add command line options to s_client/s_server.
1422 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1424 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1426 This addresses the flaw in CBC record processing discovered by
1427 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1428 at: http://www.isg.rhul.ac.uk/tls/
1430 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1431 Security Group at Royal Holloway, University of London
1432 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1433 Emilia Käsper for the initial patch.
1435 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1437 *) Return an error when checking OCSP signatures when key is NULL.
1438 This fixes a DoS attack. (CVE-2013-0166)
1441 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1442 the right response is stapled. Also change SSL_get_certificate()
1443 so it returns the certificate actually sent.
1444 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1445 (This is a backport)
1446 [Rob Stradling <rob.stradling@comodo.com>]
1448 *) Fix possible deadlock when decoding public keys.
1451 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1453 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1456 *) Sanity check record length before skipping explicit IV in DTLS
1459 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1460 fuzzing as a service testing platform.
1464 *) Initialise tkeylen properly when encrypting CMS messages.
1465 Thanks to Solar Designer of Openwall for reporting this issue.
1468 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1470 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1471 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1472 in CRYPTO_realloc_clean.
1474 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1475 issue and to Adam Langley <agl@chromium.org> for fixing it.
1477 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1479 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1481 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1482 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1483 content decryption and always return the same error. Note: this attack
1484 needs on average 2^20 messages so it only affects automated senders. The
1485 old behaviour can be reenabled in the CMS code by setting the
1486 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1487 an MMA defence is not necessary.
1488 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1489 this issue. (CVE-2012-0884)
1492 *) Fix CVE-2011-4619: make sure we really are receiving a
1493 client hello before rejecting multiple SGC restarts. Thanks to
1494 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1497 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1499 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1500 Thanks to Antonio Martin, Enterprise Secure Access Research and
1501 Development, Cisco Systems, Inc. for discovering this bug and
1502 preparing a fix. (CVE-2012-0050)
1505 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1507 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1508 of the Vaudenay padding oracle attack on CBC mode encryption
1509 which enables an efficient plaintext recovery attack against
1510 the OpenSSL implementation of DTLS. Their attack exploits timing
1511 differences arising during decryption processing. A research
1512 paper describing this attack can be found at:
1513 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1514 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1515 Security Group at Royal Holloway, University of London
1516 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1517 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1518 for preparing the fix. (CVE-2011-4108)
1519 [Robin Seggelmann, Michael Tuexen]
1521 *) Clear bytes used for block padding of SSL 3.0 records.
1523 [Adam Langley (Google)]
1525 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1526 Kadianakis <desnacked@gmail.com> for discovering this issue and
1527 Adam Langley for preparing the fix. (CVE-2011-4619)
1528 [Adam Langley (Google)]
1530 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1531 [Andrey Kulikov <amdeich@gmail.com>]
1533 *) Prevent malformed RFC3779 data triggering an assertion failure.
1534 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1535 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1536 [Rob Austein <sra@hactrn.net>]
1538 *) Improved PRNG seeding for VOS.
1539 [Paul Green <Paul.Green@stratus.com>]
1541 *) Fix ssl_ciph.c set-up race.
1542 [Adam Langley (Google)]
1544 *) Fix spurious failures in ecdsatest.c.
1545 [Emilia Käsper (Google)]
1547 *) Fix the BIO_f_buffer() implementation (which was mixing different
1548 interpretations of the '..._len' fields).
1549 [Adam Langley (Google)]
1551 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1552 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1553 threads won't reuse the same blinding coefficients.
1555 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1556 lock to call BN_BLINDING_invert_ex, and avoids one use of
1557 BN_BLINDING_update for each BN_BLINDING structure (previously,
1558 the last update always remained unused).
1559 [Emilia Käsper (Google)]
1561 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1562 [Bob Buckholz (Google)]
1564 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1566 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1567 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1568 [Kaspar Brand <ossl@velox.ch>]
1570 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1571 for multi-threaded use of ECDH. (CVE-2011-3210)
1572 [Adam Langley (Google)]
1574 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1577 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1578 signature public key algorithm by using OID xref utilities instead.
1579 Before this you could only use some ECC ciphersuites with SHA1 only.
1582 *) Add protection against ECDSA timing attacks as mentioned in the paper
1583 by Billy Bob Brumley and Nicola Tuveri, see:
1585 http://eprint.iacr.org/2011/232.pdf
1587 [Billy Bob Brumley and Nicola Tuveri]
1589 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1591 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1592 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1594 *) Fix bug in string printing code: if *any* escaping is enabled we must
1595 escape the escape character (backslash) or the resulting string is
1599 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1601 *) Disable code workaround for ancient and obsolete Netscape browsers
1602 and servers: an attacker can use it in a ciphersuite downgrade attack.
1603 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1606 *) Fixed J-PAKE implementation error, originally discovered by
1607 Sebastien Martini, further info and confirmation from Stefan
1608 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1611 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1613 *) Fix extension code to avoid race conditions which can result in a buffer
1614 overrun vulnerability: resumed sessions must not be modified as they can
1615 be shared by multiple threads. CVE-2010-3864
1618 *) Fix WIN32 build system to correctly link an ENGINE directory into
1622 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1624 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1626 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1628 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1630 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1631 context. The operation can be customised via the ctrl mechanism in
1632 case ENGINEs want to include additional functionality.
1635 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1638 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1639 output hashes compatible with older versions of OpenSSL.
1640 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1642 *) Fix compression algorithm handling: if resuming a session use the
1643 compression algorithm of the resumed session instead of determining
1644 it from client hello again. Don't allow server to change algorithm.
1647 *) Add load_crls() function to apps tidying load_certs() too. Add option
1648 to verify utility to allow additional CRLs to be included.
1651 *) Update OCSP request code to permit adding custom headers to the request:
1652 some responders need this.
1655 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1657 [Julia Lawall <julia@diku.dk>]
1659 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1660 needlessly dereferenced structures, used obsolete functions and
1661 didn't handle all updated verify codes correctly.
1664 *) Disable MD2 in the default configuration.
1667 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1668 indicate the initial BIO being pushed or popped. This makes it possible
1669 to determine whether the BIO is the one explicitly called or as a result
1670 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1671 it handles reference counts correctly and doesn't zero out the I/O bio
1672 when it is not being explicitly popped. WARNING: applications which
1673 included workarounds for the old buggy behaviour will need to be modified
1674 or they could free up already freed BIOs.
1677 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1678 renaming to all platforms (within the 0.9.8 branch, this was
1679 done conditionally on Netware platforms to avoid a name clash).
1680 [Guenter <lists@gknw.net>]
1682 *) Add ECDHE and PSK support to DTLS.
1683 [Michael Tuexen <tuexen@fh-muenster.de>]
1685 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1689 *) Add "missing" function EVP_MD_flags() (without this the only way to
1690 retrieve a digest flags is by accessing the structure directly. Update
1691 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1692 or cipher is registered as in the "from" argument. Print out all
1693 registered digests in the dgst usage message instead of manually
1694 attempting to work them out.
1697 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1698 this allows the use of compression and extensions. Change default cipher
1699 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1700 by default unless an application cipher string requests it.
1703 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1704 key ids to find matching certificates and keys but some PKCS#12 files
1705 don't follow the (somewhat unwritten) rules and this strategy fails.
1706 Now just gather all certificates together and the first private key
1707 then look for the first certificate that matches the key.
1710 *) Support use of registered digest and cipher names for dgst and cipher
1711 commands instead of having to add each one as a special case. So now
1718 openssl dgst -sha256 foo
1720 and this works for ENGINE based algorithms too.
1724 *) Update Gost ENGINE to support parameter files.
1725 [Victor B. Wagner <vitus@cryptocom.ru>]
1727 *) Support GeneralizedTime in ca utility.
1728 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1730 *) Enhance the hash format used for certificate directory links. The new
1731 form uses the canonical encoding (meaning equivalent names will work
1732 even if they aren't identical) and uses SHA1 instead of MD5. This form
1733 is incompatible with the older format and as a result c_rehash should
1734 be used to rebuild symbolic links.
1737 *) Make PKCS#8 the default write format for private keys, replacing the
1738 traditional format. This form is standardised, more secure and doesn't
1739 include an implicit MD5 dependency.
1742 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1743 committed to OpenSSL should pass this lot as a minimum.
1746 *) Add session ticket override functionality for use by EAP-FAST.
1747 [Jouni Malinen <j@w1.fi>]
1749 *) Modify HMAC functions to return a value. Since these can be implemented
1750 in an ENGINE errors can occur.
1753 *) Type-checked OBJ_bsearch_ex.
1756 *) Type-checked OBJ_bsearch. Also some constification necessitated
1757 by type-checking. Still to come: TXT_DB, bsearch(?),
1758 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1762 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1763 seconds to a tm structure directly, instead of going through OS
1764 specific date routines. This avoids any issues with OS routines such
1765 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1766 and X509_time_adj_ex() to cover the extended range. The existing
1767 X509_time_adj() is still usable and will no longer have any date issues.
1770 *) Delta CRL support. New use deltas option which will attempt to locate
1771 and search any appropriate delta CRLs available.
1773 This work was sponsored by Google.
1776 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1777 code and add additional score elements. Validate alternate CRL paths
1778 as part of the CRL checking and indicate a new error "CRL path validation
1779 error" in this case. Applications wanting additional details can use
1780 the verify callback and check the new "parent" field. If this is not
1781 NULL CRL path validation is taking place. Existing applications wont
1782 see this because it requires extended CRL support which is off by
1785 This work was sponsored by Google.
1788 *) Support for freshest CRL extension.
1790 This work was sponsored by Google.
1793 *) Initial indirect CRL support. Currently only supported in the CRLs
1794 passed directly and not via lookup. Process certificate issuer
1795 CRL entry extension and lookup CRL entries by bother issuer name
1796 and serial number. Check and process CRL issuer entry in IDP extension.
1798 This work was sponsored by Google.
1801 *) Add support for distinct certificate and CRL paths. The CRL issuer
1802 certificate is validated separately in this case. Only enabled if
1803 an extended CRL support flag is set: this flag will enable additional
1804 CRL functionality in future.
1806 This work was sponsored by Google.
1809 *) Add support for policy mappings extension.
1811 This work was sponsored by Google.
1814 *) Fixes to pathlength constraint, self issued certificate handling,
1815 policy processing to align with RFC3280 and PKITS tests.
1817 This work was sponsored by Google.
1820 *) Support for name constraints certificate extension. DN, email, DNS
1821 and URI types are currently supported.
1823 This work was sponsored by Google.
1826 *) To cater for systems that provide a pointer-based thread ID rather
1827 than numeric, deprecate the current numeric thread ID mechanism and
1828 replace it with a structure and associated callback type. This
1829 mechanism allows a numeric "hash" to be extracted from a thread ID in
1830 either case, and on platforms where pointers are larger than 'long',
1831 mixing is done to help ensure the numeric 'hash' is usable even if it
1832 can't be guaranteed unique. The default mechanism is to use "&errno"
1833 as a pointer-based thread ID to distinguish between threads.
1835 Applications that want to provide their own thread IDs should now use
1836 CRYPTO_THREADID_set_callback() to register a callback that will call
1837 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1839 Note that ERR_remove_state() is now deprecated, because it is tied
1840 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1841 to free the current thread's error state should be replaced by
1842 ERR_remove_thread_state(NULL).
1844 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1845 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1846 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1847 application was previously providing a numeric thread callback that
1848 was inappropriate for distinguishing threads, then uniqueness might
1849 have been obtained with &errno that happened immediately in the
1850 intermediate development versions of OpenSSL; this is no longer the
1851 case, the numeric thread callback will now override the automatic use
1853 [Geoff Thorpe, with help from Bodo Moeller]
1855 *) Initial support for different CRL issuing certificates. This covers a
1856 simple case where the self issued certificates in the chain exist and
1857 the real CRL issuer is higher in the existing chain.
1859 This work was sponsored by Google.
1862 *) Removed effectively defunct crypto/store from the build.
1865 *) Revamp of STACK to provide stronger type-checking. Still to come:
1866 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1867 ASN1_STRING, CONF_VALUE.
1870 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1871 RAM on SSL connections. This option can save about 34k per idle SSL.
1874 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1875 STACK, TXT_DB, bsearch, qsort.
1878 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1879 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1880 support for data, signedData, compressedData, digestedData and
1881 encryptedData, envelopedData types included. Scripts to check against
1882 RFC4134 examples draft and interop and consistency checks of many
1883 content types and variants.
1886 *) Add options to enc utility to support use of zlib compression BIO.
1889 *) Extend mk1mf to support importing of options and assembly language
1890 files from Configure script, currently only included in VC-WIN32.
1891 The assembly language rules can now optionally generate the source
1892 files from the associated perl scripts.
1895 *) Implement remaining functionality needed to support GOST ciphersuites.
1896 Interop testing has been performed using CryptoPro implementations.
1897 [Victor B. Wagner <vitus@cryptocom.ru>]
1899 *) s390x assembler pack.
1902 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1906 *) Implement Opaque PRF Input TLS extension as specified in
1907 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1908 official specification yet and no extension type assignment by
1909 IANA exists, this extension (for now) will have to be explicitly
1910 enabled when building OpenSSL by providing the extension number
1911 to use. For example, specify an option
1913 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1915 to the "config" or "Configure" script to enable the extension,
1916 assuming extension number 0x9527 (which is a completely arbitrary
1917 and unofficial assignment based on the MD5 hash of the Internet
1918 Draft). Note that by doing so, you potentially lose
1919 interoperability with other TLS implementations since these might
1920 be using the same extension number for other purposes.
1922 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1923 opaque PRF input value to use in the handshake. This will create
1924 an interal copy of the length-'len' string at 'src', and will
1925 return non-zero for success.
1927 To get more control and flexibility, provide a callback function
1930 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1931 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1935 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1938 Callback function 'cb' will be called in handshakes, and is
1939 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1940 Argument 'arg' is for application purposes (the value as given to
1941 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1942 be provided to the callback function). The callback function
1943 has to return non-zero to report success: usually 1 to use opaque
1944 PRF input just if possible, or 2 to enforce use of the opaque PRF
1945 input. In the latter case, the library will abort the handshake
1946 if opaque PRF input is not successfully negotiated.
1948 Arguments 'peerinput' and 'len' given to the callback function
1949 will always be NULL and 0 in the case of a client. A server will
1950 see the client's opaque PRF input through these variables if
1951 available (NULL and 0 otherwise). Note that if the server
1952 provides an opaque PRF input, the length must be the same as the
1953 length of the client's opaque PRF input.
1955 Note that the callback function will only be called when creating
1956 a new session (session resumption can resume whatever was
1957 previously negotiated), and will not be called in SSL 2.0
1958 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1959 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1960 for applications that need to enforce opaque PRF input.
1964 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1967 [Victor B. Wagner <vitus@cryptocom.ru>]
1969 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1970 RFC4507bis. The encrypted ticket format is an encrypted encoded
1971 SSL_SESSION structure, that way new session features are automatically
1974 If a client application caches session in an SSL_SESSION structure
1975 support is transparent because tickets are now stored in the encoded
1978 The SSL_CTX structure automatically generates keys for ticket
1979 protection in servers so again support should be possible
1980 with no application modification.
1982 If a client or server wishes to disable RFC4507 support then the option
1983 SSL_OP_NO_TICKET can be set.
1985 Add a TLS extension debugging callback to allow the contents of any client
1986 or server extensions to be examined.
1988 This work was sponsored by Google.
1991 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1992 OpenSSL should now compile cleanly on gcc 4.2
1993 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1995 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1996 support including streaming MAC support: this is required for GOST
1997 ciphersuite support.
1998 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
2000 *) Add option -stream to use PKCS#7 streaming in smime utility. New
2001 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
2002 to output in BER and PEM format.
2005 *) Experimental support for use of HMAC via EVP_PKEY interface. This
2006 allows HMAC to be handled via the EVP_DigestSign*() interface. The
2007 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
2008 ENGINE support for HMAC keys which are unextractable. New -mac and
2009 -macopt options to dgst utility.
2012 *) New option -sigopt to dgst utility. Update dgst to use
2013 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
2014 alternative signing paramaters such as X9.31 or PSS in the dgst
2018 *) Change ssl_cipher_apply_rule(), the internal function that does
2019 the work each time a ciphersuite string requests enabling
2020 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
2021 removing ("!foo+bar") a class of ciphersuites: Now it maintains
2022 the order of disabled ciphersuites such that those ciphersuites
2023 that most recently went from enabled to disabled not only stay
2024 in order with respect to each other, but also have higher priority
2025 than other disabled ciphersuites the next time ciphersuites are
2028 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
2029 the same ciphersuites as with "HIGH" alone, but in a specific
2030 order where the PSK ciphersuites come first (since they are the
2031 most recently disabled ciphersuites when "HIGH" is parsed).
2033 Also, change ssl_create_cipher_list() (using this new
2034 funcionality) such that between otherwise identical
2035 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
2039 *) Change ssl_create_cipher_list() so that it automatically
2040 arranges the ciphersuites in reasonable order before starting
2041 to process the rule string. Thus, the definition for "DEFAULT"
2042 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
2043 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
2044 This makes it much easier to arrive at a reasonable default order
2045 in applications for which anonymous ciphers are OK (meaning
2046 that you can't actually use DEFAULT).
2047 [Bodo Moeller; suggested by Victor Duchovni]
2049 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
2050 processing) into multiple integers instead of setting
2051 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
2052 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
2053 (These masks as well as the individual bit definitions are hidden
2054 away into the non-exported interface ssl/ssl_locl.h, so this
2055 change to the definition of the SSL_CIPHER structure shouldn't
2056 affect applications.) This give us more bits for each of these
2057 categories, so there is no longer a need to coagulate AES128 and
2058 AES256 into a single algorithm bit, and to coagulate Camellia128
2059 and Camellia256 into a single algorithm bit, which has led to all
2062 Thus, among other things, the kludge introduced in 0.9.7m and
2063 0.9.8e for masking out AES256 independently of AES128 or masking
2064 out Camellia256 independently of AES256 is not needed here in 0.9.9.
2066 With the change, we also introduce new ciphersuite aliases that
2067 so far were missing: "AES128", "AES256", "CAMELLIA128", and
2071 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
2072 Use the leftmost N bytes of the signature input if the input is
2073 larger than the prime q (with N being the size in bytes of q).
2076 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
2077 it yet and it is largely untested.
2080 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
2083 *) Initial incomplete changes to avoid need for function casts in OpenSSL
2084 some compilers (gcc 4.2 and later) reject their use. Safestack is
2085 reimplemented. Update ASN1 to avoid use of legacy functions.
2088 *) Win32/64 targets are linked with Winsock2.
2091 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
2092 to external functions. This can be used to increase CRL handling
2093 efficiency especially when CRLs are very large by (for example) storing
2094 the CRL revoked certificates in a database.
2097 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
2098 new CRLs added to a directory can be used. New command line option
2099 -verify_return_error to s_client and s_server. This causes real errors
2100 to be returned by the verify callback instead of carrying on no matter
2101 what. This reflects the way a "real world" verify callback would behave.
2104 *) GOST engine, supporting several GOST algorithms and public key formats.
2105 Kindly donated by Cryptocom.
2108 *) Partial support for Issuing Distribution Point CRL extension. CRLs
2109 partitioned by DP are handled but no indirect CRL or reason partitioning
2110 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
2111 selected via a scoring technique which handles IDP and AKID in CRLs.
2114 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
2115 will ultimately be used for all verify operations: this will remove the
2116 X509_STORE dependency on certificate verification and allow alternative
2117 lookup methods. X509_STORE based implementations of these two callbacks.
2120 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
2121 Modify get_crl() to find a valid (unexpired) CRL if possible.
2124 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
2125 this would be called X509_CRL_cmp() but that name is already used by
2126 a function that just compares CRL issuer names. Cache several CRL
2127 extensions in X509_CRL structure and cache CRLDP in X509.
2130 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
2131 this maps equivalent X509_NAME structures into a consistent structure.
2132 Name comparison can then be performed rapidly using memcmp().
2135 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2139 *) Allow digests to supply their own micalg string for S/MIME type using
2140 the ctrl EVP_MD_CTRL_MICALG.
2143 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2144 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2145 ctrl. It can then customise the structure before and/or after signing
2149 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2150 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2151 to free up any added signature OIDs.
2154 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2155 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2156 digest and cipher tables. New options added to openssl utility:
2157 list-message-digest-algorithms and list-cipher-algorithms.
2160 *) Change the array representation of binary polynomials: the list
2161 of degrees of non-zero coefficients is now terminated with -1.
2162 Previously it was terminated with 0, which was also part of the
2163 value; thus, the array representation was not applicable to
2164 polynomials where t^0 has coefficient zero. This change makes
2165 the array representation useful in a more general context.
2168 *) Various modifications and fixes to SSL/TLS cipher string
2169 handling. For ECC, the code now distinguishes between fixed ECDH
2170 with RSA certificates on the one hand and with ECDSA certificates
2171 on the other hand, since these are separate ciphersuites. The
2172 unused code for Fortezza ciphersuites has been removed.
2174 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2175 (not "ECDHE"). For consistency with the code for DH
2176 certificates, use of ECDH certificates is now considered ECDH
2177 authentication, not RSA or ECDSA authentication (the latter is
2178 merely the CA's signing algorithm and not actively used in the
2181 The temporary ciphersuite alias "ECCdraft" is no longer
2182 available, and ECC ciphersuites are no longer excluded from "ALL"
2183 and "DEFAULT". The following aliases now exist for RFC 4492
2184 ciphersuites, most of these by analogy with the DH case:
2186 kECDHr - ECDH cert, signed with RSA
2187 kECDHe - ECDH cert, signed with ECDSA
2188 kECDH - ECDH cert (signed with either RSA or ECDSA)
2189 kEECDH - ephemeral ECDH
2190 ECDH - ECDH cert or ephemeral ECDH
2196 AECDH - anonymous ECDH
2197 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2201 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2202 Use correct micalg parameters depending on digest(s) in signed message.
2205 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2206 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2209 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2210 an engine to register a method. Add ENGINE lookups for methods and
2211 functional reference processing.
2214 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2215 EVP_{Sign,Verify}* which allow an application to customise the signature
2219 *) New -resign option to smime utility. This adds one or more signers
2220 to an existing PKCS#7 signedData structure. Also -md option to use an
2221 alternative message digest algorithm for signing.
2224 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2225 create PKCS7 structures containing multiple signers. Update smime
2226 application to support multiple signers.
2229 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2233 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2234 Reorganize PBE internals to lookup from a static table using NIDs,
2235 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2236 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2237 PRF which will be automatically used with PBES2.
2240 *) Replace the algorithm specific calls to generate keys in "req" with the
2244 *) Update PKCS#7 enveloped data routines to use new API. This is now
2245 supported by any public key method supporting the encrypt operation. A
2246 ctrl is added to allow the public key algorithm to examine or modify
2247 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2251 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2252 a default digest type to use. In most cases this will be SHA1 but some
2253 algorithms (such as GOST) need to specify an alternative digest. The
2254 return value indicates how strong the preference is 1 means optional and
2255 2 is mandatory (that is it is the only supported type). Modify
2256 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2257 use the default md. Update openssl utilities to use the default digest
2258 type for signing if it is not explicitly indicated.
2261 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2262 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2263 signing method from the key type. This effectively removes the link
2264 between digests and public key types.
2267 *) Add an OID cross reference table and utility functions. Its purpose is to
2268 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2269 rsaEncryption. This will allow some of the algorithm specific hackery
2270 needed to use the correct OID to be removed.
2273 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2274 structures for PKCS7_sign(). They are now set up by the relevant public
2278 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2281 *) Add support for key derivation (agreement) in the API, DH method and
2285 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2286 public and private key formats. As a side effect these add additional
2287 command line functionality not previously available: DSA signatures can be
2288 generated and verified using pkeyutl and DH key support and generation in
2293 [Oliver Tappe <zooey@hirschkaefer.de>]
2295 *) New make target "install_html_docs" installs HTML renditions of the
2297 [Oliver Tappe <zooey@hirschkaefer.de>]
2299 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2300 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2301 support key and parameter generation and add initial key generation
2302 functionality for RSA.
2305 *) Add functions for main EVP_PKEY_method operations. The undocumented
2306 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2307 EVP_PKEY_{encrypt,decrypt}_old.
2310 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2311 key API, doesn't do much yet.
2314 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2315 public key algorithms. New option to openssl utility:
2316 "list-public-key-algorithms" to print out info.
2319 *) Implement the Supported Elliptic Curves Extension for
2320 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2323 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2324 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2327 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2328 utilities such as rsa, dsa, dsaparam etc except they process any key
2332 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2333 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2334 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2338 *) Initial support for pluggable public key ASN1.
2339 De-spaghettify the public key ASN1 handling. Move public and private
2340 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2341 algorithm specific handling to a single module within the relevant
2342 algorithm directory. Add functions to allow (near) opaque processing
2343 of public and private key structures.
2346 *) Implement the Supported Point Formats Extension for
2347 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2350 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2351 for the psk identity [hint] and the psk callback functions to the
2352 SSL_SESSION, SSL and SSL_CTX structure.
2355 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2359 SSL_CTX_use_psk_identity_hint
2360 SSL_get_psk_identity_hint
2361 SSL_get_psk_identity
2362 SSL_use_psk_identity_hint
2364 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2366 *) Add RFC 3161 compliant time stamp request creation, response generation
2367 and response verification functionality.
2368 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2370 *) Add initial support for TLS extensions, specifically for the server_name
2371 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2372 have new members for a host name. The SSL data structure has an
2373 additional member SSL_CTX *initial_ctx so that new sessions can be
2374 stored in that context to allow for session resumption, even after the
2375 SSL has been switched to a new SSL_CTX in reaction to a client's
2376 server_name extension.
2378 New functions (subject to change):
2380 SSL_get_servername()
2381 SSL_get_servername_type()
2384 New CTRL codes and macros (subject to change):
2386 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2387 - SSL_CTX_set_tlsext_servername_callback()
2388 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2389 - SSL_CTX_set_tlsext_servername_arg()
2390 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2392 openssl s_client has a new '-servername ...' option.
2394 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2395 '-key2 ...', '-servername_fatal' (subject to change). This allows
2396 testing the HostName extension for a specific single host name ('-cert'
2397 and '-key' remain fallbacks for handshakes without HostName
2398 negotiation). If the unrecognized_name alert has to be sent, this by
2399 default is a warning; it becomes fatal with the '-servername_fatal'
2402 [Peter Sylvester, Remy Allais, Christophe Renou]
2404 *) Whirlpool hash implementation is added.
2407 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2408 bn(64,32). Because of instruction set limitations it doesn't have
2409 any negative impact on performance. This was done mostly in order
2410 to make it possible to share assembler modules, such as bn_mul_mont
2411 implementations, between 32- and 64-bit builds without hassle.
2414 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2415 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2419 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2420 dedicated Montgomery multiplication procedure, is introduced.
2421 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2422 "64-bit" performance on certain 32-bit targets.
2425 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2426 in SSL structures. New SSL ctrl to set maximum send fragment size.
2427 Save memory by seeting the I/O buffer sizes dynamically instead of
2428 using the maximum available value.
2431 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2432 in addition to the text details.
2435 *) Very, very preliminary EXPERIMENTAL support for printing of general
2436 ASN1 structures. This currently produces rather ugly output and doesn't
2437 handle several customised structures at all.
2440 *) Integrated support for PVK file format and some related formats such
2441 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2442 these in the 'rsa' and 'dsa' utilities.
2445 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2448 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2449 place for the (very old) "NETSCAPE" format certificates which are now
2450 handled using new ASN1 code equivalents.
2453 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2454 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2455 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2458 *) Modify CRL distribution points extension code to print out previously
2459 unsupported fields. Enhance extension setting code to allow setting of
2463 *) Add print and set support for Issuing Distribution Point CRL extension.
2466 *) Change 'Configure' script to enable Camellia by default.
2469 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2471 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2473 This addresses the flaw in CBC record processing discovered by
2474 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2475 at: http://www.isg.rhul.ac.uk/tls/
2477 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2478 Security Group at Royal Holloway, University of London
2479 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2480 Emilia Käsper for the initial patch.
2482 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2484 *) Return an error when checking OCSP signatures when key is NULL.
2485 This fixes a DoS attack. (CVE-2013-0166)
2488 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2489 the right response is stapled. Also change SSL_get_certificate()
2490 so it returns the certificate actually sent.
2491 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2492 (This is a backport)
2493 [Rob Stradling <rob.stradling@comodo.com>]
2495 *) Fix possible deadlock when decoding public keys.
2498 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2500 *) Sanity check record length before skipping explicit IV in DTLS
2503 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2504 fuzzing as a service testing platform.
2508 *) Initialise tkeylen properly when encrypting CMS messages.
2509 Thanks to Solar Designer of Openwall for reporting this issue.
2512 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2514 *) The fix for CVE-2012-2110 did not take into account that the
2515 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2516 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2517 rejecting negative len parameter. (CVE-2012-2131)
2518 [Tomas Hoger <thoger@redhat.com>]
2520 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2522 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2523 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2524 in CRYPTO_realloc_clean.
2526 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2527 issue and to Adam Langley <agl@chromium.org> for fixing it.
2529 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2531 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2533 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2534 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2535 content decryption and always return the same error. Note: this attack
2536 needs on average 2^20 messages so it only affects automated senders. The
2537 old behaviour can be reenabled in the CMS code by setting the
2538 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2539 an MMA defence is not necessary.
2540 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2541 this issue. (CVE-2012-0884)
2544 *) Fix CVE-2011-4619: make sure we really are receiving a
2545 client hello before rejecting multiple SGC restarts. Thanks to
2546 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2549 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2551 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2552 Thanks to Antonio Martin, Enterprise Secure Access Research and
2553 Development, Cisco Systems, Inc. for discovering this bug and
2554 preparing a fix. (CVE-2012-0050)
2557 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2559 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2560 of the Vaudenay padding oracle attack on CBC mode encryption
2561 which enables an efficient plaintext recovery attack against
2562 the OpenSSL implementation of DTLS. Their attack exploits timing
2563 differences arising during decryption processing. A research
2564 paper describing this attack can be found at:
2565 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2566 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2567 Security Group at Royal Holloway, University of London
2568 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2569 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2570 for preparing the fix. (CVE-2011-4108)
2571 [Robin Seggelmann, Michael Tuexen]
2573 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2574 [Ben Laurie, Kasper <ekasper@google.com>]
2576 *) Clear bytes used for block padding of SSL 3.0 records.
2578 [Adam Langley (Google)]
2580 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2581 Kadianakis <desnacked@gmail.com> for discovering this issue and
2582 Adam Langley for preparing the fix. (CVE-2011-4619)
2583 [Adam Langley (Google)]
2585 *) Prevent malformed RFC3779 data triggering an assertion failure.
2586 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2587 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2588 [Rob Austein <sra@hactrn.net>]
2590 *) Fix ssl_ciph.c set-up race.
2591 [Adam Langley (Google)]
2593 *) Fix spurious failures in ecdsatest.c.
2594 [Emilia Käsper (Google)]
2596 *) Fix the BIO_f_buffer() implementation (which was mixing different
2597 interpretations of the '..._len' fields).
2598 [Adam Langley (Google)]
2600 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2601 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2602 threads won't reuse the same blinding coefficients.
2604 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2605 lock to call BN_BLINDING_invert_ex, and avoids one use of
2606 BN_BLINDING_update for each BN_BLINDING structure (previously,
2607 the last update always remained unused).
2608 [Emilia Käsper (Google)]
2610 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2611 for multi-threaded use of ECDH.
2612 [Adam Langley (Google)]
2614 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2617 *) Add protection against ECDSA timing attacks as mentioned in the paper
2618 by Billy Bob Brumley and Nicola Tuveri, see:
2620 http://eprint.iacr.org/2011/232.pdf
2622 [Billy Bob Brumley and Nicola Tuveri]
2624 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2626 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2627 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2629 *) Fix bug in string printing code: if *any* escaping is enabled we must
2630 escape the escape character (backslash) or the resulting string is
2634 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2636 *) Disable code workaround for ancient and obsolete Netscape browsers
2637 and servers: an attacker can use it in a ciphersuite downgrade attack.
2638 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2641 *) Fixed J-PAKE implementation error, originally discovered by
2642 Sebastien Martini, further info and confirmation from Stefan
2643 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2646 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2648 *) Fix extension code to avoid race conditions which can result in a buffer
2649 overrun vulnerability: resumed sessions must not be modified as they can
2650 be shared by multiple threads. CVE-2010-3864
2653 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2656 *) Don't reencode certificate when calculating signature: cache and use
2657 the original encoding instead. This makes signature verification of
2658 some broken encodings work correctly.
2661 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2662 is also one of the inputs.
2663 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2665 *) Don't repeatedly append PBE algorithms to table if they already exist.
2666 Sort table on each new add. This effectively makes the table read only
2667 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2671 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2673 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2676 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2677 access or freeing data twice (CVE-2010-0742)
2678 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2680 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2681 common in certificates and some applications which only call
2682 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2686 Reduce copying into .apps and .test in makevms.com
2687 Don't try to use blank CA certificate in CA.com
2688 Allow use of C files from original directories in maketests.com
2689 [Steven M. Schweda" <sms@antinode.info>]
2691 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2693 *) When rejecting SSL/TLS records due to an incorrect version number, never
2694 update s->server with a new major version number. As of
2695 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2696 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2697 the previous behavior could result in a read attempt at NULL when
2698 receiving specific incorrect SSL/TLS records once record payload
2699 protection is active. (CVE-2010-0740)
2700 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2702 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2703 could be crashed if the relevant tables were not present (e.g. chrooted).
2704 [Tomas Hoger <thoger@redhat.com>]
2706 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2708 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2709 [Martin Olsson, Neel Mehta]
2711 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2712 accommodate for stack sorting, always a write lock!).
2715 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2716 excessive delays in the RAND_poll(): over a minute. As a workaround
2717 include a time check in the inner Heap32Next loop too.
2720 *) The code that handled flushing of data in SSL/TLS originally used the
2721 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2722 the problem outlined in PR#1949. The fix suggested there however can
2723 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2724 of Apache). So instead simplify the code to flush unconditionally.
2725 This should be fine since flushing with no data to flush is a no op.
2728 *) Handle TLS versions 2.0 and later properly and correctly use the
2729 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2730 off ancient servers have a habit of sticking around for a while...
2733 *) Modify compression code so it frees up structures without using the
2734 ex_data callbacks. This works around a problem where some applications
2735 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2736 restarting) then use compression (e.g. SSL with compression) later.
