5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
7 *) Change default behaviour of 'openssl asn1parse' so that more
8 information is visible when viewing, e.g., a certificate:
10 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11 mode the content of non-printable OCTET STRINGs is output in a
12 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
13 avoid the appearance of a printable string.
14 [Nils Larsch <nla@trustcenter.de>]
16 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
18 EC_GROUP_set_asn1_flag()
19 EC_GROUP_get_asn1_flag()
20 EC_GROUP_set_point_conversion_form()
21 EC_GROUP_get_point_conversion_form()
22 These control ASN1 encoding details:
23 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
24 has been set to OPENSSL_EC_NAMED_CURVE.
25 - Points are encoded in uncompressed form by default; options for
26 asn1_for are as for point2oct, namely
27 POINT_CONVERSION_COMPRESSED
28 POINT_CONVERSION_UNCOMPRESSED
29 POINT_CONVERSION_HYBRID
31 Also add 'seed' and 'seed_len' members to EC_GROUP with access
35 EC_GROUP_get_seed_len()
36 This is used only for ASN1 purposes (so far).
37 [Nils Larsch <nla@trustcenter.de>]
39 *) Add 'field_type' member to EC_METHOD, which holds the NID
40 of the appropriate field type OID. The new function
41 EC_METHOD_get_field_type() returns this value.
42 [Nils Larsch <nla@trustcenter.de>]
49 providing useful interfaces to EC_POINT_point2oct() and
51 [Nils Larsch <nla@trustcenter.de>]
53 *) Change internals of the EC library so that the functions
54 EC_GROUP_set_generator()
55 EC_GROUP_get_generator()
57 EC_GROUP_get_cofactor()
58 are implemented directly in crypto/ec/ec_lib.c and not dispatched
59 to methods, which would lead to unnecessary code duplication when
60 adding different types of curves.
61 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
63 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
64 arithmetic, and such that modified wNAFs are generated
65 (which avoid length expansion in many cases).
68 *) Add a function EC_GROUP_check_discriminant() (defined via
69 EC_METHOD) that verifies that the curve discriminant is non-zero.
71 Add a function EC_GROUP_check() that makes some sanity tests
72 on a EC_GROUP, its generator and order. This includes
73 EC_GROUP_check_discriminant().
74 [Nils Larsch <nla@trustcenter.de>]
76 *) Add ECDSA in new directory crypto/ecdsa/.
78 Add applications 'openssl ecparam' and 'openssl ecdsa'
79 (these are based on 'openssl dsaparam' and 'openssl dsa').
81 ECDSA support is also included in various other files across the
82 library. Most notably,
83 - 'openssl req' now has a '-newkey ecdsa:file' option;
84 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
85 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
86 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
87 them suitable for ECDSA where domain parameters must be
88 extracted before the specific public key.
89 [Nils Larsch <nla@trustcenter.de>]
91 *) Include some named elliptic curves, and add OIDs from X9.62,
92 SECG, and WAP/WTLS. The curves can be obtained from the new
95 EC_GROUP_new_by_name()
96 Also add a 'curve_name' member to EC_GROUP objects, which can be
100 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
102 Changes between 0.9.6e and 0.9.7 [XX xxx 2002]
104 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
105 to allow version independent disabling of normally unselected ciphers,
106 which may be activated as a side-effect of selecting a single cipher.
108 (E.g., cipher list string "RSA" enables ciphersuites that are left
109 out of "ALL" because they do not provide symmetric encryption.
110 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
111 [Lutz Jaenicke, Bodo Moeller]
113 *) Add appropriate support for separate platform-dependent build
114 directories. The recommended way to make a platform-dependent
115 build directory is the following (tested on Linux), maybe with
118 # Place yourself outside of the OpenSSL source tree. In
119 # this example, the environment variable OPENSSL_SOURCE
120 # is assumed to contain the absolute OpenSSL source directory.
121 mkdir -p objtree/`uname -s`-`uname -r`-`uname -m`
122 cd objtree/`uname -s`-`uname -r`-`uname -m`
123 (cd $OPENSSL_SOURCE; find . -type f -o -type l) | while read F; do
124 mkdir -p `dirname $F`
125 ln -s $OPENSSL_SOURCE/$F $F
128 To be absolutely sure not to disturb the source tree, a "make clean"
129 is a good thing. If it isn't successfull, don't worry about it,
130 it probably means the source directory is very clean.
133 *) Make sure any ENGINE control commands make local copies of string
134 pointers passed to them whenever necessary. Otherwise it is possible
135 the caller may have overwritten (or deallocated) the original string
136 data when a later ENGINE operation tries to use the stored values.
137 [Götz Babin-Ebell <babinebell@trustcenter.de>]
139 *) Improve diagnostics in file reading and command-line digests.
140 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
142 *) Add AES modes CFB and OFB to the object database. Correct an
143 error in AES-CFB decryption.
146 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
147 allows existing EVP_CIPHER_CTX structures to be reused after
148 calling EVP_*Final(). This behaviour is used by encryption
149 BIOs and some applications. This has the side effect that
150 applications must explicitly clean up cipher contexts with
151 EVP_CIPHER_CTX_cleanup() or they will leak memory.
154 *) Check the values of dna and dnb in bn_mul_recursive before calling
155 bn_mul_comba (a non zero value means the a or b arrays do not contain
156 n2 elements) and fallback to bn_mul_normal if either is not zero.
159 *) Fix escaping of non-ASCII characters when using the -subj option
160 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
163 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
164 form for "surname", serialNumber has no short form.
165 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
166 therefore remove "mail" short name for "internet 7".
167 The OID for unique identifiers in X509 certificates is
168 x500UniqueIdentifier, not uniqueIdentifier.
169 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
172 *) Add an "init" command to the ENGINE config module and auto initialize
173 ENGINEs. Without any "init" command the ENGINE will be initialized
174 after all ctrl commands have been executed on it. If init=1 the
175 ENGINE is initailized at that point (ctrls before that point are run
176 on the uninitialized ENGINE and after on the initialized one). If
177 init=0 then the ENGINE will not be iniatialized at all.
180 *) Fix the 'app_verify_callback' interface so that the user-defined
181 argument is actually passed to the callback: In the
182 SSL_CTX_set_cert_verify_callback() prototype, the callback
183 declaration has been changed from
186 int (*cb)(X509_STORE_CTX *,void *);
187 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
188 i=s->ctx->app_verify_callback(&ctx)
189 has been changed into
190 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
192 To update applications using SSL_CTX_set_cert_verify_callback(),
193 a dummy argument can be added to their callback functions.
194 [D. K. Smetters <smetters@parc.xerox.com>]
196 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
197 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
199 *) Add and OPENSSL_LOAD_CONF define which will cause
200 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
201 This allows older applications to transparently support certain
202 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
203 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
204 load the config file and OPENSSL_add_all_algorithms_conf() which will
205 always load it have also been added.
208 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
209 Adjust NIDs and EVP layer.
210 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
212 *) Config modules support in openssl utility.
214 Most commands now load modules from the config file,
215 though in a few (such as version) this isn't done
216 because it couldn't be used for anything.
218 In the case of ca and req the config file used is
219 the same as the utility itself: that is the -config
220 command line option can be used to specify an
224 *) Move default behaviour from OPENSSL_config(). If appname is NULL
225 use "openssl_conf" if filename is NULL use default openssl config file.
228 *) Add an argument to OPENSSL_config() to allow the use of an alternative
229 config section name. Add a new flag to tolerate a missing config file
230 and move code to CONF_modules_load_file().
233 *) Support for crypto accelerator cards from Accelerated Encryption
234 Processing, www.aep.ie. (Use engine 'aep')
235 The support was copied from 0.9.6c [engine] and adapted/corrected
236 to work with the new engine framework.
237 [AEP Inc. and Richard Levitte]
239 *) Support for SureWare crypto accelerator cards from Baltimore
240 Technologies. (Use engine 'sureware')
241 The support was copied from 0.9.6c [engine] and adapted
242 to work with the new engine framework.
245 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
246 make the newer ENGINE framework commands for the CHIL engine work.
247 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
249 *) Make it possible to produce shared libraries on ReliantUNIX.
250 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
252 *) Add the configuration target debug-linux-ppro.
253 Make 'openssl rsa' use the general key loading routines
254 implemented in apps.c, and make those routines able to
255 handle the key format FORMAT_NETSCAPE and the variant
257 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
259 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
260 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
262 *) Add -keyform to rsautl, and document -engine.
263 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
265 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
266 BIO_R_NO_SUCH_FILE error code rather than the generic
267 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
272 ERR_peek_last_error_line
273 ERR_peek_last_error_line_data.
277 ERR_peek_error_line_data,
278 but report on the latest error recorded rather than the first one
279 still in the error queue.
280 [Ben Laurie, Bodo Moeller]
282 *) default_algorithms option in ENGINE config module. This allows things
284 default_algorithms = ALL
285 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
288 *) Prelminary ENGINE config module.
291 *) New experimental application configuration code.
294 *) Change the AES code to follow the same name structure as all other
295 symmetric ciphers, and behave the same way. Move everything to
296 the directory crypto/aes, thereby obsoleting crypto/rijndael.
297 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
299 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
300 [Ben Laurie and Theo de Raadt]
302 *) Add option to output public keys in req command.
303 [Massimiliano Pala madwolf@openca.org]
305 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
306 (up to about 10% better than before for P-192 and P-224).
309 *) New functions/macros
311 SSL_CTX_set_msg_callback(ctx, cb)
312 SSL_CTX_set_msg_callback_arg(ctx, arg)
313 SSL_set_msg_callback(ssl, cb)
314 SSL_set_msg_callback_arg(ssl, arg)
316 to request calling a callback function
318 void cb(int write_p, int version, int content_type,
319 const void *buf, size_t len, SSL *ssl, void *arg)
321 whenever a protocol message has been completely received
322 (write_p == 0) or sent (write_p == 1). Here 'version' is the
323 protocol version according to which the SSL library interprets
324 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
325 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
326 the content type as defined in the SSL 3.0/TLS 1.0 protocol
327 specification (change_cipher_spec(20), alert(21), handshake(22)).
328 'buf' and 'len' point to the actual message, 'ssl' to the
329 SSL object, and 'arg' is the application-defined value set by
330 SSL[_CTX]_set_msg_callback_arg().
332 'openssl s_client' and 'openssl s_server' have new '-msg' options
333 to enable a callback that displays all protocol messages.
336 *) Change the shared library support so shared libraries are built as
337 soon as the corresponding static library is finished, and thereby get
338 openssl and the test programs linked against the shared library.
339 This still only happens when the keyword "shard" has been given to
340 the configuration scripts.
342 NOTE: shared library support is still an experimental thing, and
343 backward binary compatibility is still not guaranteed.
344 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
346 *) Add support for Subject Information Access extension.
347 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
349 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
350 additional bytes when new memory had to be allocated, not just
351 when reusing an existing buffer.
354 *) New command line and configuration option 'utf8' for the req command.
355 This allows field values to be specified as UTF8 strings.
358 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
359 runs for the former and machine-readable output for the latter.
362 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
363 of the e-mail address in the DN (i.e., it will go into a certificate
364 extension only). The new configuration file option 'email_in_dn = no'
366 [Massimiliano Pala madwolf@openca.org]
368 *) Change all functions with names starting with des_ to be starting
369 with DES_ instead. Add wrappers that are compatible with libdes,
370 but are named _ossl_old_des_*. Finally, add macros that map the
371 des_* symbols to the corresponding _ossl_old_des_* if libdes
372 compatibility is desired. If OpenSSL 0.9.6c compatibility is
373 desired, the des_* symbols will be mapped to DES_*, with one
376 Since we provide two compatibility mappings, the user needs to
377 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
378 compatibility is desired. The default (i.e., when that macro
379 isn't defined) is OpenSSL 0.9.6c compatibility.
381 There are also macros that enable and disable the support of old
382 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
383 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
384 are defined, the default will apply: to support the old des routines.
386 In either case, one must include openssl/des.h to get the correct
387 definitions. Do not try to just include openssl/des_old.h, that
390 NOTE: This is a major break of an old API into a new one. Software
391 authors are encouraged to switch to the DES_ style functions. Some
392 time in the future, des_old.h and the libdes compatibility functions
393 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
394 default), and then completely removed.
397 *) Test for certificates which contain unsupported critical extensions.
398 If such a certificate is found during a verify operation it is
399 rejected by default: this behaviour can be overridden by either
400 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
401 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
402 X509_supported_extension() has also been added which returns 1 if a
403 particular extension is supported.
406 *) Modify the behaviour of EVP cipher functions in similar way to digests
407 to retain compatibility with existing code.
410 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
411 compatibility with existing code. In particular the 'ctx' parameter does
412 not have to be to be initialized before the call to EVP_DigestInit() and
413 it is tidied up after a call to EVP_DigestFinal(). New function
414 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
415 EVP_MD_CTX_copy() changed to not require the destination to be
416 initialized valid and new function EVP_MD_CTX_copy_ex() added which
417 requires the destination to be valid.
419 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
420 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
423 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
424 so that complete 'Handshake' protocol structures are kept in memory
425 instead of overwriting 'msg_type' and 'length' with 'body' data.
428 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
429 [Massimo Santin via Richard Levitte]
431 *) Major restructuring to the underlying ENGINE code. This includes
432 reduction of linker bloat, separation of pure "ENGINE" manipulation
433 (initialisation, etc) from functionality dealing with implementations
434 of specific crypto iterfaces. This change also introduces integrated
435 support for symmetric ciphers and digest implementations - so ENGINEs
436 can now accelerate these by providing EVP_CIPHER and EVP_MD
437 implementations of their own. This is detailed in crypto/engine/README
438 as it couldn't be adequately described here. However, there are a few
439 API changes worth noting - some RSA, DSA, DH, and RAND functions that
440 were changed in the original introduction of ENGINE code have now
441 reverted back - the hooking from this code to ENGINE is now a good
442 deal more passive and at run-time, operations deal directly with
443 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
444 dereferencing through an ENGINE pointer any more. Also, the ENGINE
445 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
446 they were not being used by the framework as there is no concept of a
447 BIGNUM_METHOD and they could not be generalised to the new
448 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
449 ENGINE_cpy() has been removed as it cannot be consistently defined in
453 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
456 *) Change mkdef.pl to sort symbols that get the same entry number,
457 and make sure the automatically generated functions ERR_load_*
458 become part of libeay.num as well.
461 *) New function SSL_renegotiate_pending(). This returns true once
462 renegotiation has been requested (either SSL_renegotiate() call
463 or HelloRequest/ClientHello receveived from the peer) and becomes
464 false once a handshake has been completed.
465 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
466 sends a HelloRequest, but does not ensure that a handshake takes
467 place. SSL_renegotiate_pending() is useful for checking if the
468 client has followed the request.)
471 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
472 By default, clients may request session resumption even during
473 renegotiation (if session ID contexts permit); with this option,
474 session resumption is possible only in the first handshake.
476 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
477 more bits available for options that should not be part of
478 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
481 *) Add some demos for certificate and certificate request creation.
484 *) Make maximum certificate chain size accepted from the peer application
485 settable (SSL*_get/set_max_cert_list()), as proposed by
486 "Douglas E. Engert" <deengert@anl.gov>.
489 *) Add support for shared libraries for Unixware-7
490 (Boyd Lynn Gerber <gerberb@zenez.com>).
493 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
494 be done prior to destruction. Use this to unload error strings from
495 ENGINEs that load their own error strings. NB: This adds two new API
496 functions to "get" and "set" this destroy handler in an ENGINE.
499 *) Alter all existing ENGINE implementations (except "openssl" and
500 "openbsd") to dynamically instantiate their own error strings. This
501 makes them more flexible to be built both as statically-linked ENGINEs
502 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
503 Also, add stub code to each that makes building them as self-contained
504 shared-libraries easier (see README.ENGINE).
507 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
508 implementations into applications that are completely implemented in
509 self-contained shared-libraries. The "dynamic" ENGINE exposes control
510 commands that can be used to configure what shared-library to load and
511 to control aspects of the way it is handled. Also, made an update to
512 the README.ENGINE file that brings its information up-to-date and
513 provides some information and instructions on the "dynamic" ENGINE
514 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
517 *) Make it possible to unload ranges of ERR strings with a new
518 "ERR_unload_strings" function.
521 *) Add a copy() function to EVP_MD.
524 *) Make EVP_MD routines take a context pointer instead of just the
525 md_data void pointer.
528 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
529 that the digest can only process a single chunk of data
530 (typically because it is provided by a piece of
531 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
532 is only going to provide a single chunk of data, and hence the
533 framework needn't accumulate the data for oneshot drivers.
536 *) As with "ERR", make it possible to replace the underlying "ex_data"
537 functions. This change also alters the storage and management of global
538 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
539 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
540 index counters. The API functions that use this state have been changed
541 to take a "class_index" rather than pointers to the class's local STACK
542 and counter, and there is now an API function to dynamically create new
543 classes. This centralisation allows us to (a) plug a lot of the
544 thread-safety problems that existed, and (b) makes it possible to clean
545 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
546 such data would previously have always leaked in application code and
547 workarounds were in place to make the memory debugging turn a blind eye
548 to it. Application code that doesn't use this new function will still
549 leak as before, but their memory debugging output will announce it now
550 rather than letting it slide.
552 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
553 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
554 has a return value to indicate success or failure.
557 *) Make it possible to replace the underlying "ERR" functions such that the
558 global state (2 LHASH tables and 2 locks) is only used by the "default"
559 implementation. This change also adds two functions to "get" and "set"
560 the implementation prior to it being automatically set the first time
561 any other ERR function takes place. Ie. an application can call "get",
562 pass the return value to a module it has just loaded, and that module
563 can call its own "set" function using that value. This means the
564 module's "ERR" operations will use (and modify) the error state in the
565 application and not in its own statically linked copy of OpenSSL code.
568 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
569 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
570 the operation, and provides a more encapsulated way for external code
571 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
572 to use these functions rather than manually incrementing the counts.
574 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
577 *) Add EVP test program.
580 *) Add symmetric cipher support to ENGINE. Expect the API to change!
583 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
584 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
585 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
586 These allow a CRL to be built without having to access X509_CRL fields
587 directly. Modify 'ca' application to use new functions.
590 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
591 bug workarounds. Rollback attack detection is a security feature.
592 The problem will only arise on OpenSSL servers when TLSv1 is not
593 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
594 Software authors not wanting to support TLSv1 will have special reasons
595 for their choice and can explicitly enable this option.
596 [Bodo Moeller, Lutz Jaenicke]
598 *) Rationalise EVP so it can be extended: don't include a union of
599 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
600 (similar to those existing for EVP_CIPHER_CTX).
605 EVP_MD_CTX_init(&md); /* new function call */
606 EVP_DigestInit(&md, EVP_sha1());
607 EVP_DigestUpdate(&md, in, len);
608 EVP_DigestFinal(&md, out, NULL);
609 EVP_MD_CTX_cleanup(&md); /* new function call */
613 *) Make DES key schedule conform to the usual scheme, as well as
614 correcting its structure. This means that calls to DES functions
615 now have to pass a pointer to a des_key_schedule instead of a
616 plain des_key_schedule (which was actually always a pointer
621 des_set_key_checked(..., &ks);
622 des_ncbc_encrypt(..., &ks, ...);
624 (Note that a later change renames 'des_...' into 'DES_...'.)
627 *) Initial reduction of linker bloat: the use of some functions, such as
628 PEM causes large amounts of unused functions to be linked in due to
629 poor organisation. For example pem_all.c contains every PEM function
630 which has a knock on effect of linking in large amounts of (unused)
631 ASN1 code. Grouping together similar functions and splitting unrelated
632 functions prevents this.
635 *) Cleanup of EVP macros.
638 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
642 *) Add initial OCSP responder support to ocsp application. The
643 revocation information is handled using the text based index
644 use by the ca application. The responder can either handle
645 requests generated internally, supplied in files (for example
646 via a CGI script) or using an internal minimal server.
649 *) Add configuration choices to get zlib compression for TLS.
652 *) Changes to Kerberos SSL for RFC 2712 compliance:
653 1. Implemented real KerberosWrapper, instead of just using
654 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
655 2. Implemented optional authenticator field of KerberosWrapper.
657 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
658 and authenticator structs; see crypto/krb5/.
660 Generalized Kerberos calls to support multiple Kerberos libraries.
661 [Vern Staats <staatsvr@asc.hpc.mil>,
662 Jeffrey Altman <jaltman@columbia.edu>
665 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
666 already does with RSA. testdsa.h now has 'priv_key/pub_key'
667 values for each of the key sizes rather than having just
668 parameters (and 'speed' generating keys each time).
671 *) Speed up EVP routines.
674 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
675 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
676 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
677 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
679 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
680 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
681 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
684 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
686 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
689 *) Added the OS2-EMX target.
690 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
692 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
693 to support NCONF routines in extension code. New function CONF_set_nconf()
694 to allow functions which take an NCONF to also handle the old LHASH
695 structure: this means that the old CONF compatible routines can be
696 retained (in particular wrt extensions) without having to duplicate the
697 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
700 *) Enhance the general user interface with mechanisms for inner control
701 and with possibilities to have yes/no kind of prompts.
704 *) Change all calls to low level digest routines in the library and
705 applications to use EVP. Add missing calls to HMAC_cleanup() and
706 don't assume HMAC_CTX can be copied using memcpy().
707 [Verdon Walker <VWalker@novell.com>, Steve Henson]
709 *) Add the possibility to control engines through control names but with
710 arbitrary arguments instead of just a string.
711 Change the key loaders to take a UI_METHOD instead of a callback
712 function pointer. NOTE: this breaks binary compatibility with earlier
713 versions of OpenSSL [engine].
