5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
7 *) Add 'nid' and 'asn1_flag' members to EC_GROUP, and functions
9 EC_GROUP_get_set_asn1_flag()
10 EC_GROUP_get_get_asn1_flag()
11 'nid' is an optional NID for named curves. 'asn1_flag'
12 determines the encoding to be used for ASN1 purposes.
13 [Nils Larsch <nla@trustcenter.de>]
15 *) Add 'field_type' member to EC_METHOD, which holds the NID
16 of the appropriate field type OID. The new function
17 EC_METHOD_get_field_type() returns this value.
18 [Nils Larsch <nla@trustcenter.de>]
25 providing useful interfaces to EC_POINT_point2oct() and
27 [Nils Larsch <nla@trustcenter.de>]
29 *) Change internals of the EC library so that the functions
30 EC_GROUP_set_generator()
31 EC_GROUP_get_generator()
33 EC_GROUP_get_cofactor()
34 are implemented directly in crypto/ec/ec_lib.c and not dispatched
35 to methods, which would lead to unnecessary code duplication when
36 adding different types of curves.
37 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
39 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
40 arithmetic, and such that modified wNAFs are generated
41 (which avoid length expansion in many cases).
44 *) Add a function EC_GROUP_check_discriminant() (defined via
45 EC_METHOD) that verifies that the curve discriminant is non-zero.
47 Add a function EC_GROUP_check() that makes some sanity tests
48 on a EC_GROUP, its generator and order. This includes
49 EC_GROUP_check_discriminant().
50 [Nils Larsch <nla@trustcenter.de>]
52 *) Add ECDSA in new directory crypto/ecdsa/.
54 Add applications 'openssl ecdsaparam' and 'openssl ecdsa'
55 (these are variants of 'openssl dsaparam' and 'openssl dsa').
57 ECDSA support is also included in various other files across the
58 library. Most notably,
59 - 'openssl req' now has a '-newkey ecdsa:file' option;
60 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
61 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
62 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
63 them suitable for ECDSA where domain parameters must be
64 extracted before the specific public key.
65 [Nils Larsch <nla@trustcenter.de>]
67 *) Include some named elliptic curves, and add OIDs from X9.62,
68 SECG, and WAP/WTLS. The curves can be obtained from the new
71 EC_GROUP_new_by_name()
72 Also add a 'nid' field to EC_GROUP objects, which can be accessed
76 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
78 Changes between 0.9.6d and 0.9.7 [XX xxx 2002]
80 *) Add AES modes CFB and OFB to the object database. Correct an
81 error in AES-CFB decryption.
84 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
85 allows existing EVP_CIPHER_CTX structures to be reused after
86 calling EVP_*Final(). This behaviour is used by encryption
87 BIOs and some applications. This has the side effect that
88 applications must explicitly clean up cipher contexts with
89 EVP_CIPHER_CTX_cleanup() or they will leak memory.
92 *) Check the values of dna and dnb in bn_mul_recursive before calling
93 bn_mul_comba (a non zero value means the a or b arrays do not contain
94 n2 elements) and fallback to bn_mul_normal if either is not zero.
97 *) Fix escaping of non-ASCII characters when using the -subj option
98 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
101 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
102 form for "surname", serialNumber has no short form.
103 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
104 therefore remove "mail" short name for "internet 7".
105 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
108 *) Add an "init" command to the ENGINE config module and auto initialize
109 ENGINEs. Without any "init" command the ENGINE will be initialized
110 after all ctrl commands have been executed on it. If init=1 the
111 ENGINE is initailized at that point (ctrls before that point are run
112 on the uninitialized ENGINE and after on the initialized one). If
113 init=0 then the ENGINE will not be iniatialized at all.
116 *) Fix the 'app_verify_callback' interface so that the user-defined
117 argument is actually passed to the callback: In the
118 SSL_CTX_set_cert_verify_callback() prototype, the callback
119 declaration has been changed from
122 int (*cb)(X509_STORE_CTX *,void *);
123 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
124 i=s->ctx->app_verify_callback(&ctx)
125 has been changed into
126 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
128 To update applications using SSL_CTX_set_cert_verify_callback(),
129 a dummy argument can be added to their callback functions.
130 [D. K. Smetters <smetters@parc.xerox.com>]
132 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
133 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
135 *) Add and OPENSSL_LOAD_CONF define which will cause
136 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
137 This allows older applications to transparently support certain
138 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
139 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
140 load the config file and OPENSSL_add_all_algorithms_conf() which will
141 always load it have also been added.
144 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
145 Adjust NIDs and EVP layer.
146 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
148 *) Config modules support in openssl utility.
150 Most commands now load modules from the config file,
151 though in a few (such as version) this isn't done
152 because it couldn't be used for anything.
154 In the case of ca and req the config file used is
155 the same as the utility itself: that is the -config
156 command line option can be used to specify an
160 *) Move default behaviour from OPENSSL_config(). If appname is NULL
161 use "openssl_conf" if filename is NULL use default openssl config file.
164 *) Add an argument to OPENSSL_config() to allow the use of an alternative
165 config section name. Add a new flag to tolerate a missing config file
166 and move code to CONF_modules_load_file().
169 *) Support for crypto accelerator cards from Accelerated Encryption
170 Processing, www.aep.ie. (Use engine 'aep')
171 The support was copied from 0.9.6c [engine] and adapted/corrected
172 to work with the new engine framework.
173 [AEP Inc. and Richard Levitte]
175 *) Support for SureWare crypto accelerator cards from Baltimore
176 Technologies. (Use engine 'sureware')
177 The support was copied from 0.9.6c [engine] and adapted
178 to work with the new engine framework.
181 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
182 make the newer ENGINE framework commands for the CHIL engine work.
183 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
185 *) Make it possible to produce shared libraries on ReliantUNIX.
186 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
188 *) Add the configuration target debug-linux-ppro.
189 Make 'openssl rsa' use the general key loading routines
190 implemented in apps.c, and make those routines able to
191 handle the key format FORMAT_NETSCAPE and the variant
193 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
195 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
196 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
198 *) Add -keyform to rsautl, and document -engine.
199 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
201 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
202 BIO_R_NO_SUCH_FILE error code rather than the generic
203 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
208 ERR_peek_last_error_line
209 ERR_peek_last_error_line_data.
213 ERR_peek_error_line_data,
214 but report on the latest error recorded rather than the first one
215 still in the error queue.
216 [Ben Laurie, Bodo Moeller]
218 *) default_algorithms option in ENGINE config module. This allows things
220 default_algorithms = ALL
221 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
224 *) Prelminary ENGINE config module.
227 *) New experimental application configuration code.
230 *) Change the AES code to follow the same name structure as all other
231 symmetric ciphers, and behave the same way. Move everything to
232 the directory crypto/aes, thereby obsoleting crypto/rijndael.
233 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
235 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
236 [Ben Laurie and Theo de Raadt]
238 *) Add option to output public keys in req command.
239 [Massimiliano Pala madwolf@openca.org]
241 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
242 (up to about 10% better than before for P-192 and P-224).
245 *) New functions/macros
247 SSL_CTX_set_msg_callback(ctx, cb)
248 SSL_CTX_set_msg_callback_arg(ctx, arg)
249 SSL_set_msg_callback(ssl, cb)
250 SSL_set_msg_callback_arg(ssl, arg)
252 to request calling a callback function
254 void cb(int write_p, int version, int content_type,
255 const void *buf, size_t len, SSL *ssl, void *arg)
257 whenever a protocol message has been completely received
258 (write_p == 0) or sent (write_p == 1). Here 'version' is the
259 protocol version according to which the SSL library interprets
260 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
261 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
262 the content type as defined in the SSL 3.0/TLS 1.0 protocol
263 specification (change_cipher_spec(20), alert(21), handshake(22)).
264 'buf' and 'len' point to the actual message, 'ssl' to the
265 SSL object, and 'arg' is the application-defined value set by
266 SSL[_CTX]_set_msg_callback_arg().
268 'openssl s_client' and 'openssl s_server' have new '-msg' options
269 to enable a callback that displays all protocol messages.
272 *) Change the shared library support so shared libraries are built as
273 soon as the corresponding static library is finished, and thereby get
274 openssl and the test programs linked against the shared library.
275 This still only happens when the keyword "shard" has been given to
276 the configuration scripts.
278 NOTE: shared library support is still an experimental thing, and
279 backward binary compatibility is still not guaranteed.
280 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
282 *) Add support for Subject Information Access extension.
283 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
285 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
286 additional bytes when new memory had to be allocated, not just
287 when reusing an existing buffer.
290 *) New command line and configuration option 'utf8' for the req command.
291 This allows field values to be specified as UTF8 strings.
294 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
295 runs for the former and machine-readable output for the latter.
298 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
299 of the e-mail address in the DN (i.e., it will go into a certificate
300 extension only). The new configuration file option 'email_in_dn = no'
302 [Massimiliano Pala madwolf@openca.org]
304 *) Change all functions with names starting with des_ to be starting
305 with DES_ instead. Add wrappers that are compatible with libdes,
306 but are named _ossl_old_des_*. Finally, add macros that map the
307 des_* symbols to the corresponding _ossl_old_des_* if libdes
308 compatibility is desired. If OpenSSL 0.9.6c compatibility is
309 desired, the des_* symbols will be mapped to DES_*, with one
312 Since we provide two compatibility mappings, the user needs to
313 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
314 compatibility is desired. The default (i.e., when that macro
315 isn't defined) is OpenSSL 0.9.6c compatibility.
317 There are also macros that enable and disable the support of old
318 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
319 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
320 are defined, the default will apply: to support the old des routines.
322 In either case, one must include openssl/des.h to get the correct
323 definitions. Do not try to just include openssl/des_old.h, that
326 NOTE: This is a major break of an old API into a new one. Software
327 authors are encouraged to switch to the DES_ style functions. Some
328 time in the future, des_old.h and the libdes compatibility functions
329 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
330 default), and then completely removed.
333 *) Test for certificates which contain unsupported critical extensions.
334 If such a certificate is found during a verify operation it is
335 rejected by default: this behaviour can be overridden by either
336 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
337 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
338 X509_supported_extension() has also been added which returns 1 if a
339 particular extension is supported.
342 *) Modify the behaviour of EVP cipher functions in similar way to digests
343 to retain compatibility with existing code.
346 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
347 compatibility with existing code. In particular the 'ctx' parameter does
348 not have to be to be initialized before the call to EVP_DigestInit() and
349 it is tidied up after a call to EVP_DigestFinal(). New function
350 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
351 EVP_MD_CTX_copy() changed to not require the destination to be
352 initialized valid and new function EVP_MD_CTX_copy_ex() added which
353 requires the destination to be valid.
355 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
356 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
359 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
360 so that complete 'Handshake' protocol structures are kept in memory
361 instead of overwriting 'msg_type' and 'length' with 'body' data.
364 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
365 [Massimo Santin via Richard Levitte]
367 *) Major restructuring to the underlying ENGINE code. This includes
368 reduction of linker bloat, separation of pure "ENGINE" manipulation
369 (initialisation, etc) from functionality dealing with implementations
370 of specific crypto iterfaces. This change also introduces integrated
371 support for symmetric ciphers and digest implementations - so ENGINEs
372 can now accelerate these by providing EVP_CIPHER and EVP_MD
373 implementations of their own. This is detailed in crypto/engine/README
374 as it couldn't be adequately described here. However, there are a few
375 API changes worth noting - some RSA, DSA, DH, and RAND functions that
376 were changed in the original introduction of ENGINE code have now
377 reverted back - the hooking from this code to ENGINE is now a good
378 deal more passive and at run-time, operations deal directly with
379 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
380 dereferencing through an ENGINE pointer any more. Also, the ENGINE
381 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
382 they were not being used by the framework as there is no concept of a
383 BIGNUM_METHOD and they could not be generalised to the new
384 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
385 ENGINE_cpy() has been removed as it cannot be consistently defined in
389 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
392 *) Change mkdef.pl to sort symbols that get the same entry number,
393 and make sure the automatically generated functions ERR_load_*
394 become part of libeay.num as well.
397 *) New function SSL_renegotiate_pending(). This returns true once
398 renegotiation has been requested (either SSL_renegotiate() call
399 or HelloRequest/ClientHello receveived from the peer) and becomes
400 false once a handshake has been completed.
401 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
402 sends a HelloRequest, but does not ensure that a handshake takes
403 place. SSL_renegotiate_pending() is useful for checking if the
404 client has followed the request.)
407 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
408 By default, clients may request session resumption even during
409 renegotiation (if session ID contexts permit); with this option,
410 session resumption is possible only in the first handshake.
413 *) Add some demos for certificate and certificate request creation.
416 *) Make maximum certificate chain size accepted from the peer application
417 settable (SSL*_get/set_max_cert_list()), as proposed by
418 "Douglas E. Engert" <deengert@anl.gov>.
421 *) Add support for shared libraries for Unixware-7
422 (Boyd Lynn Gerber <gerberb@zenez.com>).
425 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
426 be done prior to destruction. Use this to unload error strings from
427 ENGINEs that load their own error strings. NB: This adds two new API
428 functions to "get" and "set" this destroy handler in an ENGINE.
431 *) Alter all existing ENGINE implementations (except "openssl" and
432 "openbsd") to dynamically instantiate their own error strings. This
433 makes them more flexible to be built both as statically-linked ENGINEs
434 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
435 Also, add stub code to each that makes building them as self-contained
436 shared-libraries easier (see README.ENGINE).
439 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
440 implementations into applications that are completely implemented in
441 self-contained shared-libraries. The "dynamic" ENGINE exposes control
442 commands that can be used to configure what shared-library to load and
443 to control aspects of the way it is handled. Also, made an update to
444 the README.ENGINE file that brings its information up-to-date and
445 provides some information and instructions on the "dynamic" ENGINE
446 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
449 *) Make it possible to unload ranges of ERR strings with a new
450 "ERR_unload_strings" function.
453 *) Add a copy() function to EVP_MD.
456 *) Make EVP_MD routines take a context pointer instead of just the
457 md_data void pointer.
460 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
461 that the digest can only process a single chunk of data
462 (typically because it is provided by a piece of
463 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
464 is only going to provide a single chunk of data, and hence the
465 framework needn't accumulate the data for oneshot drivers.
468 *) As with "ERR", make it possible to replace the underlying "ex_data"
469 functions. This change also alters the storage and management of global
470 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
471 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
472 index counters. The API functions that use this state have been changed
473 to take a "class_index" rather than pointers to the class's local STACK
474 and counter, and there is now an API function to dynamically create new
475 classes. This centralisation allows us to (a) plug a lot of the
476 thread-safety problems that existed, and (b) makes it possible to clean
477 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
478 such data would previously have always leaked in application code and
479 workarounds were in place to make the memory debugging turn a blind eye
480 to it. Application code that doesn't use this new function will still
481 leak as before, but their memory debugging output will announce it now
482 rather than letting it slide.
484 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
485 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
486 has a return value to indicate success or failure.
489 *) Make it possible to replace the underlying "ERR" functions such that the
490 global state (2 LHASH tables and 2 locks) is only used by the "default"
491 implementation. This change also adds two functions to "get" and "set"
492 the implementation prior to it being automatically set the first time
493 any other ERR function takes place. Ie. an application can call "get",
494 pass the return value to a module it has just loaded, and that module
495 can call its own "set" function using that value. This means the
496 module's "ERR" operations will use (and modify) the error state in the
497 application and not in its own statically linked copy of OpenSSL code.
500 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
501 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
502 the operation, and provides a more encapsulated way for external code
503 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
504 to use these functions rather than manually incrementing the counts.
506 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
509 *) Add EVP test program.
512 *) Add symmetric cipher support to ENGINE. Expect the API to change!
515 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
516 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
517 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
518 These allow a CRL to be built without having to access X509_CRL fields
519 directly. Modify 'ca' application to use new functions.
522 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
523 bug workarounds. Rollback attack detection is a security feature.
524 The problem will only arise on OpenSSL servers when TLSv1 is not
525 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
526 Software authors not wanting to support TLSv1 will have special reasons
527 for their choice and can explicitly enable this option.
528 [Bodo Moeller, Lutz Jaenicke]
530 *) Rationalise EVP so it can be extended: don't include a union of
531 cipher/digest structures, add init/cleanup functions. This also reduces
532 the number of header dependencies.
537 EVP_MD_CTX_init(&md); /* new function call */
538 EVP_DigestInit(&md, EVP_sha1());
539 EVP_DigestUpdate(&md, in, len);
540 EVP_DigestFinal(&md, out, NULL);
541 EVP_MD_CTX_cleanup(&md); /* new function call */
545 *) Make DES key schedule conform to the usual scheme, as well as
546 correcting its structure. This means that calls to DES functions
547 now have to pass a pointer to a des_key_schedule instead of a
548 plain des_key_schedule (which was actually always a pointer
553 des_set_key_checked(..., &ks);
554 des_ncbc_encrypt(..., &ks, ...);
556 (Note that a later change renames 'des_...' into 'DES_...'.)
559 *) Initial reduction of linker bloat: the use of some functions, such as
560 PEM causes large amounts of unused functions to be linked in due to
561 poor organisation. For example pem_all.c contains every PEM function
562 which has a knock on effect of linking in large amounts of (unused)
563 ASN1 code. Grouping together similar functions and splitting unrelated
564 functions prevents this.
567 *) Cleanup of EVP macros.
570 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
574 *) Add initial OCSP responder support to ocsp application. The
575 revocation information is handled using the text based index
576 use by the ca application. The responder can either handle
577 requests generated internally, supplied in files (for example
578 via a CGI script) or using an internal minimal server.
581 *) Add configuration choices to get zlib compression for TLS.
584 *) Changes to Kerberos SSL for RFC 2712 compliance:
585 1. Implemented real KerberosWrapper, instead of just using
586 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
587 2. Implemented optional authenticator field of KerberosWrapper.
589 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
590 and authenticator structs; see crypto/krb5/.
592 Generalized Kerberos calls to support multiple Kerberos libraries.
593 [Vern Staats <staatsvr@asc.hpc.mil>,
594 Jeffrey Altman <jaltman@columbia.edu>
597 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
598 already does with RSA. testdsa.h now has 'priv_key/pub_key'
599 values for each of the key sizes rather than having just
600 parameters (and 'speed' generating keys each time).
603 *) Speed up EVP routines.
606 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
607 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
608 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
609 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
611 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
612 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
613 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
616 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
618 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
621 *) Added the OS2-EMX target.
622 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
624 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
625 to support NCONF routines in extension code. New function CONF_set_nconf()
626 to allow functions which take an NCONF to also handle the old LHASH
627 structure: this means that the old CONF compatible routines can be
628 retained (in particular wrt extensions) without having to duplicate the
629 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
632 *) Enhance the general user interface with mechanisms for inner control
633 and with possibilities to have yes/no kind of prompts.
636 *) Change all calls to low level digest routines in the library and
637 applications to use EVP. Add missing calls to HMAC_cleanup() and
638 don't assume HMAC_CTX can be copied using memcpy().
639 [Verdon Walker <VWalker@novell.com>, Steve Henson]
641 *) Add the possibility to control engines through control names but with
642 arbitrary arguments instead of just a string.
643 Change the key loaders to take a UI_METHOD instead of a callback
644 function pointer. NOTE: this breaks binary compatibility with earlier
645 versions of OpenSSL [engine].
646 Adapt the nCipher code for these new conditions and add a card insertion
650 *) Enhance the general user interface with mechanisms to better support
651 dialog box interfaces, application-defined prompts, the possibility
652 to use defaults (for example default passwords from somewhere else)
653 and interrupts/cancellations.
656 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
657 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
660 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
661 tidy up some unnecessarily weird code in 'sk_new()').
662 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
664 *) Change the key loading routines for ENGINEs to use the same kind
665 callback (pem_password_cb) as all other routines that need this
669 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
670 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
671 than this minimum value is recommended.
674 *) New random seeder for OpenVMS, using the system process statistics
675 that are easily reachable.
678 *) Windows apparently can't transparently handle global
679 variables defined in DLLs. Initialisations such as:
681 const ASN1_ITEM *it = &ASN1_INTEGER_it;
683 wont compile. This is used by the any applications that need to
684 declare their own ASN1 modules. This was fixed by adding the option
685 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
686 needed for static libraries under Win32.
689 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
690 setting of purpose and trust fields. New X509_STORE trust and
691 purpose functions and tidy up setting in other SSL functions.
694 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
695 structure. These are inherited by X509_STORE_CTX when it is
696 initialised. This allows various defaults to be set in the
697 X509_STORE structure (such as flags for CRL checking and custom
698 purpose or trust settings) for functions which only use X509_STORE_CTX
699 internally such as S/MIME.
701 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
702 trust settings if they are not set in X509_STORE. This allows X509_STORE
703 purposes and trust (in S/MIME for example) to override any set by default.
705 Add command line options for CRL checking to smime, s_client and s_server
709 *) Initial CRL based revocation checking. If the CRL checking flag(s)
710 are set then the CRL is looked up in the X509_STORE structure and
711 its validity and signature checked, then if the certificate is found
712 in the CRL the verify fails with a revoked error.
714 Various new CRL related callbacks added to X509_STORE_CTX structure.
716 Command line options added to 'verify' application to support this.
718 This needs some additional work, such as being able to handle multiple
719 CRLs with different times, extension based lookup (rather than just
720 by subject name) and ultimately more complete V2 CRL extension
724 *) Add a general user interface API (crypto/ui/). This is designed
725 to replace things like des_read_password and friends (backward
726 compatibility functions using this new API are provided).