2737 This results in significant per-connection memory leaks and
2738 has caused some security issues including CVE-2008-1678 and
2742 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2743 change when encrypting or decrypting.
2746 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2747 connect and renegotiate with servers which do not support RI.
2748 Until RI is more widely deployed this option is enabled by default.
2751 *) Add "missing" ssl ctrls to clear options and mode.
2754 *) If client attempts to renegotiate and doesn't support RI respond with
2755 a no_renegotiation alert as required by RFC5746. Some renegotiating
2756 TLS clients will continue a connection gracefully when they receive
2757 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2758 waiting for a server hello which it will never receive. Now we treat a
2759 received no_renegotiation alert as a fatal error. This is because
2760 applications requesting a renegotiation might well expect it to succeed
2761 and would have no code in place to handle the server denying it so the
2762 only safe thing to do is to terminate the connection.
2765 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2766 peer supports secure renegotiation and 0 otherwise. Print out peer
2767 renegotiation support in s_client/s_server.
2770 *) Replace the highly broken and deprecated SPKAC certification method with
2771 the updated NID creation version. This should correctly handle UTF8.
2774 *) Implement RFC5746. Re-enable renegotiation but require the extension
2775 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2776 turns out to be a bad idea. It has been replaced by
2777 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2778 SSL_CTX_set_options(). This is really not recommended unless you
2779 know what you are doing.
2780 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2782 *) Fixes to stateless session resumption handling. Use initial_ctx when
2783 issuing and attempting to decrypt tickets in case it has changed during
2784 servername handling. Use a non-zero length session ID when attempting
2785 stateless session resumption: this makes it possible to determine if
2786 a resumption has occurred immediately after receiving server hello
2787 (several places in OpenSSL subtly assume this) instead of later in
2791 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2792 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2793 fixes for a few places where the return code is not checked
2795 [Julia Lawall <julia@diku.dk>]
2797 *) Add --strict-warnings option to Configure script to include devteam
2798 warnings in other configurations.
2801 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2802 makes it possible to install openssl libraries in locations which
2803 have names other than "lib", for example "/usr/lib64" which some
2805 [Steve Henson, based on patch from Jeremy Utley]
2807 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2808 X690 8.9.12 and can produce some misleading textual output of OIDs.
2809 [Steve Henson, reported by Dan Kaminsky]
2811 *) Delete MD2 from algorithm tables. This follows the recommendation in
2812 several standards that it is not used in new applications due to
2813 several cryptographic weaknesses. For binary compatibility reasons
2814 the MD2 API is still compiled in by default.
2817 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2821 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2822 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2824 [Guenter <lists@gknw.net>]
2826 *) Fix the server certificate chain building code to use X509_verify_cert(),
2827 it used to have an ad-hoc builder which was unable to cope with anything
2828 other than a simple chain.
2829 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2831 *) Don't check self signed certificate signatures in X509_verify_cert()
2832 by default (a flag can override this): it just wastes time without
2833 adding any security. As a useful side effect self signed root CAs
2834 with non-FIPS digests are now usable in FIPS mode.
2837 *) In dtls1_process_out_of_seq_message() the check if the current message
2838 is already buffered was missing. For every new message was memory
2839 allocated, allowing an attacker to perform an denial of service attack
2840 with sending out of seq handshake messages until there is no memory
2841 left. Additionally every future messege was buffered, even if the
2842 sequence number made no sense and would be part of another handshake.
2843 So only messages with sequence numbers less than 10 in advance will be
2844 buffered. (CVE-2009-1378)
2845 [Robin Seggelmann, discovered by Daniel Mentz]
2847 *) Records are buffered if they arrive with a future epoch to be
2848 processed after finishing the corresponding handshake. There is
2849 currently no limitation to this buffer allowing an attacker to perform
2850 a DOS attack with sending records with future epochs until there is no
2851 memory left. This patch adds the pqueue_size() function to determine
2852 the size of a buffer and limits the record buffer to 100 entries.
2854 [Robin Seggelmann, discovered by Daniel Mentz]
2856 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2857 parent structure is freed. (CVE-2009-1379)
2860 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2861 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2864 [Ilya O. <vrghost@gmail.com>]
2866 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2868 *) Disable renegotiation completely - this fixes a severe security
2869 problem (CVE-2009-3555) at the cost of breaking all
2870 renegotiation. Renegotiation can be re-enabled by setting
2871 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2872 run-time. This is really not recommended unless you know what
2876 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2878 *) Don't set val to NULL when freeing up structures, it is freed up by
2879 underlying code. If sizeof(void *) > sizeof(long) this can result in
2880 zeroing past the valid field. (CVE-2009-0789)
2881 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2883 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2884 checked correctly. This would allow some invalid signed attributes to
2885 appear to verify correctly. (CVE-2009-0591)
2886 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2888 *) Reject UniversalString and BMPString types with invalid lengths. This
2889 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2890 a legal length. (CVE-2009-0590)
2893 *) Set S/MIME signing as the default purpose rather than setting it
2894 unconditionally. This allows applications to override it at the store
2898 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2899 to handle some structures.
2902 *) Improve efficiency of mem_gets: don't search whole buffer each time
2904 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2906 *) New -hex option for openssl rand.
2909 *) Print out UTF8String and NumericString when parsing ASN1.
2912 *) Support NumericString type for name components.
2915 *) Allow CC in the environment to override the automatically chosen
2916 compiler. Note that nothing is done to ensure flags work with the
2920 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2922 *) Properly check EVP_VerifyFinal() and similar return values
2924 [Ben Laurie, Bodo Moeller, Google Security Team]
2926 *) Enable TLS extensions by default.
2929 *) Allow the CHIL engine to be loaded, whether the application is
2930 multithreaded or not. (This does not release the developer from the
2931 obligation to set up the dynamic locking callbacks.)
2932 [Sander Temme <sander@temme.net>]
2934 *) Use correct exit code if there is an error in dgst command.
2935 [Steve Henson; problem pointed out by Roland Dirlewanger]
2937 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2938 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2941 *) Add experimental JPAKE support, including demo authentication in
2942 s_client and s_server.
2945 *) Set the comparison function in v3_addr_canonize().
2946 [Rob Austein <sra@hactrn.net>]
2948 *) Add support for XMPP STARTTLS in s_client.
2949 [Philip Paeps <philip@freebsd.org>]
2951 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2952 to ensure that even with this option, only ciphersuites in the
2953 server's preference list will be accepted. (Note that the option
2954 applies only when resuming a session, so the earlier behavior was
2955 just about the algorithm choice for symmetric cryptography.)
2958 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2960 *) Fix NULL pointer dereference if a DTLS server received
2961 ChangeCipherSpec as first record (CVE-2009-1386).
2964 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2965 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2968 *) The fix in 0.9.8c that supposedly got rid of unsafe
2969 double-checked locking was incomplete for RSA blinding,
2970 addressing just one layer of what turns out to have been
2971 doubly unsafe triple-checked locking.
2973 So now fix this for real by retiring the MONT_HELPER macro
2974 in crypto/rsa/rsa_eay.c.
2976 [Bodo Moeller; problem pointed out by Marius Schilder]
2978 *) Various precautionary measures:
2980 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2982 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2983 (NB: This would require knowledge of the secret session ticket key
2984 to exploit, in which case you'd be SOL either way.)
2986 - Change bn_nist.c so that it will properly handle input BIGNUMs
2987 outside the expected range.
2989 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2992 [Neel Mehta, Bodo Moeller]
2994 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2995 the load fails. Useful for distros.
2996 [Ben Laurie and the FreeBSD team]
2998 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
3001 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
3004 *) Expand ENGINE to support engine supplied SSL client certificate functions.
3006 This work was sponsored by Logica.
3009 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
3010 keystores. Support for SSL/TLS client authentication too.
3011 Not compiled unless enable-capieng specified to Configure.
3013 This work was sponsored by Logica.
3016 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
3017 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
3018 attribute creation routines such as certifcate requests and PKCS#12
3022 Changes between 0.9.8g and 0.9.8h [28 May 2008]
3024 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
3025 handshake which could lead to a cilent crash as found using the
3026 Codenomicon TLS test suite (CVE-2008-1672)
3027 [Steve Henson, Mark Cox]
3029 *) Fix double free in TLS server name extensions which could lead to
3030 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
3033 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
3035 Clear the error queue to ensure that error entries left from
3036 older function calls do not interfere with the correct operation.
3037 [Lutz Jaenicke, Erik de Castro Lopo]
3039 *) Remove root CA certificates of commercial CAs:
3041 The OpenSSL project does not recommend any specific CA and does not
3042 have any policy with respect to including or excluding any CA.
3043 Therefore it does not make any sense to ship an arbitrary selection
3044 of root CA certificates with the OpenSSL software.
3047 *) RSA OAEP patches to fix two separate invalid memory reads.
3048 The first one involves inputs when 'lzero' is greater than
3049 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
3050 before the beginning of from). The second one involves inputs where
3051 the 'db' section contains nothing but zeroes (there is a one-byte
3052 invalid read after the end of 'db').
3053 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3055 *) Partial backport from 0.9.9-dev:
3057 Introduce bn_mul_mont (dedicated Montgomery multiplication
3058 procedure) as a candidate for BIGNUM assembler implementation.
3059 While 0.9.9-dev uses assembler for various architectures, only
3060 x86_64 is available by default here in the 0.9.8 branch, and
3061 32-bit x86 is available through a compile-time setting.
3063 To try the 32-bit x86 assembler implementation, use Configure
3064 option "enable-montasm" (which exists only for this backport).
3066 As "enable-montasm" for 32-bit x86 disclaims code stability
3067 anyway, in this constellation we activate additional code
3068 backported from 0.9.9-dev for further performance improvements,
3069 namely BN_from_montgomery_word. (To enable this otherwise,
3070 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
3072 [Andy Polyakov (backport partially by Bodo Moeller)]
3074 *) Add TLS session ticket callback. This allows an application to set
3075 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
3076 values. This is useful for key rollover for example where several key
3077 sets may exist with different names.
3080 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
3081 This was broken until now in 0.9.8 releases, such that the only way
3082 a registered ENGINE could be used (assuming it initialises
3083 successfully on the host) was to explicitly set it as the default
3084 for the relevant algorithms. This is in contradiction with 0.9.7
3085 behaviour and the documentation. With this fix, when an ENGINE is
3086 registered into a given algorithm's table of implementations, the
3087 'uptodate' flag is reset so that auto-discovery will be used next
3088 time a new context for that algorithm attempts to select an
3090 [Ian Lister (tweaked by Geoff Thorpe)]
3092 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
3093 implemention in the following ways:
3095 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
3098 Lack of BER streaming support means one pass streaming processing is
3099 only supported if data is detached: setting the streaming flag is
3100 ignored for embedded content.
3102 CMS support is disabled by default and must be explicitly enabled
3103 with the enable-cms configuration option.
3106 *) Update the GMP engine glue to do direct copies between BIGNUM and
3107 mpz_t when openssl and GMP use the same limb size. Otherwise the
3108 existing "conversion via a text string export" trick is still used.
3109 [Paul Sheer <paulsheer@gmail.com>]
3111 *) Zlib compression BIO. This is a filter BIO which compressed and
3112 uncompresses any data passed through it.
3115 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
3116 RFC3394 compatible AES key wrapping.
3119 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
3120 sets string data without copying. X509_ALGOR_set0() and
3121 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
3122 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
3123 from an X509_ATTRIBUTE structure optionally checking it occurs only
3124 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
3128 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
3129 to get the expected BN_FLG_CONSTTIME behavior.
3130 [Bodo Moeller (Google)]
3134 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3135 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3136 - added some more tests to do_tests.pl
3137 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3138 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3139 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3140 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3141 - various changes to netware.pl to enable gcc-cross builds on Win32
3143 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3144 - various changes to fix missing prototype warnings
3145 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3146 - added AES, WHIRLPOOL and CPUID assembler code to build files
3147 - added missing AES assembler make rules to mk1mf.pl
3148 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3149 [Guenter Knauf <eflash@gmx.net>]
3151 *) Implement certificate status request TLS extension defined in RFC3546.
3152 A client can set the appropriate parameters and receive the encoded
3153 OCSP response via a callback. A server can query the supplied parameters
3154 and set the encoded OCSP response in the callback. Add simplified examples
3155 to s_client and s_server.
3158 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3160 *) Fix various bugs:
3161 + Binary incompatibility of ssl_ctx_st structure
3162 + DTLS interoperation with non-compliant servers
3163 + Don't call get_session_cb() without proposed session
3164 + Fix ia64 assembler code
3165 [Andy Polyakov, Steve Henson]
3167 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3169 *) DTLS Handshake overhaul. There were longstanding issues with
3170 OpenSSL DTLS implementation, which were making it impossible for
3171 RFC 4347 compliant client to communicate with OpenSSL server.
3172 Unfortunately just fixing these incompatibilities would "cut off"
3173 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3174 server keeps tolerating non RFC compliant syntax. The opposite is
3175 not true, 0.9.8f client can not communicate with earlier server.
3176 This update even addresses CVE-2007-4995.
3179 *) Changes to avoid need for function casts in OpenSSL: some compilers
3180 (gcc 4.2 and later) reject their use.
3181 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3184 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3185 RFC4507bis. The encrypted ticket format is an encrypted encoded
3186 SSL_SESSION structure, that way new session features are automatically
3189 If a client application caches session in an SSL_SESSION structure
3190 support is transparent because tickets are now stored in the encoded
3193 The SSL_CTX structure automatically generates keys for ticket
3194 protection in servers so again support should be possible
3195 with no application modification.
3197 If a client or server wishes to disable RFC4507 support then the option
3198 SSL_OP_NO_TICKET can be set.
3200 Add a TLS extension debugging callback to allow the contents of any client
3201 or server extensions to be examined.
3203 This work was sponsored by Google.
3206 *) Add initial support for TLS extensions, specifically for the server_name
3207 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3208 have new members for a host name. The SSL data structure has an
3209 additional member SSL_CTX *initial_ctx so that new sessions can be
3210 stored in that context to allow for session resumption, even after the
3211 SSL has been switched to a new SSL_CTX in reaction to a client's
3212 server_name extension.
3214 New functions (subject to change):
3216 SSL_get_servername()
3217 SSL_get_servername_type()
3220 New CTRL codes and macros (subject to change):
3222 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3223 - SSL_CTX_set_tlsext_servername_callback()
3224 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3225 - SSL_CTX_set_tlsext_servername_arg()
3226 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3228 openssl s_client has a new '-servername ...' option.
3230 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3231 '-key2 ...', '-servername_fatal' (subject to change). This allows
3232 testing the HostName extension for a specific single host name ('-cert'
3233 and '-key' remain fallbacks for handshakes without HostName
3234 negotiation). If the unrecognized_name alert has to be sent, this by
3235 default is a warning; it becomes fatal with the '-servername_fatal'
3238 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3240 *) Add AES and SSE2 assembly language support to VC++ build.
3243 *) Mitigate attack on final subtraction in Montgomery reduction.
3246 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3247 (which previously caused an internal error).
3250 *) Squeeze another 10% out of IGE mode when in != out.
3253 *) AES IGE mode speedup.
3254 [Dean Gaudet (Google)]
3256 *) Add the Korean symmetric 128-bit cipher SEED (see
3257 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3258 add SEED ciphersuites from RFC 4162:
3260 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3261 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3262 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3263 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3265 To minimize changes between patchlevels in the OpenSSL 0.9.8
3266 series, SEED remains excluded from compilation unless OpenSSL
3267 is configured with 'enable-seed'.
3268 [KISA, Bodo Moeller]
3270 *) Mitigate branch prediction attacks, which can be practical if a
3271 single processor is shared, allowing a spy process to extract
3272 information. For detailed background information, see
3273 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3274 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3275 and Necessary Software Countermeasures"). The core of the change
3276 are new versions BN_div_no_branch() and
3277 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3278 respectively, which are slower, but avoid the security-relevant
3279 conditional branches. These are automatically called by BN_div()
3280 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3281 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3282 remove a conditional branch.
3284 BN_FLG_CONSTTIME is the new name for the previous
3285 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3286 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3287 in the exponent causes BN_mod_exp_mont() to use the alternative
3288 implementation in BN_mod_exp_mont_consttime().) The old name
3289 remains as a deprecated alias.
3291 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3292 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3293 constant-time implementations for more than just exponentiation.
3294 Here too the old name is kept as a deprecated alias.
3296 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3297 the BN_BLINDING structure gets an independent copy of the
3298 modulus. This means that the previous "BIGNUM *m" argument to
3299 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3300 essentially becomes "const BIGNUM *m", although we can't actually
3301 change this in the header file before 0.9.9. It allows
3302 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3303 enable BN_FLG_CONSTTIME.
3305 [Matthew D Wood (Intel Corp)]
3307 *) In the SSL/TLS server implementation, be strict about session ID
3308 context matching (which matters if an application uses a single
3309 external cache for different purposes). Previously,
3310 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3311 set. This did ensure strict client verification, but meant that,
3312 with applications using a single external cache for quite
3313 different requirements, clients could circumvent ciphersuite
3314 restrictions for a given session ID context by starting a session
3315 in a different context.
3318 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3319 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3320 authentication-only ciphersuites.
3323 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3324 not complete and could lead to a possible single byte overflow
3325 (CVE-2007-5135) [Ben Laurie]
3327 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3329 *) Since AES128 and AES256 (and similarly Camellia128 and
3330 Camellia256) share a single mask bit in the logic of
3331 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3332 kludge to work properly if AES128 is available and AES256 isn't
3333 (or if Camellia128 is available and Camellia256 isn't).
3336 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3337 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3338 When a point or a seed is encoded in a BIT STRING, we need to
3339 prevent the removal of trailing zero bits to get the proper DER
3340 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3341 of a NamedBitList, for which trailing 0 bits need to be removed.)
3344 *) Have SSL/TLS server implementation tolerate "mismatched" record
3345 protocol version while receiving ClientHello even if the
3346 ClientHello is fragmented. (The server can't insist on the
3347 particular protocol version it has chosen before the ServerHello
3348 message has informed the client about his choice.)
3351 *) Add RFC 3779 support.
3352 [Rob Austein for ARIN, Ben Laurie]
3354 *) Load error codes if they are not already present instead of using a
3355 static variable. This allows them to be cleanly unloaded and reloaded.
3356 Improve header file function name parsing.
3359 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3360 or CAPABILITY handshake as required by RFCs.
3363 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3365 *) Introduce limits to prevent malicious keys being able to
3366 cause a denial of service. (CVE-2006-2940)
3367 [Steve Henson, Bodo Moeller]
3369 *) Fix ASN.1 parsing of certain invalid structures that can result
3370 in a denial of service. (CVE-2006-2937) [Steve Henson]
3372 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3373 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3375 *) Fix SSL client code which could crash if connecting to a
3376 malicious SSLv2 server. (CVE-2006-4343)
3377 [Tavis Ormandy and Will Drewry, Google Security Team]
3379 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3380 match only those. Before that, "AES256-SHA" would be interpreted
3381 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3382 the same strength classification in 0.9.7h) as we currently only
3383 have a single AES bit in the ciphersuite description bitmap.
3384 That change, however, also applied to ciphersuite strings such as
3385 "RC4-MD5" that intentionally matched multiple ciphersuites --
3386 namely, SSL 2.0 ciphersuites in addition to the more common ones
3387 from SSL 3.0/TLS 1.0.
3389 So we change the selection algorithm again: Naming an explicit
3390 ciphersuite selects this one ciphersuite, and any other similar
3391 ciphersuite (same bitmap) from *other* protocol versions.
3392 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3393 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3395 Since SSL 2.0 does not have any ciphersuites for which the
3396 128/256 bit distinction would be relevant, this works for now.
3397 The proper fix will be to use different bits for AES128 and
3398 AES256, which would have avoided the problems from the beginning;
3399 however, bits are scarce, so we can only do this in a new release
3400 (not just a patchlevel) when we can change the SSL_CIPHER
3401 definition to split the single 'unsigned long mask' bitmap into
3402 multiple values to extend the available space.
3406 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3408 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3409 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3411 *) Add AES IGE and biIGE modes.
3414 *) Change the Unix randomness entropy gathering to use poll() when
3415 possible instead of select(), since the latter has some
3416 undesirable limitations.
3417 [Darryl Miles via Richard Levitte and Bodo Moeller]
3419 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3420 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3421 cannot be implicitly activated as part of, e.g., the "AES" alias.
3422 However, please upgrade to OpenSSL 0.9.9[-dev] for
3423 non-experimental use of the ECC ciphersuites to get TLS extension
3424 support, which is required for curve and point format negotiation
3425 to avoid potential handshake problems.
3428 *) Disable rogue ciphersuites:
3430 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3431 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3432 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3434 The latter two were purportedly from
3435 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3438 Also deactivate the remaining ciphersuites from
3439 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3440 unofficial, and the ID has long expired.
3443 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3444 dual-core machines) and other potential thread-safety issues.
3447 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3448 versions), which is now available for royalty-free use
3449 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3450 Also, add Camellia TLS ciphersuites from RFC 4132.
3452 To minimize changes between patchlevels in the OpenSSL 0.9.8
3453 series, Camellia remains excluded from compilation unless OpenSSL
3454 is configured with 'enable-camellia'.
3457 *) Disable the padding bug check when compression is in use. The padding
3458 bug check assumes the first packet is of even length, this is not
3459 necessarily true if compresssion is enabled and can result in false
3460 positives causing handshake failure. The actual bug test is ancient
3461 code so it is hoped that implementations will either have fixed it by
3462 now or any which still have the bug do not support compression.
3465 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3467 *) When applying a cipher rule check to see if string match is an explicit
3468 cipher suite and only match that one cipher suite if it is.
3471 *) Link in manifests for VC++ if needed.
3472 [Austin Ziegler <halostatue@gmail.com>]
3474 *) Update support for ECC-based TLS ciphersuites according to
3475 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3476 TLS extensions, which are supported starting with the 0.9.9
3477 branch, not in the OpenSSL 0.9.8 branch).
3480 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3481 opaque EVP_CIPHER_CTX handling.
3484 *) Fixes and enhancements to zlib compression code. We now only use
3485 "zlib1.dll" and use the default __cdecl calling convention on Win32
3486 to conform with the standards mentioned here:
3487 http://www.zlib.net/DLL_FAQ.txt
3488 Static zlib linking now works on Windows and the new --with-zlib-include
3489 --with-zlib-lib options to Configure can be used to supply the location
3490 of the headers and library. Gracefully handle case where zlib library
3494 *) Several fixes and enhancements to the OID generation code. The old code
3495 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3496 handle numbers larger than ULONG_MAX, truncated printing and had a
3497 non standard OBJ_obj2txt() behaviour.
3500 *) Add support for building of engines under engine/ as shared libraries
3501 under VC++ build system.
3504 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3505 Hopefully, we will not see any false combination of paths any more.
3508 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3510 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3511 (part of SSL_OP_ALL). This option used to disable the
3512 countermeasure against man-in-the-middle protocol-version
3513 rollback in the SSL 2.0 server implementation, which is a bad
3514 idea. (CVE-2005-2969)
3516 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3517 for Information Security, National Institute of Advanced Industrial
3518 Science and Technology [AIST], Japan)]
3520 *) Add two function to clear and return the verify parameter flags.
3523 *) Keep cipherlists sorted in the source instead of sorting them at
3524 runtime, thus removing the need for a lock.
3527 *) Avoid some small subgroup attacks in Diffie-Hellman.
3528 [Nick Mathewson and Ben Laurie]
3530 *) Add functions for well-known primes.
3533 *) Extended Windows CE support.
3534 [Satoshi Nakamura and Andy Polyakov]
3536 *) Initialize SSL_METHOD structures at compile time instead of during
3537 runtime, thus removing the need for a lock.
3540 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3541 attempting to decrypt each encrypted key in turn. Add support to
3545 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3547 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3550 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3553 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3554 key into the same file any more.
3557 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3560 *) Add -utf8 command line and config file option to 'ca'.
3561 [Stefan <stf@udoma.org]
3563 *) Removed the macro des_crypt(), as it seems to conflict with some
3564 libraries. Use DES_crypt().
3567 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3568 involves renaming the source and generated shared-libs for
3569 both. The engines will accept the corrected or legacy ids
3570 ('ncipher' and '4758_cca' respectively) when binding. NB,
3571 this only applies when building 'shared'.
3572 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3574 *) Add attribute functions to EVP_PKEY structure. Modify
3575 PKCS12_create() to recognize a CSP name attribute and
3576 use it. Make -CSP option work again in pkcs12 utility.
3579 *) Add new functionality to the bn blinding code:
3580 - automatic re-creation of the BN_BLINDING parameters after
3581 a fixed number of uses (currently 32)
3582 - add new function for parameter creation
3583 - introduce flags to control the update behaviour of the
3584 BN_BLINDING parameters
3585 - hide BN_BLINDING structure
3586 Add a second BN_BLINDING slot to the RSA structure to improve
3587 performance when a single RSA object is shared among several
3591 *) Add support for DTLS.
3592 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3594 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3595 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3598 *) Remove buggy and incomplete DH cert support from
3599 ssl/ssl_rsa.c and ssl/s3_both.c
3602 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3603 the apps/openssl applications.
3606 *) Compile clean with "-Wall -Wmissing-prototypes
3607 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3608 DEBUG_SAFESTACK must also be set.
3611 *) Change ./Configure so that certain algorithms can be disabled by default.
3612 The new counterpiece to "no-xxx" is "enable-xxx".
3614 The patented RC5 and MDC2 algorithms will now be disabled unless
3615 "enable-rc5" and "enable-mdc2", respectively, are specified.
3617 (IDEA remains enabled despite being patented. This is because IDEA
3618 is frequently required for interoperability, and there is no license
3619 fee for non-commercial use. As before, "no-idea" can be used to
3620 avoid this algorithm.)
3624 *) Add processing of proxy certificates (see RFC 3820). This work was
3625 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3626 EGEE (Enabling Grids for E-science in Europe).
3629 *) RC4 performance overhaul on modern architectures/implementations, such
3630 as Intel P4, IA-64 and AMD64.
3633 *) New utility extract-section.pl. This can be used specify an alternative
3634 section number in a pod file instead of having to treat each file as
3635 a separate case in Makefile. This can be done by adding two lines to the
3638 =for comment openssl_section:XXX
3640 The blank line is mandatory.
3644 *) New arguments -certform, -keyform and -pass for s_client and s_server
3645 to allow alternative format key and certificate files and passphrase
3649 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3650 update associated structures and add various utility functions.
3652 Add new policy related verify parameters, include policy checking in
3653 standard verify code. Enhance 'smime' application with extra parameters
3654 to support policy checking and print out.
3657 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3658 Nehemiah processors. These extensions support AES encryption in hardware
3659 as well as RNG (though RNG support is currently disabled).
3660 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3662 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3665 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3666 [Andy Polyakov and a number of other people]
3668 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3669 implementation contributed by IBM.
3670 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3672 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3673 exponent rather than 'unsigned long'. There is a corresponding change to
3674 the new 'rsa_keygen' element of the RSA_METHOD structure.
3675 [Jelte Jansen, Geoff Thorpe]
3677 *) Functionality for creating the initial serial number file is now
3678 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3680 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3681 number file to 1, which is bound to cause problems. To avoid
3682 the problems while respecting compatibility between different 0.9.7
3683 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3684 CA.pl for serial number initialization. With the new release 0.9.8,
3685 we can fix the problem directly in the 'ca' utility.)
3688 *) Reduced header interdepencies by declaring more opaque objects in
3689 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3690 give fewer recursive includes, which could break lazy source code - so
3691 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3692 developers should define this symbol when building and using openssl to
3693 ensure they track the recommended behaviour, interfaces, [etc], but
3694 backwards-compatible behaviour prevails when this isn't defined.
3697 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3700 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3701 This will generate a random key of the appropriate length based on the
3702 cipher context. The EVP_CIPHER can provide its own random key generation
3703 routine to support keys of a specific form. This is used in the des and
3704 3des routines to generate a key of the correct parity. Update S/MIME
3705 code to use new functions and hence generate correct parity DES keys.
3706 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3707 valid (weak or incorrect parity).
3710 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3711 as looking them up. This is useful when the verified structure may contain
3712 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3713 present unless the new PKCS7_NO_CRL flag is asserted.
3716 *) Extend ASN1 oid configuration module. It now additionally accepts the
3719 shortName = some long name, 1.2.3.4
3722 *) Reimplemented the BN_CTX implementation. There is now no more static
3723 limitation on the number of variables it can handle nor the depth of the
3724 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3725 information can now expand as required, and rather than having a single
3726 static array of bignums, BN_CTX now uses a linked-list of such arrays
3727 allowing it to expand on demand whilst maintaining the usefulness of
3728 BN_CTX's "bundling".
3731 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3732 to allow all RSA operations to function using a single BN_CTX.
3735 *) Preliminary support for certificate policy evaluation and checking. This
3736 is initially intended to pass the tests outlined in "Conformance Testing
3737 of Relying Party Client Certificate Path Processing Logic" v1.07.
3740 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3741 remained unused and not that useful. A variety of other little bignum
3742 tweaks and fixes have also been made continuing on from the audit (see
3746 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3747 associated ASN1, EVP and SSL functions and old ASN1 macros.
3750 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3751 and this should never fail. So the return value from the use of
3752 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3753 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3756 *) BN_CTX_get() should return zero-valued bignums, providing the same
3757 initialised value as BN_new().
3758 [Geoff Thorpe, suggested by Ulf Möller]
3760 *) Support for inhibitAnyPolicy certificate extension.
3763 *) An audit of the BIGNUM code is underway, for which debugging code is
3764 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3765 is considered valid when processing BIGNUMs, and causes execution to
3766 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3767 further steps are taken to deliberately pollute unused data in BIGNUM
3768 structures to try and expose faulty code further on. For now, openssl will
3769 (in its default mode of operation) continue to tolerate the inconsistent
3770 forms that it has tolerated in the past, but authors and packagers should
3771 consider trying openssl and their own applications when compiled with
3772 these debugging symbols defined. It will help highlight potential bugs in
3773 their own code, and will improve the test coverage for OpenSSL itself. At
3774 some point, these tighter rules will become openssl's default to improve
3775 maintainability, though the assert()s and other overheads will remain only
3776 in debugging configurations. See bn.h for more details.
3777 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3779 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3780 that can only be obtained through BN_CTX_new() (which implicitly
3781 initialises it). The presence of this function only made it possible
3782 to overwrite an existing structure (and cause memory leaks).
3785 *) Because of the callback-based approach for implementing LHASH as a
3786 template type, lh_insert() adds opaque objects to hash-tables and
3787 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3788 to clean up those corresponding objects before destroying the hash table
3789 (and losing the object pointers). So some over-zealous constifications in
3790 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3791 objects as "const" and the lh_doall[_arg] callback wrappers are not
3792 prototyped to have "const" restrictions on the object pointers they are
3793 given (and so aren't required to cast them away any more).
3796 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3797 (speed) prefers to use its own implementation. The two implementations
3798 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3799 its object type properly exposed (MS_TM) instead of casting to/from "char
3800 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3801 aren't necessarily the greatest nomenclatures - but this is what was used
3802 internally to the implementation so I've used that for now.
3805 *) Ensure that deprecated functions do not get compiled when
3806 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3807 the self-tests were still using deprecated key-generation functions so
3808 these have been updated also.
3811 *) Reorganise PKCS#7 code to separate the digest location functionality
3812 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3813 New function PKCS7_set_digest() to set the digest type for PKCS#7
3814 digestedData type. Add additional code to correctly generate the
3815 digestedData type and add support for this type in PKCS7 initialization
3819 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3820 structure of type "other".
3823 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3824 sure the loop does correctly stop and breaking ("division by zero")
3825 modulus operations are not performed. The (pre-generated) prime
3826 table crypto/bn/bn_prime.h was already correct, but it could not be
3827 re-generated on some platforms because of the "division by zero"
3828 situation in the script.
3829 [Ralf S. Engelschall]
3831 *) Update support for ECC-based TLS ciphersuites according to
3832 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3833 SHA-1 now is only used for "small" curves (where the
3834 representation of a field element takes up to 24 bytes); for
3835 larger curves, the field element resulting from ECDH is directly
3836 used as premaster secret.
3837 [Douglas Stebila (Sun Microsystems Laboratories)]
3839 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3840 curve secp160r1 to the tests.
3841 [Douglas Stebila (Sun Microsystems Laboratories)]
3843 *) Add the possibility to load symbols globally with DSO.
3844 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3846 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3847 control of the error stack.
3850 *) Add support for STORE in ENGINE.
3853 *) Add the STORE type. The intention is to provide a common interface
3854 to certificate and key stores, be they simple file-based stores, or
3855 HSM-type store, or LDAP stores, or...
3856 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3859 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3860 pass a list of arguments to any function as well as provide a way
3861 for a function to pass data back to the caller.
3864 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3865 works like BUF_strdup() but can be used to duplicate a portion of
3866 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3870 *) Add the function sk_find_ex() which works like sk_find(), but will
3871 return an index to an element even if an exact match couldn't be
3872 found. The index is guaranteed to point at the element where the
3873 searched-for key would be inserted to preserve sorting order.
3876 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3877 takes an extra flags argument for optional functionality. Currently,
3878 the following flags are defined:
3880 OBJ_BSEARCH_VALUE_ON_NOMATCH
3881 This one gets OBJ_bsearch_ex() to return a pointer to the first
3882 element where the comparing function returns a negative or zero
3885 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3886 This one gets OBJ_bsearch_ex() to return a pointer to the first
3887 element where the comparing function returns zero. This is useful
3888 if there are more than one element where the comparing function
3892 *) Make it possible to create self-signed certificates with 'openssl ca'
3893 in such a way that the self-signed certificate becomes part of the
3894 CA database and uses the same mechanisms for serial number generation
3895 as all other certificate signing. The new flag '-selfsign' enables
3896 this functionality. Adapt CA.sh and CA.pl.in.
3899 *) Add functionality to check the public key of a certificate request
3900 against a given private. This is useful to check that a certificate
3901 request can be signed by that key (self-signing).
3904 *) Make it possible to have multiple active certificates with the same
3905 subject in the CA index file. This is done only if the keyword
3906 'unique_subject' is set to 'no' in the main CA section (default
3907 if 'CA_default') of the configuration file. The value is saved
3908 with the database itself in a separate index attribute file,
3909 named like the index file with '.attr' appended to the name.
3912 *) Generate muti valued AVAs using '+' notation in config files for
3916 *) Support for nameConstraints certificate extension.
3919 *) Support for policyConstraints certificate extension.
3922 *) Support for policyMappings certificate extension.