714 Adapt the nCipher code for these new conditions and add a card insertion
718 *) Enhance the general user interface with mechanisms to better support
719 dialog box interfaces, application-defined prompts, the possibility
720 to use defaults (for example default passwords from somewhere else)
721 and interrupts/cancellations.
724 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
725 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
728 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
729 tidy up some unnecessarily weird code in 'sk_new()').
730 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
732 *) Change the key loading routines for ENGINEs to use the same kind
733 callback (pem_password_cb) as all other routines that need this
737 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
738 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
739 than this minimum value is recommended.
742 *) New random seeder for OpenVMS, using the system process statistics
743 that are easily reachable.
746 *) Windows apparently can't transparently handle global
747 variables defined in DLLs. Initialisations such as:
749 const ASN1_ITEM *it = &ASN1_INTEGER_it;
751 wont compile. This is used by the any applications that need to
752 declare their own ASN1 modules. This was fixed by adding the option
753 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
754 needed for static libraries under Win32.
757 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
758 setting of purpose and trust fields. New X509_STORE trust and
759 purpose functions and tidy up setting in other SSL functions.
762 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
763 structure. These are inherited by X509_STORE_CTX when it is
764 initialised. This allows various defaults to be set in the
765 X509_STORE structure (such as flags for CRL checking and custom
766 purpose or trust settings) for functions which only use X509_STORE_CTX
767 internally such as S/MIME.
769 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
770 trust settings if they are not set in X509_STORE. This allows X509_STORE
771 purposes and trust (in S/MIME for example) to override any set by default.
773 Add command line options for CRL checking to smime, s_client and s_server
777 *) Initial CRL based revocation checking. If the CRL checking flag(s)
778 are set then the CRL is looked up in the X509_STORE structure and
779 its validity and signature checked, then if the certificate is found
780 in the CRL the verify fails with a revoked error.
782 Various new CRL related callbacks added to X509_STORE_CTX structure.
784 Command line options added to 'verify' application to support this.
786 This needs some additional work, such as being able to handle multiple
787 CRLs with different times, extension based lookup (rather than just
788 by subject name) and ultimately more complete V2 CRL extension
792 *) Add a general user interface API (crypto/ui/). This is designed
793 to replace things like des_read_password and friends (backward
794 compatibility functions using this new API are provided).
795 The purpose is to remove prompting functions from the DES code
796 section as well as provide for prompting through dialog boxes in
797 a window system and the like.
800 *) Add "ex_data" support to ENGINE so implementations can add state at a
801 per-structure level rather than having to store it globally.
804 *) Make it possible for ENGINE structures to be copied when retrieved by
805 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
806 This causes the "original" ENGINE structure to act like a template,
807 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
808 operational state can be localised to each ENGINE structure, despite the
809 fact they all share the same "methods". New ENGINE structures returned in
810 this case have no functional references and the return value is the single
811 structural reference. This matches the single structural reference returned
812 by ENGINE_by_id() normally, when it is incremented on the pre-existing
816 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
817 needs to match any other type at all we need to manually clear the
821 *) Changes to the "openssl engine" utility to include;
822 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
823 about an ENGINE's available control commands.
824 - executing control commands from command line arguments using the
825 '-pre' and '-post' switches. '-post' is only used if '-t' is
826 specified and the ENGINE is successfully initialised. The syntax for
827 the individual commands are colon-separated, for example;
828 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
831 *) New dynamic control command support for ENGINEs. ENGINEs can now
832 declare their own commands (numbers), names (strings), descriptions,
833 and input types for run-time discovery by calling applications. A
834 subset of these commands are implicitly classed as "executable"
835 depending on their input type, and only these can be invoked through
836 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
837 can be based on user input, config files, etc). The distinction is
838 that "executable" commands cannot return anything other than a boolean
839 result and can only support numeric or string input, whereas some
840 discoverable commands may only be for direct use through
841 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
842 pointers, or other custom uses. The "executable" commands are to
843 support parameterisations of ENGINE behaviour that can be
844 unambiguously defined by ENGINEs and used consistently across any
845 OpenSSL-based application. Commands have been added to all the
846 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
847 control over shared-library paths without source code alterations.
850 *) Changed all ENGINE implementations to dynamically allocate their
851 ENGINEs rather than declaring them statically. Apart from this being
852 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
853 this also allows the implementations to compile without using the
854 internal engine_int.h header.
857 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
858 'const' value. Any code that should be able to modify a RAND_METHOD
859 should already have non-const pointers to it (ie. they should only
860 modify their own ones).
863 *) Made a variety of little tweaks to the ENGINE code.
864 - "atalla" and "ubsec" string definitions were moved from header files
865 to C code. "nuron" string definitions were placed in variables
866 rather than hard-coded - allowing parameterisation of these values
867 later on via ctrl() commands.
868 - Removed unused "#if 0"'d code.
869 - Fixed engine list iteration code so it uses ENGINE_free() to release
870 structural references.
871 - Constified the RAND_METHOD element of ENGINE structures.
872 - Constified various get/set functions as appropriate and added
873 missing functions (including a catch-all ENGINE_cpy that duplicates
874 all ENGINE values onto a new ENGINE except reference counts/state).
875 - Removed NULL parameter checks in get/set functions. Setting a method
876 or function to NULL is a way of cancelling out a previously set
877 value. Passing a NULL ENGINE parameter is just plain stupid anyway
878 and doesn't justify the extra error symbols and code.
879 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
880 flags from engine_int.h to engine.h.
881 - Changed prototypes for ENGINE handler functions (init(), finish(),
882 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
885 *) Implement binary inversion algorithm for BN_mod_inverse in addition
886 to the algorithm using long division. The binary algorithm can be
887 used only if the modulus is odd. On 32-bit systems, it is faster
888 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
889 roughly 5-15% for 256-bit moduli), so we use it only for moduli
890 up to 450 bits. In 64-bit environments, the binary algorithm
891 appears to be advantageous for much longer moduli; here we use it
892 for moduli up to 2048 bits.
895 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
896 could not support the combine flag in choice fields.
899 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
900 extensions from a certificate request to the certificate.
903 *) Allow multiple 'certopt' and 'nameopt' options to be separated
904 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
905 file: this allows the display of the certificate about to be
906 signed to be customised, to allow certain fields to be included
907 or excluded and extension details. The old system didn't display
908 multicharacter strings properly, omitted fields not in the policy
909 and couldn't display additional details such as extensions.
912 *) Function EC_POINTs_mul for multiple scalar multiplication
913 of an arbitrary number of elliptic curve points
914 \sum scalars[i]*points[i],
915 optionally including the generator defined for the EC_GROUP:
916 scalar*generator + \sum scalars[i]*points[i].
918 EC_POINT_mul is a simple wrapper function for the typical case
919 that the point list has just one item (besides the optional
923 *) First EC_METHODs for curves over GF(p):
925 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
926 operations and provides various method functions that can also
927 operate with faster implementations of modular arithmetic.
929 EC_GFp_mont_method() reuses most functions that are part of
930 EC_GFp_simple_method, but uses Montgomery arithmetic.
932 [Bodo Moeller; point addition and point doubling
933 implementation directly derived from source code provided by
934 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
936 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
939 Curves are EC_GROUP objects (with an optional group generator)
940 based on EC_METHODs that are built into the library.
942 Points are EC_POINT objects based on EC_GROUP objects.
944 Most of the framework would be able to handle curves over arbitrary
945 finite fields, but as there are no obvious types for fields other
946 than GF(p), some functions are limited to that for now.
949 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
950 that the file contains a complete HTTP response.
953 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
954 change the def and num file printf format specifier from "%-40sXXX"
955 to "%-39s XXX". The latter will always guarantee a space after the
956 field while the former will cause them to run together if the field
957 is 40 of more characters long.
960 *) Constify the cipher and digest 'method' functions and structures
961 and modify related functions to take constant EVP_MD and EVP_CIPHER
965 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
966 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
969 *) Modify EVP_Digest*() routines so they now return values. Although the
970 internal software routines can never fail additional hardware versions
974 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
976 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
977 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
980 ERR_R_NESTED_ASN1_ERROR
982 ERR_R_MISSING_ASN1_EOS
983 were 4 .. 9, conflicting with
984 ERR_LIB_RSA (= ERR_R_RSA_LIB)
986 ERR_LIB_PEM (= ERR_R_PEM_LIB).
987 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
989 Add new error code 'ERR_R_INTERNAL_ERROR'.
992 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
996 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
997 sets the subject name for a new request or supersedes the
998 subject name in a given request. Formats that can be parsed are
999 'CN=Some Name, OU=myOU, C=IT'
1001 'CN=Some Name/OU=myOU/C=IT'.
1003 Add options '-batch' and '-verbose' to 'openssl req'.
1004 [Massimiliano Pala <madwolf@hackmasters.net>]
1006 *) Introduce the possibility to access global variables through
1007 functions on platform were that's the best way to handle exporting
1008 global variables in shared libraries. To enable this functionality,
1009 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1010 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1011 is normally done by Configure or something similar).
1013 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1014 in the source file (foo.c) like this:
1016 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1017 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1019 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1020 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1022 OPENSSL_DECLARE_GLOBAL(int,foo);
1023 #define foo OPENSSL_GLOBAL_REF(foo)
1024 OPENSSL_DECLARE_GLOBAL(double,bar);
1025 #define bar OPENSSL_GLOBAL_REF(bar)
1027 The #defines are very important, and therefore so is including the
1028 header file everywhere where the defined globals are used.
1030 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1031 of ASN.1 items, but that structure is a bit different.
1033 The largest change is in util/mkdef.pl which has been enhanced with
1034 better and easier to understand logic to choose which symbols should
1035 go into the Windows .def files as well as a number of fixes and code
1036 cleanup (among others, algorithm keywords are now sorted
1037 lexicographically to avoid constant rewrites).
1040 *) In BN_div() keep a copy of the sign of 'num' before writing the
1041 result to 'rm' because if rm==num the value will be overwritten
1042 and produce the wrong result if 'num' is negative: this caused
1043 problems with BN_mod() and BN_nnmod().
1046 *) Function OCSP_request_verify(). This checks the signature on an
1047 OCSP request and verifies the signer certificate. The signer
1048 certificate is just checked for a generic purpose and OCSP request
1052 *) Add OCSP_check_validity() function to check the validity of OCSP
1053 responses. OCSP responses are prepared in real time and may only
1054 be a few seconds old. Simply checking that the current time lies
1055 between thisUpdate and nextUpdate max reject otherwise valid responses
1056 caused by either OCSP responder or client clock inaccuracy. Instead
1057 we allow thisUpdate and nextUpdate to fall within a certain period of
1058 the current time. The age of the response can also optionally be
1059 checked. Two new options -validity_period and -status_age added to
1063 *) If signature or public key algorithm is unrecognized print out its
1064 OID rather that just UNKNOWN.
1067 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1068 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1069 ID to be generated from the issuer certificate alone which can then be
1070 passed to OCSP_id_issuer_cmp().
1073 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1074 ASN1 modules to export functions returning ASN1_ITEM pointers
1075 instead of the ASN1_ITEM structures themselves. This adds several
1076 new macros which allow the underlying ASN1 function/structure to
1077 be accessed transparently. As a result code should not use ASN1_ITEM
1078 references directly (such as &X509_it) but instead use the relevant
1079 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1080 use of the new ASN1 code on platforms where exporting structures
1081 is problematical (for example in shared libraries) but exporting
1082 functions returning pointers to structures is not.
1085 *) Add support for overriding the generation of SSL/TLS session IDs.
1086 These callbacks can be registered either in an SSL_CTX or per SSL.
1087 The purpose of this is to allow applications to control, if they wish,
1088 the arbitrary values chosen for use as session IDs, particularly as it
1089 can be useful for session caching in multiple-server environments. A
1090 command-line switch for testing this (and any client code that wishes
1091 to use such a feature) has been added to "s_server".
1092 [Geoff Thorpe, Lutz Jaenicke]
1094 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1095 of the form '#if defined(...) || defined(...) || ...' and
1096 '#if !defined(...) && !defined(...) && ...'. This also avoids
1097 the growing number of special cases it was previously handling.
1100 *) Make all configuration macros available for application by making
1101 sure they are available in opensslconf.h, by giving them names starting
1102 with "OPENSSL_" to avoid conflicts with other packages and by making
1103 sure e_os2.h will cover all platform-specific cases together with
1105 Additionally, it is now possible to define configuration/platform-
1106 specific names (called "system identities"). In the C code, these
1107 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1108 macro with the name beginning with "OPENSSL_SYS_", which is determined
1109 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1113 *) New option -set_serial to 'req' and 'x509' this allows the serial
1114 number to use to be specified on the command line. Previously self
1115 signed certificates were hard coded with serial number 0 and the
1116 CA options of 'x509' had to use a serial number in a file which was
1120 *) New options to 'ca' utility to support V2 CRL entry extensions.
1121 Currently CRL reason, invalidity date and hold instruction are
1122 supported. Add new CRL extensions to V3 code and some new objects.
1125 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1126 disable standard block padding (aka PKCS#5 padding) in the EVP
1127 API, which was previously mandatory. This means that the data is
1128 not padded in any way and so the total length much be a multiple
1129 of the block size, otherwise an error occurs.
1132 *) Initial (incomplete) OCSP SSL support.
1135 *) New function OCSP_parse_url(). This splits up a URL into its host,
1136 port and path components: primarily to parse OCSP URLs. New -url
1137 option to ocsp utility.
1140 *) New nonce behavior. The return value of OCSP_check_nonce() now
1141 reflects the various checks performed. Applications can decide
1142 whether to tolerate certain situations such as an absent nonce
1143 in a response when one was present in a request: the ocsp application
1144 just prints out a warning. New function OCSP_add1_basic_nonce()
1145 this is to allow responders to include a nonce in a response even if
1146 the request is nonce-less.
1149 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1150 skipped when using openssl x509 multiple times on a single input file,
1151 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1154 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1155 set string type: to handle setting ASN1_TIME structures. Fix ca
1156 utility to correctly initialize revocation date of CRLs.
1159 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1160 the clients preferred ciphersuites and rather use its own preferences.
1161 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1162 Internet Explorer by ensuring unchanged hash method during stepup.
1163 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1166 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1167 to aes and add a new 'exist' option to print out symbols that don't
1171 *) Additional options to ocsp utility to allow flags to be set and
1172 additional certificates supplied.
1175 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1176 OCSP client a number of certificate to only verify the response
1180 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1181 handle the new API. Currently only ECB, CBC modes supported. Add new
1184 Add TLS AES ciphersuites as described in RFC3268, "Advanced
1185 Encryption Standard (AES) Ciphersuites for Transport Layer
1186 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
1187 not enabled by default and were not part of the "ALL" ciphersuite
1188 alias because they were not yet official; they could be
1189 explicitly requested by specifying the "AESdraft" ciphersuite
1190 group alias. In the final release of OpenSSL 0.9.7, the group
1191 alias is called "AES" and is part of "ALL".)
1192 [Ben Laurie, Steve Henson, Bodo Moeller]
1194 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1195 request to response.
1198 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1199 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1200 extract information from a certificate request. OCSP_response_create()
1201 creates a response and optionally adds a basic response structure.
1202 OCSP_basic_add1_status() adds a complete single response to a basic
1203 response and returns the OCSP_SINGLERESP structure just added (to allow
1204 extensions to be included for example). OCSP_basic_add1_cert() adds a
1205 certificate to a basic response and OCSP_basic_sign() signs a basic
1206 response with various flags. New helper functions ASN1_TIME_check()
1207 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1208 (converts ASN1_TIME to GeneralizedTime).
1211 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1212 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1213 structure from a certificate. X509_pubkey_digest() digests the public_key
1214 contents: this is used in various key identifiers.
1217 *) Make sk_sort() tolerate a NULL argument.
1218 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1220 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1221 passed by the function are trusted implicitly. If any of them signed the
1222 response then it is assumed to be valid and is not verified.
1225 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1226 to data. This was previously part of the PKCS7 ASN1 code. This
1227 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1228 [Steve Henson, reported by Kenneth R. Robinette
1229 <support@securenetterm.com>]
1231 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1232 routines: without these tracing memory leaks is very painful.
1233 Fix leaks in PKCS12 and PKCS7 routines.
1236 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1237 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1238 effectively meant GeneralizedTime would never be used. Now it
1239 is initialised to -1 but X509_time_adj() now has to check the value
1240 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1241 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1242 [Steve Henson, reported by Kenneth R. Robinette
1243 <support@securenetterm.com>]
1245 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1246 result in a zero length in the ASN1_INTEGER structure which was
1247 not consistent with the structure when d2i_ASN1_INTEGER() was used
1248 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1249 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1250 where it did not print out a minus for negative ASN1_INTEGER.
1253 *) Add summary printout to ocsp utility. The various functions which
1254 convert status values to strings have been renamed to:
1255 OCSP_response_status_str(), OCSP_cert_status_str() and
1256 OCSP_crl_reason_str() and are no longer static. New options
1257 to verify nonce values and to disable verification. OCSP response
1258 printout format cleaned up.
1261 *) Add additional OCSP certificate checks. These are those specified
1262 in RFC2560. This consists of two separate checks: the CA of the
1263 certificate being checked must either be the OCSP signer certificate
1264 or the issuer of the OCSP signer certificate. In the latter case the
1265 OCSP signer certificate must contain the OCSP signing extended key
1266 usage. This check is performed by attempting to match the OCSP
1267 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1268 in the OCSP_CERTID structures of the response.
1271 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1272 and related routines. This uses the standard OpenSSL certificate
1273 verify routines to perform initial checks (just CA validity) and
1274 to obtain the certificate chain. Then additional checks will be
1275 performed on the chain. Currently the root CA is checked to see
1276 if it is explicitly trusted for OCSP signing. This is used to set
1277 a root CA as a global signing root: that is any certificate that
1278 chains to that CA is an acceptable OCSP signing certificate.
1281 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1282 extensions from a separate configuration file.
1283 As when reading extensions from the main configuration file,
1284 the '-extensions ...' option may be used for specifying the
1286 [Massimiliano Pala <madwolf@comune.modena.it>]
1288 *) New OCSP utility. Allows OCSP requests to be generated or
1289 read. The request can be sent to a responder and the output
1290 parsed, outputed or printed in text form. Not complete yet:
1291 still needs to check the OCSP response validity.
1294 *) New subcommands for 'openssl ca':
1295 'openssl ca -status <serial>' prints the status of the cert with
1296 the given serial number (according to the index file).
1297 'openssl ca -updatedb' updates the expiry status of certificates
1299 [Massimiliano Pala <madwolf@comune.modena.it>]
1301 *) New '-newreq-nodes' command option to CA.pl. This is like
1302 '-newreq', but calls 'openssl req' with the '-nodes' option
1303 so that the resulting key is not encrypted.
1304 [Damien Miller <djm@mindrot.org>]
1306 *) New configuration for the GNU Hurd.
1307 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1309 *) Initial code to implement OCSP basic response verify. This
1310 is currently incomplete. Currently just finds the signer's
1311 certificate and verifies the signature on the response.
1314 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1315 value of OPENSSLDIR. This is available via the new '-d' option
1316 to 'openssl version', and is also included in 'openssl version -a'.
1319 *) Allowing defining memory allocation callbacks that will be given
1320 file name and line number information in additional arguments
1321 (a const char* and an int). The basic functionality remains, as
1322 well as the original possibility to just replace malloc(),
1323 realloc() and free() by functions that do not know about these
1324 additional arguments. To register and find out the current
1325 settings for extended allocation functions, the following
1326 functions are provided:
1328 CRYPTO_set_mem_ex_functions
1329 CRYPTO_set_locked_mem_ex_functions
1330 CRYPTO_get_mem_ex_functions
1331 CRYPTO_get_locked_mem_ex_functions
1333 These work the same way as CRYPTO_set_mem_functions and friends.
1334 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1335 extended allocation function is enabled.
1336 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1337 a conventional allocation function is enabled.
1338 [Richard Levitte, Bodo Moeller]
1340 *) Finish off removing the remaining LHASH function pointer casts.
1341 There should no longer be any prototype-casting required when using
1342 the LHASH abstraction, and any casts that remain are "bugs". See
1343 the callback types and macros at the head of lhash.h for details
1344 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1347 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1348 If /dev/[u]random devices are not available or do not return enough
1349 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1351 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1352 /etc/entropy will be queried once each in this sequence, quering stops
1353 when enough entropy was collected without querying more sockets.
1356 *) Change the Unix RAND_poll() variant to be able to poll several
1357 random devices, as specified by DEVRANDOM, until a sufficient amount
1358 of data has been collected. We spend at most 10 ms on each file
1359 (select timeout) and read in non-blocking mode. DEVRANDOM now
1360 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1361 (previously it was just the string "/dev/urandom"), so on typical
1362 platforms the 10 ms delay will never occur.
1363 Also separate out the Unix variant to its own file, rand_unix.c.
1364 For VMS, there's a currently-empty rand_vms.c.
1367 *) Move OCSP client related routines to ocsp_cl.c. These
1368 provide utility functions which an application needing
1369 to issue a request to an OCSP responder and analyse the
1370 response will typically need: as opposed to those which an
1371 OCSP responder itself would need which will be added later.
1373 OCSP_request_sign() signs an OCSP request with an API similar
1374 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1375 response. OCSP_response_get1_basic() extracts basic response
1376 from response. OCSP_resp_find_status(): finds and extracts status
1377 information from an OCSP_CERTID structure (which will be created
1378 when the request structure is built). These are built from lower
1379 level functions which work on OCSP_SINGLERESP structures but
1380 wont normally be used unless the application wishes to examine
1381 extensions in the OCSP response for example.
1383 Replace nonce routines with a pair of functions.
1384 OCSP_request_add1_nonce() adds a nonce value and optionally
1385 generates a random value. OCSP_check_nonce() checks the
1386 validity of the nonce in an OCSP response.