727 The purpose is to remove prompting functions from the DES code
728 section as well as provide for prompting through dialog boxes in
729 a window system and the like.
732 *) Add "ex_data" support to ENGINE so implementations can add state at a
733 per-structure level rather than having to store it globally.
736 *) Make it possible for ENGINE structures to be copied when retrieved by
737 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
738 This causes the "original" ENGINE structure to act like a template,
739 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
740 operational state can be localised to each ENGINE structure, despite the
741 fact they all share the same "methods". New ENGINE structures returned in
742 this case have no functional references and the return value is the single
743 structural reference. This matches the single structural reference returned
744 by ENGINE_by_id() normally, when it is incremented on the pre-existing
748 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
749 needs to match any other type at all we need to manually clear the
753 *) Changes to the "openssl engine" utility to include;
754 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
755 about an ENGINE's available control commands.
756 - executing control commands from command line arguments using the
757 '-pre' and '-post' switches. '-post' is only used if '-t' is
758 specified and the ENGINE is successfully initialised. The syntax for
759 the individual commands are colon-separated, for example;
760 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
763 *) New dynamic control command support for ENGINEs. ENGINEs can now
764 declare their own commands (numbers), names (strings), descriptions,
765 and input types for run-time discovery by calling applications. A
766 subset of these commands are implicitly classed as "executable"
767 depending on their input type, and only these can be invoked through
768 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
769 can be based on user input, config files, etc). The distinction is
770 that "executable" commands cannot return anything other than a boolean
771 result and can only support numeric or string input, whereas some
772 discoverable commands may only be for direct use through
773 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
774 pointers, or other custom uses. The "executable" commands are to
775 support parameterisations of ENGINE behaviour that can be
776 unambiguously defined by ENGINEs and used consistently across any
777 OpenSSL-based application. Commands have been added to all the
778 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
779 control over shared-library paths without source code alterations.
782 *) Changed all ENGINE implementations to dynamically allocate their
783 ENGINEs rather than declaring them statically. Apart from this being
784 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
785 this also allows the implementations to compile without using the
786 internal engine_int.h header.
789 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
790 'const' value. Any code that should be able to modify a RAND_METHOD
791 should already have non-const pointers to it (ie. they should only
792 modify their own ones).
795 *) Made a variety of little tweaks to the ENGINE code.
796 - "atalla" and "ubsec" string definitions were moved from header files
797 to C code. "nuron" string definitions were placed in variables
798 rather than hard-coded - allowing parameterisation of these values
799 later on via ctrl() commands.
800 - Removed unused "#if 0"'d code.
801 - Fixed engine list iteration code so it uses ENGINE_free() to release
802 structural references.
803 - Constified the RAND_METHOD element of ENGINE structures.
804 - Constified various get/set functions as appropriate and added
805 missing functions (including a catch-all ENGINE_cpy that duplicates
806 all ENGINE values onto a new ENGINE except reference counts/state).
807 - Removed NULL parameter checks in get/set functions. Setting a method
808 or function to NULL is a way of cancelling out a previously set
809 value. Passing a NULL ENGINE parameter is just plain stupid anyway
810 and doesn't justify the extra error symbols and code.
811 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
812 flags from engine_int.h to engine.h.
813 - Changed prototypes for ENGINE handler functions (init(), finish(),
814 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
817 *) Implement binary inversion algorithm for BN_mod_inverse in addition
818 to the algorithm using long division. The binary algorithm can be
819 used only if the modulus is odd. On 32-bit systems, it is faster
820 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
821 roughly 5-15% for 256-bit moduli), so we use it only for moduli
822 up to 450 bits. In 64-bit environments, the binary algorithm
823 appears to be advantageous for much longer moduli; here we use it
824 for moduli up to 2048 bits.
827 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
828 could not support the combine flag in choice fields.
831 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
832 extensions from a certificate request to the certificate.
835 *) Allow multiple 'certopt' and 'nameopt' options to be separated
836 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
837 file: this allows the display of the certificate about to be
838 signed to be customised, to allow certain fields to be included
839 or excluded and extension details. The old system didn't display
840 multicharacter strings properly, omitted fields not in the policy
841 and couldn't display additional details such as extensions.
844 *) Function EC_POINTs_mul for multiple scalar multiplication
845 of an arbitrary number of elliptic curve points
846 \sum scalars[i]*points[i],
847 optionally including the generator defined for the EC_GROUP:
848 scalar*generator + \sum scalars[i]*points[i].
850 EC_POINT_mul is a simple wrapper function for the typical case
851 that the point list has just one item (besides the optional
855 *) First EC_METHODs for curves over GF(p):
857 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
858 operations and provides various method functions that can also
859 operate with faster implementations of modular arithmetic.
861 EC_GFp_mont_method() reuses most functions that are part of
862 EC_GFp_simple_method, but uses Montgomery arithmetic.
864 [Bodo Moeller; point addition and point doubling
865 implementation directly derived from source code provided by
866 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
868 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
871 Curves are EC_GROUP objects (with an optional group generator)
872 based on EC_METHODs that are built into the library.
874 Points are EC_POINT objects based on EC_GROUP objects.
876 Most of the framework would be able to handle curves over arbitrary
877 finite fields, but as there are no obvious types for fields other
878 than GF(p), some functions are limited to that for now.
881 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
882 that the file contains a complete HTTP response.
885 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
886 change the def and num file printf format specifier from "%-40sXXX"
887 to "%-39s XXX". The latter will always guarantee a space after the
888 field while the former will cause them to run together if the field
889 is 40 of more characters long.
892 *) Constify the cipher and digest 'method' functions and structures
893 and modify related functions to take constant EVP_MD and EVP_CIPHER
897 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
898 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
901 *) Modify EVP_Digest*() routines so they now return values. Although the
902 internal software routines can never fail additional hardware versions
906 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
908 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
909 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
912 ERR_R_NESTED_ASN1_ERROR
914 ERR_R_MISSING_ASN1_EOS
915 were 4 .. 9, conflicting with
916 ERR_LIB_RSA (= ERR_R_RSA_LIB)
918 ERR_LIB_PEM (= ERR_R_PEM_LIB).
919 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
921 Add new error code 'ERR_R_INTERNAL_ERROR'.
924 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
928 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
929 sets the subject name for a new request or supersedes the
930 subject name in a given request. Formats that can be parsed are
931 'CN=Some Name, OU=myOU, C=IT'
933 'CN=Some Name/OU=myOU/C=IT'.
935 Add options '-batch' and '-verbose' to 'openssl req'.
936 [Massimiliano Pala <madwolf@hackmasters.net>]
938 *) Introduce the possibility to access global variables through
939 functions on platform were that's the best way to handle exporting
940 global variables in shared libraries. To enable this functionality,
941 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
942 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
943 is normally done by Configure or something similar).
945 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
946 in the source file (foo.c) like this:
948 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
949 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
951 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
952 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
954 OPENSSL_DECLARE_GLOBAL(int,foo);
955 #define foo OPENSSL_GLOBAL_REF(foo)
956 OPENSSL_DECLARE_GLOBAL(double,bar);
957 #define bar OPENSSL_GLOBAL_REF(bar)
959 The #defines are very important, and therefore so is including the
960 header file everywhere where the defined globals are used.
962 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
963 of ASN.1 items, but that structure is a bit different.
965 The largest change is in util/mkdef.pl which has been enhanced with
966 better and easier to understand logic to choose which symbols should
967 go into the Windows .def files as well as a number of fixes and code
968 cleanup (among others, algorithm keywords are now sorted
969 lexicographically to avoid constant rewrites).
972 *) In BN_div() keep a copy of the sign of 'num' before writing the
973 result to 'rm' because if rm==num the value will be overwritten
974 and produce the wrong result if 'num' is negative: this caused
975 problems with BN_mod() and BN_nnmod().
978 *) Function OCSP_request_verify(). This checks the signature on an
979 OCSP request and verifies the signer certificate. The signer
980 certificate is just checked for a generic purpose and OCSP request
984 *) Add OCSP_check_validity() function to check the validity of OCSP
985 responses. OCSP responses are prepared in real time and may only
986 be a few seconds old. Simply checking that the current time lies
987 between thisUpdate and nextUpdate max reject otherwise valid responses
988 caused by either OCSP responder or client clock inaccuracy. Instead
989 we allow thisUpdate and nextUpdate to fall within a certain period of
990 the current time. The age of the response can also optionally be
991 checked. Two new options -validity_period and -status_age added to
995 *) If signature or public key algorithm is unrecognized print out its
996 OID rather that just UNKNOWN.
999 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1000 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1001 ID to be generated from the issuer certificate alone which can then be
1002 passed to OCSP_id_issuer_cmp().
1005 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1006 ASN1 modules to export functions returning ASN1_ITEM pointers
1007 instead of the ASN1_ITEM structures themselves. This adds several
1008 new macros which allow the underlying ASN1 function/structure to
1009 be accessed transparently. As a result code should not use ASN1_ITEM
1010 references directly (such as &X509_it) but instead use the relevant
1011 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1012 use of the new ASN1 code on platforms where exporting structures
1013 is problematical (for example in shared libraries) but exporting
1014 functions returning pointers to structures is not.
1017 *) Add support for overriding the generation of SSL/TLS session IDs.
1018 These callbacks can be registered either in an SSL_CTX or per SSL.
1019 The purpose of this is to allow applications to control, if they wish,
1020 the arbitrary values chosen for use as session IDs, particularly as it
1021 can be useful for session caching in multiple-server environments. A
1022 command-line switch for testing this (and any client code that wishes
1023 to use such a feature) has been added to "s_server".
1024 [Geoff Thorpe, Lutz Jaenicke]
1026 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
1027 of the form '#if defined(...) || defined(...) || ...' and
1028 '#if !defined(...) && !defined(...) && ...'. This also avoids
1029 the growing number of special cases it was previously handling.
1032 *) Make all configuration macros available for application by making
1033 sure they are available in opensslconf.h, by giving them names starting
1034 with "OPENSSL_" to avoid conflicts with other packages and by making
1035 sure e_os2.h will cover all platform-specific cases together with
1037 Additionally, it is now possible to define configuration/platform-
1038 specific names (called "system identities"). In the C code, these
1039 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1040 macro with the name beginning with "OPENSSL_SYS_", which is determined
1041 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1045 *) New option -set_serial to 'req' and 'x509' this allows the serial
1046 number to use to be specified on the command line. Previously self
1047 signed certificates were hard coded with serial number 0 and the
1048 CA options of 'x509' had to use a serial number in a file which was
1052 *) New options to 'ca' utility to support V2 CRL entry extensions.
1053 Currently CRL reason, invalidity date and hold instruction are
1054 supported. Add new CRL extensions to V3 code and some new objects.
1057 *) New function EVP_CIPHER_CTX_set_padding() this is used to
1058 disable standard block padding (aka PKCS#5 padding) in the EVP
1059 API, which was previously mandatory. This means that the data is
1060 not padded in any way and so the total length much be a multiple
1061 of the block size, otherwise an error occurs.
1064 *) Initial (incomplete) OCSP SSL support.
1067 *) New function OCSP_parse_url(). This splits up a URL into its host,
1068 port and path components: primarily to parse OCSP URLs. New -url
1069 option to ocsp utility.
1072 *) New nonce behavior. The return value of OCSP_check_nonce() now
1073 reflects the various checks performed. Applications can decide
1074 whether to tolerate certain situations such as an absent nonce
1075 in a response when one was present in a request: the ocsp application
1076 just prints out a warning. New function OCSP_add1_basic_nonce()
1077 this is to allow responders to include a nonce in a response even if
1078 the request is nonce-less.
1081 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1082 skipped when using openssl x509 multiple times on a single input file,
1083 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1086 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1087 set string type: to handle setting ASN1_TIME structures. Fix ca
1088 utility to correctly initialize revocation date of CRLs.
1091 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1092 the clients preferred ciphersuites and rather use its own preferences.
1093 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1094 Internet Explorer by ensuring unchanged hash method during stepup.
1095 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1098 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1099 to aes and add a new 'exist' option to print out symbols that don't
1103 *) Additional options to ocsp utility to allow flags to be set and
1104 additional certificates supplied.
1107 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
1108 OCSP client a number of certificate to only verify the response
1112 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1113 handle the new API. Currently only ECB, CBC modes supported. Add new
1116 Add TLS AES ciphersuites as described in the "AES Ciphersuites
1117 for TLS" draft-ietf-tls-ciphersuite-06.txt. As these are not yet
1118 official, they are not enabled by default and are not even part
1119 of the "ALL" ciphersuite alias; for now, they must be explicitly
1120 requested by specifying the new "AESdraft" ciphersuite alias. If
1121 you want the default ciphersuite list plus the new ciphersuites,
1122 use "DEFAULT:AESdraft:@STRENGTH".
1123 [Ben Laurie, Steve Henson, Bodo Moeller]
1125 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
1126 request to response.
1129 *) Functions for OCSP responders. OCSP_request_onereq_count(),
1130 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1131 extract information from a certificate request. OCSP_response_create()
1132 creates a response and optionally adds a basic response structure.
1133 OCSP_basic_add1_status() adds a complete single response to a basic
1134 response and returns the OCSP_SINGLERESP structure just added (to allow
1135 extensions to be included for example). OCSP_basic_add1_cert() adds a
1136 certificate to a basic response and OCSP_basic_sign() signs a basic
1137 response with various flags. New helper functions ASN1_TIME_check()
1138 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1139 (converts ASN1_TIME to GeneralizedTime).
1142 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1143 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1144 structure from a certificate. X509_pubkey_digest() digests the public_key
1145 contents: this is used in various key identifiers.
1148 *) Make sk_sort() tolerate a NULL argument.
1149 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1151 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1152 passed by the function are trusted implicitly. If any of them signed the
1153 response then it is assumed to be valid and is not verified.
1156 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1157 to data. This was previously part of the PKCS7 ASN1 code. This
1158 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1159 [Steve Henson, reported by Kenneth R. Robinette
1160 <support@securenetterm.com>]
1162 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1163 routines: without these tracing memory leaks is very painful.
1164 Fix leaks in PKCS12 and PKCS7 routines.
1167 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1168 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1169 effectively meant GeneralizedTime would never be used. Now it
1170 is initialised to -1 but X509_time_adj() now has to check the value
1171 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1172 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1173 [Steve Henson, reported by Kenneth R. Robinette
1174 <support@securenetterm.com>]
1176 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1177 result in a zero length in the ASN1_INTEGER structure which was
1178 not consistent with the structure when d2i_ASN1_INTEGER() was used
1179 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1180 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1181 where it did not print out a minus for negative ASN1_INTEGER.
1184 *) Add summary printout to ocsp utility. The various functions which
1185 convert status values to strings have been renamed to:
1186 OCSP_response_status_str(), OCSP_cert_status_str() and
1187 OCSP_crl_reason_str() and are no longer static. New options
1188 to verify nonce values and to disable verification. OCSP response
1189 printout format cleaned up.
1192 *) Add additional OCSP certificate checks. These are those specified
1193 in RFC2560. This consists of two separate checks: the CA of the
1194 certificate being checked must either be the OCSP signer certificate
1195 or the issuer of the OCSP signer certificate. In the latter case the
1196 OCSP signer certificate must contain the OCSP signing extended key
1197 usage. This check is performed by attempting to match the OCSP
1198 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1199 in the OCSP_CERTID structures of the response.
1202 *) Initial OCSP certificate verification added to OCSP_basic_verify()
1203 and related routines. This uses the standard OpenSSL certificate
1204 verify routines to perform initial checks (just CA validity) and
1205 to obtain the certificate chain. Then additional checks will be
1206 performed on the chain. Currently the root CA is checked to see
1207 if it is explicitly trusted for OCSP signing. This is used to set
1208 a root CA as a global signing root: that is any certificate that
1209 chains to that CA is an acceptable OCSP signing certificate.
1212 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1213 extensions from a separate configuration file.
1214 As when reading extensions from the main configuration file,
1215 the '-extensions ...' option may be used for specifying the
1217 [Massimiliano Pala <madwolf@comune.modena.it>]
1219 *) New OCSP utility. Allows OCSP requests to be generated or
1220 read. The request can be sent to a responder and the output
1221 parsed, outputed or printed in text form. Not complete yet:
1222 still needs to check the OCSP response validity.
1225 *) New subcommands for 'openssl ca':
1226 'openssl ca -status <serial>' prints the status of the cert with
1227 the given serial number (according to the index file).
1228 'openssl ca -updatedb' updates the expiry status of certificates
1230 [Massimiliano Pala <madwolf@comune.modena.it>]
1232 *) New '-newreq-nodes' command option to CA.pl. This is like
1233 '-newreq', but calls 'openssl req' with the '-nodes' option
1234 so that the resulting key is not encrypted.
1235 [Damien Miller <djm@mindrot.org>]
1237 *) New configuration for the GNU Hurd.
1238 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1240 *) Initial code to implement OCSP basic response verify. This
1241 is currently incomplete. Currently just finds the signer's
1242 certificate and verifies the signature on the response.
1245 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1246 value of OPENSSLDIR. This is available via the new '-d' option
1247 to 'openssl version', and is also included in 'openssl version -a'.
1250 *) Allowing defining memory allocation callbacks that will be given
1251 file name and line number information in additional arguments
1252 (a const char* and an int). The basic functionality remains, as
1253 well as the original possibility to just replace malloc(),
1254 realloc() and free() by functions that do not know about these
1255 additional arguments. To register and find out the current
1256 settings for extended allocation functions, the following
1257 functions are provided:
1259 CRYPTO_set_mem_ex_functions
1260 CRYPTO_set_locked_mem_ex_functions
1261 CRYPTO_get_mem_ex_functions
1262 CRYPTO_get_locked_mem_ex_functions
1264 These work the same way as CRYPTO_set_mem_functions and friends.
1265 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1266 extended allocation function is enabled.
1267 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1268 a conventional allocation function is enabled.
1269 [Richard Levitte, Bodo Moeller]
1271 *) Finish off removing the remaining LHASH function pointer casts.
1272 There should no longer be any prototype-casting required when using
1273 the LHASH abstraction, and any casts that remain are "bugs". See
1274 the callback types and macros at the head of lhash.h for details
1275 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1278 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1279 If /dev/[u]random devices are not available or do not return enough
1280 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1282 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1283 /etc/entropy will be queried once each in this sequence, quering stops
1284 when enough entropy was collected without querying more sockets.
1287 *) Change the Unix RAND_poll() variant to be able to poll several
1288 random devices, as specified by DEVRANDOM, until a sufficient amount
1289 of data has been collected. We spend at most 10 ms on each file
1290 (select timeout) and read in non-blocking mode. DEVRANDOM now
1291 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1292 (previously it was just the string "/dev/urandom"), so on typical
1293 platforms the 10 ms delay will never occur.
1294 Also separate out the Unix variant to its own file, rand_unix.c.
1295 For VMS, there's a currently-empty rand_vms.c.
1298 *) Move OCSP client related routines to ocsp_cl.c. These
1299 provide utility functions which an application needing
1300 to issue a request to an OCSP responder and analyse the
1301 response will typically need: as opposed to those which an
1302 OCSP responder itself would need which will be added later.
1304 OCSP_request_sign() signs an OCSP request with an API similar
1305 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1306 response. OCSP_response_get1_basic() extracts basic response
1307 from response. OCSP_resp_find_status(): finds and extracts status
1308 information from an OCSP_CERTID structure (which will be created
1309 when the request structure is built). These are built from lower
1310 level functions which work on OCSP_SINGLERESP structures but
1311 wont normally be used unless the application wishes to examine
1312 extensions in the OCSP response for example.
1314 Replace nonce routines with a pair of functions.
1315 OCSP_request_add1_nonce() adds a nonce value and optionally
1316 generates a random value. OCSP_check_nonce() checks the
1317 validity of the nonce in an OCSP response.
1320 *) Change function OCSP_request_add() to OCSP_request_add0_id().
1321 This doesn't copy the supplied OCSP_CERTID and avoids the
1322 need to free up the newly created id. Change return type
1323 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1324 This can then be used to add extensions to the request.
1325 Deleted OCSP_request_new(), since most of its functionality
1326 is now in OCSP_REQUEST_new() (and the case insensitive name
1327 clash) apart from the ability to set the request name which
1328 will be added elsewhere.
1331 *) Update OCSP API. Remove obsolete extensions argument from
1332 various functions. Extensions are now handled using the new
1333 OCSP extension code. New simple OCSP HTTP function which
1334 can be used to send requests and parse the response.
1337 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1338 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1339 uses the special reorder version of SET OF to sort the attributes
1340 and reorder them to match the encoded order. This resolves a long
1341 standing problem: a verify on a PKCS7 structure just after signing
1342 it used to fail because the attribute order did not match the
1343 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1344 it uses the received order. This is necessary to tolerate some broken
1345 software that does not order SET OF. This is handled by encoding
1346 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1347 to produce the required SET OF.
1350 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1351 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1352 files to get correct declarations of the ASN.1 item variables.
1355 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1356 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1357 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1358 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1359 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1360 ASN1_ITEM and no wrapper functions.
1363 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1364 replace the old function pointer based I/O routines. Change most of
1365 the *_d2i_bio() and *_d2i_fp() functions to use these.
1368 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1369 lines, recognice more "algorithms" that can be deselected, and make
1370 it complain about algorithm deselection that isn't recognised.
1373 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
1374 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1375 to use new functions. Add NO_ASN1_OLD which can be set to remove
1376 some old style ASN1 functions: this can be used to determine if old
1377 code will still work when these eventually go away.