3925 *) Make sure the default DSA_METHOD implementation only uses its
3926 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3927 and change its own handlers to be NULL so as to remove unnecessary
3928 indirection. This lets alternative implementations fallback to the
3929 default implementation more easily.
3932 *) Support for directoryName in GeneralName related extensions
3936 *) Make it possible to link applications using Makefile.shared.
3937 Make that possible even when linking against static libraries!
3940 *) Support for single pass processing for S/MIME signing. This now
3941 means that S/MIME signing can be done from a pipe, in addition
3942 cleartext signing (multipart/signed type) is effectively streaming
3943 and the signed data does not need to be all held in memory.
3945 This is done with a new flag PKCS7_STREAM. When this flag is set
3946 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3947 is done after the data is output (and digests calculated) in
3948 SMIME_write_PKCS7().
3951 *) Add full support for -rpath/-R, both in shared libraries and
3952 applications, at least on the platforms where it's known how
3956 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3957 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3958 will now compute a table of multiples of the generator that
3959 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3960 faster (notably in the case of a single point multiplication,
3961 scalar * generator).
3962 [Nils Larsch, Bodo Moeller]
3964 *) IPv6 support for certificate extensions. The various extensions
3965 which use the IP:a.b.c.d can now take IPv6 addresses using the
3966 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3970 *) Added an ENGINE that implements RSA by performing private key
3971 exponentiations with the GMP library. The conversions to and from
3972 GMP's mpz_t format aren't optimised nor are any montgomery forms
3973 cached, and on x86 it appears OpenSSL's own performance has caught up.
3974 However there are likely to be other architectures where GMP could
3975 provide a boost. This ENGINE is not built in by default, but it can be
3976 specified at Configure time and should be accompanied by the necessary
3977 linker additions, eg;
3978 ./config -DOPENSSL_USE_GMP -lgmp
3981 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3982 testing availability of engines with "-t" - the old behaviour is
3983 produced by increasing the feature's verbosity with "-tt".
3986 *) ECDSA routines: under certain error conditions uninitialized BN objects
3987 could be freed. Solution: make sure initialization is performed early
3988 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3992 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3993 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3994 software implementations. For DSA and DH, parameter generation can
3995 also be overriden by providing the appropriate method callbacks.
3998 *) Change the "progress" mechanism used in key-generation and
3999 primality testing to functions that take a new BN_GENCB pointer in
4000 place of callback/argument pairs. The new API functions have "_ex"
4001 postfixes and the older functions are reimplemented as wrappers for
4002 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
4003 declarations of the old functions to help (graceful) attempts to
4004 migrate to the new functions. Also, the new key-generation API
4005 functions operate on a caller-supplied key-structure and return
4006 success/failure rather than returning a key or NULL - this is to
4007 help make "keygen" another member function of RSA_METHOD etc.
4009 Example for using the new callback interface:
4011 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
4015 BN_GENCB_set(&my_cb, my_callback, my_arg);
4017 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
4018 /* For the meaning of a, b in calls to my_callback(), see the
4019 * documentation of the function that calls the callback.
4020 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
4021 * my_callback should return 1 if it wants BN_is_prime_ex()
4022 * to continue, or 0 to stop.
4027 *) Change the ZLIB compression method to be stateful, and make it
4028 available to TLS with the number defined in
4029 draft-ietf-tls-compression-04.txt.
4032 *) Add the ASN.1 structures and functions for CertificatePair, which
4033 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
4035 CertificatePair ::= SEQUENCE {
4036 forward [0] Certificate OPTIONAL,
4037 reverse [1] Certificate OPTIONAL,
4038 -- at least one of the pair shall be present -- }
4040 Also implement the PEM functions to read and write certificate
4041 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
4043 This needed to be defined, mostly for the sake of the LDAP
4044 attribute crossCertificatePair, but may prove useful elsewhere as
4048 *) Make it possible to inhibit symlinking of shared libraries in
4049 Makefile.shared, for Cygwin's sake.
4052 *) Extend the BIGNUM API by creating a function
4053 void BN_set_negative(BIGNUM *a, int neg);
4054 and a macro that behave like
4055 int BN_is_negative(const BIGNUM *a);
4057 to avoid the need to access 'a->neg' directly in applications.
4060 *) Implement fast modular reduction for pseudo-Mersenne primes
4061 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
4062 EC_GROUP_new_curve_GFp() will now automatically use this
4064 [Nils Larsch <nla@trustcenter.de>]
4066 *) Add new lock type (CRYPTO_LOCK_BN).
4069 *) Change the ENGINE framework to automatically load engines
4070 dynamically from specific directories unless they could be
4071 found to already be built in or loaded. Move all the
4072 current engines except for the cryptodev one to a new
4074 The engines in engines/ are built as shared libraries if
4075 the "shared" options was given to ./Configure or ./config.
4076 Otherwise, they are inserted in libcrypto.a.
4077 /usr/local/ssl/engines is the default directory for dynamic
4078 engines, but that can be overriden at configure time through
4079 the usual use of --prefix and/or --openssldir, and at run
4080 time with the environment variable OPENSSL_ENGINES.
4081 [Geoff Thorpe and Richard Levitte]
4083 *) Add Makefile.shared, a helper makefile to build shared
4084 libraries. Adapt Makefile.org.
4087 *) Add version info to Win32 DLLs.
4088 [Peter 'Luna' Runestig" <peter@runestig.com>]
4090 *) Add new 'medium level' PKCS#12 API. Certificates and keys
4091 can be added using this API to created arbitrary PKCS#12
4092 files while avoiding the low level API.
4094 New options to PKCS12_create(), key or cert can be NULL and
4095 will then be omitted from the output file. The encryption
4096 algorithm NIDs can be set to -1 for no encryption, the mac
4097 iteration count can be set to 0 to omit the mac.
4099 Enhance pkcs12 utility by making the -nokeys and -nocerts
4100 options work when creating a PKCS#12 file. New option -nomac
4101 to omit the mac, NONE can be set for an encryption algorithm.
4102 New code is modified to use the enhanced PKCS12_create()
4103 instead of the low level API.
4106 *) Extend ASN1 encoder to support indefinite length constructed
4107 encoding. This can output sequences tags and octet strings in
4108 this form. Modify pk7_asn1.c to support indefinite length
4109 encoding. This is experimental and needs additional code to
4110 be useful, such as an ASN1 bio and some enhanced streaming
4113 Extend template encode functionality so that tagging is passed
4114 down to the template encoder.
4117 *) Let 'openssl req' fail if an argument to '-newkey' is not
4118 recognized instead of using RSA as a default.
4121 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
4122 As these are not official, they are not included in "ALL";
4123 the "ECCdraft" ciphersuite group alias can be used to select them.
4124 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
4126 *) Add ECDH engine support.
4127 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
4129 *) Add ECDH in new directory crypto/ecdh/.
4130 [Douglas Stebila (Sun Microsystems Laboratories)]
4132 *) Let BN_rand_range() abort with an error after 100 iterations
4133 without success (which indicates a broken PRNG).
4136 *) Change BN_mod_sqrt() so that it verifies that the input value
4137 is really the square of the return value. (Previously,
4138 BN_mod_sqrt would show GIGO behaviour.)
4141 *) Add named elliptic curves over binary fields from X9.62, SECG,
4142 and WAP/WTLS; add OIDs that were still missing.
4144 [Sheueling Chang Shantz and Douglas Stebila
4145 (Sun Microsystems Laboratories)]
4147 *) Extend the EC library for elliptic curves over binary fields
4148 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4151 EC_GF2m_simple_method
4155 EC_GROUP_new_curve_GF2m
4156 EC_GROUP_set_curve_GF2m
4157 EC_GROUP_get_curve_GF2m
4158 EC_POINT_set_affine_coordinates_GF2m
4159 EC_POINT_get_affine_coordinates_GF2m
4160 EC_POINT_set_compressed_coordinates_GF2m
4162 Point compression for binary fields is disabled by default for
4163 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4166 As binary polynomials are represented as BIGNUMs, various members
4167 of the EC_GROUP and EC_POINT data structures can be shared
4168 between the implementations for prime fields and binary fields;
4169 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4170 are essentially identical to their ..._GFp counterparts.
4171 (For simplicity, the '..._GFp' prefix has been dropped from
4172 various internal method names.)
4174 An internal 'field_div' method (similar to 'field_mul' and
4175 'field_sqr') has been added; this is used only for binary fields.
4177 [Sheueling Chang Shantz and Douglas Stebila
4178 (Sun Microsystems Laboratories)]
4180 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4181 through methods ('mul', 'precompute_mult').
4183 The generic implementations (now internally called 'ec_wNAF_mul'
4184 and 'ec_wNAF_precomputed_mult') remain the default if these
4185 methods are undefined.
4187 [Sheueling Chang Shantz and Douglas Stebila
4188 (Sun Microsystems Laboratories)]
4190 *) New function EC_GROUP_get_degree, which is defined through
4191 EC_METHOD. For curves over prime fields, this returns the bit
4192 length of the modulus.
4194 [Sheueling Chang Shantz and Douglas Stebila
4195 (Sun Microsystems Laboratories)]
4197 *) New functions EC_GROUP_dup, EC_POINT_dup.
4198 (These simply call ..._new and ..._copy).
4200 [Sheueling Chang Shantz and Douglas Stebila
4201 (Sun Microsystems Laboratories)]
4203 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4204 Polynomials are represented as BIGNUMs (where the sign bit is not
4205 used) in the following functions [macros]:
4208 BN_GF2m_sub [= BN_GF2m_add]
4209 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4210 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4211 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4213 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4214 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4215 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4216 BN_GF2m_cmp [= BN_ucmp]
4218 (Note that only the 'mod' functions are actually for fields GF(2^m).
4219 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4221 For some functions, an the irreducible polynomial defining a
4222 field can be given as an 'unsigned int[]' with strictly
4223 decreasing elements giving the indices of those bits that are set;
4224 i.e., p[] represents the polynomial
4225 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4227 p[0] > p[1] > ... > p[k] = 0.
4228 This applies to the following functions:
4233 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4234 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4236 BN_GF2m_mod_sqrt_arr
4237 BN_GF2m_mod_solve_quad_arr
4241 Conversion can be performed by the following functions:
4246 bntest.c has additional tests for binary polynomial arithmetic.
4248 Two implementations for BN_GF2m_mod_div() are available.
4249 The default algorithm simply uses BN_GF2m_mod_inv() and
4250 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4251 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4252 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4254 [Sheueling Chang Shantz and Douglas Stebila
4255 (Sun Microsystems Laboratories)]
4257 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4258 functionality is disabled at compile-time.
4259 [Douglas Stebila <douglas.stebila@sun.com>]
4261 *) Change default behaviour of 'openssl asn1parse' so that more
4262 information is visible when viewing, e.g., a certificate:
4264 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4265 mode the content of non-printable OCTET STRINGs is output in a
4266 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4267 avoid the appearance of a printable string.
4268 [Nils Larsch <nla@trustcenter.de>]
4270 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4272 EC_GROUP_set_asn1_flag()
4273 EC_GROUP_get_asn1_flag()
4274 EC_GROUP_set_point_conversion_form()
4275 EC_GROUP_get_point_conversion_form()
4276 These control ASN1 encoding details:
4277 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4278 has been set to OPENSSL_EC_NAMED_CURVE.
4279 - Points are encoded in uncompressed form by default; options for
4280 asn1_for are as for point2oct, namely
4281 POINT_CONVERSION_COMPRESSED
4282 POINT_CONVERSION_UNCOMPRESSED
4283 POINT_CONVERSION_HYBRID
4285 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4288 EC_GROUP_get0_seed()
4289 EC_GROUP_get_seed_len()
4290 This is used only for ASN1 purposes (so far).
4291 [Nils Larsch <nla@trustcenter.de>]
4293 *) Add 'field_type' member to EC_METHOD, which holds the NID
4294 of the appropriate field type OID. The new function
4295 EC_METHOD_get_field_type() returns this value.
4296 [Nils Larsch <nla@trustcenter.de>]
4301 EC_POINT_point2hex()
4302 EC_POINT_hex2point()
4303 providing useful interfaces to EC_POINT_point2oct() and
4304 EC_POINT_oct2point().
4305 [Nils Larsch <nla@trustcenter.de>]
4307 *) Change internals of the EC library so that the functions
4308 EC_GROUP_set_generator()
4309 EC_GROUP_get_generator()
4310 EC_GROUP_get_order()
4311 EC_GROUP_get_cofactor()
4312 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4313 to methods, which would lead to unnecessary code duplication when
4314 adding different types of curves.
4315 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4317 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4318 arithmetic, and such that modified wNAFs are generated
4319 (which avoid length expansion in many cases).
4322 *) Add a function EC_GROUP_check_discriminant() (defined via
4323 EC_METHOD) that verifies that the curve discriminant is non-zero.
4325 Add a function EC_GROUP_check() that makes some sanity tests
4326 on a EC_GROUP, its generator and order. This includes
4327 EC_GROUP_check_discriminant().
4328 [Nils Larsch <nla@trustcenter.de>]
4330 *) Add ECDSA in new directory crypto/ecdsa/.
4332 Add applications 'openssl ecparam' and 'openssl ecdsa'
4333 (these are based on 'openssl dsaparam' and 'openssl dsa').
4335 ECDSA support is also included in various other files across the
4336 library. Most notably,
4337 - 'openssl req' now has a '-newkey ecdsa:file' option;
4338 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4339 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4340 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4341 them suitable for ECDSA where domain parameters must be
4342 extracted before the specific public key;
4343 - ECDSA engine support has been added.
4344 [Nils Larsch <nla@trustcenter.de>]
4346 *) Include some named elliptic curves, and add OIDs from X9.62,
4347 SECG, and WAP/WTLS. Each curve can be obtained from the new
4349 EC_GROUP_new_by_curve_name(),
4350 and the list of available named curves can be obtained with
4351 EC_get_builtin_curves().
4352 Also add a 'curve_name' member to EC_GROUP objects, which can be
4354 EC_GROUP_set_curve_name()
4355 EC_GROUP_get_curve_name()
4356 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4358 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4359 was actually never needed) and in BN_mul(). The removal in BN_mul()
4360 required a small change in bn_mul_part_recursive() and the addition
4361 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4362 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4363 bn_sub_words() and bn_add_words() except they take arrays with
4367 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4369 *) Cleanse PEM buffers before freeing them since they may contain
4371 [Benjamin Bennett <ben@psc.edu>]
4373 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4374 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4375 authentication-only ciphersuites.
4378 *) Since AES128 and AES256 share a single mask bit in the logic of
4379 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4380 kludge to work properly if AES128 is available and AES256 isn't.
4383 *) Expand security boundary to match 1.1.1 module.
4386 *) Remove redundant features: hash file source, editing of test vectors
4387 modify fipsld to use external fips_premain.c signature.
4390 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4391 run algorithm test programs.
4394 *) Make algorithm test programs more tolerant of whitespace.
4397 *) Have SSL/TLS server implementation tolerate "mismatched" record
4398 protocol version while receiving ClientHello even if the
4399 ClientHello is fragmented. (The server can't insist on the
4400 particular protocol version it has chosen before the ServerHello
4401 message has informed the client about his choice.)
4404 *) Load error codes if they are not already present instead of using a
4405 static variable. This allows them to be cleanly unloaded and reloaded.
4408 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4410 *) Introduce limits to prevent malicious keys being able to
4411 cause a denial of service. (CVE-2006-2940)
4412 [Steve Henson, Bodo Moeller]
4414 *) Fix ASN.1 parsing of certain invalid structures that can result
4415 in a denial of service. (CVE-2006-2937) [Steve Henson]
4417 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4418 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4420 *) Fix SSL client code which could crash if connecting to a
4421 malicious SSLv2 server. (CVE-2006-4343)
4422 [Tavis Ormandy and Will Drewry, Google Security Team]
4424 *) Change ciphersuite string processing so that an explicit
4425 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4426 will no longer include "AES128-SHA"), and any other similar
4427 ciphersuite (same bitmap) from *other* protocol versions (so that
4428 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4429 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4430 changes from 0.9.8b and 0.9.8d.
4433 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4435 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4436 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4438 *) Change the Unix randomness entropy gathering to use poll() when
4439 possible instead of select(), since the latter has some
4440 undesirable limitations.
4441 [Darryl Miles via Richard Levitte and Bodo Moeller]
4443 *) Disable rogue ciphersuites:
4445 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4446 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4447 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4449 The latter two were purportedly from
4450 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4453 Also deactive the remaining ciphersuites from
4454 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4455 unofficial, and the ID has long expired.
4458 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4459 dual-core machines) and other potential thread-safety issues.
4462 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4464 *) Adapt fipsld and the build system to link against the validated FIPS
4465 module in FIPS mode.
4468 *) Fixes for VC++ 2005 build under Windows.
4471 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4472 from a Windows bash shell such as MSYS. It is autodetected from the
4473 "config" script when run from a VC++ environment. Modify standard VC++
4474 build to use fipscanister.o from the GNU make build.
4477 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4479 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4480 The value now differs depending on if you build for FIPS or not.
4481 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4482 safely run with a non-FIPSed libcrypto, as it may crash because of
4483 the difference induced by this change.
4486 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4488 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4489 (part of SSL_OP_ALL). This option used to disable the
4490 countermeasure against man-in-the-middle protocol-version
4491 rollback in the SSL 2.0 server implementation, which is a bad
4492 idea. (CVE-2005-2969)
4494 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4495 for Information Security, National Institute of Advanced Industrial
4496 Science and Technology [AIST], Japan)]
4498 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4499 mainly for FIPS compliance and not fully integrated at this stage.
4502 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4503 the exponentiation using a fixed-length exponent. (Otherwise,
4504 the information leaked through timing could expose the secret key
4505 after many signatures; cf. Bleichenbacher's attack on DSA with
4509 *) Make a new fixed-window mod_exp implementation the default for
4510 RSA, DSA, and DH private-key operations so that the sequence of
4511 squares and multiplies and the memory access pattern are
4512 independent of the particular secret key. This will mitigate
4513 cache-timing and potential related attacks.
4515 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4516 and this is automatically used by BN_mod_exp_mont() if the new flag
4517 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4518 will use this BN flag for private exponents unless the flag
4519 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4520 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4522 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4524 *) Change the client implementation for SSLv23_method() and
4525 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4526 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4527 (Previously, the SSL 2.0 backwards compatible Client Hello
4528 message format would be used even with SSL_OP_NO_SSLv2.)
4531 *) Add support for smime-type MIME parameter in S/MIME messages which some
4535 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4536 a threadsafe manner. Modify rsa code to use new function and add calls
4537 to dsa and dh code (which had race conditions before).
4540 *) Include the fixed error library code in the C error file definitions
4541 instead of fixing them up at runtime. This keeps the error code
4542 structures constant.
4545 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4547 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4550 *) Fixes for newer kerberos headers. NB: the casts are needed because
4551 the 'length' field is signed on one version and unsigned on another
4552 with no (?) obvious way to tell the difference, without these VC++
4553 complains. Also the "definition" of FAR (blank) is no longer included
4554 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4555 some needed definitions.
4558 *) Undo Cygwin change.
4561 *) Added support for proxy certificates according to RFC 3820.
4562 Because they may be a security thread to unaware applications,
4563 they must be explicitely allowed in run-time. See
4564 docs/HOWTO/proxy_certificates.txt for further information.
4567 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4569 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4570 server and client random values. Previously
4571 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4572 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4574 This change has negligible security impact because:
4576 1. Server and client random values still have 24 bytes of pseudo random
4579 2. Server and client random values are sent in the clear in the initial
4582 3. The master secret is derived using the premaster secret (48 bytes in
4583 size for static RSA ciphersuites) as well as client server and random
4586 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4589 [Stephen Henson, reported by UK NISCC]
4591 *) Use Windows randomness collection on Cygwin.
4594 *) Fix hang in EGD/PRNGD query when communication socket is closed
4595 prematurely by EGD/PRNGD.
4596 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4598 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4601 *) Back-port of selected performance improvements from development
4602 branch, as well as improved support for PowerPC platforms.
4605 *) Add lots of checks for memory allocation failure, error codes to indicate
4606 failure and freeing up memory if a failure occurs.
4607 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4609 *) Add new -passin argument to dgst.
4612 *) Perform some character comparisons of different types in X509_NAME_cmp:
4613 this is needed for some certificates that reencode DNs into UTF8Strings
4614 (in violation of RFC3280) and can't or wont issue name rollover
4618 *) Make an explicit check during certificate validation to see that
4619 the CA setting in each certificate on the chain is correct. As a
4620 side effect always do the following basic checks on extensions,
4621 not just when there's an associated purpose to the check:
4623 - if there is an unhandled critical extension (unless the user
4624 has chosen to ignore this fault)
4625 - if the path length has been exceeded (if one is set at all)
4626 - that certain extensions fit the associated purpose (if one has
4630 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4632 *) Avoid a race condition when CRLs are checked in a multi threaded
4633 environment. This would happen due to the reordering of the revoked
4634 entries during signature checking and serial number lookup. Now the
4635 encoding is cached and the serial number sort performed under a lock.
4636 Add new STACK function sk_is_sorted().
4639 *) Add Delta CRL to the extension code.
4642 *) Various fixes to s3_pkt.c so alerts are sent properly.
4643 [David Holmes <d.holmes@f5.com>]
4645 *) Reduce the chances of duplicate issuer name and serial numbers (in
4646 violation of RFC3280) using the OpenSSL certificate creation utilities.
4647 This is done by creating a random 64 bit value for the initial serial
4648 number when a serial number file is created or when a self signed
4649 certificate is created using 'openssl req -x509'. The initial serial
4650 number file is created using 'openssl x509 -next_serial' in CA.pl
4651 rather than being initialized to 1.
4654 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4656 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4657 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4658 [Joe Orton, Steve Henson]
4660 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4662 [Joe Orton, Steve Henson]
4664 *) Make it possible to have multiple active certificates with the same
4665 subject in the CA index file. This is done only if the keyword
4666 'unique_subject' is set to 'no' in the main CA section (default
4667 if 'CA_default') of the configuration file. The value is saved
4668 with the database itself in a separate index attribute file,
4669 named like the index file with '.attr' appended to the name.
4672 *) X509 verify fixes. Disable broken certificate workarounds when
4673 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4674 keyUsage extension present. Don't accept CRLs with unhandled critical
4675 extensions: since verify currently doesn't process CRL extensions this
4676 rejects a CRL with *any* critical extensions. Add new verify error codes
4680 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4681 A clarification of RFC2560 will require the use of OCTET STRINGs and
4682 some implementations cannot handle the current raw format. Since OpenSSL
4683 copies and compares OCSP nonces as opaque blobs without any attempt at
4684 parsing them this should not create any compatibility issues.
4687 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4688 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4689 this HMAC (and other) operations are several times slower than OpenSSL
4693 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4694 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4696 *) Use the correct content when signing type "other".
4699 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4701 *) Fix various bugs revealed by running the NISCC test suite:
4703 Stop out of bounds reads in the ASN1 code when presented with
4704 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4706 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4708 If verify callback ignores invalid public key errors don't try to check
4709 certificate signature with the NULL public key.
4713 *) New -ignore_err option in ocsp application to stop the server
4714 exiting on the first error in a request.
4717 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4718 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4722 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4723 extra data after the compression methods not only for TLS 1.0
4724 but also for SSL 3.0 (as required by the specification).
4725 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4727 *) Change X509_certificate_type() to mark the key as exported/exportable
4728 when it's 512 *bits* long, not 512 bytes.
4731 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4732 blocks during encryption.
4735 *) Various fixes to base64 BIO and non blocking I/O. On write
4736 flushes were not handled properly if the BIO retried. On read
4737 data was not being buffered properly and had various logic bugs.
4738 This also affects blocking I/O when the data being decoded is a
4742 *) Various S/MIME bugfixes and compatibility changes:
4743 output correct application/pkcs7 MIME type if
4744 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4745 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4746 of files as .eml work). Correctly handle very long lines in MIME
4750 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4752 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4753 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4754 a protocol version number mismatch like a decryption error
4755 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4758 *) Turn on RSA blinding by default in the default implementation
4759 to avoid a timing attack. Applications that don't want it can call
4760 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4761 They would be ill-advised to do so in most cases.
4762 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4764 *) Change RSA blinding code so that it works when the PRNG is not
4765 seeded (in this case, the secret RSA exponent is abused as
4766 an unpredictable seed -- if it is not unpredictable, there
4767 is no point in blinding anyway). Make RSA blinding thread-safe
4768 by remembering the creator's thread ID in rsa->blinding and
4769 having all other threads use local one-time blinding factors
4770 (this requires more computation than sharing rsa->blinding, but
4771 avoids excessive locking; and if an RSA object is not shared
4772 between threads, blinding will still be very fast).
4775 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4776 ENGINE as defaults for all supported algorithms irrespective of
4777 the 'flags' parameter. 'flags' is now honoured, so applications
4778 should make sure they are passing it correctly.
4781 *) Target "mingw" now allows native Windows code to be generated in
4782 the Cygwin environment as well as with the MinGW compiler.
4785 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4787 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4788 via timing by performing a MAC computation even if incorrect
4789 block cipher padding has been found. This is a countermeasure
4790 against active attacks where the attacker has to distinguish
4791 between bad padding and a MAC verification error. (CVE-2003-0078)
4793 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4794 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4795 Martin Vuagnoux (EPFL, Ilion)]
4797 *) Make the no-err option work as intended. The intention with no-err
4798 is not to have the whole error stack handling routines removed from
4799 libcrypto, it's only intended to remove all the function name and
4800 reason texts, thereby removing some of the footprint that may not
4801 be interesting if those errors aren't displayed anyway.
4803 NOTE: it's still possible for any application or module to have it's
4804 own set of error texts inserted. The routines are there, just not
4805 used by default when no-err is given.
4808 *) Add support for FreeBSD on IA64.
4809 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4811 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4812 Kerberos function mit_des_cbc_cksum(). Before this change,
4813 the value returned by DES_cbc_cksum() was like the one from
4814 mit_des_cbc_cksum(), except the bytes were swapped.
4815 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4817 *) Allow an application to disable the automatic SSL chain building.
4818 Before this a rather primitive chain build was always performed in
4819 ssl3_output_cert_chain(): an application had no way to send the
4820 correct chain if the automatic operation produced an incorrect result.
4822 Now the chain builder is disabled if either:
4824 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4826 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4828 The reasoning behind this is that an application would not want the
4829 auto chain building to take place if extra chain certificates are
4830 present and it might also want a means of sending no additional
4831 certificates (for example the chain has two certificates and the
4835 *) Add the possibility to build without the ENGINE framework.
4836 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4838 *) Under Win32 gmtime() can return NULL: check return value in
4839 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4842 *) DSA routines: under certain error conditions uninitialized BN objects
4843 could be freed. Solution: make sure initialization is performed early
4844 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4845 Nils Larsch <nla@trustcenter.de> via PR#459)
4848 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4849 checked on reconnect on the client side, therefore session resumption
4850 could still fail with a "ssl session id is different" error. This
4851 behaviour is masked when SSL_OP_ALL is used due to
4852 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4853 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4854 followup to PR #377.
4857 *) IA-32 assembler support enhancements: unified ELF targets, support
4858 for SCO/Caldera platforms, fix for Cygwin shared build.
4861 *) Add support for FreeBSD on sparc64. As a consequence, support for
4862 FreeBSD on non-x86 processors is separate from x86 processors on
4863 the config script, much like the NetBSD support.
4864 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4866 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4868 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4871 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4872 code (06) was taken as the first octet of the session ID and the last
4873 octet was ignored consequently. As a result SSLv2 client side session
4874 caching could not have worked due to the session ID mismatch between
4876 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4880 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4881 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4885 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4886 seems that in spite of existing for more than a year, many application
4887 author have done nothing to provide the necessary callbacks, which
4888 means that this particular engine will not work properly anywhere.
4889 This is a very unfortunate situation which forces us, in the name
4890 of usability, to give the hw_ncipher.c a static lock, which is part
4892 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4893 appear in 0.9.8 or later. We EXPECT application authors to have
4894 dealt properly with this when 0.9.8 is released (unless we actually
4895 make such changes in the libcrypto locking code that changes will
4896 have to be made anyway).
4899 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4900 octets have been read, EOF or an error occurs. Without this change
4901 some truncated ASN1 structures will not produce an error.
4904 *) Disable Heimdal support, since it hasn't been fully implemented.
4905 Still give the possibility to force the use of Heimdal, but with
4906 warnings and a request that patches get sent to openssl-dev.
4909 *) Add the VC-CE target, introduce the WINCE sysname, and add
4910 INSTALL.WCE and appropriate conditionals to make it build.
4911 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4913 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4914 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4915 edit numbers of the version.
4916 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4918 *) Introduce safe string copy and catenation functions
4919 (BUF_strlcpy() and BUF_strlcat()).
4920 [Ben Laurie (CHATS) and Richard Levitte]
4922 *) Avoid using fixed-size buffers for one-line DNs.
4923 [Ben Laurie (CHATS)]
4925 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4926 resizing buffers containing secrets, and use where appropriate.
4927 [Ben Laurie (CHATS)]
4929 *) Avoid using fixed size buffers for configuration file location.
4930 [Ben Laurie (CHATS)]
4932 *) Avoid filename truncation for various CA files.
4933 [Ben Laurie (CHATS)]
4935 *) Use sizeof in preference to magic numbers.
4936 [Ben Laurie (CHATS)]
4938 *) Avoid filename truncation in cert requests.
4939 [Ben Laurie (CHATS)]
4941 *) Add assertions to check for (supposedly impossible) buffer
4943 [Ben Laurie (CHATS)]
4945 *) Don't cache truncated DNS entries in the local cache (this could
4946 potentially lead to a spoofing attack).
4947 [Ben Laurie (CHATS)]
4949 *) Fix various buffers to be large enough for hex/decimal
4950 representations in a platform independent manner.
4951 [Ben Laurie (CHATS)]
4953 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4954 resizing buffers containing secrets, and use where appropriate.
4955 [Ben Laurie (CHATS)]
4957 *) Add BIO_indent() to avoid much slightly worrying code to do
4959 [Ben Laurie (CHATS)]
4961 *) Convert sprintf()/BIO_puts() to BIO_printf().
4962 [Ben Laurie (CHATS)]
4964 *) buffer_gets() could terminate with the buffer only half
4966 [Ben Laurie (CHATS)]
4968 *) Add assertions to prevent user-supplied crypto functions from
4969 overflowing internal buffers by having large block sizes, etc.
4970 [Ben Laurie (CHATS)]
4972 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4974 [Ben Laurie (CHATS)]
4976 *) Eliminate unused copy of key in RC4.
4977 [Ben Laurie (CHATS)]
4979 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4980 [Ben Laurie (CHATS)]
4982 *) Fix off-by-one error in EGD path.
4983 [Ben Laurie (CHATS)]
4985 *) If RANDFILE path is too long, ignore instead of truncating.
4986 [Ben Laurie (CHATS)]
4988 *) Eliminate unused and incorrectly sized X.509 structure
4990 [Ben Laurie (CHATS)]
4992 *) Eliminate unused and dangerous function knumber().
4993 [Ben Laurie (CHATS)]
4995 *) Eliminate unused and dangerous structure, KSSL_ERR.
4996 [Ben Laurie (CHATS)]
4998 *) Protect against overlong session ID context length in an encoded
4999 session object. Since these are local, this does not appear to be
5001 [Ben Laurie (CHATS)]
5003 *) Change from security patch (see 0.9.6e below) that did not affect
5004 the 0.9.6 release series:
5006 Remote buffer overflow in SSL3 protocol - an attacker could
5007 supply an oversized master key in Kerberos-enabled versions.
5009 [Ben Laurie (CHATS)]
5011 *) Change the SSL kerb5 codes to match RFC 2712.
5014 *) Make -nameopt work fully for req and add -reqopt switch.
5015 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
5017 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
5018 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
5020 *) Make sure tests can be performed even if the corresponding algorithms
5021 have been removed entirely. This was also the last step to make
5022 OpenSSL compilable with DJGPP under all reasonable conditions.
5023 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
5025 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
5026 to allow version independent disabling of normally unselected ciphers,
5027 which may be activated as a side-effect of selecting a single cipher.
5029 (E.g., cipher list string "RSA" enables ciphersuites that are left
5030 out of "ALL" because they do not provide symmetric encryption.
5031 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
5032 [Lutz Jaenicke, Bodo Moeller]
5034 *) Add appropriate support for separate platform-dependent build
5035 directories. The recommended way to make a platform-dependent
5036 build directory is the following (tested on Linux), maybe with
5039 # Place yourself outside of the OpenSSL source tree. In
5040 # this example, the environment variable OPENSSL_SOURCE
5041 # is assumed to contain the absolute OpenSSL source directory.
5042 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
5043 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
5044 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
5045 mkdir -p `dirname $F`
5046 ln -s $OPENSSL_SOURCE/$F $F
5049 To be absolutely sure not to disturb the source tree, a "make clean"
5050 is a good thing. If it isn't successfull, don't worry about it,
5051 it probably means the source directory is very clean.
5054 *) Make sure any ENGINE control commands make local copies of string
5055 pointers passed to them whenever necessary. Otherwise it is possible
5056 the caller may have overwritten (or deallocated) the original string
5057 data when a later ENGINE operation tries to use the stored values.
5058 [Götz Babin-Ebell <babinebell@trustcenter.de>]
5060 *) Improve diagnostics in file reading and command-line digests.
5061 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
5063 *) Add AES modes CFB and OFB to the object database. Correct an
5064 error in AES-CFB decryption.
5067 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
5068 allows existing EVP_CIPHER_CTX structures to be reused after
5069 calling EVP_*Final(). This behaviour is used by encryption
5070 BIOs and some applications. This has the side effect that
5071 applications must explicitly clean up cipher contexts with
5072 EVP_CIPHER_CTX_cleanup() or they will leak memory.
5075 *) Check the values of dna and dnb in bn_mul_recursive before calling
5076 bn_mul_comba (a non zero value means the a or b arrays do not contain
5077 n2 elements) and fallback to bn_mul_normal if either is not zero.
5080 *) Fix escaping of non-ASCII characters when using the -subj option
5081 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
5084 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
5085 form for "surname", serialNumber has no short form.
5086 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
5087 therefore remove "mail" short name for "internet 7".
5088 The OID for unique identifiers in X509 certificates is
5089 x500UniqueIdentifier, not uniqueIdentifier.
5090 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
5093 *) Add an "init" command to the ENGINE config module and auto initialize
5094 ENGINEs. Without any "init" command the ENGINE will be initialized
5095 after all ctrl commands have been executed on it. If init=1 the
5096 ENGINE is initailized at that point (ctrls before that point are run
5097 on the uninitialized ENGINE and after on the initialized one). If
5098 init=0 then the ENGINE will not be iniatialized at all.