1389 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1390 This doesn't copy the supplied OCSP_CERTID and avoids the
1391 need to free up the newly created id. Change return type
1392 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1393 This can then be used to add extensions to the request.
1394 Deleted OCSP_request_new(), since most of its functionality
1395 is now in OCSP_REQUEST_new() (and the case insensitive name
1396 clash) apart from the ability to set the request name which
1397 will be added elsewhere.
1400 *) Update OCSP API. Remove obsolete extensions argument from
1401 various functions. Extensions are now handled using the new
1402 OCSP extension code. New simple OCSP HTTP function which
1403 can be used to send requests and parse the response.
1406 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1407 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1408 uses the special reorder version of SET OF to sort the attributes
1409 and reorder them to match the encoded order. This resolves a long
1410 standing problem: a verify on a PKCS7 structure just after signing
1411 it used to fail because the attribute order did not match the
1412 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1413 it uses the received order. This is necessary to tolerate some broken
1414 software that does not order SET OF. This is handled by encoding
1415 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1416 to produce the required SET OF.
1419 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1420 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1421 files to get correct declarations of the ASN.1 item variables.
1424 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1425 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1426 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1427 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1428 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1429 ASN1_ITEM and no wrapper functions.
1432 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1433 replace the old function pointer based I/O routines. Change most of
1434 the *_d2i_bio() and *_d2i_fp() functions to use these.
1437 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1438 lines, recognice more "algorithms" that can be deselected, and make
1439 it complain about algorithm deselection that isn't recognised.
1442 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1443 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1444 to use new functions. Add NO_ASN1_OLD which can be set to remove
1445 some old style ASN1 functions: this can be used to determine if old
1446 code will still work when these eventually go away.
1449 *) New extension functions for OCSP structures, these follow the
1450 same conventions as certificates and CRLs.
1453 *) New function X509V3_add1_i2d(). This automatically encodes and
1454 adds an extension. Its behaviour can be customised with various
1455 flags to append, replace or delete. Various wrappers added for
1456 certifcates and CRLs.
1459 *) Fix to avoid calling the underlying ASN1 print routine when
1460 an extension cannot be parsed. Correct a typo in the
1461 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1464 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1465 entries for variables.
1468 *) Add functionality to apps/openssl.c for detecting locking
1469 problems: As the program is single-threaded, all we have
1470 to do is register a locking callback using an array for
1471 storing which locks are currently held by the program.
1474 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1475 SSL_get_ex_data_X509_STORE_idx(), which is used in
1476 ssl_verify_cert_chain() and thus can be called at any time
1477 during TLS/SSL handshakes so that thread-safety is essential.
1478 Unfortunately, the ex_data design is not at all suited
1479 for multi-threaded use, so it probably should be abolished.
1482 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1483 [Broadcom, tweaked and integrated by Geoff Thorpe]
1485 *) Move common extension printing code to new function
1486 X509V3_print_extensions(). Reorganise OCSP print routines and
1487 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1490 *) New function X509_signature_print() to remove duplication in some
1494 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1495 set (this was treated exactly the same as SET OF previously). This
1496 is used to reorder the STACK representing the structure to match the
1497 encoding. This will be used to get round a problem where a PKCS7
1498 structure which was signed could not be verified because the STACK
1499 order did not reflect the encoded order.
1502 *) Reimplement the OCSP ASN1 module using the new code.
1505 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1506 for its ASN1 operations. The old style function pointers still exist
1507 for now but they will eventually go away.
1510 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1511 completely replaces the old ASN1 functionality with a table driven
1512 encoder and decoder which interprets an ASN1_ITEM structure describing
1513 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1514 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1515 has also been converted to the new form.
1518 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1519 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1520 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1521 for negative moduli.
1524 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1525 of not touching the result's sign bit.
1528 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1532 *) Changed the LHASH code to use prototypes for callbacks, and created
1533 macros to declare and implement thin (optionally static) functions
1534 that provide type-safety and avoid function pointer casting for the
1535 type-specific callbacks.
1538 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1540 [Veers Staats <staatsvr@asc.hpc.mil>,
1541 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1543 *) Reformat the FAQ so the different questions and answers can be divided
1544 in sections depending on the subject.
1547 *) Have the zlib compression code load ZLIB.DLL dynamically under
1551 *) New function BN_mod_sqrt for computing square roots modulo a prime
1552 (using the probabilistic Tonelli-Shanks algorithm unless
1553 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1554 be handled deterministically).
1555 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1557 *) Make BN_mod_inverse faster by explicitly handling small quotients
1558 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1559 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1562 *) New function BN_kronecker.
1565 *) Fix BN_gcd so that it works on negative inputs; the result is
1566 positive unless both parameters are zero.
1567 Previously something reasonably close to an infinite loop was
1568 possible because numbers could be growing instead of shrinking
1569 in the implementation of Euclid's algorithm.
1572 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1573 sign of the number in question.
1575 Fix BN_is_word(a,w) to work correctly for w == 0.
1577 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1578 because its test if the absolute value of 'a' equals 'w'.
1579 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1580 it exists mostly for use in the implementations of BN_is_zero(),
1581 BN_is_one(), and BN_is_word().
1584 *) New function BN_swap.
1587 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1588 the exponentiation functions are more likely to produce reasonable
1589 results on negative inputs.
1592 *) Change BN_mod_mul so that the result is always non-negative.
1593 Previously, it could be negative if one of the factors was negative;
1594 I don't think anyone really wanted that behaviour.
1597 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1598 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1599 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1600 and add new functions:
1609 BN_mod_lshift1_quick
1613 These functions always generate non-negative results.
1615 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1616 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1618 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1619 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1620 be reduced modulo m.
1621 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1623 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1624 was actually never needed) and in BN_mul(). The removal in BN_mul()
1625 required a small change in bn_mul_part_recursive() and the addition
1626 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1627 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1628 bn_sub_words() and bn_add_words() except they take arrays with
1632 *) In 'openssl passwd', verify passwords read from the terminal
1633 unless the '-salt' option is used (which usually means that
1634 verification would just waste user's time since the resulting
1635 hash is going to be compared with some given password hash)
1636 or the new '-noverify' option is used.
1638 This is an incompatible change, but it does not affect
1639 non-interactive use of 'openssl passwd' (passwords on the command
1640 line, '-stdin' option, '-in ...' option) and thus should not
1644 *) Remove all references to RSAref, since there's no more need for it.
1647 *) Make DSO load along a path given through an environment variable
1648 (SHLIB_PATH) with shl_load().
1651 *) Constify the ENGINE code as a result of BIGNUM constification.
1652 Also constify the RSA code and most things related to it. In a
1653 few places, most notable in the depth of the ASN.1 code, ugly
1654 casts back to non-const were required (to be solved at a later
1658 *) Make it so the openssl application has all engines loaded by default.
1661 *) Constify the BIGNUM routines a little more.
1664 *) Add the following functions:
1666 ENGINE_load_cswift()
1668 ENGINE_load_atalla()
1670 ENGINE_load_builtin_engines()
1672 That way, an application can itself choose if external engines that
1673 are built-in in OpenSSL shall ever be used or not. The benefit is
1674 that applications won't have to be linked with libdl or other dso
1675 libraries unless it's really needed.
1677 Changed 'openssl engine' to load all engines on demand.
1678 Changed the engine header files to avoid the duplication of some
1679 declarations (they differed!).
1682 *) 'openssl engine' can now list capabilities.
1685 *) Better error reporting in 'openssl engine'.
1688 *) Never call load_dh_param(NULL) in s_server.
1691 *) Add engine application. It can currently list engines by name and
1692 identity, and test if they are actually available.
1695 *) Improve RPM specification file by forcing symbolic linking and making
1696 sure the installed documentation is also owned by root.root.
1697 [Damien Miller <djm@mindrot.org>]
1699 *) Give the OpenSSL applications more possibilities to make use of
1700 keys (public as well as private) handled by engines.
1703 *) Add OCSP code that comes from CertCo.
1706 *) Add VMS support for the Rijndael code.
1709 *) Added untested support for Nuron crypto accelerator.
1712 *) Add support for external cryptographic devices. This code was
1713 previously distributed separately as the "engine" branch.
1714 [Geoff Thorpe, Richard Levitte]
1716 *) Rework the filename-translation in the DSO code. It is now possible to
1717 have far greater control over how a "name" is turned into a filename
1718 depending on the operating environment and any oddities about the
1719 different shared library filenames on each system.
1722 *) Support threads on FreeBSD-elf in Configure.
1725 *) Fix for SHA1 assembly problem with MASM: it produces
1726 warnings about corrupt line number information when assembling
1727 with debugging information. This is caused by the overlapping
1729 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1732 NCONF_get_number() has no error checking at all. As a replacement,
1733 NCONF_get_number_e() is defined (_e for "error checking") and is
1734 promoted strongly. The old NCONF_get_number is kept around for
1735 binary backward compatibility.
1736 Make it possible for methods to load from something other than a BIO,
1737 by providing a function pointer that is given a name instead of a BIO.
1738 For example, this could be used to load configuration data from an
1742 *) Fix for non blocking accept BIOs. Added new I/O special reason
1743 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1744 with non blocking I/O was not possible because no retry code was
1745 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1749 *) Added the beginnings of Rijndael support.
1752 *) Fix for bug in DirectoryString mask setting. Add support for
1753 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1754 to allow certificate printing to more controllable, additional
1755 'certopt' option to 'x509' to allow new printing options to be
1759 *) Clean old EAY MD5 hack from e_os.h.
1762 Changes between 0.9.6d and 0.9.6e [XX xxx XXXX]
1764 *) Fix cipher selection routines: ciphers without encryption had no flags
1765 for the cipher strength set and where therefore not handled correctly
1766 by the selection routines (PR #130).
1769 *) Fix EVP_dsa_sha macro.
1773 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
1774 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
1775 that was added in OpenSSL 0.9.6d.
1777 As the countermeasure turned out to be incompatible with some
1778 broken SSL implementations, the new option is part of SSL_OP_ALL.
1779 SSL_OP_ALL is usually employed when compatibility with weird SSL
1780 implementations is desired (e.g. '-bugs' option to 's_client' and
1781 's_server'), so the new option is automatically set in many
1785 Changes between 0.9.6c and 0.9.6d [9 May 2002]
1787 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
1788 encoded as NULL) with id-dsa-with-sha1.
1789 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
1791 *) Check various X509_...() return values in apps/req.c.
1792 [Nils Larsch <nla@trustcenter.de>]
1794 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
1795 an end-of-file condition would erronously be flagged, when the CRLF
1796 was just at the end of a processed block. The bug was discovered when
1797 processing data through a buffering memory BIO handing the data to a
1798 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
1799 <ptsekov@syntrex.com> and Nedelcho Stanev.
1802 *) Implement a countermeasure against a vulnerability recently found
1803 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
1804 before application data chunks to avoid the use of known IVs
1805 with data potentially chosen by the attacker.
1808 *) Fix length checks in ssl3_get_client_hello().
1811 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
1812 to prevent ssl3_read_internal() from incorrectly assuming that
1813 ssl3_read_bytes() found application data while handshake
1814 processing was enabled when in fact s->s3->in_read_app_data was
1815 merely automatically cleared during the initial handshake.
1816 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
1818 *) Fix object definitions for Private and Enterprise: they were not
1819 recognized in their shortname (=lowercase) representation. Extend
1820 obj_dat.pl to issue an error when using undefined keywords instead
1821 of silently ignoring the problem (Svenning Sorensen
1822 <sss@sss.dnsalias.net>).
1825 *) Fix DH_generate_parameters() so that it works for 'non-standard'
1826 generators, i.e. generators other than 2 and 5. (Previously, the
1827 code did not properly initialise the 'add' and 'rem' values to
1828 BN_generate_prime().)
1830 In the new general case, we do not insist that 'generator' is
1831 actually a primitive root: This requirement is rather pointless;
1832 a generator of the order-q subgroup is just as good, if not
1836 *) Map new X509 verification errors to alerts. Discovered and submitted by
1837 Tom Wu <tom@arcot.com>.
1840 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
1841 returning non-zero before the data has been completely received
1842 when using non-blocking I/O.
1843 [Bodo Moeller; problem pointed out by John Hughes]
1845 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
1846 [Ben Laurie, Lutz Jaenicke]
1848 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
1849 Yoram Zahavi <YoramZ@gilian.com>).
1852 *) Add information about CygWin 1.3 and on, and preserve proper
1853 configuration for the versions before that.
1854 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1856 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
1857 check whether we deal with a copy of a session and do not delete from
1858 the cache in this case. Problem reported by "Izhar Shoshani Levi"
1859 <izhar@checkpoint.com>.
1862 *) Do not store session data into the internal session cache, if it
1863 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
1864 flag is set). Proposed by Aslam <aslam@funk.com>.
1867 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
1871 *) Add the configuration target linux-s390x.
1872 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
1874 *) [In 0.9.6d-engine release:]
1875 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1876 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1878 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
1879 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
1880 variable as an indication that a ClientHello message has been
1881 received. As the flag value will be lost between multiple
1882 invocations of ssl3_accept when using non-blocking I/O, the
1883 function may not be aware that a handshake has actually taken
1884 place, thus preventing a new session from being added to the
1887 To avoid this problem, we now set s->new_session to 2 instead of
1888 using a local variable.
1889 [Lutz Jaenicke, Bodo Moeller]
1891 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
1892 if the SSL_R_LENGTH_MISMATCH error is detected.
1893 [Geoff Thorpe, Bodo Moeller]
1895 *) New 'shared_ldflag' column in Configure platform table.
1898 *) Fix EVP_CIPHER_mode macro.
1899 ["Dan S. Camper" <dan@bti.net>]
1901 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
1902 type, we must throw them away by setting rr->length to 0.
1903 [D P Chang <dpc@qualys.com>]
1905 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
1907 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
1908 <Dominikus.Scherkl@biodata.com>. (The previous implementation
1909 worked incorrectly for those cases where range = 10..._2 and
1910 3*range is two bits longer than range.)
1913 *) Only add signing time to PKCS7 structures if it is not already
1917 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
1918 OBJ_ld_ce should be OBJ_id_ce.
1919 Also some ip-pda OIDs in crypto/objects/objects.txt were
1920 incorrect (cf. RFC 3039).
1921 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
1923 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
1924 returns early because it has nothing to do.
1925 [Andy Schneider <andy.schneider@bjss.co.uk>]
1927 *) [In 0.9.6c-engine release:]
1928 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
1929 [Andy Schneider <andy.schneider@bjss.co.uk>]
1931 *) [In 0.9.6c-engine release:]
1932 Add support for Cryptographic Appliance's keyserver technology.
1933 (Use engine 'keyclient')
1934 [Cryptographic Appliances and Geoff Thorpe]
1936 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
1937 is called via tools/c89.sh because arguments have to be
1938 rearranged (all '-L' options must appear before the first object
1940 [Richard Shapiro <rshapiro@abinitio.com>]
1942 *) [In 0.9.6c-engine release:]
1943 Add support for Broadcom crypto accelerator cards, backported
1945 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
1947 *) [In 0.9.6c-engine release:]
1948 Add support for SureWare crypto accelerator cards from
1949 Baltimore Technologies. (Use engine 'sureware')
1950 [Baltimore Technologies and Mark Cox]
1952 *) [In 0.9.6c-engine release:]
1953 Add support for crypto accelerator cards from Accelerated
1954 Encryption Processing, www.aep.ie. (Use engine 'aep')
1955 [AEP Inc. and Mark Cox]
1957 *) Add a configuration entry for gcc on UnixWare.
1958 [Gary Benson <gbenson@redhat.com>]
1960 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
1961 messages are stored in a single piece (fixed-length part and
1962 variable-length part combined) and fix various bugs found on the way.
1965 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
1966 instead. BIO_gethostbyname() does not know what timeouts are
1967 appropriate, so entries would stay in cache even when they have
1969 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
1971 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
1972 faced with a pathologically small ClientHello fragment that does
1973 not contain client_version: Instead of aborting with an error,
1974 simply choose the highest available protocol version (i.e.,
1975 TLS 1.0 unless it is disabled). In practice, ClientHello
1976 messages are never sent like this, but this change gives us
1977 strictly correct behaviour at least for TLS.
1980 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
1981 never resets s->method to s->ctx->method when called from within
1982 one of the SSL handshake functions.
1983 [Bodo Moeller; problem pointed out by Niko Baric]
1985 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
1986 (sent using the client's version number) if client_version is
1987 smaller than the protocol version in use. Also change
1988 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
1989 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
1990 the client will at least see that alert.
1993 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
1997 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
1998 client receives HelloRequest while in a handshake.
1999 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
2001 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
2002 should end in 'break', not 'goto end' which circuments various
2003 cleanups done in state SSL_ST_OK. But session related stuff
2004 must be disabled for SSL_ST_OK in the case that we just sent a
2007 Also avoid some overhead by not calling ssl_init_wbio_buffer()
2008 before just sending a HelloRequest.
2009 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
2011 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
2012 reveal whether illegal block cipher padding was found or a MAC
2013 verification error occured. (Neither SSLerr() codes nor alerts
2014 are directly visible to potential attackers, but the information
2015 may leak via logfiles.)
2017 Similar changes are not required for the SSL 2.0 implementation
2018 because the number of padding bytes is sent in clear for SSL 2.0,
2019 and the extra bytes are just ignored. However ssl/s2_pkt.c
2020 failed to verify that the purported number of padding bytes is in
2024 *) Add OpenUNIX-8 support including shared libraries
2025 (Boyd Lynn Gerber <gerberb@zenez.com>).
2028 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
2029 'wristwatch attack' using huge encoding parameters (cf.
2030 James H. Manger's CRYPTO 2001 paper). Note that the
2031 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
2032 encoding parameters and hence was not vulnerable.
2035 *) BN_sqr() bug fix.
2036 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
2038 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
2039 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
2040 followed by modular reduction.
2041 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
2043 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
2044 equivalent based on BN_pseudo_rand() instead of BN_rand().
2047 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
2048 This function was broken, as the check for a new client hello message
2049 to handle SGC did not allow these large messages.
2050 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
2053 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
2056 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
2057 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
2060 *) Rework the configuration and shared library support for Tru64 Unix.
2061 The configuration part makes use of modern compiler features and
2062 still retains old compiler behavior for those that run older versions
2063 of the OS. The shared library support part includes a variant that
2064 uses the RPATH feature, and is available through the special
2065 configuration target "alpha-cc-rpath", which will never be selected
2067 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
2069 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
2070 with the same message size as in ssl3_get_certificate_request().
2071 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
2072 messages might inadvertently be reject as too long.
2073 [Petr Lampa <lampa@fee.vutbr.cz>]
2075 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
2078 *) Modified SSL library such that the verify_callback that has been set
2079 specificly for an SSL object with SSL_set_verify() is actually being
2080 used. Before the change, a verify_callback set with this function was
2081 ignored and the verify_callback() set in the SSL_CTX at the time of
2082 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
2083 to allow the necessary settings.
2086 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2087 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2088 done automatically (in contradiction to the requirements of the C
2089 standard). This made problems when used from OpenSSH.
2092 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2093 dh->length and always used
2095 BN_rand_range(priv_key, dh->p).
2097 BN_rand_range() is not necessary for Diffie-Hellman, and this
2098 specific range makes Diffie-Hellman unnecessarily inefficient if
2099 dh->length (recommended exponent length) is much smaller than the
2100 length of dh->p. We could use BN_rand_range() if the order of
2101 the subgroup was stored in the DH structure, but we only have
2106 BN_rand(priv_key, l, ...)
2108 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2114 RSA_eay_public_encrypt
2115 RSA_eay_private_decrypt
2116 RSA_eay_private_encrypt (signing)
2117 RSA_eay_public_decrypt (signature verification)
2119 (default implementations for RSA_public_encrypt,
2120 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2121 always reject numbers >= n.
2124 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2125 to synchronize access to 'locking_thread'. This is necessary on
2126 systems where access to 'locking_thread' (an 'unsigned long'
2127 variable) is not atomic.
2130 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2131 *before* setting the 'crypto_lock_rand' flag. The previous code had
2132 a race condition if 0 is a valid thread ID.
2133 [Travis Vitek <vitek@roguewave.com>]
2135 *) Add support for shared libraries under Irix.
2136 [Albert Chin-A-Young <china@thewrittenword.com>]
2138 *) Add configuration option to build on Linux on both big-endian and
2140 [Ralf Baechle <ralf@uni-koblenz.de>]
2142 *) Add the possibility to create shared libraries on HP-UX.
2145 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2147 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2148 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2149 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2150 PRNG state recovery was possible based on the output of
2151 one PRNG request appropriately sized to gain knowledge on
2152 'md' followed by enough consecutive 1-byte PRNG requests
2153 to traverse all of 'state'.
2155 1. When updating 'md_local' (the current thread's copy of 'md')
2156 during PRNG output generation, hash all of the previous
2157 'md_local' value, not just the half used for PRNG output.
2159 2. Make the number of bytes from 'state' included into the hash
2160 independent from the number of PRNG bytes requested.
2162 The first measure alone would be sufficient to avoid
2163 Markku-Juhani's attack. (Actually it had never occurred
2164 to me that the half of 'md_local' used for chaining was the
2165 half from which PRNG output bytes were taken -- I had always
2166 assumed that the secret half would be used.) The second
2167 measure makes sure that additional data from 'state' is never
2168 mixed into 'md_local' in small portions; this heuristically
2169 further strengthens the PRNG.
2172 *) Fix crypto/bn/asm/mips3.s.
2175 *) When only the key is given to "enc", the IV is undefined. Print out
2176 an error message in this case.
2179 *) Handle special case when X509_NAME is empty in X509 printing routines.