1380 *) New extension functions for OCSP structures, these follow the
1381 same conventions as certificates and CRLs.
1384 *) New function X509V3_add1_i2d(). This automatically encodes and
1385 adds an extension. Its behaviour can be customised with various
1386 flags to append, replace or delete. Various wrappers added for
1387 certifcates and CRLs.
1390 *) Fix to avoid calling the underlying ASN1 print routine when
1391 an extension cannot be parsed. Correct a typo in the
1392 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1395 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1396 entries for variables.
1399 *) Add functionality to apps/openssl.c for detecting locking
1400 problems: As the program is single-threaded, all we have
1401 to do is register a locking callback using an array for
1402 storing which locks are currently held by the program.
1405 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
1406 SSL_get_ex_data_X509_STORE_idx(), which is used in
1407 ssl_verify_cert_chain() and thus can be called at any time
1408 during TLS/SSL handshakes so that thread-safety is essential.
1409 Unfortunately, the ex_data design is not at all suited
1410 for multi-threaded use, so it probably should be abolished.
1413 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
1414 [Broadcom, tweaked and integrated by Geoff Thorpe]
1416 *) Move common extension printing code to new function
1417 X509V3_print_extensions(). Reorganise OCSP print routines and
1418 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1421 *) New function X509_signature_print() to remove duplication in some
1425 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
1426 set (this was treated exactly the same as SET OF previously). This
1427 is used to reorder the STACK representing the structure to match the
1428 encoding. This will be used to get round a problem where a PKCS7
1429 structure which was signed could not be verified because the STACK
1430 order did not reflect the encoded order.
1433 *) Reimplement the OCSP ASN1 module using the new code.
1436 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1437 for its ASN1 operations. The old style function pointers still exist
1438 for now but they will eventually go away.
1441 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
1442 completely replaces the old ASN1 functionality with a table driven
1443 encoder and decoder which interprets an ASN1_ITEM structure describing
1444 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1445 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1446 has also been converted to the new form.
1449 *) Change BN_mod_exp_recp so that negative moduli are tolerated
1450 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1451 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1452 for negative moduli.
1455 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
1456 of not touching the result's sign bit.
1459 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1463 *) Changed the LHASH code to use prototypes for callbacks, and created
1464 macros to declare and implement thin (optionally static) functions
1465 that provide type-safety and avoid function pointer casting for the
1466 type-specific callbacks.
1469 *) Added Kerberos Cipher Suites to be used with TLS, as written in
1471 [Veers Staats <staatsvr@asc.hpc.mil>,
1472 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1474 *) Reformat the FAQ so the different questions and answers can be divided
1475 in sections depending on the subject.
1478 *) Have the zlib compression code load ZLIB.DLL dynamically under
1482 *) New function BN_mod_sqrt for computing square roots modulo a prime
1483 (using the probabilistic Tonelli-Shanks algorithm unless
1484 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1485 be handled deterministically).
1486 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1488 *) Make BN_mod_inverse faster by explicitly handling small quotients
1489 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1490 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1493 *) New function BN_kronecker.
1496 *) Fix BN_gcd so that it works on negative inputs; the result is
1497 positive unless both parameters are zero.
1498 Previously something reasonably close to an infinite loop was
1499 possible because numbers could be growing instead of shrinking
1500 in the implementation of Euclid's algorithm.
1503 *) Fix BN_is_word() and BN_is_one() macros to take into account the
1504 sign of the number in question.
1506 Fix BN_is_word(a,w) to work correctly for w == 0.
1508 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1509 because its test if the absolute value of 'a' equals 'w'.
1510 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1511 it exists mostly for use in the implementations of BN_is_zero(),
1512 BN_is_one(), and BN_is_word().
1515 *) New function BN_swap.
1518 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1519 the exponentiation functions are more likely to produce reasonable
1520 results on negative inputs.
1523 *) Change BN_mod_mul so that the result is always non-negative.
1524 Previously, it could be negative if one of the factors was negative;
1525 I don't think anyone really wanted that behaviour.
1528 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1529 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1530 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1531 and add new functions:
1540 BN_mod_lshift1_quick
1544 These functions always generate non-negative results.
1546 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1547 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1549 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
1550 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
1551 be reduced modulo m.
1552 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1554 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1555 was actually never needed) and in BN_mul(). The removal in BN_mul()
1556 required a small change in bn_mul_part_recursive() and the addition
1557 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1558 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1559 bn_sub_words() and bn_add_words() except they take arrays with
1563 *) In 'openssl passwd', verify passwords read from the terminal
1564 unless the '-salt' option is used (which usually means that
1565 verification would just waste user's time since the resulting
1566 hash is going to be compared with some given password hash)
1567 or the new '-noverify' option is used.
1569 This is an incompatible change, but it does not affect
1570 non-interactive use of 'openssl passwd' (passwords on the command
1571 line, '-stdin' option, '-in ...' option) and thus should not
1575 *) Remove all references to RSAref, since there's no more need for it.
1578 *) Make DSO load along a path given through an environment variable
1579 (SHLIB_PATH) with shl_load().
1582 *) Constify the ENGINE code as a result of BIGNUM constification.
1583 Also constify the RSA code and most things related to it. In a
1584 few places, most notable in the depth of the ASN.1 code, ugly
1585 casts back to non-const were required (to be solved at a later
1589 *) Make it so the openssl application has all engines loaded by default.
1592 *) Constify the BIGNUM routines a little more.
1595 *) Add the following functions:
1597 ENGINE_load_cswift()
1599 ENGINE_load_atalla()
1601 ENGINE_load_builtin_engines()
1603 That way, an application can itself choose if external engines that
1604 are built-in in OpenSSL shall ever be used or not. The benefit is
1605 that applications won't have to be linked with libdl or other dso
1606 libraries unless it's really needed.
1608 Changed 'openssl engine' to load all engines on demand.
1609 Changed the engine header files to avoid the duplication of some
1610 declarations (they differed!).
1613 *) 'openssl engine' can now list capabilities.
1616 *) Better error reporting in 'openssl engine'.
1619 *) Never call load_dh_param(NULL) in s_server.
1622 *) Add engine application. It can currently list engines by name and
1623 identity, and test if they are actually available.
1626 *) Improve RPM specification file by forcing symbolic linking and making
1627 sure the installed documentation is also owned by root.root.
1628 [Damien Miller <djm@mindrot.org>]
1630 *) Give the OpenSSL applications more possibilities to make use of
1631 keys (public as well as private) handled by engines.
1634 *) Add OCSP code that comes from CertCo.
1637 *) Add VMS support for the Rijndael code.
1640 *) Added untested support for Nuron crypto accelerator.
1643 *) Add support for external cryptographic devices. This code was
1644 previously distributed separately as the "engine" branch.
1645 [Geoff Thorpe, Richard Levitte]
1647 *) Rework the filename-translation in the DSO code. It is now possible to
1648 have far greater control over how a "name" is turned into a filename
1649 depending on the operating environment and any oddities about the
1650 different shared library filenames on each system.
1653 *) Support threads on FreeBSD-elf in Configure.
1656 *) Fix for SHA1 assembly problem with MASM: it produces
1657 warnings about corrupt line number information when assembling
1658 with debugging information. This is caused by the overlapping
1660 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
1663 NCONF_get_number() has no error checking at all. As a replacement,
1664 NCONF_get_number_e() is defined (_e for "error checking") and is
1665 promoted strongly. The old NCONF_get_number is kept around for
1666 binary backward compatibility.
1667 Make it possible for methods to load from something other than a BIO,
1668 by providing a function pointer that is given a name instead of a BIO.
1669 For example, this could be used to load configuration data from an
1673 *) Fix for non blocking accept BIOs. Added new I/O special reason
1674 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
1675 with non blocking I/O was not possible because no retry code was
1676 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
1680 *) Added the beginnings of Rijndael support.
1683 *) Fix for bug in DirectoryString mask setting. Add support for
1684 X509_NAME_print_ex() in 'req' and X509_print_ex() function
1685 to allow certificate printing to more controllable, additional
1686 'certopt' option to 'x509' to allow new printing options to be
1690 *) Clean old EAY MD5 hack from e_os.h.
1693 Changes between 0.9.6d and 0.9.6e [XX xxx XXXX]
1695 *) Fix EVP_dsa_sha macro.
1698 Changes between 0.9.6c and 0.9.6d [9 May 2002]
1700 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
1701 encoded as NULL) with id-dsa-with-sha1.
1702 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
1704 *) Check various X509_...() return values in apps/req.c.
1705 [Nils Larsch <nla@trustcenter.de>]
1707 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
1708 an end-of-file condition would erronously be flagged, when the CRLF
1709 was just at the end of a processed block. The bug was discovered when
1710 processing data through a buffering memory BIO handing the data to a
1711 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
1712 <ptsekov@syntrex.com> and Nedelcho Stanev.
1715 *) Implement a countermeasure against a vulnerability recently found
1716 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
1717 before application data chunks to avoid the use of known IVs
1718 with data potentially chosen by the attacker.
1721 *) Fix length checks in ssl3_get_client_hello().
1724 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
1725 to prevent ssl3_read_internal() from incorrectly assuming that
1726 ssl3_read_bytes() found application data while handshake
1727 processing was enabled when in fact s->s3->in_read_app_data was
1728 merely automatically cleared during the initial handshake.
1729 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
1731 *) Fix object definitions for Private and Enterprise: they were not
1732 recognized in their shortname (=lowercase) representation. Extend
1733 obj_dat.pl to issue an error when using undefined keywords instead
1734 of silently ignoring the problem (Svenning Sorensen
1735 <sss@sss.dnsalias.net>).
1738 *) Fix DH_generate_parameters() so that it works for 'non-standard'
1739 generators, i.e. generators other than 2 and 5. (Previously, the
1740 code did not properly initialise the 'add' and 'rem' values to
1741 BN_generate_prime().)
1743 In the new general case, we do not insist that 'generator' is
1744 actually a primitive root: This requirement is rather pointless;
1745 a generator of the order-q subgroup is just as good, if not
1749 *) Map new X509 verification errors to alerts. Discovered and submitted by
1750 Tom Wu <tom@arcot.com>.
1753 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
1754 returning non-zero before the data has been completely received
1755 when using non-blocking I/O.
1756 [Bodo Moeller; problem pointed out by John Hughes]
1758 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
1759 [Ben Laurie, Lutz Jaenicke]
1761 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
1762 Yoram Zahavi <YoramZ@gilian.com>).
1765 *) Add information about CygWin 1.3 and on, and preserve proper
1766 configuration for the versions before that.
1767 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1769 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
1770 check whether we deal with a copy of a session and do not delete from
1771 the cache in this case. Problem reported by "Izhar Shoshani Levi"
1772 <izhar@checkpoint.com>.
1775 *) Do not store session data into the internal session cache, if it
1776 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
1777 flag is set). Proposed by Aslam <aslam@funk.com>.
1780 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
1784 *) Add the configuration target linux-s390x.
1785 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
1787 *) [In 0.9.6d-engine release:]
1788 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1789 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1791 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
1792 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
1793 variable as an indication that a ClientHello message has been
1794 received. As the flag value will be lost between multiple
1795 invocations of ssl3_accept when using non-blocking I/O, the
1796 function may not be aware that a handshake has actually taken
1797 place, thus preventing a new session from being added to the
1800 To avoid this problem, we now set s->new_session to 2 instead of
1801 using a local variable.
1802 [Lutz Jaenicke, Bodo Moeller]
1804 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
1805 if the SSL_R_LENGTH_MISMATCH error is detected.
1806 [Geoff Thorpe, Bodo Moeller]
1808 *) New 'shared_ldflag' column in Configure platform table.
1811 *) Fix EVP_CIPHER_mode macro.
1812 ["Dan S. Camper" <dan@bti.net>]
1814 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
1815 type, we must throw them away by setting rr->length to 0.
1816 [D P Chang <dpc@qualys.com>]
1818 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
1820 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
1821 <Dominikus.Scherkl@biodata.com>. (The previous implementation
1822 worked incorrectly for those cases where range = 10..._2 and
1823 3*range is two bits longer than range.)
1826 *) Only add signing time to PKCS7 structures if it is not already
1830 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
1831 OBJ_ld_ce should be OBJ_id_ce.
1832 Also some ip-pda OIDs in crypto/objects/objects.txt were
1833 incorrect (cf. RFC 3039).
1834 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
1836 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
1837 returns early because it has nothing to do.
1838 [Andy Schneider <andy.schneider@bjss.co.uk>]
1840 *) [In 0.9.6c-engine release:]
1841 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
1842 [Andy Schneider <andy.schneider@bjss.co.uk>]
1844 *) [In 0.9.6c-engine release:]
1845 Add support for Cryptographic Appliance's keyserver technology.
1846 (Use engine 'keyclient')
1847 [Cryptographic Appliances and Geoff Thorpe]
1849 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
1850 is called via tools/c89.sh because arguments have to be
1851 rearranged (all '-L' options must appear before the first object
1853 [Richard Shapiro <rshapiro@abinitio.com>]
1855 *) [In 0.9.6c-engine release:]
1856 Add support for Broadcom crypto accelerator cards, backported
1858 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
1860 *) [In 0.9.6c-engine release:]
1861 Add support for SureWare crypto accelerator cards from
1862 Baltimore Technologies. (Use engine 'sureware')
1863 [Baltimore Technologies and Mark Cox]
1865 *) [In 0.9.6c-engine release:]
1866 Add support for crypto accelerator cards from Accelerated
1867 Encryption Processing, www.aep.ie. (Use engine 'aep')
1868 [AEP Inc. and Mark Cox]
1870 *) Add a configuration entry for gcc on UnixWare.
1871 [Gary Benson <gbenson@redhat.com>]
1873 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
1874 messages are stored in a single piece (fixed-length part and
1875 variable-length part combined) and fix various bugs found on the way.
1878 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
1879 instead. BIO_gethostbyname() does not know what timeouts are
1880 appropriate, so entries would stay in cache even when they have
1882 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
1884 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
1885 faced with a pathologically small ClientHello fragment that does
1886 not contain client_version: Instead of aborting with an error,
1887 simply choose the highest available protocol version (i.e.,
1888 TLS 1.0 unless it is disabled). In practice, ClientHello
1889 messages are never sent like this, but this change gives us
1890 strictly correct behaviour at least for TLS.
1893 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
1894 never resets s->method to s->ctx->method when called from within
1895 one of the SSL handshake functions.
1896 [Bodo Moeller; problem pointed out by Niko Baric]
1898 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
1899 (sent using the client's version number) if client_version is
1900 smaller than the protocol version in use. Also change
1901 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
1902 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
1903 the client will at least see that alert.
1906 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
1910 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
1911 client receives HelloRequest while in a handshake.
1912 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
1914 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
1915 should end in 'break', not 'goto end' which circuments various
1916 cleanups done in state SSL_ST_OK. But session related stuff
1917 must be disabled for SSL_ST_OK in the case that we just sent a
1920 Also avoid some overhead by not calling ssl_init_wbio_buffer()
1921 before just sending a HelloRequest.
1922 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
1924 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
1925 reveal whether illegal block cipher padding was found or a MAC
1926 verification error occured. (Neither SSLerr() codes nor alerts
1927 are directly visible to potential attackers, but the information
1928 may leak via logfiles.)
1930 Similar changes are not required for the SSL 2.0 implementation
1931 because the number of padding bytes is sent in clear for SSL 2.0,
1932 and the extra bytes are just ignored. However ssl/s2_pkt.c
1933 failed to verify that the purported number of padding bytes is in
1937 *) Add OpenUNIX-8 support including shared libraries
1938 (Boyd Lynn Gerber <gerberb@zenez.com>).
1941 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
1942 'wristwatch attack' using huge encoding parameters (cf.
1943 James H. Manger's CRYPTO 2001 paper). Note that the
1944 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
1945 encoding parameters and hence was not vulnerable.
1948 *) BN_sqr() bug fix.
1949 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
1951 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
1952 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
1953 followed by modular reduction.
1954 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
1956 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
1957 equivalent based on BN_pseudo_rand() instead of BN_rand().
1960 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
1961 This function was broken, as the check for a new client hello message
1962 to handle SGC did not allow these large messages.
1963 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
1966 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
1969 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
1970 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
1973 *) Rework the configuration and shared library support for Tru64 Unix.
1974 The configuration part makes use of modern compiler features and
1975 still retains old compiler behavior for those that run older versions
1976 of the OS. The shared library support part includes a variant that
1977 uses the RPATH feature, and is available through the special
1978 configuration target "alpha-cc-rpath", which will never be selected
1980 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
1982 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
1983 with the same message size as in ssl3_get_certificate_request().
1984 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
1985 messages might inadvertently be reject as too long.
1986 [Petr Lampa <lampa@fee.vutbr.cz>]
1988 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
1991 *) Modified SSL library such that the verify_callback that has been set
1992 specificly for an SSL object with SSL_set_verify() is actually being
1993 used. Before the change, a verify_callback set with this function was
1994 ignored and the verify_callback() set in the SSL_CTX at the time of
1995 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
1996 to allow the necessary settings.
1999 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
2000 explicitly to NULL, as at least on Solaris 8 this seems not always to be
2001 done automatically (in contradiction to the requirements of the C
2002 standard). This made problems when used from OpenSSH.
2005 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
2006 dh->length and always used
2008 BN_rand_range(priv_key, dh->p).
2010 BN_rand_range() is not necessary for Diffie-Hellman, and this
2011 specific range makes Diffie-Hellman unnecessarily inefficient if
2012 dh->length (recommended exponent length) is much smaller than the
2013 length of dh->p. We could use BN_rand_range() if the order of
2014 the subgroup was stored in the DH structure, but we only have
2019 BN_rand(priv_key, l, ...)
2021 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
2027 RSA_eay_public_encrypt
2028 RSA_eay_private_decrypt
2029 RSA_eay_private_encrypt (signing)
2030 RSA_eay_public_decrypt (signature verification)
2032 (default implementations for RSA_public_encrypt,
2033 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
2034 always reject numbers >= n.
2037 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
2038 to synchronize access to 'locking_thread'. This is necessary on
2039 systems where access to 'locking_thread' (an 'unsigned long'
2040 variable) is not atomic.
2043 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
2044 *before* setting the 'crypto_lock_rand' flag. The previous code had
2045 a race condition if 0 is a valid thread ID.
2046 [Travis Vitek <vitek@roguewave.com>]
2048 *) Add support for shared libraries under Irix.
2049 [Albert Chin-A-Young <china@thewrittenword.com>]
2051 *) Add configuration option to build on Linux on both big-endian and
2053 [Ralf Baechle <ralf@uni-koblenz.de>]
2055 *) Add the possibility to create shared libraries on HP-UX.
2058 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
2060 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
2061 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
2062 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
2063 PRNG state recovery was possible based on the output of
2064 one PRNG request appropriately sized to gain knowledge on
2065 'md' followed by enough consecutive 1-byte PRNG requests
2066 to traverse all of 'state'.
2068 1. When updating 'md_local' (the current thread's copy of 'md')
2069 during PRNG output generation, hash all of the previous
2070 'md_local' value, not just the half used for PRNG output.
2072 2. Make the number of bytes from 'state' included into the hash
2073 independent from the number of PRNG bytes requested.
2075 The first measure alone would be sufficient to avoid
2076 Markku-Juhani's attack. (Actually it had never occurred
2077 to me that the half of 'md_local' used for chaining was the
2078 half from which PRNG output bytes were taken -- I had always
2079 assumed that the secret half would be used.) The second
2080 measure makes sure that additional data from 'state' is never
2081 mixed into 'md_local' in small portions; this heuristically
2082 further strengthens the PRNG.
2085 *) Fix crypto/bn/asm/mips3.s.
2088 *) When only the key is given to "enc", the IV is undefined. Print out
2089 an error message in this case.
2092 *) Handle special case when X509_NAME is empty in X509 printing routines.
2095 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
2096 positive and less than q.
2099 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
2100 used: it isn't thread safe and the add_lock_callback should handle
2102 [Paul Rose <Paul.Rose@bridge.com>]
2104 *) Verify that incoming data obeys the block size in
2105 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
2109 [Ulf Möller, Bodo Möller]
2111 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
2112 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
2113 when fixing the server behaviour for backwards-compatible 'client
2114 hello' messages. (Note that the attack is impractical against
2115 SSL 3.0 and TLS 1.0 anyway because length and version checking
2116 means that the probability of guessing a valid ciphertext is
2117 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
2120 Before 0.9.5, the countermeasure (hide the error by generating a
2121 random 'decryption result') did not work properly because
2122 ERR_clear_error() was missing, meaning that SSL_get_error() would
2123 detect the supposedly ignored error.
2125 Both problems are now fixed.
2128 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
2129 (previously it was 1024).
2132 *) Fix for compatibility mode trust settings: ignore trust settings
2133 unless some valid trust or reject settings are present.
2136 *) Fix for blowfish EVP: its a variable length cipher.
2139 *) Fix various bugs related to DSA S/MIME verification. Handle missing
2140 parameters in DSA public key structures and return an error in the
2141 DSA routines if parameters are absent.
2144 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
2145 in the current directory if neither $RANDFILE nor $HOME was set.
2146 RAND_file_name() in 0.9.6a returned NULL in this case. This has
2147 caused some confusion to Windows users who haven't defined $HOME.
2148 Thus RAND_file_name() is changed again: e_os.h can define a
2149 DEFAULT_HOME, which will be used if $HOME is not set.
2150 For Windows, we use "C:"; on other platforms, we still require
2151 environment variables.