5101 *) Fix the 'app_verify_callback' interface so that the user-defined
5102 argument is actually passed to the callback: In the
5103 SSL_CTX_set_cert_verify_callback() prototype, the callback
5104 declaration has been changed from
5107 int (*cb)(X509_STORE_CTX *,void *);
5108 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
5109 i=s->ctx->app_verify_callback(&ctx)
5110 has been changed into
5111 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
5113 To update applications using SSL_CTX_set_cert_verify_callback(),
5114 a dummy argument can be added to their callback functions.
5115 [D. K. Smetters <smetters@parc.xerox.com>]
5117 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
5118 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
5120 *) Add and OPENSSL_LOAD_CONF define which will cause
5121 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
5122 This allows older applications to transparently support certain
5123 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
5124 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
5125 load the config file and OPENSSL_add_all_algorithms_conf() which will
5126 always load it have also been added.
5129 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
5130 Adjust NIDs and EVP layer.
5131 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5133 *) Config modules support in openssl utility.
5135 Most commands now load modules from the config file,
5136 though in a few (such as version) this isn't done
5137 because it couldn't be used for anything.
5139 In the case of ca and req the config file used is
5140 the same as the utility itself: that is the -config
5141 command line option can be used to specify an
5145 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5146 use "openssl_conf" if filename is NULL use default openssl config file.
5149 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5150 config section name. Add a new flag to tolerate a missing config file
5151 and move code to CONF_modules_load_file().
5154 *) Support for crypto accelerator cards from Accelerated Encryption
5155 Processing, www.aep.ie. (Use engine 'aep')
5156 The support was copied from 0.9.6c [engine] and adapted/corrected
5157 to work with the new engine framework.
5158 [AEP Inc. and Richard Levitte]
5160 *) Support for SureWare crypto accelerator cards from Baltimore
5161 Technologies. (Use engine 'sureware')
5162 The support was copied from 0.9.6c [engine] and adapted
5163 to work with the new engine framework.
5166 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5167 make the newer ENGINE framework commands for the CHIL engine work.
5168 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5170 *) Make it possible to produce shared libraries on ReliantUNIX.
5171 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5173 *) Add the configuration target debug-linux-ppro.
5174 Make 'openssl rsa' use the general key loading routines
5175 implemented in apps.c, and make those routines able to
5176 handle the key format FORMAT_NETSCAPE and the variant
5178 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5180 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5181 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5183 *) Add -keyform to rsautl, and document -engine.
5184 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5186 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5187 BIO_R_NO_SUCH_FILE error code rather than the generic
5188 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5191 *) Add new functions
5193 ERR_peek_last_error_line
5194 ERR_peek_last_error_line_data.
5195 These are similar to
5198 ERR_peek_error_line_data,
5199 but report on the latest error recorded rather than the first one
5200 still in the error queue.
5201 [Ben Laurie, Bodo Moeller]
5203 *) default_algorithms option in ENGINE config module. This allows things
5205 default_algorithms = ALL
5206 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5209 *) Preliminary ENGINE config module.
5212 *) New experimental application configuration code.
5215 *) Change the AES code to follow the same name structure as all other
5216 symmetric ciphers, and behave the same way. Move everything to
5217 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5218 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5220 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5221 [Ben Laurie and Theo de Raadt]
5223 *) Add option to output public keys in req command.
5224 [Massimiliano Pala madwolf@openca.org]
5226 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5227 (up to about 10% better than before for P-192 and P-224).
5230 *) New functions/macros
5232 SSL_CTX_set_msg_callback(ctx, cb)
5233 SSL_CTX_set_msg_callback_arg(ctx, arg)
5234 SSL_set_msg_callback(ssl, cb)
5235 SSL_set_msg_callback_arg(ssl, arg)
5237 to request calling a callback function
5239 void cb(int write_p, int version, int content_type,
5240 const void *buf, size_t len, SSL *ssl, void *arg)
5242 whenever a protocol message has been completely received
5243 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5244 protocol version according to which the SSL library interprets
5245 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5246 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5247 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5248 specification (change_cipher_spec(20), alert(21), handshake(22)).
5249 'buf' and 'len' point to the actual message, 'ssl' to the
5250 SSL object, and 'arg' is the application-defined value set by
5251 SSL[_CTX]_set_msg_callback_arg().
5253 'openssl s_client' and 'openssl s_server' have new '-msg' options
5254 to enable a callback that displays all protocol messages.
5257 *) Change the shared library support so shared libraries are built as
5258 soon as the corresponding static library is finished, and thereby get
5259 openssl and the test programs linked against the shared library.
5260 This still only happens when the keyword "shard" has been given to
5261 the configuration scripts.
5263 NOTE: shared library support is still an experimental thing, and
5264 backward binary compatibility is still not guaranteed.
5265 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5267 *) Add support for Subject Information Access extension.
5268 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5270 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5271 additional bytes when new memory had to be allocated, not just
5272 when reusing an existing buffer.
5275 *) New command line and configuration option 'utf8' for the req command.
5276 This allows field values to be specified as UTF8 strings.
5279 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5280 runs for the former and machine-readable output for the latter.
5283 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5284 of the e-mail address in the DN (i.e., it will go into a certificate
5285 extension only). The new configuration file option 'email_in_dn = no'
5286 has the same effect.
5287 [Massimiliano Pala madwolf@openca.org]
5289 *) Change all functions with names starting with des_ to be starting
5290 with DES_ instead. Add wrappers that are compatible with libdes,
5291 but are named _ossl_old_des_*. Finally, add macros that map the
5292 des_* symbols to the corresponding _ossl_old_des_* if libdes
5293 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5294 desired, the des_* symbols will be mapped to DES_*, with one
5297 Since we provide two compatibility mappings, the user needs to
5298 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5299 compatibility is desired. The default (i.e., when that macro
5300 isn't defined) is OpenSSL 0.9.6c compatibility.
5302 There are also macros that enable and disable the support of old
5303 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5304 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5305 are defined, the default will apply: to support the old des routines.
5307 In either case, one must include openssl/des.h to get the correct
5308 definitions. Do not try to just include openssl/des_old.h, that
5311 NOTE: This is a major break of an old API into a new one. Software
5312 authors are encouraged to switch to the DES_ style functions. Some
5313 time in the future, des_old.h and the libdes compatibility functions
5314 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5315 default), and then completely removed.
5318 *) Test for certificates which contain unsupported critical extensions.
5319 If such a certificate is found during a verify operation it is
5320 rejected by default: this behaviour can be overridden by either
5321 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5322 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5323 X509_supported_extension() has also been added which returns 1 if a
5324 particular extension is supported.
5327 *) Modify the behaviour of EVP cipher functions in similar way to digests
5328 to retain compatibility with existing code.
5331 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5332 compatibility with existing code. In particular the 'ctx' parameter does
5333 not have to be to be initialized before the call to EVP_DigestInit() and
5334 it is tidied up after a call to EVP_DigestFinal(). New function
5335 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5336 EVP_MD_CTX_copy() changed to not require the destination to be
5337 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5338 requires the destination to be valid.
5340 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5341 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5344 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5345 so that complete 'Handshake' protocol structures are kept in memory
5346 instead of overwriting 'msg_type' and 'length' with 'body' data.
5349 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5350 [Massimo Santin via Richard Levitte]
5352 *) Major restructuring to the underlying ENGINE code. This includes
5353 reduction of linker bloat, separation of pure "ENGINE" manipulation
5354 (initialisation, etc) from functionality dealing with implementations
5355 of specific crypto iterfaces. This change also introduces integrated
5356 support for symmetric ciphers and digest implementations - so ENGINEs
5357 can now accelerate these by providing EVP_CIPHER and EVP_MD
5358 implementations of their own. This is detailed in crypto/engine/README
5359 as it couldn't be adequately described here. However, there are a few
5360 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5361 were changed in the original introduction of ENGINE code have now
5362 reverted back - the hooking from this code to ENGINE is now a good
5363 deal more passive and at run-time, operations deal directly with
5364 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5365 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5366 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5367 they were not being used by the framework as there is no concept of a
5368 BIGNUM_METHOD and they could not be generalised to the new
5369 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5370 ENGINE_cpy() has been removed as it cannot be consistently defined in
5374 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5377 *) Change mkdef.pl to sort symbols that get the same entry number,
5378 and make sure the automatically generated functions ERR_load_*
5379 become part of libeay.num as well.
5382 *) New function SSL_renegotiate_pending(). This returns true once
5383 renegotiation has been requested (either SSL_renegotiate() call
5384 or HelloRequest/ClientHello received from the peer) and becomes
5385 false once a handshake has been completed.
5386 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5387 sends a HelloRequest, but does not ensure that a handshake takes
5388 place. SSL_renegotiate_pending() is useful for checking if the
5389 client has followed the request.)
5392 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5393 By default, clients may request session resumption even during
5394 renegotiation (if session ID contexts permit); with this option,
5395 session resumption is possible only in the first handshake.
5397 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5398 more bits available for options that should not be part of
5399 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5402 *) Add some demos for certificate and certificate request creation.
5405 *) Make maximum certificate chain size accepted from the peer application
5406 settable (SSL*_get/set_max_cert_list()), as proposed by
5407 "Douglas E. Engert" <deengert@anl.gov>.
5410 *) Add support for shared libraries for Unixware-7
5411 (Boyd Lynn Gerber <gerberb@zenez.com>).
5414 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5415 be done prior to destruction. Use this to unload error strings from
5416 ENGINEs that load their own error strings. NB: This adds two new API
5417 functions to "get" and "set" this destroy handler in an ENGINE.
5420 *) Alter all existing ENGINE implementations (except "openssl" and
5421 "openbsd") to dynamically instantiate their own error strings. This
5422 makes them more flexible to be built both as statically-linked ENGINEs
5423 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5424 Also, add stub code to each that makes building them as self-contained
5425 shared-libraries easier (see README.ENGINE).
5428 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5429 implementations into applications that are completely implemented in
5430 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5431 commands that can be used to configure what shared-library to load and
5432 to control aspects of the way it is handled. Also, made an update to
5433 the README.ENGINE file that brings its information up-to-date and
5434 provides some information and instructions on the "dynamic" ENGINE
5435 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5438 *) Make it possible to unload ranges of ERR strings with a new
5439 "ERR_unload_strings" function.
5442 *) Add a copy() function to EVP_MD.
5445 *) Make EVP_MD routines take a context pointer instead of just the
5446 md_data void pointer.
5449 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5450 that the digest can only process a single chunk of data
5451 (typically because it is provided by a piece of
5452 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5453 is only going to provide a single chunk of data, and hence the
5454 framework needn't accumulate the data for oneshot drivers.
5457 *) As with "ERR", make it possible to replace the underlying "ex_data"
5458 functions. This change also alters the storage and management of global
5459 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5460 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5461 index counters. The API functions that use this state have been changed
5462 to take a "class_index" rather than pointers to the class's local STACK
5463 and counter, and there is now an API function to dynamically create new
5464 classes. This centralisation allows us to (a) plug a lot of the
5465 thread-safety problems that existed, and (b) makes it possible to clean
5466 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5467 such data would previously have always leaked in application code and
5468 workarounds were in place to make the memory debugging turn a blind eye
5469 to it. Application code that doesn't use this new function will still
5470 leak as before, but their memory debugging output will announce it now
5471 rather than letting it slide.
5473 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5474 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5475 has a return value to indicate success or failure.
5478 *) Make it possible to replace the underlying "ERR" functions such that the
5479 global state (2 LHASH tables and 2 locks) is only used by the "default"
5480 implementation. This change also adds two functions to "get" and "set"
5481 the implementation prior to it being automatically set the first time
5482 any other ERR function takes place. Ie. an application can call "get",
5483 pass the return value to a module it has just loaded, and that module
5484 can call its own "set" function using that value. This means the
5485 module's "ERR" operations will use (and modify) the error state in the
5486 application and not in its own statically linked copy of OpenSSL code.
5489 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5490 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5491 the operation, and provides a more encapsulated way for external code
5492 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5493 to use these functions rather than manually incrementing the counts.
5495 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5498 *) Add EVP test program.
5501 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5504 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5505 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5506 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5507 These allow a CRL to be built without having to access X509_CRL fields
5508 directly. Modify 'ca' application to use new functions.
5511 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5512 bug workarounds. Rollback attack detection is a security feature.
5513 The problem will only arise on OpenSSL servers when TLSv1 is not
5514 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5515 Software authors not wanting to support TLSv1 will have special reasons
5516 for their choice and can explicitly enable this option.
5517 [Bodo Moeller, Lutz Jaenicke]
5519 *) Rationalise EVP so it can be extended: don't include a union of
5520 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5521 (similar to those existing for EVP_CIPHER_CTX).
5526 EVP_MD_CTX_init(&md); /* new function call */
5527 EVP_DigestInit(&md, EVP_sha1());
5528 EVP_DigestUpdate(&md, in, len);
5529 EVP_DigestFinal(&md, out, NULL);
5530 EVP_MD_CTX_cleanup(&md); /* new function call */
5534 *) Make DES key schedule conform to the usual scheme, as well as
5535 correcting its structure. This means that calls to DES functions
5536 now have to pass a pointer to a des_key_schedule instead of a
5537 plain des_key_schedule (which was actually always a pointer
5540 des_key_schedule ks;
5542 des_set_key_checked(..., &ks);
5543 des_ncbc_encrypt(..., &ks, ...);
5545 (Note that a later change renames 'des_...' into 'DES_...'.)
5548 *) Initial reduction of linker bloat: the use of some functions, such as
5549 PEM causes large amounts of unused functions to be linked in due to
5550 poor organisation. For example pem_all.c contains every PEM function
5551 which has a knock on effect of linking in large amounts of (unused)
5552 ASN1 code. Grouping together similar functions and splitting unrelated
5553 functions prevents this.
5556 *) Cleanup of EVP macros.
5559 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5560 correct _ecb suffix.
5563 *) Add initial OCSP responder support to ocsp application. The
5564 revocation information is handled using the text based index
5565 use by the ca application. The responder can either handle
5566 requests generated internally, supplied in files (for example
5567 via a CGI script) or using an internal minimal server.
5570 *) Add configuration choices to get zlib compression for TLS.
5573 *) Changes to Kerberos SSL for RFC 2712 compliance:
5574 1. Implemented real KerberosWrapper, instead of just using
5575 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5576 2. Implemented optional authenticator field of KerberosWrapper.
5578 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5579 and authenticator structs; see crypto/krb5/.
5581 Generalized Kerberos calls to support multiple Kerberos libraries.
5582 [Vern Staats <staatsvr@asc.hpc.mil>,
5583 Jeffrey Altman <jaltman@columbia.edu>
5584 via Richard Levitte]
5586 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5587 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5588 values for each of the key sizes rather than having just
5589 parameters (and 'speed' generating keys each time).
5592 *) Speed up EVP routines.
5595 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5596 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5597 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5598 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5600 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5601 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5602 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5605 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5607 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5610 *) Added the OS2-EMX target.
5611 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5613 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5614 to support NCONF routines in extension code. New function CONF_set_nconf()
5615 to allow functions which take an NCONF to also handle the old LHASH
5616 structure: this means that the old CONF compatible routines can be
5617 retained (in particular wrt extensions) without having to duplicate the
5618 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5621 *) Enhance the general user interface with mechanisms for inner control
5622 and with possibilities to have yes/no kind of prompts.
5625 *) Change all calls to low level digest routines in the library and
5626 applications to use EVP. Add missing calls to HMAC_cleanup() and
5627 don't assume HMAC_CTX can be copied using memcpy().
5628 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5630 *) Add the possibility to control engines through control names but with
5631 arbitrary arguments instead of just a string.
5632 Change the key loaders to take a UI_METHOD instead of a callback
5633 function pointer. NOTE: this breaks binary compatibility with earlier
5634 versions of OpenSSL [engine].
5635 Adapt the nCipher code for these new conditions and add a card insertion
5639 *) Enhance the general user interface with mechanisms to better support
5640 dialog box interfaces, application-defined prompts, the possibility
5641 to use defaults (for example default passwords from somewhere else)
5642 and interrupts/cancellations.
5645 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5646 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5649 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5650 tidy up some unnecessarily weird code in 'sk_new()').
5651 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5653 *) Change the key loading routines for ENGINEs to use the same kind
5654 callback (pem_password_cb) as all other routines that need this
5658 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5659 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5660 than this minimum value is recommended.
5663 *) New random seeder for OpenVMS, using the system process statistics
5664 that are easily reachable.
5667 *) Windows apparently can't transparently handle global
5668 variables defined in DLLs. Initialisations such as:
5670 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5672 wont compile. This is used by the any applications that need to
5673 declare their own ASN1 modules. This was fixed by adding the option
5674 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5675 needed for static libraries under Win32.
5678 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5679 setting of purpose and trust fields. New X509_STORE trust and
5680 purpose functions and tidy up setting in other SSL functions.
5683 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5684 structure. These are inherited by X509_STORE_CTX when it is
5685 initialised. This allows various defaults to be set in the
5686 X509_STORE structure (such as flags for CRL checking and custom
5687 purpose or trust settings) for functions which only use X509_STORE_CTX
5688 internally such as S/MIME.
5690 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5691 trust settings if they are not set in X509_STORE. This allows X509_STORE
5692 purposes and trust (in S/MIME for example) to override any set by default.
5694 Add command line options for CRL checking to smime, s_client and s_server
5698 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5699 are set then the CRL is looked up in the X509_STORE structure and
5700 its validity and signature checked, then if the certificate is found
5701 in the CRL the verify fails with a revoked error.
5703 Various new CRL related callbacks added to X509_STORE_CTX structure.
5705 Command line options added to 'verify' application to support this.
5707 This needs some additional work, such as being able to handle multiple
5708 CRLs with different times, extension based lookup (rather than just
5709 by subject name) and ultimately more complete V2 CRL extension
5713 *) Add a general user interface API (crypto/ui/). This is designed
5714 to replace things like des_read_password and friends (backward
5715 compatibility functions using this new API are provided).
5716 The purpose is to remove prompting functions from the DES code
5717 section as well as provide for prompting through dialog boxes in
5718 a window system and the like.
5721 *) Add "ex_data" support to ENGINE so implementations can add state at a
5722 per-structure level rather than having to store it globally.
5725 *) Make it possible for ENGINE structures to be copied when retrieved by
5726 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5727 This causes the "original" ENGINE structure to act like a template,
5728 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5729 operational state can be localised to each ENGINE structure, despite the
5730 fact they all share the same "methods". New ENGINE structures returned in
5731 this case have no functional references and the return value is the single
5732 structural reference. This matches the single structural reference returned
5733 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5737 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5738 needs to match any other type at all we need to manually clear the
5742 *) Changes to the "openssl engine" utility to include;
5743 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5744 about an ENGINE's available control commands.
5745 - executing control commands from command line arguments using the
5746 '-pre' and '-post' switches. '-post' is only used if '-t' is
5747 specified and the ENGINE is successfully initialised. The syntax for
5748 the individual commands are colon-separated, for example;
5749 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5752 *) New dynamic control command support for ENGINEs. ENGINEs can now
5753 declare their own commands (numbers), names (strings), descriptions,
5754 and input types for run-time discovery by calling applications. A
5755 subset of these commands are implicitly classed as "executable"
5756 depending on their input type, and only these can be invoked through
5757 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5758 can be based on user input, config files, etc). The distinction is
5759 that "executable" commands cannot return anything other than a boolean
5760 result and can only support numeric or string input, whereas some
5761 discoverable commands may only be for direct use through
5762 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5763 pointers, or other custom uses. The "executable" commands are to
5764 support parameterisations of ENGINE behaviour that can be
5765 unambiguously defined by ENGINEs and used consistently across any
5766 OpenSSL-based application. Commands have been added to all the
5767 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5768 control over shared-library paths without source code alterations.
5771 *) Changed all ENGINE implementations to dynamically allocate their
5772 ENGINEs rather than declaring them statically. Apart from this being
5773 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5774 this also allows the implementations to compile without using the
5775 internal engine_int.h header.
5778 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5779 'const' value. Any code that should be able to modify a RAND_METHOD
5780 should already have non-const pointers to it (ie. they should only
5781 modify their own ones).
5784 *) Made a variety of little tweaks to the ENGINE code.
5785 - "atalla" and "ubsec" string definitions were moved from header files
5786 to C code. "nuron" string definitions were placed in variables
5787 rather than hard-coded - allowing parameterisation of these values
5788 later on via ctrl() commands.
5789 - Removed unused "#if 0"'d code.
5790 - Fixed engine list iteration code so it uses ENGINE_free() to release
5791 structural references.
5792 - Constified the RAND_METHOD element of ENGINE structures.
5793 - Constified various get/set functions as appropriate and added
5794 missing functions (including a catch-all ENGINE_cpy that duplicates
5795 all ENGINE values onto a new ENGINE except reference counts/state).
5796 - Removed NULL parameter checks in get/set functions. Setting a method
5797 or function to NULL is a way of cancelling out a previously set
5798 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5799 and doesn't justify the extra error symbols and code.
5800 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5801 flags from engine_int.h to engine.h.
5802 - Changed prototypes for ENGINE handler functions (init(), finish(),
5803 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5806 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5807 to the algorithm using long division. The binary algorithm can be
5808 used only if the modulus is odd. On 32-bit systems, it is faster
5809 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5810 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5811 up to 450 bits. In 64-bit environments, the binary algorithm
5812 appears to be advantageous for much longer moduli; here we use it
5813 for moduli up to 2048 bits.
5816 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5817 could not support the combine flag in choice fields.
5820 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5821 extensions from a certificate request to the certificate.
5824 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5825 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5826 file: this allows the display of the certificate about to be
5827 signed to be customised, to allow certain fields to be included
5828 or excluded and extension details. The old system didn't display
5829 multicharacter strings properly, omitted fields not in the policy
5830 and couldn't display additional details such as extensions.
5833 *) Function EC_POINTs_mul for multiple scalar multiplication
5834 of an arbitrary number of elliptic curve points
5835 \sum scalars[i]*points[i],
5836 optionally including the generator defined for the EC_GROUP:
5837 scalar*generator + \sum scalars[i]*points[i].
5839 EC_POINT_mul is a simple wrapper function for the typical case
5840 that the point list has just one item (besides the optional
5844 *) First EC_METHODs for curves over GF(p):
5846 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5847 operations and provides various method functions that can also
5848 operate with faster implementations of modular arithmetic.
5850 EC_GFp_mont_method() reuses most functions that are part of
5851 EC_GFp_simple_method, but uses Montgomery arithmetic.
5853 [Bodo Moeller; point addition and point doubling
5854 implementation directly derived from source code provided by
5855 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5857 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5858 crypto/ec/ec_lib.c):
5860 Curves are EC_GROUP objects (with an optional group generator)
5861 based on EC_METHODs that are built into the library.
5863 Points are EC_POINT objects based on EC_GROUP objects.
5865 Most of the framework would be able to handle curves over arbitrary
5866 finite fields, but as there are no obvious types for fields other
5867 than GF(p), some functions are limited to that for now.
5870 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5871 that the file contains a complete HTTP response.
5874 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5875 change the def and num file printf format specifier from "%-40sXXX"
5876 to "%-39s XXX". The latter will always guarantee a space after the
5877 field while the former will cause them to run together if the field
5878 is 40 of more characters long.
5881 *) Constify the cipher and digest 'method' functions and structures
5882 and modify related functions to take constant EVP_MD and EVP_CIPHER
5886 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5887 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5890 *) Modify EVP_Digest*() routines so they now return values. Although the
5891 internal software routines can never fail additional hardware versions
5895 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5897 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5898 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5901 ERR_R_NESTED_ASN1_ERROR
5903 ERR_R_MISSING_ASN1_EOS
5904 were 4 .. 9, conflicting with
5905 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5907 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5908 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5910 Add new error code 'ERR_R_INTERNAL_ERROR'.
5913 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5917 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5918 sets the subject name for a new request or supersedes the
5919 subject name in a given request. Formats that can be parsed are
5920 'CN=Some Name, OU=myOU, C=IT'
5922 'CN=Some Name/OU=myOU/C=IT'.
5924 Add options '-batch' and '-verbose' to 'openssl req'.
5925 [Massimiliano Pala <madwolf@hackmasters.net>]
5927 *) Introduce the possibility to access global variables through
5928 functions on platform were that's the best way to handle exporting
5929 global variables in shared libraries. To enable this functionality,
5930 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5931 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5932 is normally done by Configure or something similar).
5934 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5935 in the source file (foo.c) like this:
5937 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5938 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5940 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5941 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5943 OPENSSL_DECLARE_GLOBAL(int,foo);
5944 #define foo OPENSSL_GLOBAL_REF(foo)
5945 OPENSSL_DECLARE_GLOBAL(double,bar);
5946 #define bar OPENSSL_GLOBAL_REF(bar)
5948 The #defines are very important, and therefore so is including the
5949 header file everywhere where the defined globals are used.
5951 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5952 of ASN.1 items, but that structure is a bit different.
5954 The largest change is in util/mkdef.pl which has been enhanced with
5955 better and easier to understand logic to choose which symbols should
5956 go into the Windows .def files as well as a number of fixes and code
5957 cleanup (among others, algorithm keywords are now sorted
5958 lexicographically to avoid constant rewrites).
5961 *) In BN_div() keep a copy of the sign of 'num' before writing the
5962 result to 'rm' because if rm==num the value will be overwritten
5963 and produce the wrong result if 'num' is negative: this caused
5964 problems with BN_mod() and BN_nnmod().
5967 *) Function OCSP_request_verify(). This checks the signature on an
5968 OCSP request and verifies the signer certificate. The signer
5969 certificate is just checked for a generic purpose and OCSP request
5973 *) Add OCSP_check_validity() function to check the validity of OCSP
5974 responses. OCSP responses are prepared in real time and may only
5975 be a few seconds old. Simply checking that the current time lies
5976 between thisUpdate and nextUpdate max reject otherwise valid responses
5977 caused by either OCSP responder or client clock inaccuracy. Instead
5978 we allow thisUpdate and nextUpdate to fall within a certain period of
5979 the current time. The age of the response can also optionally be
5980 checked. Two new options -validity_period and -status_age added to
5984 *) If signature or public key algorithm is unrecognized print out its
5985 OID rather that just UNKNOWN.
5988 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5989 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5990 ID to be generated from the issuer certificate alone which can then be
5991 passed to OCSP_id_issuer_cmp().
5994 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5995 ASN1 modules to export functions returning ASN1_ITEM pointers
5996 instead of the ASN1_ITEM structures themselves. This adds several
5997 new macros which allow the underlying ASN1 function/structure to
5998 be accessed transparently. As a result code should not use ASN1_ITEM
5999 references directly (such as &X509_it) but instead use the relevant
6000 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
6001 use of the new ASN1 code on platforms where exporting structures
6002 is problematical (for example in shared libraries) but exporting
6003 functions returning pointers to structures is not.
6006 *) Add support for overriding the generation of SSL/TLS session IDs.
6007 These callbacks can be registered either in an SSL_CTX or per SSL.
6008 The purpose of this is to allow applications to control, if they wish,
6009 the arbitrary values chosen for use as session IDs, particularly as it
6010 can be useful for session caching in multiple-server environments. A
6011 command-line switch for testing this (and any client code that wishes
6012 to use such a feature) has been added to "s_server".
6013 [Geoff Thorpe, Lutz Jaenicke]
6015 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
6016 of the form '#if defined(...) || defined(...) || ...' and
6017 '#if !defined(...) && !defined(...) && ...'. This also avoids
6018 the growing number of special cases it was previously handling.
6021 *) Make all configuration macros available for application by making
6022 sure they are available in opensslconf.h, by giving them names starting
6023 with "OPENSSL_" to avoid conflicts with other packages and by making
6024 sure e_os2.h will cover all platform-specific cases together with
6026 Additionally, it is now possible to define configuration/platform-
6027 specific names (called "system identities"). In the C code, these
6028 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
6029 macro with the name beginning with "OPENSSL_SYS_", which is determined
6030 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
6034 *) New option -set_serial to 'req' and 'x509' this allows the serial
6035 number to use to be specified on the command line. Previously self
6036 signed certificates were hard coded with serial number 0 and the
6037 CA options of 'x509' had to use a serial number in a file which was
6041 *) New options to 'ca' utility to support V2 CRL entry extensions.
6042 Currently CRL reason, invalidity date and hold instruction are
6043 supported. Add new CRL extensions to V3 code and some new objects.
6046 *) New function EVP_CIPHER_CTX_set_padding() this is used to
6047 disable standard block padding (aka PKCS#5 padding) in the EVP
6048 API, which was previously mandatory. This means that the data is
6049 not padded in any way and so the total length much be a multiple
6050 of the block size, otherwise an error occurs.
6053 *) Initial (incomplete) OCSP SSL support.
6056 *) New function OCSP_parse_url(). This splits up a URL into its host,
6057 port and path components: primarily to parse OCSP URLs. New -url
6058 option to ocsp utility.
6061 *) New nonce behavior. The return value of OCSP_check_nonce() now
6062 reflects the various checks performed. Applications can decide
6063 whether to tolerate certain situations such as an absent nonce
6064 in a response when one was present in a request: the ocsp application
6065 just prints out a warning. New function OCSP_add1_basic_nonce()
6066 this is to allow responders to include a nonce in a response even if
6067 the request is nonce-less.
6070 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
6071 skipped when using openssl x509 multiple times on a single input file,
6072 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
6075 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
6076 set string type: to handle setting ASN1_TIME structures. Fix ca
6077 utility to correctly initialize revocation date of CRLs.
6080 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
6081 the clients preferred ciphersuites and rather use its own preferences.
6082 Should help to work around M$ SGC (Server Gated Cryptography) bug in
6083 Internet Explorer by ensuring unchanged hash method during stepup.
6084 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
6087 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
6088 to aes and add a new 'exist' option to print out symbols that don't
6092 *) Additional options to ocsp utility to allow flags to be set and
6093 additional certificates supplied.
6096 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
6097 OCSP client a number of certificate to only verify the response
6101 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
6102 handle the new API. Currently only ECB, CBC modes supported. Add new
6105 Add TLS AES ciphersuites as described in RFC3268, "Advanced
6106 Encryption Standard (AES) Ciphersuites for Transport Layer
6107 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
6108 not enabled by default and were not part of the "ALL" ciphersuite
6109 alias because they were not yet official; they could be
6110 explicitly requested by specifying the "AESdraft" ciphersuite
6111 group alias. In the final release of OpenSSL 0.9.7, the group
6112 alias is called "AES" and is part of "ALL".)
6113 [Ben Laurie, Steve Henson, Bodo Moeller]
6115 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
6116 request to response.
6119 *) Functions for OCSP responders. OCSP_request_onereq_count(),
6120 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
6121 extract information from a certificate request. OCSP_response_create()
6122 creates a response and optionally adds a basic response structure.
6123 OCSP_basic_add1_status() adds a complete single response to a basic
6124 response and returns the OCSP_SINGLERESP structure just added (to allow
6125 extensions to be included for example). OCSP_basic_add1_cert() adds a
6126 certificate to a basic response and OCSP_basic_sign() signs a basic
6127 response with various flags. New helper functions ASN1_TIME_check()
6128 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
6129 (converts ASN1_TIME to GeneralizedTime).
6132 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6133 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6134 structure from a certificate. X509_pubkey_digest() digests the public_key
6135 contents: this is used in various key identifiers.
6138 *) Make sk_sort() tolerate a NULL argument.
6139 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6141 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6142 passed by the function are trusted implicitly. If any of them signed the
6143 response then it is assumed to be valid and is not verified.
6146 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6147 to data. This was previously part of the PKCS7 ASN1 code. This
6148 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6149 [Steve Henson, reported by Kenneth R. Robinette
6150 <support@securenetterm.com>]
6152 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6153 routines: without these tracing memory leaks is very painful.
6154 Fix leaks in PKCS12 and PKCS7 routines.
6157 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6158 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6159 effectively meant GeneralizedTime would never be used. Now it
6160 is initialised to -1 but X509_time_adj() now has to check the value
6161 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6162 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6163 [Steve Henson, reported by Kenneth R. Robinette
6164 <support@securenetterm.com>]
6166 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6167 result in a zero length in the ASN1_INTEGER structure which was
6168 not consistent with the structure when d2i_ASN1_INTEGER() was used
6169 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6170 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6171 where it did not print out a minus for negative ASN1_INTEGER.
6174 *) Add summary printout to ocsp utility. The various functions which
6175 convert status values to strings have been renamed to:
6176 OCSP_response_status_str(), OCSP_cert_status_str() and
6177 OCSP_crl_reason_str() and are no longer static. New options
6178 to verify nonce values and to disable verification. OCSP response
6179 printout format cleaned up.
6182 *) Add additional OCSP certificate checks. These are those specified
6183 in RFC2560. This consists of two separate checks: the CA of the
6184 certificate being checked must either be the OCSP signer certificate
6185 or the issuer of the OCSP signer certificate. In the latter case the
6186 OCSP signer certificate must contain the OCSP signing extended key
6187 usage. This check is performed by attempting to match the OCSP
6188 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6189 in the OCSP_CERTID structures of the response.
6192 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6193 and related routines. This uses the standard OpenSSL certificate
6194 verify routines to perform initial checks (just CA validity) and
6195 to obtain the certificate chain. Then additional checks will be
6196 performed on the chain. Currently the root CA is checked to see
6197 if it is explicitly trusted for OCSP signing. This is used to set
6198 a root CA as a global signing root: that is any certificate that
6199 chains to that CA is an acceptable OCSP signing certificate.
6202 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6203 extensions from a separate configuration file.
6204 As when reading extensions from the main configuration file,
6205 the '-extensions ...' option may be used for specifying the
6207 [Massimiliano Pala <madwolf@comune.modena.it>]
6209 *) New OCSP utility. Allows OCSP requests to be generated or
6210 read. The request can be sent to a responder and the output
6211 parsed, outputed or printed in text form. Not complete yet:
6212 still needs to check the OCSP response validity.
6215 *) New subcommands for 'openssl ca':
6216 'openssl ca -status <serial>' prints the status of the cert with
6217 the given serial number (according to the index file).
6218 'openssl ca -updatedb' updates the expiry status of certificates
6220 [Massimiliano Pala <madwolf@comune.modena.it>]
6222 *) New '-newreq-nodes' command option to CA.pl. This is like
6223 '-newreq', but calls 'openssl req' with the '-nodes' option
6224 so that the resulting key is not encrypted.