2182 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2183 positive and less than q.
2186 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2187 used: it isn't thread safe and the add_lock_callback should handle
2189 [Paul Rose <Paul.Rose@bridge.com>]
2191 *) Verify that incoming data obeys the block size in
2192 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2196 [Ulf Möller, Bodo Möller]
2198 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2199 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2200 when fixing the server behaviour for backwards-compatible 'client
2201 hello' messages. (Note that the attack is impractical against
2202 SSL 3.0 and TLS 1.0 anyway because length and version checking
2203 means that the probability of guessing a valid ciphertext is
2204 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2207 Before 0.9.5, the countermeasure (hide the error by generating a
2208 random 'decryption result') did not work properly because
2209 ERR_clear_error() was missing, meaning that SSL_get_error() would
2210 detect the supposedly ignored error.
2212 Both problems are now fixed.
2215 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2216 (previously it was 1024).
2219 *) Fix for compatibility mode trust settings: ignore trust settings
2220 unless some valid trust or reject settings are present.
2223 *) Fix for blowfish EVP: its a variable length cipher.
2226 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2227 parameters in DSA public key structures and return an error in the
2228 DSA routines if parameters are absent.
2231 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2232 in the current directory if neither $RANDFILE nor $HOME was set.
2233 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2234 caused some confusion to Windows users who haven't defined $HOME.
2235 Thus RAND_file_name() is changed again: e_os.h can define a
2236 DEFAULT_HOME, which will be used if $HOME is not set.
2237 For Windows, we use "C:"; on other platforms, we still require
2238 environment variables.
2240 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2241 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2242 having multiple threads call RAND_poll() concurrently.
2245 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2246 combination of a flag and a thread ID variable.
2247 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2248 flag), *other* threads can enter ssleay_add_bytes without obeying
2249 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2250 that they do not hold after the first thread unsets add_do_not_lock).
2253 *) Change bctest again: '-x' expressions are not available in all
2257 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2259 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2260 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2262 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2263 the default extension for executables, if any. Also, make the perl
2264 scripts that use symlink() to test if it really exists and use "cp"
2265 if it doesn't. All this made OpenSSL compilable and installable in
2269 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2270 If SEQUENCE is length is indefinite just set c->slen to the total
2271 amount of data available.
2272 [Steve Henson, reported by shige@FreeBSD.org]
2273 [This change does not apply to 0.9.7.]
2275 *) Change bctest to avoid here-documents inside command substitution
2276 (workaround for FreeBSD /bin/sh bug).
2277 For compatibility with Ultrix, avoid shell functions (introduced
2278 in the bctest version that searches along $PATH).
2281 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2282 with des_encrypt() defined on some operating systems, like Solaris
2286 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2287 On the Importance of Eliminating Errors in Cryptographic
2288 Computations, J. Cryptology 14 (2001) 2, 101-119,
2289 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2292 *) MIPS assembler BIGNUM division bug fix.
2295 *) Disabled incorrect Alpha assembler code.
2298 *) Fix PKCS#7 decode routines so they correctly update the length
2299 after reading an EOC for the EXPLICIT tag.
2301 [This change does not apply to 0.9.7.]
2303 *) Fix bug in PKCS#12 key generation routines. This was triggered
2304 if a 3DES key was generated with a 0 initial byte. Include
2305 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2306 (but broken) behaviour.
2309 *) Enhance bctest to search for a working bc along $PATH and print
2311 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2313 *) Fix memory leaks in err.c: free err_data string if necessary;
2314 don't write to the wrong index in ERR_set_error_data.
2317 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2321 *) Replace rdtsc with _emit statements for VC++ version 5.
2322 [Jeremy Cooper <jeremy@baymoo.org>]
2324 *) Make it possible to reuse SSLv2 sessions.
2327 *) In copy_email() check for >= 0 as a return value for
2328 X509_NAME_get_index_by_NID() since 0 is a valid index.
2329 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2331 *) Avoid coredump with unsupported or invalid public keys by checking if
2332 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2333 PKCS7_verify() fails with non detached data.
2336 *) Don't use getenv in library functions when run as setuid/setgid.
2337 New function OPENSSL_issetugid().
2340 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2341 due to incorrect handling of multi-threading:
2343 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2345 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2347 3. Count how many times MemCheck_off() has been called so that
2348 nested use can be treated correctly. This also avoids
2349 inband-signalling in the previous code (which relied on the
2350 assumption that thread ID 0 is impossible).
2353 *) Add "-rand" option also to s_client and s_server.
2356 *) Fix CPU detection on Irix 6.x.
2357 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2358 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2360 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2363 [This change does not apply to 0.9.7.]
2365 *) Use the cached encoding of an X509_NAME structure rather than
2366 copying it. This is apparently the reason for the libsafe "errors"
2367 but the code is actually correct.
2370 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2371 Bleichenbacher's DSA attack.
2372 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2373 to be set and top=0 forces the highest bit to be set; top=-1 is new
2374 and leaves the highest bit random.
2375 [Ulf Moeller, Bodo Moeller]
2377 *) In the NCONF_...-based implementations for CONF_... queries
2378 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2379 a temporary CONF structure with the data component set to NULL
2380 (which gives segmentation faults in lh_retrieve).
2381 Instead, use NULL for the CONF pointer in CONF_get_string and
2382 CONF_get_number (which may use environment variables) and directly
2383 return NULL from CONF_get_section.
2386 *) Fix potential buffer overrun for EBCDIC.
2389 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2390 keyUsage if basicConstraints absent for a CA.
2393 *) Make SMIME_write_PKCS7() write mail header values with a format that
2394 is more generally accepted (no spaces before the semicolon), since
2395 some programs can't parse those values properly otherwise. Also make
2396 sure BIO's that break lines after each write do not create invalid
2400 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2401 macros previously used would not encode an empty SEQUENCE OF
2402 and break the signature.
2404 [This change does not apply to 0.9.7.]
2406 *) Zero the premaster secret after deriving the master secret in
2410 *) Add some EVP_add_digest_alias registrations (as found in
2411 OpenSSL_add_all_digests()) to SSL_library_init()
2412 aka OpenSSL_add_ssl_algorithms(). This provides improved
2413 compatibility with peers using X.509 certificates
2414 with unconventional AlgorithmIdentifier OIDs.
2417 *) Fix for Irix with NO_ASM.
2418 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2420 *) ./config script fixes.
2421 [Ulf Moeller, Richard Levitte]
2423 *) Fix 'openssl passwd -1'.
2426 *) Change PKCS12_key_gen_asc() so it can cope with non null
2427 terminated strings whose length is passed in the passlen
2428 parameter, for example from PEM callbacks. This was done
2429 by adding an extra length parameter to asc2uni().
2430 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2432 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2433 call failed, free the DSA structure.
2436 *) Fix to uni2asc() to cope with zero length Unicode strings.
2437 These are present in some PKCS#12 files.
2440 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2441 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2442 when writing a 32767 byte record.
2443 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2445 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2446 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2448 (RSA objects have a reference count access to which is protected
2449 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2450 so they are meant to be shared between threads.)
2451 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2452 "Reddie, Steven" <Steven.Reddie@ca.com>]
2454 *) Fix a deadlock in CRYPTO_mem_leaks().
2457 *) Use better test patterns in bntest.
2460 *) rand_win.c fix for Borland C.
2463 *) BN_rshift bugfix for n == 0.
2466 *) Add a 'bctest' script that checks for some known 'bc' bugs
2467 so that 'make test' does not abort just because 'bc' is broken.
2470 *) Store verify_result within SSL_SESSION also for client side to
2471 avoid potential security hole. (Re-used sessions on the client side
2472 always resulted in verify_result==X509_V_OK, not using the original
2473 result of the server certificate verification.)
2476 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2477 SSL3_RT_APPLICATION_DATA, return 0.
2478 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2482 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2483 releases, have been re-implemented by renaming the previous
2484 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2485 and ssl3_read_internal, respectively, and adding 'peek' parameters
2486 to them. The new ssl[23]_{read,peek} functions are calls to
2487 ssl[23]_read_internal with the 'peek' flag set appropriately.
2488 A 'peek' parameter has also been added to ssl3_read_bytes, which
2489 does the actual work for ssl3_read_internal.
2492 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2493 the method-specific "init()" handler. Also clean up ex_data after
2494 calling the method-specific "finish()" handler. Previously, this was
2495 happening the other way round.
2498 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2499 The previous value, 12, was not always sufficient for BN_mod_exp().
2502 *) Make sure that shared libraries get the internal name engine with
2503 the full version number and not just 0. This should mark the
2504 shared libraries as not backward compatible. Of course, this should
2505 be changed again when we can guarantee backward binary compatibility.
2508 *) Fix typo in get_cert_by_subject() in by_dir.c
2509 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2511 *) Rework the system to generate shared libraries:
2513 - Make note of the expected extension for the shared libraries and
2514 if there is a need for symbolic links from for example libcrypto.so.0
2515 to libcrypto.so.0.9.7. There is extended info in Configure for
2518 - Make as few rebuilds of the shared libraries as possible.
2520 - Still avoid linking the OpenSSL programs with the shared libraries.
2522 - When installing, install the shared libraries separately from the
2526 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2528 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2529 and not in SSL_clear because the latter is also used by the
2530 accept/connect functions; previously, the settings made by
2531 SSL_set_read_ahead would be lost during the handshake.
2532 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2534 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2535 Previously, it would create entries for disableed algorithms no
2539 *) Added several new manual pages for SSL_* function.
2542 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2544 *) In ssl23_get_client_hello, generate an error message when faced
2545 with an initial SSL 3.0/TLS record that is too small to contain the
2546 first two bytes of the ClientHello message, i.e. client_version.
2547 (Note that this is a pathologic case that probably has never happened
2548 in real life.) The previous approach was to use the version number
2549 from the record header as a substitute; but our protocol choice
2550 should not depend on that one because it is not authenticated
2551 by the Finished messages.
2554 *) More robust randomness gathering functions for Windows.
2555 [Jeffrey Altman <jaltman@columbia.edu>]
2557 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2558 not set then we don't setup the error code for issuer check errors
2559 to avoid possibly overwriting other errors which the callback does
2560 handle. If an application does set the flag then we assume it knows
2561 what it is doing and can handle the new informational codes
2565 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2566 a general "ANY" type, as such it should be able to decode anything
2567 including tagged types. However it didn't check the class so it would
2568 wrongly interpret tagged types in the same way as their universal
2569 counterpart and unknown types were just rejected. Changed so that the
2570 tagged and unknown types are handled in the same way as a SEQUENCE:
2571 that is the encoding is stored intact. There is also a new type
2572 "V_ASN1_OTHER" which is used when the class is not universal, in this
2573 case we have no idea what the actual type is so we just lump them all
2577 *) On VMS, stdout may very well lead to a file that is written to
2578 in a record-oriented fashion. That means that every write() will
2579 write a separate record, which will be read separately by the
2580 programs trying to read from it. This can be very confusing.
2582 The solution is to put a BIO filter in the way that will buffer
2583 text until a linefeed is reached, and then write everything a
2584 line at a time, so every record written will be an actual line,
2585 not chunks of lines and not (usually doesn't happen, but I've
2586 seen it once) several lines in one record. BIO_f_linebuffer() is
2589 Currently, it's a VMS-only method, because that's where it has
2590 been tested well enough.
2593 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2594 it can return incorrect results.
2595 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2596 but it was in 0.9.6-beta[12].)
2599 *) Disable the check for content being present when verifying detached
2600 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2601 include zero length content when signing messages.
2604 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2605 BIO_ctrl (for BIO pairs).
2608 *) Add DSO method for VMS.
2611 *) Bug fix: Montgomery multiplication could produce results with the
2615 *) Add RPM specification openssl.spec and modify it to build three
2616 packages. The default package contains applications, application
2617 documentation and run-time libraries. The devel package contains
2618 include files, static libraries and function documentation. The
2619 doc package contains the contents of the doc directory. The original
2620 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2623 *) Add a large number of documentation files for many SSL routines.
2624 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2626 *) Add a configuration entry for Sony News 4.
2627 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2629 *) Don't set the two most significant bits to one when generating a
2630 random number < q in the DSA library.
2633 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2634 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2635 the underlying transport is blocking) if a handshake took place.
2636 (The default behaviour is needed by applications such as s_client
2637 and s_server that use select() to determine when to use SSL_read;
2638 but for applications that know in advance when to expect data, it
2639 just makes things more complicated.)
2642 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2646 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2647 work better on such systems.
2648 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2650 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2651 Update PKCS12_parse() so it copies the friendlyName and the
2652 keyid to the certificates aux info.
2655 *) Fix bug in PKCS7_verify() which caused an infinite loop
2656 if there was more than one signature.
2657 [Sven Uszpelkat <su@celocom.de>]
2659 *) Major change in util/mkdef.pl to include extra information
2660 about each symbol, as well as presentig variables as well
2661 as functions. This change means that there's n more need
2662 to rebuild the .num files when some algorithms are excluded.
2665 *) Allow the verify time to be set by an application,
2666 rather than always using the current time.
2669 *) Phase 2 verify code reorganisation. The certificate
2670 verify code now looks up an issuer certificate by a
2671 number of criteria: subject name, authority key id
2672 and key usage. It also verifies self signed certificates
2673 by the same criteria. The main comparison function is
2674 X509_check_issued() which performs these checks.
2676 Lot of changes were necessary in order to support this
2677 without completely rewriting the lookup code.
2679 Authority and subject key identifier are now cached.
2681 The LHASH 'certs' is X509_STORE has now been replaced
2682 by a STACK_OF(X509_OBJECT). This is mainly because an
2683 LHASH can't store or retrieve multiple objects with
2684 the same hash value.
2686 As a result various functions (which were all internal
2687 use only) have changed to handle the new X509_STORE
2688 structure. This will break anything that messed round
2689 with X509_STORE internally.
2691 The functions X509_STORE_add_cert() now checks for an
2692 exact match, rather than just subject name.
2694 The X509_STORE API doesn't directly support the retrieval
2695 of multiple certificates matching a given criteria, however
2696 this can be worked round by performing a lookup first
2697 (which will fill the cache with candidate certificates)
2698 and then examining the cache for matches. This is probably
2699 the best we can do without throwing out X509_LOOKUP
2700 entirely (maybe later...).
2702 The X509_VERIFY_CTX structure has been enhanced considerably.
2704 All certificate lookup operations now go via a get_issuer()
2705 callback. Although this currently uses an X509_STORE it
2706 can be replaced by custom lookups. This is a simple way
2707 to bypass the X509_STORE hackery necessary to make this
2708 work and makes it possible to use more efficient techniques
2709 in future. A very simple version which uses a simple
2710 STACK for its trusted certificate store is also provided
2711 using X509_STORE_CTX_trusted_stack().
2713 The verify_cb() and verify() callbacks now have equivalents
2714 in the X509_STORE_CTX structure.
2716 X509_STORE_CTX also has a 'flags' field which can be used
2717 to customise the verify behaviour.
2720 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2721 excludes S/MIME capabilities.
2724 *) When a certificate request is read in keep a copy of the
2725 original encoding of the signed data and use it when outputing
2726 again. Signatures then use the original encoding rather than
2727 a decoded, encoded version which may cause problems if the
2728 request is improperly encoded.
2731 *) For consistency with other BIO_puts implementations, call
2732 buffer_write(b, ...) directly in buffer_puts instead of calling
2735 In BIO_puts, increment b->num_write as in BIO_write.
2736 [Peter.Sylvester@EdelWeb.fr]
2738 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2739 BN_zero, we may not return a BIGNUM with an array consisting of
2743 *) Avoid calling abort() from within the library when problems are
2744 detected, except if preprocessor symbols have been defined
2745 (such as REF_CHECK, BN_DEBUG etc.).
2748 *) New openssl application 'rsautl'. This utility can be
2749 used for low level RSA operations. DER public key
2750 BIO/fp routines also added.
2753 *) New Configure entry and patches for compiling on QNX 4.
2754 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2756 *) A demo state-machine implementation was sponsored by
2757 Nuron (http://www.nuron.com/) and is now available in
2758 demos/state_machine.
2761 *) New options added to the 'dgst' utility for signature
2762 generation and verification.
2765 *) Unrecognized PKCS#7 content types are now handled via a
2766 catch all ASN1_TYPE structure. This allows unsupported
2767 types to be stored as a "blob" and an application can
2768 encode and decode it manually.
2771 *) Fix various signed/unsigned issues to make a_strex.c
2773 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2775 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2776 length if passed a buffer. ASN1_INTEGER_to_BN failed
2777 if passed a NULL BN and its argument was negative.
2778 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2780 *) Modification to PKCS#7 encoding routines to output definite
2781 length encoding. Since currently the whole structures are in
2782 memory there's not real point in using indefinite length
2783 constructed encoding. However if OpenSSL is compiled with
2784 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2787 *) Added BIO_vprintf() and BIO_vsnprintf().
2790 *) Added more prefixes to parse for in the the strings written
2791 through a logging bio, to cover all the levels that are available
2792 through syslog. The prefixes are now:
2794 PANIC, EMERG, EMR => LOG_EMERG
2795 ALERT, ALR => LOG_ALERT
2796 CRIT, CRI => LOG_CRIT
2797 ERROR, ERR => LOG_ERR
2798 WARNING, WARN, WAR => LOG_WARNING
2799 NOTICE, NOTE, NOT => LOG_NOTICE
2800 INFO, INF => LOG_INFO
2801 DEBUG, DBG => LOG_DEBUG
2803 and as before, if none of those prefixes are present at the
2804 beginning of the string, LOG_ERR is chosen.
2806 On Win32, the LOG_* levels are mapped according to this:
2808 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2809 LOG_WARNING => EVENTLOG_WARNING_TYPE
2810 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2814 *) Made it possible to reconfigure with just the configuration
2815 argument "reconf" or "reconfigure". The command line arguments
2816 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2817 and are retrieved from there when reconfiguring.
2821 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2823 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2826 *) The obj_dat.pl script was messing up the sorting of object
2827 names. The reason was that it compared the quoted version
2828 of strings as a result "OCSP" > "OCSP Signing" because
2829 " > SPACE. Changed script to store unquoted versions of
2830 names and add quotes on output. It was also omitting some
2831 names from the lookup table if they were given a default
2832 value (that is if SN is missing it is given the same
2833 value as LN and vice versa), these are now added on the
2834 grounds that if an object has a name we should be able to
2835 look it up. Finally added warning output when duplicate
2836 short or long names are found.
2839 *) Changes needed for Tandem NSK.
2840 [Scott Uroff <scott@xypro.com>]
2842 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2843 RSA_padding_check_SSLv23(), special padding was never detected
2844 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2845 version rollback attacks was not effective.
2847 In s23_clnt.c, don't use special rollback-attack detection padding
2848 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2849 client; similarly, in s23_srvr.c, don't do the rollback check if
2850 SSL 2.0 is the only protocol enabled in the server.
2853 *) Make it possible to get hexdumps of unprintable data with 'openssl
2854 asn1parse'. By implication, the functions ASN1_parse_dump() and
2855 BIO_dump_indent() are added.
2858 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2859 these print out strings and name structures based on various
2860 flags including RFC2253 support and proper handling of
2861 multibyte characters. Added options to the 'x509' utility
2862 to allow the various flags to be set.
2865 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2866 Also change the functions X509_cmp_current_time() and
2867 X509_gmtime_adj() work with an ASN1_TIME structure,
2868 this will enable certificates using GeneralizedTime in validity
2869 dates to be checked.
2872 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2873 negative public key encodings) on by default,
2874 NO_NEG_PUBKEY_BUG can be set to disable it.
2877 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2878 content octets. An i2c_ASN1_OBJECT is unnecessary because
2879 the encoding can be trivially obtained from the structure.
2882 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2883 not read locks (CRYPTO_r_[un]lock).
2886 *) A first attempt at creating official support for shared
2887 libraries through configuration. I've kept it so the
2888 default is static libraries only, and the OpenSSL programs
2889 are always statically linked for now, but there are
2890 preparations for dynamic linking in place.
2891 This has been tested on Linux and Tru64.
2894 *) Randomness polling function for Win9x, as described in:
2895 Peter Gutmann, Software Generation of Practically Strong
2899 *) Fix so PRNG is seeded in req if using an already existing
2903 *) New options to smime application. -inform and -outform
2904 allow alternative formats for the S/MIME message including
2905 PEM and DER. The -content option allows the content to be
2906 specified separately. This should allow things like Netscape
2907 form signing output easier to verify.
2910 *) Fix the ASN1 encoding of tags using the 'long form'.
2913 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2914 STRING types. These convert content octets to and from the
2915 underlying type. The actual tag and length octets are
2916 already assumed to have been read in and checked. These
2917 are needed because all other string types have virtually
2918 identical handling apart from the tag. By having versions
2919 of the ASN1 functions that just operate on content octets
2920 IMPLICIT tagging can be handled properly. It also allows
2921 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2922 and ASN1_INTEGER are identical apart from the tag.
2925 *) Change the handling of OID objects as follows:
2927 - New object identifiers are inserted in objects.txt, following
2928 the syntax given in objects.README.
2929 - objects.pl is used to process obj_mac.num and create a new
2931 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2934 This is currently kind of a hack, and the perl code in objects.pl
2935 isn't very elegant, but it works as I intended. The simplest way
2936 to check that it worked correctly is to look in obj_dat.h and
2937 check the array nid_objs and make sure the objects haven't moved
2938 around (this is important!). Additions are OK, as well as
2939 consistent name changes.
2942 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2945 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2946 The given file adds to whatever has already been seeded into the
2947 random pool through the RANDFILE configuration file option or
2948 environment variable, or the default random state file.
2951 *) mkstack.pl now sorts each macro group into lexical order.