2153 *) Move 'if (!initialized) RAND_poll()' into regions protected by
2154 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
2155 having multiple threads call RAND_poll() concurrently.
2158 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
2159 combination of a flag and a thread ID variable.
2160 Otherwise while one thread is in ssleay_rand_bytes (which sets the
2161 flag), *other* threads can enter ssleay_add_bytes without obeying
2162 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
2163 that they do not hold after the first thread unsets add_do_not_lock).
2166 *) Change bctest again: '-x' expressions are not available in all
2170 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
2172 *) Fix a couple of memory leaks in PKCS7_dataDecode()
2173 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
2175 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
2176 the default extension for executables, if any. Also, make the perl
2177 scripts that use symlink() to test if it really exists and use "cp"
2178 if it doesn't. All this made OpenSSL compilable and installable in
2182 *) Fix for asn1_GetSequence() for indefinite length constructed data.
2183 If SEQUENCE is length is indefinite just set c->slen to the total
2184 amount of data available.
2185 [Steve Henson, reported by shige@FreeBSD.org]
2186 [This change does not apply to 0.9.7.]
2188 *) Change bctest to avoid here-documents inside command substitution
2189 (workaround for FreeBSD /bin/sh bug).
2190 For compatibility with Ultrix, avoid shell functions (introduced
2191 in the bctest version that searches along $PATH).
2194 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
2195 with des_encrypt() defined on some operating systems, like Solaris
2199 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
2200 On the Importance of Eliminating Errors in Cryptographic
2201 Computations, J. Cryptology 14 (2001) 2, 101-119,
2202 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
2205 *) MIPS assembler BIGNUM division bug fix.
2208 *) Disabled incorrect Alpha assembler code.
2211 *) Fix PKCS#7 decode routines so they correctly update the length
2212 after reading an EOC for the EXPLICIT tag.
2214 [This change does not apply to 0.9.7.]
2216 *) Fix bug in PKCS#12 key generation routines. This was triggered
2217 if a 3DES key was generated with a 0 initial byte. Include
2218 PKCS12_BROKEN_KEYGEN compilation option to retain the old
2219 (but broken) behaviour.
2222 *) Enhance bctest to search for a working bc along $PATH and print
2224 [Tim Rice <tim@multitalents.net> via Richard Levitte]
2226 *) Fix memory leaks in err.c: free err_data string if necessary;
2227 don't write to the wrong index in ERR_set_error_data.
2230 *) Implement ssl23_peek (analogous to ssl23_read), which previously
2234 *) Replace rdtsc with _emit statements for VC++ version 5.
2235 [Jeremy Cooper <jeremy@baymoo.org>]
2237 *) Make it possible to reuse SSLv2 sessions.
2240 *) In copy_email() check for >= 0 as a return value for
2241 X509_NAME_get_index_by_NID() since 0 is a valid index.
2242 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
2244 *) Avoid coredump with unsupported or invalid public keys by checking if
2245 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
2246 PKCS7_verify() fails with non detached data.
2249 *) Don't use getenv in library functions when run as setuid/setgid.
2250 New function OPENSSL_issetugid().
2253 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
2254 due to incorrect handling of multi-threading:
2256 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
2258 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
2260 3. Count how many times MemCheck_off() has been called so that
2261 nested use can be treated correctly. This also avoids
2262 inband-signalling in the previous code (which relied on the
2263 assumption that thread ID 0 is impossible).
2266 *) Add "-rand" option also to s_client and s_server.
2269 *) Fix CPU detection on Irix 6.x.
2270 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
2271 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2273 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
2276 [This change does not apply to 0.9.7.]
2278 *) Use the cached encoding of an X509_NAME structure rather than
2279 copying it. This is apparently the reason for the libsafe "errors"
2280 but the code is actually correct.
2283 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
2284 Bleichenbacher's DSA attack.
2285 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
2286 to be set and top=0 forces the highest bit to be set; top=-1 is new
2287 and leaves the highest bit random.
2288 [Ulf Moeller, Bodo Moeller]
2290 *) In the NCONF_...-based implementations for CONF_... queries
2291 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
2292 a temporary CONF structure with the data component set to NULL
2293 (which gives segmentation faults in lh_retrieve).
2294 Instead, use NULL for the CONF pointer in CONF_get_string and
2295 CONF_get_number (which may use environment variables) and directly
2296 return NULL from CONF_get_section.
2299 *) Fix potential buffer overrun for EBCDIC.
2302 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
2303 keyUsage if basicConstraints absent for a CA.
2306 *) Make SMIME_write_PKCS7() write mail header values with a format that
2307 is more generally accepted (no spaces before the semicolon), since
2308 some programs can't parse those values properly otherwise. Also make
2309 sure BIO's that break lines after each write do not create invalid
2313 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
2314 macros previously used would not encode an empty SEQUENCE OF
2315 and break the signature.
2317 [This change does not apply to 0.9.7.]
2319 *) Zero the premaster secret after deriving the master secret in
2323 *) Add some EVP_add_digest_alias registrations (as found in
2324 OpenSSL_add_all_digests()) to SSL_library_init()
2325 aka OpenSSL_add_ssl_algorithms(). This provides improved
2326 compatibility with peers using X.509 certificates
2327 with unconventional AlgorithmIdentifier OIDs.
2330 *) Fix for Irix with NO_ASM.
2331 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
2333 *) ./config script fixes.
2334 [Ulf Moeller, Richard Levitte]
2336 *) Fix 'openssl passwd -1'.
2339 *) Change PKCS12_key_gen_asc() so it can cope with non null
2340 terminated strings whose length is passed in the passlen
2341 parameter, for example from PEM callbacks. This was done
2342 by adding an extra length parameter to asc2uni().
2343 [Steve Henson, reported by <oddissey@samsung.co.kr>]
2345 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
2346 call failed, free the DSA structure.
2349 *) Fix to uni2asc() to cope with zero length Unicode strings.
2350 These are present in some PKCS#12 files.
2353 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
2354 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
2355 when writing a 32767 byte record.
2356 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
2358 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
2359 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
2361 (RSA objects have a reference count access to which is protected
2362 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
2363 so they are meant to be shared between threads.)
2364 [Bodo Moeller, Geoff Thorpe; original patch submitted by
2365 "Reddie, Steven" <Steven.Reddie@ca.com>]
2367 *) Fix a deadlock in CRYPTO_mem_leaks().
2370 *) Use better test patterns in bntest.
2373 *) rand_win.c fix for Borland C.
2376 *) BN_rshift bugfix for n == 0.
2379 *) Add a 'bctest' script that checks for some known 'bc' bugs
2380 so that 'make test' does not abort just because 'bc' is broken.
2383 *) Store verify_result within SSL_SESSION also for client side to
2384 avoid potential security hole. (Re-used sessions on the client side
2385 always resulted in verify_result==X509_V_OK, not using the original
2386 result of the server certificate verification.)
2389 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2390 SSL3_RT_APPLICATION_DATA, return 0.
2391 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2395 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2396 releases, have been re-implemented by renaming the previous
2397 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2398 and ssl3_read_internal, respectively, and adding 'peek' parameters
2399 to them. The new ssl[23]_{read,peek} functions are calls to
2400 ssl[23]_read_internal with the 'peek' flag set appropriately.
2401 A 'peek' parameter has also been added to ssl3_read_bytes, which
2402 does the actual work for ssl3_read_internal.
2405 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2406 the method-specific "init()" handler. Also clean up ex_data after
2407 calling the method-specific "finish()" handler. Previously, this was
2408 happening the other way round.
2411 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2412 The previous value, 12, was not always sufficient for BN_mod_exp().
2415 *) Make sure that shared libraries get the internal name engine with
2416 the full version number and not just 0. This should mark the
2417 shared libraries as not backward compatible. Of course, this should
2418 be changed again when we can guarantee backward binary compatibility.
2421 *) Fix typo in get_cert_by_subject() in by_dir.c
2422 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2424 *) Rework the system to generate shared libraries:
2426 - Make note of the expected extension for the shared libraries and
2427 if there is a need for symbolic links from for example libcrypto.so.0
2428 to libcrypto.so.0.9.7. There is extended info in Configure for
2431 - Make as few rebuilds of the shared libraries as possible.
2433 - Still avoid linking the OpenSSL programs with the shared libraries.
2435 - When installing, install the shared libraries separately from the
2439 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2441 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2442 and not in SSL_clear because the latter is also used by the
2443 accept/connect functions; previously, the settings made by
2444 SSL_set_read_ahead would be lost during the handshake.
2445 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2447 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2448 Previously, it would create entries for disableed algorithms no
2452 *) Added several new manual pages for SSL_* function.
2455 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2457 *) In ssl23_get_client_hello, generate an error message when faced
2458 with an initial SSL 3.0/TLS record that is too small to contain the
2459 first two bytes of the ClientHello message, i.e. client_version.
2460 (Note that this is a pathologic case that probably has never happened
2461 in real life.) The previous approach was to use the version number
2462 from the record header as a substitute; but our protocol choice
2463 should not depend on that one because it is not authenticated
2464 by the Finished messages.
2467 *) More robust randomness gathering functions for Windows.
2468 [Jeffrey Altman <jaltman@columbia.edu>]
2470 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2471 not set then we don't setup the error code for issuer check errors
2472 to avoid possibly overwriting other errors which the callback does
2473 handle. If an application does set the flag then we assume it knows
2474 what it is doing and can handle the new informational codes
2478 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2479 a general "ANY" type, as such it should be able to decode anything
2480 including tagged types. However it didn't check the class so it would
2481 wrongly interpret tagged types in the same way as their universal
2482 counterpart and unknown types were just rejected. Changed so that the
2483 tagged and unknown types are handled in the same way as a SEQUENCE:
2484 that is the encoding is stored intact. There is also a new type
2485 "V_ASN1_OTHER" which is used when the class is not universal, in this
2486 case we have no idea what the actual type is so we just lump them all
2490 *) On VMS, stdout may very well lead to a file that is written to
2491 in a record-oriented fashion. That means that every write() will
2492 write a separate record, which will be read separately by the
2493 programs trying to read from it. This can be very confusing.
2495 The solution is to put a BIO filter in the way that will buffer
2496 text until a linefeed is reached, and then write everything a
2497 line at a time, so every record written will be an actual line,
2498 not chunks of lines and not (usually doesn't happen, but I've
2499 seen it once) several lines in one record. BIO_f_linebuffer() is
2502 Currently, it's a VMS-only method, because that's where it has
2503 been tested well enough.
2506 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2507 it can return incorrect results.
2508 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2509 but it was in 0.9.6-beta[12].)
2512 *) Disable the check for content being present when verifying detached
2513 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2514 include zero length content when signing messages.
2517 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2518 BIO_ctrl (for BIO pairs).
2521 *) Add DSO method for VMS.
2524 *) Bug fix: Montgomery multiplication could produce results with the
2528 *) Add RPM specification openssl.spec and modify it to build three
2529 packages. The default package contains applications, application
2530 documentation and run-time libraries. The devel package contains
2531 include files, static libraries and function documentation. The
2532 doc package contains the contents of the doc directory. The original
2533 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2536 *) Add a large number of documentation files for many SSL routines.
2537 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2539 *) Add a configuration entry for Sony News 4.
2540 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2542 *) Don't set the two most significant bits to one when generating a
2543 random number < q in the DSA library.
2546 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2547 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2548 the underlying transport is blocking) if a handshake took place.
2549 (The default behaviour is needed by applications such as s_client
2550 and s_server that use select() to determine when to use SSL_read;
2551 but for applications that know in advance when to expect data, it
2552 just makes things more complicated.)
2555 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2559 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2560 work better on such systems.
2561 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2563 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2564 Update PKCS12_parse() so it copies the friendlyName and the
2565 keyid to the certificates aux info.
2568 *) Fix bug in PKCS7_verify() which caused an infinite loop
2569 if there was more than one signature.
2570 [Sven Uszpelkat <su@celocom.de>]
2572 *) Major change in util/mkdef.pl to include extra information
2573 about each symbol, as well as presentig variables as well
2574 as functions. This change means that there's n more need
2575 to rebuild the .num files when some algorithms are excluded.
2578 *) Allow the verify time to be set by an application,
2579 rather than always using the current time.
2582 *) Phase 2 verify code reorganisation. The certificate
2583 verify code now looks up an issuer certificate by a
2584 number of criteria: subject name, authority key id
2585 and key usage. It also verifies self signed certificates
2586 by the same criteria. The main comparison function is
2587 X509_check_issued() which performs these checks.
2589 Lot of changes were necessary in order to support this
2590 without completely rewriting the lookup code.
2592 Authority and subject key identifier are now cached.
2594 The LHASH 'certs' is X509_STORE has now been replaced
2595 by a STACK_OF(X509_OBJECT). This is mainly because an
2596 LHASH can't store or retrieve multiple objects with
2597 the same hash value.
2599 As a result various functions (which were all internal
2600 use only) have changed to handle the new X509_STORE
2601 structure. This will break anything that messed round
2602 with X509_STORE internally.
2604 The functions X509_STORE_add_cert() now checks for an
2605 exact match, rather than just subject name.
2607 The X509_STORE API doesn't directly support the retrieval
2608 of multiple certificates matching a given criteria, however
2609 this can be worked round by performing a lookup first
2610 (which will fill the cache with candidate certificates)
2611 and then examining the cache for matches. This is probably
2612 the best we can do without throwing out X509_LOOKUP
2613 entirely (maybe later...).
2615 The X509_VERIFY_CTX structure has been enhanced considerably.
2617 All certificate lookup operations now go via a get_issuer()
2618 callback. Although this currently uses an X509_STORE it
2619 can be replaced by custom lookups. This is a simple way
2620 to bypass the X509_STORE hackery necessary to make this
2621 work and makes it possible to use more efficient techniques
2622 in future. A very simple version which uses a simple
2623 STACK for its trusted certificate store is also provided
2624 using X509_STORE_CTX_trusted_stack().
2626 The verify_cb() and verify() callbacks now have equivalents
2627 in the X509_STORE_CTX structure.
2629 X509_STORE_CTX also has a 'flags' field which can be used
2630 to customise the verify behaviour.
2633 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2634 excludes S/MIME capabilities.
2637 *) When a certificate request is read in keep a copy of the
2638 original encoding of the signed data and use it when outputing
2639 again. Signatures then use the original encoding rather than
2640 a decoded, encoded version which may cause problems if the
2641 request is improperly encoded.
2644 *) For consistency with other BIO_puts implementations, call
2645 buffer_write(b, ...) directly in buffer_puts instead of calling
2648 In BIO_puts, increment b->num_write as in BIO_write.
2649 [Peter.Sylvester@EdelWeb.fr]
2651 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2652 BN_zero, we may not return a BIGNUM with an array consisting of
2656 *) Avoid calling abort() from within the library when problems are
2657 detected, except if preprocessor symbols have been defined
2658 (such as REF_CHECK, BN_DEBUG etc.).
2661 *) New openssl application 'rsautl'. This utility can be
2662 used for low level RSA operations. DER public key
2663 BIO/fp routines also added.
2666 *) New Configure entry and patches for compiling on QNX 4.
2667 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2669 *) A demo state-machine implementation was sponsored by
2670 Nuron (http://www.nuron.com/) and is now available in
2671 demos/state_machine.
2674 *) New options added to the 'dgst' utility for signature
2675 generation and verification.
2678 *) Unrecognized PKCS#7 content types are now handled via a
2679 catch all ASN1_TYPE structure. This allows unsupported
2680 types to be stored as a "blob" and an application can
2681 encode and decode it manually.
2684 *) Fix various signed/unsigned issues to make a_strex.c
2686 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2688 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2689 length if passed a buffer. ASN1_INTEGER_to_BN failed
2690 if passed a NULL BN and its argument was negative.
2691 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2693 *) Modification to PKCS#7 encoding routines to output definite
2694 length encoding. Since currently the whole structures are in
2695 memory there's not real point in using indefinite length
2696 constructed encoding. However if OpenSSL is compiled with
2697 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2700 *) Added BIO_vprintf() and BIO_vsnprintf().
2703 *) Added more prefixes to parse for in the the strings written
2704 through a logging bio, to cover all the levels that are available
2705 through syslog. The prefixes are now:
2707 PANIC, EMERG, EMR => LOG_EMERG
2708 ALERT, ALR => LOG_ALERT
2709 CRIT, CRI => LOG_CRIT
2710 ERROR, ERR => LOG_ERR
2711 WARNING, WARN, WAR => LOG_WARNING
2712 NOTICE, NOTE, NOT => LOG_NOTICE
2713 INFO, INF => LOG_INFO
2714 DEBUG, DBG => LOG_DEBUG
2716 and as before, if none of those prefixes are present at the
2717 beginning of the string, LOG_ERR is chosen.
2719 On Win32, the LOG_* levels are mapped according to this:
2721 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2722 LOG_WARNING => EVENTLOG_WARNING_TYPE
2723 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2727 *) Made it possible to reconfigure with just the configuration
2728 argument "reconf" or "reconfigure". The command line arguments
2729 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2730 and are retrieved from there when reconfiguring.
2734 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2736 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2739 *) The obj_dat.pl script was messing up the sorting of object
2740 names. The reason was that it compared the quoted version
2741 of strings as a result "OCSP" > "OCSP Signing" because
2742 " > SPACE. Changed script to store unquoted versions of
2743 names and add quotes on output. It was also omitting some
2744 names from the lookup table if they were given a default
2745 value (that is if SN is missing it is given the same
2746 value as LN and vice versa), these are now added on the
2747 grounds that if an object has a name we should be able to
2748 look it up. Finally added warning output when duplicate
2749 short or long names are found.
2752 *) Changes needed for Tandem NSK.
2753 [Scott Uroff <scott@xypro.com>]
2755 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2756 RSA_padding_check_SSLv23(), special padding was never detected
2757 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2758 version rollback attacks was not effective.
2760 In s23_clnt.c, don't use special rollback-attack detection padding
2761 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2762 client; similarly, in s23_srvr.c, don't do the rollback check if
2763 SSL 2.0 is the only protocol enabled in the server.
2766 *) Make it possible to get hexdumps of unprintable data with 'openssl
2767 asn1parse'. By implication, the functions ASN1_parse_dump() and
2768 BIO_dump_indent() are added.
2771 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2772 these print out strings and name structures based on various
2773 flags including RFC2253 support and proper handling of
2774 multibyte characters. Added options to the 'x509' utility
2775 to allow the various flags to be set.
2778 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2779 Also change the functions X509_cmp_current_time() and
2780 X509_gmtime_adj() work with an ASN1_TIME structure,
2781 this will enable certificates using GeneralizedTime in validity
2782 dates to be checked.
2785 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2786 negative public key encodings) on by default,
2787 NO_NEG_PUBKEY_BUG can be set to disable it.
2790 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2791 content octets. An i2c_ASN1_OBJECT is unnecessary because
2792 the encoding can be trivially obtained from the structure.
2795 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2796 not read locks (CRYPTO_r_[un]lock).
2799 *) A first attempt at creating official support for shared
2800 libraries through configuration. I've kept it so the
2801 default is static libraries only, and the OpenSSL programs
2802 are always statically linked for now, but there are
2803 preparations for dynamic linking in place.
2804 This has been tested on Linux and Tru64.
2807 *) Randomness polling function for Win9x, as described in:
2808 Peter Gutmann, Software Generation of Practically Strong
2812 *) Fix so PRNG is seeded in req if using an already existing
2816 *) New options to smime application. -inform and -outform
2817 allow alternative formats for the S/MIME message including
2818 PEM and DER. The -content option allows the content to be
2819 specified separately. This should allow things like Netscape
2820 form signing output easier to verify.
2823 *) Fix the ASN1 encoding of tags using the 'long form'.
2826 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2827 STRING types. These convert content octets to and from the
2828 underlying type. The actual tag and length octets are
2829 already assumed to have been read in and checked. These
2830 are needed because all other string types have virtually
2831 identical handling apart from the tag. By having versions
2832 of the ASN1 functions that just operate on content octets
2833 IMPLICIT tagging can be handled properly. It also allows
2834 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2835 and ASN1_INTEGER are identical apart from the tag.
2838 *) Change the handling of OID objects as follows:
2840 - New object identifiers are inserted in objects.txt, following
2841 the syntax given in objects.README.
2842 - objects.pl is used to process obj_mac.num and create a new
2844 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2847 This is currently kind of a hack, and the perl code in objects.pl
2848 isn't very elegant, but it works as I intended. The simplest way
2849 to check that it worked correctly is to look in obj_dat.h and
2850 check the array nid_objs and make sure the objects haven't moved
2851 around (this is important!). Additions are OK, as well as
2852 consistent name changes.
2855 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2858 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2859 The given file adds to whatever has already been seeded into the
2860 random pool through the RANDFILE configuration file option or
2861 environment variable, or the default random state file.
2864 *) mkstack.pl now sorts each macro group into lexical order.
2865 Previously the output order depended on the order the files
2866 appeared in the directory, resulting in needless rewriting
2870 *) Patches to make OpenSSL compile under Win32 again. Mostly
2871 work arounds for the VC++ problem that it treats func() as
2872 func(void). Also stripped out the parts of mkdef.pl that
2873 added extra typesafe functions: these no longer exist.
2876 *) Reorganisation of the stack code. The macros are now all
2877 collected in safestack.h . Each macro is defined in terms of
2878 a "stack macro" of the form SKM_<name>(type, a, b). The
2879 DEBUG_SAFESTACK is now handled in terms of function casts,
2880 this has the advantage of retaining type safety without the
2881 use of additional functions. If DEBUG_SAFESTACK is not defined
2882 then the non typesafe macros are used instead. Also modified the
2883 mkstack.pl script to handle the new form. Needs testing to see
2884 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2885 the default if no major problems. Similar behaviour for ASN1_SET_OF
2886 and PKCS12_STACK_OF.