6225 [Damien Miller <djm@mindrot.org>]
6227 *) New configuration for the GNU Hurd.
6228 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6230 *) Initial code to implement OCSP basic response verify. This
6231 is currently incomplete. Currently just finds the signer's
6232 certificate and verifies the signature on the response.
6235 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6236 value of OPENSSLDIR. This is available via the new '-d' option
6237 to 'openssl version', and is also included in 'openssl version -a'.
6240 *) Allowing defining memory allocation callbacks that will be given
6241 file name and line number information in additional arguments
6242 (a const char* and an int). The basic functionality remains, as
6243 well as the original possibility to just replace malloc(),
6244 realloc() and free() by functions that do not know about these
6245 additional arguments. To register and find out the current
6246 settings for extended allocation functions, the following
6247 functions are provided:
6249 CRYPTO_set_mem_ex_functions
6250 CRYPTO_set_locked_mem_ex_functions
6251 CRYPTO_get_mem_ex_functions
6252 CRYPTO_get_locked_mem_ex_functions
6254 These work the same way as CRYPTO_set_mem_functions and friends.
6255 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6256 extended allocation function is enabled.
6257 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6258 a conventional allocation function is enabled.
6259 [Richard Levitte, Bodo Moeller]
6261 *) Finish off removing the remaining LHASH function pointer casts.
6262 There should no longer be any prototype-casting required when using
6263 the LHASH abstraction, and any casts that remain are "bugs". See
6264 the callback types and macros at the head of lhash.h for details
6265 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6268 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6269 If /dev/[u]random devices are not available or do not return enough
6270 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6272 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6273 /etc/entropy will be queried once each in this sequence, quering stops
6274 when enough entropy was collected without querying more sockets.
6277 *) Change the Unix RAND_poll() variant to be able to poll several
6278 random devices, as specified by DEVRANDOM, until a sufficient amount
6279 of data has been collected. We spend at most 10 ms on each file
6280 (select timeout) and read in non-blocking mode. DEVRANDOM now
6281 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6282 (previously it was just the string "/dev/urandom"), so on typical
6283 platforms the 10 ms delay will never occur.
6284 Also separate out the Unix variant to its own file, rand_unix.c.
6285 For VMS, there's a currently-empty rand_vms.c.
6288 *) Move OCSP client related routines to ocsp_cl.c. These
6289 provide utility functions which an application needing
6290 to issue a request to an OCSP responder and analyse the
6291 response will typically need: as opposed to those which an
6292 OCSP responder itself would need which will be added later.
6294 OCSP_request_sign() signs an OCSP request with an API similar
6295 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6296 response. OCSP_response_get1_basic() extracts basic response
6297 from response. OCSP_resp_find_status(): finds and extracts status
6298 information from an OCSP_CERTID structure (which will be created
6299 when the request structure is built). These are built from lower
6300 level functions which work on OCSP_SINGLERESP structures but
6301 wont normally be used unless the application wishes to examine
6302 extensions in the OCSP response for example.
6304 Replace nonce routines with a pair of functions.
6305 OCSP_request_add1_nonce() adds a nonce value and optionally
6306 generates a random value. OCSP_check_nonce() checks the
6307 validity of the nonce in an OCSP response.
6310 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6311 This doesn't copy the supplied OCSP_CERTID and avoids the
6312 need to free up the newly created id. Change return type
6313 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6314 This can then be used to add extensions to the request.
6315 Deleted OCSP_request_new(), since most of its functionality
6316 is now in OCSP_REQUEST_new() (and the case insensitive name
6317 clash) apart from the ability to set the request name which
6318 will be added elsewhere.
6321 *) Update OCSP API. Remove obsolete extensions argument from
6322 various functions. Extensions are now handled using the new
6323 OCSP extension code. New simple OCSP HTTP function which
6324 can be used to send requests and parse the response.
6327 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6328 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6329 uses the special reorder version of SET OF to sort the attributes
6330 and reorder them to match the encoded order. This resolves a long
6331 standing problem: a verify on a PKCS7 structure just after signing
6332 it used to fail because the attribute order did not match the
6333 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6334 it uses the received order. This is necessary to tolerate some broken
6335 software that does not order SET OF. This is handled by encoding
6336 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6337 to produce the required SET OF.
6340 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6341 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6342 files to get correct declarations of the ASN.1 item variables.
6345 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6346 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6347 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6348 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6349 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6350 ASN1_ITEM and no wrapper functions.
6353 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6354 replace the old function pointer based I/O routines. Change most of
6355 the *_d2i_bio() and *_d2i_fp() functions to use these.
6358 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6359 lines, recognice more "algorithms" that can be deselected, and make
6360 it complain about algorithm deselection that isn't recognised.
6363 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6364 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6365 to use new functions. Add NO_ASN1_OLD which can be set to remove
6366 some old style ASN1 functions: this can be used to determine if old
6367 code will still work when these eventually go away.
6370 *) New extension functions for OCSP structures, these follow the
6371 same conventions as certificates and CRLs.
6374 *) New function X509V3_add1_i2d(). This automatically encodes and
6375 adds an extension. Its behaviour can be customised with various
6376 flags to append, replace or delete. Various wrappers added for
6377 certifcates and CRLs.
6380 *) Fix to avoid calling the underlying ASN1 print routine when
6381 an extension cannot be parsed. Correct a typo in the
6382 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6385 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6386 entries for variables.
6389 *) Add functionality to apps/openssl.c for detecting locking
6390 problems: As the program is single-threaded, all we have
6391 to do is register a locking callback using an array for
6392 storing which locks are currently held by the program.
6395 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6396 SSL_get_ex_data_X509_STORE_idx(), which is used in
6397 ssl_verify_cert_chain() and thus can be called at any time
6398 during TLS/SSL handshakes so that thread-safety is essential.
6399 Unfortunately, the ex_data design is not at all suited
6400 for multi-threaded use, so it probably should be abolished.
6403 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6404 [Broadcom, tweaked and integrated by Geoff Thorpe]
6406 *) Move common extension printing code to new function
6407 X509V3_print_extensions(). Reorganise OCSP print routines and
6408 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6411 *) New function X509_signature_print() to remove duplication in some
6415 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6416 set (this was treated exactly the same as SET OF previously). This
6417 is used to reorder the STACK representing the structure to match the
6418 encoding. This will be used to get round a problem where a PKCS7
6419 structure which was signed could not be verified because the STACK
6420 order did not reflect the encoded order.
6423 *) Reimplement the OCSP ASN1 module using the new code.
6426 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6427 for its ASN1 operations. The old style function pointers still exist
6428 for now but they will eventually go away.
6431 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6432 completely replaces the old ASN1 functionality with a table driven
6433 encoder and decoder which interprets an ASN1_ITEM structure describing
6434 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6435 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6436 has also been converted to the new form.
6439 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6440 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6441 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6442 for negative moduli.
6445 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6446 of not touching the result's sign bit.
6449 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6453 *) Changed the LHASH code to use prototypes for callbacks, and created
6454 macros to declare and implement thin (optionally static) functions
6455 that provide type-safety and avoid function pointer casting for the
6456 type-specific callbacks.
6459 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6461 [Veers Staats <staatsvr@asc.hpc.mil>,
6462 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6464 *) Reformat the FAQ so the different questions and answers can be divided
6465 in sections depending on the subject.
6468 *) Have the zlib compression code load ZLIB.DLL dynamically under
6472 *) New function BN_mod_sqrt for computing square roots modulo a prime
6473 (using the probabilistic Tonelli-Shanks algorithm unless
6474 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6475 be handled deterministically).
6476 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6478 *) Make BN_mod_inverse faster by explicitly handling small quotients
6479 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6480 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6483 *) New function BN_kronecker.
6486 *) Fix BN_gcd so that it works on negative inputs; the result is
6487 positive unless both parameters are zero.
6488 Previously something reasonably close to an infinite loop was
6489 possible because numbers could be growing instead of shrinking
6490 in the implementation of Euclid's algorithm.
6493 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6494 sign of the number in question.
6496 Fix BN_is_word(a,w) to work correctly for w == 0.
6498 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6499 because its test if the absolute value of 'a' equals 'w'.
6500 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6501 it exists mostly for use in the implementations of BN_is_zero(),
6502 BN_is_one(), and BN_is_word().
6505 *) New function BN_swap.
6508 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6509 the exponentiation functions are more likely to produce reasonable
6510 results on negative inputs.
6513 *) Change BN_mod_mul so that the result is always non-negative.
6514 Previously, it could be negative if one of the factors was negative;
6515 I don't think anyone really wanted that behaviour.
6518 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6519 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6520 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6521 and add new functions:
6530 BN_mod_lshift1_quick
6534 These functions always generate non-negative results.
6536 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6537 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6539 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6540 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6541 be reduced modulo m.
6542 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6545 The following entry accidentally appeared in the CHANGES file
6546 distributed with OpenSSL 0.9.7. The modifications described in
6547 it do *not* apply to OpenSSL 0.9.7.
6549 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6550 was actually never needed) and in BN_mul(). The removal in BN_mul()
6551 required a small change in bn_mul_part_recursive() and the addition
6552 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6553 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6554 bn_sub_words() and bn_add_words() except they take arrays with
6559 *) In 'openssl passwd', verify passwords read from the terminal
6560 unless the '-salt' option is used (which usually means that
6561 verification would just waste user's time since the resulting
6562 hash is going to be compared with some given password hash)
6563 or the new '-noverify' option is used.
6565 This is an incompatible change, but it does not affect
6566 non-interactive use of 'openssl passwd' (passwords on the command
6567 line, '-stdin' option, '-in ...' option) and thus should not
6571 *) Remove all references to RSAref, since there's no more need for it.
6574 *) Make DSO load along a path given through an environment variable
6575 (SHLIB_PATH) with shl_load().
6578 *) Constify the ENGINE code as a result of BIGNUM constification.
6579 Also constify the RSA code and most things related to it. In a
6580 few places, most notable in the depth of the ASN.1 code, ugly
6581 casts back to non-const were required (to be solved at a later
6585 *) Make it so the openssl application has all engines loaded by default.
6588 *) Constify the BIGNUM routines a little more.
6591 *) Add the following functions:
6593 ENGINE_load_cswift()
6595 ENGINE_load_atalla()
6597 ENGINE_load_builtin_engines()
6599 That way, an application can itself choose if external engines that
6600 are built-in in OpenSSL shall ever be used or not. The benefit is
6601 that applications won't have to be linked with libdl or other dso
6602 libraries unless it's really needed.
6604 Changed 'openssl engine' to load all engines on demand.
6605 Changed the engine header files to avoid the duplication of some
6606 declarations (they differed!).
6609 *) 'openssl engine' can now list capabilities.
6612 *) Better error reporting in 'openssl engine'.
6615 *) Never call load_dh_param(NULL) in s_server.
6618 *) Add engine application. It can currently list engines by name and
6619 identity, and test if they are actually available.
6622 *) Improve RPM specification file by forcing symbolic linking and making
6623 sure the installed documentation is also owned by root.root.
6624 [Damien Miller <djm@mindrot.org>]
6626 *) Give the OpenSSL applications more possibilities to make use of
6627 keys (public as well as private) handled by engines.
6630 *) Add OCSP code that comes from CertCo.
6633 *) Add VMS support for the Rijndael code.
6636 *) Added untested support for Nuron crypto accelerator.
6639 *) Add support for external cryptographic devices. This code was
6640 previously distributed separately as the "engine" branch.
6641 [Geoff Thorpe, Richard Levitte]
6643 *) Rework the filename-translation in the DSO code. It is now possible to
6644 have far greater control over how a "name" is turned into a filename
6645 depending on the operating environment and any oddities about the
6646 different shared library filenames on each system.
6649 *) Support threads on FreeBSD-elf in Configure.
6652 *) Fix for SHA1 assembly problem with MASM: it produces
6653 warnings about corrupt line number information when assembling
6654 with debugging information. This is caused by the overlapping
6656 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6659 NCONF_get_number() has no error checking at all. As a replacement,
6660 NCONF_get_number_e() is defined (_e for "error checking") and is
6661 promoted strongly. The old NCONF_get_number is kept around for
6662 binary backward compatibility.
6663 Make it possible for methods to load from something other than a BIO,
6664 by providing a function pointer that is given a name instead of a BIO.
6665 For example, this could be used to load configuration data from an
6669 *) Fix for non blocking accept BIOs. Added new I/O special reason
6670 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6671 with non blocking I/O was not possible because no retry code was
6672 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6676 *) Added the beginnings of Rijndael support.
6679 *) Fix for bug in DirectoryString mask setting. Add support for
6680 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6681 to allow certificate printing to more controllable, additional
6682 'certopt' option to 'x509' to allow new printing options to be
6686 *) Clean old EAY MD5 hack from e_os.h.
6689 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6691 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6692 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6693 [Joe Orton, Steve Henson]
6695 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6697 *) Fix additional bug revealed by the NISCC test suite:
6699 Stop bug triggering large recursion when presented with
6700 certain ASN.1 tags (CVE-2003-0851)
6703 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6705 *) Fix various bugs revealed by running the NISCC test suite:
6707 Stop out of bounds reads in the ASN1 code when presented with
6708 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6710 If verify callback ignores invalid public key errors don't try to check
6711 certificate signature with the NULL public key.
6715 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6716 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6720 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6721 extra data after the compression methods not only for TLS 1.0
6722 but also for SSL 3.0 (as required by the specification).
6723 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6725 *) Change X509_certificate_type() to mark the key as exported/exportable
6726 when it's 512 *bits* long, not 512 bytes.
6729 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6731 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6732 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6733 a protocol version number mismatch like a decryption error
6734 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6737 *) Turn on RSA blinding by default in the default implementation
6738 to avoid a timing attack. Applications that don't want it can call
6739 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6740 They would be ill-advised to do so in most cases.
6741 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6743 *) Change RSA blinding code so that it works when the PRNG is not
6744 seeded (in this case, the secret RSA exponent is abused as
6745 an unpredictable seed -- if it is not unpredictable, there
6746 is no point in blinding anyway). Make RSA blinding thread-safe
6747 by remembering the creator's thread ID in rsa->blinding and
6748 having all other threads use local one-time blinding factors
6749 (this requires more computation than sharing rsa->blinding, but
6750 avoids excessive locking; and if an RSA object is not shared
6751 between threads, blinding will still be very fast).
6754 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6756 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6757 via timing by performing a MAC computation even if incorrrect
6758 block cipher padding has been found. This is a countermeasure
6759 against active attacks where the attacker has to distinguish
6760 between bad padding and a MAC verification error. (CVE-2003-0078)
6762 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6763 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6764 Martin Vuagnoux (EPFL, Ilion)]
6766 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6768 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6769 memory from it's contents. This is done with a counter that will
6770 place alternating values in each byte. This can be used to solve
6771 two issues: 1) the removal of calls to memset() by highly optimizing
6772 compilers, and 2) cleansing with other values than 0, since those can
6773 be read through on certain media, for example a swap space on disk.
6776 *) Bugfix: client side session caching did not work with external caching,
6777 because the session->cipher setting was not restored when reloading
6778 from the external cache. This problem was masked, when
6779 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6780 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6783 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6784 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6785 [Zeev Lieber <zeev-l@yahoo.com>]
6787 *) Undo an undocumented change introduced in 0.9.6e which caused
6788 repeated calls to OpenSSL_add_all_ciphers() and
6789 OpenSSL_add_all_digests() to be ignored, even after calling
6793 *) Change the default configuration reader to deal with last line not
6794 being properly terminated.
6797 *) Change X509_NAME_cmp() so it applies the special rules on handling
6798 DN values that are of type PrintableString, as well as RDNs of type
6799 emailAddress where the value has the type ia5String.
6800 [stefank@valicert.com via Richard Levitte]
6802 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6803 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6804 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6805 the bitwise-OR of the two for use by the majority of applications
6806 wanting this behaviour, and update the docs. The documented
6807 behaviour and actual behaviour were inconsistent and had been
6808 changing anyway, so this is more a bug-fix than a behavioural
6810 [Geoff Thorpe, diagnosed by Nadav Har'El]
6812 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6813 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6816 *) Fix initialization code race conditions in
6817 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6818 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6819 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6820 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6821 ssl2_get_cipher_by_char(),
6822 ssl3_get_cipher_by_char().
6823 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6825 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6826 the cached sessions are flushed, as the remove_cb() might use ex_data
6827 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6828 (see [openssl.org #212]).
6829 [Geoff Thorpe, Lutz Jaenicke]
6831 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6832 length, instead of the encoding length to d2i_ASN1_OBJECT.
6835 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6837 *) [In 0.9.6g-engine release:]
6838 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6839 [Lynn Gazis <lgazis@rainbow.com>]
6841 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6843 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6844 and get fix the header length calculation.
6845 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6846 Alon Kantor <alonk@checkpoint.com> (and others),
6849 *) Use proper error handling instead of 'assertions' in buffer
6850 overflow checks added in 0.9.6e. This prevents DoS (the
6851 assertions could call abort()).
6852 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6854 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6856 *) Add various sanity checks to asn1_get_length() to reject
6857 the ASN1 length bytes if they exceed sizeof(long), will appear
6858 negative or the content length exceeds the length of the
6860 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6862 *) Fix cipher selection routines: ciphers without encryption had no flags
6863 for the cipher strength set and where therefore not handled correctly
6864 by the selection routines (PR #130).
6867 *) Fix EVP_dsa_sha macro.
6871 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6872 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6873 that was added in OpenSSL 0.9.6d.
6875 As the countermeasure turned out to be incompatible with some
6876 broken SSL implementations, the new option is part of SSL_OP_ALL.
6877 SSL_OP_ALL is usually employed when compatibility with weird SSL
6878 implementations is desired (e.g. '-bugs' option to 's_client' and
6879 's_server'), so the new option is automatically set in many
6883 *) Changes in security patch:
6885 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6886 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6887 Air Force Materiel Command, USAF, under agreement number
6890 *) Add various sanity checks to asn1_get_length() to reject
6891 the ASN1 length bytes if they exceed sizeof(long), will appear
6892 negative or the content length exceeds the length of the
6893 supplied buffer. (CVE-2002-0659)
6894 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6896 *) Assertions for various potential buffer overflows, not known to
6898 [Ben Laurie (CHATS)]
6900 *) Various temporary buffers to hold ASCII versions of integers were
6901 too small for 64 bit platforms. (CVE-2002-0655)
6902 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6904 *) Remote buffer overflow in SSL3 protocol - an attacker could
6905 supply an oversized session ID to a client. (CVE-2002-0656)
6906 [Ben Laurie (CHATS)]
6908 *) Remote buffer overflow in SSL2 protocol - an attacker could
6909 supply an oversized client master key. (CVE-2002-0656)
6910 [Ben Laurie (CHATS)]
6912 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6914 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6915 encoded as NULL) with id-dsa-with-sha1.
6916 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6918 *) Check various X509_...() return values in apps/req.c.
6919 [Nils Larsch <nla@trustcenter.de>]
6921 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6922 an end-of-file condition would erronously be flagged, when the CRLF
6923 was just at the end of a processed block. The bug was discovered when
6924 processing data through a buffering memory BIO handing the data to a
6925 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6926 <ptsekov@syntrex.com> and Nedelcho Stanev.
6929 *) Implement a countermeasure against a vulnerability recently found
6930 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6931 before application data chunks to avoid the use of known IVs
6932 with data potentially chosen by the attacker.
6935 *) Fix length checks in ssl3_get_client_hello().
6938 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6939 to prevent ssl3_read_internal() from incorrectly assuming that
6940 ssl3_read_bytes() found application data while handshake
6941 processing was enabled when in fact s->s3->in_read_app_data was
6942 merely automatically cleared during the initial handshake.
6943 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6945 *) Fix object definitions for Private and Enterprise: they were not
6946 recognized in their shortname (=lowercase) representation. Extend
6947 obj_dat.pl to issue an error when using undefined keywords instead
6948 of silently ignoring the problem (Svenning Sorensen
6949 <sss@sss.dnsalias.net>).
6952 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6953 generators, i.e. generators other than 2 and 5. (Previously, the
6954 code did not properly initialise the 'add' and 'rem' values to
6955 BN_generate_prime().)
6957 In the new general case, we do not insist that 'generator' is
6958 actually a primitive root: This requirement is rather pointless;
6959 a generator of the order-q subgroup is just as good, if not
6963 *) Map new X509 verification errors to alerts. Discovered and submitted by
6964 Tom Wu <tom@arcot.com>.
6967 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6968 returning non-zero before the data has been completely received
6969 when using non-blocking I/O.
6970 [Bodo Moeller; problem pointed out by John Hughes]
6972 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6973 [Ben Laurie, Lutz Jaenicke]
6975 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6976 Yoram Zahavi <YoramZ@gilian.com>).
6979 *) Add information about CygWin 1.3 and on, and preserve proper
6980 configuration for the versions before that.
6981 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6983 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6984 check whether we deal with a copy of a session and do not delete from
6985 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6986 <izhar@checkpoint.com>.
6989 *) Do not store session data into the internal session cache, if it
6990 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6991 flag is set). Proposed by Aslam <aslam@funk.com>.
6994 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6998 *) [In 0.9.6d-engine release:]
6999 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7000 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7002 *) Add the configuration target linux-s390x.
7003 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
7005 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
7006 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
7007 variable as an indication that a ClientHello message has been
7008 received. As the flag value will be lost between multiple
7009 invocations of ssl3_accept when using non-blocking I/O, the
7010 function may not be aware that a handshake has actually taken
7011 place, thus preventing a new session from being added to the
7014 To avoid this problem, we now set s->new_session to 2 instead of
7015 using a local variable.
7016 [Lutz Jaenicke, Bodo Moeller]
7018 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
7019 if the SSL_R_LENGTH_MISMATCH error is detected.
7020 [Geoff Thorpe, Bodo Moeller]
7022 *) New 'shared_ldflag' column in Configure platform table.
7025 *) Fix EVP_CIPHER_mode macro.
7026 ["Dan S. Camper" <dan@bti.net>]
7028 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
7029 type, we must throw them away by setting rr->length to 0.
7030 [D P Chang <dpc@qualys.com>]
7032 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
7034 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
7035 <Dominikus.Scherkl@biodata.com>. (The previous implementation
7036 worked incorrectly for those cases where range = 10..._2 and
7037 3*range is two bits longer than range.)
7040 *) Only add signing time to PKCS7 structures if it is not already
7044 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
7045 OBJ_ld_ce should be OBJ_id_ce.
7046 Also some ip-pda OIDs in crypto/objects/objects.txt were
7047 incorrect (cf. RFC 3039).
7048 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
7050 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
7051 returns early because it has nothing to do.
7052 [Andy Schneider <andy.schneider@bjss.co.uk>]
7054 *) [In 0.9.6c-engine release:]
7055 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
7056 [Andy Schneider <andy.schneider@bjss.co.uk>]
7058 *) [In 0.9.6c-engine release:]
7059 Add support for Cryptographic Appliance's keyserver technology.
7060 (Use engine 'keyclient')
7061 [Cryptographic Appliances and Geoff Thorpe]
7063 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
7064 is called via tools/c89.sh because arguments have to be
7065 rearranged (all '-L' options must appear before the first object
7067 [Richard Shapiro <rshapiro@abinitio.com>]
7069 *) [In 0.9.6c-engine release:]
7070 Add support for Broadcom crypto accelerator cards, backported
7072 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
7074 *) [In 0.9.6c-engine release:]
7075 Add support for SureWare crypto accelerator cards from
7076 Baltimore Technologies. (Use engine 'sureware')
7077 [Baltimore Technologies and Mark Cox]
7079 *) [In 0.9.6c-engine release:]
7080 Add support for crypto accelerator cards from Accelerated
7081 Encryption Processing, www.aep.ie. (Use engine 'aep')
7082 [AEP Inc. and Mark Cox]
7084 *) Add a configuration entry for gcc on UnixWare.
7085 [Gary Benson <gbenson@redhat.com>]
7087 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
7088 messages are stored in a single piece (fixed-length part and
7089 variable-length part combined) and fix various bugs found on the way.
7092 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
7093 instead. BIO_gethostbyname() does not know what timeouts are
7094 appropriate, so entries would stay in cache even when they have
7096 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
7098 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
7099 faced with a pathologically small ClientHello fragment that does
7100 not contain client_version: Instead of aborting with an error,
7101 simply choose the highest available protocol version (i.e.,
7102 TLS 1.0 unless it is disabled). In practice, ClientHello
7103 messages are never sent like this, but this change gives us
7104 strictly correct behaviour at least for TLS.
7107 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
7108 never resets s->method to s->ctx->method when called from within
7109 one of the SSL handshake functions.
7110 [Bodo Moeller; problem pointed out by Niko Baric]
7112 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
7113 (sent using the client's version number) if client_version is
7114 smaller than the protocol version in use. Also change
7115 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
7116 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
7117 the client will at least see that alert.
7120 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
7124 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
7125 client receives HelloRequest while in a handshake.
7126 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
7128 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
7129 should end in 'break', not 'goto end' which circumvents various
7130 cleanups done in state SSL_ST_OK. But session related stuff
7131 must be disabled for SSL_ST_OK in the case that we just sent a
7134 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7135 before just sending a HelloRequest.
7136 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7138 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7139 reveal whether illegal block cipher padding was found or a MAC
7140 verification error occured. (Neither SSLerr() codes nor alerts
7141 are directly visible to potential attackers, but the information
7142 may leak via logfiles.)
7144 Similar changes are not required for the SSL 2.0 implementation
7145 because the number of padding bytes is sent in clear for SSL 2.0,
7146 and the extra bytes are just ignored. However ssl/s2_pkt.c
7147 failed to verify that the purported number of padding bytes is in
7151 *) Add OpenUNIX-8 support including shared libraries
7152 (Boyd Lynn Gerber <gerberb@zenez.com>).
7155 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7156 'wristwatch attack' using huge encoding parameters (cf.
7157 James H. Manger's CRYPTO 2001 paper). Note that the
7158 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7159 encoding parameters and hence was not vulnerable.
7162 *) BN_sqr() bug fix.
7163 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7165 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7166 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7167 followed by modular reduction.
7168 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7170 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7171 equivalent based on BN_pseudo_rand() instead of BN_rand().
7174 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7175 This function was broken, as the check for a new client hello message
7176 to handle SGC did not allow these large messages.
7177 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7180 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7183 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7184 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7187 *) Rework the configuration and shared library support for Tru64 Unix.
7188 The configuration part makes use of modern compiler features and
7189 still retains old compiler behavior for those that run older versions
7190 of the OS. The shared library support part includes a variant that
7191 uses the RPATH feature, and is available through the special
7192 configuration target "alpha-cc-rpath", which will never be selected
7194 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7196 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7197 with the same message size as in ssl3_get_certificate_request().
7198 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7199 messages might inadvertently be reject as too long.
7200 [Petr Lampa <lampa@fee.vutbr.cz>]
7202 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7205 *) Modified SSL library such that the verify_callback that has been set
7206 specificly for an SSL object with SSL_set_verify() is actually being
7207 used. Before the change, a verify_callback set with this function was
7208 ignored and the verify_callback() set in the SSL_CTX at the time of
7209 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7210 to allow the necessary settings.
7213 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7214 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7215 done automatically (in contradiction to the requirements of the C
7216 standard). This made problems when used from OpenSSH.
7219 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7220 dh->length and always used
7222 BN_rand_range(priv_key, dh->p).
7224 BN_rand_range() is not necessary for Diffie-Hellman, and this
7225 specific range makes Diffie-Hellman unnecessarily inefficient if
7226 dh->length (recommended exponent length) is much smaller than the
7227 length of dh->p. We could use BN_rand_range() if the order of
7228 the subgroup was stored in the DH structure, but we only have
7233 BN_rand(priv_key, l, ...)
7235 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7241 RSA_eay_public_encrypt
7242 RSA_eay_private_decrypt
7243 RSA_eay_private_encrypt (signing)
7244 RSA_eay_public_decrypt (signature verification)
7246 (default implementations for RSA_public_encrypt,
7247 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7248 always reject numbers >= n.
7251 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7252 to synchronize access to 'locking_thread'. This is necessary on
7253 systems where access to 'locking_thread' (an 'unsigned long'
7254 variable) is not atomic.
7257 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7258 *before* setting the 'crypto_lock_rand' flag. The previous code had
7259 a race condition if 0 is a valid thread ID.
7260 [Travis Vitek <vitek@roguewave.com>]
7262 *) Add support for shared libraries under Irix.
7263 [Albert Chin-A-Young <china@thewrittenword.com>]
7265 *) Add configuration option to build on Linux on both big-endian and
7267 [Ralf Baechle <ralf@uni-koblenz.de>]
7269 *) Add the possibility to create shared libraries on HP-UX.
7272 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7274 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7275 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7276 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7277 PRNG state recovery was possible based on the output of
7278 one PRNG request appropriately sized to gain knowledge on
7279 'md' followed by enough consecutive 1-byte PRNG requests
7280 to traverse all of 'state'.
7282 1. When updating 'md_local' (the current thread's copy of 'md')
7283 during PRNG output generation, hash all of the previous
7284 'md_local' value, not just the half used for PRNG output.
7286 2. Make the number of bytes from 'state' included into the hash
7287 independent from the number of PRNG bytes requested.
7289 The first measure alone would be sufficient to avoid
7290 Markku-Juhani's attack. (Actually it had never occurred
7291 to me that the half of 'md_local' used for chaining was the
7292 half from which PRNG output bytes were taken -- I had always
7293 assumed that the secret half would be used.) The second
7294 measure makes sure that additional data from 'state' is never
7295 mixed into 'md_local' in small portions; this heuristically
7296 further strengthens the PRNG.
7299 *) Fix crypto/bn/asm/mips3.s.
7302 *) When only the key is given to "enc", the IV is undefined. Print out
7303 an error message in this case.
7306 *) Handle special case when X509_NAME is empty in X509 printing routines.
7309 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7310 positive and less than q.
7313 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7314 used: it isn't thread safe and the add_lock_callback should handle
7316 [Paul Rose <Paul.Rose@bridge.com>]
7318 *) Verify that incoming data obeys the block size in
7319 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7323 [Ulf Möller, Bodo Möller]
7325 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7326 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7327 when fixing the server behaviour for backwards-compatible 'client
7328 hello' messages. (Note that the attack is impractical against
7329 SSL 3.0 and TLS 1.0 anyway because length and version checking
7330 means that the probability of guessing a valid ciphertext is
7331 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7334 Before 0.9.5, the countermeasure (hide the error by generating a
7335 random 'decryption result') did not work properly because
7336 ERR_clear_error() was missing, meaning that SSL_get_error() would
7337 detect the supposedly ignored error.
7339 Both problems are now fixed.
7342 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7343 (previously it was 1024).
7346 *) Fix for compatibility mode trust settings: ignore trust settings
7347 unless some valid trust or reject settings are present.
7350 *) Fix for blowfish EVP: its a variable length cipher.
7353 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7354 parameters in DSA public key structures and return an error in the
7355 DSA routines if parameters are absent.
7358 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7359 in the current directory if neither $RANDFILE nor $HOME was set.
7360 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7361 caused some confusion to Windows users who haven't defined $HOME.
7362 Thus RAND_file_name() is changed again: e_os.h can define a
7363 DEFAULT_HOME, which will be used if $HOME is not set.
7364 For Windows, we use "C:"; on other platforms, we still require
7365 environment variables.
7367 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7368 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7369 having multiple threads call RAND_poll() concurrently.
7372 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7373 combination of a flag and a thread ID variable.
7374 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7375 flag), *other* threads can enter ssleay_add_bytes without obeying
7376 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7377 that they do not hold after the first thread unsets add_do_not_lock).
7380 *) Change bctest again: '-x' expressions are not available in all
7384 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7386 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7387 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7389 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7390 the default extension for executables, if any. Also, make the perl
7391 scripts that use symlink() to test if it really exists and use "cp"
7392 if it doesn't. All this made OpenSSL compilable and installable in
7396 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7397 If SEQUENCE is length is indefinite just set c->slen to the total
7398 amount of data available.
7399 [Steve Henson, reported by shige@FreeBSD.org]
7400 [This change does not apply to 0.9.7.]
7402 *) Change bctest to avoid here-documents inside command substitution
7403 (workaround for FreeBSD /bin/sh bug).
7404 For compatibility with Ultrix, avoid shell functions (introduced
7405 in the bctest version that searches along $PATH).
7408 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7409 with des_encrypt() defined on some operating systems, like Solaris
7413 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7414 On the Importance of Eliminating Errors in Cryptographic
7415 Computations, J. Cryptology 14 (2001) 2, 101-119,
7416 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7419 *) MIPS assembler BIGNUM division bug fix.
7422 *) Disabled incorrect Alpha assembler code.
7425 *) Fix PKCS#7 decode routines so they correctly update the length
7426 after reading an EOC for the EXPLICIT tag.
7428 [This change does not apply to 0.9.7.]
7430 *) Fix bug in PKCS#12 key generation routines. This was triggered
7431 if a 3DES key was generated with a 0 initial byte. Include
7432 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7433 (but broken) behaviour.
7436 *) Enhance bctest to search for a working bc along $PATH and print
7438 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7440 *) Fix memory leaks in err.c: free err_data string if necessary;
7441 don't write to the wrong index in ERR_set_error_data.
7444 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7448 *) Replace rdtsc with _emit statements for VC++ version 5.
7449 [Jeremy Cooper <jeremy@baymoo.org>]
7451 *) Make it possible to reuse SSLv2 sessions.
7454 *) In copy_email() check for >= 0 as a return value for
7455 X509_NAME_get_index_by_NID() since 0 is a valid index.
7456 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7458 *) Avoid coredump with unsupported or invalid public keys by checking if
7459 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7460 PKCS7_verify() fails with non detached data.
7463 *) Don't use getenv in library functions when run as setuid/setgid.
7464 New function OPENSSL_issetugid().
7467 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7468 due to incorrect handling of multi-threading:
7470 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7472 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7474 3. Count how many times MemCheck_off() has been called so that
7475 nested use can be treated correctly. This also avoids
7476 inband-signalling in the previous code (which relied on the
7477 assumption that thread ID 0 is impossible).
7480 *) Add "-rand" option also to s_client and s_server.
7483 *) Fix CPU detection on Irix 6.x.
7484 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7485 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7487 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7490 [This change does not apply to 0.9.7.]
7492 *) Use the cached encoding of an X509_NAME structure rather than
7493 copying it. This is apparently the reason for the libsafe "errors"
7494 but the code is actually correct.
7497 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7498 Bleichenbacher's DSA attack.