2952 Previously the output order depended on the order the files
2953 appeared in the directory, resulting in needless rewriting
2957 *) Patches to make OpenSSL compile under Win32 again. Mostly
2958 work arounds for the VC++ problem that it treats func() as
2959 func(void). Also stripped out the parts of mkdef.pl that
2960 added extra typesafe functions: these no longer exist.
2963 *) Reorganisation of the stack code. The macros are now all
2964 collected in safestack.h . Each macro is defined in terms of
2965 a "stack macro" of the form SKM_<name>(type, a, b). The
2966 DEBUG_SAFESTACK is now handled in terms of function casts,
2967 this has the advantage of retaining type safety without the
2968 use of additional functions. If DEBUG_SAFESTACK is not defined
2969 then the non typesafe macros are used instead. Also modified the
2970 mkstack.pl script to handle the new form. Needs testing to see
2971 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2972 the default if no major problems. Similar behaviour for ASN1_SET_OF
2973 and PKCS12_STACK_OF.
2976 *) When some versions of IIS use the 'NET' form of private key the
2977 key derivation algorithm is different. Normally MD5(password) is
2978 used as a 128 bit RC4 key. In the modified case
2979 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2980 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2981 as the old Netscape_RSA functions except they have an additional
2982 'sgckey' parameter which uses the modified algorithm. Also added
2983 an -sgckey command line option to the rsa utility. Thanks to
2984 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2985 algorithm to openssl-dev.
2988 *) The evp_local.h macros were using 'c.##kname' which resulted in
2989 invalid expansion on some systems (SCO 5.0.5 for example).
2990 Corrected to 'c.kname'.
2991 [Phillip Porch <root@theporch.com>]
2993 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2994 a STACK of email addresses from a certificate or request, these look
2995 in the subject name and the subject alternative name extensions and
2996 omit any duplicate addresses.
2999 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
3000 This makes DSA verification about 2 % faster.
3003 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
3004 (meaning that now 2^5 values will be precomputed, which is only 4 KB
3005 plus overhead for 1024 bit moduli).
3006 This makes exponentiations about 0.5 % faster for 1024 bit
3007 exponents (as measured by "openssl speed rsa2048").
3010 *) Rename memory handling macros to avoid conflicts with other
3012 Malloc => OPENSSL_malloc
3013 Malloc_locked => OPENSSL_malloc_locked
3014 Realloc => OPENSSL_realloc
3015 Free => OPENSSL_free
3018 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
3019 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
3022 *) CygWin32 support.
3023 [John Jarvie <jjarvie@newsguy.com>]
3025 *) The type-safe stack code has been rejigged. It is now only compiled
3026 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
3027 by default all type-specific stack functions are "#define"d back to
3028 standard stack functions. This results in more streamlined output
3029 but retains the type-safety checking possibilities of the original
3033 *) The STACK code has been cleaned up, and certain type declarations
3034 that didn't make a lot of sense have been brought in line. This has
3035 also involved a cleanup of sorts in safestack.h to more correctly
3036 map type-safe stack functions onto their plain stack counterparts.
3037 This work has also resulted in a variety of "const"ifications of
3038 lots of the code, especially "_cmp" operations which should normally
3039 be prototyped with "const" parameters anyway.
3042 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
3043 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
3044 (The PRNG state consists of two parts, the large pool 'state' and 'md',
3045 where all of 'md' is used each time the PRNG is used, but 'state'
3046 is used only indexed by a cyclic counter. As entropy may not be
3047 well distributed from the beginning, 'md' is important as a
3048 chaining variable. However, the output function chains only half
3049 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
3050 all of 'md', and seeding with STATE_SIZE dummy bytes will result
3051 in all of 'state' being rewritten, with the new values depending
3052 on virtually all of 'md'. This overcomes the 80 bit limitation.)
3055 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
3056 the handshake is continued after ssl_verify_cert_chain();
3057 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
3058 can lead to 'unexplainable' connection aborts later.
3059 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
3061 *) Major EVP API cipher revision.
3062 Add hooks for extra EVP features. This allows various cipher
3063 parameters to be set in the EVP interface. Support added for variable
3064 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
3065 setting of RC2 and RC5 parameters.
3067 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
3070 Remove lots of duplicated code from the EVP library. For example *every*
3071 cipher init() function handles the 'iv' in the same way according to the
3072 cipher mode. They also all do nothing if the 'key' parameter is NULL and
3073 for CFB and OFB modes they zero ctx->num.
3075 New functionality allows removal of S/MIME code RC2 hack.
3077 Most of the routines have the same form and so can be declared in terms
3080 By shifting this to the top level EVP_CipherInit() it can be removed from
3081 all individual ciphers. If the cipher wants to handle IVs or keys
3082 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
3085 Change lots of functions like EVP_EncryptUpdate() to now return a
3086 value: although software versions of the algorithms cannot fail
3087 any installed hardware versions can.
3090 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3091 this option is set, tolerate broken clients that send the negotiated
3092 protocol version number instead of the requested protocol version
3096 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3097 i.e. non-zero for export ciphersuites, zero otherwise.
3098 Previous versions had this flag inverted, inconsistent with
3099 rsa_tmp_cb (..._TMP_RSA_CB).
3100 [Bodo Moeller; problem reported by Amit Chopra]
3102 *) Add missing DSA library text string. Work around for some IIS
3103 key files with invalid SEQUENCE encoding.
3106 *) Add a document (doc/standards.txt) that list all kinds of standards
3107 and so on that are implemented in OpenSSL.
3110 *) Enhance c_rehash script. Old version would mishandle certificates
3111 with the same subject name hash and wouldn't handle CRLs at all.
3112 Added -fingerprint option to crl utility, to support new c_rehash
3116 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3119 *) Fix for SSL server purpose checking. Server checking was
3120 rejecting certificates which had extended key usage present
3121 but no ssl client purpose.
3122 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3124 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3125 is a little unclear about how a blank password is handled.
3126 Since the password in encoded as a BMPString with terminating
3127 double NULL a zero length password would end up as just the
3128 double NULL. However no password at all is different and is
3129 handled differently in the PKCS#12 key generation code. NS
3130 treats a blank password as zero length. MSIE treats it as no
3131 password on export: but it will try both on import. We now do
3132 the same: PKCS12_parse() tries zero length and no password if
3133 the password is set to "" or NULL (NULL is now a valid password:
3134 it wasn't before) as does the pkcs12 application.
3137 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3138 perror when PEM_read_bio_X509_REQ fails, the error message must
3139 be obtained from the error queue.
3142 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3143 it in ERR_remove_state if appropriate, and change ERR_get_state
3144 accordingly to avoid race conditions (this is necessary because
3145 thread_hash is no longer constant once set).
3148 *) Bugfix for linux-elf makefile.one.
3151 *) RSA_get_default_method() will now cause a default
3152 RSA_METHOD to be chosen if one doesn't exist already.
3153 Previously this was only set during a call to RSA_new()
3154 or RSA_new_method(NULL) meaning it was possible for
3155 RSA_get_default_method() to return NULL.
3158 *) Added native name translation to the existing DSO code
3159 that will convert (if the flag to do so is set) filenames
3160 that are sufficiently small and have no path information
3161 into a canonical native form. Eg. "blah" converted to
3162 "libblah.so" or "blah.dll" etc.
3165 *) New function ERR_error_string_n(e, buf, len) which is like
3166 ERR_error_string(e, buf), but writes at most 'len' bytes
3167 including the 0 terminator. For ERR_error_string_n, 'buf'
3169 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3171 *) CONF library reworked to become more general. A new CONF
3172 configuration file reader "class" is implemented as well as a
3173 new functions (NCONF_*, for "New CONF") to handle it. The now
3174 old CONF_* functions are still there, but are reimplemented to
3175 work in terms of the new functions. Also, a set of functions
3176 to handle the internal storage of the configuration data is
3177 provided to make it easier to write new configuration file
3178 reader "classes" (I can definitely see something reading a
3179 configuration file in XML format, for example), called _CONF_*,
3180 or "the configuration storage API"...
3182 The new configuration file reading functions are:
3184 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3185 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3187 NCONF_default, NCONF_WIN32
3189 NCONF_dump_fp, NCONF_dump_bio
3191 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3192 NCONF_new creates a new CONF object. This works in the same way
3193 as other interfaces in OpenSSL, like the BIO interface.
3194 NCONF_dump_* dump the internal storage of the configuration file,
3195 which is useful for debugging. All other functions take the same
3196 arguments as the old CONF_* functions wth the exception of the
3197 first that must be a `CONF *' instead of a `LHASH *'.
3199 To make it easer to use the new classes with the old CONF_* functions,
3200 the function CONF_set_default_method is provided.
3203 *) Add '-tls1' option to 'openssl ciphers', which was already
3204 mentioned in the documentation but had not been implemented.
3205 (This option is not yet really useful because even the additional
3206 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3209 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3210 OpenSSL-based applications) load shared libraries and bind to
3211 them in a portable way.
3212 [Geoff Thorpe, with contributions from Richard Levitte]
3214 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3216 *) Make sure _lrotl and _lrotr are only used with MSVC.
3218 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3219 (the default implementation of RAND_status).
3221 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3222 to '-clrext' (= clear extensions), as intended and documented.
3223 [Bodo Moeller; inconsistency pointed out by Michael Attili
3226 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3227 was larger than the MD block size.
3228 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3230 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3231 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3232 using the passed key: if the passed key was a private key the result
3233 of X509_print(), for example, would be to print out all the private key
3237 *) des_quad_cksum() byte order bug fix.
3238 [Ulf Möller, using the problem description in krb4-0.9.7, where
3239 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3241 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3243 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3245 *) For easily testing in shell scripts whether some command
3246 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3247 returns with exit code 0 iff no command of the given name is available.
3248 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3249 the output goes to stdout and nothing is printed to stderr.
3250 Additional arguments are always ignored.
3252 Since for each cipher there is a command of the same name,
3253 the 'no-cipher' compilation switches can be tested this way.
3255 ('openssl no-XXX' is not able to detect pseudo-commands such
3256 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3259 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3262 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3263 is set; it will be thrown away anyway because each handshake creates
3265 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3266 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3267 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3268 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3271 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3272 'Q' and 'R' lose their special meanings (quit/renegotiate).
3273 This is part of what -quiet does; unlike -quiet, -ign_eof
3274 does not suppress any output.
3277 *) Add compatibility options to the purpose and trust code. The
3278 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3279 accepts a certificate or CA, this was the previous behaviour,
3280 with all the associated security issues.
3282 X509_TRUST_COMPAT is the old trust behaviour: only and
3283 automatically trust self signed roots in certificate store. A
3284 new trust setting X509_TRUST_DEFAULT is used to specify that
3285 a purpose has no associated trust setting and it should instead
3286 use the value in the default purpose.
3289 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3290 and fix a memory leak.
3293 *) In util/mkerr.pl (which implements 'make errors'), preserve
3294 reason strings from the previous version of the .c file, as
3295 the default to have only downcase letters (and digits) in
3296 automatically generated reasons codes is not always appropriate.
3299 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3300 using strerror. Previously, ERR_reason_error_string() returned
3301 library names as reason strings for SYSerr; but SYSerr is a special
3302 case where small numbers are errno values, not library numbers.
3305 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3306 converts DSA parameters into DH parameters. (When creating parameters,
3307 DSA_generate_parameters is used.)
3310 *) Include 'length' (recommended exponent length) in C code generated
3311 by 'openssl dhparam -C'.
3314 *) The second argument to set_label in perlasm was already being used
3315 so couldn't be used as a "file scope" flag. Moved to third argument
3319 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3320 instead of RAND_bytes for encryption IVs and salts.
3323 *) Include RAND_status() into RAND_METHOD instead of implementing
3324 it only for md_rand.c Otherwise replacing the PRNG by calling
3325 RAND_set_rand_method would be impossible.
3328 *) Don't let DSA_generate_key() enter an infinite loop if the random
3329 number generation fails.
3332 *) New 'rand' application for creating pseudo-random output.
3335 *) Added configuration support for Linux/IA64
3336 [Rolf Haberrecker <rolf@suse.de>]
3338 *) Assembler module support for Mingw32.
3341 *) Shared library support for HPUX (in shlib/).
3342 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3344 *) Shared library support for Solaris gcc.
3345 [Lutz Behnke <behnke@trustcenter.de>]
3347 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3349 *) PKCS7_encrypt() was adding text MIME headers twice because they
3350 were added manually and by SMIME_crlf_copy().
3353 *) In bntest.c don't call BN_rand with zero bits argument.
3354 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3356 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3357 case was implemented. This caused BN_div_recp() to fail occasionally.
3360 *) Add an optional second argument to the set_label() in the perl
3361 assembly language builder. If this argument exists and is set
3362 to 1 it signals that the assembler should use a symbol whose
3363 scope is the entire file, not just the current function. This
3364 is needed with MASM which uses the format label:: for this scope.
3365 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3367 *) Change the ASN1 types so they are typedefs by default. Before
3368 almost all types were #define'd to ASN1_STRING which was causing
3369 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3373 *) Change names of new functions to the new get1/get0 naming
3374 convention: After 'get1', the caller owns a reference count
3375 and has to call ..._free; 'get0' returns a pointer to some
3376 data structure without incrementing reference counters.
3377 (Some of the existing 'get' functions increment a reference
3378 counter, some don't.)
3379 Similarly, 'set1' and 'add1' functions increase reference
3380 counters or duplicate objects.
3383 *) Allow for the possibility of temp RSA key generation failure:
3384 the code used to assume it always worked and crashed on failure.
3387 *) Fix potential buffer overrun problem in BIO_printf().
3388 [Ulf Möller, using public domain code by Patrick Powell; problem
3389 pointed out by David Sacerdote <das33@cornell.edu>]
3391 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3392 RAND_egd() and RAND_status(). In the command line application,
3393 the EGD socket can be specified like a seed file using RANDFILE
3397 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3398 Some CAs (e.g. Verisign) distribute certificates in this form.
3401 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3402 list to exclude them. This means that no special compilation option
3403 is needed to use anonymous DH: it just needs to be included in the
3407 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3408 EVP_MD_type. The old functionality is available in a new macro called
3409 EVP_MD_md(). Change code that uses it and update docs.
3412 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3413 where the 'void *' argument is replaced by a function pointer argument.
3414 Previously 'void *' was abused to point to functions, which works on
3415 many platforms, but is not correct. As these functions are usually
3416 called by macros defined in OpenSSL header files, most source code
3417 should work without changes.
3420 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3421 sections with information on -D... compiler switches used for
3422 compiling the library so that applications can see them. To enable
3423 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3424 must be defined. E.g.,
3425 #define OPENSSL_ALGORITHM_DEFINES
3426 #include <openssl/opensslconf.h>
3427 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3428 [Richard Levitte, Ulf and Bodo Möller]
3430 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3434 *) Change the 'other' type in certificate aux info to a STACK_OF
3435 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3436 the required ASN1 format: arbitrary types determined by an OID.
3439 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3440 argument to 'req'. This is not because the function is newer or
3441 better than others it just uses the work 'NEW' in the certificate
3442 request header lines. Some software needs this.
3445 *) Reorganise password command line arguments: now passwords can be
3446 obtained from various sources. Delete the PEM_cb function and make
3447 it the default behaviour: i.e. if the callback is NULL and the
3448 usrdata argument is not NULL interpret it as a null terminated pass
3449 phrase. If usrdata and the callback are NULL then the pass phrase
3450 is prompted for as usual.
3453 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3454 the support is automatically enabled. The resulting binaries will
3455 autodetect the card and use it if present.
3456 [Ben Laurie and Compaq Inc.]
3458 *) Work around for Netscape hang bug. This sends certificate request
3459 and server done in one record. Since this is perfectly legal in the
3460 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3461 the bugs/SSLv3 entry for more info.
3464 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3467 *) Add -rand argument to smime and pkcs12 applications and read/write
3471 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3474 *) Add command line password options to the remaining applications.
3477 *) Bug fix for BN_div_recp() for numerators with an even number of
3481 *) More tests in bntest.c, and changed test_bn output.
3484 *) ./config recognizes MacOS X now.
3487 *) Bug fix for BN_div() when the first words of num and divsor are
3488 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3491 *) Add support for various broken PKCS#8 formats, and command line
3492 options to produce them.
3495 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3496 get temporary BIGNUMs from a BN_CTX.
3499 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3503 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3504 include a #define from the old name to the new. The original intent
3505 was that statically linked binaries could for example just call
3506 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3507 link with digests. This never worked becayse SSLeay_add_all_digests()
3508 and SSLeay_add_all_ciphers() were in the same source file so calling
3509 one would link with the other. They are now in separate source files.
3512 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3515 *) Use a less unusual form of the Miller-Rabin primality test (it used
3516 a binary algorithm for exponentiation integrated into the Miller-Rabin
3517 loop, our standard modexp algorithms are faster).
3520 *) Support for the EBCDIC character set completed.
3521 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3523 *) Source code cleanups: use const where appropriate, eliminate casts,
3524 use void * instead of char * in lhash.
3527 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3528 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3529 this the server could overwrite ephemeral keys that the client
3533 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3534 using 50 iterations of the Rabin-Miller test.
3536 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3537 iterations of the Rabin-Miller test as required by the appendix
3538 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3539 As BN_is_prime_fasttest includes trial division, DSA parameter
3540 generation becomes much faster.
3542 This implies a change for the callback functions in DSA_is_prime
3543 and DSA_generate_parameters: The callback function is called once
3544 for each positive witness in the Rabin-Miller test, not just
3545 occasionally in the inner loop; and the parameters to the
3546 callback function now provide an iteration count for the outer
3547 loop rather than for the current invocation of the inner loop.
3548 DSA_generate_parameters additionally can call the callback
3549 function with an 'iteration count' of -1, meaning that a
3550 candidate has passed the trial division test (when q is generated
3551 from an application-provided seed, trial division is skipped).
3554 *) New function BN_is_prime_fasttest that optionally does trial
3555 division before starting the Rabin-Miller test and has
3556 an additional BN_CTX * argument (whereas BN_is_prime always
3557 has to allocate at least one BN_CTX).
3558 'callback(1, -1, cb_arg)' is called when a number has passed the
3559 trial division stage.
3562 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3566 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3569 *) New function BN_pseudo_rand().
3572 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3573 bignum version of BN_from_montgomery() with the working code from
3574 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3578 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3579 made it impossible to use the same SSL_SESSION data structure in
3580 SSL2 clients in multiple threads.
3583 *) The return value of RAND_load_file() no longer counts bytes obtained
3584 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3585 to seed the PRNG (previously an explicit byte count was required).
3586 [Ulf Möller, Bodo Möller]
3588 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3589 used (char *) instead of (void *) and had casts all over the place.
3592 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3595 *) Retain source code compatibility for BN_prime_checks macro:
3596 BN_is_prime(..., BN_prime_checks, ...) now uses
3597 BN_prime_checks_for_size to determine the appropriate number of
3598 Rabin-Miller iterations.
3601 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3602 DH_CHECK_P_NOT_SAFE_PRIME.
3603 (Check if this is true? OpenPGP calls them "strong".)
3606 *) Merge the functionality of "dh" and "gendh" programs into a new program
3607 "dhparam". The old programs are retained for now but will handle DH keys
3608 (instead of parameters) in future.
3611 *) Make the ciphers, s_server and s_client programs check the return values
3612 when a new cipher list is set.
3615 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3616 ciphers. Before when the 56bit ciphers were enabled the sorting was
3619 The syntax for the cipher sorting has been extended to support sorting by
3620 cipher-strength (using the strength_bits hard coded in the tables).
3621 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3623 Fix a bug in the cipher-command parser: when supplying a cipher command
3624 string with an "undefined" symbol (neither command nor alphanumeric
3625 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3626 an error is flagged.
3628 Due to the strength-sorting extension, the code of the
3629 ssl_create_cipher_list() function was completely rearranged. I hope that
3630 the readability was also increased :-)
3631 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3633 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3634 for the first serial number and places 2 in the serial number file. This
3635 avoids problems when the root CA is created with serial number zero and
3636 the first user certificate has the same issuer name and serial number
3640 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3641 the new code. Add documentation for this stuff.
3644 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3645 X509_*() to X509at_*() on the grounds that they don't handle X509
3646 structures and behave in an analagous way to the X509v3 functions:
3647 they shouldn't be called directly but wrapper functions should be used
3650 So we also now have some wrapper functions that call the X509at functions
3651 when passed certificate requests. (TO DO: similar things can be done with
3652 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3653 things. Some of these need some d2i or i2d and print functionality
3654 because they handle more complex structures.)
3657 *) Add missing #ifndefs that caused missing symbols when building libssl
3658 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3659 NO_RSA in ssl/s2*.c.
3660 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3662 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3663 has a return value which indicates the quality of the random data
3664 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3665 error queue. New function RAND_pseudo_bytes() generates output that is
3666 guaranteed to be unique but not unpredictable. RAND_add is like
3667 RAND_seed, but takes an extra argument for an entropy estimate
3668 (RAND_seed always assumes full entropy).
3671 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3672 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3673 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3674 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3675 false-positive rate of at most 2^-80 for random input.
3678 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3681 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3682 in the 0.9.5 release), this returns the chain
3683 from an X509_CTX structure with a dup of the stack and all
3684 the X509 reference counts upped: so the stack will exist
3685 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3688 Also make SSL_SESSION_print() print out the verify return
3692 *) Add manpage for the pkcs12 command. Also change the default
3693 behaviour so MAC iteration counts are used unless the new
3694 -nomaciter option is used. This improves file security and
3695 only older versions of MSIE (4.0 for example) need it.
3698 *) Honor the no-xxx Configure options when creating .DEF files.
3701 *) Add PKCS#10 attributes to field table: challengePassword,
3702 unstructuredName and unstructuredAddress. These are taken from
3703 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3704 international characters are used.