2889 *) When some versions of IIS use the 'NET' form of private key the
2890 key derivation algorithm is different. Normally MD5(password) is
2891 used as a 128 bit RC4 key. In the modified case
2892 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2893 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2894 as the old Netscape_RSA functions except they have an additional
2895 'sgckey' parameter which uses the modified algorithm. Also added
2896 an -sgckey command line option to the rsa utility. Thanks to
2897 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2898 algorithm to openssl-dev.
2901 *) The evp_local.h macros were using 'c.##kname' which resulted in
2902 invalid expansion on some systems (SCO 5.0.5 for example).
2903 Corrected to 'c.kname'.
2904 [Phillip Porch <root@theporch.com>]
2906 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2907 a STACK of email addresses from a certificate or request, these look
2908 in the subject name and the subject alternative name extensions and
2909 omit any duplicate addresses.
2912 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2913 This makes DSA verification about 2 % faster.
2916 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2917 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2918 plus overhead for 1024 bit moduli).
2919 This makes exponentiations about 0.5 % faster for 1024 bit
2920 exponents (as measured by "openssl speed rsa2048").
2923 *) Rename memory handling macros to avoid conflicts with other
2925 Malloc => OPENSSL_malloc
2926 Malloc_locked => OPENSSL_malloc_locked
2927 Realloc => OPENSSL_realloc
2928 Free => OPENSSL_free
2931 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2932 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2935 *) CygWin32 support.
2936 [John Jarvie <jjarvie@newsguy.com>]
2938 *) The type-safe stack code has been rejigged. It is now only compiled
2939 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2940 by default all type-specific stack functions are "#define"d back to
2941 standard stack functions. This results in more streamlined output
2942 but retains the type-safety checking possibilities of the original
2946 *) The STACK code has been cleaned up, and certain type declarations
2947 that didn't make a lot of sense have been brought in line. This has
2948 also involved a cleanup of sorts in safestack.h to more correctly
2949 map type-safe stack functions onto their plain stack counterparts.
2950 This work has also resulted in a variety of "const"ifications of
2951 lots of the code, especially "_cmp" operations which should normally
2952 be prototyped with "const" parameters anyway.
2955 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2956 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2957 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2958 where all of 'md' is used each time the PRNG is used, but 'state'
2959 is used only indexed by a cyclic counter. As entropy may not be
2960 well distributed from the beginning, 'md' is important as a
2961 chaining variable. However, the output function chains only half
2962 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2963 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2964 in all of 'state' being rewritten, with the new values depending
2965 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2968 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2969 the handshake is continued after ssl_verify_cert_chain();
2970 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2971 can lead to 'unexplainable' connection aborts later.
2972 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2974 *) Major EVP API cipher revision.
2975 Add hooks for extra EVP features. This allows various cipher
2976 parameters to be set in the EVP interface. Support added for variable
2977 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2978 setting of RC2 and RC5 parameters.
2980 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2983 Remove lots of duplicated code from the EVP library. For example *every*
2984 cipher init() function handles the 'iv' in the same way according to the
2985 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2986 for CFB and OFB modes they zero ctx->num.
2988 New functionality allows removal of S/MIME code RC2 hack.
2990 Most of the routines have the same form and so can be declared in terms
2993 By shifting this to the top level EVP_CipherInit() it can be removed from
2994 all individual ciphers. If the cipher wants to handle IVs or keys
2995 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2998 Change lots of functions like EVP_EncryptUpdate() to now return a
2999 value: although software versions of the algorithms cannot fail
3000 any installed hardware versions can.
3003 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
3004 this option is set, tolerate broken clients that send the negotiated
3005 protocol version number instead of the requested protocol version
3009 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
3010 i.e. non-zero for export ciphersuites, zero otherwise.
3011 Previous versions had this flag inverted, inconsistent with
3012 rsa_tmp_cb (..._TMP_RSA_CB).
3013 [Bodo Moeller; problem reported by Amit Chopra]
3015 *) Add missing DSA library text string. Work around for some IIS
3016 key files with invalid SEQUENCE encoding.
3019 *) Add a document (doc/standards.txt) that list all kinds of standards
3020 and so on that are implemented in OpenSSL.
3023 *) Enhance c_rehash script. Old version would mishandle certificates
3024 with the same subject name hash and wouldn't handle CRLs at all.
3025 Added -fingerprint option to crl utility, to support new c_rehash
3029 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
3032 *) Fix for SSL server purpose checking. Server checking was
3033 rejecting certificates which had extended key usage present
3034 but no ssl client purpose.
3035 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
3037 *) Make PKCS#12 code work with no password. The PKCS#12 spec
3038 is a little unclear about how a blank password is handled.
3039 Since the password in encoded as a BMPString with terminating
3040 double NULL a zero length password would end up as just the
3041 double NULL. However no password at all is different and is
3042 handled differently in the PKCS#12 key generation code. NS
3043 treats a blank password as zero length. MSIE treats it as no
3044 password on export: but it will try both on import. We now do
3045 the same: PKCS12_parse() tries zero length and no password if
3046 the password is set to "" or NULL (NULL is now a valid password:
3047 it wasn't before) as does the pkcs12 application.
3050 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
3051 perror when PEM_read_bio_X509_REQ fails, the error message must
3052 be obtained from the error queue.
3055 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
3056 it in ERR_remove_state if appropriate, and change ERR_get_state
3057 accordingly to avoid race conditions (this is necessary because
3058 thread_hash is no longer constant once set).
3061 *) Bugfix for linux-elf makefile.one.
3064 *) RSA_get_default_method() will now cause a default
3065 RSA_METHOD to be chosen if one doesn't exist already.
3066 Previously this was only set during a call to RSA_new()
3067 or RSA_new_method(NULL) meaning it was possible for
3068 RSA_get_default_method() to return NULL.
3071 *) Added native name translation to the existing DSO code
3072 that will convert (if the flag to do so is set) filenames
3073 that are sufficiently small and have no path information
3074 into a canonical native form. Eg. "blah" converted to
3075 "libblah.so" or "blah.dll" etc.
3078 *) New function ERR_error_string_n(e, buf, len) which is like
3079 ERR_error_string(e, buf), but writes at most 'len' bytes
3080 including the 0 terminator. For ERR_error_string_n, 'buf'
3082 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
3084 *) CONF library reworked to become more general. A new CONF
3085 configuration file reader "class" is implemented as well as a
3086 new functions (NCONF_*, for "New CONF") to handle it. The now
3087 old CONF_* functions are still there, but are reimplemented to
3088 work in terms of the new functions. Also, a set of functions
3089 to handle the internal storage of the configuration data is
3090 provided to make it easier to write new configuration file
3091 reader "classes" (I can definitely see something reading a
3092 configuration file in XML format, for example), called _CONF_*,
3093 or "the configuration storage API"...
3095 The new configuration file reading functions are:
3097 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
3098 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
3100 NCONF_default, NCONF_WIN32
3102 NCONF_dump_fp, NCONF_dump_bio
3104 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
3105 NCONF_new creates a new CONF object. This works in the same way
3106 as other interfaces in OpenSSL, like the BIO interface.
3107 NCONF_dump_* dump the internal storage of the configuration file,
3108 which is useful for debugging. All other functions take the same
3109 arguments as the old CONF_* functions wth the exception of the
3110 first that must be a `CONF *' instead of a `LHASH *'.
3112 To make it easer to use the new classes with the old CONF_* functions,
3113 the function CONF_set_default_method is provided.
3116 *) Add '-tls1' option to 'openssl ciphers', which was already
3117 mentioned in the documentation but had not been implemented.
3118 (This option is not yet really useful because even the additional
3119 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
3122 *) Initial DSO code added into libcrypto for letting OpenSSL (and
3123 OpenSSL-based applications) load shared libraries and bind to
3124 them in a portable way.
3125 [Geoff Thorpe, with contributions from Richard Levitte]
3127 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
3129 *) Make sure _lrotl and _lrotr are only used with MSVC.
3131 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
3132 (the default implementation of RAND_status).
3134 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
3135 to '-clrext' (= clear extensions), as intended and documented.
3136 [Bodo Moeller; inconsistency pointed out by Michael Attili
3139 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3140 was larger than the MD block size.
3141 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3143 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3144 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3145 using the passed key: if the passed key was a private key the result
3146 of X509_print(), for example, would be to print out all the private key
3150 *) des_quad_cksum() byte order bug fix.
3151 [Ulf Möller, using the problem description in krb4-0.9.7, where
3152 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3154 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3156 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3158 *) For easily testing in shell scripts whether some command
3159 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3160 returns with exit code 0 iff no command of the given name is available.
3161 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3162 the output goes to stdout and nothing is printed to stderr.
3163 Additional arguments are always ignored.
3165 Since for each cipher there is a command of the same name,
3166 the 'no-cipher' compilation switches can be tested this way.
3168 ('openssl no-XXX' is not able to detect pseudo-commands such
3169 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3172 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3175 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3176 is set; it will be thrown away anyway because each handshake creates
3178 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3179 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3180 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3181 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3184 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3185 'Q' and 'R' lose their special meanings (quit/renegotiate).
3186 This is part of what -quiet does; unlike -quiet, -ign_eof
3187 does not suppress any output.
3190 *) Add compatibility options to the purpose and trust code. The
3191 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3192 accepts a certificate or CA, this was the previous behaviour,
3193 with all the associated security issues.
3195 X509_TRUST_COMPAT is the old trust behaviour: only and
3196 automatically trust self signed roots in certificate store. A
3197 new trust setting X509_TRUST_DEFAULT is used to specify that
3198 a purpose has no associated trust setting and it should instead
3199 use the value in the default purpose.
3202 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3203 and fix a memory leak.
3206 *) In util/mkerr.pl (which implements 'make errors'), preserve
3207 reason strings from the previous version of the .c file, as
3208 the default to have only downcase letters (and digits) in
3209 automatically generated reasons codes is not always appropriate.
3212 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3213 using strerror. Previously, ERR_reason_error_string() returned
3214 library names as reason strings for SYSerr; but SYSerr is a special
3215 case where small numbers are errno values, not library numbers.
3218 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3219 converts DSA parameters into DH parameters. (When creating parameters,
3220 DSA_generate_parameters is used.)
3223 *) Include 'length' (recommended exponent length) in C code generated
3224 by 'openssl dhparam -C'.
3227 *) The second argument to set_label in perlasm was already being used
3228 so couldn't be used as a "file scope" flag. Moved to third argument
3232 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3233 instead of RAND_bytes for encryption IVs and salts.
3236 *) Include RAND_status() into RAND_METHOD instead of implementing
3237 it only for md_rand.c Otherwise replacing the PRNG by calling
3238 RAND_set_rand_method would be impossible.
3241 *) Don't let DSA_generate_key() enter an infinite loop if the random
3242 number generation fails.
3245 *) New 'rand' application for creating pseudo-random output.
3248 *) Added configuration support for Linux/IA64
3249 [Rolf Haberrecker <rolf@suse.de>]
3251 *) Assembler module support for Mingw32.
3254 *) Shared library support for HPUX (in shlib/).
3255 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3257 *) Shared library support for Solaris gcc.
3258 [Lutz Behnke <behnke@trustcenter.de>]
3260 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3262 *) PKCS7_encrypt() was adding text MIME headers twice because they
3263 were added manually and by SMIME_crlf_copy().
3266 *) In bntest.c don't call BN_rand with zero bits argument.
3267 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3269 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3270 case was implemented. This caused BN_div_recp() to fail occasionally.
3273 *) Add an optional second argument to the set_label() in the perl
3274 assembly language builder. If this argument exists and is set
3275 to 1 it signals that the assembler should use a symbol whose
3276 scope is the entire file, not just the current function. This
3277 is needed with MASM which uses the format label:: for this scope.
3278 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3280 *) Change the ASN1 types so they are typedefs by default. Before
3281 almost all types were #define'd to ASN1_STRING which was causing
3282 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3286 *) Change names of new functions to the new get1/get0 naming
3287 convention: After 'get1', the caller owns a reference count
3288 and has to call ..._free; 'get0' returns a pointer to some
3289 data structure without incrementing reference counters.
3290 (Some of the existing 'get' functions increment a reference
3291 counter, some don't.)
3292 Similarly, 'set1' and 'add1' functions increase reference
3293 counters or duplicate objects.
3296 *) Allow for the possibility of temp RSA key generation failure:
3297 the code used to assume it always worked and crashed on failure.
3300 *) Fix potential buffer overrun problem in BIO_printf().
3301 [Ulf Möller, using public domain code by Patrick Powell; problem
3302 pointed out by David Sacerdote <das33@cornell.edu>]
3304 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3305 RAND_egd() and RAND_status(). In the command line application,
3306 the EGD socket can be specified like a seed file using RANDFILE
3310 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3311 Some CAs (e.g. Verisign) distribute certificates in this form.
3314 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3315 list to exclude them. This means that no special compilation option
3316 is needed to use anonymous DH: it just needs to be included in the
3320 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3321 EVP_MD_type. The old functionality is available in a new macro called
3322 EVP_MD_md(). Change code that uses it and update docs.
3325 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3326 where the 'void *' argument is replaced by a function pointer argument.
3327 Previously 'void *' was abused to point to functions, which works on
3328 many platforms, but is not correct. As these functions are usually
3329 called by macros defined in OpenSSL header files, most source code
3330 should work without changes.
3333 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3334 sections with information on -D... compiler switches used for
3335 compiling the library so that applications can see them. To enable
3336 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3337 must be defined. E.g.,
3338 #define OPENSSL_ALGORITHM_DEFINES
3339 #include <openssl/opensslconf.h>
3340 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3341 [Richard Levitte, Ulf and Bodo Möller]
3343 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3347 *) Change the 'other' type in certificate aux info to a STACK_OF
3348 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3349 the required ASN1 format: arbitrary types determined by an OID.
3352 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3353 argument to 'req'. This is not because the function is newer or
3354 better than others it just uses the work 'NEW' in the certificate
3355 request header lines. Some software needs this.
3358 *) Reorganise password command line arguments: now passwords can be
3359 obtained from various sources. Delete the PEM_cb function and make
3360 it the default behaviour: i.e. if the callback is NULL and the
3361 usrdata argument is not NULL interpret it as a null terminated pass
3362 phrase. If usrdata and the callback are NULL then the pass phrase
3363 is prompted for as usual.
3366 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3367 the support is automatically enabled. The resulting binaries will
3368 autodetect the card and use it if present.
3369 [Ben Laurie and Compaq Inc.]
3371 *) Work around for Netscape hang bug. This sends certificate request
3372 and server done in one record. Since this is perfectly legal in the
3373 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3374 the bugs/SSLv3 entry for more info.
3377 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3380 *) Add -rand argument to smime and pkcs12 applications and read/write
3384 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3387 *) Add command line password options to the remaining applications.
3390 *) Bug fix for BN_div_recp() for numerators with an even number of
3394 *) More tests in bntest.c, and changed test_bn output.
3397 *) ./config recognizes MacOS X now.
3400 *) Bug fix for BN_div() when the first words of num and divsor are
3401 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3404 *) Add support for various broken PKCS#8 formats, and command line
3405 options to produce them.
3408 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3409 get temporary BIGNUMs from a BN_CTX.
3412 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3416 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3417 include a #define from the old name to the new. The original intent
3418 was that statically linked binaries could for example just call
3419 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3420 link with digests. This never worked becayse SSLeay_add_all_digests()
3421 and SSLeay_add_all_ciphers() were in the same source file so calling
3422 one would link with the other. They are now in separate source files.
3425 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3428 *) Use a less unusual form of the Miller-Rabin primality test (it used
3429 a binary algorithm for exponentiation integrated into the Miller-Rabin
3430 loop, our standard modexp algorithms are faster).
3433 *) Support for the EBCDIC character set completed.
3434 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3436 *) Source code cleanups: use const where appropriate, eliminate casts,
3437 use void * instead of char * in lhash.
3440 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3441 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3442 this the server could overwrite ephemeral keys that the client
3446 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3447 using 50 iterations of the Rabin-Miller test.
3449 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3450 iterations of the Rabin-Miller test as required by the appendix
3451 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3452 As BN_is_prime_fasttest includes trial division, DSA parameter
3453 generation becomes much faster.
3455 This implies a change for the callback functions in DSA_is_prime
3456 and DSA_generate_parameters: The callback function is called once
3457 for each positive witness in the Rabin-Miller test, not just
3458 occasionally in the inner loop; and the parameters to the
3459 callback function now provide an iteration count for the outer
3460 loop rather than for the current invocation of the inner loop.
3461 DSA_generate_parameters additionally can call the callback
3462 function with an 'iteration count' of -1, meaning that a
3463 candidate has passed the trial division test (when q is generated
3464 from an application-provided seed, trial division is skipped).
3467 *) New function BN_is_prime_fasttest that optionally does trial
3468 division before starting the Rabin-Miller test and has
3469 an additional BN_CTX * argument (whereas BN_is_prime always
3470 has to allocate at least one BN_CTX).
3471 'callback(1, -1, cb_arg)' is called when a number has passed the
3472 trial division stage.
3475 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3479 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3482 *) New function BN_pseudo_rand().
3485 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3486 bignum version of BN_from_montgomery() with the working code from
3487 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3491 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3492 made it impossible to use the same SSL_SESSION data structure in
3493 SSL2 clients in multiple threads.
3496 *) The return value of RAND_load_file() no longer counts bytes obtained
3497 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3498 to seed the PRNG (previously an explicit byte count was required).
3499 [Ulf Möller, Bodo Möller]
3501 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3502 used (char *) instead of (void *) and had casts all over the place.
3505 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3508 *) Retain source code compatibility for BN_prime_checks macro:
3509 BN_is_prime(..., BN_prime_checks, ...) now uses
3510 BN_prime_checks_for_size to determine the appropriate number of
3511 Rabin-Miller iterations.
3514 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3515 DH_CHECK_P_NOT_SAFE_PRIME.
3516 (Check if this is true? OpenPGP calls them "strong".)
3519 *) Merge the functionality of "dh" and "gendh" programs into a new program
3520 "dhparam". The old programs are retained for now but will handle DH keys
3521 (instead of parameters) in future.
3524 *) Make the ciphers, s_server and s_client programs check the return values
3525 when a new cipher list is set.
3528 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3529 ciphers. Before when the 56bit ciphers were enabled the sorting was
3532 The syntax for the cipher sorting has been extended to support sorting by
3533 cipher-strength (using the strength_bits hard coded in the tables).
3534 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3536 Fix a bug in the cipher-command parser: when supplying a cipher command
3537 string with an "undefined" symbol (neither command nor alphanumeric
3538 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3539 an error is flagged.
3541 Due to the strength-sorting extension, the code of the
3542 ssl_create_cipher_list() function was completely rearranged. I hope that
3543 the readability was also increased :-)
3544 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3546 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3547 for the first serial number and places 2 in the serial number file. This
3548 avoids problems when the root CA is created with serial number zero and
3549 the first user certificate has the same issuer name and serial number
3553 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3554 the new code. Add documentation for this stuff.
3557 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3558 X509_*() to X509at_*() on the grounds that they don't handle X509
3559 structures and behave in an analagous way to the X509v3 functions:
3560 they shouldn't be called directly but wrapper functions should be used
3563 So we also now have some wrapper functions that call the X509at functions
3564 when passed certificate requests. (TO DO: similar things can be done with
3565 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3566 things. Some of these need some d2i or i2d and print functionality
3567 because they handle more complex structures.)
3570 *) Add missing #ifndefs that caused missing symbols when building libssl
3571 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3572 NO_RSA in ssl/s2*.c.
3573 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3575 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3576 has a return value which indicates the quality of the random data
3577 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3578 error queue. New function RAND_pseudo_bytes() generates output that is
3579 guaranteed to be unique but not unpredictable. RAND_add is like
3580 RAND_seed, but takes an extra argument for an entropy estimate
3581 (RAND_seed always assumes full entropy).
3584 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3585 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3586 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3587 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3588 false-positive rate of at most 2^-80 for random input.
3591 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3594 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3595 in the 0.9.5 release), this returns the chain
3596 from an X509_CTX structure with a dup of the stack and all
3597 the X509 reference counts upped: so the stack will exist
3598 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3601 Also make SSL_SESSION_print() print out the verify return
3605 *) Add manpage for the pkcs12 command. Also change the default
3606 behaviour so MAC iteration counts are used unless the new
3607 -nomaciter option is used. This improves file security and
3608 only older versions of MSIE (4.0 for example) need it.
3611 *) Honor the no-xxx Configure options when creating .DEF files.
3614 *) Add PKCS#10 attributes to field table: challengePassword,
3615 unstructuredName and unstructuredAddress. These are taken from
3616 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3617 international characters are used.
3619 More changes to X509_ATTRIBUTE code: allow the setting of types
3620 based on strings. Remove the 'loc' parameter when adding
3621 attributes because these will be a SET OF encoding which is sorted
3625 *) Initial changes to the 'req' utility to allow request generation
3626 automation. This will allow an application to just generate a template
3627 file containing all the field values and have req construct the
3630 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3631 used all over the place including certificate requests and PKCS#7
3632 structures. They are currently handled manually where necessary with
3633 some primitive wrappers for PKCS#7. The new functions behave in a
3634 manner analogous to the X509 extension functions: they allow
3635 attributes to be looked up by NID and added.