7499 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7500 to be set and top=0 forces the highest bit to be set; top=-1 is new
7501 and leaves the highest bit random.
7502 [Ulf Moeller, Bodo Moeller]
7504 *) In the NCONF_...-based implementations for CONF_... queries
7505 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7506 a temporary CONF structure with the data component set to NULL
7507 (which gives segmentation faults in lh_retrieve).
7508 Instead, use NULL for the CONF pointer in CONF_get_string and
7509 CONF_get_number (which may use environment variables) and directly
7510 return NULL from CONF_get_section.
7513 *) Fix potential buffer overrun for EBCDIC.
7516 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7517 keyUsage if basicConstraints absent for a CA.
7520 *) Make SMIME_write_PKCS7() write mail header values with a format that
7521 is more generally accepted (no spaces before the semicolon), since
7522 some programs can't parse those values properly otherwise. Also make
7523 sure BIO's that break lines after each write do not create invalid
7527 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7528 macros previously used would not encode an empty SEQUENCE OF
7529 and break the signature.
7531 [This change does not apply to 0.9.7.]
7533 *) Zero the premaster secret after deriving the master secret in
7537 *) Add some EVP_add_digest_alias registrations (as found in
7538 OpenSSL_add_all_digests()) to SSL_library_init()
7539 aka OpenSSL_add_ssl_algorithms(). This provides improved
7540 compatibility with peers using X.509 certificates
7541 with unconventional AlgorithmIdentifier OIDs.
7544 *) Fix for Irix with NO_ASM.
7545 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7547 *) ./config script fixes.
7548 [Ulf Moeller, Richard Levitte]
7550 *) Fix 'openssl passwd -1'.
7553 *) Change PKCS12_key_gen_asc() so it can cope with non null
7554 terminated strings whose length is passed in the passlen
7555 parameter, for example from PEM callbacks. This was done
7556 by adding an extra length parameter to asc2uni().
7557 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7559 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7560 call failed, free the DSA structure.
7563 *) Fix to uni2asc() to cope with zero length Unicode strings.
7564 These are present in some PKCS#12 files.
7567 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7568 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7569 when writing a 32767 byte record.
7570 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7572 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7573 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7575 (RSA objects have a reference count access to which is protected
7576 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7577 so they are meant to be shared between threads.)
7578 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7579 "Reddie, Steven" <Steven.Reddie@ca.com>]
7581 *) Fix a deadlock in CRYPTO_mem_leaks().
7584 *) Use better test patterns in bntest.
7587 *) rand_win.c fix for Borland C.
7590 *) BN_rshift bugfix for n == 0.
7593 *) Add a 'bctest' script that checks for some known 'bc' bugs
7594 so that 'make test' does not abort just because 'bc' is broken.
7597 *) Store verify_result within SSL_SESSION also for client side to
7598 avoid potential security hole. (Re-used sessions on the client side
7599 always resulted in verify_result==X509_V_OK, not using the original
7600 result of the server certificate verification.)
7603 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7604 SSL3_RT_APPLICATION_DATA, return 0.
7605 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7609 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7610 releases, have been re-implemented by renaming the previous
7611 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7612 and ssl3_read_internal, respectively, and adding 'peek' parameters
7613 to them. The new ssl[23]_{read,peek} functions are calls to
7614 ssl[23]_read_internal with the 'peek' flag set appropriately.
7615 A 'peek' parameter has also been added to ssl3_read_bytes, which
7616 does the actual work for ssl3_read_internal.
7619 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7620 the method-specific "init()" handler. Also clean up ex_data after
7621 calling the method-specific "finish()" handler. Previously, this was
7622 happening the other way round.
7625 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7626 The previous value, 12, was not always sufficient for BN_mod_exp().
7629 *) Make sure that shared libraries get the internal name engine with
7630 the full version number and not just 0. This should mark the
7631 shared libraries as not backward compatible. Of course, this should
7632 be changed again when we can guarantee backward binary compatibility.
7635 *) Fix typo in get_cert_by_subject() in by_dir.c
7636 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7638 *) Rework the system to generate shared libraries:
7640 - Make note of the expected extension for the shared libraries and
7641 if there is a need for symbolic links from for example libcrypto.so.0
7642 to libcrypto.so.0.9.7. There is extended info in Configure for
7645 - Make as few rebuilds of the shared libraries as possible.
7647 - Still avoid linking the OpenSSL programs with the shared libraries.
7649 - When installing, install the shared libraries separately from the
7653 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7655 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7656 and not in SSL_clear because the latter is also used by the
7657 accept/connect functions; previously, the settings made by
7658 SSL_set_read_ahead would be lost during the handshake.
7659 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7661 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7662 Previously, it would create entries for disableed algorithms no
7666 *) Added several new manual pages for SSL_* function.
7669 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7671 *) In ssl23_get_client_hello, generate an error message when faced
7672 with an initial SSL 3.0/TLS record that is too small to contain the
7673 first two bytes of the ClientHello message, i.e. client_version.
7674 (Note that this is a pathologic case that probably has never happened
7675 in real life.) The previous approach was to use the version number
7676 from the record header as a substitute; but our protocol choice
7677 should not depend on that one because it is not authenticated
7678 by the Finished messages.
7681 *) More robust randomness gathering functions for Windows.
7682 [Jeffrey Altman <jaltman@columbia.edu>]
7684 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7685 not set then we don't setup the error code for issuer check errors
7686 to avoid possibly overwriting other errors which the callback does
7687 handle. If an application does set the flag then we assume it knows
7688 what it is doing and can handle the new informational codes
7692 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7693 a general "ANY" type, as such it should be able to decode anything
7694 including tagged types. However it didn't check the class so it would
7695 wrongly interpret tagged types in the same way as their universal
7696 counterpart and unknown types were just rejected. Changed so that the
7697 tagged and unknown types are handled in the same way as a SEQUENCE:
7698 that is the encoding is stored intact. There is also a new type
7699 "V_ASN1_OTHER" which is used when the class is not universal, in this
7700 case we have no idea what the actual type is so we just lump them all
7704 *) On VMS, stdout may very well lead to a file that is written to
7705 in a record-oriented fashion. That means that every write() will
7706 write a separate record, which will be read separately by the
7707 programs trying to read from it. This can be very confusing.
7709 The solution is to put a BIO filter in the way that will buffer
7710 text until a linefeed is reached, and then write everything a
7711 line at a time, so every record written will be an actual line,
7712 not chunks of lines and not (usually doesn't happen, but I've
7713 seen it once) several lines in one record. BIO_f_linebuffer() is
7716 Currently, it's a VMS-only method, because that's where it has
7717 been tested well enough.
7720 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7721 it can return incorrect results.
7722 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7723 but it was in 0.9.6-beta[12].)
7726 *) Disable the check for content being present when verifying detached
7727 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7728 include zero length content when signing messages.
7731 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7732 BIO_ctrl (for BIO pairs).
7735 *) Add DSO method for VMS.
7738 *) Bug fix: Montgomery multiplication could produce results with the
7742 *) Add RPM specification openssl.spec and modify it to build three
7743 packages. The default package contains applications, application
7744 documentation and run-time libraries. The devel package contains
7745 include files, static libraries and function documentation. The
7746 doc package contains the contents of the doc directory. The original
7747 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7750 *) Add a large number of documentation files for many SSL routines.
7751 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7753 *) Add a configuration entry for Sony News 4.
7754 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7756 *) Don't set the two most significant bits to one when generating a
7757 random number < q in the DSA library.
7760 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7761 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7762 the underlying transport is blocking) if a handshake took place.
7763 (The default behaviour is needed by applications such as s_client
7764 and s_server that use select() to determine when to use SSL_read;
7765 but for applications that know in advance when to expect data, it
7766 just makes things more complicated.)
7769 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7773 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7774 work better on such systems.
7775 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7777 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7778 Update PKCS12_parse() so it copies the friendlyName and the
7779 keyid to the certificates aux info.
7782 *) Fix bug in PKCS7_verify() which caused an infinite loop
7783 if there was more than one signature.
7784 [Sven Uszpelkat <su@celocom.de>]
7786 *) Major change in util/mkdef.pl to include extra information
7787 about each symbol, as well as presenting variables as well
7788 as functions. This change means that there's n more need
7789 to rebuild the .num files when some algorithms are excluded.
7792 *) Allow the verify time to be set by an application,
7793 rather than always using the current time.
7796 *) Phase 2 verify code reorganisation. The certificate
7797 verify code now looks up an issuer certificate by a
7798 number of criteria: subject name, authority key id
7799 and key usage. It also verifies self signed certificates
7800 by the same criteria. The main comparison function is
7801 X509_check_issued() which performs these checks.
7803 Lot of changes were necessary in order to support this
7804 without completely rewriting the lookup code.
7806 Authority and subject key identifier are now cached.
7808 The LHASH 'certs' is X509_STORE has now been replaced
7809 by a STACK_OF(X509_OBJECT). This is mainly because an
7810 LHASH can't store or retrieve multiple objects with
7811 the same hash value.
7813 As a result various functions (which were all internal
7814 use only) have changed to handle the new X509_STORE
7815 structure. This will break anything that messed round
7816 with X509_STORE internally.
7818 The functions X509_STORE_add_cert() now checks for an
7819 exact match, rather than just subject name.
7821 The X509_STORE API doesn't directly support the retrieval
7822 of multiple certificates matching a given criteria, however
7823 this can be worked round by performing a lookup first
7824 (which will fill the cache with candidate certificates)
7825 and then examining the cache for matches. This is probably
7826 the best we can do without throwing out X509_LOOKUP
7827 entirely (maybe later...).
7829 The X509_VERIFY_CTX structure has been enhanced considerably.
7831 All certificate lookup operations now go via a get_issuer()
7832 callback. Although this currently uses an X509_STORE it
7833 can be replaced by custom lookups. This is a simple way
7834 to bypass the X509_STORE hackery necessary to make this
7835 work and makes it possible to use more efficient techniques
7836 in future. A very simple version which uses a simple
7837 STACK for its trusted certificate store is also provided
7838 using X509_STORE_CTX_trusted_stack().
7840 The verify_cb() and verify() callbacks now have equivalents
7841 in the X509_STORE_CTX structure.
7843 X509_STORE_CTX also has a 'flags' field which can be used
7844 to customise the verify behaviour.
7847 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7848 excludes S/MIME capabilities.
7851 *) When a certificate request is read in keep a copy of the
7852 original encoding of the signed data and use it when outputing
7853 again. Signatures then use the original encoding rather than
7854 a decoded, encoded version which may cause problems if the
7855 request is improperly encoded.
7858 *) For consistency with other BIO_puts implementations, call
7859 buffer_write(b, ...) directly in buffer_puts instead of calling
7862 In BIO_puts, increment b->num_write as in BIO_write.
7863 [Peter.Sylvester@EdelWeb.fr]
7865 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7866 BN_zero, we may not return a BIGNUM with an array consisting of
7870 *) Avoid calling abort() from within the library when problems are
7871 detected, except if preprocessor symbols have been defined
7872 (such as REF_CHECK, BN_DEBUG etc.).
7875 *) New openssl application 'rsautl'. This utility can be
7876 used for low level RSA operations. DER public key
7877 BIO/fp routines also added.
7880 *) New Configure entry and patches for compiling on QNX 4.
7881 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7883 *) A demo state-machine implementation was sponsored by
7884 Nuron (http://www.nuron.com/) and is now available in
7885 demos/state_machine.
7888 *) New options added to the 'dgst' utility for signature
7889 generation and verification.
7892 *) Unrecognized PKCS#7 content types are now handled via a
7893 catch all ASN1_TYPE structure. This allows unsupported
7894 types to be stored as a "blob" and an application can
7895 encode and decode it manually.
7898 *) Fix various signed/unsigned issues to make a_strex.c
7900 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7902 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7903 length if passed a buffer. ASN1_INTEGER_to_BN failed
7904 if passed a NULL BN and its argument was negative.
7905 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7907 *) Modification to PKCS#7 encoding routines to output definite
7908 length encoding. Since currently the whole structures are in
7909 memory there's not real point in using indefinite length
7910 constructed encoding. However if OpenSSL is compiled with
7911 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7914 *) Added BIO_vprintf() and BIO_vsnprintf().
7917 *) Added more prefixes to parse for in the the strings written
7918 through a logging bio, to cover all the levels that are available
7919 through syslog. The prefixes are now:
7921 PANIC, EMERG, EMR => LOG_EMERG
7922 ALERT, ALR => LOG_ALERT
7923 CRIT, CRI => LOG_CRIT
7924 ERROR, ERR => LOG_ERR
7925 WARNING, WARN, WAR => LOG_WARNING
7926 NOTICE, NOTE, NOT => LOG_NOTICE
7927 INFO, INF => LOG_INFO
7928 DEBUG, DBG => LOG_DEBUG
7930 and as before, if none of those prefixes are present at the
7931 beginning of the string, LOG_ERR is chosen.
7933 On Win32, the LOG_* levels are mapped according to this:
7935 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7936 LOG_WARNING => EVENTLOG_WARNING_TYPE
7937 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7941 *) Made it possible to reconfigure with just the configuration
7942 argument "reconf" or "reconfigure". The command line arguments
7943 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7944 and are retrieved from there when reconfiguring.
7948 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7950 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7953 *) The obj_dat.pl script was messing up the sorting of object
7954 names. The reason was that it compared the quoted version
7955 of strings as a result "OCSP" > "OCSP Signing" because
7956 " > SPACE. Changed script to store unquoted versions of
7957 names and add quotes on output. It was also omitting some
7958 names from the lookup table if they were given a default
7959 value (that is if SN is missing it is given the same
7960 value as LN and vice versa), these are now added on the
7961 grounds that if an object has a name we should be able to
7962 look it up. Finally added warning output when duplicate
7963 short or long names are found.
7966 *) Changes needed for Tandem NSK.
7967 [Scott Uroff <scott@xypro.com>]
7969 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7970 RSA_padding_check_SSLv23(), special padding was never detected
7971 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7972 version rollback attacks was not effective.
7974 In s23_clnt.c, don't use special rollback-attack detection padding
7975 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7976 client; similarly, in s23_srvr.c, don't do the rollback check if
7977 SSL 2.0 is the only protocol enabled in the server.
7980 *) Make it possible to get hexdumps of unprintable data with 'openssl
7981 asn1parse'. By implication, the functions ASN1_parse_dump() and
7982 BIO_dump_indent() are added.
7985 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7986 these print out strings and name structures based on various
7987 flags including RFC2253 support and proper handling of
7988 multibyte characters. Added options to the 'x509' utility
7989 to allow the various flags to be set.
7992 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7993 Also change the functions X509_cmp_current_time() and
7994 X509_gmtime_adj() work with an ASN1_TIME structure,
7995 this will enable certificates using GeneralizedTime in validity
7996 dates to be checked.
7999 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
8000 negative public key encodings) on by default,
8001 NO_NEG_PUBKEY_BUG can be set to disable it.
8004 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
8005 content octets. An i2c_ASN1_OBJECT is unnecessary because
8006 the encoding can be trivially obtained from the structure.
8009 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
8010 not read locks (CRYPTO_r_[un]lock).
8013 *) A first attempt at creating official support for shared
8014 libraries through configuration. I've kept it so the
8015 default is static libraries only, and the OpenSSL programs
8016 are always statically linked for now, but there are
8017 preparations for dynamic linking in place.
8018 This has been tested on Linux and Tru64.
8021 *) Randomness polling function for Win9x, as described in:
8022 Peter Gutmann, Software Generation of Practically Strong
8026 *) Fix so PRNG is seeded in req if using an already existing
8030 *) New options to smime application. -inform and -outform
8031 allow alternative formats for the S/MIME message including
8032 PEM and DER. The -content option allows the content to be
8033 specified separately. This should allow things like Netscape
8034 form signing output easier to verify.
8037 *) Fix the ASN1 encoding of tags using the 'long form'.
8040 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
8041 STRING types. These convert content octets to and from the
8042 underlying type. The actual tag and length octets are
8043 already assumed to have been read in and checked. These
8044 are needed because all other string types have virtually
8045 identical handling apart from the tag. By having versions
8046 of the ASN1 functions that just operate on content octets
8047 IMPLICIT tagging can be handled properly. It also allows
8048 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
8049 and ASN1_INTEGER are identical apart from the tag.
8052 *) Change the handling of OID objects as follows:
8054 - New object identifiers are inserted in objects.txt, following
8055 the syntax given in objects.README.
8056 - objects.pl is used to process obj_mac.num and create a new
8058 - obj_dat.pl is used to create a new obj_dat.h, using the data in
8061 This is currently kind of a hack, and the perl code in objects.pl
8062 isn't very elegant, but it works as I intended. The simplest way
8063 to check that it worked correctly is to look in obj_dat.h and
8064 check the array nid_objs and make sure the objects haven't moved
8065 around (this is important!). Additions are OK, as well as
8066 consistent name changes.
8069 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
8072 *) Addition of the command line parameter '-rand file' to 'openssl req'.
8073 The given file adds to whatever has already been seeded into the
8074 random pool through the RANDFILE configuration file option or
8075 environment variable, or the default random state file.
8078 *) mkstack.pl now sorts each macro group into lexical order.
8079 Previously the output order depended on the order the files
8080 appeared in the directory, resulting in needless rewriting
8084 *) Patches to make OpenSSL compile under Win32 again. Mostly
8085 work arounds for the VC++ problem that it treats func() as
8086 func(void). Also stripped out the parts of mkdef.pl that
8087 added extra typesafe functions: these no longer exist.
8090 *) Reorganisation of the stack code. The macros are now all
8091 collected in safestack.h . Each macro is defined in terms of
8092 a "stack macro" of the form SKM_<name>(type, a, b). The
8093 DEBUG_SAFESTACK is now handled in terms of function casts,
8094 this has the advantage of retaining type safety without the
8095 use of additional functions. If DEBUG_SAFESTACK is not defined
8096 then the non typesafe macros are used instead. Also modified the
8097 mkstack.pl script to handle the new form. Needs testing to see
8098 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
8099 the default if no major problems. Similar behaviour for ASN1_SET_OF
8100 and PKCS12_STACK_OF.
8103 *) When some versions of IIS use the 'NET' form of private key the
8104 key derivation algorithm is different. Normally MD5(password) is
8105 used as a 128 bit RC4 key. In the modified case
8106 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
8107 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
8108 as the old Netscape_RSA functions except they have an additional
8109 'sgckey' parameter which uses the modified algorithm. Also added
8110 an -sgckey command line option to the rsa utility. Thanks to
8111 Adrian Peck <bertie@ncipher.com> for posting details of the modified
8112 algorithm to openssl-dev.
8115 *) The evp_local.h macros were using 'c.##kname' which resulted in
8116 invalid expansion on some systems (SCO 5.0.5 for example).
8117 Corrected to 'c.kname'.
8118 [Phillip Porch <root@theporch.com>]
8120 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
8121 a STACK of email addresses from a certificate or request, these look
8122 in the subject name and the subject alternative name extensions and
8123 omit any duplicate addresses.
8126 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
8127 This makes DSA verification about 2 % faster.
8130 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
8131 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8132 plus overhead for 1024 bit moduli).
8133 This makes exponentiations about 0.5 % faster for 1024 bit
8134 exponents (as measured by "openssl speed rsa2048").
8137 *) Rename memory handling macros to avoid conflicts with other
8139 Malloc => OPENSSL_malloc
8140 Malloc_locked => OPENSSL_malloc_locked
8141 Realloc => OPENSSL_realloc
8142 Free => OPENSSL_free
8145 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8146 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8149 *) CygWin32 support.
8150 [John Jarvie <jjarvie@newsguy.com>]
8152 *) The type-safe stack code has been rejigged. It is now only compiled
8153 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8154 by default all type-specific stack functions are "#define"d back to
8155 standard stack functions. This results in more streamlined output
8156 but retains the type-safety checking possibilities of the original
8160 *) The STACK code has been cleaned up, and certain type declarations
8161 that didn't make a lot of sense have been brought in line. This has
8162 also involved a cleanup of sorts in safestack.h to more correctly
8163 map type-safe stack functions onto their plain stack counterparts.
8164 This work has also resulted in a variety of "const"ifications of
8165 lots of the code, especially "_cmp" operations which should normally
8166 be prototyped with "const" parameters anyway.
8169 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8170 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8171 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8172 where all of 'md' is used each time the PRNG is used, but 'state'
8173 is used only indexed by a cyclic counter. As entropy may not be
8174 well distributed from the beginning, 'md' is important as a
8175 chaining variable. However, the output function chains only half
8176 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8177 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8178 in all of 'state' being rewritten, with the new values depending
8179 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8182 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8183 the handshake is continued after ssl_verify_cert_chain();
8184 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8185 can lead to 'unexplainable' connection aborts later.
8186 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8188 *) Major EVP API cipher revision.
8189 Add hooks for extra EVP features. This allows various cipher
8190 parameters to be set in the EVP interface. Support added for variable
8191 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8192 setting of RC2 and RC5 parameters.
8194 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8197 Remove lots of duplicated code from the EVP library. For example *every*
8198 cipher init() function handles the 'iv' in the same way according to the
8199 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8200 for CFB and OFB modes they zero ctx->num.
8202 New functionality allows removal of S/MIME code RC2 hack.
8204 Most of the routines have the same form and so can be declared in terms
8207 By shifting this to the top level EVP_CipherInit() it can be removed from
8208 all individual ciphers. If the cipher wants to handle IVs or keys
8209 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8212 Change lots of functions like EVP_EncryptUpdate() to now return a
8213 value: although software versions of the algorithms cannot fail
8214 any installed hardware versions can.
8217 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8218 this option is set, tolerate broken clients that send the negotiated
8219 protocol version number instead of the requested protocol version
8223 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8224 i.e. non-zero for export ciphersuites, zero otherwise.
8225 Previous versions had this flag inverted, inconsistent with
8226 rsa_tmp_cb (..._TMP_RSA_CB).
8227 [Bodo Moeller; problem reported by Amit Chopra]
8229 *) Add missing DSA library text string. Work around for some IIS
8230 key files with invalid SEQUENCE encoding.
8233 *) Add a document (doc/standards.txt) that list all kinds of standards
8234 and so on that are implemented in OpenSSL.
8237 *) Enhance c_rehash script. Old version would mishandle certificates
8238 with the same subject name hash and wouldn't handle CRLs at all.
8239 Added -fingerprint option to crl utility, to support new c_rehash
8243 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8246 *) Fix for SSL server purpose checking. Server checking was
8247 rejecting certificates which had extended key usage present
8248 but no ssl client purpose.
8249 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8251 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8252 is a little unclear about how a blank password is handled.
8253 Since the password in encoded as a BMPString with terminating
8254 double NULL a zero length password would end up as just the
8255 double NULL. However no password at all is different and is
8256 handled differently in the PKCS#12 key generation code. NS
8257 treats a blank password as zero length. MSIE treats it as no
8258 password on export: but it will try both on import. We now do
8259 the same: PKCS12_parse() tries zero length and no password if
8260 the password is set to "" or NULL (NULL is now a valid password:
8261 it wasn't before) as does the pkcs12 application.
8264 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8265 perror when PEM_read_bio_X509_REQ fails, the error message must
8266 be obtained from the error queue.
8269 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8270 it in ERR_remove_state if appropriate, and change ERR_get_state
8271 accordingly to avoid race conditions (this is necessary because
8272 thread_hash is no longer constant once set).
8275 *) Bugfix for linux-elf makefile.one.
8278 *) RSA_get_default_method() will now cause a default
8279 RSA_METHOD to be chosen if one doesn't exist already.
8280 Previously this was only set during a call to RSA_new()
8281 or RSA_new_method(NULL) meaning it was possible for
8282 RSA_get_default_method() to return NULL.
8285 *) Added native name translation to the existing DSO code
8286 that will convert (if the flag to do so is set) filenames
8287 that are sufficiently small and have no path information
8288 into a canonical native form. Eg. "blah" converted to
8289 "libblah.so" or "blah.dll" etc.
8292 *) New function ERR_error_string_n(e, buf, len) which is like
8293 ERR_error_string(e, buf), but writes at most 'len' bytes
8294 including the 0 terminator. For ERR_error_string_n, 'buf'
8296 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8298 *) CONF library reworked to become more general. A new CONF
8299 configuration file reader "class" is implemented as well as a
8300 new functions (NCONF_*, for "New CONF") to handle it. The now
8301 old CONF_* functions are still there, but are reimplemented to
8302 work in terms of the new functions. Also, a set of functions
8303 to handle the internal storage of the configuration data is
8304 provided to make it easier to write new configuration file
8305 reader "classes" (I can definitely see something reading a
8306 configuration file in XML format, for example), called _CONF_*,
8307 or "the configuration storage API"...
8309 The new configuration file reading functions are:
8311 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8312 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8314 NCONF_default, NCONF_WIN32
8316 NCONF_dump_fp, NCONF_dump_bio
8318 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8319 NCONF_new creates a new CONF object. This works in the same way
8320 as other interfaces in OpenSSL, like the BIO interface.
8321 NCONF_dump_* dump the internal storage of the configuration file,
8322 which is useful for debugging. All other functions take the same
8323 arguments as the old CONF_* functions wth the exception of the
8324 first that must be a `CONF *' instead of a `LHASH *'.
8326 To make it easer to use the new classes with the old CONF_* functions,
8327 the function CONF_set_default_method is provided.
8330 *) Add '-tls1' option to 'openssl ciphers', which was already
8331 mentioned in the documentation but had not been implemented.
8332 (This option is not yet really useful because even the additional
8333 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8336 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8337 OpenSSL-based applications) load shared libraries and bind to
8338 them in a portable way.
8339 [Geoff Thorpe, with contributions from Richard Levitte]
8341 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8343 *) Make sure _lrotl and _lrotr are only used with MSVC.
8345 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8346 (the default implementation of RAND_status).
8348 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8349 to '-clrext' (= clear extensions), as intended and documented.
8350 [Bodo Moeller; inconsistency pointed out by Michael Attili
8353 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8354 was larger than the MD block size.
8355 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8357 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8358 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8359 using the passed key: if the passed key was a private key the result
8360 of X509_print(), for example, would be to print out all the private key
8364 *) des_quad_cksum() byte order bug fix.
8365 [Ulf Möller, using the problem description in krb4-0.9.7, where
8366 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8368 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8370 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8372 *) For easily testing in shell scripts whether some command
8373 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8374 returns with exit code 0 iff no command of the given name is available.
8375 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8376 the output goes to stdout and nothing is printed to stderr.
8377 Additional arguments are always ignored.
8379 Since for each cipher there is a command of the same name,
8380 the 'no-cipher' compilation switches can be tested this way.
8382 ('openssl no-XXX' is not able to detect pseudo-commands such
8383 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8386 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8389 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8390 is set; it will be thrown away anyway because each handshake creates
8392 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8393 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8394 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8395 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8398 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8399 'Q' and 'R' lose their special meanings (quit/renegotiate).
8400 This is part of what -quiet does; unlike -quiet, -ign_eof
8401 does not suppress any output.
8404 *) Add compatibility options to the purpose and trust code. The
8405 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8406 accepts a certificate or CA, this was the previous behaviour,
8407 with all the associated security issues.
8409 X509_TRUST_COMPAT is the old trust behaviour: only and
8410 automatically trust self signed roots in certificate store. A
8411 new trust setting X509_TRUST_DEFAULT is used to specify that
8412 a purpose has no associated trust setting and it should instead
8413 use the value in the default purpose.
8416 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8417 and fix a memory leak.
8420 *) In util/mkerr.pl (which implements 'make errors'), preserve
8421 reason strings from the previous version of the .c file, as
8422 the default to have only downcase letters (and digits) in
8423 automatically generated reasons codes is not always appropriate.
8426 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8427 using strerror. Previously, ERR_reason_error_string() returned
8428 library names as reason strings for SYSerr; but SYSerr is a special
8429 case where small numbers are errno values, not library numbers.
8432 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8433 converts DSA parameters into DH parameters. (When creating parameters,
8434 DSA_generate_parameters is used.)
8437 *) Include 'length' (recommended exponent length) in C code generated
8438 by 'openssl dhparam -C'.
8441 *) The second argument to set_label in perlasm was already being used
8442 so couldn't be used as a "file scope" flag. Moved to third argument
8446 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8447 instead of RAND_bytes for encryption IVs and salts.
8450 *) Include RAND_status() into RAND_METHOD instead of implementing
8451 it only for md_rand.c Otherwise replacing the PRNG by calling
8452 RAND_set_rand_method would be impossible.
8455 *) Don't let DSA_generate_key() enter an infinite loop if the random
8456 number generation fails.
8459 *) New 'rand' application for creating pseudo-random output.
8462 *) Added configuration support for Linux/IA64
8463 [Rolf Haberrecker <rolf@suse.de>]
8465 *) Assembler module support for Mingw32.
8468 *) Shared library support for HPUX (in shlib/).
8469 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8471 *) Shared library support for Solaris gcc.
8472 [Lutz Behnke <behnke@trustcenter.de>]
8474 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8476 *) PKCS7_encrypt() was adding text MIME headers twice because they
8477 were added manually and by SMIME_crlf_copy().
8480 *) In bntest.c don't call BN_rand with zero bits argument.
8481 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8483 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8484 case was implemented. This caused BN_div_recp() to fail occasionally.
8487 *) Add an optional second argument to the set_label() in the perl
8488 assembly language builder. If this argument exists and is set
8489 to 1 it signals that the assembler should use a symbol whose
8490 scope is the entire file, not just the current function. This
8491 is needed with MASM which uses the format label:: for this scope.
8492 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8494 *) Change the ASN1 types so they are typedefs by default. Before
8495 almost all types were #define'd to ASN1_STRING which was causing
8496 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8500 *) Change names of new functions to the new get1/get0 naming
8501 convention: After 'get1', the caller owns a reference count
8502 and has to call ..._free; 'get0' returns a pointer to some
8503 data structure without incrementing reference counters.
8504 (Some of the existing 'get' functions increment a reference
8505 counter, some don't.)
8506 Similarly, 'set1' and 'add1' functions increase reference
8507 counters or duplicate objects.
8510 *) Allow for the possibility of temp RSA key generation failure:
8511 the code used to assume it always worked and crashed on failure.
8514 *) Fix potential buffer overrun problem in BIO_printf().
8515 [Ulf Möller, using public domain code by Patrick Powell; problem
8516 pointed out by David Sacerdote <das33@cornell.edu>]
8518 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8519 RAND_egd() and RAND_status(). In the command line application,
8520 the EGD socket can be specified like a seed file using RANDFILE
8524 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8525 Some CAs (e.g. Verisign) distribute certificates in this form.
8528 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8529 list to exclude them. This means that no special compilation option
8530 is needed to use anonymous DH: it just needs to be included in the
8534 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8535 EVP_MD_type. The old functionality is available in a new macro called
8536 EVP_MD_md(). Change code that uses it and update docs.
8539 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8540 where the 'void *' argument is replaced by a function pointer argument.
8541 Previously 'void *' was abused to point to functions, which works on
8542 many platforms, but is not correct. As these functions are usually
8543 called by macros defined in OpenSSL header files, most source code
8544 should work without changes.
8547 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8548 sections with information on -D... compiler switches used for
8549 compiling the library so that applications can see them. To enable
8550 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8551 must be defined. E.g.,
8552 #define OPENSSL_ALGORITHM_DEFINES
8553 #include <openssl/opensslconf.h>
8554 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8555 [Richard Levitte, Ulf and Bodo Möller]
8557 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8561 *) Change the 'other' type in certificate aux info to a STACK_OF
8562 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8563 the required ASN1 format: arbitrary types determined by an OID.
8566 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8567 argument to 'req'. This is not because the function is newer or
8568 better than others it just uses the work 'NEW' in the certificate
8569 request header lines. Some software needs this.
8572 *) Reorganise password command line arguments: now passwords can be
8573 obtained from various sources. Delete the PEM_cb function and make
8574 it the default behaviour: i.e. if the callback is NULL and the
8575 usrdata argument is not NULL interpret it as a null terminated pass
8576 phrase. If usrdata and the callback are NULL then the pass phrase
8577 is prompted for as usual.
8580 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8581 the support is automatically enabled. The resulting binaries will
8582 autodetect the card and use it if present.
8583 [Ben Laurie and Compaq Inc.]
8585 *) Work around for Netscape hang bug. This sends certificate request
8586 and server done in one record. Since this is perfectly legal in the
8587 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8588 the bugs/SSLv3 entry for more info.
8591 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8594 *) Add -rand argument to smime and pkcs12 applications and read/write
8598 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8601 *) Add command line password options to the remaining applications.
8604 *) Bug fix for BN_div_recp() for numerators with an even number of
8608 *) More tests in bntest.c, and changed test_bn output.
8611 *) ./config recognizes MacOS X now.
8614 *) Bug fix for BN_div() when the first words of num and divsor are
8615 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8618 *) Add support for various broken PKCS#8 formats, and command line
8619 options to produce them.
8622 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8623 get temporary BIGNUMs from a BN_CTX.
8626 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8630 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8631 include a #define from the old name to the new. The original intent
8632 was that statically linked binaries could for example just call
8633 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8634 link with digests. This never worked because SSLeay_add_all_digests()
8635 and SSLeay_add_all_ciphers() were in the same source file so calling
8636 one would link with the other. They are now in separate source files.
8639 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8642 *) Use a less unusual form of the Miller-Rabin primality test (it used
8643 a binary algorithm for exponentiation integrated into the Miller-Rabin
8644 loop, our standard modexp algorithms are faster).
8647 *) Support for the EBCDIC character set completed.
8648 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8650 *) Source code cleanups: use const where appropriate, eliminate casts,
8651 use void * instead of char * in lhash.
8654 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8655 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8656 this the server could overwrite ephemeral keys that the client
8660 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8661 using 50 iterations of the Rabin-Miller test.
8663 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8664 iterations of the Rabin-Miller test as required by the appendix
8665 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8666 As BN_is_prime_fasttest includes trial division, DSA parameter
8667 generation becomes much faster.
8669 This implies a change for the callback functions in DSA_is_prime
8670 and DSA_generate_parameters: The callback function is called once
8671 for each positive witness in the Rabin-Miller test, not just
8672 occasionally in the inner loop; and the parameters to the
8673 callback function now provide an iteration count for the outer
8674 loop rather than for the current invocation of the inner loop.
8675 DSA_generate_parameters additionally can call the callback
8676 function with an 'iteration count' of -1, meaning that a
8677 candidate has passed the trial division test (when q is generated
8678 from an application-provided seed, trial division is skipped).
8681 *) New function BN_is_prime_fasttest that optionally does trial
8682 division before starting the Rabin-Miller test and has
8683 an additional BN_CTX * argument (whereas BN_is_prime always
8684 has to allocate at least one BN_CTX).