3706 More changes to X509_ATTRIBUTE code: allow the setting of types
3707 based on strings. Remove the 'loc' parameter when adding
3708 attributes because these will be a SET OF encoding which is sorted
3712 *) Initial changes to the 'req' utility to allow request generation
3713 automation. This will allow an application to just generate a template
3714 file containing all the field values and have req construct the
3717 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3718 used all over the place including certificate requests and PKCS#7
3719 structures. They are currently handled manually where necessary with
3720 some primitive wrappers for PKCS#7. The new functions behave in a
3721 manner analogous to the X509 extension functions: they allow
3722 attributes to be looked up by NID and added.
3724 Later something similar to the X509V3 code would be desirable to
3725 automatically handle the encoding, decoding and printing of the
3726 more complex types. The string types like challengePassword can
3727 be handled by the string table functions.
3729 Also modified the multi byte string table handling. Now there is
3730 a 'global mask' which masks out certain types. The table itself
3731 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3732 is useful when for example there is only one permissible type
3733 (as in countryName) and using the mask might result in no valid
3737 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3738 SSL_get_peer_finished to allow applications to obtain the latest
3739 Finished messages sent to the peer or expected from the peer,
3740 respectively. (SSL_get_peer_finished is usually the Finished message
3741 actually received from the peer, otherwise the protocol will be aborted.)
3743 As the Finished message are message digests of the complete handshake
3744 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3745 be used for external authentication procedures when the authentication
3746 provided by SSL/TLS is not desired or is not enough.
3749 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3750 the host supports BWX extension and if Compaq C is present on the
3751 $PATH. Just exploiting of the BWX extension results in 20-30%
3752 performance kick for some algorithms, e.g. DES and RC4 to mention
3753 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3757 *) Add support for MS "fast SGC". This is arguably a violation of the
3758 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3759 weak crypto and after checking the certificate is SGC a second one
3760 with strong crypto. MS SGC stops the first handshake after receiving
3761 the server certificate message and sends a second client hello. Since
3762 a server will typically do all the time consuming operations before
3763 expecting any further messages from the client (server key exchange
3764 is the most expensive) there is little difference between the two.
3766 To get OpenSSL to support MS SGC we have to permit a second client
3767 hello message after we have sent server done. In addition we have to
3768 reset the MAC if we do get this second client hello.
3771 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3772 if a DER encoded private key is RSA or DSA traditional format. Changed
3773 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3774 format DER encoded private key. Newer code should use PKCS#8 format which
3775 has the key type encoded in the ASN1 structure. Added DER private key
3776 support to pkcs8 application.
3779 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3780 ciphersuites has been selected (as required by the SSL 3/TLS 1
3781 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3782 is set, we interpret this as a request to violate the specification
3783 (the worst that can happen is a handshake failure, and 'correct'
3784 behaviour would result in a handshake failure anyway).
3787 *) In SSL_CTX_add_session, take into account that there might be multiple
3788 SSL_SESSION structures with the same session ID (e.g. when two threads
3789 concurrently obtain them from an external cache).
3790 The internal cache can handle only one SSL_SESSION with a given ID,
3791 so if there's a conflict, we now throw out the old one to achieve
3795 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3796 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3797 some routines that use cipher OIDs: some ciphers do not have OIDs
3798 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3802 *) Simplify the trust setting structure and code. Now we just have
3803 two sequences of OIDs for trusted and rejected settings. These will
3804 typically have values the same as the extended key usage extension
3805 and any application specific purposes.
3807 The trust checking code now has a default behaviour: it will just
3808 check for an object with the same NID as the passed id. Functions can
3809 be provided to override either the default behaviour or the behaviour
3810 for a given id. SSL client, server and email already have functions
3811 in place for compatibility: they check the NID and also return "trusted"
3812 if the certificate is self signed.
3815 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3816 traditional format into an EVP_PKEY structure.
3819 *) Add a password callback function PEM_cb() which either prompts for
3820 a password if usr_data is NULL or otherwise assumes it is a null
3821 terminated password. Allow passwords to be passed on command line
3822 environment or config files in a few more utilities.
3825 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3826 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3827 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3828 Update documentation.
3831 *) Support for ASN1 "NULL" type. This could be handled before by using
3832 ASN1_TYPE but there wasn't any function that would try to read a NULL
3833 and produce an error if it couldn't. For compatibility we also have
3834 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3835 don't allocate anything because they don't need to.
3838 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3840 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3842 *) Rebuild of the memory allocation routines used by OpenSSL code and
3843 possibly others as well. The purpose is to make an interface that
3844 provide hooks so anyone can build a separate set of allocation and
3845 deallocation routines to be used by OpenSSL, for example memory
3846 pool implementations, or something else, which was previously hard
3847 since Malloc(), Realloc() and Free() were defined as macros having
3848 the values malloc, realloc and free, respectively (except for Win32
3849 compilations). The same is provided for memory debugging code.
3850 OpenSSL already comes with functionality to find memory leaks, but
3851 this gives people a chance to debug other memory problems.
3853 With these changes, a new set of functions and macros have appeared:
3855 CRYPTO_set_mem_debug_functions() [F]
3856 CRYPTO_get_mem_debug_functions() [F]
3857 CRYPTO_dbg_set_options() [F]
3858 CRYPTO_dbg_get_options() [F]
3859 CRYPTO_malloc_debug_init() [M]
3861 The memory debug functions are NULL by default, unless the library
3862 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3863 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3864 gives the standard debugging functions that come with OpenSSL) or
3865 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3866 provided by the library user) must be used. When the standard
3867 debugging functions are used, CRYPTO_dbg_set_options can be used to
3868 request additional information:
3869 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3870 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3872 Also, things like CRYPTO_set_mem_functions will always give the
3873 expected result (the new set of functions is used for allocation
3874 and deallocation) at all times, regardless of platform and compiler
3877 To finish it up, some functions that were never use in any other
3878 way than through macros have a new API and new semantic:
3881 CRYPTO_dbg_realloc()
3884 All macros of value have retained their old syntax.
3885 [Richard Levitte and Bodo Moeller]
3887 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3888 ordering of SMIMECapabilities wasn't in "strength order" and there
3889 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3893 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3894 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3895 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3897 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3898 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3899 functionality to handle multipart/signed properly) and a utility
3900 called 'smime' to call all this stuff. This is based on code I
3901 originally wrote for Celo who have kindly allowed it to be
3902 included in OpenSSL.
3905 *) Add variants des_set_key_checked and des_set_key_unchecked of
3906 des_set_key (aka des_key_sched). Global variable des_check_key
3907 decides which of these is called by des_set_key; this way
3908 des_check_key behaves as it always did, but applications and
3909 the library itself, which was buggy for des_check_key == 1,
3910 have a cleaner way to pick the version they need.
3913 *) New function PKCS12_newpass() which changes the password of a
3917 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3918 dynamic mix. In both cases the ids can be used as an index into the
3919 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3920 functions so they accept a list of the field values and the
3921 application doesn't need to directly manipulate the X509_TRUST
3925 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3929 *) Modify the way the V3 extension code looks up extensions. This now
3930 works in a similar way to the object code: we have some "standard"
3931 extensions in a static table which is searched with OBJ_bsearch()
3932 and the application can add dynamic ones if needed. The file
3933 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3934 updated whenever a new extension is added to the core code and kept
3935 in ext_nid order. There is a simple program 'tabtest.c' which checks
3936 this. New extensions are not added too often so this file can readily
3937 be maintained manually.
3939 There are two big advantages in doing things this way. The extensions
3940 can be looked up immediately and no longer need to be "added" using
3941 X509V3_add_standard_extensions(): this function now does nothing.
3942 [Side note: I get *lots* of email saying the extension code doesn't
3943 work because people forget to call this function]
3944 Also no dynamic allocation is done unless new extensions are added:
3945 so if we don't add custom extensions there is no need to call
3946 X509V3_EXT_cleanup().
3949 *) Modify enc utility's salting as follows: make salting the default. Add a
3950 magic header, so unsalted files fail gracefully instead of just decrypting
3951 to garbage. This is because not salting is a big security hole, so people
3952 should be discouraged from doing it.
3955 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3956 digest to be passed on the command line but it only used this
3957 parameter when signing a certificate. Modified so all relevant
3958 operations are affected by the digest parameter including the
3959 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3960 DSA key was used because it didn't fix the digest.
3963 *) Initial certificate chain verify code. Currently tests the untrusted
3964 certificates for consistency with the verify purpose (which is set
3965 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3967 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3968 this is because it will reject chains with invalid extensions whereas
3969 every previous version of OpenSSL and SSLeay made no checks at all.
3971 Trust code: checks the root CA for the relevant trust settings. Trust
3972 settings have an initial value consistent with the verify purpose: e.g.
3973 if the verify purpose is for SSL client use it expects the CA to be
3974 trusted for SSL client use. However the default value can be changed to
3975 permit custom trust settings: one example of this would be to only trust
3976 certificates from a specific "secure" set of CAs.
3978 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3979 which should be used for version portability: especially since the
3980 verify structure is likely to change more often now.
3982 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3983 to set them. If not set then assume SSL clients will verify SSL servers
3986 Two new options to the verify program: -untrusted allows a set of
3987 untrusted certificates to be passed in and -purpose which sets the
3988 intended purpose of the certificate. If a purpose is set then the
3989 new chain verify code is used to check extension consistency.
3992 *) Support for the authority information access extension.
3995 *) Modify RSA and DSA PEM read routines to transparently handle
3996 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3997 public keys in a format compatible with certificate
3998 SubjectPublicKeyInfo structures. Unfortunately there were already
3999 functions called *_PublicKey_* which used various odd formats so
4000 these are retained for compatibility: however the DSA variants were
4001 never in a public release so they have been deleted. Changed dsa/rsa
4002 utilities to handle the new format: note no releases ever handled public
4003 keys so we should be OK.
4005 The primary motivation for this change is to avoid the same fiasco
4006 that dogs private keys: there are several incompatible private key
4007 formats some of which are standard and some OpenSSL specific and
4008 require various evil hacks to allow partial transparent handling and
4009 even then it doesn't work with DER formats. Given the option anything
4010 other than PKCS#8 should be dumped: but the other formats have to
4011 stay in the name of compatibility.
4013 With public keys and the benefit of hindsight one standard format
4014 is used which works with EVP_PKEY, RSA or DSA structures: though
4015 it clearly returns an error if you try to read the wrong kind of key.
4017 Added a -pubkey option to the 'x509' utility to output the public key.
4018 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
4019 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
4020 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
4021 that do the same as the EVP_PKEY_assign_*() except they up the
4022 reference count of the added key (they don't "swallow" the
4026 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
4027 CRLs would fail if the file contained no certificates or no CRLs:
4028 added a new function to read in both types and return the number
4029 read: this means that if none are read it will be an error. The
4030 DER versions of the certificate and CRL reader would always fail
4031 because it isn't possible to mix certificates and CRLs in DER format
4032 without choking one or the other routine. Changed this to just read
4033 a certificate: this is the best we can do. Also modified the code
4034 in apps/verify.c to take notice of return codes: it was previously
4035 attempting to read in certificates from NULL pointers and ignoring
4036 any errors: this is one reason why the cert and CRL reader seemed
4037 to work. It doesn't check return codes from the default certificate
4038 routines: these may well fail if the certificates aren't installed.
4041 *) Code to support otherName option in GeneralName.
4044 *) First update to verify code. Change the verify utility
4045 so it warns if it is passed a self signed certificate:
4046 for consistency with the normal behaviour. X509_verify
4047 has been modified to it will now verify a self signed
4048 certificate if *exactly* the same certificate appears
4049 in the store: it was previously impossible to trust a
4050 single self signed certificate. This means that:
4051 openssl verify ss.pem
4052 now gives a warning about a self signed certificate but
4053 openssl verify -CAfile ss.pem ss.pem
4057 *) For servers, store verify_result in SSL_SESSION data structure
4058 (and add it to external session representation).
4059 This is needed when client certificate verifications fails,
4060 but an application-provided verification callback (set by
4061 SSL_CTX_set_cert_verify_callback) allows accepting the session
4062 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
4063 but returns 1): When the session is reused, we have to set
4064 ssl->verify_result to the appropriate error code to avoid
4066 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
4068 *) Fix a bug in the new PKCS#7 code: it didn't consider the
4069 case in PKCS7_dataInit() where the signed PKCS7 structure
4070 didn't contain any existing data because it was being created.
4071 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
4073 *) Add a salt to the key derivation routines in enc.c. This
4074 forms the first 8 bytes of the encrypted file. Also add a
4075 -S option to allow a salt to be input on the command line.
4078 *) New function X509_cmp(). Oddly enough there wasn't a function
4079 to compare two certificates. We do this by working out the SHA1
4080 hash and comparing that. X509_cmp() will be needed by the trust
4084 *) SSL_get1_session() is like SSL_get_session(), but increments
4085 the reference count in the SSL_SESSION returned.
4086 [Geoff Thorpe <geoff@eu.c2.net>]
4088 *) Fix for 'req': it was adding a null to request attributes.
4089 Also change the X509_LOOKUP and X509_INFO code to handle
4090 certificate auxiliary information.
4093 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4097 *) Add the possibility to add extra information to the memory leak
4098 detecting output, to form tracebacks, showing from where each
4099 allocation was originated: CRYPTO_push_info("constant string") adds
4100 the string plus current file name and line number to a per-thread
4101 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4102 is like calling CYRPTO_pop_info() until the stack is empty.
4103 Also updated memory leak detection code to be multi-thread-safe.
4106 *) Add options -text and -noout to pkcs7 utility and delete the
4107 encryption options which never did anything. Update docs.
4110 *) Add options to some of the utilities to allow the pass phrase
4111 to be included on either the command line (not recommended on
4112 OSes like Unix) or read from the environment. Update the
4113 manpages and fix a few bugs.
4116 *) Add a few manpages for some of the openssl commands.
4119 *) Fix the -revoke option in ca. It was freeing up memory twice,
4120 leaking and not finding already revoked certificates.
4123 *) Extensive changes to support certificate auxiliary information.
4124 This involves the use of X509_CERT_AUX structure and X509_AUX
4125 functions. An X509_AUX function such as PEM_read_X509_AUX()
4126 can still read in a certificate file in the usual way but it
4127 will also read in any additional "auxiliary information". By
4128 doing things this way a fair degree of compatibility can be
4129 retained: existing certificates can have this information added
4130 using the new 'x509' options.
4132 Current auxiliary information includes an "alias" and some trust
4133 settings. The trust settings will ultimately be used in enhanced
4134 certificate chain verification routines: currently a certificate
4135 can only be trusted if it is self signed and then it is trusted
4139 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4140 The problem was that one of the replacement routines had not been working
4141 since SSLeay releases. For now the offending routine has been replaced
4142 with non-optimised assembler. Even so, this now gives around 95%
4143 performance improvement for 1024 bit RSA signs.
4146 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4147 handling. Most clients have the effective key size in bits equal to
4148 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4149 A few however don't do this and instead use the size of the decrypted key
4150 to determine the RC2 key length and the AlgorithmIdentifier to determine
4151 the effective key length. In this case the effective key length can still
4152 be 40 bits but the key length can be 168 bits for example. This is fixed
4153 by manually forcing an RC2 key into the EVP_PKEY structure because the
4154 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4155 the key length and effective key length are equal.
4158 *) Add a bunch of functions that should simplify the creation of
4159 X509_NAME structures. Now you should be able to do:
4160 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4161 and have it automatically work out the correct field type and fill in
4162 the structures. The more adventurous can try:
4163 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4164 and it will (hopefully) work out the correct multibyte encoding.
4167 *) Change the 'req' utility to use the new field handling and multibyte
4168 copy routines. Before the DN field creation was handled in an ad hoc
4169 way in req, ca, and x509 which was rather broken and didn't support
4170 BMPStrings or UTF8Strings. Since some software doesn't implement
4171 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4172 using the dirstring_type option. See the new comment in the default
4173 openssl.cnf for more info.
4176 *) Make crypto/rand/md_rand.c more robust:
4177 - Assure unique random numbers after fork().
4178 - Make sure that concurrent threads access the global counter and
4179 md serializably so that we never lose entropy in them
4180 or use exactly the same state in multiple threads.
4181 Access to the large state is not always serializable because
4182 the additional locking could be a performance killer, and
4183 md should be large enough anyway.
4186 *) New file apps/app_rand.c with commonly needed functionality
4187 for handling the random seed file.
4189 Use the random seed file in some applications that previously did not:
4191 dsaparam -genkey (which also ignored its '-rand' option),
4194 x509 (when signing).
4195 Except on systems with /dev/urandom, it is crucial to have a random
4196 seed file at least for key creation, DSA signing, and for DH exchanges;
4197 for RSA signatures we could do without one.
4199 gendh and gendsa (unlike genrsa) used to read only the first byte
4200 of each file listed in the '-rand' option. The function as previously
4201 found in genrsa is now in app_rand.c and is used by all programs
4202 that support '-rand'.
4205 *) In RAND_write_file, use mode 0600 for creating files;
4206 don't just chmod when it may be too late.
4209 *) Report an error from X509_STORE_load_locations
4210 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4213 *) New function ASN1_mbstring_copy() this copies a string in either
4214 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4215 into an ASN1_STRING type. A mask of permissible types is passed
4216 and it chooses the "minimal" type to use or an error if not type
4220 *) Add function equivalents to the various macros in asn1.h. The old
4221 macros are retained with an M_ prefix. Code inside the library can
4222 use the M_ macros. External code (including the openssl utility)
4223 should *NOT* in order to be "shared library friendly".
4226 *) Add various functions that can check a certificate's extensions
4227 to see if it usable for various purposes such as SSL client,
4228 server or S/MIME and CAs of these types. This is currently
4229 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4230 verification. Also added a -purpose flag to x509 utility to
4231 print out all the purposes.
4234 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4238 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4239 for, obtain and decode and extension and obtain its critical flag.
4240 This allows all the necessary extension code to be handled in a
4241 single function call.
4244 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4245 platforms. See crypto/rc4/rc4_enc.c for further details.
4248 *) New -noout option to asn1parse. This causes no output to be produced
4249 its main use is when combined with -strparse and -out to extract data
4250 from a file (which may not be in ASN.1 format).
4253 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4254 when producing the local key id.
4255 [Richard Levitte <levitte@stacken.kth.se>]
4257 *) New option -dhparam in s_server. This allows a DH parameter file to be
4258 stated explicitly. If it is not stated then it tries the first server
4259 certificate file. The previous behaviour hard coded the filename
4263 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4264 a public key to be input or output. For example:
4265 openssl rsa -in key.pem -pubout -out pubkey.pem
4266 Also added necessary DSA public key functions to handle this.
4269 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4270 in the message. This was handled by allowing
4271 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4272 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4274 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4275 to the end of the strings whereas this didn't. This would cause problems
4276 if strings read with d2i_ASN1_bytes() were later modified.
4277 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4279 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4280 data and it contains EOF it will end up returning an error. This is
4281 caused by input 46 bytes long. The cause is due to the way base64
4282 BIOs find the start of base64 encoded data. They do this by trying a
4283 trial decode on each line until they find one that works. When they
4284 do a flag is set and it starts again knowing it can pass all the
4285 data directly through the decoder. Unfortunately it doesn't reset
4286 the context it uses. This means that if EOF is reached an attempt
4287 is made to pass two EOFs through the context and this causes the
4288 resulting error. This can also cause other problems as well. As is
4289 usual with these problems it takes *ages* to find and the fix is
4290 trivial: move one line.
4291 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4293 *) Ugly workaround to get s_client and s_server working under Windows. The
4294 old code wouldn't work because it needed to select() on sockets and the
4295 tty (for keypresses and to see if data could be written). Win32 only
4296 supports select() on sockets so we select() with a 1s timeout on the
4297 sockets and then see if any characters are waiting to be read, if none
4298 are present then we retry, we also assume we can always write data to
4299 the tty. This isn't nice because the code then blocks until we've
4300 received a complete line of data and it is effectively polling the
4301 keyboard at 1s intervals: however it's quite a bit better than not
4302 working at all :-) A dedicated Windows application might handle this
4303 with an event loop for example.
4306 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4307 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4308 will be called when RSA_sign() and RSA_verify() are used. This is useful
4309 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4310 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4311 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4312 This necessitated the support of an extra signature type NID_md5_sha1
4313 for SSL signatures and modifications to the SSL library to use it instead
4314 of calling RSA_public_decrypt() and RSA_private_encrypt().
4317 *) Add new -verify -CAfile and -CApath options to the crl program, these
4318 will lookup a CRL issuers certificate and verify the signature in a
4319 similar way to the verify program. Tidy up the crl program so it
4320 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4321 less strict. It will now permit CRL extensions even if it is not
4322 a V2 CRL: this will allow it to tolerate some broken CRLs.
4325 *) Initialize all non-automatic variables each time one of the openssl
4326 sub-programs is started (this is necessary as they may be started
4327 multiple times from the "OpenSSL>" prompt).
4328 [Lennart Bang, Bodo Moeller]
4330 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4331 removing all other RSA functionality (this is what NO_RSA does). This
4332 is so (for example) those in the US can disable those operations covered
4333 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4337 *) Non-copying interface to BIO pairs.
4338 (still largely untested)
4341 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4342 ASCII string. This was handled independently in various places before.
4345 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4346 UTF8 strings a character at a time.