3637 Later something similar to the X509V3 code would be desirable to
3638 automatically handle the encoding, decoding and printing of the
3639 more complex types. The string types like challengePassword can
3640 be handled by the string table functions.
3642 Also modified the multi byte string table handling. Now there is
3643 a 'global mask' which masks out certain types. The table itself
3644 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3645 is useful when for example there is only one permissible type
3646 (as in countryName) and using the mask might result in no valid
3650 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3651 SSL_get_peer_finished to allow applications to obtain the latest
3652 Finished messages sent to the peer or expected from the peer,
3653 respectively. (SSL_get_peer_finished is usually the Finished message
3654 actually received from the peer, otherwise the protocol will be aborted.)
3656 As the Finished message are message digests of the complete handshake
3657 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3658 be used for external authentication procedures when the authentication
3659 provided by SSL/TLS is not desired or is not enough.
3662 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3663 the host supports BWX extension and if Compaq C is present on the
3664 $PATH. Just exploiting of the BWX extension results in 20-30%
3665 performance kick for some algorithms, e.g. DES and RC4 to mention
3666 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3670 *) Add support for MS "fast SGC". This is arguably a violation of the
3671 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3672 weak crypto and after checking the certificate is SGC a second one
3673 with strong crypto. MS SGC stops the first handshake after receiving
3674 the server certificate message and sends a second client hello. Since
3675 a server will typically do all the time consuming operations before
3676 expecting any further messages from the client (server key exchange
3677 is the most expensive) there is little difference between the two.
3679 To get OpenSSL to support MS SGC we have to permit a second client
3680 hello message after we have sent server done. In addition we have to
3681 reset the MAC if we do get this second client hello.
3684 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3685 if a DER encoded private key is RSA or DSA traditional format. Changed
3686 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3687 format DER encoded private key. Newer code should use PKCS#8 format which
3688 has the key type encoded in the ASN1 structure. Added DER private key
3689 support to pkcs8 application.
3692 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3693 ciphersuites has been selected (as required by the SSL 3/TLS 1
3694 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3695 is set, we interpret this as a request to violate the specification
3696 (the worst that can happen is a handshake failure, and 'correct'
3697 behaviour would result in a handshake failure anyway).
3700 *) In SSL_CTX_add_session, take into account that there might be multiple
3701 SSL_SESSION structures with the same session ID (e.g. when two threads
3702 concurrently obtain them from an external cache).
3703 The internal cache can handle only one SSL_SESSION with a given ID,
3704 so if there's a conflict, we now throw out the old one to achieve
3708 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3709 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3710 some routines that use cipher OIDs: some ciphers do not have OIDs
3711 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3715 *) Simplify the trust setting structure and code. Now we just have
3716 two sequences of OIDs for trusted and rejected settings. These will
3717 typically have values the same as the extended key usage extension
3718 and any application specific purposes.
3720 The trust checking code now has a default behaviour: it will just
3721 check for an object with the same NID as the passed id. Functions can
3722 be provided to override either the default behaviour or the behaviour
3723 for a given id. SSL client, server and email already have functions
3724 in place for compatibility: they check the NID and also return "trusted"
3725 if the certificate is self signed.
3728 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3729 traditional format into an EVP_PKEY structure.
3732 *) Add a password callback function PEM_cb() which either prompts for
3733 a password if usr_data is NULL or otherwise assumes it is a null
3734 terminated password. Allow passwords to be passed on command line
3735 environment or config files in a few more utilities.
3738 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3739 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3740 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3741 Update documentation.
3744 *) Support for ASN1 "NULL" type. This could be handled before by using
3745 ASN1_TYPE but there wasn't any function that would try to read a NULL
3746 and produce an error if it couldn't. For compatibility we also have
3747 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3748 don't allocate anything because they don't need to.
3751 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3753 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3755 *) Rebuild of the memory allocation routines used by OpenSSL code and
3756 possibly others as well. The purpose is to make an interface that
3757 provide hooks so anyone can build a separate set of allocation and
3758 deallocation routines to be used by OpenSSL, for example memory
3759 pool implementations, or something else, which was previously hard
3760 since Malloc(), Realloc() and Free() were defined as macros having
3761 the values malloc, realloc and free, respectively (except for Win32
3762 compilations). The same is provided for memory debugging code.
3763 OpenSSL already comes with functionality to find memory leaks, but
3764 this gives people a chance to debug other memory problems.
3766 With these changes, a new set of functions and macros have appeared:
3768 CRYPTO_set_mem_debug_functions() [F]
3769 CRYPTO_get_mem_debug_functions() [F]
3770 CRYPTO_dbg_set_options() [F]
3771 CRYPTO_dbg_get_options() [F]
3772 CRYPTO_malloc_debug_init() [M]
3774 The memory debug functions are NULL by default, unless the library
3775 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3776 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3777 gives the standard debugging functions that come with OpenSSL) or
3778 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3779 provided by the library user) must be used. When the standard
3780 debugging functions are used, CRYPTO_dbg_set_options can be used to
3781 request additional information:
3782 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3783 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3785 Also, things like CRYPTO_set_mem_functions will always give the
3786 expected result (the new set of functions is used for allocation
3787 and deallocation) at all times, regardless of platform and compiler
3790 To finish it up, some functions that were never use in any other
3791 way than through macros have a new API and new semantic:
3794 CRYPTO_dbg_realloc()
3797 All macros of value have retained their old syntax.
3798 [Richard Levitte and Bodo Moeller]
3800 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3801 ordering of SMIMECapabilities wasn't in "strength order" and there
3802 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3806 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3807 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3808 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3810 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3811 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3812 functionality to handle multipart/signed properly) and a utility
3813 called 'smime' to call all this stuff. This is based on code I
3814 originally wrote for Celo who have kindly allowed it to be
3815 included in OpenSSL.
3818 *) Add variants des_set_key_checked and des_set_key_unchecked of
3819 des_set_key (aka des_key_sched). Global variable des_check_key
3820 decides which of these is called by des_set_key; this way
3821 des_check_key behaves as it always did, but applications and
3822 the library itself, which was buggy for des_check_key == 1,
3823 have a cleaner way to pick the version they need.
3826 *) New function PKCS12_newpass() which changes the password of a
3830 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3831 dynamic mix. In both cases the ids can be used as an index into the
3832 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3833 functions so they accept a list of the field values and the
3834 application doesn't need to directly manipulate the X509_TRUST
3838 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3842 *) Modify the way the V3 extension code looks up extensions. This now
3843 works in a similar way to the object code: we have some "standard"
3844 extensions in a static table which is searched with OBJ_bsearch()
3845 and the application can add dynamic ones if needed. The file
3846 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3847 updated whenever a new extension is added to the core code and kept
3848 in ext_nid order. There is a simple program 'tabtest.c' which checks
3849 this. New extensions are not added too often so this file can readily
3850 be maintained manually.
3852 There are two big advantages in doing things this way. The extensions
3853 can be looked up immediately and no longer need to be "added" using
3854 X509V3_add_standard_extensions(): this function now does nothing.
3855 [Side note: I get *lots* of email saying the extension code doesn't
3856 work because people forget to call this function]
3857 Also no dynamic allocation is done unless new extensions are added:
3858 so if we don't add custom extensions there is no need to call
3859 X509V3_EXT_cleanup().
3862 *) Modify enc utility's salting as follows: make salting the default. Add a
3863 magic header, so unsalted files fail gracefully instead of just decrypting
3864 to garbage. This is because not salting is a big security hole, so people
3865 should be discouraged from doing it.
3868 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3869 digest to be passed on the command line but it only used this
3870 parameter when signing a certificate. Modified so all relevant
3871 operations are affected by the digest parameter including the
3872 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3873 DSA key was used because it didn't fix the digest.
3876 *) Initial certificate chain verify code. Currently tests the untrusted
3877 certificates for consistency with the verify purpose (which is set
3878 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3880 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3881 this is because it will reject chains with invalid extensions whereas
3882 every previous version of OpenSSL and SSLeay made no checks at all.
3884 Trust code: checks the root CA for the relevant trust settings. Trust
3885 settings have an initial value consistent with the verify purpose: e.g.
3886 if the verify purpose is for SSL client use it expects the CA to be
3887 trusted for SSL client use. However the default value can be changed to
3888 permit custom trust settings: one example of this would be to only trust
3889 certificates from a specific "secure" set of CAs.
3891 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3892 which should be used for version portability: especially since the
3893 verify structure is likely to change more often now.
3895 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3896 to set them. If not set then assume SSL clients will verify SSL servers
3899 Two new options to the verify program: -untrusted allows a set of
3900 untrusted certificates to be passed in and -purpose which sets the
3901 intended purpose of the certificate. If a purpose is set then the
3902 new chain verify code is used to check extension consistency.
3905 *) Support for the authority information access extension.
3908 *) Modify RSA and DSA PEM read routines to transparently handle
3909 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3910 public keys in a format compatible with certificate
3911 SubjectPublicKeyInfo structures. Unfortunately there were already
3912 functions called *_PublicKey_* which used various odd formats so
3913 these are retained for compatibility: however the DSA variants were
3914 never in a public release so they have been deleted. Changed dsa/rsa
3915 utilities to handle the new format: note no releases ever handled public
3916 keys so we should be OK.
3918 The primary motivation for this change is to avoid the same fiasco
3919 that dogs private keys: there are several incompatible private key
3920 formats some of which are standard and some OpenSSL specific and
3921 require various evil hacks to allow partial transparent handling and
3922 even then it doesn't work with DER formats. Given the option anything
3923 other than PKCS#8 should be dumped: but the other formats have to
3924 stay in the name of compatibility.
3926 With public keys and the benefit of hindsight one standard format
3927 is used which works with EVP_PKEY, RSA or DSA structures: though
3928 it clearly returns an error if you try to read the wrong kind of key.
3930 Added a -pubkey option to the 'x509' utility to output the public key.
3931 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3932 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3933 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3934 that do the same as the EVP_PKEY_assign_*() except they up the
3935 reference count of the added key (they don't "swallow" the
3939 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3940 CRLs would fail if the file contained no certificates or no CRLs:
3941 added a new function to read in both types and return the number
3942 read: this means that if none are read it will be an error. The
3943 DER versions of the certificate and CRL reader would always fail
3944 because it isn't possible to mix certificates and CRLs in DER format
3945 without choking one or the other routine. Changed this to just read
3946 a certificate: this is the best we can do. Also modified the code
3947 in apps/verify.c to take notice of return codes: it was previously
3948 attempting to read in certificates from NULL pointers and ignoring
3949 any errors: this is one reason why the cert and CRL reader seemed
3950 to work. It doesn't check return codes from the default certificate
3951 routines: these may well fail if the certificates aren't installed.
3954 *) Code to support otherName option in GeneralName.
3957 *) First update to verify code. Change the verify utility
3958 so it warns if it is passed a self signed certificate:
3959 for consistency with the normal behaviour. X509_verify
3960 has been modified to it will now verify a self signed
3961 certificate if *exactly* the same certificate appears
3962 in the store: it was previously impossible to trust a
3963 single self signed certificate. This means that:
3964 openssl verify ss.pem
3965 now gives a warning about a self signed certificate but
3966 openssl verify -CAfile ss.pem ss.pem
3970 *) For servers, store verify_result in SSL_SESSION data structure
3971 (and add it to external session representation).
3972 This is needed when client certificate verifications fails,
3973 but an application-provided verification callback (set by
3974 SSL_CTX_set_cert_verify_callback) allows accepting the session
3975 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3976 but returns 1): When the session is reused, we have to set
3977 ssl->verify_result to the appropriate error code to avoid
3979 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3981 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3982 case in PKCS7_dataInit() where the signed PKCS7 structure
3983 didn't contain any existing data because it was being created.
3984 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3986 *) Add a salt to the key derivation routines in enc.c. This
3987 forms the first 8 bytes of the encrypted file. Also add a
3988 -S option to allow a salt to be input on the command line.
3991 *) New function X509_cmp(). Oddly enough there wasn't a function
3992 to compare two certificates. We do this by working out the SHA1
3993 hash and comparing that. X509_cmp() will be needed by the trust
3997 *) SSL_get1_session() is like SSL_get_session(), but increments
3998 the reference count in the SSL_SESSION returned.
3999 [Geoff Thorpe <geoff@eu.c2.net>]
4001 *) Fix for 'req': it was adding a null to request attributes.
4002 Also change the X509_LOOKUP and X509_INFO code to handle
4003 certificate auxiliary information.
4006 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
4010 *) Add the possibility to add extra information to the memory leak
4011 detecting output, to form tracebacks, showing from where each
4012 allocation was originated: CRYPTO_push_info("constant string") adds
4013 the string plus current file name and line number to a per-thread
4014 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
4015 is like calling CYRPTO_pop_info() until the stack is empty.
4016 Also updated memory leak detection code to be multi-thread-safe.
4019 *) Add options -text and -noout to pkcs7 utility and delete the
4020 encryption options which never did anything. Update docs.
4023 *) Add options to some of the utilities to allow the pass phrase
4024 to be included on either the command line (not recommended on
4025 OSes like Unix) or read from the environment. Update the
4026 manpages and fix a few bugs.
4029 *) Add a few manpages for some of the openssl commands.
4032 *) Fix the -revoke option in ca. It was freeing up memory twice,
4033 leaking and not finding already revoked certificates.
4036 *) Extensive changes to support certificate auxiliary information.
4037 This involves the use of X509_CERT_AUX structure and X509_AUX
4038 functions. An X509_AUX function such as PEM_read_X509_AUX()
4039 can still read in a certificate file in the usual way but it
4040 will also read in any additional "auxiliary information". By
4041 doing things this way a fair degree of compatibility can be
4042 retained: existing certificates can have this information added
4043 using the new 'x509' options.
4045 Current auxiliary information includes an "alias" and some trust
4046 settings. The trust settings will ultimately be used in enhanced
4047 certificate chain verification routines: currently a certificate
4048 can only be trusted if it is self signed and then it is trusted
4052 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
4053 The problem was that one of the replacement routines had not been working
4054 since SSLeay releases. For now the offending routine has been replaced
4055 with non-optimised assembler. Even so, this now gives around 95%
4056 performance improvement for 1024 bit RSA signs.
4059 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
4060 handling. Most clients have the effective key size in bits equal to
4061 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
4062 A few however don't do this and instead use the size of the decrypted key
4063 to determine the RC2 key length and the AlgorithmIdentifier to determine
4064 the effective key length. In this case the effective key length can still
4065 be 40 bits but the key length can be 168 bits for example. This is fixed
4066 by manually forcing an RC2 key into the EVP_PKEY structure because the
4067 EVP code can't currently handle unusual RC2 key sizes: it always assumes
4068 the key length and effective key length are equal.
4071 *) Add a bunch of functions that should simplify the creation of
4072 X509_NAME structures. Now you should be able to do:
4073 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
4074 and have it automatically work out the correct field type and fill in
4075 the structures. The more adventurous can try:
4076 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
4077 and it will (hopefully) work out the correct multibyte encoding.
4080 *) Change the 'req' utility to use the new field handling and multibyte
4081 copy routines. Before the DN field creation was handled in an ad hoc
4082 way in req, ca, and x509 which was rather broken and didn't support
4083 BMPStrings or UTF8Strings. Since some software doesn't implement
4084 BMPStrings or UTF8Strings yet, they can be enabled using the config file
4085 using the dirstring_type option. See the new comment in the default
4086 openssl.cnf for more info.
4089 *) Make crypto/rand/md_rand.c more robust:
4090 - Assure unique random numbers after fork().
4091 - Make sure that concurrent threads access the global counter and
4092 md serializably so that we never lose entropy in them
4093 or use exactly the same state in multiple threads.
4094 Access to the large state is not always serializable because
4095 the additional locking could be a performance killer, and
4096 md should be large enough anyway.
4099 *) New file apps/app_rand.c with commonly needed functionality
4100 for handling the random seed file.
4102 Use the random seed file in some applications that previously did not:
4104 dsaparam -genkey (which also ignored its '-rand' option),
4107 x509 (when signing).
4108 Except on systems with /dev/urandom, it is crucial to have a random
4109 seed file at least for key creation, DSA signing, and for DH exchanges;
4110 for RSA signatures we could do without one.
4112 gendh and gendsa (unlike genrsa) used to read only the first byte
4113 of each file listed in the '-rand' option. The function as previously
4114 found in genrsa is now in app_rand.c and is used by all programs
4115 that support '-rand'.
4118 *) In RAND_write_file, use mode 0600 for creating files;
4119 don't just chmod when it may be too late.
4122 *) Report an error from X509_STORE_load_locations
4123 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
4126 *) New function ASN1_mbstring_copy() this copies a string in either
4127 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
4128 into an ASN1_STRING type. A mask of permissible types is passed
4129 and it chooses the "minimal" type to use or an error if not type
4133 *) Add function equivalents to the various macros in asn1.h. The old
4134 macros are retained with an M_ prefix. Code inside the library can
4135 use the M_ macros. External code (including the openssl utility)
4136 should *NOT* in order to be "shared library friendly".
4139 *) Add various functions that can check a certificate's extensions
4140 to see if it usable for various purposes such as SSL client,
4141 server or S/MIME and CAs of these types. This is currently
4142 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4143 verification. Also added a -purpose flag to x509 utility to
4144 print out all the purposes.
4147 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4151 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4152 for, obtain and decode and extension and obtain its critical flag.
4153 This allows all the necessary extension code to be handled in a
4154 single function call.
4157 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4158 platforms. See crypto/rc4/rc4_enc.c for further details.
4161 *) New -noout option to asn1parse. This causes no output to be produced
4162 its main use is when combined with -strparse and -out to extract data
4163 from a file (which may not be in ASN.1 format).
4166 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4167 when producing the local key id.
4168 [Richard Levitte <levitte@stacken.kth.se>]
4170 *) New option -dhparam in s_server. This allows a DH parameter file to be
4171 stated explicitly. If it is not stated then it tries the first server
4172 certificate file. The previous behaviour hard coded the filename
4176 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4177 a public key to be input or output. For example:
4178 openssl rsa -in key.pem -pubout -out pubkey.pem
4179 Also added necessary DSA public key functions to handle this.
4182 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4183 in the message. This was handled by allowing
4184 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4185 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4187 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4188 to the end of the strings whereas this didn't. This would cause problems
4189 if strings read with d2i_ASN1_bytes() were later modified.
4190 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4192 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4193 data and it contains EOF it will end up returning an error. This is
4194 caused by input 46 bytes long. The cause is due to the way base64
4195 BIOs find the start of base64 encoded data. They do this by trying a
4196 trial decode on each line until they find one that works. When they
4197 do a flag is set and it starts again knowing it can pass all the
4198 data directly through the decoder. Unfortunately it doesn't reset
4199 the context it uses. This means that if EOF is reached an attempt
4200 is made to pass two EOFs through the context and this causes the
4201 resulting error. This can also cause other problems as well. As is
4202 usual with these problems it takes *ages* to find and the fix is
4203 trivial: move one line.
4204 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4206 *) Ugly workaround to get s_client and s_server working under Windows. The
4207 old code wouldn't work because it needed to select() on sockets and the
4208 tty (for keypresses and to see if data could be written). Win32 only
4209 supports select() on sockets so we select() with a 1s timeout on the
4210 sockets and then see if any characters are waiting to be read, if none
4211 are present then we retry, we also assume we can always write data to
4212 the tty. This isn't nice because the code then blocks until we've
4213 received a complete line of data and it is effectively polling the
4214 keyboard at 1s intervals: however it's quite a bit better than not
4215 working at all :-) A dedicated Windows application might handle this
4216 with an event loop for example.
4219 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4220 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4221 will be called when RSA_sign() and RSA_verify() are used. This is useful
4222 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4223 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4224 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4225 This necessitated the support of an extra signature type NID_md5_sha1
4226 for SSL signatures and modifications to the SSL library to use it instead
4227 of calling RSA_public_decrypt() and RSA_private_encrypt().
4230 *) Add new -verify -CAfile and -CApath options to the crl program, these
4231 will lookup a CRL issuers certificate and verify the signature in a
4232 similar way to the verify program. Tidy up the crl program so it
4233 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4234 less strict. It will now permit CRL extensions even if it is not
4235 a V2 CRL: this will allow it to tolerate some broken CRLs.
4238 *) Initialize all non-automatic variables each time one of the openssl
4239 sub-programs is started (this is necessary as they may be started
4240 multiple times from the "OpenSSL>" prompt).
4241 [Lennart Bang, Bodo Moeller]
4243 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4244 removing all other RSA functionality (this is what NO_RSA does). This
4245 is so (for example) those in the US can disable those operations covered
4246 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4250 *) Non-copying interface to BIO pairs.
4251 (still largely untested)
4254 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4255 ASCII string. This was handled independently in various places before.
4258 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4259 UTF8 strings a character at a time.
4262 *) Use client_version from client hello to select the protocol
4263 (s23_srvr.c) and for RSA client key exchange verification
4264 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4267 *) Add various utility functions to handle SPKACs, these were previously
4268 handled by poking round in the structure internals. Added new function
4269 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4270 print, verify and generate SPKACs. Based on an original idea from
4271 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4274 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4277 *) Allow the config file extension section to be overwritten on the
4278 command line. Based on an original idea from Massimiliano Pala
4279 <madwolf@comune.modena.it>. The new option is called -extensions
4280 and can be applied to ca, req and x509. Also -reqexts to override
4281 the request extensions in req and -crlexts to override the crl extensions
4285 *) Add new feature to the SPKAC handling in ca. Now you can include
4286 the same field multiple times by preceding it by "XXXX." for example:
4289 this is the same syntax as used in the req config file.