8685 'callback(1, -1, cb_arg)' is called when a number has passed the
8686 trial division stage.
8689 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8693 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8696 *) New function BN_pseudo_rand().
8699 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8700 bignum version of BN_from_montgomery() with the working code from
8701 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8705 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8706 made it impossible to use the same SSL_SESSION data structure in
8707 SSL2 clients in multiple threads.
8710 *) The return value of RAND_load_file() no longer counts bytes obtained
8711 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8712 to seed the PRNG (previously an explicit byte count was required).
8713 [Ulf Möller, Bodo Möller]
8715 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8716 used (char *) instead of (void *) and had casts all over the place.
8719 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8722 *) Retain source code compatibility for BN_prime_checks macro:
8723 BN_is_prime(..., BN_prime_checks, ...) now uses
8724 BN_prime_checks_for_size to determine the appropriate number of
8725 Rabin-Miller iterations.
8728 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8729 DH_CHECK_P_NOT_SAFE_PRIME.
8730 (Check if this is true? OpenPGP calls them "strong".)
8733 *) Merge the functionality of "dh" and "gendh" programs into a new program
8734 "dhparam". The old programs are retained for now but will handle DH keys
8735 (instead of parameters) in future.
8738 *) Make the ciphers, s_server and s_client programs check the return values
8739 when a new cipher list is set.
8742 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8743 ciphers. Before when the 56bit ciphers were enabled the sorting was
8746 The syntax for the cipher sorting has been extended to support sorting by
8747 cipher-strength (using the strength_bits hard coded in the tables).
8748 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8750 Fix a bug in the cipher-command parser: when supplying a cipher command
8751 string with an "undefined" symbol (neither command nor alphanumeric
8752 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8753 an error is flagged.
8755 Due to the strength-sorting extension, the code of the
8756 ssl_create_cipher_list() function was completely rearranged. I hope that
8757 the readability was also increased :-)
8758 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8760 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8761 for the first serial number and places 2 in the serial number file. This
8762 avoids problems when the root CA is created with serial number zero and
8763 the first user certificate has the same issuer name and serial number
8767 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8768 the new code. Add documentation for this stuff.
8771 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8772 X509_*() to X509at_*() on the grounds that they don't handle X509
8773 structures and behave in an analogous way to the X509v3 functions:
8774 they shouldn't be called directly but wrapper functions should be used
8777 So we also now have some wrapper functions that call the X509at functions
8778 when passed certificate requests. (TO DO: similar things can be done with
8779 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8780 things. Some of these need some d2i or i2d and print functionality
8781 because they handle more complex structures.)
8784 *) Add missing #ifndefs that caused missing symbols when building libssl
8785 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8786 NO_RSA in ssl/s2*.c.
8787 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8789 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8790 has a return value which indicates the quality of the random data
8791 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8792 error queue. New function RAND_pseudo_bytes() generates output that is
8793 guaranteed to be unique but not unpredictable. RAND_add is like
8794 RAND_seed, but takes an extra argument for an entropy estimate
8795 (RAND_seed always assumes full entropy).
8798 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8799 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8800 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8801 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8802 false-positive rate of at most 2^-80 for random input.
8805 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8808 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8809 in the 0.9.5 release), this returns the chain
8810 from an X509_CTX structure with a dup of the stack and all
8811 the X509 reference counts upped: so the stack will exist
8812 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8815 Also make SSL_SESSION_print() print out the verify return
8819 *) Add manpage for the pkcs12 command. Also change the default
8820 behaviour so MAC iteration counts are used unless the new
8821 -nomaciter option is used. This improves file security and
8822 only older versions of MSIE (4.0 for example) need it.
8825 *) Honor the no-xxx Configure options when creating .DEF files.
8828 *) Add PKCS#10 attributes to field table: challengePassword,
8829 unstructuredName and unstructuredAddress. These are taken from
8830 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8831 international characters are used.
8833 More changes to X509_ATTRIBUTE code: allow the setting of types
8834 based on strings. Remove the 'loc' parameter when adding
8835 attributes because these will be a SET OF encoding which is sorted
8839 *) Initial changes to the 'req' utility to allow request generation
8840 automation. This will allow an application to just generate a template
8841 file containing all the field values and have req construct the
8844 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8845 used all over the place including certificate requests and PKCS#7
8846 structures. They are currently handled manually where necessary with
8847 some primitive wrappers for PKCS#7. The new functions behave in a
8848 manner analogous to the X509 extension functions: they allow
8849 attributes to be looked up by NID and added.
8851 Later something similar to the X509V3 code would be desirable to
8852 automatically handle the encoding, decoding and printing of the
8853 more complex types. The string types like challengePassword can
8854 be handled by the string table functions.
8856 Also modified the multi byte string table handling. Now there is
8857 a 'global mask' which masks out certain types. The table itself
8858 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8859 is useful when for example there is only one permissible type
8860 (as in countryName) and using the mask might result in no valid
8864 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8865 SSL_get_peer_finished to allow applications to obtain the latest
8866 Finished messages sent to the peer or expected from the peer,
8867 respectively. (SSL_get_peer_finished is usually the Finished message
8868 actually received from the peer, otherwise the protocol will be aborted.)
8870 As the Finished message are message digests of the complete handshake
8871 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8872 be used for external authentication procedures when the authentication
8873 provided by SSL/TLS is not desired or is not enough.
8876 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8877 the host supports BWX extension and if Compaq C is present on the
8878 $PATH. Just exploiting of the BWX extension results in 20-30%
8879 performance kick for some algorithms, e.g. DES and RC4 to mention
8880 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8884 *) Add support for MS "fast SGC". This is arguably a violation of the
8885 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8886 weak crypto and after checking the certificate is SGC a second one
8887 with strong crypto. MS SGC stops the first handshake after receiving
8888 the server certificate message and sends a second client hello. Since
8889 a server will typically do all the time consuming operations before
8890 expecting any further messages from the client (server key exchange
8891 is the most expensive) there is little difference between the two.
8893 To get OpenSSL to support MS SGC we have to permit a second client
8894 hello message after we have sent server done. In addition we have to
8895 reset the MAC if we do get this second client hello.
8898 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8899 if a DER encoded private key is RSA or DSA traditional format. Changed
8900 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8901 format DER encoded private key. Newer code should use PKCS#8 format which
8902 has the key type encoded in the ASN1 structure. Added DER private key
8903 support to pkcs8 application.
8906 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8907 ciphersuites has been selected (as required by the SSL 3/TLS 1
8908 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8909 is set, we interpret this as a request to violate the specification
8910 (the worst that can happen is a handshake failure, and 'correct'
8911 behaviour would result in a handshake failure anyway).
8914 *) In SSL_CTX_add_session, take into account that there might be multiple
8915 SSL_SESSION structures with the same session ID (e.g. when two threads
8916 concurrently obtain them from an external cache).
8917 The internal cache can handle only one SSL_SESSION with a given ID,
8918 so if there's a conflict, we now throw out the old one to achieve
8922 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8923 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8924 some routines that use cipher OIDs: some ciphers do not have OIDs
8925 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8929 *) Simplify the trust setting structure and code. Now we just have
8930 two sequences of OIDs for trusted and rejected settings. These will
8931 typically have values the same as the extended key usage extension
8932 and any application specific purposes.
8934 The trust checking code now has a default behaviour: it will just
8935 check for an object with the same NID as the passed id. Functions can
8936 be provided to override either the default behaviour or the behaviour
8937 for a given id. SSL client, server and email already have functions
8938 in place for compatibility: they check the NID and also return "trusted"
8939 if the certificate is self signed.
8942 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8943 traditional format into an EVP_PKEY structure.
8946 *) Add a password callback function PEM_cb() which either prompts for
8947 a password if usr_data is NULL or otherwise assumes it is a null
8948 terminated password. Allow passwords to be passed on command line
8949 environment or config files in a few more utilities.
8952 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8953 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8954 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8955 Update documentation.
8958 *) Support for ASN1 "NULL" type. This could be handled before by using
8959 ASN1_TYPE but there wasn't any function that would try to read a NULL
8960 and produce an error if it couldn't. For compatibility we also have
8961 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8962 don't allocate anything because they don't need to.
8965 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8967 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8969 *) Rebuild of the memory allocation routines used by OpenSSL code and
8970 possibly others as well. The purpose is to make an interface that
8971 provide hooks so anyone can build a separate set of allocation and
8972 deallocation routines to be used by OpenSSL, for example memory
8973 pool implementations, or something else, which was previously hard
8974 since Malloc(), Realloc() and Free() were defined as macros having
8975 the values malloc, realloc and free, respectively (except for Win32
8976 compilations). The same is provided for memory debugging code.
8977 OpenSSL already comes with functionality to find memory leaks, but
8978 this gives people a chance to debug other memory problems.
8980 With these changes, a new set of functions and macros have appeared:
8982 CRYPTO_set_mem_debug_functions() [F]
8983 CRYPTO_get_mem_debug_functions() [F]
8984 CRYPTO_dbg_set_options() [F]
8985 CRYPTO_dbg_get_options() [F]
8986 CRYPTO_malloc_debug_init() [M]
8988 The memory debug functions are NULL by default, unless the library
8989 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8990 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8991 gives the standard debugging functions that come with OpenSSL) or
8992 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8993 provided by the library user) must be used. When the standard
8994 debugging functions are used, CRYPTO_dbg_set_options can be used to
8995 request additional information:
8996 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8997 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8999 Also, things like CRYPTO_set_mem_functions will always give the
9000 expected result (the new set of functions is used for allocation
9001 and deallocation) at all times, regardless of platform and compiler
9004 To finish it up, some functions that were never use in any other
9005 way than through macros have a new API and new semantic:
9008 CRYPTO_dbg_realloc()
9011 All macros of value have retained their old syntax.
9012 [Richard Levitte and Bodo Moeller]
9014 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
9015 ordering of SMIMECapabilities wasn't in "strength order" and there
9016 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
9020 *) Some ASN1 types with illegal zero length encoding (INTEGER,
9021 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
9022 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
9024 *) Merge in my S/MIME library for OpenSSL. This provides a simple
9025 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
9026 functionality to handle multipart/signed properly) and a utility
9027 called 'smime' to call all this stuff. This is based on code I
9028 originally wrote for Celo who have kindly allowed it to be
9029 included in OpenSSL.
9032 *) Add variants des_set_key_checked and des_set_key_unchecked of
9033 des_set_key (aka des_key_sched). Global variable des_check_key
9034 decides which of these is called by des_set_key; this way
9035 des_check_key behaves as it always did, but applications and
9036 the library itself, which was buggy for des_check_key == 1,
9037 have a cleaner way to pick the version they need.
9040 *) New function PKCS12_newpass() which changes the password of a
9044 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
9045 dynamic mix. In both cases the ids can be used as an index into the
9046 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
9047 functions so they accept a list of the field values and the
9048 application doesn't need to directly manipulate the X509_TRUST
9052 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
9056 *) Modify the way the V3 extension code looks up extensions. This now
9057 works in a similar way to the object code: we have some "standard"
9058 extensions in a static table which is searched with OBJ_bsearch()
9059 and the application can add dynamic ones if needed. The file
9060 crypto/x509v3/ext_dat.h now has the info: this file needs to be
9061 updated whenever a new extension is added to the core code and kept
9062 in ext_nid order. There is a simple program 'tabtest.c' which checks
9063 this. New extensions are not added too often so this file can readily
9064 be maintained manually.
9066 There are two big advantages in doing things this way. The extensions
9067 can be looked up immediately and no longer need to be "added" using
9068 X509V3_add_standard_extensions(): this function now does nothing.
9069 [Side note: I get *lots* of email saying the extension code doesn't
9070 work because people forget to call this function]
9071 Also no dynamic allocation is done unless new extensions are added:
9072 so if we don't add custom extensions there is no need to call
9073 X509V3_EXT_cleanup().
9076 *) Modify enc utility's salting as follows: make salting the default. Add a
9077 magic header, so unsalted files fail gracefully instead of just decrypting
9078 to garbage. This is because not salting is a big security hole, so people
9079 should be discouraged from doing it.
9082 *) Fixes and enhancements to the 'x509' utility. It allowed a message
9083 digest to be passed on the command line but it only used this
9084 parameter when signing a certificate. Modified so all relevant
9085 operations are affected by the digest parameter including the
9086 -fingerprint and -x509toreq options. Also -x509toreq choked if a
9087 DSA key was used because it didn't fix the digest.
9090 *) Initial certificate chain verify code. Currently tests the untrusted
9091 certificates for consistency with the verify purpose (which is set
9092 when the X509_STORE_CTX structure is set up) and checks the pathlength.
9094 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
9095 this is because it will reject chains with invalid extensions whereas
9096 every previous version of OpenSSL and SSLeay made no checks at all.
9098 Trust code: checks the root CA for the relevant trust settings. Trust
9099 settings have an initial value consistent with the verify purpose: e.g.
9100 if the verify purpose is for SSL client use it expects the CA to be
9101 trusted for SSL client use. However the default value can be changed to
9102 permit custom trust settings: one example of this would be to only trust
9103 certificates from a specific "secure" set of CAs.
9105 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
9106 which should be used for version portability: especially since the
9107 verify structure is likely to change more often now.
9109 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
9110 to set them. If not set then assume SSL clients will verify SSL servers
9113 Two new options to the verify program: -untrusted allows a set of
9114 untrusted certificates to be passed in and -purpose which sets the
9115 intended purpose of the certificate. If a purpose is set then the
9116 new chain verify code is used to check extension consistency.
9119 *) Support for the authority information access extension.
9122 *) Modify RSA and DSA PEM read routines to transparently handle
9123 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
9124 public keys in a format compatible with certificate
9125 SubjectPublicKeyInfo structures. Unfortunately there were already
9126 functions called *_PublicKey_* which used various odd formats so
9127 these are retained for compatibility: however the DSA variants were
9128 never in a public release so they have been deleted. Changed dsa/rsa
9129 utilities to handle the new format: note no releases ever handled public
9130 keys so we should be OK.
9132 The primary motivation for this change is to avoid the same fiasco
9133 that dogs private keys: there are several incompatible private key
9134 formats some of which are standard and some OpenSSL specific and
9135 require various evil hacks to allow partial transparent handling and
9136 even then it doesn't work with DER formats. Given the option anything
9137 other than PKCS#8 should be dumped: but the other formats have to
9138 stay in the name of compatibility.
9140 With public keys and the benefit of hindsight one standard format
9141 is used which works with EVP_PKEY, RSA or DSA structures: though
9142 it clearly returns an error if you try to read the wrong kind of key.
9144 Added a -pubkey option to the 'x509' utility to output the public key.
9145 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9146 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9147 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9148 that do the same as the EVP_PKEY_assign_*() except they up the
9149 reference count of the added key (they don't "swallow" the
9153 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9154 CRLs would fail if the file contained no certificates or no CRLs:
9155 added a new function to read in both types and return the number
9156 read: this means that if none are read it will be an error. The
9157 DER versions of the certificate and CRL reader would always fail
9158 because it isn't possible to mix certificates and CRLs in DER format
9159 without choking one or the other routine. Changed this to just read
9160 a certificate: this is the best we can do. Also modified the code
9161 in apps/verify.c to take notice of return codes: it was previously
9162 attempting to read in certificates from NULL pointers and ignoring
9163 any errors: this is one reason why the cert and CRL reader seemed
9164 to work. It doesn't check return codes from the default certificate
9165 routines: these may well fail if the certificates aren't installed.
9168 *) Code to support otherName option in GeneralName.
9171 *) First update to verify code. Change the verify utility
9172 so it warns if it is passed a self signed certificate:
9173 for consistency with the normal behaviour. X509_verify
9174 has been modified to it will now verify a self signed
9175 certificate if *exactly* the same certificate appears
9176 in the store: it was previously impossible to trust a
9177 single self signed certificate. This means that:
9178 openssl verify ss.pem
9179 now gives a warning about a self signed certificate but
9180 openssl verify -CAfile ss.pem ss.pem
9184 *) For servers, store verify_result in SSL_SESSION data structure
9185 (and add it to external session representation).
9186 This is needed when client certificate verifications fails,
9187 but an application-provided verification callback (set by
9188 SSL_CTX_set_cert_verify_callback) allows accepting the session
9189 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9190 but returns 1): When the session is reused, we have to set
9191 ssl->verify_result to the appropriate error code to avoid
9193 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9195 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9196 case in PKCS7_dataInit() where the signed PKCS7 structure
9197 didn't contain any existing data because it was being created.
9198 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9200 *) Add a salt to the key derivation routines in enc.c. This
9201 forms the first 8 bytes of the encrypted file. Also add a
9202 -S option to allow a salt to be input on the command line.
9205 *) New function X509_cmp(). Oddly enough there wasn't a function
9206 to compare two certificates. We do this by working out the SHA1
9207 hash and comparing that. X509_cmp() will be needed by the trust
9211 *) SSL_get1_session() is like SSL_get_session(), but increments
9212 the reference count in the SSL_SESSION returned.
9213 [Geoff Thorpe <geoff@eu.c2.net>]
9215 *) Fix for 'req': it was adding a null to request attributes.
9216 Also change the X509_LOOKUP and X509_INFO code to handle
9217 certificate auxiliary information.
9220 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9224 *) Add the possibility to add extra information to the memory leak
9225 detecting output, to form tracebacks, showing from where each
9226 allocation was originated: CRYPTO_push_info("constant string") adds
9227 the string plus current file name and line number to a per-thread
9228 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9229 is like calling CYRPTO_pop_info() until the stack is empty.
9230 Also updated memory leak detection code to be multi-thread-safe.
9233 *) Add options -text and -noout to pkcs7 utility and delete the
9234 encryption options which never did anything. Update docs.
9237 *) Add options to some of the utilities to allow the pass phrase
9238 to be included on either the command line (not recommended on
9239 OSes like Unix) or read from the environment. Update the
9240 manpages and fix a few bugs.
9243 *) Add a few manpages for some of the openssl commands.
9246 *) Fix the -revoke option in ca. It was freeing up memory twice,
9247 leaking and not finding already revoked certificates.
9250 *) Extensive changes to support certificate auxiliary information.
9251 This involves the use of X509_CERT_AUX structure and X509_AUX
9252 functions. An X509_AUX function such as PEM_read_X509_AUX()
9253 can still read in a certificate file in the usual way but it
9254 will also read in any additional "auxiliary information". By
9255 doing things this way a fair degree of compatibility can be
9256 retained: existing certificates can have this information added
9257 using the new 'x509' options.
9259 Current auxiliary information includes an "alias" and some trust
9260 settings. The trust settings will ultimately be used in enhanced
9261 certificate chain verification routines: currently a certificate
9262 can only be trusted if it is self signed and then it is trusted
9266 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9267 The problem was that one of the replacement routines had not been working
9268 since SSLeay releases. For now the offending routine has been replaced
9269 with non-optimised assembler. Even so, this now gives around 95%
9270 performance improvement for 1024 bit RSA signs.
9273 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9274 handling. Most clients have the effective key size in bits equal to
9275 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9276 A few however don't do this and instead use the size of the decrypted key
9277 to determine the RC2 key length and the AlgorithmIdentifier to determine
9278 the effective key length. In this case the effective key length can still
9279 be 40 bits but the key length can be 168 bits for example. This is fixed
9280 by manually forcing an RC2 key into the EVP_PKEY structure because the
9281 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9282 the key length and effective key length are equal.
9285 *) Add a bunch of functions that should simplify the creation of
9286 X509_NAME structures. Now you should be able to do:
9287 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9288 and have it automatically work out the correct field type and fill in
9289 the structures. The more adventurous can try:
9290 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9291 and it will (hopefully) work out the correct multibyte encoding.
9294 *) Change the 'req' utility to use the new field handling and multibyte
9295 copy routines. Before the DN field creation was handled in an ad hoc
9296 way in req, ca, and x509 which was rather broken and didn't support
9297 BMPStrings or UTF8Strings. Since some software doesn't implement
9298 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9299 using the dirstring_type option. See the new comment in the default
9300 openssl.cnf for more info.
9303 *) Make crypto/rand/md_rand.c more robust:
9304 - Assure unique random numbers after fork().
9305 - Make sure that concurrent threads access the global counter and
9306 md serializably so that we never lose entropy in them
9307 or use exactly the same state in multiple threads.
9308 Access to the large state is not always serializable because
9309 the additional locking could be a performance killer, and
9310 md should be large enough anyway.
9313 *) New file apps/app_rand.c with commonly needed functionality
9314 for handling the random seed file.
9316 Use the random seed file in some applications that previously did not:
9318 dsaparam -genkey (which also ignored its '-rand' option),
9321 x509 (when signing).
9322 Except on systems with /dev/urandom, it is crucial to have a random
9323 seed file at least for key creation, DSA signing, and for DH exchanges;
9324 for RSA signatures we could do without one.
9326 gendh and gendsa (unlike genrsa) used to read only the first byte
9327 of each file listed in the '-rand' option. The function as previously
9328 found in genrsa is now in app_rand.c and is used by all programs
9329 that support '-rand'.
9332 *) In RAND_write_file, use mode 0600 for creating files;
9333 don't just chmod when it may be too late.
9336 *) Report an error from X509_STORE_load_locations
9337 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9340 *) New function ASN1_mbstring_copy() this copies a string in either
9341 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9342 into an ASN1_STRING type. A mask of permissible types is passed
9343 and it chooses the "minimal" type to use or an error if not type
9347 *) Add function equivalents to the various macros in asn1.h. The old
9348 macros are retained with an M_ prefix. Code inside the library can
9349 use the M_ macros. External code (including the openssl utility)
9350 should *NOT* in order to be "shared library friendly".
9353 *) Add various functions that can check a certificate's extensions
9354 to see if it usable for various purposes such as SSL client,
9355 server or S/MIME and CAs of these types. This is currently
9356 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9357 verification. Also added a -purpose flag to x509 utility to
9358 print out all the purposes.
9361 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9365 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9366 for, obtain and decode and extension and obtain its critical flag.
9367 This allows all the necessary extension code to be handled in a
9368 single function call.
9371 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9372 platforms. See crypto/rc4/rc4_enc.c for further details.
9375 *) New -noout option to asn1parse. This causes no output to be produced
9376 its main use is when combined with -strparse and -out to extract data
9377 from a file (which may not be in ASN.1 format).
9380 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9381 when producing the local key id.
9382 [Richard Levitte <levitte@stacken.kth.se>]
9384 *) New option -dhparam in s_server. This allows a DH parameter file to be
9385 stated explicitly. If it is not stated then it tries the first server
9386 certificate file. The previous behaviour hard coded the filename
9390 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9391 a public key to be input or output. For example:
9392 openssl rsa -in key.pem -pubout -out pubkey.pem
9393 Also added necessary DSA public key functions to handle this.
9396 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9397 in the message. This was handled by allowing
9398 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9399 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9401 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9402 to the end of the strings whereas this didn't. This would cause problems
9403 if strings read with d2i_ASN1_bytes() were later modified.
9404 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9406 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9407 data and it contains EOF it will end up returning an error. This is
9408 caused by input 46 bytes long. The cause is due to the way base64
9409 BIOs find the start of base64 encoded data. They do this by trying a
9410 trial decode on each line until they find one that works. When they
9411 do a flag is set and it starts again knowing it can pass all the
9412 data directly through the decoder. Unfortunately it doesn't reset
9413 the context it uses. This means that if EOF is reached an attempt
9414 is made to pass two EOFs through the context and this causes the
9415 resulting error. This can also cause other problems as well. As is
9416 usual with these problems it takes *ages* to find and the fix is
9417 trivial: move one line.
9418 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9420 *) Ugly workaround to get s_client and s_server working under Windows. The
9421 old code wouldn't work because it needed to select() on sockets and the
9422 tty (for keypresses and to see if data could be written). Win32 only
9423 supports select() on sockets so we select() with a 1s timeout on the
9424 sockets and then see if any characters are waiting to be read, if none
9425 are present then we retry, we also assume we can always write data to
9426 the tty. This isn't nice because the code then blocks until we've
9427 received a complete line of data and it is effectively polling the
9428 keyboard at 1s intervals: however it's quite a bit better than not
9429 working at all :-) A dedicated Windows application might handle this
9430 with an event loop for example.
9433 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9434 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9435 will be called when RSA_sign() and RSA_verify() are used. This is useful
9436 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9437 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9438 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9439 This necessitated the support of an extra signature type NID_md5_sha1
9440 for SSL signatures and modifications to the SSL library to use it instead
9441 of calling RSA_public_decrypt() and RSA_private_encrypt().
9444 *) Add new -verify -CAfile and -CApath options to the crl program, these
9445 will lookup a CRL issuers certificate and verify the signature in a
9446 similar way to the verify program. Tidy up the crl program so it
9447 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9448 less strict. It will now permit CRL extensions even if it is not
9449 a V2 CRL: this will allow it to tolerate some broken CRLs.
9452 *) Initialize all non-automatic variables each time one of the openssl
9453 sub-programs is started (this is necessary as they may be started
9454 multiple times from the "OpenSSL>" prompt).
9455 [Lennart Bang, Bodo Moeller]
9457 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9458 removing all other RSA functionality (this is what NO_RSA does). This
9459 is so (for example) those in the US can disable those operations covered
9460 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9464 *) Non-copying interface to BIO pairs.
9465 (still largely untested)
9468 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9469 ASCII string. This was handled independently in various places before.
9472 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9473 UTF8 strings a character at a time.
9476 *) Use client_version from client hello to select the protocol
9477 (s23_srvr.c) and for RSA client key exchange verification
9478 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9481 *) Add various utility functions to handle SPKACs, these were previously
9482 handled by poking round in the structure internals. Added new function
9483 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9484 print, verify and generate SPKACs. Based on an original idea from
9485 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9488 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9491 *) Allow the config file extension section to be overwritten on the
9492 command line. Based on an original idea from Massimiliano Pala
9493 <madwolf@comune.modena.it>. The new option is called -extensions
9494 and can be applied to ca, req and x509. Also -reqexts to override
9495 the request extensions in req and -crlexts to override the crl extensions
9499 *) Add new feature to the SPKAC handling in ca. Now you can include
9500 the same field multiple times by preceding it by "XXXX." for example:
9503 this is the same syntax as used in the req config file.
9506 *) Allow certificate extensions to be added to certificate requests. These
9507 are specified in a 'req_extensions' option of the req section of the
9508 config file. They can be printed out with the -text option to req but
9509 are otherwise ignored at present.
9512 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9513 data read consists of only the final block it would not decrypted because
9514 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9515 A misplaced 'break' also meant the decrypted final block might not be
9516 copied until the next read.
9519 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9520 a few extra parameters to the DH structure: these will be useful if
9521 for example we want the value of 'q' or implement X9.42 DH.
9524 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9525 provides hooks that allow the default DSA functions or functions on a
9526 "per key" basis to be replaced. This allows hardware acceleration and
9527 hardware key storage to be handled without major modification to the
9528 library. Also added low level modexp hooks and CRYPTO_EX structure and
9529 associated functions.
9532 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9533 as "read only": it can't be written to and the buffer it points to will
9534 not be freed. Reading from a read only BIO is much more efficient than
9535 a normal memory BIO. This was added because there are several times when
9536 an area of memory needs to be read from a BIO. The previous method was
9537 to create a memory BIO and write the data to it, this results in two
9538 copies of the data and an O(n^2) reading algorithm. There is a new
9539 function BIO_new_mem_buf() which creates a read only memory BIO from
9540 an area of memory. Also modified the PKCS#7 routines to use read only
9544 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9545 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9546 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9547 but a retry condition occured while trying to read the rest.
9550 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9551 NID_pkcs7_encrypted by default: this was wrong since this should almost
9552 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9553 the encrypted data type: this is a more sensible place to put it and it
9554 allows the PKCS#12 code to be tidied up that duplicated this
9558 *) Changed obj_dat.pl script so it takes its input and output files on
9559 the command line. This should avoid shell escape redirection problems
9563 *) Initial support for certificate extension requests, these are included
9564 in things like Xenroll certificate requests. Included functions to allow
9565 extensions to be obtained and added.
9568 *) -crlf option to s_client and s_server for sending newlines as
9569 CRLF (as required by many protocols).
9572 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9574 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9575 [Ralf S. Engelschall]
9577 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9578 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9580 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9584 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9585 DH parameters/keys (q is lost during that conversion, but the resulting
9586 DH parameters contain its length).
9588 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9589 much faster than DH_generate_parameters (which creates parameters
9590 where p = 2*q + 1), and also the smaller q makes DH computations
9591 much more efficient (160-bit exponentiation instead of 1024-bit
9592 exponentiation); so this provides a convenient way to support DHE
9593 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9594 utter importance to use
9595 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9597 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9598 when such DH parameters are used, because otherwise small subgroup
9599 attacks may become possible!
9602 *) Avoid memory leak in i2d_DHparams.
9605 *) Allow the -k option to be used more than once in the enc program:
9606 this allows the same encrypted message to be read by multiple recipients.
9609 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9610 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9611 it will always use the numerical form of the OID, even if it has a short
9615 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9616 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9617 otherwise bn_mod_exp was called. In the case of hardware keys for example
9618 no private key components need be present and it might store extra data
9619 in the RSA structure, which cannot be accessed from bn_mod_exp.
9620 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9621 private key operations.
9624 *) Added support for SPARC Linux.
9627 *) pem_password_cb function type incompatibly changed from
9628 typedef int pem_password_cb(char *buf, int size, int rwflag);
9630 ....(char *buf, int size, int rwflag, void *userdata);
9631 so that applications can pass data to their callbacks:
9632 The PEM[_ASN1]_{read,write}... functions and macros now take an
9633 additional void * argument, which is just handed through whenever
9634 the password callback is called.
9635 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9637 New function SSL_CTX_set_default_passwd_cb_userdata.
9639 Compatibility note: As many C implementations push function arguments
9640 onto the stack in reverse order, the new library version is likely to
9641 interoperate with programs that have been compiled with the old
9642 pem_password_cb definition (PEM_whatever takes some data that
9643 happens to be on the stack as its last argument, and the callback
9644 just ignores this garbage); but there is no guarantee whatsoever that
9647 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9648 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9649 problems not only on Windows, but also on some Unix platforms.
9650 To avoid problematic command lines, these definitions are now in an
9651 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9652 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9655 *) MIPS III/IV assembler module is reimplemented.
9658 *) More DES library cleanups: remove references to srand/rand and
9659 delete an unused file.
9662 *) Add support for the the free Netwide assembler (NASM) under Win32,
9663 since not many people have MASM (ml) and it can be hard to obtain.
9664 This is currently experimental but it seems to work OK and pass all
9665 the tests. Check out INSTALL.W32 for info.
9668 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9669 without temporary keys kept an extra copy of the server key,
9670 and connections with temporary keys did not free everything in case
9674 *) New function RSA_check_key and new openssl rsa option -check
9675 for verifying the consistency of RSA keys.
9676 [Ulf Moeller, Bodo Moeller]
9678 *) Various changes to make Win32 compile work:
9679 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9680 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9681 comparison" warnings.
9682 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9685 *) Add a debugging option to PKCS#5 v2 key generation function: when
9686 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9687 derived keys are printed to stderr.
9690 *) Copy the flags in ASN1_STRING_dup().
9691 [Roman E. Pavlov <pre@mo.msk.ru>]
9693 *) The x509 application mishandled signing requests containing DSA
9694 keys when the signing key was also DSA and the parameters didn't match.
9696 It was supposed to omit the parameters when they matched the signing key:
9697 the verifying software was then supposed to automatically use the CA's
9698 parameters if they were absent from the end user certificate.
9700 Omitting parameters is no longer recommended. The test was also
9701 the wrong way round! This was probably due to unusual behaviour in
9702 EVP_cmp_parameters() which returns 1 if the parameters match.
9703 This meant that parameters were omitted when they *didn't* match and
9704 the certificate was useless. Certificates signed with 'ca' didn't have
9706 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9708 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9709 The interface is as follows:
9710 Applications can use
9711 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9712 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9713 "off" is now the default.
9714 The library internally uses
9715 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9716 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9717 to disable memory-checking temporarily.
9719 Some inconsistent states that previously were possible (and were
9720 even the default) are now avoided.
9722 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9723 with each memory chunk allocated; this is occasionally more helpful
9724 than just having a counter.
9726 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9728 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9732 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9733 which largely parallels "options", but is for changing API behaviour,
9734 whereas "options" are about protocol behaviour.
9735 Initial "mode" flags are:
9737 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9738 a single record has been written.
9739 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9740 retries use the same buffer location.
9741 (But all of the contents must be
9745 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9748 *) Fix problems with no-hmac etc.
9749 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9751 *) New functions RSA_get_default_method(), RSA_set_method() and
9752 RSA_get_method(). These allows replacement of RSA_METHODs without having
9753 to mess around with the internals of an RSA structure.
9756 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9757 Also really enable memory leak checks in openssl.c and in some
9759 [Chad C. Mulligan, Bodo Moeller]
9761 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9762 up the length of negative integers. This has now been simplified to just
9763 store the length when it is first determined and use it later, rather
9764 than trying to keep track of where data is copied and updating it to
9766 [Steve Henson, reported by Brien Wheeler
9767 <bwheeler@authentica-security.com>]
9769 *) Add a new function PKCS7_signatureVerify. This allows the verification
9770 of a PKCS#7 signature but with the signing certificate passed to the
9771 function itself. This contrasts with PKCS7_dataVerify which assumes the
9772 certificate is present in the PKCS#7 structure. This isn't always the
9773 case: certificates can be omitted from a PKCS#7 structure and be
9774 distributed by "out of band" means (such as a certificate database).
9777 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9778 function prototypes in pem.h, also change util/mkdef.pl to add the
9779 necessary function names.
9782 *) mk1mf.pl (used by Windows builds) did not properly read the
9783 options set by Configure in the top level Makefile, and Configure
9784 was not even able to write more than one option correctly.
9785 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9788 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9789 file to be loaded from a BIO or FILE pointer. The BIO version will
9790 for example allow memory BIOs to contain config info.
9793 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9794 Whoever hopes to achieve shared-library compatibility across versions
9795 must use this, not the compile-time macro.
9796 (Exercise 0.9.4: Which is the minimum library version required by
9798 Note: All this applies only to multi-threaded programs, others don't
9802 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9803 through a BIO pair triggered the default case, i.e.
9804 SSLerr(...,SSL_R_UNKNOWN_STATE).
9807 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9808 can use the SSL library even if none of the specific BIOs is
9812 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9813 for the encoded length.
9814 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9816 *) Add initial documentation of the X509V3 functions.