4349 *) Use client_version from client hello to select the protocol
4350 (s23_srvr.c) and for RSA client key exchange verification
4351 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4354 *) Add various utility functions to handle SPKACs, these were previously
4355 handled by poking round in the structure internals. Added new function
4356 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4357 print, verify and generate SPKACs. Based on an original idea from
4358 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4361 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4364 *) Allow the config file extension section to be overwritten on the
4365 command line. Based on an original idea from Massimiliano Pala
4366 <madwolf@comune.modena.it>. The new option is called -extensions
4367 and can be applied to ca, req and x509. Also -reqexts to override
4368 the request extensions in req and -crlexts to override the crl extensions
4372 *) Add new feature to the SPKAC handling in ca. Now you can include
4373 the same field multiple times by preceding it by "XXXX." for example:
4376 this is the same syntax as used in the req config file.
4379 *) Allow certificate extensions to be added to certificate requests. These
4380 are specified in a 'req_extensions' option of the req section of the
4381 config file. They can be printed out with the -text option to req but
4382 are otherwise ignored at present.
4385 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4386 data read consists of only the final block it would not decrypted because
4387 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4388 A misplaced 'break' also meant the decrypted final block might not be
4389 copied until the next read.
4392 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4393 a few extra parameters to the DH structure: these will be useful if
4394 for example we want the value of 'q' or implement X9.42 DH.
4397 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4398 provides hooks that allow the default DSA functions or functions on a
4399 "per key" basis to be replaced. This allows hardware acceleration and
4400 hardware key storage to be handled without major modification to the
4401 library. Also added low level modexp hooks and CRYPTO_EX structure and
4402 associated functions.
4405 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4406 as "read only": it can't be written to and the buffer it points to will
4407 not be freed. Reading from a read only BIO is much more efficient than
4408 a normal memory BIO. This was added because there are several times when
4409 an area of memory needs to be read from a BIO. The previous method was
4410 to create a memory BIO and write the data to it, this results in two
4411 copies of the data and an O(n^2) reading algorithm. There is a new
4412 function BIO_new_mem_buf() which creates a read only memory BIO from
4413 an area of memory. Also modified the PKCS#7 routines to use read only
4417 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4418 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4419 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4420 but a retry condition occured while trying to read the rest.
4423 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4424 NID_pkcs7_encrypted by default: this was wrong since this should almost
4425 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4426 the encrypted data type: this is a more sensible place to put it and it
4427 allows the PKCS#12 code to be tidied up that duplicated this
4431 *) Changed obj_dat.pl script so it takes its input and output files on
4432 the command line. This should avoid shell escape redirection problems
4436 *) Initial support for certificate extension requests, these are included
4437 in things like Xenroll certificate requests. Included functions to allow
4438 extensions to be obtained and added.
4441 *) -crlf option to s_client and s_server for sending newlines as
4442 CRLF (as required by many protocols).
4445 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4447 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4448 [Ralf S. Engelschall]
4450 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4451 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4453 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4457 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4458 DH parameters/keys (q is lost during that conversion, but the resulting
4459 DH parameters contain its length).
4461 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4462 much faster than DH_generate_parameters (which creates parameters
4463 where p = 2*q + 1), and also the smaller q makes DH computations
4464 much more efficient (160-bit exponentiation instead of 1024-bit
4465 exponentiation); so this provides a convenient way to support DHE
4466 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4467 utter importance to use
4468 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4470 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4471 when such DH parameters are used, because otherwise small subgroup
4472 attacks may become possible!
4475 *) Avoid memory leak in i2d_DHparams.
4478 *) Allow the -k option to be used more than once in the enc program:
4479 this allows the same encrypted message to be read by multiple recipients.
4482 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4483 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4484 it will always use the numerical form of the OID, even if it has a short
4488 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4489 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4490 otherwise bn_mod_exp was called. In the case of hardware keys for example
4491 no private key components need be present and it might store extra data
4492 in the RSA structure, which cannot be accessed from bn_mod_exp.
4493 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4494 private key operations.
4497 *) Added support for SPARC Linux.
4500 *) pem_password_cb function type incompatibly changed from
4501 typedef int pem_password_cb(char *buf, int size, int rwflag);
4503 ....(char *buf, int size, int rwflag, void *userdata);
4504 so that applications can pass data to their callbacks:
4505 The PEM[_ASN1]_{read,write}... functions and macros now take an
4506 additional void * argument, which is just handed through whenever
4507 the password callback is called.
4508 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4510 New function SSL_CTX_set_default_passwd_cb_userdata.
4512 Compatibility note: As many C implementations push function arguments
4513 onto the stack in reverse order, the new library version is likely to
4514 interoperate with programs that have been compiled with the old
4515 pem_password_cb definition (PEM_whatever takes some data that
4516 happens to be on the stack as its last argument, and the callback
4517 just ignores this garbage); but there is no guarantee whatsoever that
4520 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4521 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4522 problems not only on Windows, but also on some Unix platforms.
4523 To avoid problematic command lines, these definitions are now in an
4524 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4525 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4528 *) MIPS III/IV assembler module is reimplemented.
4531 *) More DES library cleanups: remove references to srand/rand and
4532 delete an unused file.
4535 *) Add support for the the free Netwide assembler (NASM) under Win32,
4536 since not many people have MASM (ml) and it can be hard to obtain.
4537 This is currently experimental but it seems to work OK and pass all
4538 the tests. Check out INSTALL.W32 for info.
4541 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4542 without temporary keys kept an extra copy of the server key,
4543 and connections with temporary keys did not free everything in case
4547 *) New function RSA_check_key and new openssl rsa option -check
4548 for verifying the consistency of RSA keys.
4549 [Ulf Moeller, Bodo Moeller]
4551 *) Various changes to make Win32 compile work:
4552 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4553 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4554 comparison" warnings.
4555 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4558 *) Add a debugging option to PKCS#5 v2 key generation function: when
4559 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4560 derived keys are printed to stderr.
4563 *) Copy the flags in ASN1_STRING_dup().
4564 [Roman E. Pavlov <pre@mo.msk.ru>]
4566 *) The x509 application mishandled signing requests containing DSA
4567 keys when the signing key was also DSA and the parameters didn't match.
4569 It was supposed to omit the parameters when they matched the signing key:
4570 the verifying software was then supposed to automatically use the CA's
4571 parameters if they were absent from the end user certificate.
4573 Omitting parameters is no longer recommended. The test was also
4574 the wrong way round! This was probably due to unusual behaviour in
4575 EVP_cmp_parameters() which returns 1 if the parameters match.
4576 This meant that parameters were omitted when they *didn't* match and
4577 the certificate was useless. Certificates signed with 'ca' didn't have
4579 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4581 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4582 The interface is as follows:
4583 Applications can use
4584 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4585 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4586 "off" is now the default.
4587 The library internally uses
4588 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4589 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4590 to disable memory-checking temporarily.
4592 Some inconsistent states that previously were possible (and were
4593 even the default) are now avoided.
4595 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4596 with each memory chunk allocated; this is occasionally more helpful
4597 than just having a counter.
4599 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4601 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4605 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4606 which largely parallels "options", but is for changing API behaviour,
4607 whereas "options" are about protocol behaviour.
4608 Initial "mode" flags are:
4610 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4611 a single record has been written.
4612 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4613 retries use the same buffer location.
4614 (But all of the contents must be
4618 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4621 *) Fix problems with no-hmac etc.
4622 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4624 *) New functions RSA_get_default_method(), RSA_set_method() and
4625 RSA_get_method(). These allows replacement of RSA_METHODs without having
4626 to mess around with the internals of an RSA structure.
4629 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4630 Also really enable memory leak checks in openssl.c and in some
4632 [Chad C. Mulligan, Bodo Moeller]
4634 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4635 up the length of negative integers. This has now been simplified to just
4636 store the length when it is first determined and use it later, rather
4637 than trying to keep track of where data is copied and updating it to
4639 [Steve Henson, reported by Brien Wheeler
4640 <bwheeler@authentica-security.com>]
4642 *) Add a new function PKCS7_signatureVerify. This allows the verification
4643 of a PKCS#7 signature but with the signing certificate passed to the
4644 function itself. This contrasts with PKCS7_dataVerify which assumes the
4645 certificate is present in the PKCS#7 structure. This isn't always the
4646 case: certificates can be omitted from a PKCS#7 structure and be
4647 distributed by "out of band" means (such as a certificate database).
4650 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4651 function prototypes in pem.h, also change util/mkdef.pl to add the
4652 necessary function names.
4655 *) mk1mf.pl (used by Windows builds) did not properly read the
4656 options set by Configure in the top level Makefile, and Configure
4657 was not even able to write more than one option correctly.
4658 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4661 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4662 file to be loaded from a BIO or FILE pointer. The BIO version will
4663 for example allow memory BIOs to contain config info.
4666 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4667 Whoever hopes to achieve shared-library compatibility across versions
4668 must use this, not the compile-time macro.
4669 (Exercise 0.9.4: Which is the minimum library version required by
4671 Note: All this applies only to multi-threaded programs, others don't
4675 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4676 through a BIO pair triggered the default case, i.e.
4677 SSLerr(...,SSL_R_UNKNOWN_STATE).
4680 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4681 can use the SSL library even if none of the specific BIOs is
4685 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4686 for the encoded length.
4687 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4689 *) Add initial documentation of the X509V3 functions.
4692 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4693 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4694 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4695 secure PKCS#8 private key format with a high iteration count.
4698 *) Fix determination of Perl interpreter: A perl or perl5
4699 _directory_ in $PATH was also accepted as the interpreter.
4700 [Ralf S. Engelschall]
4702 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4703 wrong with it but it was very old and did things like calling
4704 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4708 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4709 to use the new extension code.
4712 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4713 with macros. This should make it easier to change their form, add extra
4714 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4718 *) Add to configuration table a new entry that can specify an alternative
4719 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4720 according to Mark Crispin <MRC@Panda.COM>.
4724 *) DES CBC did not update the IV. Weird.
4727 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4728 Changing the behaviour of the former might break existing programs --
4729 where IV updating is needed, des_ncbc_encrypt can be used.
4732 *) When bntest is run from "make test" it drives bc to check its
4733 calculations, as well as internally checking them. If an internal check
4734 fails, it needs to cause bc to give a non-zero result or make test carries
4735 on without noticing the failure. Fixed.
4738 *) DES library cleanups.
4741 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4742 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4743 ciphers. NOTE: although the key derivation function has been verified
4744 against some published test vectors it has not been extensively tested
4745 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4749 *) Instead of "mkdir -p", which is not fully portable, use new
4750 Perl script "util/mkdir-p.pl".
4753 *) Rewrite the way password based encryption (PBE) is handled. It used to
4754 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4755 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4756 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4757 the 'parameter' field of the AlgorithmIdentifier is passed to the
4758 underlying key generation function so it must do its own ASN1 parsing.
4759 This has also changed the EVP_PBE_CipherInit() function which now has a
4760 'parameter' argument instead of literal salt and iteration count values
4761 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4764 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4765 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4766 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4767 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4768 value was just used as a "magic string" and not used directly its
4769 value doesn't matter.
4772 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4776 *) "linux-sparc64" configuration (ultrapenguin).
4777 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4778 "linux-sparc" configuration.
4779 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4781 *) config now generates no-xxx options for missing ciphers.
4784 *) Support the EBCDIC character set (work in progress).
4785 File ebcdic.c not yet included because it has a different license.
4786 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4788 *) Support BS2000/OSD-POSIX.
4789 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4791 *) Make callbacks for key generation use void * instead of char *.
4794 *) Make S/MIME samples compile (not yet tested).
4797 *) Additional typesafe stacks.
4800 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4804 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4806 *) New configuration variant "sco5-gcc".
4808 *) Updated some demos.
4809 [Sean O Riordain, Wade Scholine]
4811 *) Add missing BIO_free at exit of pkcs12 application.
4814 *) Fix memory leak in conf.c.
4817 *) Updates for Win32 to assembler version of MD5.
4820 *) Set #! path to perl in apps/der_chop to where we found it
4821 instead of using a fixed path.
4824 *) SHA library changes for irix64-mips4-cc.
4827 *) Improvements for VMS support.
4831 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4833 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4834 This also avoids the problems with SC4.2 and unpatched SC5.
4835 [Andy Polyakov <appro@fy.chalmers.se>]
4837 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4838 These are required because of the typesafe stack would otherwise break
4839 existing code. If old code used a structure member which used to be STACK
4840 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4841 sk_num or sk_value it would produce an error because the num, data members
4842 are not present in STACK_OF. Now it just produces a warning. sk_set
4843 replaces the old method of assigning a value to sk_value
4844 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4845 that does this will no longer work (and should use sk_set instead) but
4846 this could be regarded as a "questionable" behaviour anyway.
4849 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4850 correctly handle encrypted S/MIME data.
4853 *) Change type of various DES function arguments from des_cblock
4854 (which means, in function argument declarations, pointer to char)
4855 to des_cblock * (meaning pointer to array with 8 char elements),
4856 which allows the compiler to do more typechecking; it was like
4857 that back in SSLeay, but with lots of ugly casts.
4859 Introduce new type const_des_cblock.
4862 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4863 problems: find RecipientInfo structure that matches recipient certificate
4864 and initialise the ASN1 structures properly based on passed cipher.
4867 *) Belatedly make the BN tests actually check the results.
4870 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4871 to and from BNs: it was completely broken. New compilation option
4872 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4873 key elements as negative integers.
4876 *) Reorganize and speed up MD5.
4877 [Andy Polyakov <appro@fy.chalmers.se>]
4880 [Richard Levitte <richard@levitte.org>]
4882 *) New option -out to asn1parse to allow the parsed structure to be
4883 output to a file. This is most useful when combined with the -strparse
4884 option to examine the output of things like OCTET STRINGS.
4887 *) Make SSL library a little more fool-proof by not requiring any longer
4888 that SSL_set_{accept,connect}_state be called before
4889 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4890 in many applications because usually everything *appeared* to work as
4891 intended anyway -- now it really works as intended).
4894 *) Move openssl.cnf out of lib/.
4897 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4898 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4899 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4900 [Ralf S. Engelschall]
4902 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4903 handle PKCS#7 enveloped data properly.
4904 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4906 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4907 copying pointers. The cert_st handling is changed by this in
4908 various ways (and thus what used to be known as ctx->default_cert
4909 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4910 any longer when s->cert does not give us what we need).
4911 ssl_cert_instantiate becomes obsolete by this change.
4912 As soon as we've got the new code right (possibly it already is?),
4913 we have solved a couple of bugs of the earlier code where s->cert
4914 was used as if it could not have been shared with other SSL structures.
4916 Note that using the SSL API in certain dirty ways now will result
4917 in different behaviour than observed with earlier library versions:
4918 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4919 does not influence s as it used to.
4921 In order to clean up things more thoroughly, inside SSL_SESSION
4922 we don't use CERT any longer, but a new structure SESS_CERT
4923 that holds per-session data (if available); currently, this is
4924 the peer's certificate chain and, for clients, the server's certificate
4925 and temporary key. CERT holds only those values that can have
4926 meaningful defaults in an SSL_CTX.
4929 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4930 from the internal representation. Various PKCS#7 fixes: remove some
4931 evil casts and set the enc_dig_alg field properly based on the signing
4935 *) Allow PKCS#12 password to be set from the command line or the
4936 environment. Let 'ca' get its config file name from the environment
4937 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4941 *) Allow certificate policies extension to use an IA5STRING for the
4942 organization field. This is contrary to the PKIX definition but
4943 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4947 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4948 without disallowing inline assembler and the like for non-pedantic builds.
4951 *) Support Borland C++ builder.
4952 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4957 *) SHA-1 cleanups and performance enhancements.
4958 [Andy Polyakov <appro@fy.chalmers.se>]
4960 *) Sparc v8plus assembler for the bignum library.
4961 [Andy Polyakov <appro@fy.chalmers.se>]
4963 *) Accept any -xxx and +xxx compiler options in Configure.
4966 *) Update HPUX configuration.
4969 *) Add missing sk_<type>_unshift() function to safestack.h
4970 [Ralf S. Engelschall]
4972 *) New function SSL_CTX_use_certificate_chain_file that sets the
4973 "extra_cert"s in addition to the certificate. (This makes sense
4974 only for "PEM" format files, as chains as a whole are not
4978 *) Support verify_depth from the SSL API.
4979 x509_vfy.c had what can be considered an off-by-one-error:
4980 Its depth (which was not part of the external interface)
4981 was actually counting the number of certificates in a chain;
4982 now it really counts the depth.
4985 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4986 instead of X509err, which often resulted in confusing error
4987 messages since the error codes are not globally unique
4988 (e.g. an alleged error in ssl3_accept when a certificate
4989 didn't match the private key).
4991 *) New function SSL_CTX_set_session_id_context that allows to set a default
4992 value (so that you don't need SSL_set_session_id_context for each
4993 connection using the SSL_CTX).
4996 *) OAEP decoding bug fix.
4999 *) Support INSTALL_PREFIX for package builders, as proposed by
5003 *) New Configure options "threads" and "no-threads". For systems
5004 where the proper compiler options are known (currently Solaris
5005 and Linux), "threads" is the default.
5008 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
5011 *) Install various scripts to $(OPENSSLDIR)/misc, not to
5012 $(INSTALLTOP)/bin -- they shouldn't clutter directories
5013 such as /usr/local/bin.
5016 *) "make linux-shared" to build shared libraries.
5017 [Niels Poppe <niels@netbox.org>]
5019 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
5022 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
5023 extension adding in x509 utility.
5026 *) Remove NOPROTO sections and error code comments.
5029 *) Partial rewrite of the DEF file generator to now parse the ANSI
5033 *) New Configure options --prefix=DIR and --openssldir=DIR.
5036 *) Complete rewrite of the error code script(s). It is all now handled
5037 by one script at the top level which handles error code gathering,
5038 header rewriting and C source file generation. It should be much better
5039 than the old method: it now uses a modified version of Ulf's parser to
5040 read the ANSI prototypes in all header files (thus the old K&R definitions
5041 aren't needed for error creation any more) and do a better job of
5042 translating function codes into names. The old 'ASN1 error code imbedded
5043 in a comment' is no longer necessary and it doesn't use .err files which
5044 have now been deleted. Also the error code call doesn't have to appear all
5045 on one line (which resulted in some large lines...).
5048 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
5051 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
5052 0 (which usually indicates a closed connection), but continue reading.
5055 *) Fix some race conditions.
5058 *) Add support for CRL distribution points extension. Add Certificate
5059 Policies and CRL distribution points documentation.
5062 *) Move the autogenerated header file parts to crypto/opensslconf.h.
5065 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
5066 8 of keying material. Merlin has also confirmed interop with this fix
5067 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
5068 [Merlin Hughes <merlin@baltimore.ie>]
5070 *) Fix lots of warnings.
5071 [Richard Levitte <levitte@stacken.kth.se>]
5073 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
5074 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
5075 [Richard Levitte <levitte@stacken.kth.se>]
5077 *) Fix problems with sizeof(long) == 8.
5078 [Andy Polyakov <appro@fy.chalmers.se>]
5080 *) Change functions to ANSI C.
5083 *) Fix typos in error codes.
5084 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
5086 *) Remove defunct assembler files from Configure.
5089 *) SPARC v8 assembler BIGNUM implementation.
5090 [Andy Polyakov <appro@fy.chalmers.se>]
5092 *) Support for Certificate Policies extension: both print and set.
5093 Various additions to support the r2i method this uses.
5096 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5097 return a const string when you are expecting an allocated buffer.
5100 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5101 types DirectoryString and DisplayText.
5104 *) Add code to allow r2i extensions to access the configuration database,
5105 add an LHASH database driver and add several ctx helper functions.
5108 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5109 fail when they extended the size of a BIGNUM.
5112 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5113 support typesafe stack.
5116 *) Fix typo in SSL_[gs]et_options().
5117 [Nils Frostberg <nils@medcom.se>]
5119 *) Delete various functions and files that belonged to the (now obsolete)
5120 old X509V3 handling code.
5123 *) New Configure option "rsaref".
5126 *) Don't auto-generate pem.h.
5129 *) Introduce type-safe ASN.1 SETs.
5132 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5133 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5135 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5136 that links with OpenSSL (well at least cause lots of warnings), but fear
5137 not: the conversion is trivial, and it eliminates loads of evil casts. A
5138 few STACKed things have been converted already. Feel free to convert more.
5139 In the fullness of time, I'll do away with the STACK type altogether.
5142 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5143 specified in <certfile> by updating the entry in the index.txt file.
5144 This way one no longer has to edit the index.txt file manually for
5145 revoking a certificate. The -revoke option does the gory details now.
5146 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5148 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5149 `-text' option at all and this way the `-noout -text' combination was
5150 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5151 [Ralf S. Engelschall]
5153 *) Make sure a corresponding plain text error message exists for the
5154 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5155 verify callback function determined that a certificate was revoked.
5156 [Ralf S. Engelschall]
5158 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5159 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5160 all available cipers including rc5, which was forgotten until now.
5161 In order to let the testing shell script know which algorithms
5162 are available, a new (up to now undocumented) command
5163 "openssl list-cipher-commands" is used.
5166 *) Bugfix: s_client occasionally would sleep in select() when
5167 it should have checked SSL_pending() first.
5170 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5171 the raw DSA values prior to ASN.1 encoding.
5174 *) Tweaks to Configure
5175 [Niels Poppe <niels@netbox.org>]
5177 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5181 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5184 *) New config option to avoid instructions that are illegal on the 80386.
5185 The default code is faster, but requires at least a 486.
5188 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5189 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5190 same as SSL2_VERSION anyway.
5193 *) New "-showcerts" option for s_client.
5196 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5197 application. Various cleanups and fixes.
5200 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5201 modify error routines to work internally. Add error codes and PBE init
5202 to library startup routines.
5205 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5206 packing functions to asn1 and evp. Changed function names and error
5207 codes along the way.
5210 *) PKCS12 integration: and so it begins... First of several patches to
5211 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5212 objects to objects.h
5215 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5216 and display support for Thawte strong extranet extension.