4292 *) Allow certificate extensions to be added to certificate requests. These
4293 are specified in a 'req_extensions' option of the req section of the
4294 config file. They can be printed out with the -text option to req but
4295 are otherwise ignored at present.
4298 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4299 data read consists of only the final block it would not decrypted because
4300 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4301 A misplaced 'break' also meant the decrypted final block might not be
4302 copied until the next read.
4305 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4306 a few extra parameters to the DH structure: these will be useful if
4307 for example we want the value of 'q' or implement X9.42 DH.
4310 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4311 provides hooks that allow the default DSA functions or functions on a
4312 "per key" basis to be replaced. This allows hardware acceleration and
4313 hardware key storage to be handled without major modification to the
4314 library. Also added low level modexp hooks and CRYPTO_EX structure and
4315 associated functions.
4318 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4319 as "read only": it can't be written to and the buffer it points to will
4320 not be freed. Reading from a read only BIO is much more efficient than
4321 a normal memory BIO. This was added because there are several times when
4322 an area of memory needs to be read from a BIO. The previous method was
4323 to create a memory BIO and write the data to it, this results in two
4324 copies of the data and an O(n^2) reading algorithm. There is a new
4325 function BIO_new_mem_buf() which creates a read only memory BIO from
4326 an area of memory. Also modified the PKCS#7 routines to use read only
4330 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4331 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4332 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4333 but a retry condition occured while trying to read the rest.
4336 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4337 NID_pkcs7_encrypted by default: this was wrong since this should almost
4338 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4339 the encrypted data type: this is a more sensible place to put it and it
4340 allows the PKCS#12 code to be tidied up that duplicated this
4344 *) Changed obj_dat.pl script so it takes its input and output files on
4345 the command line. This should avoid shell escape redirection problems
4349 *) Initial support for certificate extension requests, these are included
4350 in things like Xenroll certificate requests. Included functions to allow
4351 extensions to be obtained and added.
4354 *) -crlf option to s_client and s_server for sending newlines as
4355 CRLF (as required by many protocols).
4358 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4360 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4361 [Ralf S. Engelschall]
4363 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4364 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4366 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4370 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4371 DH parameters/keys (q is lost during that conversion, but the resulting
4372 DH parameters contain its length).
4374 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4375 much faster than DH_generate_parameters (which creates parameters
4376 where p = 2*q + 1), and also the smaller q makes DH computations
4377 much more efficient (160-bit exponentiation instead of 1024-bit
4378 exponentiation); so this provides a convenient way to support DHE
4379 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4380 utter importance to use
4381 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4383 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4384 when such DH parameters are used, because otherwise small subgroup
4385 attacks may become possible!
4388 *) Avoid memory leak in i2d_DHparams.
4391 *) Allow the -k option to be used more than once in the enc program:
4392 this allows the same encrypted message to be read by multiple recipients.
4395 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4396 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4397 it will always use the numerical form of the OID, even if it has a short
4401 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4402 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4403 otherwise bn_mod_exp was called. In the case of hardware keys for example
4404 no private key components need be present and it might store extra data
4405 in the RSA structure, which cannot be accessed from bn_mod_exp.
4406 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4407 private key operations.
4410 *) Added support for SPARC Linux.
4413 *) pem_password_cb function type incompatibly changed from
4414 typedef int pem_password_cb(char *buf, int size, int rwflag);
4416 ....(char *buf, int size, int rwflag, void *userdata);
4417 so that applications can pass data to their callbacks:
4418 The PEM[_ASN1]_{read,write}... functions and macros now take an
4419 additional void * argument, which is just handed through whenever
4420 the password callback is called.
4421 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4423 New function SSL_CTX_set_default_passwd_cb_userdata.
4425 Compatibility note: As many C implementations push function arguments
4426 onto the stack in reverse order, the new library version is likely to
4427 interoperate with programs that have been compiled with the old
4428 pem_password_cb definition (PEM_whatever takes some data that
4429 happens to be on the stack as its last argument, and the callback
4430 just ignores this garbage); but there is no guarantee whatsoever that
4433 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4434 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4435 problems not only on Windows, but also on some Unix platforms.
4436 To avoid problematic command lines, these definitions are now in an
4437 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4438 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4441 *) MIPS III/IV assembler module is reimplemented.
4444 *) More DES library cleanups: remove references to srand/rand and
4445 delete an unused file.
4448 *) Add support for the the free Netwide assembler (NASM) under Win32,
4449 since not many people have MASM (ml) and it can be hard to obtain.
4450 This is currently experimental but it seems to work OK and pass all
4451 the tests. Check out INSTALL.W32 for info.
4454 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4455 without temporary keys kept an extra copy of the server key,
4456 and connections with temporary keys did not free everything in case
4460 *) New function RSA_check_key and new openssl rsa option -check
4461 for verifying the consistency of RSA keys.
4462 [Ulf Moeller, Bodo Moeller]
4464 *) Various changes to make Win32 compile work:
4465 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4466 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4467 comparison" warnings.
4468 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4471 *) Add a debugging option to PKCS#5 v2 key generation function: when
4472 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4473 derived keys are printed to stderr.
4476 *) Copy the flags in ASN1_STRING_dup().
4477 [Roman E. Pavlov <pre@mo.msk.ru>]
4479 *) The x509 application mishandled signing requests containing DSA
4480 keys when the signing key was also DSA and the parameters didn't match.
4482 It was supposed to omit the parameters when they matched the signing key:
4483 the verifying software was then supposed to automatically use the CA's
4484 parameters if they were absent from the end user certificate.
4486 Omitting parameters is no longer recommended. The test was also
4487 the wrong way round! This was probably due to unusual behaviour in
4488 EVP_cmp_parameters() which returns 1 if the parameters match.
4489 This meant that parameters were omitted when they *didn't* match and
4490 the certificate was useless. Certificates signed with 'ca' didn't have
4492 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4494 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4495 The interface is as follows:
4496 Applications can use
4497 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4498 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4499 "off" is now the default.
4500 The library internally uses
4501 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4502 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4503 to disable memory-checking temporarily.
4505 Some inconsistent states that previously were possible (and were
4506 even the default) are now avoided.
4508 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4509 with each memory chunk allocated; this is occasionally more helpful
4510 than just having a counter.
4512 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4514 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4518 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4519 which largely parallels "options", but is for changing API behaviour,
4520 whereas "options" are about protocol behaviour.
4521 Initial "mode" flags are:
4523 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4524 a single record has been written.
4525 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4526 retries use the same buffer location.
4527 (But all of the contents must be
4531 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4534 *) Fix problems with no-hmac etc.
4535 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4537 *) New functions RSA_get_default_method(), RSA_set_method() and
4538 RSA_get_method(). These allows replacement of RSA_METHODs without having
4539 to mess around with the internals of an RSA structure.
4542 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4543 Also really enable memory leak checks in openssl.c and in some
4545 [Chad C. Mulligan, Bodo Moeller]
4547 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4548 up the length of negative integers. This has now been simplified to just
4549 store the length when it is first determined and use it later, rather
4550 than trying to keep track of where data is copied and updating it to
4552 [Steve Henson, reported by Brien Wheeler
4553 <bwheeler@authentica-security.com>]
4555 *) Add a new function PKCS7_signatureVerify. This allows the verification
4556 of a PKCS#7 signature but with the signing certificate passed to the
4557 function itself. This contrasts with PKCS7_dataVerify which assumes the
4558 certificate is present in the PKCS#7 structure. This isn't always the
4559 case: certificates can be omitted from a PKCS#7 structure and be
4560 distributed by "out of band" means (such as a certificate database).
4563 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4564 function prototypes in pem.h, also change util/mkdef.pl to add the
4565 necessary function names.
4568 *) mk1mf.pl (used by Windows builds) did not properly read the
4569 options set by Configure in the top level Makefile, and Configure
4570 was not even able to write more than one option correctly.
4571 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4574 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4575 file to be loaded from a BIO or FILE pointer. The BIO version will
4576 for example allow memory BIOs to contain config info.
4579 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4580 Whoever hopes to achieve shared-library compatibility across versions
4581 must use this, not the compile-time macro.
4582 (Exercise 0.9.4: Which is the minimum library version required by
4584 Note: All this applies only to multi-threaded programs, others don't
4588 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4589 through a BIO pair triggered the default case, i.e.
4590 SSLerr(...,SSL_R_UNKNOWN_STATE).
4593 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4594 can use the SSL library even if none of the specific BIOs is
4598 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4599 for the encoded length.
4600 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4602 *) Add initial documentation of the X509V3 functions.
4605 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4606 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4607 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4608 secure PKCS#8 private key format with a high iteration count.
4611 *) Fix determination of Perl interpreter: A perl or perl5
4612 _directory_ in $PATH was also accepted as the interpreter.
4613 [Ralf S. Engelschall]
4615 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4616 wrong with it but it was very old and did things like calling
4617 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4621 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4622 to use the new extension code.
4625 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4626 with macros. This should make it easier to change their form, add extra
4627 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4631 *) Add to configuration table a new entry that can specify an alternative
4632 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4633 according to Mark Crispin <MRC@Panda.COM>.
4637 *) DES CBC did not update the IV. Weird.
4640 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4641 Changing the behaviour of the former might break existing programs --
4642 where IV updating is needed, des_ncbc_encrypt can be used.
4645 *) When bntest is run from "make test" it drives bc to check its
4646 calculations, as well as internally checking them. If an internal check
4647 fails, it needs to cause bc to give a non-zero result or make test carries
4648 on without noticing the failure. Fixed.
4651 *) DES library cleanups.
4654 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4655 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4656 ciphers. NOTE: although the key derivation function has been verified
4657 against some published test vectors it has not been extensively tested
4658 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4662 *) Instead of "mkdir -p", which is not fully portable, use new
4663 Perl script "util/mkdir-p.pl".
4666 *) Rewrite the way password based encryption (PBE) is handled. It used to
4667 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4668 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4669 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4670 the 'parameter' field of the AlgorithmIdentifier is passed to the
4671 underlying key generation function so it must do its own ASN1 parsing.
4672 This has also changed the EVP_PBE_CipherInit() function which now has a
4673 'parameter' argument instead of literal salt and iteration count values
4674 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4677 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4678 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4679 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4680 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4681 value was just used as a "magic string" and not used directly its
4682 value doesn't matter.
4685 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4689 *) "linux-sparc64" configuration (ultrapenguin).
4690 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4691 "linux-sparc" configuration.
4692 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4694 *) config now generates no-xxx options for missing ciphers.
4697 *) Support the EBCDIC character set (work in progress).
4698 File ebcdic.c not yet included because it has a different license.
4699 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4701 *) Support BS2000/OSD-POSIX.
4702 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4704 *) Make callbacks for key generation use void * instead of char *.
4707 *) Make S/MIME samples compile (not yet tested).
4710 *) Additional typesafe stacks.
4713 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4717 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4719 *) New configuration variant "sco5-gcc".
4721 *) Updated some demos.
4722 [Sean O Riordain, Wade Scholine]
4724 *) Add missing BIO_free at exit of pkcs12 application.
4727 *) Fix memory leak in conf.c.
4730 *) Updates for Win32 to assembler version of MD5.
4733 *) Set #! path to perl in apps/der_chop to where we found it
4734 instead of using a fixed path.
4737 *) SHA library changes for irix64-mips4-cc.
4740 *) Improvements for VMS support.
4744 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4746 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4747 This also avoids the problems with SC4.2 and unpatched SC5.
4748 [Andy Polyakov <appro@fy.chalmers.se>]
4750 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4751 These are required because of the typesafe stack would otherwise break
4752 existing code. If old code used a structure member which used to be STACK
4753 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4754 sk_num or sk_value it would produce an error because the num, data members
4755 are not present in STACK_OF. Now it just produces a warning. sk_set
4756 replaces the old method of assigning a value to sk_value
4757 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4758 that does this will no longer work (and should use sk_set instead) but
4759 this could be regarded as a "questionable" behaviour anyway.
4762 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4763 correctly handle encrypted S/MIME data.
4766 *) Change type of various DES function arguments from des_cblock
4767 (which means, in function argument declarations, pointer to char)
4768 to des_cblock * (meaning pointer to array with 8 char elements),
4769 which allows the compiler to do more typechecking; it was like
4770 that back in SSLeay, but with lots of ugly casts.
4772 Introduce new type const_des_cblock.
4775 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4776 problems: find RecipientInfo structure that matches recipient certificate
4777 and initialise the ASN1 structures properly based on passed cipher.
4780 *) Belatedly make the BN tests actually check the results.
4783 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4784 to and from BNs: it was completely broken. New compilation option
4785 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4786 key elements as negative integers.
4789 *) Reorganize and speed up MD5.
4790 [Andy Polyakov <appro@fy.chalmers.se>]
4793 [Richard Levitte <richard@levitte.org>]
4795 *) New option -out to asn1parse to allow the parsed structure to be
4796 output to a file. This is most useful when combined with the -strparse
4797 option to examine the output of things like OCTET STRINGS.
4800 *) Make SSL library a little more fool-proof by not requiring any longer
4801 that SSL_set_{accept,connect}_state be called before
4802 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4803 in many applications because usually everything *appeared* to work as
4804 intended anyway -- now it really works as intended).
4807 *) Move openssl.cnf out of lib/.
4810 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4811 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4812 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4813 [Ralf S. Engelschall]
4815 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4816 handle PKCS#7 enveloped data properly.
4817 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4819 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4820 copying pointers. The cert_st handling is changed by this in
4821 various ways (and thus what used to be known as ctx->default_cert
4822 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4823 any longer when s->cert does not give us what we need).
4824 ssl_cert_instantiate becomes obsolete by this change.
4825 As soon as we've got the new code right (possibly it already is?),
4826 we have solved a couple of bugs of the earlier code where s->cert
4827 was used as if it could not have been shared with other SSL structures.
4829 Note that using the SSL API in certain dirty ways now will result
4830 in different behaviour than observed with earlier library versions:
4831 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4832 does not influence s as it used to.
4834 In order to clean up things more thoroughly, inside SSL_SESSION
4835 we don't use CERT any longer, but a new structure SESS_CERT
4836 that holds per-session data (if available); currently, this is
4837 the peer's certificate chain and, for clients, the server's certificate
4838 and temporary key. CERT holds only those values that can have
4839 meaningful defaults in an SSL_CTX.
4842 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4843 from the internal representation. Various PKCS#7 fixes: remove some
4844 evil casts and set the enc_dig_alg field properly based on the signing
4848 *) Allow PKCS#12 password to be set from the command line or the
4849 environment. Let 'ca' get its config file name from the environment
4850 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4854 *) Allow certificate policies extension to use an IA5STRING for the
4855 organization field. This is contrary to the PKIX definition but
4856 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4860 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4861 without disallowing inline assembler and the like for non-pedantic builds.
4864 *) Support Borland C++ builder.
4865 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4870 *) SHA-1 cleanups and performance enhancements.
4871 [Andy Polyakov <appro@fy.chalmers.se>]
4873 *) Sparc v8plus assembler for the bignum library.
4874 [Andy Polyakov <appro@fy.chalmers.se>]
4876 *) Accept any -xxx and +xxx compiler options in Configure.
4879 *) Update HPUX configuration.
4882 *) Add missing sk_<type>_unshift() function to safestack.h
4883 [Ralf S. Engelschall]
4885 *) New function SSL_CTX_use_certificate_chain_file that sets the
4886 "extra_cert"s in addition to the certificate. (This makes sense
4887 only for "PEM" format files, as chains as a whole are not
4891 *) Support verify_depth from the SSL API.
4892 x509_vfy.c had what can be considered an off-by-one-error:
4893 Its depth (which was not part of the external interface)
4894 was actually counting the number of certificates in a chain;
4895 now it really counts the depth.
4898 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4899 instead of X509err, which often resulted in confusing error
4900 messages since the error codes are not globally unique
4901 (e.g. an alleged error in ssl3_accept when a certificate
4902 didn't match the private key).
4904 *) New function SSL_CTX_set_session_id_context that allows to set a default
4905 value (so that you don't need SSL_set_session_id_context for each
4906 connection using the SSL_CTX).
4909 *) OAEP decoding bug fix.
4912 *) Support INSTALL_PREFIX for package builders, as proposed by
4916 *) New Configure options "threads" and "no-threads". For systems
4917 where the proper compiler options are known (currently Solaris
4918 and Linux), "threads" is the default.
4921 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4924 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4925 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4926 such as /usr/local/bin.
4929 *) "make linux-shared" to build shared libraries.
4930 [Niels Poppe <niels@netbox.org>]
4932 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4935 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4936 extension adding in x509 utility.
4939 *) Remove NOPROTO sections and error code comments.
4942 *) Partial rewrite of the DEF file generator to now parse the ANSI
4946 *) New Configure options --prefix=DIR and --openssldir=DIR.
4949 *) Complete rewrite of the error code script(s). It is all now handled
4950 by one script at the top level which handles error code gathering,
4951 header rewriting and C source file generation. It should be much better
4952 than the old method: it now uses a modified version of Ulf's parser to
4953 read the ANSI prototypes in all header files (thus the old K&R definitions
4954 aren't needed for error creation any more) and do a better job of
4955 translating function codes into names. The old 'ASN1 error code imbedded
4956 in a comment' is no longer necessary and it doesn't use .err files which
4957 have now been deleted. Also the error code call doesn't have to appear all
4958 on one line (which resulted in some large lines...).
4961 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4964 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4965 0 (which usually indicates a closed connection), but continue reading.
4968 *) Fix some race conditions.
4971 *) Add support for CRL distribution points extension. Add Certificate
4972 Policies and CRL distribution points documentation.
4975 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4978 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4979 8 of keying material. Merlin has also confirmed interop with this fix
4980 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4981 [Merlin Hughes <merlin@baltimore.ie>]
4983 *) Fix lots of warnings.
4984 [Richard Levitte <levitte@stacken.kth.se>]
4986 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4987 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4988 [Richard Levitte <levitte@stacken.kth.se>]
4990 *) Fix problems with sizeof(long) == 8.
4991 [Andy Polyakov <appro@fy.chalmers.se>]
4993 *) Change functions to ANSI C.
4996 *) Fix typos in error codes.
4997 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4999 *) Remove defunct assembler files from Configure.
5002 *) SPARC v8 assembler BIGNUM implementation.
5003 [Andy Polyakov <appro@fy.chalmers.se>]
5005 *) Support for Certificate Policies extension: both print and set.
5006 Various additions to support the r2i method this uses.
5009 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
5010 return a const string when you are expecting an allocated buffer.
5013 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
5014 types DirectoryString and DisplayText.
5017 *) Add code to allow r2i extensions to access the configuration database,
5018 add an LHASH database driver and add several ctx helper functions.
5021 *) Fix an evil bug in bn_expand2() which caused various BN functions to
5022 fail when they extended the size of a BIGNUM.
5025 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
5026 support typesafe stack.
5029 *) Fix typo in SSL_[gs]et_options().
5030 [Nils Frostberg <nils@medcom.se>]
5032 *) Delete various functions and files that belonged to the (now obsolete)
5033 old X509V3 handling code.
5036 *) New Configure option "rsaref".
5039 *) Don't auto-generate pem.h.
5042 *) Introduce type-safe ASN.1 SETs.
5045 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
5046 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
5048 *) Introduce type-safe STACKs. This will almost certainly break lots of code
5049 that links with OpenSSL (well at least cause lots of warnings), but fear
5050 not: the conversion is trivial, and it eliminates loads of evil casts. A
5051 few STACKed things have been converted already. Feel free to convert more.
5052 In the fullness of time, I'll do away with the STACK type altogether.
5055 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
5056 specified in <certfile> by updating the entry in the index.txt file.
5057 This way one no longer has to edit the index.txt file manually for
5058 revoking a certificate. The -revoke option does the gory details now.
5059 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
5061 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
5062 `-text' option at all and this way the `-noout -text' combination was
5063 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
5064 [Ralf S. Engelschall]
5066 *) Make sure a corresponding plain text error message exists for the
5067 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
5068 verify callback function determined that a certificate was revoked.
5069 [Ralf S. Engelschall]
5071 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
5072 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
5073 all available cipers including rc5, which was forgotten until now.
5074 In order to let the testing shell script know which algorithms
5075 are available, a new (up to now undocumented) command
5076 "openssl list-cipher-commands" is used.
5079 *) Bugfix: s_client occasionally would sleep in select() when
5080 it should have checked SSL_pending() first.
5083 *) New functions DSA_do_sign and DSA_do_verify to provide access to
5084 the raw DSA values prior to ASN.1 encoding.
5087 *) Tweaks to Configure
5088 [Niels Poppe <niels@netbox.org>]
5090 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
5094 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
5097 *) New config option to avoid instructions that are illegal on the 80386.
5098 The default code is faster, but requires at least a 486.
5101 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
5102 SSL2_SERVER_VERSION (not used at all) macros, which are now the
5103 same as SSL2_VERSION anyway.
5106 *) New "-showcerts" option for s_client.
5109 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
5110 application. Various cleanups and fixes.
5113 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
5114 modify error routines to work internally. Add error codes and PBE init
5115 to library startup routines.
5118 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
5119 packing functions to asn1 and evp. Changed function names and error
5120 codes along the way.
5123 *) PKCS12 integration: and so it begins... First of several patches to
5124 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
5125 objects to objects.h
5128 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
5129 and display support for Thawte strong extranet extension.
5132 *) Add LinuxPPC support.
5133 [Jeff Dubrule <igor@pobox.org>]
5135 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5136 bn_div_words in alpha.s.