9819 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9820 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9821 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9822 secure PKCS#8 private key format with a high iteration count.
9825 *) Fix determination of Perl interpreter: A perl or perl5
9826 _directory_ in $PATH was also accepted as the interpreter.
9827 [Ralf S. Engelschall]
9829 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9830 wrong with it but it was very old and did things like calling
9831 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9835 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9836 to use the new extension code.
9839 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9840 with macros. This should make it easier to change their form, add extra
9841 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9845 *) Add to configuration table a new entry that can specify an alternative
9846 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9847 according to Mark Crispin <MRC@Panda.COM>.
9851 *) DES CBC did not update the IV. Weird.
9854 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9855 Changing the behaviour of the former might break existing programs --
9856 where IV updating is needed, des_ncbc_encrypt can be used.
9859 *) When bntest is run from "make test" it drives bc to check its
9860 calculations, as well as internally checking them. If an internal check
9861 fails, it needs to cause bc to give a non-zero result or make test carries
9862 on without noticing the failure. Fixed.
9865 *) DES library cleanups.
9868 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9869 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9870 ciphers. NOTE: although the key derivation function has been verified
9871 against some published test vectors it has not been extensively tested
9872 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9876 *) Instead of "mkdir -p", which is not fully portable, use new
9877 Perl script "util/mkdir-p.pl".
9880 *) Rewrite the way password based encryption (PBE) is handled. It used to
9881 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9882 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9883 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9884 the 'parameter' field of the AlgorithmIdentifier is passed to the
9885 underlying key generation function so it must do its own ASN1 parsing.
9886 This has also changed the EVP_PBE_CipherInit() function which now has a
9887 'parameter' argument instead of literal salt and iteration count values
9888 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9891 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9892 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9893 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9894 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9895 value was just used as a "magic string" and not used directly its
9896 value doesn't matter.
9899 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9903 *) "linux-sparc64" configuration (ultrapenguin).
9904 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9905 "linux-sparc" configuration.
9906 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9908 *) config now generates no-xxx options for missing ciphers.
9911 *) Support the EBCDIC character set (work in progress).
9912 File ebcdic.c not yet included because it has a different license.
9913 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9915 *) Support BS2000/OSD-POSIX.
9916 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9918 *) Make callbacks for key generation use void * instead of char *.
9921 *) Make S/MIME samples compile (not yet tested).
9924 *) Additional typesafe stacks.
9927 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9931 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9933 *) New configuration variant "sco5-gcc".
9935 *) Updated some demos.
9936 [Sean O Riordain, Wade Scholine]
9938 *) Add missing BIO_free at exit of pkcs12 application.
9941 *) Fix memory leak in conf.c.
9944 *) Updates for Win32 to assembler version of MD5.
9947 *) Set #! path to perl in apps/der_chop to where we found it
9948 instead of using a fixed path.
9951 *) SHA library changes for irix64-mips4-cc.
9954 *) Improvements for VMS support.
9958 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9960 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9961 This also avoids the problems with SC4.2 and unpatched SC5.
9962 [Andy Polyakov <appro@fy.chalmers.se>]
9964 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9965 These are required because of the typesafe stack would otherwise break
9966 existing code. If old code used a structure member which used to be STACK
9967 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9968 sk_num or sk_value it would produce an error because the num, data members
9969 are not present in STACK_OF. Now it just produces a warning. sk_set
9970 replaces the old method of assigning a value to sk_value
9971 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9972 that does this will no longer work (and should use sk_set instead) but
9973 this could be regarded as a "questionable" behaviour anyway.
9976 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9977 correctly handle encrypted S/MIME data.
9980 *) Change type of various DES function arguments from des_cblock
9981 (which means, in function argument declarations, pointer to char)
9982 to des_cblock * (meaning pointer to array with 8 char elements),
9983 which allows the compiler to do more typechecking; it was like
9984 that back in SSLeay, but with lots of ugly casts.
9986 Introduce new type const_des_cblock.
9989 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9990 problems: find RecipientInfo structure that matches recipient certificate
9991 and initialise the ASN1 structures properly based on passed cipher.
9994 *) Belatedly make the BN tests actually check the results.
9997 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9998 to and from BNs: it was completely broken. New compilation option
9999 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
10000 key elements as negative integers.
10003 *) Reorganize and speed up MD5.
10004 [Andy Polyakov <appro@fy.chalmers.se>]
10007 [Richard Levitte <richard@levitte.org>]
10009 *) New option -out to asn1parse to allow the parsed structure to be
10010 output to a file. This is most useful when combined with the -strparse
10011 option to examine the output of things like OCTET STRINGS.
10014 *) Make SSL library a little more fool-proof by not requiring any longer
10015 that SSL_set_{accept,connect}_state be called before
10016 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
10017 in many applications because usually everything *appeared* to work as
10018 intended anyway -- now it really works as intended).
10021 *) Move openssl.cnf out of lib/.
10024 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
10025 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
10026 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
10027 [Ralf S. Engelschall]
10029 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
10030 handle PKCS#7 enveloped data properly.
10031 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
10033 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
10034 copying pointers. The cert_st handling is changed by this in
10035 various ways (and thus what used to be known as ctx->default_cert
10036 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
10037 any longer when s->cert does not give us what we need).
10038 ssl_cert_instantiate becomes obsolete by this change.
10039 As soon as we've got the new code right (possibly it already is?),
10040 we have solved a couple of bugs of the earlier code where s->cert
10041 was used as if it could not have been shared with other SSL structures.
10043 Note that using the SSL API in certain dirty ways now will result
10044 in different behaviour than observed with earlier library versions:
10045 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
10046 does not influence s as it used to.
10048 In order to clean up things more thoroughly, inside SSL_SESSION
10049 we don't use CERT any longer, but a new structure SESS_CERT
10050 that holds per-session data (if available); currently, this is
10051 the peer's certificate chain and, for clients, the server's certificate
10052 and temporary key. CERT holds only those values that can have
10053 meaningful defaults in an SSL_CTX.
10056 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
10057 from the internal representation. Various PKCS#7 fixes: remove some
10058 evil casts and set the enc_dig_alg field properly based on the signing
10062 *) Allow PKCS#12 password to be set from the command line or the
10063 environment. Let 'ca' get its config file name from the environment
10064 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
10068 *) Allow certificate policies extension to use an IA5STRING for the
10069 organization field. This is contrary to the PKIX definition but
10070 VeriSign uses it and IE5 only recognises this form. Document 'x509'
10074 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
10075 without disallowing inline assembler and the like for non-pedantic builds.
10078 *) Support Borland C++ builder.
10079 [Janez Jere <jj@void.si>, modified by Ulf Möller]
10081 *) Support Mingw32.
10084 *) SHA-1 cleanups and performance enhancements.
10085 [Andy Polyakov <appro@fy.chalmers.se>]
10087 *) Sparc v8plus assembler for the bignum library.
10088 [Andy Polyakov <appro@fy.chalmers.se>]
10090 *) Accept any -xxx and +xxx compiler options in Configure.
10093 *) Update HPUX configuration.
10096 *) Add missing sk_<type>_unshift() function to safestack.h
10097 [Ralf S. Engelschall]
10099 *) New function SSL_CTX_use_certificate_chain_file that sets the
10100 "extra_cert"s in addition to the certificate. (This makes sense
10101 only for "PEM" format files, as chains as a whole are not
10105 *) Support verify_depth from the SSL API.
10106 x509_vfy.c had what can be considered an off-by-one-error:
10107 Its depth (which was not part of the external interface)
10108 was actually counting the number of certificates in a chain;
10109 now it really counts the depth.
10112 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
10113 instead of X509err, which often resulted in confusing error
10114 messages since the error codes are not globally unique
10115 (e.g. an alleged error in ssl3_accept when a certificate
10116 didn't match the private key).
10118 *) New function SSL_CTX_set_session_id_context that allows to set a default
10119 value (so that you don't need SSL_set_session_id_context for each
10120 connection using the SSL_CTX).
10123 *) OAEP decoding bug fix.
10126 *) Support INSTALL_PREFIX for package builders, as proposed by
10130 *) New Configure options "threads" and "no-threads". For systems
10131 where the proper compiler options are known (currently Solaris
10132 and Linux), "threads" is the default.
10135 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10138 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10139 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10140 such as /usr/local/bin.
10143 *) "make linux-shared" to build shared libraries.
10144 [Niels Poppe <niels@netbox.org>]
10146 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10149 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10150 extension adding in x509 utility.
10153 *) Remove NOPROTO sections and error code comments.
10156 *) Partial rewrite of the DEF file generator to now parse the ANSI
10160 *) New Configure options --prefix=DIR and --openssldir=DIR.
10163 *) Complete rewrite of the error code script(s). It is all now handled
10164 by one script at the top level which handles error code gathering,
10165 header rewriting and C source file generation. It should be much better
10166 than the old method: it now uses a modified version of Ulf's parser to
10167 read the ANSI prototypes in all header files (thus the old K&R definitions
10168 aren't needed for error creation any more) and do a better job of
10169 translating function codes into names. The old 'ASN1 error code imbedded
10170 in a comment' is no longer necessary and it doesn't use .err files which
10171 have now been deleted. Also the error code call doesn't have to appear all
10172 on one line (which resulted in some large lines...).
10175 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10178 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10179 0 (which usually indicates a closed connection), but continue reading.
10182 *) Fix some race conditions.
10185 *) Add support for CRL distribution points extension. Add Certificate
10186 Policies and CRL distribution points documentation.
10189 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10192 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10193 8 of keying material. Merlin has also confirmed interop with this fix
10194 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10195 [Merlin Hughes <merlin@baltimore.ie>]
10197 *) Fix lots of warnings.
10198 [Richard Levitte <levitte@stacken.kth.se>]
10200 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10201 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10202 [Richard Levitte <levitte@stacken.kth.se>]
10204 *) Fix problems with sizeof(long) == 8.
10205 [Andy Polyakov <appro@fy.chalmers.se>]
10207 *) Change functions to ANSI C.
10210 *) Fix typos in error codes.
10211 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10213 *) Remove defunct assembler files from Configure.
10216 *) SPARC v8 assembler BIGNUM implementation.
10217 [Andy Polyakov <appro@fy.chalmers.se>]
10219 *) Support for Certificate Policies extension: both print and set.
10220 Various additions to support the r2i method this uses.
10223 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10224 return a const string when you are expecting an allocated buffer.
10227 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10228 types DirectoryString and DisplayText.
10231 *) Add code to allow r2i extensions to access the configuration database,
10232 add an LHASH database driver and add several ctx helper functions.
10235 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10236 fail when they extended the size of a BIGNUM.
10239 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10240 support typesafe stack.
10243 *) Fix typo in SSL_[gs]et_options().
10244 [Nils Frostberg <nils@medcom.se>]
10246 *) Delete various functions and files that belonged to the (now obsolete)
10247 old X509V3 handling code.
10250 *) New Configure option "rsaref".
10253 *) Don't auto-generate pem.h.
10256 *) Introduce type-safe ASN.1 SETs.
10259 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10260 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10262 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10263 that links with OpenSSL (well at least cause lots of warnings), but fear
10264 not: the conversion is trivial, and it eliminates loads of evil casts. A
10265 few STACKed things have been converted already. Feel free to convert more.
10266 In the fullness of time, I'll do away with the STACK type altogether.
10269 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10270 specified in <certfile> by updating the entry in the index.txt file.
10271 This way one no longer has to edit the index.txt file manually for
10272 revoking a certificate. The -revoke option does the gory details now.
10273 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10275 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10276 `-text' option at all and this way the `-noout -text' combination was
10277 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10278 [Ralf S. Engelschall]
10280 *) Make sure a corresponding plain text error message exists for the
10281 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10282 verify callback function determined that a certificate was revoked.
10283 [Ralf S. Engelschall]
10285 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10286 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10287 all available cipers including rc5, which was forgotten until now.
10288 In order to let the testing shell script know which algorithms
10289 are available, a new (up to now undocumented) command
10290 "openssl list-cipher-commands" is used.
10293 *) Bugfix: s_client occasionally would sleep in select() when
10294 it should have checked SSL_pending() first.
10297 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10298 the raw DSA values prior to ASN.1 encoding.
10301 *) Tweaks to Configure
10302 [Niels Poppe <niels@netbox.org>]
10304 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10308 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10311 *) New config option to avoid instructions that are illegal on the 80386.
10312 The default code is faster, but requires at least a 486.
10315 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10316 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10317 same as SSL2_VERSION anyway.
10320 *) New "-showcerts" option for s_client.
10323 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10324 application. Various cleanups and fixes.
10327 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10328 modify error routines to work internally. Add error codes and PBE init
10329 to library startup routines.
10332 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10333 packing functions to asn1 and evp. Changed function names and error
10334 codes along the way.
10337 *) PKCS12 integration: and so it begins... First of several patches to
10338 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10339 objects to objects.h
10342 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10343 and display support for Thawte strong extranet extension.
10346 *) Add LinuxPPC support.
10347 [Jeff Dubrule <igor@pobox.org>]
10349 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10350 bn_div_words in alpha.s.
10351 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10353 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10354 OAEP isn't supported when OpenSSL is built with RSAref.
10355 [Ulf Moeller <ulf@fitug.de>]
10357 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10358 so they no longer are missing under -DNOPROTO.
10359 [Soren S. Jorvang <soren@t.dk>]
10362 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10364 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10365 doesn't work when the session is reused. Coming soon!
10368 *) Fix a security hole, that allows sessions to be reused in the wrong
10369 context thus bypassing client cert protection! All software that uses
10370 client certs and session caches in multiple contexts NEEDS PATCHING to
10371 allow session reuse! A fuller solution is in the works.
10372 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10374 *) Some more source tree cleanups (removed obsolete files
10375 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10376 permission on "config" script to be executable) and a fix for the INSTALL
10378 [Ulf Moeller <ulf@fitug.de>]
10380 *) Remove some legacy and erroneous uses of malloc, free instead of
10382 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10384 *) Make rsa_oaep_test return non-zero on error.
10385 [Ulf Moeller <ulf@fitug.de>]
10387 *) Add support for native Solaris shared libraries. Configure
10388 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10389 if someone would make that last step automatic.
10390 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10392 *) ctx_size was not built with the right compiler during "make links". Fixed.
10395 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10396 except NULL ciphers". This means the default cipher list will no longer
10397 enable NULL ciphers. They need to be specifically enabled e.g. with
10398 the string "DEFAULT:eNULL".
10401 *) Fix to RSA private encryption routines: if p < q then it would
10402 occasionally produce an invalid result. This will only happen with
10403 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10406 *) Be less restrictive and allow also `perl util/perlpath.pl
10407 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10408 because this way one can also use an interpreter named `perl5' (which is
10409 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10410 installed as `perl').
10411 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10413 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10414 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10416 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10417 advapi32.lib to Win32 build and change the pem test comparison
10418 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10419 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10420 and crypto/des/ede_cbcm_enc.c.
10423 *) DES quad checksum was broken on big-endian architectures. Fixed.
10426 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10427 Win32 test batch file so it (might) work again. The Win32 test batch file
10428 is horrible: I feel ill....
10431 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10432 in e_os.h. Audit of header files to check ANSI and non ANSI
10433 sections: 10 functions were absent from non ANSI section and not exported
10434 from Windows DLLs. Fixed up libeay.num for new functions.
10437 *) Make `openssl version' output lines consistent.
10438 [Ralf S. Engelschall]
10440 *) Fix Win32 symbol export lists for BIO functions: Added
10441 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10442 to ms/libeay{16,32}.def.
10443 [Ralf S. Engelschall]
10445 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10446 fine under Unix and passes some trivial tests I've now added. But the
10447 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10448 added to make sure no one expects that this stuff really works in the
10449 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10450 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10452 [Ralf S. Engelschall]
10454 *) Fix the generation of two part addresses in perl.
10455 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10457 *) Add config entry for Linux on MIPS.
10458 [John Tobey <jtobey@channel1.com>]
10460 *) Make links whenever Configure is run, unless we are on Windoze.
10463 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10464 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10468 *) Add a useful kludge to allow package maintainers to specify compiler and
10469 other platforms details on the command line without having to patch the
10470 Configure script everytime: One now can use ``perl Configure
10471 <id>:<details>'', i.e. platform ids are allowed to have details appended
10472 to them (separated by colons). This is treated as there would be a static
10473 pre-configured entry in Configure's %table under key <id> with value
10474 <details> and ``perl Configure <id>'' is called. So, when you want to
10475 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10476 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10477 now, which overrides the FreeBSD-elf entry on-the-fly.
10478 [Ralf S. Engelschall]
10480 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10483 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10484 on the `perl Configure ...' command line. This way one can compile
10485 OpenSSL libraries with Position Independent Code (PIC) which is needed
10486 for linking it into DSOs.
10487 [Ralf S. Engelschall]
10489 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10493 *) Cleaned up the LICENSE document: The official contact for any license
10494 questions now is the OpenSSL core team under openssl-core@openssl.org.
10495 And add a paragraph about the dual-license situation to make sure people
10496 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10497 to the OpenSSL toolkit.
10498 [Ralf S. Engelschall]
10500 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10501 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10502 Additionally cleaned up the `make links' target: Remove unnecessary
10503 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10504 to speed processing and no longer clutter the display with confusing
10505 stuff. Instead only the actually done links are displayed.
10506 [Ralf S. Engelschall]
10508 *) Permit null encryption ciphersuites, used for authentication only. It used
10509 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10510 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10514 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10515 signed attributes when verifying signatures (this would break them),
10516 the detached data encoding was wrong and public keys obtained using
10517 X509_get_pubkey() weren't freed.
10520 *) Add text documentation for the BUFFER functions. Also added a work around
10521 to a Win95 console bug. This was triggered by the password read stuff: the
10522 last character typed gets carried over to the next fread(). If you were
10523 generating a new cert request using 'req' for example then the last
10524 character of the passphrase would be CR which would then enter the first
10528 *) Added the new `Includes OpenSSL Cryptography Software' button as
10529 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10530 button and can be used by applications based on OpenSSL to show the
10531 relationship to the OpenSSL project.
10532 [Ralf S. Engelschall]
10534 *) Remove confusing variables in function signatures in files
10535 ssl/ssl_lib.c and ssl/ssl.h.
10536 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10538 *) Don't install bss_file.c under PREFIX/include/
10539 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10541 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10542 functions that return function pointers and has support for NT specific
10543 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10544 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10545 unsigned to signed types: this was killing the Win32 compile.
10548 *) Add new certificate file to stack functions,
10549 SSL_add_dir_cert_subjects_to_stack() and
10550 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10551 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10552 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10553 This means that Apache-SSL and similar packages don't have to mess around
10554 to add as many CAs as they want to the preferred list.
10557 *) Experiment with doxygen documentation. Currently only partially applied to
10559 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10560 openssl.doxy as the configuration file.
10563 *) Get rid of remaining C++-style comments which strict C compilers hate.
10564 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10566 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10567 compiled in by default: it has problems with large keys.
10570 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10571 DH private keys and/or callback functions which directly correspond to
10572 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10573 is needed for applications which have to configure certificates on a
10574 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10576 For the RSA certificate situation is makes no difference, but
10577 for the DSA certificate situation this fixes the "no shared cipher"
10578 problem where the OpenSSL cipher selection procedure failed because the
10579 temporary keys were not overtaken from the context and the API provided
10580 no way to reconfigure them.
10581 The new functions now let applications reconfigure the stuff and they
10582 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10583 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10584 non-public-API function ssl_cert_instantiate() is used as a helper
10585 function and also to reduce code redundancy inside ssl_rsa.c.
10586 [Ralf S. Engelschall]
10588 *) Move s_server -dcert and -dkey options out of the undocumented feature
10589 area because they are useful for the DSA situation and should be
10590 recognized by the users.
10591 [Ralf S. Engelschall]
10593 *) Fix the cipher decision scheme for export ciphers: the export bits are
10594 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10595 SSL_EXP_MASK. So, the original variable has to be used instead of the
10596 already masked variable.
10597 [Richard Levitte <levitte@stacken.kth.se>]
10599 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10600 [Richard Levitte <levitte@stacken.kth.se>]
10602 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10603 from `int' to `unsigned int' because it's a length and initialized by
10604 EVP_DigestFinal() which expects an `unsigned int *'.
10605 [Richard Levitte <levitte@stacken.kth.se>]
10607 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10608 script. Instead use the usual Shell->Perl transition trick.
10609 [Ralf S. Engelschall]
10611 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10612 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10613 -noout -modulus' as it's already the case for `openssl rsa -noout
10614 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10615 currently the public key is printed (a decision which was already done by
10616 `openssl dsa -modulus' in the past) which serves a similar purpose.
10617 Additionally the NO_RSA no longer completely removes the whole -modulus
10618 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10620 [Ralf S. Engelschall]
10622 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10623 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10624 [Arne Ansper <arne@ats.cyber.ee>]
10626 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10627 to be added. Now both 'req' and 'ca' can use new objects defined in the
10631 *) Add cool BIO that does syslog (or event log on NT).
10632 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10634 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10635 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10636 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10637 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10640 *) Add preliminary config info for new extension code.
10643 *) Make RSA_NO_PADDING really use no padding.
10644 [Ulf Moeller <ulf@fitug.de>]
10646 *) Generate errors when private/public key check is done.
10649 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10650 for some CRL extensions and new objects added.
10653 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10654 key usage extension and fuller support for authority key id.
10657 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10658 padding method for RSA, which is recommended for new applications in PKCS
10659 #1 v2.0 (RFC 2437, October 1998).
10660 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10661 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10662 against Bleichbacher's attack on RSA.
10663 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10666 *) Updates to the new SSL compression code
10667 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10669 *) Fix so that the version number in the master secret, when passed
10670 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10671 (because the server will not accept higher), that the version number
10672 is 0x03,0x01, not 0x03,0x00
10673 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10675 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10676 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10677 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10680 *) Support for RAW extensions where an arbitrary extension can be
10681 created by including its DER encoding. See apps/openssl.cnf for
10685 *) Make sure latest Perl versions don't interpret some generated C array
10686 code as Perl array code in the crypto/err/err_genc.pl script.
10687 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10689 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10690 not many people have the assembler. Various Win32 compilation fixes and
10691 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10692 build instructions.
10695 *) Modify configure script 'Configure' to automatically create crypto/date.h
10696 file under Win32 and also build pem.h from pem.org. New script
10697 util/mkfiles.pl to create the MINFO file on environments that can't do a
10698 'make files': perl util/mkfiles.pl >MINFO should work.
10701 *) Major rework of DES function declarations, in the pursuit of correctness
10702 and purity. As a result, many evil casts evaporated, and some weirdness,
10703 too. You may find this causes warnings in your code. Zapping your evil
10704 casts will probably fix them. Mostly.
10707 *) Fix for a typo in asn1.h. Bug fix to object creation script
10708 obj_dat.pl. It considered a zero in an object definition to mean
10709 "end of object": none of the objects in objects.h have any zeros
10710 so it wasn't spotted.
10711 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10713 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10714 Masking (CBCM). In the absence of test vectors, the best I have been able
10715 to do is check that the decrypt undoes the encrypt, so far. Send me test
10716 vectors if you have them.
10719 *) Correct calculation of key length for export ciphers (too much space was
10720 allocated for null ciphers). This has not been tested!
10723 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10724 message is now correct (it understands "crypto" and "ssl" on its
10725 command line). There is also now an "update" option. This will update
10726 the util/ssleay.num and util/libeay.num files with any new functions.
10728 perl util/mkdef.pl crypto ssl update
10729 it will update them.
10732 *) Overhauled the Perl interface (perl/*):
10733 - ported BN stuff to OpenSSL's different BN library
10734 - made the perl/ source tree CVS-aware
10735 - renamed the package from SSLeay to OpenSSL (the files still contain
10736 their history because I've copied them in the repository)
10737 - removed obsolete files (the test scripts will be replaced
10738 by better Test::Harness variants in the future)
10739 [Ralf S. Engelschall]
10741 *) First cut for a very conservative source tree cleanup:
10742 1. merge various obsolete readme texts into doc/ssleay.txt
10743 where we collect the old documents and readme texts.
10744 2. remove the first part of files where I'm already sure that we no
10745 longer need them because of three reasons: either they are just temporary
10746 files which were left by Eric or they are preserved original files where
10747 I've verified that the diff is also available in the CVS via "cvs diff
10748 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10749 the crypto/md/ stuff).
10750 [Ralf S. Engelschall]
10752 *) More extension code. Incomplete support for subject and issuer alt
10753 name, issuer and authority key id. Change the i2v function parameters
10754 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10755 what that's for :-) Fix to ASN1 macro which messed up
10756 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10759 *) Preliminary support for ENUMERATED type. This is largely copied from the
10763 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10764 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10766 *) Make sure `make rehash' target really finds the `openssl' program.
10767 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10769 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10770 like to hear about it if this slows down other processors.
10773 *) Add CygWin32 platform information to Configure script.
10774 [Alan Batie <batie@aahz.jf.intel.com>]
10776 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10777 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10779 *) New program nseq to manipulate netscape certificate sequences
10782 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10786 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10787 but the BN code had some problems that would cause failures when
10788 doing certificate verification and some other functions.
10789 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10791 *) Add ASN1 and PEM code to support netscape certificate sequences.
10794 *) Add ASN1 and PEM code to support netscape certificate sequences.
10797 *) Add several PKIX and private extended key usage OIDs.
10800 *) Modify the 'ca' program to handle the new extension code. Modify
10801 openssl.cnf for new extension format, add comments.
10804 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10805 and add a sample to openssl.cnf so req -x509 now adds appropriate
10809 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10810 error code, add initial support to X509_print() and x509 application.
10813 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10814 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10815 stuff is currently isolated and isn't even compiled yet.
10818 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10819 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10820 Removed the versions check from X509 routines when loading extensions:
10821 this allows certain broken certificates that don't set the version
10822 properly to be processed.
10825 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10826 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10827 can still be regenerated with "make depend".
10830 *) Spelling mistake in C version of CAST-128.
10831 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10833 *) Changes to the error generation code. The perl script err-code.pl
10834 now reads in the old error codes and retains the old numbers, only
10835 adding new ones if necessary. It also only changes the .err files if new
10836 codes are added. The makefiles have been modified to only insert errors
10837 when needed (to avoid needlessly modifying header files). This is done
10838 by only inserting errors if the .err file is newer than the auto generated
10839 C file. To rebuild all the error codes from scratch (the old behaviour)
10840 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10841 or delete all the .err files.
10844 *) CAST-128 was incorrectly implemented for short keys. The C version has
10845 been fixed, but is untested. The assembler versions are also fixed, but
10846 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10847 to regenerate it if needed.
10848 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10849 Hagino <itojun@kame.net>]
10851 *) File was opened incorrectly in randfile.c.
10852 [Ulf Möller <ulf@fitug.de>]
10854 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10855 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10856 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10857 al: it's just almost always a UTCTime. Note this patch adds new error
10858 codes so do a "make errors" if there are problems.
10861 *) Correct Linux 1 recognition in config.
10862 [Ulf Möller <ulf@fitug.de>]
10864 *) Remove pointless MD5 hash when using DSA keys in ca.
10865 [Anonymous <nobody@replay.com>]
10867 *) Generate an error if given an empty string as a cert directory. Also
10868 generate an error if handed NULL (previously returned 0 to indicate an
10869 error, but didn't set one).
10870 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10872 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10875 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10876 parameters. This was causing a warning which killed off the Win32 compile.
10879 *) Remove C++ style comments from crypto/bn/bn_local.h.
10880 [Neil Costigan <neil.costigan@celocom.com>]
10882 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10883 based on a text string, looking up short and long names and finally
10884 "dot" format. The "dot" format stuff didn't work. Added new function
10885 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10886 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10887 OID is not part of the table.
10890 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10891 X509_LOOKUP_by_alias().
10894 *) Sort openssl functions by name.
10897 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10898 encryption from sample DSA keys (in case anyone is interested the password
10902 *) Make _all_ *_free functions accept a NULL pointer.
10903 [Frans Heymans <fheymans@isaserver.be>]
10905 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10907 [Anonymous <nobody@replay.com>]
10909 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10910 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10912 *) Don't blow it for numeric -newkey arguments to apps/req.
10913 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10915 *) Temp key "for export" tests were wrong in s3_srvr.c.
10916 [Anonymous <nobody@replay.com>]
10918 *) Add prototype for temp key callback functions
10919 SSL_CTX_set_tmp_{rsa,dh}_callback().
10922 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10923 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10926 *) X509_name_add_entry() freed the wrong thing after an error.
10927 [Arne Ansper <arne@ats.cyber.ee>]
10929 *) rsa_eay.c would attempt to free a NULL context.
10930 [Arne Ansper <arne@ats.cyber.ee>]
10932 *) BIO_s_socket() had a broken should_retry() on Windoze.
10933 [Arne Ansper <arne@ats.cyber.ee>]
10935 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10936 [Arne Ansper <arne@ats.cyber.ee>]
10938 *) Make sure the already existing X509_STORE->depth variable is initialized
10939 in X509_STORE_new(), but document the fact that this variable is still
10940 unused in the certificate verification process.
10941 [Ralf S. Engelschall]
10943 *) Fix the various library and apps files to free up pkeys obtained from
10944 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10947 *) Fix reference counting in X509_PUBKEY_get(). This makes
10948 demos/maurice/example2.c work, amongst others, probably.
10949 [Steve Henson and Ben Laurie]
10951 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10952 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10953 are no longer created. This way we have a single and consistent command
10954 line interface `openssl <command>', similar to `cvs <command>'.
10955 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10957 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10958 BIT STRING wrapper always have zero unused bits.
10961 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10964 *) Make the top-level INSTALL documentation easier to understand.
10967 *) Makefiles updated to exit if an error occurs in a sub-directory
10968 make (including if user presses ^C) [Paul Sutton]
10970 *) Make Montgomery context stuff explicit in RSA data structure.
10973 *) Fix build order of pem and err to allow for generated pem.h.
10976 *) Fix renumbering bug in X509_NAME_delete_entry().
10979 *) Enhanced the err-ins.pl script so it makes the error library number
10980 global and can add a library name. This is needed for external ASN1 and
10981 other error libraries.
10984 *) Fixed sk_insert which never worked properly.
10987 *) Fix ASN1 macros so they can handle indefinite length constructed
10988 EXPLICIT tags. Some non standard certificates use these: they can now
10992 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10993 into a single doc/ssleay.txt bundle. This way the information is still
10994 preserved but no longer messes up this directory. Now it's new room for
10995 the new set of documentation files.
10996 [Ralf S. Engelschall]
10998 *) SETs were incorrectly DER encoded. This was a major pain, because they
10999 shared code with SEQUENCEs, which aren't coded the same. This means that
11000 almost everything to do with SETs or SEQUENCEs has either changed name or
11001 number of arguments.
11002 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
11004 *) Fix test data to work with the above.
11007 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
11008 was already fixed by Eric for 0.9.1 it seems.
11009 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
11011 *) Autodetect FreeBSD3.
11014 *) Fix various bugs in Configure. This affects the following platforms:
11018 unixware-2.0-pentium
11022 *) Eliminate generated files from CVS. Reorder tests to regenerate files
11023 before they are needed.
11026 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
11030 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
11032 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
11033 changed SSLeay to OpenSSL in version strings.
11034 [Ralf S. Engelschall]
11036 *) Some fixups to the top-level documents.
11039 *) Fixed the nasty bug where rsaref.h was not found under compile-time
11040 because the symlink to include/ was missing.
11041 [Ralf S. Engelschall]
11043 *) Incorporated the popular no-RSA/DSA-only patches
11044 which allow to compile a RSA-free SSLeay.
11045 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
11047 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
11048 when "ssleay" is still not found.
11049 [Ralf S. Engelschall]
11051 *) Added more platforms to Configure: Cray T3E, HPUX 11,
11052 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
11054 *) Updated the README file.
11055 [Ralf S. Engelschall]
11057 *) Added various .cvsignore files in the CVS repository subdirs
11058 to make a "cvs update" really silent.
11059 [Ralf S. Engelschall]
11061 *) Recompiled the error-definition header files and added
11062 missing symbols to the Win32 linker tables.
11063 [Ralf S. Engelschall]
11065 *) Cleaned up the top-level documents;
11066 o new files: CHANGES and LICENSE
11067 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
11068 o merged COPYRIGHT into LICENSE
11069 o removed obsolete TODO file
11070 o renamed MICROSOFT to INSTALL.W32
11071 [Ralf S. Engelschall]
11073 *) Removed dummy files from the 0.9.1b source tree:
11074 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
11075 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
11076 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
11077 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
11078 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
11079 [Ralf S. Engelschall]
11081 *) Added various platform portability fixes.
11084 *) The Genesis of the OpenSSL rpject:
11085 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
11086 Young and Tim J. Hudson created while they were working for C2Net until
11088 [The OpenSSL Project]
11091 Changes between 0.9.0b and 0.9.1b [not released]
11093 *) Updated a few CA certificates under certs/
11096 *) Changed some BIGNUM api stuff.
11099 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
11100 DGUX x86, Linux Alpha, etc.
11103 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
11104 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
11108 *) Add -strparse option to asn1pars program which parses nested
11110 [Dr Stephen Henson <shenson@bigfoot.com>]
11112 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
11115 *) DSA fix for "ca" program.
11118 *) Added "-genkey" option to "dsaparam" program.
11121 *) Added RIPE MD160 (rmd160) message digest.
11124 *) Added -a (all) option to "ssleay version" command.
11127 *) Added PLATFORM define which is the id given to Configure.
11130 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
11133 *) Extended the ASN.1 parser routines.
11136 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11139 *) Added a BN_CTX to the BN library.
11142 *) Fixed the weak key values in DES library
11145 *) Changed API in EVP library for cipher aliases.
11148 *) Added support for RC2/64bit cipher.
11151 *) Converted the lhash library to the crypto/mem.c functions.
11154 *) Added more recognized ASN.1 object ids.
11157 *) Added more RSA padding checks for SSL/TLS.
11160 *) Added BIO proxy/filter functionality.
11163 *) Added extra_certs to SSL_CTX which can be used
11164 send extra CA certificates to the client in the CA cert chain sending
11165 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11168 *) Now Fortezza is denied in the authentication phase because
11169 this is key exchange mechanism is not supported by SSLeay at all.
11172 *) Additional PKCS1 checks.
11175 *) Support the string "TLSv1" for all TLS v1 ciphers.
11178 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11179 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11182 *) Fixed a few memory leaks.
11185 *) Fixed various code and comment typos.
11188 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11189 bytes sent in the client random.
11190 [Edward Bishop <ebishop@spyglass.com>]