5219 *) Add LinuxPPC support.
5220 [Jeff Dubrule <igor@pobox.org>]
5222 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5223 bn_div_words in alpha.s.
5224 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5226 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5227 OAEP isn't supported when OpenSSL is built with RSAref.
5228 [Ulf Moeller <ulf@fitug.de>]
5230 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5231 so they no longer are missing under -DNOPROTO.
5232 [Soren S. Jorvang <soren@t.dk>]
5235 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5237 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5238 doesn't work when the session is reused. Coming soon!
5241 *) Fix a security hole, that allows sessions to be reused in the wrong
5242 context thus bypassing client cert protection! All software that uses
5243 client certs and session caches in multiple contexts NEEDS PATCHING to
5244 allow session reuse! A fuller solution is in the works.
5245 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5247 *) Some more source tree cleanups (removed obsolete files
5248 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5249 permission on "config" script to be executable) and a fix for the INSTALL
5251 [Ulf Moeller <ulf@fitug.de>]
5253 *) Remove some legacy and erroneous uses of malloc, free instead of
5255 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5257 *) Make rsa_oaep_test return non-zero on error.
5258 [Ulf Moeller <ulf@fitug.de>]
5260 *) Add support for native Solaris shared libraries. Configure
5261 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5262 if someone would make that last step automatic.
5263 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5265 *) ctx_size was not built with the right compiler during "make links". Fixed.
5268 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5269 except NULL ciphers". This means the default cipher list will no longer
5270 enable NULL ciphers. They need to be specifically enabled e.g. with
5271 the string "DEFAULT:eNULL".
5274 *) Fix to RSA private encryption routines: if p < q then it would
5275 occasionally produce an invalid result. This will only happen with
5276 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5279 *) Be less restrictive and allow also `perl util/perlpath.pl
5280 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5281 because this way one can also use an interpreter named `perl5' (which is
5282 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5283 installed as `perl').
5284 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5286 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5287 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5289 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5290 advapi32.lib to Win32 build and change the pem test comparision
5291 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5292 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5293 and crypto/des/ede_cbcm_enc.c.
5296 *) DES quad checksum was broken on big-endian architectures. Fixed.
5299 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5300 Win32 test batch file so it (might) work again. The Win32 test batch file
5301 is horrible: I feel ill....
5304 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5305 in e_os.h. Audit of header files to check ANSI and non ANSI
5306 sections: 10 functions were absent from non ANSI section and not exported
5307 from Windows DLLs. Fixed up libeay.num for new functions.
5310 *) Make `openssl version' output lines consistent.
5311 [Ralf S. Engelschall]
5313 *) Fix Win32 symbol export lists for BIO functions: Added
5314 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5315 to ms/libeay{16,32}.def.
5316 [Ralf S. Engelschall]
5318 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5319 fine under Unix and passes some trivial tests I've now added. But the
5320 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5321 added to make sure no one expects that this stuff really works in the
5322 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5323 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5325 [Ralf S. Engelschall]
5327 *) Fix the generation of two part addresses in perl.
5328 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5330 *) Add config entry for Linux on MIPS.
5331 [John Tobey <jtobey@channel1.com>]
5333 *) Make links whenever Configure is run, unless we are on Windoze.
5336 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5337 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5341 *) Add a useful kludge to allow package maintainers to specify compiler and
5342 other platforms details on the command line without having to patch the
5343 Configure script everytime: One now can use ``perl Configure
5344 <id>:<details>'', i.e. platform ids are allowed to have details appended
5345 to them (seperated by colons). This is treated as there would be a static
5346 pre-configured entry in Configure's %table under key <id> with value
5347 <details> and ``perl Configure <id>'' is called. So, when you want to
5348 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5349 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5350 now, which overrides the FreeBSD-elf entry on-the-fly.
5351 [Ralf S. Engelschall]
5353 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5356 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5357 on the `perl Configure ...' command line. This way one can compile
5358 OpenSSL libraries with Position Independent Code (PIC) which is needed
5359 for linking it into DSOs.
5360 [Ralf S. Engelschall]
5362 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5366 *) Cleaned up the LICENSE document: The official contact for any license
5367 questions now is the OpenSSL core team under openssl-core@openssl.org.
5368 And add a paragraph about the dual-license situation to make sure people
5369 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5370 to the OpenSSL toolkit.
5371 [Ralf S. Engelschall]
5373 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5374 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5375 Additonally cleaned up the `make links' target: Remove unnecessary
5376 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5377 to speed processing and no longer clutter the display with confusing
5378 stuff. Instead only the actually done links are displayed.
5379 [Ralf S. Engelschall]
5381 *) Permit null encryption ciphersuites, used for authentication only. It used
5382 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5383 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5387 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5388 signed attributes when verifying signatures (this would break them),
5389 the detached data encoding was wrong and public keys obtained using
5390 X509_get_pubkey() weren't freed.
5393 *) Add text documentation for the BUFFER functions. Also added a work around
5394 to a Win95 console bug. This was triggered by the password read stuff: the
5395 last character typed gets carried over to the next fread(). If you were
5396 generating a new cert request using 'req' for example then the last
5397 character of the passphrase would be CR which would then enter the first
5401 *) Added the new `Includes OpenSSL Cryptography Software' button as
5402 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5403 button and can be used by applications based on OpenSSL to show the
5404 relationship to the OpenSSL project.
5405 [Ralf S. Engelschall]
5407 *) Remove confusing variables in function signatures in files
5408 ssl/ssl_lib.c and ssl/ssl.h.
5409 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5411 *) Don't install bss_file.c under PREFIX/include/
5412 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5414 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5415 functions that return function pointers and has support for NT specific
5416 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5417 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5418 unsigned to signed types: this was killing the Win32 compile.
5421 *) Add new certificate file to stack functions,
5422 SSL_add_dir_cert_subjects_to_stack() and
5423 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5424 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5425 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5426 This means that Apache-SSL and similar packages don't have to mess around
5427 to add as many CAs as they want to the preferred list.
5430 *) Experiment with doxygen documentation. Currently only partially applied to
5432 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5433 openssl.doxy as the configuration file.
5436 *) Get rid of remaining C++-style comments which strict C compilers hate.
5437 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5439 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5440 compiled in by default: it has problems with large keys.
5443 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5444 DH private keys and/or callback functions which directly correspond to
5445 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5446 is needed for applications which have to configure certificates on a
5447 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5449 For the RSA certificate situation is makes no difference, but
5450 for the DSA certificate situation this fixes the "no shared cipher"
5451 problem where the OpenSSL cipher selection procedure failed because the
5452 temporary keys were not overtaken from the context and the API provided
5453 no way to reconfigure them.
5454 The new functions now let applications reconfigure the stuff and they
5455 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5456 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5457 non-public-API function ssl_cert_instantiate() is used as a helper
5458 function and also to reduce code redundancy inside ssl_rsa.c.
5459 [Ralf S. Engelschall]
5461 *) Move s_server -dcert and -dkey options out of the undocumented feature
5462 area because they are useful for the DSA situation and should be
5463 recognized by the users.
5464 [Ralf S. Engelschall]
5466 *) Fix the cipher decision scheme for export ciphers: the export bits are
5467 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5468 SSL_EXP_MASK. So, the original variable has to be used instead of the
5469 already masked variable.
5470 [Richard Levitte <levitte@stacken.kth.se>]
5472 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5473 [Richard Levitte <levitte@stacken.kth.se>]
5475 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5476 from `int' to `unsigned int' because it's a length and initialized by
5477 EVP_DigestFinal() which expects an `unsigned int *'.
5478 [Richard Levitte <levitte@stacken.kth.se>]
5480 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5481 script. Instead use the usual Shell->Perl transition trick.
5482 [Ralf S. Engelschall]
5484 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5485 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5486 -noout -modulus' as it's already the case for `openssl rsa -noout
5487 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5488 currently the public key is printed (a decision which was already done by
5489 `openssl dsa -modulus' in the past) which serves a similar purpose.
5490 Additionally the NO_RSA no longer completely removes the whole -modulus
5491 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5493 [Ralf S. Engelschall]
5495 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5496 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5497 [Arne Ansper <arne@ats.cyber.ee>]
5499 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5500 to be added. Now both 'req' and 'ca' can use new objects defined in the
5504 *) Add cool BIO that does syslog (or event log on NT).
5505 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5507 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5508 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5509 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5510 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5513 *) Add preliminary config info for new extension code.
5516 *) Make RSA_NO_PADDING really use no padding.
5517 [Ulf Moeller <ulf@fitug.de>]
5519 *) Generate errors when private/public key check is done.
5522 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5523 for some CRL extensions and new objects added.
5526 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5527 key usage extension and fuller support for authority key id.
5530 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5531 padding method for RSA, which is recommended for new applications in PKCS
5532 #1 v2.0 (RFC 2437, October 1998).
5533 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5534 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5535 against Bleichbacher's attack on RSA.
5536 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5539 *) Updates to the new SSL compression code
5540 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5542 *) Fix so that the version number in the master secret, when passed
5543 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5544 (because the server will not accept higher), that the version number
5545 is 0x03,0x01, not 0x03,0x00
5546 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5548 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5549 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5550 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5553 *) Support for RAW extensions where an arbitrary extension can be
5554 created by including its DER encoding. See apps/openssl.cnf for
5558 *) Make sure latest Perl versions don't interpret some generated C array
5559 code as Perl array code in the crypto/err/err_genc.pl script.
5560 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5562 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5563 not many people have the assembler. Various Win32 compilation fixes and
5564 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5568 *) Modify configure script 'Configure' to automatically create crypto/date.h
5569 file under Win32 and also build pem.h from pem.org. New script
5570 util/mkfiles.pl to create the MINFO file on environments that can't do a
5571 'make files': perl util/mkfiles.pl >MINFO should work.
5574 *) Major rework of DES function declarations, in the pursuit of correctness
5575 and purity. As a result, many evil casts evaporated, and some weirdness,
5576 too. You may find this causes warnings in your code. Zapping your evil
5577 casts will probably fix them. Mostly.
5580 *) Fix for a typo in asn1.h. Bug fix to object creation script
5581 obj_dat.pl. It considered a zero in an object definition to mean
5582 "end of object": none of the objects in objects.h have any zeros
5583 so it wasn't spotted.
5584 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5586 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5587 Masking (CBCM). In the absence of test vectors, the best I have been able
5588 to do is check that the decrypt undoes the encrypt, so far. Send me test
5589 vectors if you have them.
5592 *) Correct calculation of key length for export ciphers (too much space was
5593 allocated for null ciphers). This has not been tested!
5596 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5597 message is now correct (it understands "crypto" and "ssl" on its
5598 command line). There is also now an "update" option. This will update
5599 the util/ssleay.num and util/libeay.num files with any new functions.
5601 perl util/mkdef.pl crypto ssl update
5602 it will update them.
5605 *) Overhauled the Perl interface (perl/*):
5606 - ported BN stuff to OpenSSL's different BN library
5607 - made the perl/ source tree CVS-aware
5608 - renamed the package from SSLeay to OpenSSL (the files still contain
5609 their history because I've copied them in the repository)
5610 - removed obsolete files (the test scripts will be replaced
5611 by better Test::Harness variants in the future)
5612 [Ralf S. Engelschall]
5614 *) First cut for a very conservative source tree cleanup:
5615 1. merge various obsolete readme texts into doc/ssleay.txt
5616 where we collect the old documents and readme texts.
5617 2. remove the first part of files where I'm already sure that we no
5618 longer need them because of three reasons: either they are just temporary
5619 files which were left by Eric or they are preserved original files where
5620 I've verified that the diff is also available in the CVS via "cvs diff
5621 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5622 the crypto/md/ stuff).
5623 [Ralf S. Engelschall]
5625 *) More extension code. Incomplete support for subject and issuer alt
5626 name, issuer and authority key id. Change the i2v function parameters
5627 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5628 what that's for :-) Fix to ASN1 macro which messed up
5629 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5632 *) Preliminary support for ENUMERATED type. This is largely copied from the
5636 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5637 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5639 *) Make sure `make rehash' target really finds the `openssl' program.
5640 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5642 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5643 like to hear about it if this slows down other processors.
5646 *) Add CygWin32 platform information to Configure script.
5647 [Alan Batie <batie@aahz.jf.intel.com>]
5649 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5650 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5652 *) New program nseq to manipulate netscape certificate sequences
5655 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5659 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5660 but the BN code had some problems that would cause failures when
5661 doing certificate verification and some other functions.
5662 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5664 *) Add ASN1 and PEM code to support netscape certificate sequences.
5667 *) Add ASN1 and PEM code to support netscape certificate sequences.
5670 *) Add several PKIX and private extended key usage OIDs.
5673 *) Modify the 'ca' program to handle the new extension code. Modify
5674 openssl.cnf for new extension format, add comments.
5677 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5678 and add a sample to openssl.cnf so req -x509 now adds appropriate
5682 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5683 error code, add initial support to X509_print() and x509 application.
5686 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5687 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5688 stuff is currently isolated and isn't even compiled yet.
5691 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5692 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5693 Removed the versions check from X509 routines when loading extensions:
5694 this allows certain broken certificates that don't set the version
5695 properly to be processed.
5698 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5699 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5700 can still be regenerated with "make depend".
5703 *) Spelling mistake in C version of CAST-128.
5704 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5706 *) Changes to the error generation code. The perl script err-code.pl
5707 now reads in the old error codes and retains the old numbers, only
5708 adding new ones if necessary. It also only changes the .err files if new
5709 codes are added. The makefiles have been modified to only insert errors
5710 when needed (to avoid needlessly modifying header files). This is done
5711 by only inserting errors if the .err file is newer than the auto generated
5712 C file. To rebuild all the error codes from scratch (the old behaviour)
5713 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5714 or delete all the .err files.
5717 *) CAST-128 was incorrectly implemented for short keys. The C version has
5718 been fixed, but is untested. The assembler versions are also fixed, but
5719 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5720 to regenerate it if needed.
5721 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5722 Hagino <itojun@kame.net>]
5724 *) File was opened incorrectly in randfile.c.
5725 [Ulf Möller <ulf@fitug.de>]
5727 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5728 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5729 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5730 al: it's just almost always a UTCTime. Note this patch adds new error
5731 codes so do a "make errors" if there are problems.
5734 *) Correct Linux 1 recognition in config.
5735 [Ulf Möller <ulf@fitug.de>]
5737 *) Remove pointless MD5 hash when using DSA keys in ca.
5738 [Anonymous <nobody@replay.com>]
5740 *) Generate an error if given an empty string as a cert directory. Also
5741 generate an error if handed NULL (previously returned 0 to indicate an
5742 error, but didn't set one).
5743 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5745 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5748 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5749 parameters. This was causing a warning which killed off the Win32 compile.
5752 *) Remove C++ style comments from crypto/bn/bn_local.h.
5753 [Neil Costigan <neil.costigan@celocom.com>]
5755 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5756 based on a text string, looking up short and long names and finally
5757 "dot" format. The "dot" format stuff didn't work. Added new function
5758 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5759 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5760 OID is not part of the table.
5763 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5764 X509_LOOKUP_by_alias().
5767 *) Sort openssl functions by name.
5770 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5771 encryption from sample DSA keys (in case anyone is interested the password
5775 *) Make _all_ *_free functions accept a NULL pointer.
5776 [Frans Heymans <fheymans@isaserver.be>]
5778 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5780 [Anonymous <nobody@replay.com>]
5782 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5783 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5785 *) Don't blow it for numeric -newkey arguments to apps/req.
5786 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5788 *) Temp key "for export" tests were wrong in s3_srvr.c.
5789 [Anonymous <nobody@replay.com>]
5791 *) Add prototype for temp key callback functions
5792 SSL_CTX_set_tmp_{rsa,dh}_callback().
5795 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5796 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5799 *) X509_name_add_entry() freed the wrong thing after an error.
5800 [Arne Ansper <arne@ats.cyber.ee>]
5802 *) rsa_eay.c would attempt to free a NULL context.
5803 [Arne Ansper <arne@ats.cyber.ee>]
5805 *) BIO_s_socket() had a broken should_retry() on Windoze.
5806 [Arne Ansper <arne@ats.cyber.ee>]
5808 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5809 [Arne Ansper <arne@ats.cyber.ee>]
5811 *) Make sure the already existing X509_STORE->depth variable is initialized
5812 in X509_STORE_new(), but document the fact that this variable is still
5813 unused in the certificate verification process.
5814 [Ralf S. Engelschall]
5816 *) Fix the various library and apps files to free up pkeys obtained from
5817 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5820 *) Fix reference counting in X509_PUBKEY_get(). This makes
5821 demos/maurice/example2.c work, amongst others, probably.
5822 [Steve Henson and Ben Laurie]
5824 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5825 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5826 are no longer created. This way we have a single and consistent command
5827 line interface `openssl <command>', similar to `cvs <command>'.
5828 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5830 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5831 BIT STRING wrapper always have zero unused bits.
5834 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5837 *) Make the top-level INSTALL documentation easier to understand.
5840 *) Makefiles updated to exit if an error occurs in a sub-directory
5841 make (including if user presses ^C) [Paul Sutton]
5843 *) Make Montgomery context stuff explicit in RSA data structure.
5846 *) Fix build order of pem and err to allow for generated pem.h.
5849 *) Fix renumbering bug in X509_NAME_delete_entry().
5852 *) Enhanced the err-ins.pl script so it makes the error library number
5853 global and can add a library name. This is needed for external ASN1 and
5854 other error libraries.
5857 *) Fixed sk_insert which never worked properly.
5860 *) Fix ASN1 macros so they can handle indefinite length construted
5861 EXPLICIT tags. Some non standard certificates use these: they can now
5865 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5866 into a single doc/ssleay.txt bundle. This way the information is still
5867 preserved but no longer messes up this directory. Now it's new room for
5868 the new set of documenation files.
5869 [Ralf S. Engelschall]
5871 *) SETs were incorrectly DER encoded. This was a major pain, because they
5872 shared code with SEQUENCEs, which aren't coded the same. This means that
5873 almost everything to do with SETs or SEQUENCEs has either changed name or
5874 number of arguments.
5875 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5877 *) Fix test data to work with the above.
5880 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5881 was already fixed by Eric for 0.9.1 it seems.
5882 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5884 *) Autodetect FreeBSD3.
5887 *) Fix various bugs in Configure. This affects the following platforms:
5891 unixware-2.0-pentium
5895 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5896 before they are needed.
5899 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5903 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5905 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5906 changed SSLeay to OpenSSL in version strings.
5907 [Ralf S. Engelschall]
5909 *) Some fixups to the top-level documents.
5912 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5913 because the symlink to include/ was missing.
5914 [Ralf S. Engelschall]
5916 *) Incorporated the popular no-RSA/DSA-only patches
5917 which allow to compile a RSA-free SSLeay.
5918 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5920 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5921 when "ssleay" is still not found.
5922 [Ralf S. Engelschall]
5924 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5925 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5927 *) Updated the README file.
5928 [Ralf S. Engelschall]
5930 *) Added various .cvsignore files in the CVS repository subdirs
5931 to make a "cvs update" really silent.
5932 [Ralf S. Engelschall]
5934 *) Recompiled the error-definition header files and added
5935 missing symbols to the Win32 linker tables.
5936 [Ralf S. Engelschall]
5938 *) Cleaned up the top-level documents;
5939 o new files: CHANGES and LICENSE
5940 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5941 o merged COPYRIGHT into LICENSE
5942 o removed obsolete TODO file
5943 o renamed MICROSOFT to INSTALL.W32
5944 [Ralf S. Engelschall]
5946 *) Removed dummy files from the 0.9.1b source tree:
5947 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5948 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5949 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5950 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5951 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5952 [Ralf S. Engelschall]
5954 *) Added various platform portability fixes.
5957 *) The Genesis of the OpenSSL rpject:
5958 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5959 Young and Tim J. Hudson created while they were working for C2Net until
5961 [The OpenSSL Project]
5964 Changes between 0.9.0b and 0.9.1b [not released]
5966 *) Updated a few CA certificates under certs/
5969 *) Changed some BIGNUM api stuff.
5972 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5973 DGUX x86, Linux Alpha, etc.
5976 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5977 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5981 *) Add -strparse option to asn1pars program which parses nested
5983 [Dr Stephen Henson <shenson@bigfoot.com>]
5985 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5988 *) DSA fix for "ca" program.
5991 *) Added "-genkey" option to "dsaparam" program.
5994 *) Added RIPE MD160 (rmd160) message digest.
5997 *) Added -a (all) option to "ssleay version" command.
6000 *) Added PLATFORM define which is the id given to Configure.
6003 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
6006 *) Extended the ASN.1 parser routines.
6009 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
6012 *) Added a BN_CTX to the BN library.
6015 *) Fixed the weak key values in DES library
6018 *) Changed API in EVP library for cipher aliases.
6021 *) Added support for RC2/64bit cipher.
6024 *) Converted the lhash library to the crypto/mem.c functions.
6027 *) Added more recognized ASN.1 object ids.
6030 *) Added more RSA padding checks for SSL/TLS.
6033 *) Added BIO proxy/filter functionality.
6036 *) Added extra_certs to SSL_CTX which can be used
6037 send extra CA certificates to the client in the CA cert chain sending
6038 process. It can be configured with SSL_CTX_add_extra_chain_cert().
6041 *) Now Fortezza is denied in the authentication phase because
6042 this is key exchange mechanism is not supported by SSLeay at all.
6045 *) Additional PKCS1 checks.
6048 *) Support the string "TLSv1" for all TLS v1 ciphers.
6051 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
6052 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
6055 *) Fixed a few memory leaks.
6058 *) Fixed various code and comment typos.
6061 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
6062 bytes sent in the client random.
6063 [Edward Bishop <ebishop@spyglass.com>]