5137 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5139 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5140 OAEP isn't supported when OpenSSL is built with RSAref.
5141 [Ulf Moeller <ulf@fitug.de>]
5143 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5144 so they no longer are missing under -DNOPROTO.
5145 [Soren S. Jorvang <soren@t.dk>]
5148 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5150 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5151 doesn't work when the session is reused. Coming soon!
5154 *) Fix a security hole, that allows sessions to be reused in the wrong
5155 context thus bypassing client cert protection! All software that uses
5156 client certs and session caches in multiple contexts NEEDS PATCHING to
5157 allow session reuse! A fuller solution is in the works.
5158 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5160 *) Some more source tree cleanups (removed obsolete files
5161 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5162 permission on "config" script to be executable) and a fix for the INSTALL
5164 [Ulf Moeller <ulf@fitug.de>]
5166 *) Remove some legacy and erroneous uses of malloc, free instead of
5168 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5170 *) Make rsa_oaep_test return non-zero on error.
5171 [Ulf Moeller <ulf@fitug.de>]
5173 *) Add support for native Solaris shared libraries. Configure
5174 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5175 if someone would make that last step automatic.
5176 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5178 *) ctx_size was not built with the right compiler during "make links". Fixed.
5181 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5182 except NULL ciphers". This means the default cipher list will no longer
5183 enable NULL ciphers. They need to be specifically enabled e.g. with
5184 the string "DEFAULT:eNULL".
5187 *) Fix to RSA private encryption routines: if p < q then it would
5188 occasionally produce an invalid result. This will only happen with
5189 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5192 *) Be less restrictive and allow also `perl util/perlpath.pl
5193 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5194 because this way one can also use an interpreter named `perl5' (which is
5195 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5196 installed as `perl').
5197 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5199 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5200 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5202 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5203 advapi32.lib to Win32 build and change the pem test comparision
5204 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5205 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5206 and crypto/des/ede_cbcm_enc.c.
5209 *) DES quad checksum was broken on big-endian architectures. Fixed.
5212 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5213 Win32 test batch file so it (might) work again. The Win32 test batch file
5214 is horrible: I feel ill....
5217 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5218 in e_os.h. Audit of header files to check ANSI and non ANSI
5219 sections: 10 functions were absent from non ANSI section and not exported
5220 from Windows DLLs. Fixed up libeay.num for new functions.
5223 *) Make `openssl version' output lines consistent.
5224 [Ralf S. Engelschall]
5226 *) Fix Win32 symbol export lists for BIO functions: Added
5227 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5228 to ms/libeay{16,32}.def.
5229 [Ralf S. Engelschall]
5231 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5232 fine under Unix and passes some trivial tests I've now added. But the
5233 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5234 added to make sure no one expects that this stuff really works in the
5235 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5236 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5238 [Ralf S. Engelschall]
5240 *) Fix the generation of two part addresses in perl.
5241 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5243 *) Add config entry for Linux on MIPS.
5244 [John Tobey <jtobey@channel1.com>]
5246 *) Make links whenever Configure is run, unless we are on Windoze.
5249 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5250 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5254 *) Add a useful kludge to allow package maintainers to specify compiler and
5255 other platforms details on the command line without having to patch the
5256 Configure script everytime: One now can use ``perl Configure
5257 <id>:<details>'', i.e. platform ids are allowed to have details appended
5258 to them (seperated by colons). This is treated as there would be a static
5259 pre-configured entry in Configure's %table under key <id> with value
5260 <details> and ``perl Configure <id>'' is called. So, when you want to
5261 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5262 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5263 now, which overrides the FreeBSD-elf entry on-the-fly.
5264 [Ralf S. Engelschall]
5266 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5269 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5270 on the `perl Configure ...' command line. This way one can compile
5271 OpenSSL libraries with Position Independent Code (PIC) which is needed
5272 for linking it into DSOs.
5273 [Ralf S. Engelschall]
5275 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5279 *) Cleaned up the LICENSE document: The official contact for any license
5280 questions now is the OpenSSL core team under openssl-core@openssl.org.
5281 And add a paragraph about the dual-license situation to make sure people
5282 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5283 to the OpenSSL toolkit.
5284 [Ralf S. Engelschall]
5286 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5287 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5288 Additonally cleaned up the `make links' target: Remove unnecessary
5289 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5290 to speed processing and no longer clutter the display with confusing
5291 stuff. Instead only the actually done links are displayed.
5292 [Ralf S. Engelschall]
5294 *) Permit null encryption ciphersuites, used for authentication only. It used
5295 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5296 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5300 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5301 signed attributes when verifying signatures (this would break them),
5302 the detached data encoding was wrong and public keys obtained using
5303 X509_get_pubkey() weren't freed.
5306 *) Add text documentation for the BUFFER functions. Also added a work around
5307 to a Win95 console bug. This was triggered by the password read stuff: the
5308 last character typed gets carried over to the next fread(). If you were
5309 generating a new cert request using 'req' for example then the last
5310 character of the passphrase would be CR which would then enter the first
5314 *) Added the new `Includes OpenSSL Cryptography Software' button as
5315 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5316 button and can be used by applications based on OpenSSL to show the
5317 relationship to the OpenSSL project.
5318 [Ralf S. Engelschall]
5320 *) Remove confusing variables in function signatures in files
5321 ssl/ssl_lib.c and ssl/ssl.h.
5322 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5324 *) Don't install bss_file.c under PREFIX/include/
5325 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5327 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5328 functions that return function pointers and has support for NT specific
5329 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5330 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5331 unsigned to signed types: this was killing the Win32 compile.
5334 *) Add new certificate file to stack functions,
5335 SSL_add_dir_cert_subjects_to_stack() and
5336 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5337 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5338 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5339 This means that Apache-SSL and similar packages don't have to mess around
5340 to add as many CAs as they want to the preferred list.
5343 *) Experiment with doxygen documentation. Currently only partially applied to
5345 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5346 openssl.doxy as the configuration file.
5349 *) Get rid of remaining C++-style comments which strict C compilers hate.
5350 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5352 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5353 compiled in by default: it has problems with large keys.
5356 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5357 DH private keys and/or callback functions which directly correspond to
5358 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5359 is needed for applications which have to configure certificates on a
5360 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5362 For the RSA certificate situation is makes no difference, but
5363 for the DSA certificate situation this fixes the "no shared cipher"
5364 problem where the OpenSSL cipher selection procedure failed because the
5365 temporary keys were not overtaken from the context and the API provided
5366 no way to reconfigure them.
5367 The new functions now let applications reconfigure the stuff and they
5368 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5369 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5370 non-public-API function ssl_cert_instantiate() is used as a helper
5371 function and also to reduce code redundancy inside ssl_rsa.c.
5372 [Ralf S. Engelschall]
5374 *) Move s_server -dcert and -dkey options out of the undocumented feature
5375 area because they are useful for the DSA situation and should be
5376 recognized by the users.
5377 [Ralf S. Engelschall]
5379 *) Fix the cipher decision scheme for export ciphers: the export bits are
5380 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5381 SSL_EXP_MASK. So, the original variable has to be used instead of the
5382 already masked variable.
5383 [Richard Levitte <levitte@stacken.kth.se>]
5385 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5386 [Richard Levitte <levitte@stacken.kth.se>]
5388 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5389 from `int' to `unsigned int' because it's a length and initialized by
5390 EVP_DigestFinal() which expects an `unsigned int *'.
5391 [Richard Levitte <levitte@stacken.kth.se>]
5393 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5394 script. Instead use the usual Shell->Perl transition trick.
5395 [Ralf S. Engelschall]
5397 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5398 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5399 -noout -modulus' as it's already the case for `openssl rsa -noout
5400 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5401 currently the public key is printed (a decision which was already done by
5402 `openssl dsa -modulus' in the past) which serves a similar purpose.
5403 Additionally the NO_RSA no longer completely removes the whole -modulus
5404 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5406 [Ralf S. Engelschall]
5408 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5409 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5410 [Arne Ansper <arne@ats.cyber.ee>]
5412 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5413 to be added. Now both 'req' and 'ca' can use new objects defined in the
5417 *) Add cool BIO that does syslog (or event log on NT).
5418 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5420 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5421 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5422 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5423 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5426 *) Add preliminary config info for new extension code.
5429 *) Make RSA_NO_PADDING really use no padding.
5430 [Ulf Moeller <ulf@fitug.de>]
5432 *) Generate errors when private/public key check is done.
5435 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5436 for some CRL extensions and new objects added.
5439 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5440 key usage extension and fuller support for authority key id.
5443 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5444 padding method for RSA, which is recommended for new applications in PKCS
5445 #1 v2.0 (RFC 2437, October 1998).
5446 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5447 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5448 against Bleichbacher's attack on RSA.
5449 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5452 *) Updates to the new SSL compression code
5453 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5455 *) Fix so that the version number in the master secret, when passed
5456 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5457 (because the server will not accept higher), that the version number
5458 is 0x03,0x01, not 0x03,0x00
5459 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5461 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5462 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5463 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5466 *) Support for RAW extensions where an arbitrary extension can be
5467 created by including its DER encoding. See apps/openssl.cnf for
5471 *) Make sure latest Perl versions don't interpret some generated C array
5472 code as Perl array code in the crypto/err/err_genc.pl script.
5473 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5475 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5476 not many people have the assembler. Various Win32 compilation fixes and
5477 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5481 *) Modify configure script 'Configure' to automatically create crypto/date.h
5482 file under Win32 and also build pem.h from pem.org. New script
5483 util/mkfiles.pl to create the MINFO file on environments that can't do a
5484 'make files': perl util/mkfiles.pl >MINFO should work.
5487 *) Major rework of DES function declarations, in the pursuit of correctness
5488 and purity. As a result, many evil casts evaporated, and some weirdness,
5489 too. You may find this causes warnings in your code. Zapping your evil
5490 casts will probably fix them. Mostly.
5493 *) Fix for a typo in asn1.h. Bug fix to object creation script
5494 obj_dat.pl. It considered a zero in an object definition to mean
5495 "end of object": none of the objects in objects.h have any zeros
5496 so it wasn't spotted.
5497 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5499 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5500 Masking (CBCM). In the absence of test vectors, the best I have been able
5501 to do is check that the decrypt undoes the encrypt, so far. Send me test
5502 vectors if you have them.
5505 *) Correct calculation of key length for export ciphers (too much space was
5506 allocated for null ciphers). This has not been tested!
5509 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5510 message is now correct (it understands "crypto" and "ssl" on its
5511 command line). There is also now an "update" option. This will update
5512 the util/ssleay.num and util/libeay.num files with any new functions.
5514 perl util/mkdef.pl crypto ssl update
5515 it will update them.
5518 *) Overhauled the Perl interface (perl/*):
5519 - ported BN stuff to OpenSSL's different BN library
5520 - made the perl/ source tree CVS-aware
5521 - renamed the package from SSLeay to OpenSSL (the files still contain
5522 their history because I've copied them in the repository)
5523 - removed obsolete files (the test scripts will be replaced
5524 by better Test::Harness variants in the future)
5525 [Ralf S. Engelschall]
5527 *) First cut for a very conservative source tree cleanup:
5528 1. merge various obsolete readme texts into doc/ssleay.txt
5529 where we collect the old documents and readme texts.
5530 2. remove the first part of files where I'm already sure that we no
5531 longer need them because of three reasons: either they are just temporary
5532 files which were left by Eric or they are preserved original files where
5533 I've verified that the diff is also available in the CVS via "cvs diff
5534 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5535 the crypto/md/ stuff).
5536 [Ralf S. Engelschall]
5538 *) More extension code. Incomplete support for subject and issuer alt
5539 name, issuer and authority key id. Change the i2v function parameters
5540 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5541 what that's for :-) Fix to ASN1 macro which messed up
5542 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5545 *) Preliminary support for ENUMERATED type. This is largely copied from the
5549 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5550 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5552 *) Make sure `make rehash' target really finds the `openssl' program.
5553 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5555 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5556 like to hear about it if this slows down other processors.
5559 *) Add CygWin32 platform information to Configure script.
5560 [Alan Batie <batie@aahz.jf.intel.com>]
5562 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5563 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5565 *) New program nseq to manipulate netscape certificate sequences
5568 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5572 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5573 but the BN code had some problems that would cause failures when
5574 doing certificate verification and some other functions.
5575 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5577 *) Add ASN1 and PEM code to support netscape certificate sequences.
5580 *) Add ASN1 and PEM code to support netscape certificate sequences.
5583 *) Add several PKIX and private extended key usage OIDs.
5586 *) Modify the 'ca' program to handle the new extension code. Modify
5587 openssl.cnf for new extension format, add comments.
5590 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5591 and add a sample to openssl.cnf so req -x509 now adds appropriate
5595 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5596 error code, add initial support to X509_print() and x509 application.
5599 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5600 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5601 stuff is currently isolated and isn't even compiled yet.
5604 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5605 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5606 Removed the versions check from X509 routines when loading extensions:
5607 this allows certain broken certificates that don't set the version
5608 properly to be processed.
5611 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5612 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5613 can still be regenerated with "make depend".
5616 *) Spelling mistake in C version of CAST-128.
5617 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5619 *) Changes to the error generation code. The perl script err-code.pl
5620 now reads in the old error codes and retains the old numbers, only
5621 adding new ones if necessary. It also only changes the .err files if new
5622 codes are added. The makefiles have been modified to only insert errors
5623 when needed (to avoid needlessly modifying header files). This is done
5624 by only inserting errors if the .err file is newer than the auto generated
5625 C file. To rebuild all the error codes from scratch (the old behaviour)
5626 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5627 or delete all the .err files.
5630 *) CAST-128 was incorrectly implemented for short keys. The C version has
5631 been fixed, but is untested. The assembler versions are also fixed, but
5632 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5633 to regenerate it if needed.
5634 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5635 Hagino <itojun@kame.net>]
5637 *) File was opened incorrectly in randfile.c.
5638 [Ulf Möller <ulf@fitug.de>]
5640 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5641 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5642 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5643 al: it's just almost always a UTCTime. Note this patch adds new error
5644 codes so do a "make errors" if there are problems.
5647 *) Correct Linux 1 recognition in config.
5648 [Ulf Möller <ulf@fitug.de>]
5650 *) Remove pointless MD5 hash when using DSA keys in ca.
5651 [Anonymous <nobody@replay.com>]
5653 *) Generate an error if given an empty string as a cert directory. Also
5654 generate an error if handed NULL (previously returned 0 to indicate an
5655 error, but didn't set one).
5656 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5658 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5661 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5662 parameters. This was causing a warning which killed off the Win32 compile.
5665 *) Remove C++ style comments from crypto/bn/bn_local.h.
5666 [Neil Costigan <neil.costigan@celocom.com>]
5668 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5669 based on a text string, looking up short and long names and finally
5670 "dot" format. The "dot" format stuff didn't work. Added new function
5671 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5672 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5673 OID is not part of the table.
5676 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5677 X509_LOOKUP_by_alias().
5680 *) Sort openssl functions by name.
5683 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5684 encryption from sample DSA keys (in case anyone is interested the password
5688 *) Make _all_ *_free functions accept a NULL pointer.
5689 [Frans Heymans <fheymans@isaserver.be>]
5691 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5693 [Anonymous <nobody@replay.com>]
5695 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5696 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5698 *) Don't blow it for numeric -newkey arguments to apps/req.
5699 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5701 *) Temp key "for export" tests were wrong in s3_srvr.c.
5702 [Anonymous <nobody@replay.com>]
5704 *) Add prototype for temp key callback functions
5705 SSL_CTX_set_tmp_{rsa,dh}_callback().
5708 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5709 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5712 *) X509_name_add_entry() freed the wrong thing after an error.
5713 [Arne Ansper <arne@ats.cyber.ee>]
5715 *) rsa_eay.c would attempt to free a NULL context.
5716 [Arne Ansper <arne@ats.cyber.ee>]
5718 *) BIO_s_socket() had a broken should_retry() on Windoze.
5719 [Arne Ansper <arne@ats.cyber.ee>]
5721 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5722 [Arne Ansper <arne@ats.cyber.ee>]
5724 *) Make sure the already existing X509_STORE->depth variable is initialized
5725 in X509_STORE_new(), but document the fact that this variable is still
5726 unused in the certificate verification process.
5727 [Ralf S. Engelschall]
5729 *) Fix the various library and apps files to free up pkeys obtained from
5730 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5733 *) Fix reference counting in X509_PUBKEY_get(). This makes
5734 demos/maurice/example2.c work, amongst others, probably.
5735 [Steve Henson and Ben Laurie]
5737 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5738 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5739 are no longer created. This way we have a single and consistent command
5740 line interface `openssl <command>', similar to `cvs <command>'.
5741 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5743 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5744 BIT STRING wrapper always have zero unused bits.
5747 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5750 *) Make the top-level INSTALL documentation easier to understand.
5753 *) Makefiles updated to exit if an error occurs in a sub-directory
5754 make (including if user presses ^C) [Paul Sutton]
5756 *) Make Montgomery context stuff explicit in RSA data structure.
5759 *) Fix build order of pem and err to allow for generated pem.h.
5762 *) Fix renumbering bug in X509_NAME_delete_entry().
5765 *) Enhanced the err-ins.pl script so it makes the error library number
5766 global and can add a library name. This is needed for external ASN1 and
5767 other error libraries.
5770 *) Fixed sk_insert which never worked properly.
5773 *) Fix ASN1 macros so they can handle indefinite length construted
5774 EXPLICIT tags. Some non standard certificates use these: they can now
5778 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5779 into a single doc/ssleay.txt bundle. This way the information is still
5780 preserved but no longer messes up this directory. Now it's new room for
5781 the new set of documenation files.
5782 [Ralf S. Engelschall]
5784 *) SETs were incorrectly DER encoded. This was a major pain, because they
5785 shared code with SEQUENCEs, which aren't coded the same. This means that
5786 almost everything to do with SETs or SEQUENCEs has either changed name or
5787 number of arguments.
5788 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5790 *) Fix test data to work with the above.
5793 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5794 was already fixed by Eric for 0.9.1 it seems.
5795 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5797 *) Autodetect FreeBSD3.
5800 *) Fix various bugs in Configure. This affects the following platforms:
5804 unixware-2.0-pentium
5808 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5809 before they are needed.
5812 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5816 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5818 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5819 changed SSLeay to OpenSSL in version strings.
5820 [Ralf S. Engelschall]
5822 *) Some fixups to the top-level documents.
5825 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5826 because the symlink to include/ was missing.
5827 [Ralf S. Engelschall]
5829 *) Incorporated the popular no-RSA/DSA-only patches
5830 which allow to compile a RSA-free SSLeay.
5831 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5833 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5834 when "ssleay" is still not found.
5835 [Ralf S. Engelschall]
5837 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5838 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5840 *) Updated the README file.
5841 [Ralf S. Engelschall]
5843 *) Added various .cvsignore files in the CVS repository subdirs
5844 to make a "cvs update" really silent.
5845 [Ralf S. Engelschall]
5847 *) Recompiled the error-definition header files and added
5848 missing symbols to the Win32 linker tables.
5849 [Ralf S. Engelschall]
5851 *) Cleaned up the top-level documents;
5852 o new files: CHANGES and LICENSE
5853 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5854 o merged COPYRIGHT into LICENSE
5855 o removed obsolete TODO file
5856 o renamed MICROSOFT to INSTALL.W32
5857 [Ralf S. Engelschall]
5859 *) Removed dummy files from the 0.9.1b source tree:
5860 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5861 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5862 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5863 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5864 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5865 [Ralf S. Engelschall]
5867 *) Added various platform portability fixes.
5870 *) The Genesis of the OpenSSL rpject:
5871 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5872 Young and Tim J. Hudson created while they were working for C2Net until
5874 [The OpenSSL Project]
5877 Changes between 0.9.0b and 0.9.1b [not released]
5879 *) Updated a few CA certificates under certs/
5882 *) Changed some BIGNUM api stuff.
5885 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5886 DGUX x86, Linux Alpha, etc.
5889 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5890 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5894 *) Add -strparse option to asn1pars program which parses nested
5896 [Dr Stephen Henson <shenson@bigfoot.com>]
5898 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5901 *) DSA fix for "ca" program.
5904 *) Added "-genkey" option to "dsaparam" program.
5907 *) Added RIPE MD160 (rmd160) message digest.
5910 *) Added -a (all) option to "ssleay version" command.
5913 *) Added PLATFORM define which is the id given to Configure.
5916 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5919 *) Extended the ASN.1 parser routines.
5922 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5925 *) Added a BN_CTX to the BN library.
5928 *) Fixed the weak key values in DES library
5931 *) Changed API in EVP library for cipher aliases.
5934 *) Added support for RC2/64bit cipher.
5937 *) Converted the lhash library to the crypto/mem.c functions.
5940 *) Added more recognized ASN.1 object ids.
5943 *) Added more RSA padding checks for SSL/TLS.
5946 *) Added BIO proxy/filter functionality.
5949 *) Added extra_certs to SSL_CTX which can be used
5950 send extra CA certificates to the client in the CA cert chain sending
5951 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5954 *) Now Fortezza is denied in the authentication phase because
5955 this is key exchange mechanism is not supported by SSLeay at all.
5958 *) Additional PKCS1 checks.
5961 *) Support the string "TLSv1" for all TLS v1 ciphers.
5964 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5965 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5968 *) Fixed a few memory leaks.
5971 *) Fixed various code and comment typos.
5974 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5975 bytes sent in the client random.
5976 [Edward Bishop <ebishop@spyglass.com>]